ll I B ii I H^HHIH >i BMBHHBBH !'.! lull! llllll in LIBRARY OF THE UNIVERSITY OF CALIFORNIA. LIBRARY VARIATION IN ANIMALS AND PLANTS BY H. M. VERNON, M. A., M. D. Fellow of Magdalen College, Oxford NEW YORK HENRY HOLT AND COMPANY 1903 GENERAL COPYRIGHT, 1902, BY HENRY HOLT & CO T BIO 1 RA 6 PKEFACE. IN this little book I have endeavoured to give a brief account of the subject of " Variation," so far as the present state of our knowledge admits. Though I did not in any degree aim at giving a complete representa- tion of the subject, yet I hope that most of the more important and more recent work has been included. I have not treated Variation of Plants so fully as that of Animals, and from lack of thorough acquaintance with the literature, have probably made some omissions of real importance to the adequate comprehension of the subject in its bearing on living organisms taken as a whole. For such I offer my apologies. I have pur- posely avoided any higher mathematics in discussing the facts of variation, as it seemed out of place in a book of this character. An obvious criticism upon the contents of the book will be that I have given greater prominence to my own researches than their intrinsic importance warrants. To this I frankly plead guilty, urging in extenuation that I did not intend to write a text-book in the ordi- nary sense of the term. Thus some of the hypotheses and interpretations of facts which I have given are my own personal opinions, and by no means current views held in general acceptation. Also a few of the data published in the latter part of Chapter VI., and those on " identical twins " in Chapter IV., are here published m 11270? iv PREFACE. for the first time, and are for this reason given rather in extenso. As regards the illustrations, I am indebted to the Royal Society for the blocks of Figs. 17, 18, 20, and 21, whilst Figs. 3, 6, 9, 16, 19, and 25 have been copied from illustrations in their publications. Figs. 22, 23, and 24 are copied from Davenport's " Experimental Morphology," and Figs. 7 and 8 from Bateson's " Ma- terials for the Study of Variation." The rest are either original, or from sources indicated in the text. I desire to take this opportunity of expressing my ob- ligations to Mr. E. S. Goodrich for his kindness in read- ing over the manuscript, and to Professors W. F. R. Weldon and S. H. Vines, for their useful advice and suggestions. CONTENTS. PART I. THE FACTS OF VARIATION. CHAPTER I. THE MEASUREMENT OF VARIATION. PAGE Variation studied from the mathematical standpoint Vari- ation of birds diagrammatically represented Distribution of crab measurements Normal curve of error Its relation ^ to binomial curve Variations in animals and plants subject to Law of Frequency of Error Measurement of variation in terms of Probable Error, Arithmetic Meant Error, and Error of Mean Square Examples of Asym- / matrical series Representation of these by a generalised mathematical expression, 1 CHAPTER H. DIMORPHISM AND DISCONTINUOUS VARIATION. Dimorphism in the earwig and in the crab How to dis- tinguish between species and varieties, as instanced by dimorphism in certain fishes, and in a marsh plant Poly- morphism in plants Series of Fibonacci Discontinuous variation in animals as regards vertebrae, ribs, mammae, teeth, digits, and other characters Homceosis De Vries' Theory of Mutation Dimorphism may be due to internal causes, or the result of divergent evolution Physiological Selection Infertility between varieties, . , .37 vi CONTENTS. CHAPTER III. CORRELATED VARIATIONS. The measurement of correlation Galton's function Cor- relation between various organs in man, in local races of the shrimp, and in crabs Comparison between primitive and civilised races of man Correlation between morpho- logical characters and the reproductive system Genetic Selection in man Especial fertility of type forms in cer- tain plants Evolution in the Peppered moth Parallel variation Importance of mathematical treatment of varia- tion, 72 PART II. THE CAUSES OF VARIATION. CHAPTER IV. BLASTOGENIC VARIATIONS. The ultimate cause of blastogenic variation Effect of stale- ness and of comparative maturity of sex-cells on the characters of organisms Amphimixis Identical twins Transplantation of ova in the rabbit Law of Ancestral Heredity in man and in the Basset hound Regression towards mediocrity Exclusive inheritance Homotyposis, 101 CHAPTER V. BLASTOGENIC VARIATIONS (Continued). Reversion; commonest in crossed races, as of the pigeon and fowl; its theoretical explanation Prepotency; in the trotting horse and in man; probably due in large part to inbreedingMendel's Law of Hybridisation, and its range Natural and artificial plant hybrids Animal hybrids Sports ; probably of different origin to normal variations Artificial production of monsters Telegony; probably non-existent Parthenogenesis in an Ostracod and in Daphnia Does sexual reproduction induce vari- ability? Relation of variability of individual to variability of race Asexual reproduction in plants Bud-variation, 138 CONTENTS. vii CHAPTER VI. CERTAIN LAWS OF VARIATION. PAGE Effect of environment on growth diminishes rapidly from time of impregnation onwards Reaction of an organism to environment dependent on nature of organism Rapidly diminishing rate of growth in man and in the guinea-pig with progress of development Variability also diminishes with growth Effect on growth once produced, probably never eradicated Increased variability of sparrow and of periwinkle in America Relation between variability and want of adaptation to environment Variability of migra tory and non-migratory birds Does domestication increase 1 Y\ variability? 3 fSKF CHAPTER VH. THE EFFECT OF TEMPERATURE AND OF LIGHT. Variations and modifications Effect of temperature on growth of frog Optimum temperature of growth in plants Effect of temperature on size of sea-urchin larvae, of Lepidoptera, and of Mollusca Seasonal dimorphism in certain Lepidoptera in its relation to temperature Temper- ature differences giving rise also to local races, sports, and phylogenetic forms Critical period of reaction to temperature Effect of Arctic climate on coat of mammals Effect of darkness and of light on growth of plants Effect of sunlight and of diffused light How far does pigmen- tation of animals depend on exposure to light? Cave animals Illumination of under surface of flounder Effect of light and of darkness on Molluscs Variable protective resemblance in the frog, in fish, and in larvae and pupae of certain Lepidoptera, 223 CHAPTER VIII. THE EFFECT OF MOISTURE AND OF SALINITY. Effect of humidity of soil on plant growth Effect of dry and moist surroundings on characters of plants Desert plants and Aquatic plants Effect of moisture on Lepi- viii CONTENTS. PAGE doptera and on Molluscs Characters of maritime plants probably due to saline environment Conversion of A. salina into A. milJiausenii and into Branchipus Effect of increased salinity on characters of the cockle Influence of salinity on rate of growth of Tubularians, and on size of sea-urchin larvse, 260 CHAPTER IX. THE EFFECT OF FOOD AND OF PRODUCTS OF METABOLISM. Effect of artificial manures on growth of crops Effect of nutrition on plant variation Development of bees and of aphides in relation to food Influence of nature of food on wing markings of certain Lepidoptera Dependence of colour of larvae on plant pigments Influence of food on growth of tadpoles Plumage of certain birds altered by abnormal diet Quality of food influences organs of digestion Every organism probably has specific metab- olism, which has especially adverse action on its own growth Products of metabolism may stimulate growth- Effects of small quantities of urea, uric acid, and ammonium salts Influence of volume and of surface area of water on growth of pond snail Influence of surface area on growth of tadpole Effects of increasing quantities of metabolic products on characters of a snail, and of a Crustacean, . 281 CHAPTER X. THE EFFECTS OF CONDITIONS OF LIFE IN GENERAL. Local conditions of life perhaps the cause of local races, but proof of this as a rule impossible American and European trees compared Alpine and Arctic plants Effects of cultivation Local races of oysters and of snails Lepi- doptera in Malay Archipelago Local races of shrimps, of mackerel, and of herring North American birds and mammals Action of climate on goats and on rabbits Effect of domestication on rabbits, pigeons, fowls, and ducks, 310 CONTENTS. ix PART III. VARIATION IN ITS RELATION TO EVOLUTION. CHAPTER XI. THE ACTION OF NATURAL SELECTION ON VARIATIONS. PAGE Proof of Natural Selection in the crab, and in the sparrow Selection in man Evolution of the mouse Inheritance of acquired characters seems to be shown by cumulative effects of conditions of life, as European climate acting on Ameri- can maize; domestication acting on wild turkeys and ducks; changed climate acting on sheep and dogs Environment may act on germ-plasm through specific excretions and secretions Cases of inherited effects of use and disuse, and of epilepsy, accounted for Somatic variations may increase variability, and so afford Natural Selection a better handle to work upon, . . . *' 335 CHAPTER XII. ADAPTIVE VARIATIONS. Adaptability a fundamental property of protoplasm In- stances of adaptive variations in plants Acclimatisation of Protozoa to high temperature, to poisons, to mechanical stimuli, to saline solutions Acclimatisation of fresh-water Mollusca to salt water, and of various marine animals to fresh water Acclimatisation of Mammals to vegetable poisons, and to toxins Sum total of somatic variations always in direction of adaptation Somatic variations of importance in evolution, but they can effect little without Natural Selection Germinal Selection, . .371 VARIATION IN ANIMALS AND PLANTS. PART I. THE FACTS OF VAKIAT1OK CHAPTEK I. THE MEASUEEMENT OF VARIATION. Variation studied from the mathematical standpoint Variation of birds diagraramatically represented Distribution of crab measure- ments Normal curve of error Its relation to binomial curve Variations in animals and plants subject to Law of Frequency of Error Measurement of variation in terms of Probable Error, Arithmetic Mean Error, and Error, of Mean Square Examples of asymmetrical series Representation of these by a generalised mathematical expression. IF a number of individuals of any species be com- pared, it will be found that they all show differences from each other either in size, shape, colour, relation of parts, or other characteristics; in fact, no two of them are exactly alike. Even if offspring be compared with their own parents, similar, though on the whole not such marked, differences will present themselves. These differences constitute what is known as Varia- tion, and it is into the facts of this variation, and its im- 2 THE MEASUREMENT OF VARIATION. portance as the corner stone of the whole fabric of Evo- lution, that we shall briefly inquire in the following pages. In his " Origin of Species," Darwin clearly recog- nised the fundamental importance of the existence of variation, for without it there could evidently be no such thing as evolution. In his " Variation of Animals and Plants," also, he brought together an enormous mass of material concerning the facts of variation, though unfortunately this dealt almost exclusively with organisms in a condition of domestication. Still, there was sufficient evidence even then to show that wild ani- mals and plants are also subject to variation, though Darwin probably did not fully recognise how consider- able and universal this variation is. As to the causes of variation, Darwin did not hazard many conjectures. To do so would have been premature^ and from actual lack of knowledge almost impossible. For many years after the publication of Darwin's work, the additions to our knowledge of the subject of variation were exceed- ingly small. Scientists seemed to rest content with the material he had collected, and to theorise on this alone, rather than to test their theories by a search after fresh facts and data. Within the last decade, however, the importance of the scientific study of variation has be- gun to be more thoroughly recognised, and has resulted in its being attacked with considerable vigour from sev- eral entirely different points of view. Investigations from the mathematical side have shown that many of the apparently disconnected facts of variation can be expressed with ease and lucidity by exact mathematical expressions, and that much material which has hitherto THE MEASUREMENT OF VARIATION. 3 been regarded as quite outside all law was in reality amenable to treatment according to the well-known. Laws of Chance. Again, investigations from the ex- perimental side have suggested much concerning the causes of variations, both genetic and somatic. Still again, a fresh burst of activity in the collection of data regarding the actual facts of variation, more especially in respect of organisms found in a state of nature, has shown us how much in this branch of the subject there remains for us yet to learn. Perhaps the keynote of most of the recent work on variation lies in the recognition of the fact that almost all the problems to be solved must be attacked from a numerical standpoint. It is no longer sufficient to say that such and such a kind of variation is frequently or occasionally found. It is necessary to know the exact amount of the variation, so far as it is measur- able, and the exact proportion of cases in which it occurs. Only by obtaining data of this kind can we hope to ascertain with any certainty the probable degree of importance of any particular vari- ation in the evolution of a species, and whether such evolution is actually taking place at the present day. No apology is therefore needed for the fre- quent introduction of figures into the study of ques- tions of variation. Rather is this necessary if one should attempt to found theories and deduce conclusions from generalised statements and opinions, unsupported by such evidence. To say that any particular organ is very variable means but little, for so much depends upon the personal opinion of the observer as to what constitutes a great and what a slight variation. But 4 THE MEASUREMENT OF VARIATION. supposing it be said that out of a large number of indi- viduals half varied in size by 5 per cent, from the average of the whole, then there is afforded a numerical expression of the degree of variation, which can readily be compared with similar expressions concerning the variability of other parts of the same organism, and with those of quite distinct organisms. Let us first of all, therefore, examine one or two simple series of measurements made on a group of indi- viduals of a species, so as to get some idea of the actual differences exhibited by the varying characters, or, as they have been termed, the variants. Some of the most striking are those obtained by J. A. Allen,* concerning the variation in certain mammals and winter birds of East Florida. Of a species of squirrel (Sciurus caroli- nensis), for instance, 28 individuals were measured, and these . neasurements are reproduced to scale in the accompanying diagram. Here the animals are ar- ranged in order according to the length of their body in inches, and the corresponding values for the head, tail, and forefoot are given on the same ordinates. By means of this diagram, the magnitude of each and all of the measurements made can be read off at a glance. The body was on an average 9.15 inches long, but the extreme values were 8.25 and 10.20 inches, or respect- ively 9.8 per cent, and 11.5 per cent, less and greater than the mean. The tail measurements were even more variable than this, the extremes varying from 6.75 to 8.75 inches, or by respectively 14.3 per cent, and 11 per cent, from the mean. In the forefoot the range of variation was less, and in the head smaller still; but * Bulletin Museum Comp. Zo51,, Harvard, 1871. Body Tail Head Fore foot FIG. 1. Variation of Sciurus carolinensis. 6 THE MEASUREMENT OF VARIATION. there was never any constancy, every animal varying in respect of each of the measurements made. This is a point of fundamental importance, which cannot be too thoroughly grasped. Every organism varies in re- spect of all its characters, whatever be their nature. The amount of this variation differs greatly, as these results well show, but it is always present in a greater or less degree. Another fact which this diagram brings out very clearly is the comparative independence of these measurements. Because the body of one animal is longer than another, it by no means necessarily fol- lows that the head or tail is longer also. A superficial glance at this diagram might, indeed, lead one to sup- pose that the various parts of the body were absolutely independent of each other. But this we know not to be the case. Between most parts and organs there is a greater or less degree of correlation, so that, on an aver- age, animals with a longer body may have a longer head and longer tail than animals with a shorter body. A careful examination of the diagram will show that, on the whole, though with numerous exceptions, the curves for head, tail, and foot do slope very slightly upwards from left to right, though nothing like as much as the curve for body lengths. Some degree of correlation is therefore present, though it is only slight. We know that frequently it may be very great indeed, as for in- stance between the two fore limbs or two hind limbs of a quadruped, and very considerable between a fore and a hind limb; but into this question we must not enter now. Almost innumerable diagrams of a similar nature to the above might be given, but this is scarcely neces- THE MEASUREMENT OF VARIATION. 7 sary. All that they would demonstrate would be the fact that variation of a similar nature though of a varying degree is present in all organisms, to what- ever class of the Animal or Vegetable Kingdom they belong. Should more evidence of this kind be de- sired, the reader is referred to Wallace's book on " Dar- winism " (Chapter III). Here an admirable series of diagrams is given, illustrating the variation in several species of lizards, birds, and mammals. The diagram given above is modelled on the plan adopted by Wal- lace, and still earlier by Galton, as the one best adapted for bringing before the eye the facts of individual variability. In the above diagram the measurements of only 28 different individuals are given, and hence we are not able to gather much as to the distribution of the differ- ent measurements about their means. Supposing that instead of tens, fifties or hundreds of the animals had been measured, what should we expect to find? Would there or would there not be just as many animals with a very long or very short body length, as with a moder- ately long or moderately short one, or as with a nearly average one? Such a question as this is also best an- swered by reproducing the measurements diagram- matically, though in this case they must be arranged on a different system. In the accompanying diagram, Fig. 2, 65 measurements of the wing of Sterna hirundo, recorded in the above-mentioned paper of Allen, are plotted out. Here each dot represents one measure- ment, all the measurements between 10.46 and 10.55 inches being placed over the number 10.5, and so on. The mean of all the measurements is 10.49 inches, and 8 THE MEASUREMENT OF VARIATION. we see in the diagram that the most frequently occur- ring measurement is one of 10.5 inches. Wing lengths smaller or greater than the mean occur less and less fre- quently, in rough proportion to their degree of devia- tion from it, so that finally, beyond the extreme devia- tions of 9.6 and 11.7 inches, no measurements were ob- served at all. The number of observations here plotted out is ob- viously much too small to yield at all a regular series, but it is quite sufficient to show that the 9-5 10-0 10-5 11-0 11-5 FIG. 2. Wing of Sterna Mrundo. measurements are by no means evenly distributed through the whole range of their variation. There is a most conspicuous collection of them, or heaping up, in the region of the mean measurement. Supposing the number of observations were increased, then one would expect as a general rule to get a more and more even series; in fact, to get a fairly accurate idea as to the kind of series obtainable, supposing an in- finite number of observations were made. In Fig. 3 is plotted out a curve representing the distribution of 1923 measurements made by Warren * on a certain dimension, viz., the carapace breadth of the crab Por- *Proc. Roy. Soc., vol. Ix. p. 225. I CO I I a II a" I C nte oic Ve tica 10 THE MEASUREMENT OF VARIATION. tunus depurator. In order to get rid as far as possible of the factor of size, and obtain a measure of the varia- bility apart from this, each measurement was calculated as a fraction on that of the carapace length of the crab taken as 1000. The numbers on the abscissa line therefore represent 1230, 1240, etc., thousandths of the total length. The figures on the central ordinate represent the numbers of individuals of each particular dimension. For instance, one may gather that 16 indi- viduals had a post-spinous length of 1260, 172 of them one of 12 97, and so on. If this curve be compared with the general contour of the previous figure, it will be seen at a glance that there is a much more regular rise and fall, especially in regard to the extreme measurements. In fact, it does not differ very greatly from the dotted line curve upon which it is superposed, and supposing the number of observations had been greater, one would expect the approximation to be still closer; supposing it had been infinitely great, one would expect the two curves to be identical. Now this dotted line is a probability curve, or a diagrammatic representation of the Law of Fre- quency of Error, of which the mathematical expression* was first deduced by Gauss at the beginning of the last century. It would be out of place to attempt to reproduce its mathematical proof here, but perhaps a concrete instance may help to bring home to the non- mathematical reader the fact that variability does obey * This expression is y = ke -W a , or taking k and h each as unity, y = s , where e is the base of Naperian logarithms, and y an ordi- 6 " nate erected from any point on the abscissa, distant x from the middle ordinate. THE MEASUREMENT OF VARIATION. 11 the laws of chance. Supposing a group of developing organisms be taken, of which the growth can be affected in a favourable or an unfavourable manner by their surroundings. Let us suppose that there are twenty different agencies, each of which would produce an equal, favourable effect on growth, and twenty which would produce just as great an effect in the opposite direction. Suppose also that each organism is sub- jected to only half of these forty different agencies; then it would follow, according to the laws of chance, that a larger number of the organisms would be acted upon by 10 favourable and 10 unfavourable agencies, than by any other combination; i. e., they would, on our hypothesis, remain absolutely unaffected in their growth. A somewhat smaller number would be acted upon by 11 favourable and 9 unfavourable agencies, or on the whole, would have their growth slightly in- creased. A still smaller proportion would be acted on by 12 favourable and 8 unfavourable agencies, or would have their growth rather more increased. Finally the number of organisms acted on by 20 favourable and unfavourable agencies would be extraordinarily small, but in this case the effect on growth would be extremely large. Similar relationships, only in the reverse direc- tion, would of course be found in those cases in which the number of unfavourable agencies exceeded the number of favourable. If desired, the proportional numbers of organisms acted on by all the different com- binations of agencies may be readily determined by ex- panding the binomial (-J + J) 20 . It is found, for in- stance, that for each single time the organisms are acted on by the whole 20 favourable agencies, they are 12 THE MEASUREMENT OF VARIATION. acted on 190 times by 18 favourable and 2 unfavour- able, 15,504 times by 15 favourable and 5 unfavour- able, and no less than 184,756 times by 10 favourable and 10 unfavourable. Let us consider that the organ- isms acted on by 20 favourable and unfavourable agencies have their size increased by 20 per cent., those acted on by 15 favourable and 5 unfavourable by 15 5 = 10 per cent., and so on. If now these percentage increments and decrements be plotted out at equal dis- tances on a base line, and ordinates corresponding to the theoretical frequencies erected from each, then by joining these ordinates we shall obtain a curve which is practically identical in form with the dotted line curve given in Fig. 3; i. e., with the probability curve of the law of frequency of error. Thus, by a simple arith- metical method, we can obtain a series approximating more and more closely to the probability curve, the greater the number of times the expression (J + J) is expanded. Expanded 20 times, the average error is less than .5 per cent., and for a greater number of times it becomes rapidly smaller and smaller. The deviations in the dimensions of organisms are thus distributed about their mean in a symmetrical manner, in accordance with the law of frequency of error. This is true not of one or two characteristics of an organism, but probably, in the majority of cases, of nearly all of them. The dependence of variation on the Laws of Probability was first demonstrated by Que- telet * in the case of height and chest measurements of soldiers. These he showed to group themselves in ac- cordance with the ordinates of a binomial curve. * " Lettres sur la theorie des probabilites," Brussels, 1846. THE MEASUREMENT OF VARIATION. 13 Subsequently * he proved that a similar relationship was true not only for the height, weight, strength, lon- gevity, and other physical qualities of man, but also for his intellectual and moral qualities, such as age at mar- riage, age of criminals, and so on. He considered also that these laws extend to the whole Animal and Vege- table Kingdoms, though he did not give proofs of this hypothesis. In confirmation and extension of Quetelet's results, the observations of Mr. Francis Galton f may be quoted. These were made at the Anthropometric Laboratory of the International Health Exhibition of 1884, upon from 489 to 1788 men and women. It was found that the variations in height, span of arms, weight, breathing capacity, strength of pull, strength of squeeze, swiftness of blow, and keenness of sight all con- formed in their distribution to the Law of Error. With regard to the lower animals, Professor Weldon has made measurements on the carapace, post-spinous por- tion of carapace, length of the sixth abdominal tergum, and length of telson, in the case of two to five local races of shrimps, and obtained a similar result. He has also made no less than eleven different series of measurements on 999 female crabs (Carcinus mcenas) obtained from Plymouth Sound, and a similar number on 999 specimens obtained from the Bay of Naples. Twenty series of frequencies of deviation from the average were thereby obtained, and were found in every * " Anthropometrie," p. 257, 1870. f " Natural Inheritance," p. 201. t Proc. Roy. Soc., xlvii. p. 445, 1889, and Proc. Roy. Soc., li. p. 2, 1892. Proc. Roy. Soc., liv. p. 318, 1893. 14 THE MEASUREMENT OF VARIATION. case but one to conform to that required by the Law of Error. The single exception to the general rule will be referred to in the next chapter. Again, H. Thomp- son * made twenty-two different measurements on 1000 adult prawns, and found the variations in every case but one to correspond more or less accurately with the law. E. Warren f made seven different measurements on 2300 male crabs (Portunus depuraior), obtained from Plymouth, and found that the variations very nearly corresponded to the law. Duncker $ made eight series of determinations on the number of spines and rays in the fins of 1900 specimens of the fish Acerina cernua, and found that with one slight exception the variations obeyed the general law. Of the twelve series of meas- urements made on 1120 specimens of the flounder (Pleuronectes flesus), however, only six were quite sym- metrical and in accordance with the law. Finally the author || made 9850 measurements on the plutei or larvsB of a sea-urchin, Strongylocentrotus lividus, and found that the variations in size corresponded very closely indeed with the law. The lengths of the arms of these plutei were calculated as percentages on the length of body, and were found in the case of the oral arm lengths to correspond closely with theory, but in the case of the anal arm lengths, there was some slight divergence. With regard to the variation of plants, our accurate *Proc. Roy. Soc., Iv. p. 234, 1894. fProc. Roy. Soc., Ix. p. 221. t Biologischen Centralblatt, xvii. p. 785, 1897. gZool. Anzeig., xxiii. p. 141. flPhil. Trans. 1895, B. p. 613. THE MEASUREMENT OF VARIATION. 15 knowledge is derived chiefly from the work of Ludwig, De Vries, and Vb'chting. The majority of variations hitherto examined have not been found to be at all ac- curately in accordance with the law of frequency of error, for reasons which will be referred to later. However, in the case of one or two local races of Torilis anthriscus (hedge parsley) examined by Ludwig*, the distribution of the frequencies of the numbers of branches in the main umbels more or less conforms, and the same is true for the numbers of ray florets in a pure race of Chrysanthemum segetum (corn marigold) examined by De Vries. f Again H. VochtingJ has recently examined the anomalies occurring in 61,736 flowers of Linaria spuria (toadflax), obtained in differ- ent years and from different sources. He determined the proportions of the various forms of peloric flowers and anomalous zygomorphic forms, of flowers of vary- ing structure and with various numbers of spurs, and came to the conclusion that their distribution followed the law of error. For instance, the numbers of flowers in each inflorescence showed the following variations : Number of flowers, 234 5 678 Frequency, 1 6 283 61,060 221 9 1 Percent., .0016 .0097 .459 99.153 .358 .014 .0016 Here we see that though more than 99 per cent, of the flowers exhibited the normal pentamerous form, yet the variations from this normal are very evenly distributed on either side of it. The distribution of the numbers in all the peloric flowers (i. e., regular *Bot. Centralb., vol. Ixiv. p. 40. f Arch. f. Entwickelungsmechanik, ii. p. 52, 1896. \ Jahrb. f. wiss. Bot., Bd. xxxi. p. 391, 1898. 16 THE MEASUREMENT OF VARIATION. flowers, instead of the normal irregular ones) was as follows : Number of flowers, 234 5 Frequency, 1 2 43 810 Percent., .109 .219 4.720 88.913 678 52 2 1 5.708 .219 .109 From these two series a very interesting relationship declares itself, which may for convenience be referred to here, though it properly comes under the heading of "correlated variations." Thus, as the following fig- ures show, we find that the probability of occurrence of a peloric flower increases according to the amount of deviation of the number of flowers on a stalk from the normal pentamerous form, or that the less often a particular number of flowers occurs, the more fre- quently does it produce peloric flowers : per cent, of the 5 flower form have peloric flowers. .132 15.19 23.53 22.22 33.33 100.00 5 4 6 7 3 2 and 8 Of English observers, J. H. Pledge * has determined the variations in the numbers of petals, stamens, and carpels in 1000 specimens of Ranunculus repens (creep- ing crowfoot), the distributions of all but the numbers of petals agreeing fairly closely with the probability integral. For instance, the numbers of sepals varied thus: Sepals, 34 567 Frequency, 1 20 959 18 2 We see, therefore, that in the majority of the characteristics of the various organisms investigated, *Nat. Science, vol. x. p. 323, and vol. xii. p. 179, 1898. THE MEASUREMENT OF VARIATION. 17 especially those belonging to the Animal Kingdom, the variations are distributed about their mean in accordance with the Law of Error. It is scarcely necessary to point out, however, that the actual range of the variations is exceedingly variable, and that the general contour of the curves, supposing the results are expressed in that way, must be equally variable. The greater the variability of any char- acteristic, the more spread out, or flattened, must be the curve representing the frequencies of its devia- tions. If, therefore, results were invariably expressed in the form of curves, and if, by multiplying each series of measurements by some factor, the central ordinate were always brought to the same height, then it would follow that the variability of each char- acteristic would be accurately represented by the ex- tent of spread of the curve. In order to obtain an in- dex of the variability of any characteristic, we must accordingly adopt some convenient method of deter- mining the degree of spread of its curve. One of the simplest of these methods, and one widely employed by English statisticians, is to determine the so-called Probable Error. The meaning of this term is best explained by reference to the accompanying diagram of a curve of frequency of error. The ordinate drawn through the middle of the curve is spoken of by Mr. Galton as the Median, and is denoted by the symbol M. In symmetrical curves it is identical with the ordinary arithmetic mean or average, and in this sense is called the Centroid Vertical. It is the middle value of the whole series of observations, which are symmetrically distributed on each side of it. That is to say, 50 per 18 THE MEASUEEMENT OF VARIATION. cent, of all the observations fall below it in magnitude, and 50 per cent, above it. The actual number of ob- servations made is obviously represented by the area of the figure enclosed by the curve and the abscissa line, or the so-called " polygon of variation." The area to the left of the median corresponds to the half of the observations of less magnitude than the average, and SAQ, M Q,A'S' FIG. 4. Normal Curve of Error. that to the right, of those of greater magnitude. Now let two other ordinates, Q l and Q 3 , be erected so as to divide each of these areas into equal halves. We now have four areas representing four numerically equal groups; i. e., all the observations of small magni- tude from to 25 per cent, of the whole; those of greater magnitude, from 25 per cent, to 50 per cent, of the whole ; those of greater magnitude than the average, representing 50 per cent, to 75 per cent., and finally those of greatest magnitude, representing the remain- ing 25 per cent. Half of all the observations there- fore exceed the limits of these ordinates Qi and Q 3 , and half of them fall between or within them; so the dis- tance on the abscissa line from M to Q l or M to Q 3 , is THE MEASUREMENT OF VARIATION. 19 called the " Probable Error " of variation. In a per- fectly normal curve, these values are equal in value and opposite in sign, but as no experimental result is perfect, they usually differ slightly in amount. A mean between the two is therefore taken, and this is denoted by the symbol Q. For the practical determination of the probable error, however, it is quite unnecessary to plot out the results in the form of a curve. The method adopted is best illustrated by a concrete instance. In the ac- companying table are given the results obtained by Mr. Galton * for the strength of pull, as of an archer with a bow, of 519 males, aged 23 to 26 : PERCENTAGES. STRENGTH OP PULL. OBSERVED. NUMBER OP CASES SUMS FROM BE- OBSERVED. GINNING. Under 50 Ibs. 10 2 2 60 42 8 10 70 140 27 37 80 168 33 70 90 113 21 91 100 22 4 95 Above 100 24 5 100 Here we see that the numbers of actual cases in each group are given in the second column, and that they are calculated as percentages in the third column. They are summed from the beginning in the fourth column, and we thereby gather that whilst only 37 per cent, of all the men had a strength of pull under 70 Ibs., 70 per cent, of them had one under 80 Ibs. It can be calcu- *" Natural Inheritance," p. 199. 20 THE MEASUREMENT OF VARIATION. lated, therefore, that 50 per cent, of them had a strength of pull under 74 Ibs., or, in Mr. Galton's nota- tion, the strength of pull at Grade 50, was under 74 Ibs. This, then, is the average strength of pull, or M , of the whole group. Fifty per cent, of the men pulled less than this amount, and 50 per cent, of them more. Similarly, also, one can calculate that 25 per cent, of the men would have a pull of less than 66 Ibs., and 75 per cent, one of greater amount, whilst 75 per cent, would have one of less than 82 Ibs., and 25 per cent, one of greater. That is to say, the strengths of pull at Grades 25 and 75 were respectively 66 and 82 Ibs. The prob- able error of variation in pull, or Q J? is therefore equal to 74 66 = 8 Ibs., and also to Q 3 , or 82 74 = 8 Ibs., whilst the mean value which is always in practice 8-4-8 adopted as the probable error, or Q, is ~ = 8 Ibs. This probable error is 10.8 per cent, on the magnitude of the average strength of pull, and this value accu- rately represents the variability of this group of men in respect of this particular characteristic. Supposing an- other group were found to have a probable error of only 5.4 per cent, on the magnitude of the average, then one would be justified in saying that th^ir variability, or range of variation, was only half as great; or if it had been 21.6 per cent., then twice as great. This relative probable error is therefore a con- venient index of variability of any characteristic. A few examples may be quoted in order to give an idea as to its range. From the anthropometric data obtained by Mr. Galton, it is calculated that the index was 2.50 per cent, for man's stature, and 2.52 per cent. THE MEASUREMENT OF VARIATION. 21 for woman's; 2.92 per cent, for the span of arms of both man and woman; but no less than 6.89 per cent, for man's weight, and 8.89 per cent, for woman's. That is to say, weight is more than twice as variable as the other two characteristics. In most of Weldon's shrimp and crab measurements the amount of variability was considerably smaller, but this was partly due to the fact that the element of size was largely excluded by first of all calculating all measurements as thousandths of the body and carapace lengths respectively. In 1000 shrimps from Plymouth, the total carapace length had a relative probable error of only 1.82 per cent., the post-spinous carapace length one of 1.97 per cent., the sixth abdominal tergum one of 1.93 per cent., and the telson one of 2.36 per cent. In 999 crabs obtained from Naples, the value was only 1.07 per cent, for the total breadth of carapace, and from 1 to 2 per cent, for several of the other measurements made, but in the carpopodite of the right chela it rose to 3.63 per cent., and in the proximal portion of the chela to no less than 5.77 per cent. These last were, however, quite ex- ceptionally large degrees of variation. Still, in the sea- urchin larvae measured by the author, the variability was found to be greater than in any of these instances recorded in the higher animals, it being 6.1 per cent, for the body length, 9.4 per cent, for the oral arm length, and 11.3 per cent, for the anal arm length. We have seen that the degree of correspondence of the variations in any characteristic with the law of error can be determined by plotting out the results in the form of a curve, but this is clearly a somewhat laborious process. It is much simpler and more con- 22 THE MEASUREMENT OF VARIATION. venient to compare the experimental and theoretical values directly by a numerical method. This is done by extending the method of grades referred to above. In addition to determining the magnitude of the char- acteristics at grades 25, 50, and 75, one determines it also at grades 5, 10, 20, 30, and so on, or determines the values having respectively 5, 10, 20, 30 per cent., etc., of all the measurements below them in magnitude, and 95, 90, 80, 70 per cent., etc., above them in magni- tude. Let the median, or value at grade 50, be now subtracted from the values at all the other grades, and the numbers so obtained be divided by the probable error, or 1 = ! . We then obtain a series of values at the various grades, in terms of the probable error taken as unity; so that, whatever had been the magnitude of the median, and of the probable error, the values are now directly comparable with the theoretical values cal- culated from the probability integral. These theoretical values are given in the first line of the subjoined table : GRADE. 5 10 20 25 30 40 50 60 70 75 80 90 95 THEORETICAL VALUES 2.44 1.90 1.25 1.00 .78 .38 .00 .38 .78 1.00 1.25 1.90 2.44 9443 anthropometric measurements. 2.44 1.87 1.84 1.00 .77 .40 ,00 .38 .75 .98 1.21 1.92 2.47 400 shrimp carapace length measurements 9850 measurements of 2.42 1.86 1.22 1.00 .79 .39 .00 .32 .71 1.00 1.28 2.10 2.63 sea-urchin larvae, . 2.51 1.92 1.25 1.01 .79 .38 .00 .37 .77 .99 1.24 1.90 2.46 Beneath them are given the means of the values ob- tained by Mr. Galton for 18 different series of measure- ments on men and women, the total number of observa- tions made being 9443. In the individual series the THE MEASUREMENT OF VARIATION. 23 deviations from the theoretical values were of course greater, but these differences almost completely neu> tralise each other in the general mean. Indeed the correspondence is extraordinarily close ? considering the very mixed nature of the faculties measured, viz., three linear measurements, one of weight, one of capacity, two of strength, one of vision, and one of swiftness. The next series of values is that obtained by Professor "Weldon for 400 shrimps. It is given to show that in the case of a comparatively small number of observa- tions, the correspondence between fact and theory may be very close indeed. Finally, in the bottom line of the table are given the values obtained by the author for 9850 measurements on the body length of sea-urchin larvae. Here the correspondence is closer even than in the anthropometric measurements, the average differ- ence being only .014, as against .0175. In order to express the variability of a characteristic, we are by no means limited to the method of determin- ing the probable error. A much older method is that of the arithmetic mean error, or average deviation. This value consists of the mean of all the deviations, both positive and negative, from the general mean. For in- stance, to determine the arithmetic mean error of the following series of 16 7, 8, 8, 9, 9, 9, 10, 10, 10, 10, 11, 11, 11, 12, 12, 13. figures, one calculates the general mean, viz., 10, and determines their deviations from it. These are 3, 2, 2, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 2, 2, 3. Added together these equal 20, so that the arithmetic 20 mean error is = 1.25. In practice, it is sometimes 24 THE MEASUREMENT OF VARIATION. simpler to separate all the numbers into two groups, one containing all the values greater than the general mean, and the other all those less than the mean. Then the arithmetic mean error is half the difference between the mean of each group. When, as in the present instance, several of the values are identical with the mean, half of them must be put in each group. The mean of one 70 group is now = 8.75, and of the other group 90 g- = 11.25, and the arithmetic mean error is 11.25-8.75 2 This method of estimating variability has frequently been employed in recent times, especially in America. Thus Minot * used it for comparing the variability of guinea-pigs at various periods of their growth. Brew- ster f used it for calculating the amount of varia- tion in a number of head, face, and limb measurements which were made by Weisbach J on individuals of 23 different races of men. In the general mean are in- cluded the measurements of 195 individuals, represent- ing 20 different races. The following are some of the mean values for the arithmetical mean error, calculated as percentages on the mean size : Nose length, 9.49 per cent. Head length, 2.44 per cent. " breadth, 7.57 * " breadth, 2.78 " height, 15.2 ' Upper arm length, 6.50 Forehead height, 10.4 ' Forearm length, 3.85 Under jaw length, 4.81 ' Upper leg length, 5.00 Mouth breadth, 5.18 * Lower " 5.04 Foot length, 5.92 * J. Physiol., xii. p. 138, 1891. fProc. Amer. Acad. Arts Sci., xxxii. p. 268, 1897. j Zeitschrift f. Ethnologic, Bd. ix. Supplement, 1878. THE MEASUREMENT OF VARIATION. 25 Here we see that most of the face measurements are far more variable than most of the head and limb measurements; that of the nose height, for instance, being six times as great as that of the head length. The high value which is universally accorded to facial proportions as a means of personal identification thus receives its numerical justification. On comparing the variability of the measurements in the individuals of eight different races, it was found to be more or less the same in each case. If the nose of a Jew is a very variable organ, so is that of a Slav, a Magyar, or a Chinaman. Davenport and Bullard * used the method of arith- metic mean error in the 4000 enumerations which they made of the Mullerian glands in the forelegs of swine. These glands vary in number from to 10, the average being 3.53. The arithmetic mean error was 1.41 in male swine, and 1.38 in female swine, or the variability was 2.5 per cent, greater in the one case than the other. Again Garstang f has used it to estimate the variability of various local races of the mackerel. There is still another method of estimating varia- bility, which is more accurate than either of the two mentioned, but which until recently has not been used so frequently as they were, because of the labour of ap- plying it. This is the method of Mean or Least Squares. One determines the deviations from the average in the same way as for the arithmetic mean error, but then squares each of them, takes the sum of these squares, *Proc. Amer. Acad. Arts Sci., xxxii. p. 87, 1896. f Jour. Marine Biol. Asso., vol. v. p. 235, 1898. 26 THE MEASUREMENT OP VARIATION. divides by the number of observations, and takes the square root of the quotient. Thus: (or ff) = Where n = number of observations, and v = a devia- tion from the average. For instance, to determine the variability of the following series, representing the frequencies of the numbers of veins in 26 leaves col- lected from different parts of a beech tree,* we find the Number of veins, 15 16 17 18 19 20 Frequencies, 147941 mean (IT. 5), determine the deviations from it in each direction, and square them. Then the variability will be represented by the square root of the following ex- pression: (2.5)' X 1 + (1.5) a X 4+(.5) X 7 + (.5)' X 9 + (1.5)'X 4 +(2.5)*X 1 i. e., by 1.15. This index of variability, or " Error of Mean Square," is termed by Professor Pearson the " Standard Devia- tion," or #, and its percentage ratio on the mean the " Coefficient of Variation." It has been made use of by Warren in the crab measurements already referred to, and also in a very elaborate research f on the variability of the skeleton of the Naquada race, a people that existed in Egypt about 3500 B. c. It has also been employed by Weldon, whilst Pearson al- most invariably adopts it. Duncker $ has expressed his * Vide K. Pearson's " Grammar of Science," p. 382, fPhil. Trans. 1898, B. p. 135. \ Biol. Centralblatt, xvii. p. 785, 1897. THE MEASUREMENT OF VARIATION. 27 results on the variability of the fin rays of certain fishes in terms both of the mean error and the error of mean square. Again, from data obtained by Petersen, Bum- pus, Weldon, and himself, Duncker * has calculated the error of mean square, and obtained the following values for the number of fin rays in certain fishes : DORSAL FIN ANAL TIN Me Me Pleuronectes flesus, Baltic, 39.46 1.4838 North Sea, 41.56 1.7739 " Plymouth, 61.72 2.3895 43.61 1.6026 americanus, 65.06 2.4467 48.62 1.8188 Wtombus maximus, 62.98 2.2533 45.86 1.6792 And the following for the number of rostral teeth: DORSAL FIN M. e Palcsmonetes varians, 4.3137 .8627 1.6948 .4799 iris, 8.2819 .8145 2.9781 .4477 These results show that though the average values of a character may differ considerably even in the local races of the same species, yet the indices of variability may remain fairly constant, not only in these, but also in different species, and perhaps even in different genera and families. Thus the two species of Palcz- monetes vary by respectively 92.0 per cent, and 75.7 per cent, in the number of rostral teeth in their dorsal and anal fins, but by only 5.9 per cent, and 7.2 per cent, in their indices of variability. Arguing from these data, Duncker f concludes that one has no right to ac- cept the " coefficient of variation " of an organ as the *Nat. Science, xv. p. 328, 1899. f Amer. Nat., xxxiv. p. 621, 1900. 28 THE MEASUREMENT OF VARIATION. absolute measure of its variability, as Verschaeffelt,* Brewster,f and others have done. He thinks that the indices of variability alone may be of morphological significance, for in this case, at least, they are obviously independent of the mean values of the characters. How far Duncker's view is to be accepted can only be determined by further enquiry. Doubtless it will be found to hold good occasionally, but I think that the great weight of evidence at present available, especially as regards measurements of size and shape, rather than those of numbers of organs, is in favour of the alterna- tive hypothesis. The three indices of variability above referred to are by no means numerically equivalent. They bear the following relations to each other: Probable error, 1.000 Corresponding grades, 25. 0, 75.0 Arithmetic mean error, 1.183 " " 21. 2, 78. 8 Error of mean square, 1.483 " " 16.0, 84.0 Thus the error of mean square is nearly half as large again as the probable error, and therefore includes a proportionately larger percentage of the deviations from the mean within its limits. On the frequency curve given a few pages back are drawn dotted line ordinates A, A' and S, S', which enclose areas of the variation polygon corresponding to these " mean error " and " error of mean square " indices of variability. The " probable error " index is in some ways the most convenient of the three, as it is the smallest, and in- cludes within its limits just half of all the variants. As the error of mean square is held to be a more accurate *Ber. d. deutsch. bot. Ges., xii. p. 350. f Loc. cit. THE MEASUREMENT OF VARIATION. 29 method of estimating variability, however, the plan is sometimes adopted of determining this index, and then reducing it to terms of probable error by multiplying by .6745. Similarly an arithmetic mean error may be reduced to terms of probable error by multiplying by .8453. It will have been noticed that in the series of meas- urements from time to time referred to, a few excep- tions to the general law of distribution of variations were mentioned. In these cases the variations were not distributed evenly about the middle ordinate, but the curve of distribution was asymmetrical, or skew. Such series as these are by no means uncommon, espe- cially in the case of plant statistics. For instance, De Vries * found that the number of petals in the butter- cup varied between 5 and 10, the frequency of distribu- tion being as follows : Number of petals, 5 6 7 8 9 10 11 Frequency observed, 133 55 23 7 2 20 Theory, 136.9 48.5 22.6 9.6 3.4 .8 .2 Here flowers with the smallest number of petals occur the most, and those with the largest number the least, frequently. The values marked " Theory " in this and the next series will be referred to later. Again De Yries cultivated a variety of clover in which the axis is very frequently prolonged beyond the head of the flower, and bears from one to ten blossoms. The following were the frequencies of occurrence of flowers with none of these blossoms, or with various numbers of them : * Ber. d. deutschen bot. Gesellschaft, xii. p. 203, 1894. 30 THE MEASUREMENT OF VARIATION. High blossoms, 01 23456789 10 Frequency obsd. 325 83 66 51 36 36 18 7611 Theory, 303.2 106.1 70.0 49.3 35.2 24.9 17.1 11.0 6.3 2.8 .5 J. H. Pledge * observed the following frequencies in the numbers of petals in Ranunculus repens : Number of petals, 4 5 6 7 8 9 10 11 12 13 Frequency, 8 706 145 72 38 15 7 7 1 1 Again E. T. Browne f found the following variations in the number of tentaculocysts in the ephyra and adult forms of the medusa Aurelia aurita : Tentaculocysts, 45678 9 10 11 12 13 14 15 Percentage in 1136 ephyrse, .09 .5 3.0 79.1 6.7 5.4 3.1 1.4 .2 .09 Percentage in 3000 adult Aurelia, .1 .1.74.178.9 6.3 4.8 3.0 1.4 .4 .1 .1 In each case the normal eight tentaculocyst form com- prised nearly four-fifths of the whole, but the great majority of the abnormal forms possessed more than eight tentaculocysts, only 3.6 to 5.0 per cent, of them having less. Now the distribution of frequencies in these and somewhat similar asymmetric series obviously occurs ac- cording to some orderly plan, but can a mathematical expression be obtained to represent them? This had been found impossible till within the last few years, when Professor Pearson $ took up the subject, and showed that such series, if composed of homogeneous material, could often be fitted most exactly with curves calculated in accordance with a single generalised *Nat. Science, vol. xii. p. 179, 1898. fQ. J. Microsc. Sci., vol. xxxvii. p. 245, 1895, and Biometrika, I. p. 90, 1901. JPhil. Trans. 1895, A. p. 343, THE MEASUREMENT OF VARIATION. 31 mathematical expression. We saw a few pages back that an expansion of the binomial (^ + ^) for 20 or more times gave a series of values which differed very slightly in their frequencies from that required by the Law of Error. Supposing, now, a binomial in which the two terms are unequal is expanded, then obviously an asymmetrical series of values is obtained. For in- stance, instead of (-J + J) let (f + i) or ( +, ) be expanded, and series are obtained of which the diagram- matic representations are given in the two curves to the left of the accompanying figure. The symmetrical curve represents the expansion of (i + i), the expres- sion being in each case expanded ten times. The areas enclosed between each of these curves and the base line, or the so-called polygons of variation, are obviously of exactly equal extent, in that the sum of the two terms expanded is in each case equal to unity. It follows, therefore, that these asymmetrical series can be represented by the expansion of the expression (P + (?)"* Supposing that n is infinitely large, then curves representing the expansion would stretch out to an unlimited extent in each direction, and though con- stantly approaching nearer and nearer to the abscissa, would never touch it. Supposing n is some finite num- ber, as 20 or 40, then obviously the series is finite also, and its curve is limited in extent. If the two terms of the binomial are unequal, then the curve approaches the * The algebraical expansion of this expression is: (p + q) n = -{ngn - 32 THE MEASUREMENT OF VARIATION. abscissa much more rapidly on one side than on the other, and so, for practical purposes, by taking various values for p y q, and n, we can represent series of the 1 284 ff 6 7 8 9 10 11 FIG. 5. Types of binomial curves. (After Duncker.) 12 following five types by means of the above generalised expression: I. Asymmetrical curves limited on both sides. II. Symmetrical III. Asymmetrical " " " one side, unlimited on the other. IV. Asymmetrical curves, unlimited on both sides. V. Symmetrical The normal curve of error belongs to this last type. Pearson has also pointed out that the abnormal fre- quency curves which cannot be represented by a point- THE MEASUREMENT OF VARIATION. 33 binomial may be the resultant of two or more normal curves, which differ in the position of their axes, or their areas, or their degree of spread, or in all three of these respects.* To return to the curves in Fig. 5, we see that the centroid vertical of the symmetrical curve corresponds to the summit of the curve, or is identical with the maximum ordinate or mode, as it is sometimes called. In the asymmetrical curves, however, this is not the case, but the more asymmetrical the curve, the greater is the distance between the two. The ratio between this distance and the index of variability adopted (such as the error of mean square), gives a convenient " index of asymmetry " of the curve. It is to be noted also that in asymmetrical curves the median, or middle value of the whole series, such that 50 per cent, of the values are below it in magnitude and 50 per cent, above it, no longer coincides with the arithmetic mean. It lies somewhere between the centroid vertical and the maximum ordinate. As to the practical application of this method of fit- ting series of variation frequencies with curves, Pro- fessor Pearson gives numerous instances in the above cited memoir. Fig. 6 will serve to afford some idea as to the types of frequency curves actually met with in practical statistics. Type <* represents the above-men- tioned series of frequencies which De Yries obtained for the petals of buttercups, and high blossoms of clover. It also represents infantile mortality statistics. Type ft represents the relation of scarlet fever and diphtheria mortality to age; type y that of scarlet fever and *Proc. Roy. Soc., liv. p. 329. THE MEASUREMENT OF VARIATION. typhus fever cases to age; type d that of typhoid fever cases to age, and also senile mortality statistics. Fi- nally, type s represents various slight degrees of skew- ness which are frequently found to occur even in anthro- a FIG. 6. Types of Skew curves. pometric and other series which had previously been thought to be quite symmetrical. Most of the series of deviation frequencies obtained by Warren * for vari- ous crab measurements were found by him to be better fitted by skew curves than by absolutely symmetrical ones. Again, of the twelve series of measurements made by Duncker f on the Flounder (Pleuronectes flesus), the six which showed regular variations (number of rays in dorsal, anal, and pectoral fins) were found to give very *Proc. Roy. Soc., Ix. p. 221. f Wissenschaf tliche Meeresuntersuchungen aus der biologischeu Anstalt auf Helgoland, Bd. iii. p. 339, 1900. THE MEASUREMENT OF VARIATION. 35 slightly asymmetrical curves of variation, the varia- bility of the bilateral homologous measurements being always, with one exception, slightly higher on the blind than on the eye side of the fish. Still the departure of some of the curves from the normal Gaussian curve was only very slight. The degree of difference between the actual frequencies obtained in any series of measure- ments and the theoretical frequencies calculated from the type of curve found to show the closest agreement with them, is best represented by determining the per- centage difference of each actual frequency from each theoretical frequency, and then taking an (arithmetical) average of the whole. This average percentage differ- ence of theoretical and actual values may be repre- sented by the sign A . In the case of four of the above series of measurements, the A was only 2.20 per cent, when the measurements were compared with a normal Gaussian curve, and 1.82 per cent, when compared with a slightly asymmetrical curve (Pearson's Type IV). The " fit " was therefore better with the asymmetrical curve, but only very slightly so. It is open to question, therefore, whether any practical object is gained by estimating exceedingly slight degrees of asymmetry in series of measurements. The labour of so doing is very considerable, and it may well be doubted whether it would not be more profitably employed in making more extended series of observations, and subjecting them to less rigid examination. Some recent observations of Miss Hefferan * are instructive in this connection. These were made upon the frequency of distribution of the numbers of teeth on the jaw of an annelid, Nereis * Biol. Bulletin, vol. ii. p. 129, 1900. 36 THE MEASUREMENT OF VARIATION. limbata. Four hundred individuals were examined, and it was found that, as regards the distribution of the total number of teeth, the left total fell into a curve of Pearson's Type I, whilst the right total was of Type IV. However, by dropping out a single individual from the series, it was found that the curve was thrown from Type IV to Type I. As Miss Hefferan points out, this raises a serious question as to the biological importance of the distinction between Pearson's Type I and Type IV.* * Should any further information regarding these asymmetrical curves be desired, the reader should consult Professor Pearson's memoir on the subject, or, if he is not a mathematician, then a recently published book by Davenport on " Statistical Methods,"* and also a paper by Duncker on "Die Methode der Variations- statistik " f may be referred to. Both of these are said to be written specially for biologists. I must mention my special indebtedness to Duncker's paper, which has been drawn upon freely in writing the last few pages of the present chapter. * New York and London, 1899. fArch. f. Entwickelungsmechanik, Bd, viii. p, 112, CHAPTER II. DIMOEPHISM AND DISCONTINUOUS VAKIATION. Dimorphism in the earwig and in the crab How to distinguish be- tween species and varieties, as instanced by dimorphism in certain fishes, and in a marsh plant Polymorphism in plants Series of Fibonacci Discontinuous variation in animals as regards vertebrae, ribs, mammae, teeth, digits, and other characters Homoeosis De Vries' Theory of Mutation Dimorphism may be due to internal causes, or the result of divergent evolution Physiological Selection Infertility between varieties. WE have seen that the distribution of variations about their mean is in many cases quite symmetrical, whilst in other cases in which it is asymmetrical it still takes place according to some orderly arrangement, for which a mathematical expression can be found. There is still a third group of cases, however, in which the curve of distribution is, as a rule, very asymmetrical, but for which, even if symmetrical, no single general mathe- matical expression is obtainable. A study of such curves has taught us that the cause is frequently refer- able to the fact that our material is not homogeneous; that, in fact, we have a mixture of varying numbers of two or more groups of individuals differing in mean size and range of variation. For instance, Bateson * measured the length of the forceps of 583 specimens of the common earwig, Forficula auricularia, which had * Materials for the Study of Variation," p. 41. 37 38 DISCONTINUOUS VARIATION. been collected at random in one day in the Fame Islands. Only mature males with elytra fully devel- oped were measured. The range of variation was from 2.5 to 9.0 mm., the various lengths occurring with a frequency indicated by the accompanying curve. Here 120 100 60 40 20 \ 4567 .Length ofi Forceps in - -30- 580 5UO 600 610 620 630 (JM 650. 660 t>70 60 6.90 700 FIG. 9. Distribution of frontal breadths of Carcinua mcenas. served values, and so supports Weldon's hypothesis. It is somewhat curious that of all the 22 series of measure- ments made by Weldon on Naples and Plymouth crabs, this was the only characteristic in respect of which dimorphism was exhibited. As an explanation of it, Giard * has suggested that one of the two groups owed its altered frontal breadth to the presence of an internal parasite, Portunion mcenadis. Thus he measured five specimens of C. mcenas infested by this parasite, and found that their mean relative frontal breadth very * Comptes Rendus, cxviii. p. 870, 1894. DISCONTINUOUS VARIATION. 41 nearly corresponded to the lower mean value of Wei- don's crabs (viz., 630.32 as against 630.62). Giard thinks that the dimorphism in the length of the forceps of the earwig observed by Bateson can be similarly ex- plained, for the short individuals appear to be infested with Gregarines, and the longer ones not. He does not wish to insist, however, that all dimorphism is the result of parasitic influence, but merely that it may be so in certain instances. It is obvious, indeed, that be- tween two absolutely distinct varieties or species, and between pure monomorphic forms, all intermediate stages may exist. But how are these intermediate stages to be classified? When is one justified in assum- ing the existence of two distinct species, and when of only one species with an increased range of variation and perhaps a tendency to split up? To overcome this difficulty Davenport and Blankinship * have suggested that in order to decide in any given case whether we are dealing with two or more confluent species, or only with varieties, the following procedure should be adopted: First of all one should determine the most distinctive character of the members of the group, and after mak- ing a series of measurements in respect of this char- acter, plot out a curve showing the relative frequency of occurrence of each measurement. Supposing that in this way a double humped curve is obtained, then the degree, of isolation of the constituent races is esti- mated by measuring the depth of the depression be- tween the two humps, from the level of the maximum of the lower hump. This depth may be expressed as a percentage on the length of the maximum ordinate, or * Science, N. S. vol. vii. p. 685, 42 DISCONTINUOUS VAEIATION. mode, of the lower hump. The value so obtained is termed by Davenport and Blankinship the " Index of Isolation." They suggest that if this index be over 50 per cent., then one should agree to look upon the two groups as distinct species; if under 50 per cent., then only as varieties. As an example, they adduce a case of two doubtful species of fishes, Leuciscus balteatus, and L. hydrophlox, which differ in the number of rays in the anal fin. On plotting out the frequencies of oc- w 50 40 30 20 10 I \ 1 \ I \ / '\ / \ ^ / \ *-~~^ ->^ / / ^> ^- 12 14 16 18 20 22 FIG. 10. Distribution of fin rays in Leuciscus. currence of the various numbers of rays in 194 indi- viduals, the curve given in Fig. 10 was obtained. In this case the index of isolation is exactly 50 per cent., hence we are just at the limit of species and varieties. The importance of determining which is the most distinctive character, before drawing any conclusions from the indices of isolation found, is well shown by an- other case adduced by these authors. It concerns the marsh plant Typha, which is found in the eastern United States. Seven characters regarded as probably specific were measured in about 250 specimens, which had been collected at distances of about one metre apart across the swamps in which it occurred. The variation curves obtained in the case of the stem height, DISCONTINUOUS VARIATION. 43 diameter of stem taken at half the height, and the width of the largest leaf at its widest part, are reproduced in Figs. 11, 12, and 13. Here we see that the stem height shows no differentiation, the curve being more or less symmetrical. The mid-stem diameter shows a slight second hump, but this is obviously insufficient to indicate the presence of two species. The leaf width, 40- 20- 10- NO.: 7 Indv. r Dm.t 8 9 1011121314151617181920 21 Stem-Height FIG. 11. Distribution of Stem-Heights in Typha. however, shows a marked differentiation, the index of isolation being 75 per cent. Of the other characters measured, the diameter of the stem at its base, the diameter of the pistillate spike, and the interval be- tween the staminate and pistillate spikes, had indices of isolation of respectively 79, 89, and 83 per cent., or showed even greater differentiation than the leaf width. The curve for the pistillate spike length was, however, symmetrical. Thus this plant shows distinct differen- 44 DISCONTINUOUS VARIATION. tiation into two species in respect of four out of the seven characters measured, and so is obviously, in the authors' opinion, to be regarded as composed of two more or less confluent species. This " precise criterion of species " suggested by Davenport and Blankinship has much to recommend it, but probably it would generally be considered that an index of isolation of only 50 per cent, is too small a dif- ference to merit specific distinction. Perhaps it would 70 40 20 4 6 8 10 12mm. Mid-Stem Diameter FIG. 12. Distribution of Mid-Stem Diameters in TypTia. be better, therefore, to increase it to 90 or 95 per cent. In any case, it must, I think, be admitted that a slight degree of confluency, or overlapping of the curves of variation, ought not to compel one to assume the exist- ence of only a single species, though this is the view which has been generally held in the past. On the other hand, the fact of a group of organisms showing absolutely discontinuous variation in respect of some apparently unimportant characteristic ought not to DISCONTINUOUS VAKIATION. 45 compel one to regard it as composed of two distinct species. It is, of course, often impossible to tell whether any given characteristic is important or not, and hence we must recognise that a really precise and univer- FIG. 13. Distribution of Leaf- Widths in Typha. sally applicable definition of a species is, and always must be, unattainable. It is probable that variation series in the Vegetable Kingdom often give double humped or multiple humped curves, even if the material examined is as homogeneous as it is possible to obtain it. Possibly, if only individuals of the same stock were examined, they would be found to give single humped curves, but if material collected from different parts of the same district, or even of the same field, is to be regarded as composed of so many local races or sub-varieties, then the determination of the variations in many plant 46 DISCONTINUOUS VARIATION. species would become an almost hopeless task. The degree to which local races may vary is well shown by some of Ludwig's determinations. For instance,* four groups of specimens of Torilis anthriscus, obtained from various spots near Schmalkalden, had the follow- ing numbers of branches of the main umbels : 34 5 67 89 10 11 12 13 Total Group I. 1 11 18 45 28 20 5 4 1 133 " II. 1 5 8 13 25 12 5 2 71 i III. 5 7 9 8 12 8 1 1 51 " IV. 7 60 213 152 46 18 3 1 500 In the first group, the 8 branch form occurred the most frequently, the variations from this number being distributed more or less evenly around it. In the next group, the 10 branch form was the commonest; in tlie third group the 8 and 10 branch forms were both com- mon, whilst in the last group of all, collected in a wood at Wolfsberg, near to Schmalkalden, a quite distinct race having five branches presented itself. In determining the variations of a plant, therefore, it is probably best to obtain a very large amount of ma- terial, from various sources, and submit this to examina- tion. Though the curve thereby obtained may be very composite, yet at least it will indicate something as to the range of variation of the flower in many of its local races, and also what number of branches or other parts occur on the whole most frequently. For in- stance, Ludwig f has had made enumerations of the number of ray florets in 17,000 specimens of the Ox-eye Daisy, Chrysanthemum leucanthemum. The material *Bot. Centralb., vol. Ixiv. p. 40. fBot. Centralb., vol. Ixiv. p. 1, 1895. DISCONTINUOUS VAKIATION. 47 was collected from various sources, between the years 1890-95, and was examined by various people. The number of florets varied from 7 to 43, the following being the frequency of their occurrence: N o . kj g g 1 i i | | g 1 1 P I 1 fc D | I 1 m & 7 2 16 479 25 602 34 346 8 9 17 525 26 614 35 186 9 13 18 625 27 375 36 64 10 36 19 856 28 377 37 28 11 65 20 1568 29 294 38 16 12 148 21 3650 30 196 39 16 13 427 22 1790 31 183 40 14 14 383 23 1147 32 187 41 15 455 24 812 33 307 42 3 43 2 Those results are plotted out in the form of a curve in Fig. 14. Here we see that the 21 floret form occurs by far the most frequently, but that there are also secondary smaller maxima or humps on the curve for 13, 26, and 34 floret forms. This curve thus gives one a good idea both as to the range of variation of the num- ber of florets in this flower, and also as to the most fre- quently occurring forms. Enumerations of small num- bers of specimens of local races showed the 13 floret form to be the commonest form in one case, and the 34 floret form in another, but these results obviously fail to give a true idea of the variation of the plant. Similar enumerations of the florets of various other species of the Composite? showed the following to be the most frequently occurring numbers of ray florets, 48 DISCONTINUOUS VAEIATION. the absolute maximum being in each case indicated by thick type : * Chrysanthemum leucanthemum, 13 21 34 26 " inodorum, 13 21 " segetum . 13 21 Anthemis arvensis, Cotula, " tinctoria, Achillea ptarmica, Senecio nemorensis, Fuchsii, 5 8 13 8 13 21 8 13 10 3 5 3 5 Taking the observations as a whole, we see that the most frequently occurring numbers are the following: 3 5 8 10 13 21 26 34 In various species of the Umbellifercz the following are the most frequently occurring numbers of petals: 3 5 (10 15 20 25) 8 13 21 34 That is to say, in each case the numbers follow the so- called series of Fibonacci, viz., 1, 2, 3, 5, 8, 13, 21, 34, etc., in which each number is the sum of the two before it; or else they follow some multiples of these numbers. Ludwig states that this relationship is by no means limited to the two orders of plants mentioned, but that it extends to other members of the Vegetable Kingdom, and probably also to the Animal Kingdom. Perhaps the most striking result that Ludwig t ob- tained was for the number of petals of one of the Prim- roses, Primula officinalis. This varied from 1 to 22 in the sample of 1170 flowers examined. These flowers were all obtained from a single meadow (near Wieda), and were therefore as homogeneous as it was possible to *Bot. Centralb., vol. Ixiv. p. 100. |Ber. d. deutsch. bot. Gesell., xiv. p. 204, 1896. DISCONTINUOUS VARIATION. 49 obtain them. The frequency of occurrence of the vari- ous forms is indicated in Fig. 15, given below. Here it will be seen that the curve has five most dis- 8500 85 40 Number of Ray=-f lorets. FIG. 14. Distribution of Ray-florets in the Ox-eye daisy. tinct maxima, corresponding to 3, 5, 8, 10, and 13 petal forms. That these multi-humped curves are actually due to a mingling of two or more local races is proved by an 50 DISCONTINUOUS VARIATION. interesting experiment of De Vries.* He sowed the mixed seed of Chrysanthemum segetum obtained from twenty different gardens. The topmost flowers of the 024 6 8 10 12 14 16 18 20 Kumbei: of petals PIG. 15. Distribution of Petals in Primula officinalis. chief stem of each of the 97 healthy plants obtained were examined, and were found to contain the follow- ing numbers of ray florets: Ray florets, 12 13 14 15 16 17 18 19 Frequencies, 114 13 4 6 9 7 10 *Arch. f. Entwickelungsmechanik, Bd. ii. p. 52, 1896. 21 20 DISCONTINUOUS VARIATION. 51 Thus there were obviously two forms present, a 13 ray form and a 21 ray form. The seeds from the 12 and 13 ray forms were collected and sown next year, the flowers obtained therefrom having the following num- bers of florets: Ray florets, 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Frequencies, 2 107 13 94 25 7712030 That is to say, all trace of the 21 ray form had been eliminated, and a nearly pure 13 ray form obtained. That this was so was proved by sowing the seed of some of the 12 rayed plants obtained on this occasion in the following year. It was then found that the frequencies of occurrence of flowers with various numbers of rays remained practically unchanged. But how do these cases of what Bateson has termed discontinuous variation arise ? In one or two of the in- stances quoted we saw that the two humps of the curve of variation scarcely overlapped at all. In the case of Primula they were all of them sharply defined, but there was still a good deal of fusion, whilst in the ray florets of the Ox-eye Daisy the fusion was greater still. Finally, in the frontal breadths of Naples crabs the fusion was complete, and the existence of dimorphism was shown only by the asymmetry of the curve. It would be possible to multiply instances of such curves as these, in which the fusion ran through all stages of completeness and incompleteness, but those quoted are quite sufficient for our purpose. They suffice to show that all stages of fusion may be met with, and so incline one to the opinion that the later stages, in which the two or more humps of the curve overlap little if at all, are 52 DISCONTINUOUS VARIATION. but more advanced stages of those curves in which the fusion is nearly complete. In fact they seem to indi- cate that if only the ancestry of such varying organisms could be traced backwards continuously, it would be found that at no period was there any sudden change from continuity to discontinuity; that a condition of absolute dimorphism, or formation of two new species, was merely evolved by very gradual and almost imper- ceptible steps from the original pure monomorphism. This is, I believe, the opinion held by the majority of naturalists at the present day as to the origin of by far the larger number of cases of dimorphism, but dissen- tient voices have not been entirely wanting. Thus Gal- ton * is of the opinion that the aberrant or discontinu- ous variations generally known as sports may be of con- siderable significance in evolution. Because evolution may proceed by minute steps, he considers that it does not by any means follow that it must so proceed. Again, within recent years the orthodox view has been ably combated by Bateson f in his book on Variation. In this work he has collected a very large number of in- stances of discontinuous variation, or variations in re- spect of certain organs or parts, which have suddenly arisen in a complete and perfect state, without, as a rule, the occurrence of any intermediate stages. If, therefore, argues Bateson, such instances of discontinu- ous variation undoubtedly occur, is it not possible that the Discontinuity of Species which is so striking a fact amongst living organisms is a consequence and expres- sion of this discontinuity of variation? Thus the view * " Natural Inheritance," p. 32, 1889. f " Materials for the Study of Variation," London, 1894. DISCONTINUOUS VARIATION. 53 hitherto generally held, since Darwin first gave ex- pression to it, is that almost all variations are very slight, and form a continuous series. It is only by their very slow accumulation, therefore, under the action of Natural Selection and other agencies, that species as we know them have been evolved. This view of Bateson's is so striking and important that it behoves us to ex- amine it at some little length. It will enable us to ob- tain a clearer idea of Bateson's views if we indicate his system of classification. Thus he points out that varia- tions are divisible into two classes, substantive and meristic. Substantive variations are variations occur- ring in the actual constitution or substance of the parts themselves. Meristic variations, on the other hand, are those which relate to the number of parts in organisms. For instance, the flower of the Narcissus is commonly divided into six parts, but through meristic variation it may be divided into seven parts, or only four. Never- theless, there is in such a case no perceptible change in the tissue or substance of which the parts are made up. On the other hand many Narcissi, N. corbularia, for example, are known in two colours, one a dark yel- low, and the other a sulphur yellow, though the number of parts and pattern of the flowers are identi- cal. This is, therefore, an example of a substantive variation. Bateson considers that there can be no doubt that these two classes of variation are essentially distinct from each other. It is obvious that all cases of meris- tic variation are also cases of discontinuous variation, whilst cases of substantive variation are much more fre- quently continuous than discontinuous. It is to be noted 54 DISCONTINUOUS VAEIATION. that these discontinuous meristic variations are not only large, but they are complete and perfect. But after all one would scarcely expect anything else. Between a six petal and a seven petal flower it is scarcely pos- sible to imagine such a thing as a really intermediate stage. Even if one found a flower with six normal petals and a seventh abnormally small one, or five nor- mal ones and a sixth in process of dividing into two, one would scarcely be justified in regarding it as an in- termediate form, for the flower would no longer be sym- metrical. Perhaps the most interesting part of Bateson's work lies in the cases which he has collected of what he terms Homceosis. By this he means those variations which consist in the assumption by one member of a meristic series, of the form and characters proper to other mem- bers of the series. For instance, Kraatz has described a saw-fly, Cimbex axillaris, having the peripheral parts of the left antenna developed as a foot, the right an- tenna being normal. Kriechbaumer has described a nearly similar condition in a Humble-bee, Bombus variabilis. Bateson has himself described a crab, Cancer pagurus, having the right third maxillipede de- veloped as a chela. Milne-Edwards has described an- other crab, Palinurus penicillatus, in which the left eye bore an antenna-like flagellum several centimetres in length, growing up from the surface of the eye. The eye stalk appears to have been of normal shape, but re- duced in size. Other instances somewhat similar to these are adduced, but it is unnecessary to quote them here. It may be mentioned, however, that most inter- esting examples of this form of variation have been DISCONTINUOUS VARIATION. 55 recently obtained by Herbst.* He found in no less than ten different species of Crustacea, belonging to four different families, that if an eye stalk were totally extirpated, there always grew up in its place a hetero- morphic new structure, like an antennula, which bore olfactory hairs. If, however, only the eye were re- moved, and the stalk together with the ganglion left, instead of the antennula there arose the beginnings of a new eye. A similar result to this was obtained on ex- tirpation of either stalk or eye in Porcellana platycheles, presumably because in this species the stalk contains no ganglion. The chief contents of Bateson's book may be very briefly summarised, in order that the reader may gather some idea as to the kind of evidence on which Bateson founds his argument. A considerable body of evidence is given concerning variations in the numbers of ver- tebrae and ribs, the most important conclusions being that the variations are considerable, especially in some types such as Simia satyrus, the Bradypodidse, and Bombinator igneus, and that imperfect vertebrae are very rare. Turning to Invertebrates, it is shown that among Oligochaeta and Hirudinea, certain forms, e. g., Perionyx excavatus and PacJiydrilus sphagnetorum, have great variability, whilst others, such as the com- mon earthworm, rarely vary. Both forward and back- ward Homoeosis may occur, forms which normally have the male pores on the 15th segment having them on the 16th, or on the 13th. Returning to Vertebrates, evi- dence is next adduced concerning cervical fistulae and auricles, and supernumerary mammae. Variations in * Arch. f. Entwickelungsmechanik, Bd. ix. p. 215, 1899. 56 DISCONTINUOUS VARIATION. teeth are dealt with in very great detail, and conclu- sions are drawn as to the comparative frequency of dental variation in various animals. The animals show- ing the greatest frequency of extra teeth are- domestic dogs, Anthropoid apes, and the Phocidse. It is espe- cially noticeable that the variability of domestic animals in respect of teeth is not markedly in excess of that seen in wild forms. Thus, though supernumerary teeth are more common in domestic dogs and cats than in wild Canidse and Eelidse, they are not more so than in An- thropoid apes and in the Phocidse. With respect to the question of symmetry, the evidence shows that dental variation may be symmetrical on the two sides, but that much more frequently it is not so. Other evidence is given concerning the division of teeth, the presence and absence of first premolars and last molars, the least size of particular teeth, and other subjects. Variations in the number of digits are treated more fully than any other subject discussed, though the evi- dence adduced is stated to bear rather on morphological conceptions than in any direct manner on the problem of Species. It is found that the frequency of digital variation is immensely greater in some classes of species than in others. Thus the horse shows many recorded cases, but the ass none at all. Variation is common in the cat, pig, fowl, and pheasant, but rare in the dog, sheep, and in most birds. In the cat, ox, horse, pig, and in man, the digital variation approaches to particular forms, and has in it something distinctive. Digital variation is sometimes symmetrical, but more often asymmetrical. In other chapters of the book is found a considerable DISCONTINUOUS VARIATION. 57 mass of data concerning the repetition and division of appendages in insects and Crustacea, colour markings and colour variations in Lepidoptera, variations in the number of legs of different species of Peripatus^ the occurrence of double monsters, and various other sub- jects, but to these it is unnecessary to refer here. Sufficient have been mentioned to indicate the general nature and scope of the evidence, so that we are enabled to enquire how far, if at all, it can warrant Bateson's hypothesis as to the origin and production of discon- tinuity in species. We see that most of the evidence con- cerns obvious abnormalities, generally in the direction of increase in the number of parts, which have arisen suddenly and apparently spontaneously. In practically no case has any new structure arisen, but only a repe- tition or misplacement of those already present, and so it is difficult to understand how really new structures and organs could have originated, even if it be admitted that such abnormalities are of very frequent occur- rence, and that they could succeed in permanently establishing themselves. But first of all it is neces- sary to point out that the large majority of these abnor- malities are extraordinarily rare, occurring perhaps not once in 100,000 or once in a million cases. What chance have they, then, of establishing themselves on a permanent footing? Bateson remarks, " An error more far-reaching and mischievous is the doctrine that a new variation must immediately be swamped," but he fails to adduce one tittle of evidence to prove that it is an error at all. This is simply because no such evidence exists. It is true that some animals are prepotent over others in procreating their characteristics, and their 58 DISCONTINUOUS VARIATION. abnormalities if they possess them, but this prepotency is quite limited in its range. Darwin in his " Variation of Animals and Plants under Domestication " * men- tions an instance of transmission of supernumerary digits through five generations, whilst in other cases they have reappeared after an interval of even three generations. " But," says Darwin, " we must not over- estimate the force of inheritance. Dr. Struthers as- serts that cases of non-inheritance and of the first ap- pearance of additional digits in unaffected families are much more frequent than cases of inheritance." Unless much stronger evidence than that hitherto ad- vanced be obtained, it therefore follows that, according to the known laws of inheritance, suddenly occurring variations, unless artificially selected, must inevitably be swamped by intercrossing and disappear. Suppos- ing, on the other hand, any such variation is artificially isolated, and bred in and in with its own offspring, then it may be possible to establish a distinct race, bearing in undiminished degree all the abnormal characteristics of the original variety. For instance, Darwin thus records the origin of the ancon sheep :f "In 1791 a ram-lamb was born in Massachusetts, having short crooked legs and a long back, like a turnspit dog. From this one lamb the otter or ancon semi-monstrous breed was raised ; as these sheep could not leap over the fences, it was thought that they would be valuable. The sheep are remarkable for transmitting their char- acter so truly that Colonel Humphreys never heard of ' but one questionable case ' of an ancon ram and ewe *Vol. i. p. 457, Ed. ii. \Ibid., vol. i. p. 104. DISCONTINUOUS VARIATION. 59 not producing ancon offspring." Again, Darwin says: * " It is certain that the ancon and the mauchamp breeds of sheep, and almost certain that the niata cattle, turn- spit and pug-dogs, jumper and frizzled fowls, short- faced tumbler pigeons, hook-billed ducks, etc., suddenly appeared in nearly the same state as we now see them. So it has been with many cultivated plants." Ajid then he adds, " The frequency of these cases is likely to lead to the false belief that natural species have often originated in the same abrupt manner. But we have no evidence of the appearance, or at least of the con- tinued procreation, under nature, of abrupt modifica- tions of structure." We see therefore that, though Darwin brought for- ward much more powerful and convincing instances of discontinuous variation than those cited by Bateson, he held them to be quite inadequate to account in any way for the discontinuity observed in species. Under the title of " Die Mutationstheorie" De Vries has recently promulgated views concerning the origin of species which are somewhat similar to those held by Bateson. The evidence he adduces in support of them is chiefly derived from observations of his own on flower- ing plants, and even if his theoretical views be entirely rejected, there can be no doubt as to the intrinsic interest and importance of the observations themselves. According to the theory of mutation, the qualities of organisms are built up of individual units sharply defined from each other. When, in the course of evolution, one species arises from another, it follows that the change takes place by a distinct step or jump, *Ittd., vol. ii. p. 409. 60 DISCONTINUOUS VARIATION. i. e., is discontinuous, and does not occur gradually. Such mutations may take place in all directions, but probably they only occur from time to time, due, per- haps, to the periodical action of fixed causes. They are distinct from the slight differences observed in local races and varieties, for these can be produced gradually by artificial selection and changed condi- tions of environment. Also Natural Selection can only lead to the formation of such local races, it being powerless to bring about true mutations. The varia- tion which leads to the formation of new species, there- fore, is essentially discontinuous, not continuous. In order to obtain evidence in support of his theory, De Yries has cultivated over 100 different species of plants, but only one of them, (Enothera Lamarckiana, showed the desired mutations. This plant was origi- nally brought to Europe from America, and kept under cultivation. It has since run wild, and De Vries ob- tained the stock of nine plants which formed his first generation from a field near Hilversum. Unfortu- nately, the true origin of the plant is obscure. In Britton and Brown's recently issued " Flora of the United States," no reference whatever is made to it as a wild species. Hence it is probably a garden variety of (Enothera biennis (Evening Primrose), and may be a hybrid plant, whilst the mutations obtained by De Yries may merely be partial or complete reversions to the original ancestors of the plant. To obtain these mutations, De Yries cultivated the plant through eight generations, and during this time, obtained over 50,000 specimens. Of these, 834 showed characters which sharply differentiated them from the normal 0. La- DISCONTINUOUS VARIATION. 61 marckiana. De Yries classified them as follows: 350 0. oblonga; 229 0. lata; 158 0. nanella; 56 0. albida; 32 0. rubrinervis; 8 0. scintillans; and 1 0. gig as. Of these new " species," oblonga, albida, rubrinervis, nanella, and gigas remained absolutely constant in sub- sequent generations, when crossed among themselves, or self -fertilised, in the case of 0. gigas. 0. scintillans was not nearly so constant, the offspring yielding only about a third of the parent form, and the rest of them being Lamarckiana, oblonga , and lata. 0. albida bred quite constant, but the plants were weak, and not very fertile. De Yries looks upon his sports as true species, and not varieties, for he says that varieties differ from their parent species in only one or two characters, whilst species differ from their nearest allies in almost all their characters. Thus in comparison with the parent form, 0. Lamarckiana, gigas was stronger and albida was weaker, both forms having broader and shorter leaves. The flowers of gigas were larger, those of rubrinervis darker yellow, those of oblonga and scintillans smaller, and those of albida paler. The cuticle of albida was rough. The bosses on the leaves of lata were increased, and on those of scintillans diminished. The forma- tion of pollen was increased in rubrinervis and dimin- ished in scintillans. The seeds of gigas were larger, and those of scintillans smaller; those of rubrinervis more abundant, and those of lata more scanty. By artificial selection De Yries obtained in one in- stance * what he regards as a true mutation. This was in the case of Linaria vulgaris (yellow toadflax). * " Die Mutationstbeorie," p. 552, 62 DISCONTINUOUS VAEIATION. Starting with plants which had one or two peloric flowers, he bred and selected them through several gen- erations, and ultimately obtained some entirely peloric plants. These plants were most of them sterile, but a few yielded seeds, and ultimately De Yries obtained a peloric race, Linaria vulgaris peloria, only 10 per cent, of the seeds of which reverted to hemipeloric plants (i. e., plants with some peloric and some non-peloric flowers). According to De Yries, this new race is a true mutation, because the pure peloric plants from which it was derived arose suddenly, and apparently capriciously, from hemipeloric parents. He also ob- tained new races of other plants by artificial selection extending through several generations, but these he regards only as varieties, and not true species, in that their formation was gradual. However, there seems to be no valid ground for sharply differentiating them in this manner. For instance, in the case of the five- leaved clover race (Tri folium pratense quinque folium) obtained by him,* he started breeding with two natu- rally occurring clover plants which had four leaflets to their leaves, and in the case of one leaf, five leaflets. It is difficult to understand why these naturally occur- ring plants should not be regarded as true mutations, just as much as the peloric race above mentioned, or why, indeed, the naturally occurring hemipeloric plants from which the peloric race was obtained were not like- wise true mutations. In the case of the clover, the breeding was continued through several generations, the seed of only the few plants richest in four or more leaflets being preserved * LOG. cit., p. 437. DISCONTINUOUS VARIATION. 63 for Bowing the following year. The proportion of four leaflet plants steadily increased, and in the fourth gen- eration the most widely diverging plants had the fol- lowing (percentage) proportions of leaves with from 3 to 7 leaflets. TOTAL NUMBER OF LEAVES COUNTED. Number of leaflets, 3 4 5 6 7 Normal plant, 17 16 37 14 16 172 Atavistic " 75 19 5 1 216 Extreme variation, 12 9 22 17 40 97 The plant considered as " normal " was obviously of a five leaflet type, the numbers of leaves with 3 and 4 leaflets, and those with 6 and 7 leaflets, being dis- tributed symmetrically around it. A comparison of the numbers of leaflets in the " atavistic plant " with those in the " extreme variation " is interesting, as showing the range of variation possible in plants of the same stock. As an instance of the formation of a variety, De Vries' experiments with Chrysanthemum segetum grandiflorum may be quoted.* Starting in 1896 with plants which had 21 ray florets occurring most fre- quently in their capitula, and none of which had more than 23 florets, he picked out each year the two or three plants richest in florets for breeding with, and sowed their seed the following year. In 1897 a single flower was obtained having 34 florets, but the 21 floret form was still the commonest. In 1898 one of the flowers had 48 florets, the commonest forms now having 26 or 34 florets. In 1899 one had 67 florets, the com- monest forms having 26 or 33 to 35 florets, and in 1900 *Loc. cit.,p. 523. 64 DISCONTINUOUS VAEIATION. one had 101 florets, the commonest form having 47 florets. Let us now return to the cases of dimorphism men- tioned at the beginning of the chapter. Instances were there adduced in which the dimorphism was slight, fairly marked, or so great that the two forms scarcely overlapped at all. To what may such dimorphism be due? Bateson points out that a dimorphic condition may have arisen from a previous monomorphic one, or it may always have been present since the character was first acquired. As already stated, the first view is the one which finds favour in the eyes of most biologists, but on the other hand there is a certain amount of evi- dence to show that the second view may hold good, at least in some cases. It is a well-known fact that when two breeds are crossed, their characters do not always blend, but are transmitted in an unmodified state to the offspring from one or from both parents. For instance, in breeding game fowls, if one crosses a black with a white game, birds of both breeds of the clearest colour are obtained. " Sir R. Heron crossed during many years white, black, brown, and fawn-coloured Angora rabbits, and never once got these colours mingled in the same animal, but often got all four colours in the same litter." * Again, Miss E. A. Saunders f has recently made observations on Biscutella lavigata, a cruciferous plant occurring as a perennial herb in the alpine and sub-alpine regions of middle and southern Europe. It was observed by Bateson that this species exhibits two distinct forms, which exist side by side, the one hairy * " Animals and Plants," vol. ii. p. 70. fProc. Roy. Soc., vol. Ixii. p. 11. DISCONTINUOUS VARIATION. 65 and the other smooth or glabrous. Intermediate forms were also found, but these were scarce. Some of the ripe seeds of these plants were obtained, and grown in England. A portion of the seedlings were derived from cross-fertilised seeds of known origin, and it was found that though there was to a certain extent a blend- ing of parental characters as regards hairiness and smoothness in the offspring of plants of dissimilar types, giving rise to intermediate forms, yet this intermediate condition was found only quite exceptionally among full-grown individuals. It was much more common in the young plants, but as these grew older, their leaves became smooth, and hence almost all the plants were ultimately either hairy or glabrous; that is to say, they varied discontinuously. Supposing a dimorphic condition is due to internal causes, or to the fact that it is the nature of the plant to vary in this way around two " positions of organic stability," as Galton has termed them, rather than around one such position, then it would seem almost impossible to get further to the root of such causes. Supposing, on the other hand, as is probably true in the majority of instances, this dimorphic condition has been derived from a previous monomorphic one, then we may hold a more reasonable hope of being able to eluci- date the cause or causes of this evolution from one con- dition to another. The problem of the splitting up of species was recognised by Darwin to be one of immense importance, and he discussed it at some length in the " Origin of Species." * The chief cause of divergence of character he attributed to the circumstance that *Ed. vi. p. 86. 66 DISCONTINUOUS VARIATION. " the more diversified the descendants from any one species become in structure, constitution, and habits, by so much the more will they be better enabled to seize on many and widely diversified places in the polity of nature, and so be enabled to increase in numbers." He also attached considerable importance to geograph- ical isolation of a portion of a species, as an element in the modification of species through natural selection. Though Darwin's principle of diversification of structure is doubtless a very true one, yet it does not in itself contain sufficient clue as to why a species should split up into two or more varieties. Thus, if by some means these actually arose, but both continued to in- habit the same area, it is difficult to understand why intercrossing should not rapidly reduce them to the single species from which they took their origin. It was to overcome this difficulty that Romanes suggested his theory of " Physiological Selection." * This theory is founded on the fact that individuals of a species, though fertile with some, may be perfectly sterile with other individuals, and this apparently independent of any differences of form, colour, or structure. Romanes thought that this incompatibility might run through a whole race or strain, and so a group of individuals of a species be in a physiological sense isolated from the rest, and therefore able to vary independently, without hav- ing their newly acquired characteristics swamped by intercrossing. As Wallace has very clearly shown, t this theory, in the form originally proposed by its au- *Journ. Linn. Soc. (ZoSl.), vol. xix. p. 337, 1886. Also "Dar- win and after Darwin," vol. iii. p. 41. f" Darwinism," p. 181. DISCONTINUOUS VARIATION. 67 thor, cannot stand, at least for members of the Animal Kingdom in which there is no promiscuous union of the sexes. For instance, if 10 per cent, of the members of a species are thus physiologically isolated, so as to be fertile inter se, but sterile when crossed with any of the other members of the species, and if the interbreeding take place purely according to the laws of chance, then on an average only one-tenth of these 10 per cent, will happen to pair with individuals with which they are fertile, and the remaining nine-tenths will form abso- lutely sterile unions. Thus this physiologically isolated section will never be able to increase in numbers and establish itself. In the case of flowering plants which are fertilised by insects, each of which perhaps visits ten or more flowers in a journey, Fletcher Moulton has shown * that Wallace's objection does not hold, as the dimunition in fertility in such a case is practically neg- ligeable. In his more recent discussion of the theory Romanes somewhat modified his views, and laid more stress on the fact that the mutual sterility may have been slight at first, and have been subject to a gradual development, it acting as a segregating cause in a de- gree proportional to its completeness. However, he makes no suggestion as to why and how such physio- logical incompatibility should arise, other than general ones such as the influence of food and climate, and spontaneous variability of the reproductive system. It seems to me that the origin of this physio- logical barrier which so generally exists between species can be most readily accounted for by assuming that in some cases at least it is, as Romanes * " Darwin and after Darwin," vol. iii. p. 165. 68 DISCONTINUOUS VARIATION. suggests, slowly evolved from an originally almost imperceptible degree of infertility, but that this takes place only simultaneously with the evolution of mor- phological character, in consequence of some form of isolation. Thus, suppose a number of individuals of a species become for a time separated from the remainder of the species by a geographical barrier, by migration, or some other cause of isolation, whereby they are enabled to vary independently of the gen- eral stock in response to changed conditions of life. Then as they gradually become more and more diver- gent from the parent stock in respect of morphological characters, it is highly probable that they may concur- rently perhaps from the direct action of the body tissues on the reproductive system diverge also in re- spect of physiological characters. Should any of them now happen to meet and intercross with individuals of the parent stock, or even if they should occupy the same breeding area again, their newly acquired mor- phological characters would no longer be in danger of being swamped, for the simple reason that few or no hybrid offspring would result from such crossing. In the case of the higher animals, also, it is probable that individuals of different varieties or sub-species, once these are formed, instinctively tend to breed amongst themselves, and hence the chance of production of hy- brid offspring is still further diminished. Thus Dar- win records * that in Paraguay it is believed " that the native horses of the same colour and size prefer asso- ciating with each other, and that the horses which have been imported from Entre Eios and Banda Oriental * Ibid., vol. ii. p. 80. DISCONTINUOUS VARIATION. 69 into Paraguay likewise prefer associating together." Again, " It has been observed, in a district stocked with heavy Lincolnshire and light Norfolk sheep, that both kinds, though bred together, when turned out, in a short time separate to a sheep." Still again, with respect to fallow-deer, " Mr. Bennett states that the dark and pale coloured herds, which have long been kept together in the Forest of Dean, in High Meadow Woods, and in the New Forest, have never been known to mingle." Dar- win adduces other similar instances, in the case of the dog, horse, sheep, rabbit, and pigeon] hence there can be little doubt of the genuineness of the phenomenon, even though it is not based on very exact observation. Supposing that the above view is correct, it follows that between at least some varieties there must exist a greater or less degree of sterility. Of course it is not necessary that divergence of morphological character should always be accompanied by corresponding diver- gence of physiological character; but merely that this is sometimes the case. Upon this point Darwin has col- lected a considerable amount of evidence in his " Ani- mals and Plants." * One or two of the cases there cited may be quoted here. Gartner found that a variety of dwarf maize, bearing yellow seed, showed a considerably diminished fertility with a tall maize bear- ing red seed, though both varieties were perfectly fer- tile when crossed inter se. Again, in the genus Ver- bascum, numerous experiments were made by Gartner with the white and yellow varieties of V. lychnitis and V. blatteria, when he found that crosses between simi- larly coloured flowers yielded more seed than those be- *Vol. ii. p. 82. 70 DISCONTINUOUS VARIATION. tween dissimilarly coloured ones. These experiments have been repeated and extended by Scott with con- firmatory results. Still better evidence than that quoted by Darwin has been obtained by Jordan in a laborious research on various species of plants annuals and perennials, bul- bous and aquatic, trees and shrubs extending over thirty years. Jordan found that when a Linnean species is indigenous to a country, and is of common occurrence, it is represented by more or less numerous and perfectly constant varieties, all growing in intimate association with one another. It was found that in many hundreds of cases these varieties, though they differed but slightly in morphological characters, came true to seed, but were always more or less infertile when crossed inter se* With regard to members of the Animal Kingdom, there is very little evidence indeed. The following anthropological data may perhaps be held valid. From statistics collected in Prussia between 1875 and 1890, it was found that Protestants, Catholics, and Jews, when marrying among themselves, had, on an average, respectively 4.35, 5.24, and 4.21 children. When, how- ever, the husband was a Jew and the wife a Protestant or Catholic, the numbers of children were only 1.58 and 1.38 respectively: and when the wife was a Jewess and the husband a Protestant or Catholic, only 1.78 and 1.66 respectively. t Whether this apparent partial sterility was due to differences of race or to social rea- sons, it is impossible to say. Again, Professor Broca t * Quoted from Romanes, ibid., p. 86. f Quoted from Mayo Smith's " Statistics of Sociology," p. 115. \ " On the Phenomena of Hybridity in the Genus Homo," 1864. DISCONTINUOUS VARIATION. 71 has brought forward evidence that some races of man show diminished fertility together. In order to obtain further evidence, the author * made numerous observations on the effects of crossing the colour varieties of the sea-urchins Splicer echinus granularis and Strongylocentrotus lividus. With the former organism the numbers of blastulse and of larvae produced on crossing dissimilar colour varieties were distinctly smaller than for similar varieties. In the most marked instance, the similar color varieties yielded on an average 98.5 per cent, of blastulse, and 73 per cent, of larvaa, whilst the dissimilar yielded 68 per cent, of blastulaa and only 15.6 per cent, of larvaa. More- over these latter larvaa were 4.5 per cent, smaller than the others. In the case of Strongylocentrotus, however, where the colour varieties are much less pronounced, there was very little difference of fertility. There can be no doubt, therefore, that certain varie- ties show a greater or less degree of mutual infertility, though this is doubtless not nearly so marked, or of such frequent occurrence, as in the case of species. Whatever view be taken as to the cause of such infer- tility, and its relation to divergent evolution, these in- stances quoted have also an intrinsic value. They show that just as the deviations from the average in respect of morphological characters may form double humped curves, so the deviations in respect of physiological characters may show corresponding irregularities, though of course it is impossible to measure them exactly and construct their curves of variation. * Phil. Trans. 1898, B. p. 511. CHAPTER III. CORRELATED VARIATIONS. The measurement of correlation Gallon's function Correlation between various organs in man, in local races of the shrimp, and in crabs Comparison between primitive and civilised races of man Correlation between morphological characters and the reproduct- ive system Genetic Selection in man Especial fertility of type forms in certain plants Evolution in the Peppered moth Paral- lel variation Importance of mathematical treatment of variation. ALL parts of an organism are to a certain extent re- lated to each other, so that when one part varies other parts vary simultaneously in a greater or less degree. That is to say, variations are correlated. The most marked and obvious correlation is that existing between homologous parts. The symmetry of the correspond* ing or homologous organs on the right and left sides of the body, which is present in most animals, represents a very close degree of correlation. But even in this case the correlation is not constant or complete. Thus the two arms and the two legs of a man resemble each other very closely indeed, but careful measurement shows that the resemblance is not absolute. Again, the arms, as a rule, vary in length more or less in the same proportion as the legs, but personal experience will probably recall instances to the contrary, in which the length of the limbs was quite disproportionate. Be- tween the arms and the legs, therefore, the degree of 72 CORRELATED VARIATIONS. 73 correlation is obviously less close than between arm and arm, or leg and leg. Still again, personal experience teaches us that there is correlation between even the length of the face and that of the limbs. Tall men as a rule have longer faces than short men; or, a more striking instance, greyhounds have long heads and long legs on the one hand, as compared with bull-dogs with short heads and short legs on the other. Between the length of face and length of limb, however, it is clear that there is a less degree of correlation than between length of arm and of leg, and between certain other organs of the body the connection must be less intimate still. It follows, therefore, that between the various parts there may exist all degrees of correlation, stretch- ing from an almost perfect degree of resemblance down to an absolute lack of it. We must also recognise the existence of negative correlation, in which the variation of one part in one direction is accompanied by a greater or less degree of variation of another part in the oppo- site direction. Here again we may experience all de- grees of negative correlation, just as of positive. In- stances of negative correlation are much less frequent than those of positive, and the only one known to me in the case of man is that recently discovered by Professor Pearson.* It was found that between stature and head index the correlation is distinctly negative, or that brachycephalic or relatively broad-headed persons are slightly shorter than dolichocephalic or narrow- headed. From what has been said it is clear that a bald state- ment that in such and such a case one part or organ is *Proc. Roy. Soc., Ixvi. p. 23, 1900. 74 CORRELATED VARIATIONS. correlated with another conveys no exact meaning. Such a statement must vary according to the notion of the observer as to what does and what does not consti- tute correlation. In order to obtain reliable and com- parable data concerning the degree of correlation, it is necessary to obtain a mathematical expression for it, just as one was found to be necessary for indicating the range of a variation. The fundamental theorems of correlation were for the first time exhaustively discussed by Bravais * more than half a century ago, but a more convenient and improved method of obtaining an ex- pression was first indicated by Galton, and he termed it the correlation constant, or r. It is now more generally known as " Galton's function." The principle on which this constant is determined is best explained by a concrete instance, viz., one given by Galton in the original paper in which he explained his method.f Galton's data are anthropometric ones, ob- tained at his own laboratory, and consist of several measurements made on 350 males of 21 years and up- wards. For instance, Galton found that the average relation between stature and cubit, or distance between the elbow of the bent arm and the tip of the middle finger, was as 100 to 3Y. In determining the correla- tion between these two measurements, however, it is obvious that it is not possible to compare the absolute amount of variation of the one with the absolute amount of the other, or even the proportionate amounts, but * "Analyse mathematique sur les probabilites des erreurs de situ- ation d'un point." Memoires par divers Savans, T. ix., Paris, 1846, p. 255. fProc. Roy. Soc., vol. xlv. p. 135, 1888. CORRELATED VARIATIONS. 75 we must first transmute them into units dependent on their respective scales of variability. We shall thus cause a long or a short cubit and an equally long or short stature, as compared to the general run of cubits and statures, to be designated by identical scale values. The most convenient unit to employ is the value of the probable error of each group. The probable error of the cubit is .56 inch = 1.42 cm.; and of the stature, 1.75 inch = 4.44 cm. Therefore each of the measure- ments of the cubit must be transmuted into terms of a new scale, in which each unit = .56 inch, and each of the measurements of the stature into those in which each unit 1.75 inch. After this has been done, we shall find that on an average each deviation in the stature of say 1 unit from the mean is not accompanied by a similar deviation of 1 unit in the cubit, but by only .8 of a unit. Conversely it is found that in a similar manner each deviation in the cubit of 1 unit from the mean is accompanied by only .8 of a unit of de- viation in the stature. The degree of correlation, or r, between the one organ and the other, is therefore said to be .8. If the correlation had been perfect, then this r would have been equal to 1, and if it had been entirely wanting, then it would have been 0. Comparison with other data shows that a correlation of .8 is a high one, not often surpassed. The other correlation constants determined by Galton are the following: MEAN r. Stature and head length, 35 Stature and middle finger, 7 Middle finger and cubit, 85 Head length and head breadth, 45 Stature and height of knee, ..... .9 Cubit and height of knee, 8 76 CORRELATED VARIATIONS. Here we see that the maximum amount of correlation was observed between stature and height of knee, and the minimum between stature and head length. Even in this latter instance, however, the correlation was fairly marked. Thus a constant of .35 indicates that in men 1.75 inch, or 1 unit, above the mean stature, the QH length of head will on an average be ^-^ X .19 =.0665 -L . UU inch above the mean, .19 inch being the probable error of variation of the head length. The height of knee, on the other hand, would on an average be no less than Q :p-r X .80 = .72 inch greater. The various medians or middlemost values and probable errors found by Gal- ton are as follows : MEDIAN. PROBABLE ERROR. DIMENSION. INCH. CENTIM. INCH. CBNTIM. Head length, 7.62 19.35 .19 .48 Head breadth, 6.00 15.24 .18 .46 Stature, 67.20 170.69 1.75 4.44 Left middle finger, 4.54 11.53 .15 .38 Left cubit, 18.05 45.70 .56 1.42 Height of rt. knee, 20.50 52.00 .80 2.03 In order to determine the degree of correlation be- tween any two organs, it is therefore necessary to adopt the following procedure. Sort out all the individuals into groups such as, for instance, in the case of stature, those varying from 64 to 65, 65 to 66, 66 to 67 inches, and so on, and then determine in each of these groups the mean of all the deviations from the average of the organ of which the correlation with stature is to be determined. Thus, in the group of individuals 64 to 65 inches high, the average difference of all the indi- vidual cubit measurements from the median of the CORRELATED VARIATIONS. 77 cubit (18.05 inches), is about .6 inch. Let the devia- tion of each value for stature from its median (67.20 inches) be now divided by the probable error of varia- tion of stature (i. e., 1.75 inch), and each associated mean deviation of cubit be divided by its probable error (i. e., .56 inch). Then, by dividing each of these terms for cubit by the corresponding term for stature, a series of values is obtained, each representing the amount of correlation between the various degrees of stature and the cubit. These values would be approximately equal in amount if a very large number of observations were made, but with only moderate numbers they vary very considerably. A mean of all of them may be called r, or the average degree of correlation between cubit in relation to stature. In a similar manner the individuals must be split up into groups in respect of cubit, and the associated deviation of stature determined. Another series of correlation values will be obtained, represent- ing stature in relation to cubit, of which the mean may be called r a . This value is found to be approximately equal to r l} and the mean of r^ and r a is called r, or the correlation constant. The degree to which these individual correlation values vary is best shown by means of a diagram. The one given in Fig. 16 is taken from Professor Weldon's paper on correlated variations in Crangon vulgaris* and represents the correlation between the post-spinous carapace length and the total carapace length in Plymouth shrimps. The mean value of r found was .81. In this dia- gram, the deviations of the organ whose value is fixed *Proc. Roy. Soc., li. p. 2, 1892. 78 CORRELATED VARIATIONS. are measured along the ordinates, they varying on an average between the extremes of + 3.19 and 3.01, whilst the mean deviations of the associated organ are + 2 *1 -1 -9 2 -1 + 2 + 2 -1 -S piQ. 16. Correlation between post-spinous carapace length and total carapace length of shrimp. measured along the abscissae, they varying on an aver- age from + 2.92 to 2.17. The crosses in the dia- gram indicate the values obtained when the carapace length was fixed, and the circles those when the post- CORRELATED VARIATIONS. 79 spinous portion was fixed. The line drawn through them indicates the ratio .81. Every point should theoretically lie on it, and it will be seen that they do, as a matter of fact, lie very closely around it. This correlation constant of .81 was obtained by Professor Weldon for shrimps collected at a particular locality, viz., Plymouth. Similar determinations were also made for shrimps obtained from other localities, with the following results : In Plymouth r .81 (1000 individuals examined) In Southport .85 ( 800 " " ) In Roscoff .80 ( 500 " " ) In Sheerness .85 ( 380 " " ) In Helder .83 ( 300 " " ) The approach to identity between these values is very striking, the differences appearing to be within the probable error of each determination. There seems a reasonable ground for assuming, therefore, that the degree of correlation between the two particular organs measured is practically constant in all the races exam- ined. The correlation between other organs was also estimated, but this was in each instance very much slighter, and in the case of the telson and sixth abdom- inal tergum, it was negative. Considering the degree of independence of these organs, as shown by the small- ness of their correlation constants, the similarity be- tween the values for the two local races is probably as close as could be expected. Hence, as both the organs measured and the samples of shrimps examined were chosen by chance, any result which holds for all these organs through all these races may be reasonably ex- pected to prove generally true of all organs through the whole species. 80 COERELATED VARIATIONS. CARAPACE LENGTH CARAPACE LENGTH TELSON AND AND TERGUM VI. AND TELSON. TERGUM VI. Plymouth, .09 .18 -.11 Southport, .06 .14 -.09 Professor Weldon points out that the above results lead us to hope that it may be possible to determine con- stants for any species of animal which would " give an altogether new kind of knowledge of the physiological connection between the various organs of animals, while a study of those relations which remain constant through large groups of species would give an idea, at- tainable at present in no other way, of the functional correlations between various organs which have led to the establishment of the great sub-divisions of the ani- mal kingdom." In a subsequent paper,* Professor Weldon deter- mined no less than 23 different correlation constants, between various pairs of organs in 1000 adult female crabs (Carcinus mcenas), collected in Plymouth Sound, and in another 1000 collected in the Bay of Naples. He found that there was as a rule a remarkable degree of correspondence between the values of r derived from an investigation of the same pair of organs in the two races examined. There were in some cases consider- able differences between the values, it is true, but Wel- don considers that these were in no case sufficient to justify the assertion that the degree of correlation is really different in the two cases. It should be men- tioned, however, that Professor Pearson f does not *Proc. Roy. Soc., liv. p. 318. f Phil. Trans. 1896, A. p. 267. CORRELATED VARIATIONS. 81 agree with this conclusion, but thinks that the differ- ences observed are too large to justify such an assump- tion. At Professor Weldon's suggestion, Mr. E. Warren * undertook similar measurements on 2300 specimens of another crab, Portunus depurator, also obtained from Plymouth. The accompanying table gives the results obtained by Warren, and some of those obtained by Weldon: C. MCENAS, PORTUNUS C. MOENAS, PLYMOUTH DEPURATOR, ORGANS. NAPLES RACK. RACE. PLYMOUTH. Total breadth and frontal breadth, .08 .10 .14 R. antero-lateral, .66 .65 .67 " " R. dentary margin, .50 .55 .56 Frontal breadth and R. antero-lat., .29 .24 .30 " " R. dentary margin, .23 .18 .03 L. dentary margin, .26 .20 .01 R. antero-lateral and L. antero-lat., .76 .78 .86 R. dentary margin, .71 .78 .80 " " L. dentary margin, .60 .70 .74 On glancing through this table, it will be seen that, with two exceptions, the values for the two races of C. mcenas differ from one another nearly as much as they do from the constants of Portunus, an animal be- longing to a different genus. It is probable, however, that the larger differences in this latter animal do indi- cate real differences in the correlation constants, asso- ciated, perhaps, with changes in habit or environment. For example, it is conceivable that a crab which swims might find it advantageous to be more symmetrical than one which only crawls between the tide marks. Por- tunus does swim to a certain extent, and we see that the correlation of the two sides of its body is greater than in the essentially shore-living Carcinus. *Proc. Roy. Soc., Ix. p. 221. 82 COREELATED VAEIATIONS. Another most laborious research undertaken at Pro- fessor Weldon's suggestion is that of H. Thompson,* on the correlation of certain external parts of the prawn, Palcemon serratus. Twenty-two measurements were made on 1000 adult females, and from these the value of Galton's function was calculated for 56 pairs of organs. As might be expected, the degree of corre- lation was highest between the paired organs; e. g., .94 between the right and left squames. Also there was a strong correlation between the terga of adjacent ab- dominal segments, their values ranging between .58 and .71. Of other recent work on correlation, that by Miss Lee and Professor Pearson f may be briefly alluded to. This consists in a comparison of measurements on cer- tain long bones of about 40 male and 25 female skele- tons of the Aino race (a primitive tribe dwelling in Japan), with corresponding measurements of 50 male and 50 female skeletons of the modern French race. It was found that the transition from the uncivilised to the civilised condition is accompanied by well-marked changes in the sexual relationships; primitive man and woman being more nearly equal in size, variability, and correlation than highly civilised man and woman. Civilised man has gained in size on woman, but this has been accompanied by a relative loss in variability and the correlation of parts. The general result of in- creased civilisation is to increase the absolute size and amount of variation. In females, also, the degree of correlation is increased, but in males this remains sta- * Proc. Roy. Soc. , Iv. p. 234. fProc. Roy. Soc., Ixi. p. 343. CORRELATED VARIATIONS. 83 tionary. It is therefore impossible to say that civilised woman is nearer to the primitive type than civilised man, for while civilised man differs more from the primitive type than civilised woman, so far as absolute size is concerned, he has made only about half her progress in variation, and hardly any progress in corre- lation. The absolute amount of correlation is very high, as the following figures show: HALES. FEMALES. ORGANS. AINO. FRENCH. AINO. FRENCH. Femur and tibia, .83 .81 .85 .89 humerus, .86 .84 .87 .87 radius, .79 .74 .70 .78 Clavicle and humerus, .44 .63 Humerus and radius, .78 .85 .74 .85 Tibia and fibula, .89 .96 .97 .98 Humerus and ulna, .77 .77 .75 .86 Radius and ulna, .98 .88 .98 .92 Here we see that the Galtonian constant, r, was in most instances above .8. Between the tibia and fibula it averaged .95, and between radius and ulna .94, so that in these cases the correlation was almost absolute. In the Naquada race investigated by Warren * the correlation between the lengths of the long bones in males (as measured in about 60 skeletons) was distinctly higher than for Aino males, but in females (measured in about 90 skeletons) it was either the same or was lower than in Aino females. Probably correlation is to some extent affected by sex in most animals. Thus Duncker f determined the cor- relation coefficients of 40 pairs of measurements in male * Vide Phil. Trans. 1898, B. p. 178. f Wissenschaf tliche Meeresuntersuchungen aus der biologischen Anstalt auf Helgoland, Bd. iii. p. 351. 84 CORRELATED VARIATIONS. and female flounders (Pleuronectes flesus), and the values obtained showed that the correlation was affected by sex in 17 out of the 40 instances. The coefficients were greater in the male than in the female fish in 11 instances, and in the female than in the male in 6 in- stances. Several of the pairs of bilateral homologous measurements (such as the numbers of rays in the right and left pectoral and ventral fins) showed distinctly lower correlation constants than were shown by the corresponding pairs of measurements in the sym- metrical fish Acerina cernua and Coitus gdbis. This was doubtless due to their possessing slight differences of function in the asymmetrical fish. It should be mentioned that Pearson, "Warren, and Duncker employed a somewhat modified and improved formula for determining these correlation constants,* as compared with that originally suggested by Galton. G. O. Yule,f and also Pearson and Filon,$ have recently shown that the correlation can be determined in the case of skew variation, as well as of normal variation. All these results may be taken to show that every part and organ of the body is correlated with every other part in a greater or less degree, though such cor- relation may sometimes be of the negative order. The immense importance in evolutionary processes of such correlation, whereby when one organ becomes modified by the action of an agency such as Natural Selection, others are modified also, is sufficiently, obvious to need no discussion. *Phil. Trans., 1896, A. p. 264. fProc. Roy. Soc., Ix. p. 477, 1897. {Phil. Trans., 1898, A. p. 229. CORRELATED VARIATIONS. 85 Besides these instances in which the degree of corre- lation can be expressed in numerical equivalents, there remain other cases in which such expression is difficult or impossible. In these it must accordingly be de- fined in general terms. Darwin has collected a large number of such cases in his " Animals and Plants," * but it is not necessary to reproduce more than a few of the most striking of them here. For instance, Teget- mier has stated that young pigeons of all breeds which when mature have white, yellow, silver, blue, or dun- coloured plumage, are born almost naked; whereas pigeons of other colours are clothed with plenty of down. Darwin himself has noticed that in feather- footed pigeons, not only does the exterior surface sup- port a row of long feathers, like wing feathers, but the very same digits which in the wing are completely united by skin become partially united by skin in the feet. Again, Polish fowls have a large tuft of feathers on their heads, and their skulls are perforated by numerous holes. That this deficiency of bone is in some way connected with the tuft of feathers is clear from the fact of tufted ducks and geese likewise having per- forated skulls. Constitutional peculiarities are some- times correlated with colour in a most curious and inter- esting manner. For instance, Beddoe has shown that a relation exists between liability to consumption and the colour of the hair, eyes, and skin. As regards ani- mals, white terriers suffer most from distemper, white chickens from a parasitic worm in their tracheae, white pigs from scorching in the -sun, and white cattle from flies. Again, all the hogs in Virginia, excepting those *Chap. xxv. 86 COEEELATED VAEIATIONS. of a black colour, suffer severely from eating the root of Lachnanthes tinctoria. Similarly, buckwheat when in flower is highly injurious to white or white-spotted pigs, if they are exposed to the heat of the sun, but quite in- nocuous to black pigs. These few instances suffice to show how widespread and apparently capricious may be the range of correla- tion. Until careful observations have been made, ac- companied when possible by measurements, one can never on a priori grounds assume that there is no cor- relation between particular organs or parts of an organ- ism, and that an agency acting on one part may not at the same time be thereby indirectly modifying another. In their effects on the modification of species, prob- ably by far the most important cases of correlation are those in which the reproductive system is concerned. Until recently comparatively little attention was paid to such phenomena, and probably, even now, they are far from being estimated at their true value by many biologists. Perhaps the reason of this lies in the fact that the physiological condition of an organism, or its relative degree of sexual compatibility with other or- ganisms of its own and of different species, is so exceedingly difficult, if not impossible, to estimate. Thus the degree of fertility cannot be tested in more than one or two instances with each individual or- ganism at least in the higher animals and so the desired information can only be acquired by carry- ing out most lengthy and laborious series of observa- tions. There are sufficient data at our disposal, how- ever, to indicate that the reproductive system is no less subject to variation than any other part of the CORRELATED VARIATIONS. 8? organism, and indeed is probably a good deal more so. Supposing that the quality of fertility is correlated with some particular character or characters more than it is with other characters, then it follows that more in- dividuals bearing the character in question will be born and propagate their kind, and so, in course of time, the whole race will be modified in this direction. This prin- ciple has been termed by its discoverer, Professor Pear- son, " Reproductive " or " Genetic Selection." Its existence as a real factor in evolution depends on the validity of the assumption that the characteristic of fertility is inherited. That this is so, Professor Pear- son, in conjunction with Miss Alice Lee and Mr. L. Bramley-Moore,* has recently proved by a most labori- ous research on inheritance in man and in the thor- oughbred race-horse. Their results show that fertility is undoubtedly inherited from mother to daughter, and also from father to son. It was also found that a woman's fertility is as highly correlated with that of her paternal as with that of her maternal grandmother. In other words, the latent character fertility in the woman is transmitted through the male line, and with an intensity which approximates to that required by the law of ancestral heredity. Again, it was deduced that fecundity in the brood-mare is inherited from dam to mare, and also from grand-dam to mare through the dam. Also the latent quality of fecundity in the brood- mare is inherited through the sire, and by the stallion from his sire. In these latter two cases, the degree of inheritance approaches fairly closely to that required by Galton's law of ancestral heredity, but, in the two for- *Proc. Roy. Soc.,lxiv. p. 163, 1899. 88 CORRELATED VARIATIONS. mer, it is much less. As, therefore, fertility is proved to be inherited in man, and fecundity in the horse, it is probable that both these characters are inherited in all classes of life. The importance of this theory of Genetic Selection will, perhaps, be better realised by quoting a concrete case concerning man, this being the only one in which statistics are at present available. Working on data con- cerning 4000 families, principally of the Anglo-Saxon race, and 1842 families of the Danish race, Professor Pearson * determined that there is a sensible correla- tion (about .18) between fertility and height in mothers of daughters. Supposing genetic selection to have been unchecked by natural selection, say for forty genera- tions, the mean height of women would have been raised about 3J inches. A factor which would alter stature by about three inches in 1000 years is clearly capable of producing very considerable results in the long periods during which evolution may be supposed to have been at work. The importance of the influence of genetic selection in the case of man is also shown by the fact that, as proved by these statistics, less than a quar- ter of one generation, by reason of their fertility, pro- duce more than half of the next generation. Correla- tion between fertility and any mental or physical char- acteristic must therefore work a progressive change. For example, arguing from the class fertility statistics which have been determined among the population of Copenhagen, it is gathered that the artisan class pro- duce a larger proportion of children than the profes- sional classes, as their gross fertility is greater, and *Proc. Roy. Soc., lix. p. 301, 1896. CORRELATED VARIATIONS. 83 their marriage rate is so much higher. This increased fertility is somewhat counteracted by their greater death rate, but it would nevertheless appear that the population will ultimately reproduce itself from the artisan classes. Definite evidence of evolution under natural condi- tions as the result of genetic selection has not been ob- tained, but this may be simply because it has never occurred to anybody to look for it. Professor Pear- son * has, however, shown the existence of a highly interesting and important relationship between fer- tility and morphological characters in certain plants. He counted the number of stigmatic bands on the 4443 seed-capsules obtained from 176 Shirley poppies grow- ing in a single garden, and found the following fre- quencies : Bands, 5678 9 10 11 12 13 14 15 16 17 18 19 Frequency, 1 11 32 56 148 363 628 925 954 709 397 155 51 12 1 Here we see that the 12 and 13 band forms were the most common, the 11 and 14 ones less so, the 10 and 15 still less, and so on. To his surprise^ Professor Pear- son found that whilst the commonest or type capsules contained a very large number of seeds, the 11 and 14 band forms contained distinctly less, and the 10 and 15 ones very few seeds indeed, whilst the capsules with very few or very many bands contained practically no seeds. A repetition of these observations on the wild poppy gave a very similar result, and this was likewise the case with the seed capsules of" a number of plants of Nigella Hispanica. The distribution of the seg- mentation on 3212 capsules was as follows: * " Grammar of Science," Ed. ii. p. 443. 90 CORRELATED VARIATIONS. No. of Segments, 234 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Frequency, 10 7 20 303 412 534 1552 223 59 35 43 6 6 2 In this plant Pearson found the 8 segment capsules to be highly fertile, whilst in the 10, 11, and 12 segment capsules he could find hardly any seed at all. Six and 7 segment capsules were only moderately fertile, and those with 5 segments or less were practically sterile. These experiments Professor Pearson holds to illustrate a very important law, namely, " Fertility is not uni- formly distributed among all individuals, but for stable races there is a strong tendency for the character of maximum fertility to become one with the character which is the type." It follows, therefore, if this prin- ciple is generally true, that stable races are very largely the product of the typical or most frequently occurring members, and not of all the individual members in pro- portion to their numbers. There would seem to be a constant tendency to keep the type uniform and limit the variability in either direction as much as possible. Doubtless under changed conditions of environment, the relative fertility might also become changed, and in consequence a gradual evolution result. A similar relationship between fertility and type form has been noticed by Davenport * in one of the Hydromedusse, Pseudoclytia pentata. This organism differs from all other Hydromedusse in that it normally has five radial canals, instead of four. A. G. Mayer t has examined the variation of the species, and obtained the frequencies given in the subjoined table. From these values we gather that the four and six canal forms *Biometrika, i. p. 255, 1902. \ Science Bulletin of the Brooklyn Museum, vol. i. COERELATED VARIATIONS. 91 are comparatively common, whilst other abnormalities are rare. Typically one gonad (or reproductive organ) occurs on each radial canal, but on an average about one in seven of them fails to develop. In atypical forms, however, as can be gathered from the table, the proportion failing to develop is found to be larger and larger, the further the departure of the individual from the type. In such forms as depart from the normal radial symmetry, even if they still possess five rays, the partial sterility is very much increased. PER CENT. OF GONADS FAILING TO DEVELOP. NUMBER OP FREQUENCY OF IRREGULAR CANALS. OCCURRENCE. ALL INDIVIDUALS. INDIVIDUALS. 2 1 3 8 25.0 44.5 4 56 21.8 31.0 5 860 15.6 37.0 6 64 18.6 31.0 7 6 21.5 43.0 8 1 A very interesting case of variation described by Bate- son * may perhaps be ascribed to the action of genetic selection, though there is no direct evidence to warrant the assumption. It concerns the Peppered moth, Amphidasys betularia. A striking black variety of this insect, A. doubledayaria, was first met with as a rarity in 1840-50. Since then its numbers have gradually in- creased, till in 1870 about equal numbers of the pure type and of its variety occurred at Monmouth, whilst a few years later the typical form had entirely disap- peared. At Chester none but black forms have been met with for many years. In the south of England, however, the typical form is still alone present. Bate- * Science Progress, vol. vi. p. 557, 1897. 92 CORRELATED VARIATIONS. son suggests that this gradual replacement of a type by its variety is probably due to success in the struggle for existence of this particular dark strain, but it may equally well be accounted for by supposing that it is the result of a greater fertility. Intermediate strains are not unknown, for in Belgium it seems clear that one has succeeded in establishing itself, and in England it is probable that they were once more common than they are now. Intermediate forms are said to be plentiful also in the Rhenish Provinces and Westphalia, and the same is true of the black forms. Bateson says there is no doubt that the black variety existed at an early stage in the transformation, side by side with the light one. The course of events has not been that the insects of each successive district have become more and more tinged with black, till they culminated in A. double- dayaria, but rather that this variety, or less often one of the intermediate forms, spread into or at least ap- peared in the area, and either coexists with the type or has replaced it. The few breeding experiments thus far made on the moth show that there is an imperfect blending of type and variety. Steinert raised from a black wild female a brood of 75 typical and 90 varietal forms, but there were no really intermediate ones, though two of the examples classed as betularia were darker than the nor- mal. That this female had paired with a typical male the progeny thus resembling either the one parent or the other, but not both is shown by the following ex j periment, also recorded by Bateson.* W. H. B. Fletcher tied out a black female, which had been reared * Science Progress, vol. vii. p. 53, 1898. CORRELATED VARIATIONS. 93 in captivity. This was at Worthing, where the typical male form is the only one known, so that the brood ob- tained were most certainly produced from a cross of the two varieties. The offspring were sharply divided into 10 male and 8 female betularidj and 6 male and 5 fe- male doubledayaria. Conversely to the modification of a community in one direction by reason of the increased productiveness of certain of its members, we may experience modifica- tion in the opposite direction by reason of a decreased productiveness. For instance, Dr. Beddoe * states that there is a good deal of evidence as to the greater liability of blonds than of brunets to certain classes of disease. At least this is so in America, as has been shown by Baxter, f Thus it would appear that the blonds have less chance than the brunets of contribut- ing their due proportion to the next generation, and so they must be relatively diminishing in numbers. That this is the case is supported by the fact that of Ameri- cans accepted for service in the army a greater propor- tion were brunets than of the English, Irish, and Germans accepted. Thus: Among the Americans were 66 light and 34 dark complexioned " English " 70 " " 30 " " Irish " 70 " " 30 ' " Germans " 69 " " 29 " The fact that most species are to some extent mu- tually sterile, whilst their hybrid offspring are almost invariably so, proves that the physiological condition of * Science Progress, vol. v. p. 384, 1896. f "Medical Statistics of the Provost-Marshal-General's Bureau," Washington, 1875. 94 CORRELATED VARIATIONS. the reproductive system is closely correlated with the physiological condition of the organism taken as a whole. Recent physiological research has taught us that most organs in the body, in addition to their more obvious functions, have an internal secretion which passes into the circulation of the body, and there exerts some important, but unexplained, influence on the gen- eral metabolism of the tissues. Deprivation of such internal secretion, by extirpation of one of these organs, may speedily upset the normal working of many or most of the other tissues of the body, and ultimately re- sult in death. Every organ and tissue of the body probably reacts on every other organ, and modifies its physiological condition, and thereby may ultimately produce structural changes. The reproductive system is apparently more sensitive than most other organs, and hence is very readily affected by any changes in the condition of the organism as a whole. For instance, it is well known that most animals refuse to breed in con- finement, though they can be kept for many years in a condition of perfect health. The changed conditions of life must therefore have acted on the organism as a whole, so as to modify certain of its internal secretions, and these, reacting on the reproductive system, have brought about the observed sterility. If a considerable change in conditions of life produces complete sterility, it seems highly probable that slight changes in such conditions may produce a partial sterility, or a differ- ential fertility. Thus organisms in a state of nature, if exposed to a change of climate, the result of migra- tion or stress of weather, or to a change in their food, may have the physiological condition of their repro- CORRELATED VARIATIONS. 95 ductive system slightly altered, whereby the principles of Genetic Selection and in the case of plants of Physiological Selection, may become effective, and modify or split up the species. Just as the condition of the organism as a whole may modify that of the reproductive system, so may the con- dition of the reproductive system modify that of the organism. The difference between the spirit and ap- pearance of castrated animals and th#t of normal ani- mals is sufficiently well known to need no remark. Such differences must be due in large part to the lack of internal secretion from the organs of reproduction. Striking as is the influence of the reproductive organs on the physiological condition of an animal, that upon the morphological structure is even more noteworthy. That castrated male animals fail to develop their sec- ondary sexual characteristics, is notorious. Thus, if the operation be performed upon a young cock, he never crows again; the comb, wattles, and spurs do not grow to their full size, and the hackles assume an inter- mediate appearance between true hackles and the feathers of the hen. Conversely, it is well known that a large number of female birds, such as fowls, various pheasants, partridges, pea-hens, and ducks, when old or diseased, or when operated upon, assume many or all of the secondary male characters of their species. Water- ton gives a curious case of a hen which had ceased lay- ing, and had assumed the plumage, voice, spurs, and warlike disposition of the cock. Again, the females of two kinds of deer, when old, have been known to ac- quire horns.* "Animals and Plants," vol. ii. p. 26. 96 CORRELATED VARIATIONS. Analogous or Parallel Variation. This term has been used by Darwin to indicate that similar characters occasionally make their appearance in several varieties or races descended from the same species, and more rarely in the offspring of widely distinct species.* It is unnecessary to make more than very brief mention of this subject, because, as Darwin points out, the majority of observed cases such as the occasional appearance of black wing bars in the various breeds of pigeon, and of stripes on the legs of the ass and of various races of the horse are evidently due to reversion. The others are probably mere coincidences, and of no scientific value. Among these latter, Darwin mentions the fact that many trees belonging to quite different orders have pro- duced pendulous and f astigate varieties. A multitude of plants have yielded varieties with deeply cut leaves. Several varieties of melon resemble other species in im- portant characters. Thus one variety has fruit so like, both externally and internally, the fruit of the cucum- ber, as hardly to be distinguished from it. In animals, again, we find feather-footed races of the fowl, pigeon, and canary bird. Horses of the most different races present the same range of colour. Many sub-varieties of the pigeon have reversed and somewhat lengthened feathers on the back parts of the head. In connection with this subject of parallel variation, Walsh's " Law of Equable Variability " f may be men- tioned. This states that " if any given character is very variable in one species of a group, it will tend to be variable in allied species ; and if any given character * " Animals and Plants," vol. ii. p. 340. f Proc. Entomolog. Soc. Philadelphia, p. 213, 1863. CORRELATED VARIATIONS. 97 is perfectly constant in one species of a group, it will tend to be constant in allied species." The general truth of this law seems highly probable on the face of it, because most allied species have presumably split off from their common ancestor at no very remote period, and so would still resemble each other more or less closely in respect of variability, just as they do in re- spect of morphological structure. The results, already quoted, of Weldon and Warren for the correlation of various organs in Carcinus and Portunus, afford direct experimental evidence in support of this law. In conclusion, it may not be out of place to make one or two brief remarks as to the general bearing of the evidence which has been adduced in these three chap- ters concerning the " facts of variation." I believe that they include most of the more important and more recent contributions to our knowledge of the subject, especially those in which the results have been ex- pressed in exact numerical terms. It may very likely be objected that insufficient importance has been at- tached to the kind of information which Darwin col- lected so thoroughly and in such profusion in his work on " Variation in Animals and Plants under Domestica- tion." The reason of this is twofold. In the first place it seemed unnecessary to refer at great length and with much frequency to data with which most seri- ous students of Biology must be already conversant: whilst in the second place, comparatively little infor- mation of this kind has, as far as I am aware, been recorded in scientific journals since Darwin's time. It has, indeed, been recognised that the facts of variation 98 CORRELATED VARIATIONS. attain a much higher and more permanent value in pro- portion as they are expressed in exact numerical terms. To say that some organism or part of an organism is more variable than another is very unsatisfactory, com- pared with the statement that the variabilities of cer- tain characters in the one are of such and such values, and in the other, of certain other values. From such data as these we can compare the variabilities of all the variants exactly, both with each other and with any other variants, and determine what relation, if any, they bear to their systematic importance. We can tell if the variations obey the normal law of error, or if they are asymmetrical in their distribution. In this latter case, we may be able to discover the existence of a tendency to divergence or splitting up of a species, in its very earliest stages. Repetition of our observa- tions at some future period would thus become a sub- ject of especial interest, as we might in such a case hope to detect some change both in the mean values and in the distribution of the variations, indicating that the evolutionary process was still progressing, and the di- rection in which the progress was being made. Fur- ther, as we shall see in a subsequent chapter, by deter- mining the average characters of groups of individuals which have been subjected for some period to the struggle for existence, and comparing them with the characters of other individuals which have not been ex- posed to such a struggle, or by comparing the characters of individuals, which owing to the severity of the strug- gle for existence had been actually eliminated, with the characters of the survivors, we are able to obtain CORRELATED VARIATIONS. 99 numerical proof of the action of Natural Selection. Still again, by comparing the characters of individuals with those of their parents and of their offspring, we are enabled to work out with exactness the origin and transmission of such characters, and so elucidate the Laws of Heredity. PART II. THE CAUSES OF VAKIATIOK CHAPTEE IV. BLASTOGENIC VARIATIONS. The ultimate cause of blastogenic variationEffect of staleness and of comparative maturity of sex-cells on the characters of organisms Amphimixis Identical twins Transplantation of ova in the rabbit Law of Ancestral Heredity in man and in the Basset hound Regression towards mediocrity Exclusive inheritance. Homotyposis. AKGUING from his theory of the continuity of the germ-plasm, first suggested in 1883,* Weismann came to the conclusion that acquired characters were not transmissible. Such acquired characters are due to the direct influence of the environment upon the body tissues of an organism, or are variations of somatogenic origin. Opposed to these are variations due directly to certain peculiarities of the germ-plasm 2 or variations of blastoger^ic origin, which are, on the contrary, hereditary or transmissible. Thus, according to Weismann, all variations are, in respect of their origin, sharply divisi- ble into these two groups, whilst in respect of trans- missibility they are equally distinct. In the present *Vide " Essays on Heredity," Oxford, 1889, p. 71. 101 102 BLASTOGENIC VARIATIONS. chapter these blastogeriic, genetic, or germinal varia- tions will be discussed, whilst somatogenic variations will be treated of later. The cause of hereditary variation Weismann ascribes to the direct effect of external influences on the so- called biophors and determinants of the germ-plasm.* These biophors Weismann defines as " bearers of vitality/ 7 or the smallest units of protoplasm which exhibit the primary forces of assimilation^anct metab- olism, growth, and multiplication by fission. All pro- toplasm, both the nucleus and body of cells, is made up of these units. They are the bearers of the qualities or characters of the cells. Determinants, on the other hand, are groups of biophors, and are the particles of germ-plasm corresponding to and determining the cells or groups of cells which are independently variable from the germ onwards. The reaction of the germ-plasm to external influ- ences is primarily one of nutrition. The biophors and determinants are supposed to be subject to continual changes of composition during their almost uninter- rupted growth, and these very minute fluctuations are the primary cause of the greater deviations of the de- terminants, which are finally observed in the form of individual variations. The growing determinants must originally differ to some slight extent in the composi- tion of their biophors, as otherwise inequalities of nutri- tion could never effect any transformation, but could only alter their rate of growth. Slight as are these deviations in the determinants effected by inequalities of nutrition, they are nevertheless of great significance, * The Germ-Plasm," London, 1893, p. 415. BLASTOGENIC VARIATIONS. 103 as, by a process of accumulation, they form the material from which the visible individual variations are pro- duced by means of " amphimixis." Amphimixis is that form of reproduction which is found in all the higher organisms, and which consists in the mingling of two individuals or their germs; i. e., the so-called sexual reproduction. The term is also ap- plied to a similar phenomenon occurring amongst uni- cellular organisms, i. e., to conjugation. In this case reproduction is not a necessary or even usual concomi- tant, but takes place independently. To amphimixis Weismann attributes the constant occurrence of indi- vidual variability, although he recognises that it is not the primary cause of this variability; but rather the process furnishes an inexhaustible supply of fresh com- binations of individual variations. Thus the germ- plasm of a new individual produced by amphimixis never receives more than half the ids of each parent, and these are differently selected and arranged in each case. By an id, it may be remarked, "Weismann means a group of determinants which contains all the deter- mining elements of the species, though in a manner peculiar to the individual. Blastogenic variations are thus, according to Weis- mann, primarily dependent on two chief factors: (1) Inequalities of nutrition acting on the individual con- stituents of the germ-plasm; (2) Amphimixis. It behoves us to examine these two factors more closely, and see how far they are supported by experi- mental evidence. It is, on the face of it, impossible to put Weismann's hypothesis of the reaction of determi- nants to inequalities of nutrition to a practical test, but 104 BLASTOGENIC VAEIATIONS. we can at least enquire into what is known about the in- fluence of nutrition on the germ-plasm as a whole. In fact, we can see how far the parental plasms are indi- vidually capable of being affected by changes of nutri- tion, so as, on subsequent mingling in sexual union, to give rise to appreciable changes in the resulting off- spring. Important as this subject is, the direct experi- mental evidence available upon it is distinctly meagre. It is, for instance, probable that the children of a father whose tissues, and therefore his sex-cells, are saturated with alcohol or the products of some disease, are smaller and less well formed than those of normal parents, but there are no satisfactory data to support it. Similar evidence with reference to the female sex- cells is obviously not available, as any effects produced on offspring would probably in chief part arise during embryonic development, or after, and not before, fer- tilisation. The evidence obtained as to the influence of nutri- tion on the evolution of sex is only indirectly related to the problem under discussion. Most of it goes to show that increased feeding of young organisms tends to- wards the production of a larger proportion of females,* and hence, as male and female sex-cells cannot be considered entirely equivalent, it follows that an effect is produced on the germ-plasm. Yet there is no evidence to show that the offspring of fe- males which arose in spite of bad feeding differ in any way from those of females produced in consequence of good feeding. Nevertheless it seems probable that, as nutrition has some influence in determining the sexual * Vide Geddes and Thomson's " Evolution of Sex," p. 41, 1889. BLASTOGENIC VARIATIONS. 105 character of the germ-plasm in a developing animal after fertilisation, it may also have some influence if it be brought to bear on the parental germ-plasms before fertilisation. It seems likely, in fact, that a highly nourished ovum, as compared with one less favourably conditioned, will tend rather to a female than a male development. Evidence bearing more directly on the question at issue has recently been obtained by the author, in a research on the effect of staleness of the sex-cells on development.* The method of experiment was to keep the ova or spermatozoa, or both, of the sea-urchin Strongylocentrotus lividus for varying numbers of hours in sea water before permitting fertilisation, and after eight days' development to measure the length of the larvae and see if they differed in size from normal larvae. As other observations on larvae obtained from these artificial fertilisations will be referred to later, the experimental procedure adopted may be briefly in- dicated. It consisted in shaking pieces of the ovaries and testes of several ripe specimens of the Echinoid in small jars of water, and mixing portions of their con- tents either immediately, or after a given number of hours. The mixed solutions were allowed to stand for an hour, and were then poured into large jars holding from 2 to 4 litres of sea-water. These were placed in a tank of running water, whereby the temperature was kept nearly constant, it varying less than a degree dur- ing twenty-four hours and not more than two degrees during the whole course of the experiment. The fer- tilised ova were allowed to develop for eight days, as by *Proc. Roy. Soc., vol. Ixv. p. 350, 1899. 106 BLASTOGENIC VAEIATIONS. that time the arms of the larvae or plutei have attained their maximum length, whilst the body has practically ceased growing. The larvae were then killed by adding corrosive sublimate. They were collected and pre- served in 80 per cent, alcohol, and were subsequently mounted in glycerine and measured under the micro- FIGS. 17 AND 18. Larvae of Strongylocentrotus limdus. scope by means of a micrometer eyepiece. The body- length, AB, was always measured in 50 different larvae, and a mean taken. In many cases also the anal arm- length, AC, and sometimes also the oral arm-length, AD were measured as well, and these measurements calcu- lated as percentages on the body length. In order to determine the effect of staleness of the sex-cells on the size of the larvae, five series of measure- ments had to be made, viz., (1) of the normal larvae ob- tained from the fresh ova fertilised with the fresh BLASTOGENIC VARIATIONS. 107 sperm, (2) those from stale ova and stale sperm, (3) from stale ova and fresh sperm obtained from another freshly opened Echinoid, (4) from fresh ova and stale sperm, (5) and lastly, from the ova and sperm of the freshly opened Echinoids. It was, of course, impos- sible to get an exact basis of comparison for the larvae obtained from one stale and one fresh sexual product. The best possible was to take a mean between the size of the original normal larvae, and that of the larvae ob- tained from the fresh sexual products used f o* fertilis- ing the stale products. The larvae obtained with both sexual products stale are, of course, accurately com- parable with the original normal larvae. In the ac- companying table an example is given of the mean per- centage differences in the size of the larvae obtained with fresh and stale products, from the original nor- mal larvae in the one case, and from the mean between the original and fresh normal larvae in the other two FERTILIZATION MADE AFTER CONDITION OF SEXUAL CELLS. 9 HRS. 24 HRS. 33 HRS. 45 HRS. Stale 9, stale $ -0.2 +1.9 + 1.1 - 1.9 Fresh 9, stale $ +7.1 +3.7 +10.9 + 1.5 Stale $ , fresh $ -2.8 -3.0 + 2.0 -15.9 cases. In this experiment, which was the most com- plete made, the fertilisations were performed after keeping the sexual products for respectively 9, 24, 33, and 45 hours. It will be seen that the larvae obtained when both sexual cells were stale were of practically 108 BLASTOGENIC VAEIATIONS. the same size as when they were fresh, even after they had been kept 33 or 45 hours. The larvae from fresh ova and stale sperm, on the other hand, were in each case distinctly larger than the normal, they differing on an average by '+ 5.8 per cent., whilst those from stale ova and fresh sperm were distinctly smaller, they differ- ing by 4.9 per cent. In one of these latter observa- tions there was, for some unknown reason, a slight in- crease in size, but there can be no doubt that on the whole the tendency was towards diminution. On tak- ing means of all the values obtained in this and in other similar experiments, it was found that as an average of eight observations, the stale $ stale $ larvae were di- minished by .7 per cent. in size; as an average of eleven observations, the fresh 9 stale $ larvae were increased by 4.0 per cent., and as an average of ten observations, the stale $ fresh $ larvae were diminished by 6.9 per cent. There can be no doubt, therefore, that variations in the degree of freshness of the sexual cells, that is to say, in the comparative state of nutrition of the germ- plasm as a whole, do have a very appreciable effect upon the size of the subsequently developing larvae. It is to be particularly noticed that the effect produced differs entirely according to the sex-cells acted upon, and hence affords distinct evidence of the possibility that different portions of the same sex-cell may also react differently to one and the same change of nutrition. Perhaps a more convincing proof of the influence of the nutritional condition of the sex-cells on the offspring they produce is afforded by certain other observations on these larvae. On two separate occasions * series of *Phil. Trans. 1895, B. p. 585, and 1898, B. p. 483. BLASTOGENIC VARIATIONS. 109 artificial fertilisations were carried out at short inter- vals over periods of several months, and the larvse al- lowed to develop under conditions which were probably nearly constant, except as regards temperature. For this varying factor a correction could be easily applied. In spite of the constancy of environmental conditions, however, the size of the larvss showed very marked Apr. Apr. M y June July July Aug. Sept. Oct. Nov. Mov. Dec. Jan. Ifet 35th 19th 12th 6th .80th, 23cd 16th JOth Scd jWih ;30th 13th FIG. 19. Seasonal variation in size of larvae. variations. The range of these variations may be gathered from the accompanying diagram. Here the ordinates represent the mean body lengths of the larvse in micrometer eyepiece units (of which 152.3 are equivalent to 1 mm.), and the abscissas the time of year at which the fertilisations were made. It will be seen that in April and May the larvae were on an average about 34 units in length, but that then they began steadily to dwindle down in size, so that in June they were about 31 units, and in July and August only about 110 BLASTOGENIC VARIATIONS. 28 units. From this minimum they then rose rapidly, so that in September they were about 32 units, in Octo- ber and November 34 units, and in December and Janu- ary 35 units in length. The extreme variations were from 36.80 to 24.49, the larvae of this latter length, ob- tained on July 9, being no less than 33.4 per cent, smaller than the former. This extraordinary seasonal variation in the size of the larvae is probably very closely, if not entirely, de- pendent on changes in the maturity and nutritional con- dition of the sexual products. Thus, of the specimens of Strongylocentrotus obtained in the winter months, al- most every individual contained ripe sexual products in large quantities, whilst of those obtained in the sum- mer months, not more than about one in four yielded any ova or sperm at all on shaking the ovaries and testes in water, and occasionally twenty or more indi- viduals were opened before any ripe sperm was ob- tained. Again, the best of the specimens obtained in the summer months did not contain nearly so much of the ripe sexual products as they did in the winter. That this effect of season on the condition of the sex- cells is more far-reaching than is implied in a mere diminution of size in the resulting offspring, is proved by some observations on the crossing of this species of sea-urchin with another species, viz., 8 phcer echinus granularis. Hybrids between Splicer echinus ova and Strongylocentrotus sperm can probably be obtained, though it may be only after several attempts, at all times of the year. It was found, however, that their structure was by no means constant.* The majority of *Phil. Trans. 1898, B. p. 470. BLASTOGENIC VARIATIONS. Ill the hybrids obtained in May, June, and July were of the almost pure Sphcerechinus type, of which an example is given in Fig. 20; but about a third of them or less were of the intermediate or Strongylocentrotus type, of which an example is given in Fig. 21. In November, on the other hand, only about a sixth of the hybrid larvae were of the Sphcerechinus or maternal type, and FIG. 20. Sphcerechinus FIG. 21. Sphcerechinus $ Stron- larva. gylocentrotus $ larva. five-sixths of the paternal type. Finally, in December and January, all the larvae were of the paternal type. These so-called paternal larvae in almost all cases showed obvious traces of their hybrid origin, but they were evidently much more inclined to the Strongylocen- trotus than to the Sphcerechinus type. Combining this series of observations wi'th that just recorded, we therefore find that in the summer months, when the Strongylocentrotus sperm is in a condition of minimum maturity, the Sphcerechinus $ Strongylo- centrotus $ hybrids are chiefly of the Sphcerechinus type. As, however, the maturity of the sperm in- creases, it is able to transform first a portion and then 112 BLASTOGENIC VAKIATIONS. the whole of the hybrid larvae from the Sphcerechinus to the Strongylocentrotus type. A repetition of these crossing experiments in a subsequent year * confirmed the conclusion that the summer hybrids were more in- clined to the Sphcerechinus type than the winter ones, though on this occasion they were only very rarely found to approach to the pure Splicer echinus type. The reciprocal cross of Strongylocentrotus ova with Splicer echinus sperm illustrates still another way in which the sex-cells may be affected by changes in ma- turity and nutrition. Thus during April, May, and June a fair number of the ova were cross-fertilised, though no plutei were obtained: but in July and August some 47 per cent, of the ova were fertilised, and 29 per cent, of them survived to the eight days pluteus stage. In November and December, on the other hand, with one exception, not only were no plutei obtained, but, as a rule, not a single ovum was cross-fertilised. In other words, the Strongylocentrotus $ Sphcerechinus $ hy- brid is only formed at a time when the Strongylocen- trotus ova have reached their minimum maturity. The observations made upon these sea-urchin larvse thus afford conclusive evidence that changes in the con- ditions of nutrition of the sex-cells produced by keeping them in sea water may affect the size of the larvae both in a positive and negative direction, whilst changes of nutrition dependent on season may produce a much more considerable effect on size, and may in some in- stances so alter the nature of the germ-plasm as to give rise to most marked variations of structure in the re- sulting hybrid offspring, and in other instances largely * Arch, f . Entwickelungsmechanik, Bd. ix. p. 464, 1900. BLASTOGENIC VARIATIONS. 113 to abolish the normal resistance to cross-fertilisa- tion. If differences of nutrition in the parental germ- plasms as a whole can produce such profound effects on the offspring to which they give rise, then, supposing it is possible that various individual portions of the germ- plasm are capable of being more or less independently affected by inequalities of nutrition, there seems no rea- son to doubt that they may show a similar reaction, and so give rise to variations in the individual parts of an organism which they represent or " determine." Experiments upon higher organisms are, of course, very much more difficult to carry out than those upon sea-urchin larvae, but nevertheless Professor Ewart * has been able to bring to a successful conclusion some experiments upon rabbits. Thus he found " that if a well-matured rabbit doe is prematurely (i. e., some time before ovulation is due) fertilised by a buck of a differ- ent strain, the young take after the sire; when the fer- tilisation takes place at the usual time, some of the young resemble the buck, some the doe, whilst some present new characters or reproduce, more or less accu- rately, one or more of the ancestors. When, however, the mating is delayed for about thirty hours beyond the normal time, all the young, as a rule, resemble the doe. It may hence be inferred that in mammals, as in echinoderms, the characters of the offspring are related to the condition of the germ-cells at the moment of conjugation, the offspring resulting from the union of equally ripe germ-cells differing from the offspring de- veloped from the conjugation of ripe and unripe germ- * Presidential Address before the British Association. Vide Na- ture, vol. Ixiv, p. 482, 1901. 114 BLASTOGENIC VAEIATIONS. cells, and still more from the union of fresh and over- ripe germ-cells." Upon plants, as far as I am aware, no direct experiments have been made, but some obser- vations of De Vries * upon (Enothera Lamarckiana bear closely upon the matter. Some seeds of this plant had been kept for 5J years before sowing, and it was then found that only about 1 per cent, of them germi- nated, instead of the usual 14 per cent, or so. Of the seedlings obtained, however, about 40 per cent, were " mutations," whilst the proportion of mutations ob- tained from fresh seeds was only 1 to 5 per cent. Weismann's second factor in the production of varia- tions is the so-called amphimixis, or sexual reproduction in multicellular organisms, and conjugation in unicel- lular. That this is one of the chief causes of variation was maintained by W. K. Brooks f some years ago. Basing his theory on Darwin's hypothesis of Pan- genesis, he considered that as every " gemmule " of the spermatozoon united with that particle of the ovum which is destined to give rise in the offspring to the cell which corresponds to the one which produced the germ or gemmule, then such a cell will be a hybrid, and will therefore tend to vary. In his opinion the egg-cell is the conservative principle which controls the transmis- sion of purely racial or specific characters, whereas the sperm cell is the progressive element which causes variation. To what extent are we justified in assuming that this process of amphimixis does furnish an inexhaustible * Die Mutationstheorie," p. 360. f " The Law of Heredity: A Study of the Cause of Variation and the Origin of Living Organisms," Baltimore, 1883. BLASTOOENIC VARIATIONS. 115 supply of fresh combinations of individual variations, as Weismann maintains? Exact numerical evidence upon the point is, indeed, almost entirely wanting, but one's own everyday experience is all in favour of its validity. Thus one knows that animals of the same litter, which during embryonic development must have been exposed to very nearly equal environmental conditions, differ al- most as much from each other as from animals of for- mer litters, and in many cases not very much less than from animals in the litters of entirely different parents. ISTow, as this phenomenon is one of almost universal occurrence, it cannot be maintained that the observed variations may be brought about by chance differences of environmental conditions acting during development. They must obviously be in chief part the result of dif- ferences in the individual sex-cells from which the off- spring took their rise. So important is this conclusion that it was enunciated by Victor Hensen as a funda- mental law of amphigonic heredity. This has been thus worded by Weismann:* " The individual is deter- mined at the time of fertilisation, or, in other words, the individuality of an organism results from the fact that the germ-plasm is composed of the paternal and maternal ids which are brought together in the egg- cell." Very interesting evidence in favour of this law is furnished by cases of identical human twins. It has long been known that whilst the larger number of twins show no greater resemblance to each other than do children of the same parents born consecutively, a * " Germ-plasm," p. 253. 116 BLASTOGENIC VARIATIONS. certain proportion exhibit a most striking resemblance, which, although not perfect, is much closer than has ever been observed in children born successively. These Weismann speaks of as " identical " twins. He says there is every reason to suppose that such twins are derived from a single ovum and spermatozoon, whilst dissimilar twins are derived from two ova, which must, of course, have been fertilised by two different sperma- tozoa. If this is actually the case, it furnishes a proof that heredity is potentially decided at the time of ferti- lisation. Interesting cases of identical twins have been re- corded by Galton in his book on " Inquiries into Hu- man Faculty." With reference to disease, for instance, it was found that both twins were apt to sicken at the same time in 9 out of the 35 cases collected. Either their illnesses were non-contagious, or, if contagious, the twins caught them simultaneously. The mental and moral resemblance between the twins was just as close as the physical. An instance cited by Dr. J. Moreau * well illustrates this. His case was one of twin brothers who had been confined on account of monomania. They were physically so alike as to be easily mistaken for one another, and as regards their moral condition they had exactly the same dominant ideas; they both con- sidered themselves subject to the same imaginary per- secutions ; they both had hallucinations of hearing ; both were melancholy and morose. Unfortunately, Galton did not obtain any exact anthropometric data. Weismann has obtained one series of measurements, however, viz., for twin brothers * "Psychologic Morbide," Paris, 1859, p. 172. BLASTOGENIC VARIATIONS. 117 seventeen years of age.* The following are the measurements made: PER CENT. TWIN A. TWIN B. DIFFERENCE. Stature, 172cm. 170cm. 1.2 Left arm, 74 74 0.0 Right arm 71 74 4.1 Left upper arm, .... 27 27.5 1.8 Forearm 27 26 3.8 These slight differences are probably due to the effect of external influences acting during the course of de- velopment, or are somatic, as distinguished from blas- togenic, variations. In a series of measurements on twin brothers obtained by the author, the resemblance was very much closer, the difference in no case reaching even 1 per cent. These brothers, aged twenty-three, were extraordi- narily alike in physiognomy, and, moreover, they had both suffered at the same times from the same diseases, viz., bronchitis, measles, chicken-pox, mumps, and influenza. The slightly smaller one of the two had had a rather more severe attack of bronchitis than his brother, when a year and a half old, and so, perhaps, but for this, the physical resemblance would have been even closer: PER CENT. TWIN A. TWIN B. DIFFERENCE. Standing height 173.00cm. 172.67cm. -.19 Sitting height (from seat of chair), 88.03 87.87 .18 Span of arms, .... 179.90 179.88 -.01 Length of right mid-finger (from metacarpo-phalangeal joint), 10.99 10.98 .09 Span of hand, .... 21.33 21.30 -.14 Length of skull (occipital protuber- ance to base of nose), . . 18.52 18.42 .54 Maximum breadth of skull, . . 15.13 15.01 .80 f "Germ-plasm," p. 253. 118 BLASTOGENIC VARIATIONS. The finger prints, though bearing some resemblance, were nevertheless easily distinguishable. Hence in this case Galton's finger-print method would serve for an identification, whilst Bertillon's anthropometric sys- tem would be useless. The very slight modifications produced in these twins by the action of environment during growth is prob- ably explained by the fact that they had always been brought up together, and so exposed to practically the same conditions all their lives. In another series of measurements upon twin brothers (aged twelve), the differences observed were somewhat greater, and the facial resemblance was like- wise not quite so marked as in the previous case. The boys had both had scarlet fever, chicken-pox, and measles at the same times. Of the measurements given, it will be seen that the span of arms and length of fore- arm showed the greatest differences. As before, one twin was slightly smaller than the other in respect of every measurement made : PER CENT. TWIN A. TWIN B. DIFFERENCE. Standing height 143.22cm. 142.62cm. - .42 Sitting height 72.93 72.87 - .08 Span of arms 150.73 148.81 -1.27 Elbow to tip of mid-finger, . . . 39.59 38.93 1.67 Length of right mid-finger, . . . 10.06 10.03 .30 Circumference of head (over occi- pital protuberance and 2.5 cm. above eyebrows), .... 50.23 49.97 .52 From the vegetable kingdom Weismann also quotes an instance in support of Hensen's law. Thus, in re- spect of two species of Oxalis, " the flowers of the dif- ferent hybrids were by no means quite similar, but BLASTOGENIC VARIATIONS. 119 three principal forms could be distinguished according to the combination of colours in the flowers. The flowers of the same hybrid, however, resembled each other in the most minute details. One plant, for in- stance, had violet petals of a rather pinker tint than those of one of the parent species, and all the petals were strongly tinged with red on one and the same lateral margin. As far as I could observe, all the flowers were similarly coloured on this stock. On an- other stock, all the sepals had brown rims, and on a third there was a narrow dark orange-coloured band in the centre of each flower. In these cases, therefore, the combination of the colours of the parents which ap- peared in the petals of the hybrids must have been de- cided at the time of fertilisation." * That the influence which the maternal fluids canj exert on an embryo during intra-uterine development is f at best very slight seems at first sight to be proved by the experiments of Heape f on the transplantation of rabbits' ova. In the first successful experiment, two segmenting ova were obtained from an Angora doe rab-^ bit which had been fertilised by an Angora buck thirty- two hours previously, and were immediately transferred to the upper end of the fallopian tube of a Belgian hare rabbit which had been fertilised three hours before by a buck of the same breed as herself. In due course this Belgian hare doe gave birth to six young. Four of these resembled herself and her mate, but the other two were undoubted Angoras. The Angora young were characterised by the possession of the long silky * " Germ-plasm," p. 256. fProc. Roy. Soc., xlviii. p. 457, 1890. 120 BLASTOGENIC VARIATIONS. hair peculiar to the breed, and were true albinoes, like their Angora parents. They also possessed the char- acteristic habit of slowly swaying the head from side to side when they looked at one. Both of the Angora young were born bigger and stronger than any of the other young, and they all along maintained their su- premacy in this direction. Heape could observe no sign in the Angora young of any Belgian hare strain, and the Belgian hare young showed no likeness to their foster-brothers. In a subsequent paper,* Heape records another suc- cessful experiment. In this a Belgian hare doe was covered by a Belgian hare buck, and shortly after the segmenting ova obtained from a Dutch doe which had been covered by a Dutch buck twenty-four hours pre- viously were transferred to her fallopian tube. The Belgian hare doe gave birth to seven young, of which five Avere Belgian hares, and two very irregularly marked Dutch. It was found, however, on putting the same Dutch buck which had been used in this experi- ment to a thoroughbred Dutch doe, most, if not all, of the litter resulting were as badly marked as the Dutch foster-children. Hence it is not necessary to suppose that the foster-mother was the cause of the irregularity. From this and other evidence Heape considers he is justified in concluding that the uterine foster-mother exerts no modifying influence upon her foster-children, in so far as can be tested by the examination of a single generation. Romanes t has however remarked, that inasmuch as rabbits, when crossed in the ordinary way, *Proc. Roy. Soc., Ixii. p. 178, 1897. f "Darwin and after Darwin," vol. ii. p. 146. BLASTOGENIC VARIATIONS. 121 never throw intermediate characters, the result of Heape's experiment is without significance, as far as it bears on the inheritance of acquired characters. Heape considers that Romanes is mistaken in this view, for he has himself obtained experimental evidence to show that some of the young obtained by crossing are of an intermediate character. It is nevertheless true that in the majority of cases the young are apparently pure bred of one type or the other, and hence the value which ought to attach to Heape's experiments, so far as they relate to the production of somatic variations by change of environmental conditions during embryonic develop- ment, is probably not very great. The evidence so far available seems to render it highly probable, therefore, that the major part of the variation exhibited by organisms is of blastogenic rather than somatic origin. It is due more to dif- ferences in the germ cells than to external influences acting during ontogeny. If it be found possible to col- lect considerable series of anthropometric measure- ments of identical twins, and to compare them, as re- gards variability, with similar measurements on dissimi- lar twins, then we may hope to obtain some adequate conception as to what proportion of the variation ex- hibited by adult individuals is due to external influences acting during pre- and post-natal existence. Also by comparing the variability of dissimilar twins with that of members of families produced in the normal manner of one at a birth, we may hope to determine what changes, if any, are produced by the slight differences in the maternal fluids which must doubtless exist dur- ing the development of the different offspring. 122 BLASTOGENIC VARIATIONS. The variations of offspring are therefore largely pro- duced by_ the mingling of % dissimilar parental germ- plasms, so that the offspring do not closely resemble either each_ other t or their parents. But there must clearly be a relation of some sort between them. As to the extent of this relation, we are chiefly indebted for our knowledge to the labours of Mr. Francis Galton. In his work on " Natural Inheritance," he has analysed a very large number of anthropometric data which were collected by himself specially for the purpose. The most important of them consist of records of the stature, eye-colour, artistic faculty, and condition of health of the various members of some 150 distinct families, extending over three or more generations. Arguing from these data, he concluded that on an aver- age each parent contributed to the characters of his or her offspring J of their amount, or both parents to- gether contributed a half; whilst each grandparent con- tributed T \, or the four grandparents together J, and so on; but he considered his data insufficient to warrant him in extending the sequence to more distant genera- tions. Some years later, Galton obtained other more fav- ourable data.* These enabled him to ascertain the con- tributions of ancestors to offspring with much greater exactness, and warranted him in formulating a Law of Ancestral Heredity, which there is some reason for thinking may prove to be universally applicable to bisexual descent. The data consisted in long series 'of records of the colours of a large pedigree stock of Basset hounds, extending through many successive *Proc. Roy. Soc., vol. Ixi. p. 401, 1897. BLASTOGENIC VARIATIONS. 123 generations. These records were preserved by Sir Everett Millais, who had originated the stock of hounds some twenty years ago. The Bassets are dwarf blood- hounds, of two and only two recognised varieties of colour. They are either white with large blotches ranging between red and yellow, or they may, in addi- tion, be marked with more or less black. In the former case they are technically known as " lemon and white," and in the latter case as " tricolour." Transitional cases between these two forms are very rare. No less than 817 hounds of known colour, all descended from parents of known colour, were available as material. In 567 out of these 817 the colours of all four grand- parents were known, and in 188 cases the colours of all eight great-grandparents were known as well. It was found that 79 per cent, of the parents of tricolour hounds were tricolour, whilst 56 per cent, of the parents of lemon and white hounds were tricolour. Hence from these values the contributions of unknown ancestors could easily be calculated. Working from these numerous data, Galton was able to confirm en- tirely his previous conclusions regarding heredity, and extend them in the direction then hinted at. He proved that the two parents do contribute between them one- half or (0.5) of the total heritage of the offspring: whilst the four grandparents contribute one-quarter, or (0.5) 2 : the eight great-grandparents one-eighth, or (0.5) 3 , and so on. Thus the sum of the ancestral con- tributions is expressed by the series [ (0.5) + (0.5) 2 ;-f- (0.5) 3 + . . . etc.], which, being equal to 1, accounts for the whole heritage. The same statement may be put in a different form, by saying that each parent con- 124 BLASTOGENIC VAEIATIONS. tributes on an average one-quarter, or (0.5) 2 , each grandparent one-sixteenth, or (0.5) 4 , and so on, or that the occupier of each ancestral place in the nth degree, whatever be the value of n, contributes (0.5) 2n of the heritage. It is unnecessary to quote the numerical details ad- duced by Galton, but two final results may be men- tioned just to show how close was the approximation be- tween fact and theory. Thus in one series 387 tri- colour offspring were obtained from certain parents of known colour, themselves the offspring of parents of known colour. On the law of heredity, the number of tricolour offspring should have been 391. In the other series, the colours of the great-grandparents were known in addition, and in this case the approximation was even closer. One hundred and eighty-one tri- colour offspring were obtained, as against the calculated number of 180. Galton points out that there is nothing in his statis- tical law to contradict the generally accepted view that the chief, if not the sole, line of descent runs from germ to germ, and not from person to person. The person on the whole may be accepted as a fair representative of the germ, and so statistical laws which apply to per- sons would apply to germs also. Now the law is strictly consonant with the ' observed binary subdivi- sions of the germ cells, and the concomitant extrusion and loss of one-half of the several contributions from each of the two parents to the germ cell of the off- spring. Galton's law has been shown by Pearson * to be even *Proc. Roy. Soc., Ixii. p. 386, BLASTOGENIC VARIATIONS. 125 more fundamental and far-reaching than its author claimed it to be. Thus he says, " If Mr. Galton's law can be firmly established, it is a complete solution, at any rate to a first approximation, of the whole problem of heredity." Professor Pearson points out that by means of it we are enabled to find the coefficients of cor- relation between an individual and any individual an- cestor, and that these coefficients in their turn will suffice to determine all inheritance, whether direct or collateral. As regards the relation of this law of heredity to variations produced by amphimixis, it is necessary to emphasize one fact, viz., that it concerns only the j average contributions of ancestors to offspring, and not \ the absolute contributions. Within what limits the contributions of each parent and grandparent to the heritage of a child may vary, nothing whatsoever is known. It is possible that they may be very wide in- deed, and everyday experience tends to give colour to this view. How trite is the expression that such and such a child is the " image " of his father or mother, whilst instances are no less common in which it is diffi- cult, if not impossible, to trace any distinct resem- blance between parent and child. Such cases as these, even if they could be substantiated by exact physical measurements, would in reality prove but little. It would be impossible to make accurate comparisons of all the tissues and organs of the body, and of the cells composing them, and it might be that these unexamined and unexaminable portions of the organism in reality possessed a very close correlation with the correspond- ing parental tissues. The average degree of correla- 126 BLASTOGENIC VARIATIONS. tion for all the tissues in the body might thus be just as great as for an individual who to all appearances closely resembled his parents, but in this case chiefly in exter- nal characters and not internal. If amphimixis be so largely responsible for the varia- tions observed in offspring, what, then, are the rules which govern the amount and range of these variations ? For the answer to this question we are again primarily indebted to the labours of Mr. Galton. He set himself to determine the exact average relationships between the two parents and their offspring. It might be thought that this was so simple and obvious as to ren- der it waste of time to put it to the test of experiment. It might be thought, in fact, that the average char- acters of offspring are a mean between those of the parents. But this is far from being the case. As Gal- ton first showed, by means of extensive observations on the size of sweet-pea seeds obtained from plants which had been grown from seeds of known size, the average characters of the offspring show a considerable regres- sion towards the mean characters of the race. That is to say, in the present instance, the size of the filial seeds was, on an average, more mediocre than that of the parent seeds. By means of his data, above referred to, concerning the stature of families, Galton was able not only to thoroughly substantiate this phenomenon of re- gression, but to calculate with some degree of exactness the actual amount of regression occurring between various kinsmen. For example, he found that if parents were sorted into groups according to their stature, then the stature of their sons, on an average, deviated only two-thirds as much from the mean stature 1 UN BLASTOGENIC VARIATIONS. 127 of the general population as theirs did. Thus, if the mid-parental stature (the average between the statures of the man and the transmuted stature of the woman) be 72 inches, or 3| inches greater than the mean stature of the whole population, then the average stature of their sons will be only 3f X f = 2^ inches greater than the mean. If the mid-parental stature be 66 inches, or 2J inches less than the mean stature, then the average filial stature will be 66| inches, or only 1 inches less than the general mean stature. In addition to calculat- ing the regression between parents and sons, and grand- parents and grandsons, Galton calculated it for col- lateral relationships, as between uncles and nephews, and brothers and brothers. Many of the data recorded by Galton in his " Natural Inheritance " were worked over again by Pearson in his memoir on " ^Regression, Heredity, and Pan- mixia/' * and various improvements in statistical methods suggested. The mathematical measure of re- gression, or coefficient of regression, he defined to be " the ratio of mean deviation of offspring of selected parents from the mean of all the offspring to the devia- tion of the selected parents from the mean of all the parents." It is to be noticed that according to this definition, the deviation of the offspring ought to be measured from the mean of the offspring of the general population, and not of the whole population, both parents and offspring, for thereby factors such as secu- lar natural selection and reproductive selection are allowed for. In this memoir, when discussing coefficients of re- *Phil. Trans. 1896, A. p. 253. 128 BLASTOGENIC VAKIATIONS. gression, Pearson came to the conclusion that there was not at that time sufficient ground for forming any defi- nite conclusion as to the manner in which lineal and col- lateral heredity were related. Thus it did not appear necessary to him that the coefficient of the former should be half that of the latter, as Galton had sup- posed. On attacking the problem a second time,* however, Pearson succeeded in proving that they were connected, according to a mathematically ascertainable relationship, so that, starting from Galton's law of heredity, it was possible to calculate the coefficients of regression or correlation between an individual and any of his kinsmen, either direct or collateral. Thus Pear- son calculated the coefficient of regression between mid- parent and son to be .6, or somewhat less than that found by Galton. Between a single parent and a son, it would, therefore, be .3. Between grandparent and grandson it was .15, between great-grandparent and great-grandson .075, and so on. Between brothers it was .4, or considerably less than the coefficient found by Galton. Nevertheless this value confirms Galton's conclusion that brothers are more closely related to each other by blood than are fathers and sons. It may be pointed out that in a stable population the coefficients of regression and of correlation between an individual and an ancestor are one and the same thing. If, however, the population is not stable, so that the variability of the offspring differs from the variability of the parents, then these coefficients also differ slightly. The importance of this extension of Galton's law can- not be rated too highly, for by its means the whole *Proc. Roy. Soc., Ixii. p. 386. BLASTOGENIC VAEIATIONS. 129 theory of heredity is rendered simple, straightforward, and luminous. Pearson points out that we no longer need to know the characters of parents, grandparents, etc., to test the law, for any single relationship, near or far, direct or collateral, will bring its cjuota of evidence for or against it. Galton's principle of " Eegression towards Medi-J ocrity " has been spoken of occasionally as if it werd| something abnormal and unexpected; something, in- deed, unexplained and inexplicable. It is clearly noth- ing of the kind, however, but only what might readily be deduced from his law of Ancestral Heredity, sup- posing that this and this alone were known to us. Thus we have seen that offspring derive certain por- tions of their heritage from their grandparents and more remote ancestors, and as these are likely to be, on an average, more mediocre than their parents, they water down the parental characters transmitted to the offspring. Supposing all the grandparents and more remote ancestors of any given parents were absolutely mediocre, then, as the offspring receive only half their heritage from these parents, they would exhibit their characters in only half strength, or the coefficient of regression would be .5, and not .6. The reason why the regression reaches, on an average, the higher figure, is of course that the grandparents and other ancestors are not, as a rule, absolutely mediocre. They possess the characters exhibited by the parents, though in a diminished degree. Grandparents regress on parents to just the same extent as offspring do. It may perhaps be enquired how it is that, if off- spring are on an average more mediocre than their 130 BLASTOGENIC VAEIATIONS. parents, the variability of the race does not become less and less for each generation, and so finally be reduced to zero. Why this is not the case is perhaps most readily grasped by examining a statistical table of the relations between parent and offspring in respect of some character. The table here given is reduced from a larger one given by Galton in his " Natural Inherit- ance " (p. 208), and represents the numbers of adult children of various heights born of 205 mid-parents of various heights. For instance, we see that the 17 mid- M fiS egi ti 5 w NUMBER OF ADULT CHILDREN OP HEIGHTS *i ? n *% ill Si* below 63 63-65 65-67 67-69 69-71 71-73 73 & over MEDIAl HEIGH1 CHILDB 5 73 & over 1 3 17 71-73 4 8 18 22 10 70.6 63 69-71 1 18 23 90 42 15- 69.1 82 67-69 4 37 92 131 126 37 3 67.9 32 65-67 4 22 37 49 29 3 67.0 6 below 65 3 14 9 8 3 65.6 parents 71 to 73 inches in height had between them 62 children, of whom the most frequently occurring were also 71 to 73 inches, or the same height as their parents. Children of from 69 to 71 inches were, however, nearly as frequent, so that, on the whole, it is obvious that the stature of the children was more mediocre than that of the parents. The median, or middle height of all these 62 values, was, in fact, only 70.6 inches, or 1.4 inch less than the median height of the mid-parents. The median of the children of mid-parents 69 to 71 inches in height, was .9 inch less than their median; whilst in the children of mid-parents varying from 67 to 69 inches it was only .1 inch less, for the median of these BLASTOGENIC VARIATIONS. 131 mid-parents, viz., 68 inches, was very nearly that of the whole population, and so obviously the filial height could undergo no regression, but would be practically the same value. The values in this table thus illus- trate the existence of regression, but they also indicate that the offspring produced by these mid-parents are, as a whole, no less variable than they themselves are. The offspring are, in fact, more variable, as a mid-parental stature, being the mean of two parental statures, is ob- viously, on an average, less variable than either stature individually. Thus the mid-parents vary roughly be- tween about 74 and 64 inches, but the children between 75 and 62 inches. This table therefore teaches us that though the children are, on an average, more mediocre than their parents, yet the general variability of the race is not diminished. The reason why the variability remains undiminished may be seen by studying the com- ponents of the vertical columns of the table. For ex- ample, with reference to children 71 to 73 inches in height, we see that mid-parents of 71 inches and up- wards contribute proportionately more of these tall children than do any other mid-parents, but still mid- parents of 69 to 71 inches contribute (proportionately) a good many, and parents of 67 to 69 inches no small number. Even mid-parents of 65 to 67 inches con- tribute a very minute number of these children, who are thus no less than 6 inches taller than their parents. By these several contributions, therefore, the number of tall and similarly of other children is kept up to the same level in each generation. One may accord- ingly sum up the contents of this table as follows: Tall parents have many tall children, a moderate number 132 BLASTOGENIC VARIATIONS. of medium children, and a very small number of short children; medium parents have many medium children, and moderate numbers of tall and short children; short parents have many short children, a moderate number of medium children, and a very small number of tall children. As was first pointed out by Mr. Galton,* characters such as stature and eye-colour offer a distinct contrast in their hereditary behaviour, for whilst " Parents of different statures usually transmit a blended heritage to their children, parents of different Eye-colours usually transmit an alternative heritage. . . If one parent has a light Eye-colour and the other a dark Eye- colour, some of the children will, as a rule, be light and the rest dark: they will seldom be medium eye-col- oured." Thus eye-colour is a case of more or less ex- clusive inheritance, or inheritance by the offspring of the whole of the character of one parent and none of that of the other. Obviously, therefore, for such in- heritance the law of ancestral heredity does not at first sight appear to hold. Supposing the offspring are equally likely to take after one parent or the other, then the coefficient of regression between parent and offspring will be .5, instead of .3, as in the case of blended inheritance : between grandparent and offspring it will be .25, instead of .15, and so on. Nevertheless it is probable that the law of ancestral heredity is just as true for one form of inheritance as for the other, only from the mere fact of the inheritance being ex- clusive, it does not reveal itself in the same way. Sup- posing there is no alternative between, for instance, * " Natural Inheritance," p. 139. BLASTOGENIC VARIATIONS. 133 light and dark eye-colour, or in animals, light and dark coat colour, then we can imagine the various light and dark heritages from each of the parents, grandparents, and more remote ancestors to be summated and bal- anced against each other in each individual, and, which- ever reach the higher figure, be it by ever so little an amount, to be thereby enabled to originate exclusively the character to which they correspond. The constitu- tion of the germ-plasm of a light or dark-coloured ani- mal cannot be inferred, therefore, unless the colour of its ancestors be known, for it may contain anything from just over half right up to the full number possible of " light " or " dark " determinants. That exclusive inheritance obeys the law of heredity in the same manner as blended inheritance seems to be shown by the fact that the striking proof of the law re- ferred to above was obtained by Galton for an almost exclusively inherited character, viz., coat colour in Bas- set hounds. Galton believed also that his data for eye-colour in man afforded considerable support to the law in question. Professor Pearson,* however, seems to regard exclusive inheritance as distinct from blended inheritance, and to look upon it as governed by a Law of Reversion, and not by the law of ancestral heredity. Arguing from this law, we may suppose that 25 per cent, of the offspring show the full character of either parent, 2 5 per cent, of them exhibit or revert to the full character of each of the four grandparents, -f f per cent, revert to the full character of each of the eight great-grandparents, and so on. However, the whole *Proc. Roy. Soc., Ixvi. p. 140; also " Grammar of Science," pp. 486-496; also Phil. Trans. Roy. Soc. 1901, A. p. 79. 134 BLASTOGENIC VARIATIONS. question is fraught with doubt and difficulty, and greatly lacks the experimental data necessary for put- ting it to an adequate practical test. Hence, for the present, it is best to regard the matter as still sub judice. The law of heredity and regression which we may consider to have been substantiated for sweet peas, for Basset hounds, and for man, may justifiably be ex- tended to other organisms as well. It seems probable that it is, in fact, to use Mr. Galton's words, " univer- sally applicable to bisexual descent." As already stated, however, it should always be borne in mind that it deals only with average amounts, and not absolute amounts. Though the law is of great value in the breeding of pedigree stock, it is not exact enough to enable one to predict with any accuracy the characters of the unborn offspring of known parents, or even of known grandparents as well as parents. Nevertheless the mere fact of such a law of average inheritance being demonstrable, indicates triumphantly how fun- damentally important is the constitution of the germ- plasm in the determination of variations. We see, then, that Weismann's conclusions as to the chief factors concerned in the origin of blastogenic variations are in the main confirmed, so far as it is pos- sible to put them to an experimental test. It is of course impossible to obtain experimental proof of the actual existence of biophors, determinants, and ids in the germ-plasm, but it is scarcely possible to account for the facts of heredity without making some such hypothesis. The Law of Ancestral Heredity proves that all ancestors, however remote, are able to leave BLASTOGENIC VARIATIONS. 135 the impress of their individuality upon the sex-cells, in diminishing proportion according to their remoteness. Such a fact can only be accounted for by assuming the existence, in the germ-plasm, of definite units carrying definite characters, and the regular halving in the aver- age strength or amount of such characters during the reducing division of the nuclear matter of the sex-cells which precedes each act of sexual reproduction. It was stated above that Professor Pearson calcu- lated the correlation constant between brothers to be .4. In a remarkable memoir recently published Pro- fessor Pearson * and his collaborators have collected together all the statistics at present available, as to fra- ternal correlation in the horse, the dog, and in daphnia, as well as in man, and have found the mean of the con- stants deduced from 19 series of observations to be .4479. Individual constants range from .6934 down to .1973, but doubtless some of the extreme values in either direction are, for various reasons, invalid. It seems probable, therefore, that fraternal correlation, whether it concerns stature, cephalic index, eye-colour, or longevity in man, or coat colour in the dog and horse, may be taken to fluctuate about a mean value of .4 to .5. The greater part of this memoir, however, concerns correlation in the vegetable kingdom. Professor Pear- son points out that the individual puts forth a number of like organs, such as blood corpuscles, spermatozoa, petals of the flower, leaves of the trees, which are un- differentiated, but that nevertheless there is a consider- able amount of variation among these " undifferenti- * Phil. Trans. 1901, A. p. 285. 136 BLASTOGENIC VARIATIONS. ated like organs/ 7 or " homotypes." It is found, how- ever, that the variability of these like organs in an in- dividual is less than that of similar like organs in all the members of a race (it being as a rule 80 to 90 per cent, as great), and that therefore there is a considerable cor- relation between them. The principle that like organs are correlated, or that the undiiferentiated like organs of individuals have a certain degree of resemblance, Professor Pearson speaks of as Jiomotyposis. Professor Pearson and his collaborators have deter- mined the degree of homotyposis in 22 distinct series, and have determined, for instance, the numbers of leaf- lets on the leaf of the Ash (26 leaves being taken from each of 329 trees), the number of veins in the leaf of the Spanish Chestnut (26 leaves from 204 trees), and in that of the Beech (26 leaves from 100 trees), the prickles on the leaf of the Holly (26 leaves from 299 trees), the stigmatic bands on the seed capsules of poppies (10,435 capsules, taken from 1064 plants), the sori on 8 to 12 fronds of each of 101 Hartstongue ferns, etc. The mean correlation for all the 22 series was .4570, or practically the same value as was obtained for fraternal correlation. The extreme values ranged from .6311 to .1733, but there are numerous causes which will account, at least in part, for these wide de- viations from the average. Supposing that any of the organs measured had undergone a certain amount of differentiation or splitting up in various directions (and this, it must be remembered, is always possible, as there is no real criterion as to whether any given organ is really undiiferentiated, or differentiated), this would generally result in a great reduction in the correlation; BLASTOGENIC VAEIATIONS. 137 whilst a heterogeneity of material, such as a mixture of two different local races, would tend, as a rule, to raise correlation.* Also the environmental factor, and the difficulty of ensuring that all individuals are of the same age, or in the same state of development, must be borne in mind. Professor Pearson therefore considers that he is justified in assuming that the intensity of pure homo- typosis throughout the vegetable kingdom probably lies between A and .5, and as this is the mean value for fraternal correlation, he believes that " heredity is really only a phase of the wider factor of homotyposis." In a criticism of Pearson's conclusions, Bateson t draws attention to the fact that it is difficult or impos- sible to distinguish between chance variation occurring between members of a series, and actual differentiation, which may be present in greater or less degree. He therefore considers that the average value of the homo- typosis coefficient has no significance. However, Pear- son states that the " diversity due to differentiation. . . is the result of dominating factors which can be isolated and described," though he does not attempt this in detail in his present memoir. To what extent he will be able to accomplish it, and so ultimately obtain the true correlation constants of absolutely undifferentiated like organs, remains to be seen. J *L. c., p. 292. fProc. Roy. Soc., Ixix. p. 193. J See also rejoinder by Professor Pearson in Biometrika, i. p. 320, 1902. CHAPTER V. BLASTOGENIC VARIATIONS (Continued). Reversion; commonest in crossed races, as of the pigeon and fowl; its theoretical explanation Prepotency; in the trotting horse and in man; probably due in large part to inbreeding Mendel's Law of Hybridisation, and its range Natural and artificial plant hybrids Animal hybrids Sports; probably of different origin to normal variations Artificial production of monsters Telegony; probably non-existent Parthenogenesis in an Ostracod and in Daphnia Does sexual reproduction induce variability? Relation of varia- bility of individual to variability of race Asexual reproduction in plants Bud-variation. IN the last chapter we saw that the average char- acters of offspring are inherited from their ancestors in accordance with a simple and definite law, but it re- mains for us to discuss several phenomena related to this law, some of which, indeed, appear to afford a par- tial contradiction of it. These are the phenomena of reversion, prepotency, the appearance of sports, and certain cases of hybridism. The variations which show themselves in connection with such phenomena, though doubtless of less importance than those already dis- cussed, are nevertheless in some instances considerable, and of not infrequent occurrence. They therefore merit a fairly full discussion. It is impossible, how- ever, to illustrate this with many exact numerical data, simply because these do not exist. One must as a rule remain content to quote the descriptive evidence of 138 BLASTOGENIC VAEIATIONS. 139 breeders and others, who seldom troubled to substan- tiate their views by measurements and figures. The phenomenon of Reversion or Atavism has long been recognised, not only by agriculturalists and breeders, but also by others who have witnessed its oc- currence in members of the human race. One of the simplest instances of reversion is that of a child or a lower animal resembling a grandparent more closely than its immediate parents. Much more remarkable, however, are those instances in which the resemblance is to a remote ancestor, or to some distant member in a collateral line (supposing, of course, that these be held to be properly substantiated). Cases of reversion are very frequent in respect of secondary sexual characters, as when a son resembles his maternal grandsire more closely than his paternal in some such attribute as a peculiarity of the beard, in the case of man; of the horns, in the case of the bull; and of the hackles or comb in the cock. Also it is well known that certain diseases, such as haemophilia and colour-blindness, are frequently transmitted to male offspring through a woman who herself remains unaffected. For most of our knowledge on the subject of rever- sion we are indebted to the labours of, Charles Darwin, who obtained most valuable experimental evidence him- self, besides collecting from most varied sources the re- sults obtained by others. One of the most striking in- stances he records is that of a pointer bitch,* which pro- duced seven puppies. Four of these were marked with blue and white, which is so unusual a colour with pointers that the animal was thought to have played * "Animals and Plants," ii. p. 8. 140 BLASTOGENIC VARIATIONS, false with the greyhounds, and all but one of the litter were destroyed. Two years later, this young dog was seen by a friend of the owner, and he declared him to be the image of his old pointer bitch, the only blue and white pointer of pure descent he had ever seen. On close enquiry, it was proved that the dog was the great- great- grandson of the bitch, and so, on Galton's law, it should have received only fg- part of its heritage from her. Another even more remarkable instance is that of a calf which was coloured in a very peculiar manner, its legs, belly, and part of the tail being white, and the remainder black. Its great-great-great-great-grand- father was coloured in the same peculiar manner, but all the intermediate offspring were black. Hence the calf reverted in its colour markings to an ancestor from which it should have drawn only ^^ part of its heritage. It is when two distinct races are crossed that the tendency in the offspring to reversion most often de- clares itself. No examples are more striking than those obtained by Darwin in the case of the domestic pigeon. For instance,* he paired a mongrel female Barb-fantail with a mongrel male Barb-spot, neither of these mongrels having the least blue about them. " Nevertheless the offspring from these two mongrels was of exactly the same blue tint as that of the wild rock-pigeon from the Shetland Islands over the whole back and wings; the double black wing bars were equally conspicuous; the tail was exactly alike in all its characters, and the croup was pure white; the head, however, was tinted with a shade of red, evidently de- *L. c. t i. p. 209. BLASTOGENIC VARIATIONS. 141 rived from the Spot, and was of a paler blue than in the rock-pigeon, as was the stomach. So that two black Barbs, a red Spot, and a white Fantail, as the four purely-bred grandparents, produced a bird exhibiting the general blue colour, together with every character- istic mark, the wild Columba livia" Professor J. C. Ewart, in the breeding experiments he has recently beep carrying out at Penycuik,* has ob- tained an equally striking case of reversion in the case of the pigeon. He crossed a pure white Fantail cock with the offspring of an Owl and an Archangel. One of the young of this complex pair had the colouration of the Shetland rock-pigeon, whilst- the other resembled the Indian rock-pigeon in having a blue croup and the front part of the wings chequered. In this second bird there was complete reversion as to colour, and in the first, wherever measurements were possible, there was practically complete reversion also as to form. The tail feathers were twelve in number and showed but the faintest indications of any colour inheritance from their immediate parents. An additional point of interest lay in the fact that in disposition the bird seemed wilder and more shy than the domestic breeds usually are. Many other instances might be quoted from Darwin and others to prove that this tendency to the production of offspring of a blue colour, with the same charac- teristic marks as Columba livia, is present in all the chief domestic races of pigeon. It shows itself more especially when these domestic races are crossed, but may even appear occasionally in the purely bred races, * " The Penycuik Experiments," Edinburgh, 1899. 142 BLASTOGENIC VARIATIONS. Similar phenomena show themselves in other domestic animals besides the pigeon, though they are seldom so striking or so clear. Thus in some cases the wild an- cestor or ancestors are quite unknown, and hence one is debarred from coming to any certain conclusions as to whether reversion is present or not. The Game fowl, however, and probably most other domestic breeds of fowl, may with considerable confidence be de- rived from the jungle fowl, Gallus lankiva. Now purely bred Game, Malay, Cochin, Dorking, Bantam, and Silk fowls may frequently or occasionally be met with, which are almost identical in plumage with the wild Gallus bankiva. The most striking instance ob- tained by Darwin * is one in which a glossy green-black Spanish cock was crossed with a diminutive white Silk hen. Both of these breeds are ancient, and have long been known to breed true. All the offspring from this cross were coal black, and all plainly showed their parentage in having blackish combs and bones; but as the young cocks grew, one became a gorgeous bird, closely resembling the wild G. bankiva, but with the red feathers rather darker. In all but a few details there was the closest resemblance. In recent years a series of interesting observations has been carried out by von Guiata upon mice.f Fifty-five Japanese waltzing mice were crossed with white mice belonging to a race bred by Weismann for eight years, and these crosses were continued through *L. c. ii. p. 253. fBer. Naturf. Ges. zu Freiburg, Bd. x. p. 317, 1898, and Bd. xi. p. 131, 1900. Review by Davenport in Biol. Bulletin, ii. p. 121, 1901. BLASTOGENIC VARIATIONS. 143 seven generations. Japanese waltzing mice are mostly black and white, i. e., piebald, in colour, but their crosses and reciprocal crosses with the albino race yielded a most unexpected result. The whole of the offspring produced were of a gray colour, indistin- guishable in respect either of colour or of size from the common house mouse. The waltzing action was en- tirely wanting, the reversion being apparently com- plete. Heacke had obtained a similar result on cross- ing the same races.* In the third generation, how- ever, the type was broken, for the 44 young produced by the mating of 4 pairs of the reverting gray mice consisted of 8 waltzers (albino, spotted, gray, and black), 11 pure albinos, and 25 gray mice. In the sub- sequent generations, the albinos and also the gray and the spotted mice were found to breed true. Gray mice crossed with white yielded mostly gray offspring, but a certain number of waltzers. Of the occurrence of reversion there can thus be no question. In fact, its appearance in the offspring of crossed races is by no means an infrequent phenomenon. The reversion of hybrids and mongrels to one of their pure parent forms, after an interval of two or more generations, is especially common. Hence it would seem that the act of crossing in itself gives an impulse towards reversion. Why and how this is the case must be more or less a matter of conjecture. Indeed, this is equally true for all the phenomena of reversion, but I think that a brief consideration of certain presumptions regarding the germ-plasm as the bearer of hereditary- characters will show that, after all, we are not dealing *Biol. Central., Bd. xv. p. 44, 1895. 144 BLASTOGENIC VARIATIONS. with anything more mysterious and remarkable than is found in most of the phenomena of nature. Thus tak- ing it for granted that each of the parts of an organism capable of independent variation from the germ on- wards has a definite representative or determinant of some sort in the germ-plasm, what proportion does the mass of the determinants of, say, all the characters which distinguish a pouter or a fantail pigeon from a rock pigeon, bear to the mass of the determinants which represent the species pigeon, as such? Let us suppose that the average total differences between the char- acters of species of the same genus be counted as one unit, what would be the number of units corresponding to differences between the characters of genera, fam- ilies, orders, and so on? No two biologists would judge alike, and of course it is impossible to estimate them; but, for the sake of our argument, let us attempt some sort of rough numerical estimate as to what these dif- ferences might be. Let us assume that, if species on an average differ by one unit in the sum total of char- acters, genera differ by three units, and families by per- haps ten units. Orders might differ by 25 units, classes by 50 units, and phyla by 100 units. Therefore we assume that an individual of one phylum, in the sum total of its characters, is 100 times more different from an individual of another phylum than is one species from another of the same genus. The difference be- tween the highest Vertebrate and the lowest Protophyte would probably be considered to be perhaps ten times greater than this, but let that pass. Let us take it that the sum total of characters represented by any species of pigeon is 100 units, of which the total characters pe- BLASTOGENIC VARIATIONS. 145 culiar to a rock or other pigeon, as such, is one unit. Also let it be granted that the characters separating any variety of the pigeon from the ancestral rock pigeon are of the same value as those separating species of the same genus, namely, one unit. Now in the gradual course of evolution of a domestic variety of pigeon from a rock pigeon, we may assume that the total amount of germ- plasm bearing hereditary characters has remained prac- tically constant, and hence, as one unit of determinants has been added on to the rock pigeon germ-plasm, one must have disappeared. Now did this unit of deter- minants corresponding to the characters of the domes- tic variety of pigeon replace that of the rock pigeon, or was it superimposed on it ? Embryology seems to teach us that once any character is, as it were, laid down in the germ-plasm, it is fixed there, and as a rule only very slowly dwindles away by a process of gradual dilution by subsequent ontogenetic stages. It seems reasonable to assume, therefore, that the determinants are re- placed in proportion to the relative amounts of them present, or that, on an average, -ffo of the replaced unit concern the sum total of hereditary characters which go to constitute the species pigeon, and y^ those peculiar to the species blue rock pigeon. The germ-plasm of a domestic pigeon will therefore be made up of 1 unit of determinants corresponding to the char- acters domestic pigeon, .99 of a unit corresponding to the characters Hue rock pigeon, and 98.01 units cor- responding to the characters species pigeon. It there- fore follows that the hereditary characters of the an- cestral rock pigeon are almost as strongly represented in the germ-plasm of a domestic pigeon as they were 146 BLASTOGENIC VARIATIONS. originally, only that they seldom have an opportunity of showing themselves. They are covered up by the more recently acquired characters, and it is only under ex- ceptional circumstances that they are able to reveal themselves. When, for instance, two distinct races of pigeon, such as a pouter and a f antail, are crossed, then the offspring would on an average receive .5 of a unit of determinants corresponding to each of the special group of characters pouter and f antail, the same .99 of a unit corresponding to the characters blue rock pigeon, and 98.01 units corresponding to the characters species pigeon. If, then, the determinants of pouter and fan- tail do not to any great extent correspond, what wonder is it that they more or less neutralise each other, and allow the blue rock pigeon determinants to gain the upper hand, and show their presence ? This view of the constitution of the germ-plasm may at first sight seem contrary to the law of ancestral heredity, but in reality it is not so. A man may receive a quarter of his hereditary characters from each parent, and a sixteenth from each grandparent, but all except a very minute proportion of these characters are com- mon to all men, they being, in fact, the characters proper to the species Homo sapiens, as such. Instead of a quarter of a unit from each parent, a man in reality receives only a hundredth or a thousandth of a unit of characters peculiar to the parent as such, all the rest being the characters common to all members of the race. Even this minute fraction of a unit does not in any way represent characters acquired by the parent during his life-time, but is itself built up of proportions of peculiar characters received from his parents, grand- BLASTOGENIC VARIATIONS. 147 parents and other ancestors in accordance with the law of heredity. It seems, then, that the sudden reappearance of an- cestral characters ought not to be regarded as a very remarkable phenomenon, but certain other cases of re- version offer a greater difficulty. Thus cases such as that above mentioned of a calf reverting to the colour marking of an ancestor six generations back, if of at all frequent occurrence, are truly remarkable. If of only very infrequent occurrence, however, they may perhaps be ascribed to a mere coincidence, or to like conditions of environment having acted on both ancestor and de- scendant, and produced like results. Prepotency. The phenomenon of prepotency of cer- tain individuals, races, and species in the transmission of their characters is a very common one, and it merits our consideration, in that it is an important factor in the production of variations. As a rule, the offspring of dissimilar parents are in most respects of an inter- mediate character. Frequently, however, they more or less closely resemble one parent in one part, and the other parent in another part. Less seldom they show a much closer resemblance to one parent than to the other, or may apparently resemble one parent in every respect, to the entire exclusion of the other parent. We may here be dealing with true cases of prepotency, or it may be that the characters in question are for some unknown reason unable to blend, and so be neces- sarily transmissible only in toto from one parent to the other. For instance, it is well known that certain do- mestic animals, such as the cat, show only a few well- defined differences of colour marking, such as white, 148 BLASTOQENIC VARIATIONS. black, tabby, and tortoise-shell, and though they breed promiscuously, very seldom throw intermediate colours. Again, Sir R. Heron crossed during many years white, black, brown, and fawn-coloured rabbits, and never once got these colours mingled in the same animal, but often got all four colours in the same litter.* All the off- spring of dissimilarly coloured parents may therefore resemble either parent, or some resemble one and others the other, possibly quite apart from any ques- tion of prepotency. Of undoubted cases of prepotency, but few have been recorded with much detail or exactness. Of those col- lected by Darwin, the most striking is that of a famous black grayhound,f which " invariably got all his pup- pies black, no matter what was the colour of the bitch " ; but this dog " had a preponderance of black in his blood both on the sire and dam side," a point which will be referred to again later. Again, the famous bull Fav- ourite is believed to have had a prepotent influence on the shorthorn race. The male Manx cat appears to be prepotent in transmitting his tailless condition. Pro- fessor EwartJ has recorded a few additional cases of prepotency. Thus a well-known breeder of highly bred ponies used to boast that he had a filly so prepotent through inbreeding that, though she were sent to the best Clydesdale stallion in Scotland, she would throw a colt showing no cart-horse blood, provided always that the Clydesdale was not also the product of inbreeding. Again, Professor Ewart points out that Jews, as a race, * " Animals and Plants," ii, p. 70. \L. c.,ii. p. 40. \ " Penycuik Experiments/' p. xli. BLASTOGENIC VARIATIONS. 149 are strongly prepotent, probably because they are purer bred than other races. A numerical estimate of the frequency with which different grades of prepotency are distributed appears to have been attempted for the first time quite recently by Mr. Galton.* From data given in Wallace's Year Book of American Trotting Horses, he has determined the numbers of offspring of a certain standard, pro- duced by various sires and dams. A standard per- former is a horse which has succeeded in trotting a mile in 2 min. 30 seconds or less, or in pacing (ambling) a mile in 2 min. 25 seconds or less. Data concerning the offspring of 716 sires and 494 dams were available, and the following were the percentage proportions of " standard performers " produced by them. NUMBER OF STANDARD PERFORMERS PRODUCED BT A SINGLE PARENT, SIRE OR DAM. 1 2 3 4 5 6 to 11 to 21 to 31 to 41 to 51 and Total Parents 10 20 30 40 50 above Sires 46 17 10 7 3 9 4 1 1 1 1 100 Dams 50 35 10 3 1 1 - - - - - 100 This table would seem to show that the prepotency of certain sires is enormous, even allowing for the tend- ency of breeders to send the best mares to the best horses. Thus the horse Happy Medium had 92 dis- tinguished offspring, and Electioneer no less than 154. The same results are indicated by the produce of the dams, though the figures are less striking owing to the relative fewness of their offspring. A sire produces * Nature, vol. Iviii. p. 246, 1898. 150 BLASTOGENIC VARIATIONS. 30 foals annually, but a dam only one, hence the pro- duction of respectively 7, 8, and 9 standard performers by three mares is very remarkable. Professor Pear- son,* however, does not accept the high degree of pre- potency which these figures seem to indicate. He points out that the fact of certain sires producing such a preponderance of standard performers is largely due to their exceptional pedigrees. It is also due to the second-rate stallions being given far less chance of pro- ducing performers, in that the mares sent them are often inferior, or past their most intense fecundity, as well as being fewer in number. In discussing the law of heredity in the last chapter, it was tacitly assumed that the heritage from each parent was the same, or that both parents were equi- potent. This does not seem to be necessarily the case, however, as Professor Pearson finds that in man the father is slightly prepotent over the mother for the off- spring of both sexes, f Thus a determination of the coefficient of correlation in respect of stature and of head index, gave the following figures: Coeff. of correlation Father and son (Middle class English) .396 (for stature) " daughter " " " .360 " Mother and son " " " .302 " " daughter " " " .284 " " " son (N. American Indians) .370 (for head index) " daughter " " .300 " " The average correlation between stature of father and of offspring was thus .378, and between that of mother and of offspring .293, or 22.5 per cent. less. The * Nature, vol. Iviii. p. 292. f " Grammar of Science," p. 458. BLASTOGENIC VARIATIONS. 151 average correlation between parent and offspring was thus .335, instead of the theoretical .3. Pearson thinks this high value may be due to assortive mating.* The number of data available for calculating these constants was not very great, so that they cannot be accepted as final, but there seems little doubt of the existence of a small degree of male prepotency. It should be no- ticed, also, that the intensity of the heredity is stronger in the son than in the daughter, and this not only for stature in the English race, but also for cephalic index in the North American Indians. A similar relation was found in respect of eye-colour, hence Pearson con- siders that in man it may be a general rule for the male to inherit more than the female. A comparison of other coefficients of correlation seemed to show that the hereditary resemblance be- tween brother and brother, or sister and sister, is greater than that between brother and sister. This was true for stature, head index, and eye-colour in man, and also for coat-colour in thoroughbred-race horses. It would therefore follow that inheritance in a line through one sex is prepotent over inheritance with a change of sex, or that, for instance, a man would resemble his paternal more closely than his maternal grandfather. It is not to be imagined that prepotency of the male over the female is in any way a general law. Thus in thoroughbred horses, sire and dam are equipotent in the transmission of coat-colour. In Basset hounds, on the other hand, Galtonf found that the female was pre- potent over the male in transmitting colour in about the *L. c.,p. 457. fProc. Roy. Soc., Ixi. p. 404, 1897. 152 BLASTOGENIC VARIATIONS. proportion of 6 to 5. "Whatever the degree of pre- potency of one parent over another and with similarly bred stocks it is probably never very great we must conclude that the average contribution of both parents together still remains at a half, that of grandparents at a quarter, and so on. It is only the relative propor- tions contributed by the two sexes which differ. Thus it will be remembered that it was the Basset hound data which afforded Galton such valuable evidence in support of his law. But what view are we to take of the more striking instances of prepotency mentioned above? Are they also conformable to the law of heredity, or are they ab- normal and exceptional? Galton himself has come to the conclusion that high prepotency does not arise through normal variation, but must rank as a highly heritable sport. As has been mentioned in a previous chapter, there is no adequate proof that sports transmit their characters more persistently than other varia- tions, and in any case it is probably unnecessary to as- sume that prepotency is other than a special case of the law of heredity. Thus we saw in the above-mentioned case of the black grajrhound, that the dog had a prepon- derance of black in his blood, both on the sire and dam side, whilst both the instances of prepotency mentioned by Professor Ewart seem largely attributable to in- breeding. This inbreeding, according to Professor Ewart, induces prepotency by fixing the characters of the particular variety selected. But what is really meant by fixing a character? To adequately com- prehend the meaning of the term it is only neces- sary to examine Galton's law of ancestral heredity a BLASTOGENIC VARIATIONS. 15$ little more in detail than we did in the last chapter, and it will then be obvious that a character becomes more and more fixed in the offspring, the more and more fully it is represented in the parents, grandparents, and more remote ancestors. Thus, according to the law, offspring receive a half of their heritage from their parents, a quarter from their grandparents, and so forth. But of what is this half and this quarter made up ? Obviously half of the parental half heritage, or a quarter in all, was received from their parents, and a quarter of a half, or an eighth in all, from their grandparents, and so on; whilst, as regards the quarter heritage received by the offspring from their grandparents, a half of it, or an eighth in all, was received from their parents, and so on. By tracing back the heritages in this way, it is therefore possible to calculate the absolute amounts of any character or strain present in offspring, as distin- guished from the relative amounts; relative, that is, to those present in the parents and other ancestors. For instance, supposing the parents and parents only had been selected in respect of any particular character, the condition of the previous ancestors being entirely un- known, then, as we have already seen, the offspring will exhibit these exceptional characters in only .6 of their full strength, or will have regressed to this extent to- wards the general mean of the race. Supposing the grandparents have been selected in respect of the same characters, as well as the parents, then Pearson * has calculated that the offspring will exhibit the characters in .8049 of their full strength; if the great-grand- parents also, then in .9027 of their strength; and if still *Proc. Roy. Soc., Ixii, p. 39! 154 BLASTOGENIC VARIATIONS. three other generations back be selected, then in .9879 of their full strength. That is to say, " after six gener- ations of selection the selected individuals will, without further selection, breed true to the selected type within nearly 1 per cent, of its value." In fact, prac- tically all regression towards mediocrity will have been weeded out. Supposing now, some variety of a species which had been bred true to its varietal characters for only two generations were crossed with another variety of the same species which had been bred true to Us characters for six, then the resulting offspring 8049 would receive ' ~ = .4025 of the characters of one & 9879 parent, = . 4940 of those of the more thorough- <0 bred parent, and .1035 of unknown blood. Knowing as we do that many characters show little or no tend- ency to blend, it would not be very remarkable if the offspring resembled the more thoroughbred parent to the partial or almost complete exclusion of the ill bred. That is to say, the one parent would prove itself strongly prepotent, simply through its characters hav- ing become fixed through inbreeding. Hybridisation. Though, as a rule, intercrossing be- tween different varieties of the same species tends to produce uniformity of character, yet it may also very frequently lead to the production of increased varia- bility, not only by the partial or complete absence of blending of the parental characters, but also by the ap- pearance of seemingly fresh characters, due to rever- sion or some other cause. Our knowledge of the laws governing hybridisation BLASTOGENIC VARIATIONS. 155 is chiefly derived from observations on plants, for by reason of the ease and success with which they are carried out, and the scientific and practical results ob- tained, these altogether outweigh the comparatively few observations which have been made on animals. Though much of the evidence obtained is variable and contradictory, yet some of it has afforded results of striking lucidity. Especially is this the case as regards what may be termed Mendel's Law of Hybridisation. Though discovered as long ago as 1865,* this important generalisation has passed almost unnoticed until the last year or two, when it was independently re-discov- ered and confirmed by De Yries, by Correns, and by Tschermak. Mendel's observations extended over eight years, during which over 10,000 plants were ex- amined. Most of them were made upon the different varieties of the pea, Pisum sativum. The varieties employed differed in respect of (1) the form of the ripe seeds, these being either nearly round, or angular and wrinkled; (2) the colour of the cotyledons, these being various shades of yellow or green; (3) the colour of the seed coat, this being either white, gray, or brown; (4) the form of the ripe pods, these being either simply in- flated, or deeply constricted between the seeds; (5) the colour of the unripe pods, this being yellow, or light to dark green; (6) the position of the flowers, either axial or terminal; (7) the length of stem. On uniting each of these two differentiating char- acters by cross-fertilisation, the hybrids obtained in each case were found to resemble only one of their * Abhandl. d. naturforsch. Ver. in Brttnn, Bd. iv. 1865. Trans- lated by W. Bateson in J. Roy. Horticult. Soc., xxvi. p. 1, 1901, 156 BLASTOGENIC VARIATIONS. parent forms, and to show little or no trace of the other. The characters thus appearing were termed by Mendel dominant, and the characters becoming latent in the process, recessive. In the next generation, how- ever, the seeds from these dominant hybrids betrayed their mixed origin, for instead of maintaining the pure dominant character, on an average one out of every four of the plants or seeds obtained reverted to the re- cessive parent form. The following are the actual numbers of plants and seeds examined by Mendel in respect of the various differentiating characters above mentioned: Proportion of Dominant to Recessive. (1) 253hybr (2) 258 (3) 925 (4) 1187 (5) 580 (6) 858 (7) 1064 ds (gave 7324 seeds) ( " 8023 " ) Mean 2.96 3.01 3.15 2.95: 2.82: 3.14: 2.84: 2.98 It will be seen that the proportion of 3:1 is fairly evenly maintained in respect of all the characters ob- served. The observations on the next and succeeding genera- tions afforded an even more remarkable result than this, for they proved that the recessive forms obtained in the second generation were absolutely pure. Thus the seeds obtained by crossing them amongst each other, or by self-fertilisation, yielded offspring which never showed any trace of the dominant grand-parental char- acters. The dominant forms, on the other hand, which of course were self -fertilised, underwent a further split- ting up. A third of them yielded plants which in sub- BLASTOGENIC VAEIATIONS. 157 sequent generations proved themselves to be pure domi- nant forms, whilst two thirds of them still retained their hybrid nature, as was shown by their yielding, in the next generation, recessive and dominant forms in the proportion of 1:3. The gradual resolution of the original hybrid forms into pure parental forms may be represented diagrammatically thus,* it being assumed that 64 hybrids with yellow cotyledons had been pro- duced by the crossing of parental forms having respect- ively green and yellow cotyledons. Parents I. Gen. green 64 yellow yellow We see that of the plants produced by crossing the original 64 yellow hybrids haphazard amongst them- selves, a quarter are of the pure green form, a quarter of the pure yellow form, and a half of them hybrids with the yellow character dominating. On crossing these hybrids among themselves, we see that in each subsequent generation their number is reduced by half, till in the seventh generation only 1 of the original 64 hybrids would be still remaining. The explanation of this result was clearly laid down by Mendel, he supposing that the cross-bred plant pro- duced pollen grains and ovules, each of which bore only one of the alternative varietal characters, and not both. If D and R represent the two characters present in dif- * Modified from Correns (Ber. d. deutsch. bot. Gesell., xvii. p. 162, 1900. II. Gen. 16 green r [32 yellow 1 4S\ I* [ 16 yellow III. Gen. 16 green 8 green f ( 16 yellow j 4 (8 yellow 16 yellow IV. Gen. V. Gen. 16 green 16 green 8 green 8 green 4 green 4 green ( 2 green { 8 yellow { 4 yellow (by.) ( 2 yellow 4 yellow 4 yellow 8 yellow 8 yellow 16 yellow 16 yellow 158 BLASTOGENIC VARIATIONS. f erent ovules of the hybrids, and d and r those in pol- len grains, then on crossing these hybrids haphazard, the germ cells giving rise to the next generation will unite so as to form Dd + Dr + dR -\-Rr. Now Men- del found that it was perfectly immaterial whether the dominant character belonged to the male or the female plant, and so it follows that we should get twice as many similar hybrid forms (Dr and dR) as of pure dominant or pure recessive. If parental forms possessing two or more differentiat- ing characters be crossed, the law of alternative herit- age continues to hold, though it necessarily becomes somewhat more complicated. For instance, Mendel crossed seed parents with round seeds (A), and yellow cotyledons (.?), with pollen from plants having angular seeds (a), and green cotyledons (6). The hybrids would therefore consist of plants with germ cells hav- ing the characters AB, Ab, Ba, and db. These hybrids, on crossing haphazard, would yield the following: (AB + Ab + Ba + o&) 2 = A*B* + AW +2A*Bb + 2AB*a + 38 35 65 60 4 ABdb +2 Aab* + 2 Ba?b + BW + S 6 9 138 67 68 28 30 The figures underneath indicate the actual numbers of plants obtained by Mendel from the 556 seeds yielded by the 15 original hybrid plants. The average numbers with two, three, and four characters are re- spectively 34, 65, and 138, or very nearly in the theo- retical proportion of 1 :2 :4. Mendel even took the immense trouble to cross parents differing in respect of three characters, and he found that the offspring of the resulting hybrids with BLASTOGENIC VARIATIONS. 159 3, 4, 5, and 6 characters were on an average respectively as 10 : 19 : 43 : 78, or very nearly as 1: 2: 4: 8. He also confirmed his law by some observations on PJiase- olus vulgaris and P. nanus, but the crossings of P. nanus 9 with P. multiflorus gave only a partial result, whilst those on Hieracium did not agree at all. In the light of the hybridisation experiments of Kolreuter, Gartner, and others, Mendel recognised that his law was by no means universally applicable. It obviously can only apply to cases of exclusive inheritance, and not to those of blended or mixed inheritance. De Yries * has made similar observations to those of Mendel upon varieties of no less than 15 different species of plants, and in every case found that the pro- portion of recessive forms obtained in the second gen- eration was approximately the theoretical 25 per cent. When the observations were continued through other generations, the results likewise agreed with theory. Tschermakf repeated MendePs observations upon the different varieties of Pisum sativum, and with some of them obtained a similar result. However, he found that in some other cases $ the law did not hold. Correns also experimented with varie- ties of the pea, and he found that whilst some of the characters obeyed Mendel's law, others, such as the colour of the skin of the seed, did not. He ob- tained a similar result If on crossing Matthiola incana *Ber. d. deutsch. Botan. Gesell., xviii. p. 83, 1900. Translation in J. Roy. Horticult. Soc., xxv. p. 243, 1901. fBer. d. deutsch. Bot. Ges., xviii. p. 232, 1900. JBer. d. deutsch. Bot. Ges., xix. p. 35, 1901; also Zeitschr. f. d. landwr. Versuchewesen in Oesterr., iii. p. 465, 1900. Ber. d. deutsch. Bot. Ges., xviii. p. 158, 1900. IBot. Centralb., Ixxxiv. p. 97, 1900. 160 BLASTOGENIC VARIATIONS. and M . glabra. Thus some of the characters, such as the colour of the flower petals, remained fixed in the hybrids. Also, on the whole, the hybrids had a greater resemblance to the female than to the male parent. In the second generation flowers of new colours, viz., white and red, appeared, in addition to the yellow-white and violet flowers exhibited by the parents. Correns also made a number of crosses between undoubted species (e. g., Cirsium palustre .+. spinosissimum, Achillea macrophylla + moschata, Car ex echinata +, fcetida, etc.) and he is doubtful whether one of these hybrids showed a single really dominant character. It was quite obvious that almost all the characters which served to differentiate the parents were present, in greater or less degree, in the hybrids. Correns con- cludes, therefore, that almost without exception the domination of a character shows itself only in crosses between varieties, whilst the hybrids of true species show the characters of both species, though in dimin- ished degree. It has been pointed out by Weldon * that Mendel's results are partly vitiated by the fact that he quite neg- lected the ancestry of the plants with which he started his cross-fertilisations. Weldon also adduces a con- siderable body of evidence to show that the separation of the seed characters into definite dominant and re- cessive types by no means invariably holds good. The offspring of cross-bred peas may continue to contain a large percentage of intermediate forms, even as long as 25 generations after the crossing. f *Biometrica, i. p. 228, 1902. * For further evidence concerning the Law see Report to Evolution Committee of Royal Society by Miss Saunders and W. Bateson, 1902. BLASTOGENIC VARIATIONS. 161 According to Darwin, variability is especially induced if mongrels are repeatedly crossed with either pure parent form, whilst the crossing of different species may lead to much wider variatio'n than the crossing of varieties. The hybrids produced on the first cross are, as a rule, fairly constant in their characters, but if these hybrids be crossed again, or crossed with either pure parent form, then a very considerable variability may result. " He who wishes," says Kolreuter, " to obtain an endless number of varieties from hybrids, should cross and recross them." * Again Darwin t says that cross-bred animals " for breeding are found utterly useless; for though they may themselves be uniform in character, they yield during many generations aston- ishingly diversified offspring." Indeed it would seem that entirely new characters may be produced by this means. For instance, Kolreuter says that hybrids in the genus Mirabilis vary almost infinitely, and he de- scribes new and singular characters in the seeds, an- thers, and cotyledons. Professor Lecoq also asserts that many of the hybrids from Mirabilis jalapa and multi-flora might easily be mistaken for distinct species. Again, Herbert J has described certain hybrid Khodo- dendrons as being unlike all others in foliage, just as if they were a separate species. According to Focke, the hybrid may be related to the parent forms in three different ways: (1) there may * " Animals and Plants," ii. p. 254. \L. c., ii. p. 74. j Science Progress, vol. vii. p. 185, 1898. "Die Pflanzen-Mischlingen," Berlin, 1881. Quoted by Weis- mann, " Germ Plasm," p. 261. 162 BLASTOGENIC VARIATIONS. be a strict mean in all parts; (2) the paternal or mater- nal characters may predominate; (3) the paternal char- acters may predominate in some parts of the hybrid, and the maternal in others. The first-mentioned con- dition is by far the most frequent. For instance, K61- reuter states that the hybrid between Nicotiana rustica 9 and N. paniculata $ (tAVO species of tobacco plant) is exactly intermediate between the parent forms. On the other hand, the hybrid between N. paniculata 9 and N. vincoeflora $ bears so close a resemblance to the second of these species that the characters of N. paniculata can hardly be recognised at all. An in- stance of the third class is occasionally found in the cross between N. rustica 2 and N. paniculata $ , the blossoms resembling one parent species, and the leaves the other. Again, Milardet * has obtained a series of non-separating crosses by the union of Fragaria, Rubus, etc. They resembled either the male or the female parent. De Vries f has obtained a similar result with (Enothera muricata 9 X biennis, which displayed the paternal character. Crosses between (Enothera La- mar clciana $ and 0. nanella $ gave progeny which always displayed two types, the maternal and paternal, but these occurred in very varying ratios. Crosses of 0. lata 9 and 0. Lamarckiana $ also yielded progeny of both parental types. Plant hybrids are of considerably more frequent oc- currence in nature than animal hybrids, and, by virtue of the fertility which they often possess are of dis- *Mem. Soc. Sc. Phys. et. Nat. Bordeaux, vol. iv. p. 1, 1894. fBer. d. deutscli. Bot. Ges., xviii. p. 435, 1900. Translation in J. Roy. Hort. Soc., xxv. p. 249, 1901. BLASTOGENIC VARIATIONS. 163 tinct importance as a source of variations. Thus Ben- nett,* in a paper on Hybridity in Plants, makes the fol- lowing remark: " There seems, however, scarcely to be room for doubt that in some of our abundant wild genera, such as Rubus, 8alix, and Hieracium, hybridity is not uncommon in nature. It has long been known that in some genera, such as Passiflora, and in some Orchidese, the ovules appear to be even more readily fertilised by pollen of a different species. W. Focke now states that this is also the case with the species of Lilium belonging to the group lulbiferum, and with some species of Hemerocallis; and J. H. Wilson affirms the same respecting the Cape genus Albuca, also belong- ing to the Liliacese." Again, Rolfe f states that Nar- cissus incomparabilis is known to be wild in France, and that Herbert found that on crossing a daffodil with pol- len of N. poeticus, the seedlings yielded flowers identi- cal with those of N. incomparabilis. Similarly, by crossing the Daffodil with the Jonquil, Herbert suc- ceeded in producing N. odorus. Again, Engleheart has proved the hybrid origin of N. biflorus by crossing N. poeticus with the pollen of N. tazetta, he obtain- ing seedlings identical with wild forms. Also he reconstructed N. pulcJiellus Salisb., by crossing N. triandrus with the Jonquil, the seedlings proving absolutely identical with the wild plant. Further " Digitalis supplies some wild hybrids whose origin has been artificially demonstrated. For example, D. pur- purascens y Both, has been reconstructed by crossing and recrossing D. lutea and D. purpurea; and D. media, *Nat. Sci.,vol. ii. p. 208. \ J. Roy. Horticult. Soc., xxiv. p. 181, 1900. 164 BLASTOGENIC VARIATIONS. Roth, in the same way from D. purpurea and D. am- bigua (grandiflora) . . . D. ambigua has also been crossed with D. purpurea and with D. lanata, in each case yielding hybrids which also occur wild." Rolfe also records that Wichura succeeded in raising artificially no less than eight hybrid willows identical with those which had long been known in the wild state, and Linton has added at least six others. For instance, Salix rubra was obtained by crossing 8. purpurea with the pollen of S. viminalis. Kerner is of the opinion that species may be produced by hybridisation. In his " Natural History of Plants " he gives instances of these hybrid races. To quote Rolfe, " A hybrid between Medicago falcata and sativa, known as M. media, is widely cultivated as a fodder plant, and is propagated from seed. Salvia betonicce- folia, a hybrid from 8. nemorosa and nutans, is as com- mon as its parents in grassland in Central Hungary. Betula alpestrisj a hybrid between B. alba and nana, is abundant in the Jura, Scandinavia, and in North Russia, here and there whole copses of it being found. Nigri- tella suaveolens, a hybrid between N. angustifolia and Gymnadenia conopsea, is abundant in some Swiss locali- ties, hundreds of plants sometimes occurring in a single meadow. Hybrids between the Primrose and Cowslip occur in thousands in upland meadows in the Eastern Alps." Again, in some localities in the Tyrol the hybrid Rhododendron intermedium exists side by side with its parent forms, R. ferrugineum and R. hirsutum, it sometimes being commoner than they are. Also it seeds freely, and comes true to seed, and so fulfils all BLASTOGENIC VARIATIONS. 165 the requirements of a species. The same is true of Salvia sylvestris, a hybrid from 8. nemorosa and fira- tensis, which abounds in dry meadows all over the low country south of Vienna, and of Nuphar intermedium, a hybrid from N. luteum and pumilum, which occurs in the Black Forest, Russia, Sweden, and other localities. It appears that a hybrid is sometimes found in company with one parent only, or with one in one locality and both in another; or sometimes even where both are absent. Kerner estimated that something like a thousand natural hybrids have been found in Europe during the last forty years, but of these hybrids only a fraction survive and multiply. As regards artificial hybrids, Hurst * has compiled a list of genera from various authorities, and from his own observations, and he finds that 91 distinct genera are recorded in which fertile hybrids are known. In only three, viz., Ribes, Polemonium, and Digitalis, were the hybrids all quite infertile, and in none of them had many experiments been made. Hurst also remarks that " during the past seven years Mr. Reginald Young has been crossing inter se some 30 distinct species and 53 distinct hybrids in the genus Paphiopedilum (Pfitz),and has . . . carefully recorded no less than 849 crosses. Of these, taken together, 80.2 per cent, have proved fertile, i. e., produced good seeds. Of 263 crosses between distinct species, 95 per cent, were fertile. This seems to show that in this genus crosses between distinct species are almost, if not quite, as fertile as crosses between varieties of the same * J. Roy. Horticult. Soc., xxiv. p. 90, 1900. 166 BLASTOQENIC VARIATIONS. species; while in crosses in which a hybrid was con- cerned in the parentage, out of 586, only 73.5 per cent, proved fertile, showing that crosses with hybrids, though fertile to a high degree, are yet rather less fer- tile than crosses between species. . . A further analysis of the figures shows that while hybrids crossed with the pollen of pure species give 91.8 per cent, fertile, yet pure species crossed with the pollen of hybrids give but 60 per cent, fertile." That is to say, the decline in the fertility of the hybrids is due in a large measure to the loss of power in the pollen of the hybrids. This decline in power of the male element has been noticed before in other plants by Darwin, Focke, and others. Rimpau * has made series of experiments on the crossing of some of our common agricultural plants, and, amongst other results, obtained ten artificial and nine natural hybrids in wheat, and two artificial and six natural hybrids in barley. His most striking result of all was to obtain a fertile hybrid between wheat and rye, plants belonging to different genera. Again, Hurst states f that amongst Orchids no less than 150 bigeneric crosses are recorded. Bigeneric hybrids have also been recorded J between Philesia and Lapigeria, between Urceolina and Eucharis, between numerous genera of Gresneracese, etc. Finally a cross has been described from Digitalis ambigua (Scrophulariacese) by pollen of Sinningia speciosa (Gesneracese) ; i. e., a binordinal hybrid. * ' ' Kreutzungsproducte landwirth. Cultur-pflanzen," Berlin , 1891. \Loc. cit. i Nature, vol. Ixiv. p. 447, 1901. Maund's " Botanic Garden," v. p. 468. BLASTOGENIC VARIATIONS. 167 Upon members of the Animal Kingdom very few ex- tensive and systematic crossing experiments have been made. The most complete are those of Standfuss,* on various races and species among the Lepidoptera. Standfuss' general conclusion is that on crossing the normal form of a species with a gradually formed local race of the same species, a series of more or less inter- mediate forms results. For example, on crossing Calli- morpha dominula $ with the variety persona 9, the issue resulting were of a very variable form, more or less intermediate, but somewhat more closely resem- bling the type than the variety. In the reciprocal cross, the insects, on the whole, also came nearer to (7. dominula than to the variety, but not so much as be- fore. When species were crossed Standf uss found that the hybrid form lay between the extreme parental forms, but was not strictly intermediate. Arguing from his experiments on crossing various species of Saturnia, Standfuss concludes that the adult offspring are more similar to the male parent than to the female, the extent of approximation depending on the relative age of the two species. Crosses of the male hybrids with the parent forms were in some cases proved to be fertile, and hence there is no reason why such forms should not establish themselves under natural condi- tions. Thus Dr. Dixey, in a very good resume of Standfuss' researches,f says with reference to these hy- brids, " Since the product of this kind of crossing is not found to show a complete reversion to the type of the female parent, it is possible that the existence of vari- *" Handbuch der palaartischen Gross-Schmetterlinge," Jena, 1896. t Science Progr., vol. vii. p. 185, 1898. 168 BLASTOGENIC VARIATIONS. ous intermediate forms in such genera as Melitcea, Zygoma, and Agrotis may be accounted for in this man- ner. Cases of simple pairing between distinct species of the two former genera have been observed by the author [Standfuss] in nature." Upon Echinoids, the author has made numerous crosses and reciprocal crosses.* Eight species were worked with, and of the 56 possible crosses, 41 were at- tempted. Of these, 22 yielded larvae of 8 days' growth. In only one cross did any of the larvae incline towards the paternal type, and the majority of those then ob- tained were more or less intermediate. In nine other crosses also they were more or less intermediate in char- acter, whilst in the remaining twelve they were of the maternal type. A few of the larvae exhibited char- acters which were not present in either parent. Upon members of the Mammalian and Avian King- doms, a very large number of crossing experiments have been made, and frequently with success, but the observations are not sufficiently extensive to admit of generalisations. The most interesting experiments of recent years are those of Professor Ewart, upon zebra hybrids.f By crossing mares of various sizes (11 to 15 hands) with a zebra stallion, nine hybrids were obtained altogether. Also Professor Ewart had in his possession three hybrids out of zebra mares, one having for his sire a donkey, whilst the other two were sired by ponies. The hybrids showed a " curious blending of characters, derived apparently partly from their actual and partly *Phil. Trans. Roy. Soc., 1898, B. p. 483, and Arch, f . Entwick- elungsmechanik. , Bd. ix. p. 468, 1900. f " The Penycuik Experiments." BLASTOGENIC VARIATIONS. 169 from their remote ancestors. . . Some of the hybrids in make and disposition strongly suggest their zebra sire, others their respective dams; but even the most zebra-like in form are utterly unlike their sire in their markings." In some respects, also, the hybrids were intermediate between their parents. As to the causes of the different relationships be- tween parental and hybrid characters, we are almost entirely in the dark. Weismann has endeavoured to account for them on his theory of the germ-plasm, but his explanation is purely theoretical and from its nature incapable of experimental verification. The observa- tions of the author on sea-urchin hybrids, and of Pro- fessor Ewart on crosses between varieties of rabbits, throw a little light on the subject, for they show that the characters of the hybrids may be considerably in- fluenced by the seasonal condition of the parental sex- cells, and thereby seem to indicate that the compara- tive degrees of nutrition of the sex-cells, and perhaps also of their constituent parts, may be a very important factor. One should also bear in mind that, as was demonstrated by Mendel in the case of certain plant hybrids, some of the parental characters may remain latent in the hybrid offspring, and only reveal their presence in subsequent generations. The existence of latency is also shown by secondary sexual characters. In every female all the secondary male characters, and in every male all the secondary female characters, ap- parently exist in a latent state, ready to be evolved under certain conditions, such as the removal of the ovaries or testes. The variability of hybrids may there- fore be due not only to their having received varying 170 BLASTOGENIC VARIATIONS. and unequal amounts of the different characters from their parents, to these either partially or entirely refus- ing to blend, but also to some of the characters received remaining latent, or to characters latent in the parents revealing themselves in the offspring. Sports. Instances of so-called sports, or suddenly occurring aberrant variations, have been given in the second chapter, but nothing was said of their origin. To what are we to attribute this ? Are they to be regarded as normal, only somewhat exaggerated, variations, or are they something essentially different? The more general opinion probably inclines to the latter view, as there are several facts which it is difficult to reconcile with the former. It is said, for instance, that sports, as distinguished from varieties, are much more stable; that they may be transmitted to successive generations with considerable persistence and in undiminished strength. Galton has suggested * that whilst organ- isms showing ordinary variations are grouped round one " position of organic stability," towards which the off- spring in the next generation tend to regress, sports are centred round a different position of stability, and are not merely a strained modification of the original type. They therefore have little tendency to revert to Ais original type, but are capable of propagating their freshly acquired characters more or less undiminished, and so giving rise to fresh races. Galton considers that the results which he has obtained in his detailed study of human finger-prints f afford strong evidence in sup- port of his view. These patterns, formed by the papil- * Vide "Natural Inheritance," p. 30; also "Mind," p. 362, 1894. fPhil. Trans. 1891, B. BLASTOQENIC VARIATIONS. 171 lary ridges on the bulbs of the fingers, are the most per- sistent of all the external characters that have yet been examined. They are found to fall in three definite and widely different classes. Each of these is a true race in the sense in which that word was defined, transitional forms being rare and the typical forms being frequent. Galton thinks that the continual appearance of these well-marked and very distinct patterns proves the reality of the alleged positions of organic stability. A clear distinction between sports and varieties seems to show itself also amongst the Lepidoptera. Thus Standfuss * found that when a sport is crossed with its parent form, the issue is sharply divided in both sexes into specimens resembling either the sport or the nor- mal form. There are no true intermediate forms, though occasionally forms are observed in which the characters are unsymmetrically mixed. When the nor- mal form of a species is crossed with a gradually formed local race, however, a series of intermediate forms is obtained. We have seen also that De Yries, in his ex- periments on plants, claims to have found a wide dif- ference between mere varieties, and true sports such as were obtained from (Enothera Lamarckiana. As already mentioned, sports have been stated to be much more persistent in propagating their aberrant characters than normal varieties, but the evidence in favour of such a generalised statement is quite in- sufficient. There are certainly a few instances which strongly support it, but there are a good many more which entirely fail to do so. Of the former, the in- *"Handbuch der palaartischen Gross-Schmetterlinge," Jena, 1896. 172 BLASTOGENIC VARIATIONS. stances of the ancon or otter sheep * and japanned or black-shouldered peacocks,f quoted by Darwin, are the most striking. The originator of the ancon breed of sheep was a single ram, born in Massachusetts in 1791. Ancon rams and ewes invariably produced ancon off- spring, whilst when crossed with other breeds the off- spring resembled either parent, and only very excep- tionally yielded intermediate forms. Japanned pea- cocks, which differ conspicuously from the common peacock in colouring, appear suddenly in flocks of the common kind. Though smaller and weaker birds, they have been known in two instances to increase, and finally extinguish the previously existing breed. They would therefore seem to have been strongly prepotent. That sports may be no more transmissible than other variations seems to be true in the case of polydac- tylism, for Dr. Struthers asserts that cases of non- inheritance and of the first appearance of additional digits in unaffected families are much more frequent than cases of inheritance. { Again, Galton regards as sports the mental arithmeticians and eminent musi- cians who are occasionally born into families which in previous generations have shown no signs of such ex- ceptional characters. Though these characters may be transmitted to descendants, yet this is the excep- tion, and not the rule. The subservience of sports to the law of hereditary transmission is well shown by some observations of Standfuss on Lepidop- tera. In 1888 a normal female Aglia tau was crossed * " Animals and Plants," i. p. 104. \L. c., i. p. 305. JZ. c.,i. p. 458. BLASTOGENIC VARIATIONS. 173 with a dark aberrant form or sport of this species, Aglia lugens, which had been interbred for two genera- tions. In 1889 some of the lugens, both male and female, obtained from this cross, were crossed with normal tau specimens. About half the offspring ob- tained resembled one parent and half the other, inter- mediate forms being absent. On breeding some of the 1889 9 and $ lug ens together, however, their off- spring consisted of about 36 per cent, of tau, and 64 per cent, of lugens forms. In 1890 some of these lugens were bred together, and their offspring consisted almost entirely of lugens, only 11 per cent, being of the tau form. In this latter case, therefore, both parents and all four grandparents were lugens; in the 1889 off- spring, both parents but only two grandparents, and in the 1888 offspring only one parent and two grand- parents. If sports be of an essentially different nature to nor- mal variations, as the somewhat insufficient evidence available may perhaps be taken to indicate, how is it that they arise ? Apparently they occur spontaneously, but doubtless some exciting cause must exist, invisible though it may be. The artificial production of mon- sters seems to throw some light on the subject, and hence a brief reference to them may be made. These monsters or malformations probably differ from sports only in degree, and not in kind. Hence, if the means adopted for their artificial production are such as may occur under natural conditions, it seems possible, and even probable, that sports themselves may owe their origin to similar agencies. For instance, Dareste, as long ago as 1877, described numerous experiments on 174 BLASTOGENIC VARIATIONS. the effects of placing fowls' eggs vertically instead of horizontally during development, of keeping them slightly above or below the normal temperature of incu- bation, of heating different parts of the egg unequally, and of modifying the conditions of respiration by var- nishing part of the shell.* Various considerable mal- formations were produced, but these were more or less the same, whatever the conditions that produced them. Professor Windle,t who has extended these investiga- tions and determined the effects of various other agencies, as electricity and magnetism^ on development, came to a similar conclusion. He considered that these disturbing agents act, in the majority of cases, on that part of the developing organisation which is concerned with the formation of the vascular system of the em- bryo, and so indirectly produce the malformations ob- served. The suggested connection between considerable mal- formations and sports has not as yet been borne out by Dareste's researches, although observations have been made with the object of finding it. In these observa- tions the conditions found to produce considerable mal- formations were reduced in strength, in the hope of thereby obtaining only slight anomalies, compatible with continued existence and the procreation of off- spring. Unfortunately the domestic fowl, which was invariably made use of, is unsuitable for such observa- tions. The type is so diversified that the experimenter who obtains some variety can never be certain whether * " Recherches sur la Production artificielle des Monstrosites," Paris, 1877; second edition, 1891. fProc. Birmingham Phil. Soc., vii. p. 220, 1890. BLASTOGENIC VARIATIONS. 175 it should be attributed to the conditions of experiment, or to some physiological cause arising in the egg itself. To test this question with some chance of success, the eggs of some species which varies but little ought to be employed; e. g., some wild species. But in this case it would be very difficult to obtain sufficient material. In the case of certain Lepidoptera, however, the arti- ficial production of sports has been successfully accom- plished by Standfuss.* By keeping the pupae of V. cardui (Painted Lady) at a high temperature for a short period, he succeeded in producing a small number of specimens of the aberrant form elymi, a form which is occasionally found under natural conditions. Again a low temperature, acting on pupae of V. io (Peacock), produced a variety ab. fischeri, which exhibits a reduc- tion in the number of the blue scales on both fore and hind wings. In these and other characters there seemed to be an approach to the type of V. urticce. These and other observations seem to justify Standfuss' conclusion that many of the aberrations occurring in nature may likewise have arisen through the influence of abnormal temperature conditions. Telegony. The term telegony, or so called infection of the germ, is applied to certain cases apparently show- ing the influence of a previous fertilisation on the structure of the subsequent offspring. The test case, always quoted in support of the existence of this phe- nomenon, is that of Lord Morton's mare. This animal bore a hybrid to a quagga, and subsequently produced two colts by a Black Arabian horse. These colts, both in the hair of their manes, their partial dun colour, and *The Entomologist, vol. xxviii. p. 145, 1895. 176 BLASTOGENIC VARIATIONS. striping on the legs, strongly resembled the quagga. Other cases have been quoted in support of this phe- nomenon, but it is unnecessary to mention them here, for none of them are absolutely convincing. The reader who wishes for details of these cases should con- sult a useful paper by Finn.* In his Penycuik experi- ments, Professor Ewart has made a number of attempts to obtain evidence of the phenomenon, but so far with entirely negative results. Sir Everett Millais made a considerably larger series of experiments, on a variety of animals, but was equally unsuccessful. Many Ger- man breeders also believe telegony as yet unproven. Finally, Professor Pearson f has shown that exact statistical examination of appropriate data gives no support whatever to the hypothesis. Pearson's method of testing the question was to determine whether younger children are more closely correlated to their parents in respect of some character such as stature, than older children. Supposing the male parent were able to exert any influence on the ma- ternal tissues, and so indirectly on the offspring, then clearly this influence would be greater for the younger children than for the older children. As Pearson recognises, it is possible that telegony, if it occurs at all, is due to the abnormal preservation of the male sex cells of an earlier union, and in such a case his method would afford no evidence one way or the other. Probably, therefore, no such thing as telegony exists. In any case it is so exceedingly rare that, as a possible source of variations, it may be neglected. *Nat. Sci., vol. iii. p. 436. fProc. Roy. Soc., Ix. p. 273. BLASTOGENIC VAEIATIONS. 177 Parthenogenesis. We saw in the last chapter that Weismann regarded sexual reproduction as a potent factor in the production of variations, in that it af- forded inexhaustible supplies of fresh combinations of the individual variations already represented in the mingling germ-plasms. We should accordingly con- clude that when such sexual union is wanting, as in parthenogenetically produced animals, the amount of variation will be smaller, and that parent and offspring will more closely resemble each other. The evidence upon this point is exceedingly slight, but what there is perhaps tends rather to support this deduction. Thus Weismann made a series of observations, extending over eight years, upon a small ostracod, Cypris reptans. This organism exists as two well-marked varieties, one being coloured yellow, with five small green spots on each side of the shell, and the other seemingly dark green, owing to the great enlargement of these spots.* Both varieties are produced parthenogenetically in the neighbourhood of Freiburg, males never being found. Females of each variety were isolated, fed well, and allowed to multiply for many generations. It was found that " the descendants of the same mother re- sembled one another as well as the parent with which the experiment began, even as regards minute details of the markings. The differences were mostly as small as those which may be observed in identical human twins." Even after many generations no modification showed itself, so that colonies were obtained which could not be distinguished from their ancestors 40 gen- erations back. In three different instances, however, * ' Germ-Plasm," p. 344. 178 BLASTOGENIC VARIATIONS. some of the dark green variety appeared in broods of the typical yellow variety, and in one instance some of the yellow variety in broods of the dark green. These sudden transformations could not have been due to ex- ternal circumstances, as the two forms appeared in the same aquaria, under precisely the same conditions. Weismann attributes them to reversion. Evidence telling in the opposite direction to Weis- mann's has recently been obtained by Warren,* and as it is based on exact statistical measurements, one is in- clined at first sight to attach greater weight to it. The observations consisted in measurements of the total length of body to base of spine, and of the length of the protopodite of the second antenna of the right side, in 23 female Daphnia magna, and their 96 partheno- genetically produced offspring. As these animals con- tinue to grow throughout life, the second dimension was expressed in terms of the first, before calculating its variability. Its error of mean square, or standard deviation, was found to be 2.22 in the mothers, and 2.95 in the offspring. That is to say, the offspring were distinctly more variable than the mothers, and even the offspring of a single mother were found to be on an average more variable than all the mothers put to- gether. As the mothers had in a way been selected, only those which produced offspring being chosen, the daughters would be expected to be somewhat more vari- able, but in any case the variability was considerable. Again, it was found that the coefficient of correlation between mother and offspring was .446, whilst the co- efficient of regression of offspring on mothers was .619. *Proc. Roy. Soc., Ixv. p. 154, 1899. BLASTOGENIC VARIATIONS. 179 Now it has already been found, in the case of stature in man, that the correlation between mid-parent (i. e., mean between male and transmuted female) and off- spring is .424, and the regression of offspring on mid- parents .6;* hence Warren's values seem to show that the parthenogenetic mother acts as a mid-parent to her offspring, and not as a single parent, and also that these offspring exhibit regression towards the mean race type, just as sexually produced individuals do. As Warren himself points out, however, his evidence is not con- clusive. Thus the number of individuals measured was comparatively small, and also it would seem that Daphnia is a very unreliable organism to work with. It is so exceedingly sensitive to its environmental con- ditions f very considerable variations being produced by comparatively slight changes that these data de- rived from it can only be accepted with considerable reserve. A further series of observations was made by War- ren J upon Aphides (Hyalopterus irirhodus). Sixty parents and their 368 children were measured, and also 30 grandparents and their 291 grandchildren. War- ren found that the coefficients of parental and grand- * It has been stated in the previous chapter that the coefficients of correlation and of regression between single parent and offspring are practically the same thing, and are equal to .3. The coefficient of correlation between mid-parent and offspring is, however, /y/2 x .3 = .424, because the mid-parent, being the mean of two parents, is less variable than the single parents (in the proportion of 1 to -/o). The coefficient of regression of offspring on mid-parents, is, however, twice that of offspring on a single parent, i. e., is .6. \Vide Q. J. Microsc. Sci., vol. xliii. p. 199, 1900. t Biometrika, i. p. 129, 1902. 180 BLASTOGENIC VARIATIONS. parental correlation showed no marked difference from those obtained in sexual reproduction, just as in the case of his daphnia observations. If anything, there was a decrease in the correlation, on passing from sexual to parthenogenetic forms, rather than the in- crease we should expect. However, the variability of the individuals of a brood was found to be only about 60 per cent, of the racial variability; -i.e., distinctly less than in sexual reproduction. Also the mean coefficient of fraternal correlation for aphis and daphnia was .66, or considerably higher than the mean value of .45 ob- tained by Pearson * for fraternal correlation among sexual forms. Warren's general conclusion may be summed up in the words: " The question as to whether we have a real difference between parthenogenetic and sexual offspring can only be decided by further investi- gation both on aphis and other forms." In the light of Weismann's observations, which were carried on for such a number of generations, we seem entitled to con- clude that probably a real difference will be found to exist between them. Arguing partly from Warren's observations, and partly from others of his own, Professor Pearson f has come to the conclusion that, " whatever be the function of sex in evolution, it is not the production of greater variability." Thus he says that the individual contains in itself a variability which is 80 to 90 per cent, of the variability of the race, and which it can exhibit quite independently of sexual union, e. g., as in this case of parthenogenesis. As instances of individual varia- * Phil. Trans. 1901. A. p. 285. f " Grammar of Science," p. 474. BLASTOGENIC VARIATIONS. 181 bility, he refers to the stigmatic bands on the seed cap- sules of Shirley poppies. These vary in number from about 7 to 18, the most commonly occurring number being 12. The variability of a large number of indi- viduals (as expressed by the error of mean square), which were taken as a good sample of the whole race, was found to be 1.885. The average variability of the bands in the capsules obtained from each of 300 differ- ent plants, or the individual variability, was, however, only 15 per cent. less. Again, with reference to the number of leaflets on the compound leaf of the ash, the individual variability was only 8 per cent, less than the racial variability. In a recent paper,* Professor Pearson and his co- workers have determined the relationship between racial and individual variability in a number of other plant species. Enumerations were made of the veins in the leaf of the Spanish Chestnut and the Beech, of the prickles on Holly leaves, the sori on the fronds of Hartstongue ferns, the seeds in the pods of Broom plants, etc., and measurements of the length and breadth of ivy leaves and of the gills of mushrooms. On an average, the individual variability was found to be about 87 per cent, of the racial, it varying in the dif- ferent series of observations between 77 and 98 per cent. Now, even admitting that in these instances the individual variability is only slightly smaller than the racial, it does not appear to me that Professor Pearson is entitled to his contention, for all these highly variable individuals were, of course, produced as the result of sexual union in their immediate or remote progenitors. * Phil. Trans, 1901, A. p. 285. 182 BLASTOGENIC VARIATIONS. Such union may have been the starting point of con- siderably increased variation, which was never lost, even through innumerable subsequent asexual genera- tions. Thus Professor Pearson has shown that if the ancestors of individuals be selected so as to be abso- lutely similar in character for an indefinite number of generations back, such individuals will still have a variability of upwards of 89 per cent, of that of the original race. Though produced sexually, these indi- viduals are in reality comparable to asexually repro- duced forms, as by hypothesis no new characters were introduced by any of their ancestors. "Whether the difference between racial and individual variability is as small as Pearson maintains, or not, de- pends solely on what is meant by the word " race." If " species," in the generally accepted sense, is meant, then the view is certainly incorrect. If, however, a group of individuals is implied, all of which have been exposed during several generations to practically identical con- ditions of environment, then the view must be admitted. It is of little practical value, however, as may be real- ised by quoting certain data which Pearson has himself brought forward in another connection.* Thus the variability in the number of stigmatic bands in a sample of the wild poppy, Papaver Ehceas, collected in a corn field at the foot of the Chiltern Hills, was found to be 1.473, that in two individual poppy plants being, on an average, 1.166, or 20.9 per cent. less. Another sample was collected in some fields at the top of the Chilterns, and in this case the variability was 1.770, or 20.2 per cent, greater than in the other sample. Moreover, the * " Grammar of Science." p. 387. BLASTOGENIC VARIATIONS. 183 mean number of bands was also greater, it being 10.04 as against 9.84. In a third sample collected from still another locality, the variability was 1.455, but the mean number of bands was only 8.77. Supposing, therefore, equal numbers of specimens had been collected from all three localities and combined, the variability would have been about double the average variability of the individual groups of plants. Supposing samples had been collected from numerous and more widely sepa- rated localities, so as to get a representative sample of the whole species, then doubtless the variability would have been much greater still. Individual variability may therefore be only slightly smaller than local racial variability, but it is very much smaller than specific variability. What is true for plants is true also for animals. Supposing that in the case of the middle classes of Eng- lish society, the average variability of the stature of all the offspring is only about 10 per cent, more than that of the offspring of individual parents, then it is clear that if we were to include also representatives of the lower and of the upper classes in our sample, the aver- age variability would be somewhat greater, perhaps 12 per cent. If we were to include representatives in due proportions from all the continental nations, then the variability might be 25 per cent, or more in excess, and if from all the nations of the world, with African pyg- mies on the one hand, and Patagonian giants on the other, then it might be 50 per cent, greater, or even more. Asexual Reproduction in Plants. In plants asexual reproduction is very much more common than in ani- 184 BLASTOGENIC VAEIATIONS. mals, and though, until the above-mentioned memoir was published, there was practically no statistical evi- dence as to the range of variation then experienced, as compared with that found in sexually reproduced forms, there was available the common knowledge of every horticulturist and nurseryman, were he scientifically trained or otherwise. Thus it had been thoroughly well established that asexually produced forms, i. e., grafts, cuttings, offsets, and tubers, are characterised by a very much greater constancy than sexually produced forms, e. g., seedlings. For concrete instances I cannot do bet- ter than quote from those given by Mr. Sedgwick in his recent Presidential Address before the British Asso- ciation.* For example, in the asexual propagation of the potato by tubers, the plants, be they, for instance, of the Magnum Bonum variety, give rise to plants exactly resembling their parent in foliage, flowers, and tubers; if they be of the Snowdrop variety, the foliage, flowers, habit, and tubers are also similar, and are totally dif- ferent from those of the Magnum Bonum. " In this way a favourable variety of potato can be reproduced to almost any extent with all its peculiarities of earli- ness or lateness, pastiness or mealiness, power of re- sisting disease and so forth. By asexual reproduction the exact facsimile of the parent may always be ob- tained, provided the conditions remain the same." Supposing, on the other hand, one tries to raise Magnum Bonum plants from seed, in all probability not one of the seedlings will exactly reproduce the parents; they will all be different, both in properties of keeping, re- sisting disease, and so forth. " Again, take the apple : * Nature, vol. xl. p. 502, 1899. BLASTOGENIC VARIATIONS. 185 if you sow the seed of the Blenheim Orange and raise young apple trees, you will not get a Blenheim Orange. All your plants will be different, and probably not one will give you apples with the peculiar excellence of the parent. If you want to propagate your Blen- heim Orange and increase the number of your trees, you must proceed by grafting or by striking cuttings." In the face of such evidence as this, it seems impos- sible for Professor Pearson to maintain his belief that the function of sex in evolution " is not the production of greater variability." At the same time, his results above quoted show the incorrectness of the view some- times held, that variability is quite insignificant in asexually, as compared with that in sexually, reproduced forms. Statistically measured, it is only 10 to, say, 50 per cent, less, though when this amount is translated into differences of foliage and flowers, or of quality of fruit, it seems at first sight much more considerable. Bud-Variation. Considerable variations may arise asexually in cases of so-called bud-variation. This term was used by Darwin to designate the sudden changes in structure and appearance which occasionally occur in the flower-buds or leaf-buds of full-grown plants. Such changes are known to gardeners as sports, but, as we have already seen, this term is now used to include all suddenly arising discontinuous vari- ations. One of the best known and most striking instances of bud-variation is that of the nectarine, which occasion- ally appears on full-grown peach trees which have pre- 186 BLASTOGENIC VARIATIONS. viously borne peaches alone. This is the more remark- able as most varieties of both the peach and the nec- tarine reproduce themselves truly by seed. Again, nectarine stones occasionally yield peach trees, and a single instance is recorded of a full-grown nectarine tree bearing perfect peaches.* Numerous other in- stances of bud-variation have been observed in the plum, cherry, vine, gooseberry, currant, and other fruits, but it is unnecessary to refer to these here. In flowering plants, also, many cases have been recorded of a whole plant, or a single branch or bud, suddenly producing flowers different from the proper type in colour, form, size, or other character. For instance, a Chrysanthemum, raised from seed, produced by bud- variation six distinct varieties, five differing in colour, and one in foliage.f The common double moss-rose probably took its origin from the Provence rose by bud-variation. The leaves and shoots may be modi- fied by bud-variation as well as the flowers, and several varieties of trees have probably originated in this manner. ( As to the cause of bud-variation, we are in the ma- jority of cases entirely ignorant. Darwin attributes I many of the cases to reversion to characters previously p/ present, but which have been lost for a longer or shorter \time. Other cases he attributes to the plants being of crossed parentage, and to the buds reverting to one of the two parent forms. There are still many cases left, ^however, in which what appear to be absolutely new characters present themselves. These can only be at- * " Animals and Plants," 1. p. 362. f " Animals and Plants," i. p. 440. BLASTOGENIC VARIATIONS. 187 tributed to so-called spontaneous variabilify. Though in individual instances it may be difficult or impossible to assign any reason for the sudden appearance of such " spontaneous variability," yet various observations in- cline one to believe that it is probably, after all, only a special instance of variations due to changed conditions of life. Thus it is noticeable that all plants which have yielded bud-variations have likewise varied greatly by seed. They seem, in fact, to possess an inherent varia- bility. Again, almost all plants showing bud-varia^ tion have been highly cultivated for long periods, in many soils and under different climates. On the other hand, plants living under their natural conditions are very rarely subject to bud-variation. In some in- stances, as when all the fruit on a purple plum tree sud- denly becomes yellow, or all the fruit on a double- flowered almond suddenly becomes peach-like, we seem to perceive a direct result of changed condi- tions of life; but more often than not we are com- pelled to conclude that the connection is only an indirect one. t Darwin points out * that it is " a singular and inex- plicable fact that when plaUts vary by buds, the varia- tions, though they occur with comparative rarity, are^ often, or even generally, strongly pronounced." In plants raised from seed, however, the variations are al- most infinitely numerous, but their differences are gen- erally slight. Bud-variations clearly seem, therefore, to be true discontinuous variations, and not merely exaggerated normal variations. As to the ultimate cause of their production, we are as completely in the *L. c.,i. p. 443. + 188 BLASTOGENIC VARIATIONS. dark as we were for the analogous phenomena observed in sexually produced forms. We can only conclude, as we did then, that the process of development at some point takes on a new and abnormal departure, the direct or indirect result of changes of environmental condi- tions. CHAPTER VI. CERTAIN LAWS OF VARIATION. Effect of environment on growth diminishes rapidly from time of impregnation onwards Reaction of an organism to environment dependent on nature of organism Rapidly diminishing rate of growth in man and in the guinea-pig with progress in develop- ment Variability also diminishes with growth Effect on growth once produced, probably never eradicated Increased variability of sparrow and of periwinkle in America Relation between vari- ability and want of adaptation to environment Variability of migratory and non-migratory birds Does domestication increase variability? BEFORE entering on the discussion of the causes of so-called somatogenic variations, i. e., of acquired char- acters, it will be well to examine at some little length certain more or less general laws and conditions which control their acquisition and retention. This is the more necessary, as the matter has received such very little attention hitherto. It seems to have been more or less tacitly assumed that external conditions act equally powerfully at all periods in the growth of a de- veloping organism, whilst the persistence or otherwise of any effect, once produced, has scarcely been debated at all. In what way, then, does a developing organism react in its growth to the conditions of its surroundings? It would probably be concluded that any given change of environmental condition would produce more effect in 190 CERTAIN LAWS OF VARIATION. the earlier stages of development than in the later, but what is the numerical expression of this difference? Such an expression can be obtained in two ways: di- rectly, as the result of experiment ; and indirectly, from certain considerations as to the rate of growth and per- sistence of variations. To determine the effect of environment on growth, almost any organism can be made use of, but it would obviously be exceedingly troublesome and laborious to work with the higher organisms, such as mammals. In them the earliest stages of embryonic development would be especially difficult to reach. With many of the lower organisms, however, all such difficulties are avoided, and an inexhaustible supply of material can readily be obtained. For these reasons the author at- tempted to investigate the question at issue by observa- tions on the larvae of sea-urchins. The method of ex- periment has already been indicated in Chapter IV., so that it is unnecessary to refer to it again. The object in view was to determine the permanent effect of some abnormal environmental condition, act- ing at various periods of development, on the size of the larvae. It was found that their growth practically ceased after 6 or 8 days, and hence any effect then found to be present was fixed and ineradicable, so far as the larval stage of the organism was concerned. The most convenient environmental condition to work with proved to be temperature, for it is easily controlled, and the effect produced may be considerable. Thus larvae kept at 10 C. during growth are some 25 per cent, smaller than those kept at 20 C.* *Phil. Trans. 1898, B. p. 481. CERTAIN LAWS OF VARIATION. 191 To determine the effect of temperature acting at the time of impregnation, portions of the ova and sperma- tozoa were shaken from the ovaries and testes in small beakers of sea-water. After bringing these to the re- quired abnormal temperature, their contents were mixed, and the mixed solution kept at the same tempera- ture for, in most cases, an hour. It was then poured into a jar holding 2 to 4 litres of sea-water at the normal temperature. The ova, now fertilised, were allowed to develop under as constant conditions as possible for 8 days, and the larvae were then killed and measured in groups of 50, as already mentioned. Other ova, kept at the time of impregnation at a normal instead of an abnormal temperature, were allowed to develop under otherwise exactly similar conditions, and so afforded " control " or " normal " larvae, against which the effect produced in the other larvae by exposure to the abnor- mal temperature could be determined. In the accom- panying table the results obtained in the various obser- vations are collected: * *83* H> 'ilig H p N P P P PERCENTAGE DIMINUTION PRODUCED IN SIZE OF LARY^E. I! 2 PH H H 3 1 hour about 8 11 25.5 about 19 8.5, 1.8, 8.3, 0.0, 1.2, 2.5, 8.7, 2.5, 3.6, 4.2. 3.3, 13.8. 4.7. + 1.0, 5.1, 9.4, 6.0. 4.1* 5.9 1 or 3 min. " 8 .3, 6.9, 3.6, 2.4. 3.3 " 25.5 " 10.6, 1.9, 2.7. 5.1 10 seconds 8 or 25.5 2.4, 2.7, + .2, 1.9. 1.7 Here we see that exposure for an hour to a tempera- ture of about 8 C. at the time of impregnation, instead * Vide Phil. Trans. 1895, B. p. 582, and Proc. Roy. Soc., vol. xlvii. p. 85, 1900. 192 CERTAIN LAWS OF VARIATION. of one of 19 C., produced, in ten observations, an aver- age diminution of 4.1 per cent, in the size of the larvae. Temperatures a few degrees above the normal acted even more unfavourably, one of 25.5 producing, in seven observations, an average diminution of 5.9 per cent. It follows, therefore, that at the time of their impregnation, the ova are most extraordinarily sensi- tive to the temperature of their surroundings, be it ab- normally high or abnormally low. Further observa- tions showed that they were also very sensitive to an- other condition, viz., salinity of the water, though not to the same extent as to temperature. It seems very probable, therefore, that at this period they are very sensitive to all conditions of environment, whatever their nature. The results contained in the lower half of the table are even more remarkable than those in the upper. Thus, if the ova were kept at about 8 or at 25.5 for only one to three minutes after the mingling of the solutions containing the ova and spermatozoa, and after this short period were poured into jars of water at nor- mal temperature, an average diminution in size of re- spectively 3.3 and 5.1 per cent, was effected! The ob- servations made are not very numerous or regular, and hence not much importance can be attached to the actual figures, but one is justified in concluding that the effect produced is not so very much smaller than when the period of exposure to the abnormal temperature g_o S < *& Kg J 1 ! ^*Jr 4C. 288 hours 2 4C. 471 hours A 8 210 4 8.75 192 1 9.5 139.2 5 12.12' 126 2 13 96 2 16 60 3 14 90 2 22 27.5 2 18 60 3 24 25.5 2 22 40 6 26 P 21.5 1 Within the last few years, a considerable number of observations have been made. Thus Lillie and Knowl- ton f experimented with the ova of Amblystoma tigri- num and of the frog Rana virescens. The time of de- velopment from the first or second cleavage to the last *Phil. Trans. 1850, p. 431. Zo61. Bulletin, vol. i. p. 179. 226 THE EFFECT OF TEMPERATURE stage of disappearance of the yolk plug was determined, these being the most sharply marked periods. The temperatures varied between 4 and 26 C., several ob- servations being made at each temperature, and means taken. From tbe above table we may gather that at the 24 23 22 21 2019 181716 15 1413 12''ll' > 10 9' > 8 7 6 5 4 3 2 1 I?IG. 22. Effect of temperature on growth of tadpole. highest temperatures employed the rate of develop- ment was respectively 7.2 and 21.9 times more rapid than at the lowest. In the accompanying figure are reproduced the re- AND OF LIGHT. 227 suits obtained by O. Hertwig * upon the ova of Mana fusca. The curves represent the number of days re- quired by the ova, kept at different temperatures, to reach certain definite stages. The ordinates indicate the time in days after fertilisation, and the abscissae the temperature. For the lowest curve the stage to be reached was that of a gastrula with the blastopore clos- ing in, and we gather from this curve that the time re- quired at a temperature of 1 C. was 23 days; at a tem- perature of 6, 4.9 days, and at a temperature of 24, only a single day. Stage II was that of an embryo having a rudimentary medullary plate, with its edges rising and separated by a broad cleft; Stage III that of an embryo with a closed medullary tube, and with a distinctly marked head: Stage IV that of a more elon- gated embryo with an obvious tail, but with gills not formed; Stage V that of an embryo 5 mm. long, with rudiments of gills; Stage VI that of an embryo 7.5 mm. long, with well-developed gill tufts and tail 3.5 mm. long; Stage YII that of an embryo 9 mm. long, with a tail 5 mm. long. The curves representing the times of growth to all these more advanced stages are very simi- lar to each other and to the first curve. In the observations thus far quoted, the highest tem- perature employed was 26, and this proved to be also the most favourable temperature for growth. Higher temperatures still may produce an adverse influence, as we have already observed in the case of Echinoid larvae. For the growth of tadpoles' tails, Lillie and Knowlton found the " optimum " temperature to be 30, the rate of increase in length then being 10.6 times greater than *Arch. f. mik. Anat., li. p. 319, 1898. 228 THE EFFECT OF TEMPERATURE at 10. At 31 to 34.9, however, the rate was only 9.0 times greater. Better instances of the more and more unfavourable influence of increasing high temperature are found amongst plants, as in them the optimum temperature is much further removed from the " maxi- mum " temperature (the highest temperature at which growth can take place at all) than it is in animals. The following table shows the increments in the length of the hypocotyls of various plants in a period of 48 hours, as determined by Koppen and by De Vries:* KOPPEN. DE VRIES. 2 03 1 -B 00 M 9 ad 'ia fli 2 j a *3 GO 1! P-i-H g. 9J5 ^ a I || a.| i I" 5 II 1 m Jl fi-S 13 14 1C 9.1 mm. 5.0 mm. 15'. 1 3.8 mm. 5.9 mm. 1.5mm. 18.0 11.6 8.3 1.1 mm. 21.6 24.9 38.0 20.5 23.5 31.0 30.0 10.8 26.6 54.1 53.9 29.6 27.4 52.0 71.9 44.8 28.5 50.1 40.4 26.5 30.2 43.8 38.5 64.6 30.6 44.1 44.6 39.9 33.5 14.2 23 69.5 33.9 30.2 26.9 28.1 36.5 12.6 8.7 20.7 37.2 10.0 0.0 9.2 The optimum temperature was about 27 for every plant but one, viz., Zea mats, and in this case it was 33.5. The rate of growth at the optimum was, in the various plants, respectively 5.9, 10.8, 63.2, 13.7, 12.2, and 29.9 times greater than at the lowest temperature at which it was observed, and respectively 4.3, 6.2, 3.4, * Quoted from Vines' " Physiology of Plants," p. 293. AND OF LIGHT. 5.2, and 4.9 times greater than at the highest tempera- ture. In one plant, however, the temperature of 37.2 was sufficient to stop all growth, so that this was a slightly supra-maximal temperature. That the increased growth produced by warmth leads, at least in some cases, to an actual increase in the size of an organism, is proved by some f urjbher observations of the author on Echinoid larvse. 'No special determina- tions were made as to the effect of temperature on growth, but it was noticed that, whilst at a temperature of 13.8, the ova took about 22 hours to reach the free swimming blastula stage, they took only about 5 hours at 24. We may assume, then, that the rate of growth increases rapidly with temperature. The effect of various temperatures on the actual size of the larvae may be gathered from the following table:* STRONGYLOCEN- SPII^ERECHINUS. ECHINUS MICROTU- TROTUS. BERCULATUS. TEMPERATURE OP DEVELOPMENT. t! el < J3 S"8) 51 *4 11 if 11.4 100.0 100.0 100.0 100.0 100.0 100.0 15.9 113.5 143.4 109.4 287.0 113.4 116.3 20. 4* 120.6 156.8 104.6 327.2 124.5 106.6 23.7 122.5 149.1 100.6 386.7 123.9 113.7 Larvae of three different species were allowed to develop at four different temperatures, and measured after 8 days' growth in respect of both their body length and their anal arm length, and this latter measurement was calculated as a percentage on the former. In the case *Phil. Trans. 1898, B. p. 479. 230 THE EFFECT OF TEMPERATURE of both Strongylocentrotus and Echinus plutei, the body length increases considerably with the tempera- ture up to 20.4, when it is more than 20 per cent, greater than in plutei grown at 11.4, but it is prac- tically unaffected by a further rise. In Sphcsr echinus, on the other hand, the optimum temperature appears to be 15.9, and a further rise of temperature acts unfav- ourably. The effects on the arm lengths differ con- siderably more than those on the body lengths. In Strongylocentrotus this dimension is half as long again at 20.4 as it is at 11.4, but in Echinus it is only very little affected. In the Sph&rechinus pluteus, on the other hand, it is nearly four times longer at 23.7 than it is at 11.4. Each organism, therefore, in respect of each portion measured, reacts in a different manner to changes in the temperature of development. Some observations of Standfuss * upon the larvae of certain Lepidoptera, show that the effect of tempera- ture on growth is not necessarily in the direction of in- creased size. Thus he found that whtai, as was often the case, the larval period was shortened by raising the temperature, the size of the adult insects resulting therefrom was correspondingly reduced. For example, a pair of A. fasciata, of which the wings measured re- spectively 46 and 48 mm. across, produced three speci- mens measuring only 36 to 39 mm., when the larval stage was reduced to 68 to 87 days, and the pupal to 15 to 20 days, by subjection to a temperature of 25 to 30. On the other hand, some other eggs from the same original pair of A. fasciata, which, though exposed *The Entomologist, vol., xxviii. p. 69, 1895. (Translated from the German by Dr. F. A. Dixey.) AND OF LIGHT. 231 to the same high temperature, developed more slowly the larval period taking 142 to 163 days, and the pupal 25 to 31 days yielded specimens having a wing meas- urement of 55 to 57 mm. It would seem, therefore, that high temperature may so hurry forward the time of onset of the metamorphosis from larva to pupa, that there is insufficient opportunity for adequate larval feeding and growth, and a consequent dwarfing of the adult imago. If, however, there be no curtailment of the normal period of feeding, the high temperature may produce a considerable increase in the size of the individuals. These conclusions are supported by sev- eral other observations. The permanent effects of temperature on size are probably very considerable among many of the Mol- lusca. Thus it has often been noticed that snails living in cold and exposed positions are considerably dwarfed in comparison with those living in warmer regions, but, as far as I am aware, no exact comparisons have been made. Even if this had been the case, it would not be permissible to ascribe the differences to the direct result of temperature, as this might have acted indirectly, through the vegetation. Certain observations of Mo- bius * on marine Mollusca seem, however, to demon- strate the direct effects of temperature changes. Thus it was noticed that the Molluscs in the Eastern basin of the Baltic are much more stunted than those in the Western. For instance, Mytilus edulis is only 3 to 4 cm. long near Gothland, whereas at Kiel it attains a length of 8 to 9 cm. Also in the Eastern basin the cal- careous layers of certain shells such as M ya arenaria * Report on " Pomerania " Expedition, p. 138. 232 THE EFFECT OF TEMPERATURE are extremely thin. " These remarkable variations are, no doubt, to a large extent due to the violent changes of temperature which are experienced in the Baltic, and by which the steady development of the animals in question is thrown out of gear. The same species occur on the coast of Greenland and Iceland, where they at- tain a considerably larger size than in the Baltic, in spite of the lower mean temperature, probably because their development is not interrupted by any sudden change from cold to heat, or vice-versa."* The influence of a low temperature on the colour of marine Mollusca seems to be indicated by the observa- tions of Fischer f on the shells of the west coast of South America. Numerous species of these shells ex- hibit a remarkable degree of melanism, and it seems highly probable that " this concurrence of specific melanism (which stands quite alone in the world) is due to the cold polar current which impinges on the Chilian coasts, for the same genera occur on the opposite shores of the continent without exhibiting any trace of this mournful characteristic." $ It is very improbable, however, that this melanism is the direct result of the cold current. If so, why should it not be observed in other parts of the world, which are similarly visited by cold currents? More interesting and unequivocal effects of tempera- ture are afforded by the numerous experiments which have been made upon the wing colours and markings of * Quoted from Cooke, "Cambridge Natural History," vol. iii. p. 84. f Journ. de Conchy!. , xxiii. p. 105, 1875. \L. c., p. 85 AND OF LIGHT. 233 Lepidoptera. It has been known for more than sixty years that the two butterflies Vanessa levana and V. prorsa, formerly regarded as different species, are but seasonal forms of one and the same species. Thus V. levana emerges in the spring, breeds immediately, and produces adult V. prorsa progeny in the same summer. The progeny of these insects pass the winter as chrysa- lids, and emerge the next spring as V. levana. The levana form is characterised by a yellow and black pat- tern on the upper side of the wings, whilst the prorsa form has black wings with a broad white transverse band. The lower surfaces differ only slightly. It is a natural supposition that these changes of colour marking are dependent in some way on tempera- ture, and Dorf meister * proved that this is actually the case. By the application of warmth to the pupae he succeeded in producing prorsa out of the offspring of prorsa, and by the application of cold he obtained from levana not the pure levana form, but one intermediate between it and prorsa. This intermediate form, which has occasionally been observed in nature, is termed V. porima. These experiments were repeated and ex- tended by Weismann, and by employing a greater de- gree of cold he succeeded in obtaining levana from levana; but he found that prorsa was only exceptionally reared from prorsa by the application of heat. The mode of action of the temperature is not so clear as might at first sight be imagined. The simplest ex- planation is to attribute the effect to the direct influ- ence of the warmth and cold, and this view of the in- *Mitt. des naturwiss. Vereins fiir Steiermark, 1864. See also Elmer's " Organic Evolution," English Ed., p. 116, etseq. THE EFFECT OF TEMPERATURE fluence of warmth is actually held by Eimer.* Accord- ing to Weismann, however, the action, both for warmth and cold, is an indirect one. The change of the prog- eny of levana back to levana through the influence of cold he attributes to reversion to the ancestral form, for there is practically no doubt that levana is phylogeneti- cally the older form of the two. He considers that prorsa has slowly arisen through the gradual increase in the warmth of the climate, or is a seasonally adaptive form, and that its occasional production from the prog- eny of prorsa is due to the high temperature unduly stimulating the development of the prorsa " determi- nants." A clearer case of the direct influence of warmth and cold is afforded by Polyommatus phlceas, the Small Cop- per butterfly. By exposure of the pupse to various temperatures, Merrifieldf obtained the following re- sults : TEMPERATURE. TIMB OP EMERGENCE. COLORING OP SPECIMENS. 27 32 C. 6 days Spots large, not sharply defined; dusky suffusion of fore wings. about 21 p 1115 days Spots smaller; copper colour more vivid; black more intense. about 14 Q 2223 days Copper colour still more vivid; copper band on hind wings broader. about 7' 5759 days Effects intensified. .5;thenl3 p 10 weeks; Extreme effects, especially in smallness of then 5 wks. spots and breadth of coppery band on hind wings. Here we see that the temperatures ranged from about 30 C., or 85 F., to just above the freezing point. The * " Organic Evolution," p. 122. f Trans. Ent. Soc. 1893, p. 55. AND OF LIGHT. 235 times of emergence of the butterflies from the chrysa- lis varied from 6 days to no less than 15 weeks, and probably if the low temperature had been continued in this latter case, the time would have extended to many months, or there may have been no emergence at all. It will be seen that the principal effects produced by warmth are a dusky suffusion of the fore wings, and by cold an intensity of colouring in both the coppery and dark parts, the enlargement of the copper band on the hind wings being an especially marked feature. In fact these " cold " specimens were very similar to those caught in England, Germany, and similar latitudes, whereas the " warm " specimens were similar to the variety eleus, which is found in Southern Europe. Merrifield therefore came to the conclusion that the difference in the appearance of these local forms " is not necessarily to be attributed to the existence of races of different colouring, but may be owing to the differ- ence between the temperatures to which the individuals are exposed in the two climates." Weismann has shown,* however, that the modifications cannot be en- tirely due to the direct effects of temperature. Thus none of the specimens obtained by exposing pupse of a German stock to high temperature were so dusted with black as the darkest forms of the southern variety eleus, whilst conversely, none of the specimens obtained by exposing the pupse of a Neapolitan stock to a low tem- perature were so light-coloured as the ordinary Ger- man form. "The German and Neapolitan forms are therefore constitutionally distinct, the former tending much more strongly towards a pure reddish-gold, and * " Germ-Plasm," p. 399. 236 THE EFFECT OF TEMPERATURE the latter towards a black colouration." Weismann thinks that the two varieties may have originated owing to a gradual cumulative influence of the climate, the slight effects of one summer or winter having been transmitted and added to from generation to genera- tion. Such a cumulative effect can be accounted for satisfactorily by supposing that the temperature not only affects the " primary constituents " of the wings of each individual i. e., a part of the soma but also the corresponding " determinants " of the germ-plasm contained in the germ cells of the animal. Arguing from experiments on about 5000 pupse, Standfuss * has endeavoured to classify under five dif- ferent headings the effects which temperature changes may produce in Lepidoptera. (1) They may give rise to seasonal forms having a similar aspect to those occurring among the palsearctic fauna at certain definite seasons of the year. For in- stance, pupse of Vanessa c-album (Comma butterfly), kept at 37 C., gave origin to the light coloured, yel- lowish brown form of butterfly, especially pale on the under surface, whilst those kept in a refrigerator pro- duced the form with a considerably darker under side, in many cases mingled with a moss-green tint. Also this form had much more sharply defined markings, and a more deeply indented margin to the wings. Both these forms, be it noticed, occur in nature at the present time. Again, by exposing pupse of P. machaon (Swal- low-tail), to a temperature of 37 C. ? insects were ob- tained which bore a perfect resemblance to those that * The Entomologist, vol. xxviii. pp. 69, 102, and 145. (Translated from the German by Dr. F. A. Dixey.) AND OF LIGHT. 237 fly in August in the neighbourhood of Antioch and Jerusalem. Pupae kept at 5 to 8, however, yielded the common Swiss and German form of butterfly ob- tained from hibernated pupae. (2) Local forms and races such as occur constantly in certain definite localities may be produced. For in- stance, exposure of pupae of V. urticce (Small Tortoise- shell) to warmth produced specimens somewhat similar to the variety ichnusa, whilst cold produced some speci- mens which strongly recall the North American V. milberti, and others which were indistinguishable from the northern variety polaris. Again, warmth acting on pupse of V. cardui (Painted Lady), gave an extraordi- narily pale form, like those found in very different parts of the tropics. Cold, on the other hand, gave specimens with a very recognisable darkening of the whole insect, such as is exhibited by a form found in Lapland. (3) There may arise forms of an entirely similar as- pect to some which are also found exceptionally under natural conditions, i. e., aberrations. For instance, warmth, acting for a brief period on V. cardui, produced a few specimens of the aberrant form elymi. Cold, acting on pupse of V. io (Peacock), produced a variety ab. fischeri, which exhibits a reduction in the number of the blue scales on both fore and hind wings. In these and other characters an approach to the type of V. urticcE is perceived. Such observations as these sug- gest that a large number of the aberrations occurring in nature may have actually arisen through the influ- ence of abnormal temperature conditions. (4) There may be produced phylogenetic forms; 238 THE EFFECT OF TEMPERATURE forms, that is, which are nowhere to be found on the earth at the present day, but which may have existed at past epochs. Such a result may have been effected through modification of temperature conditions hav- ing taken place in the actual habitat of the species, or from the species having migrated to a more southerly or northerly region. The variety fiscberi of V. io, just mentioned, is probably a phylogenetic form. The same may be true of a variety rcederi of V. antiopa (Camber- well Beauty), which Standfuss obtained by keeping the pupae in a refrigerator. Again, exposure of the pupae of V. atalanta (Red Admiral) to warmth, produced specimens approximating towards V. callirrhoe and its local forms, such as var. vulcanix, which are found in the Canaries: i. e., to forms which may resemble the common ancestor of these species. Other forms were produced which may perhaps be destined to arise in the future, in that they are further removed from the type of related species, instead of approximating to them, like the true phylogenetic forms. For instance, the widely diverging specimens obtained in a few instances by the action of warmth on V. antiopa, may belong to this class. This variety has been named daubi by Stand- fuss. (5) Finally, there is still a small unexplained residue of modifications produced by temperature changes. This possibly represents the direct reaction of the indi- vidual species, completely independent of, and uncon- trolled by, any inherited developmental tendency. It will be noticed that the principle of reversion is called in by Standfuss to account for one of his five groups, but Weismann, Dixey, Fischer, and others are AND OF LIGHT. 239 inclined to extend its scope to some of the other cases as well. Thus Weismann formerly made use of it to account for seasonal dimorphism, though now he rather withdraws this opinion.* According to Fischer, t both very low and very high temperatures are equally capa- ble of determining reversion by acting simply as ex- citants. A moderate elevation of the temperature, on the contrary, may give rise to new modifications which are not phylogenetic, but which actually occur in warm climates. Dixey,lf arguing especially from Merrifield's observations on V. atalanta, and Merrifield himself, from these and other observations, have come to the conclusion that reversion may be occasioned by ex- posure both to high and to low temperatures, but that the kind of effect produced is different in the two cases. Eimer is of the opinion that only cold has the power of causing a reversion to an ancestral form, the effect of warmth being " evidently a direct effect." || In sup- port of his views, he refers to Weismann's experiments on Pieris napi (Green-veined White), and V. levana- prorsa. The former butterfly occurs in a summer and a winter form, the winter being the darker. There is also a variety of P. napi, viz., bryonice, which is found in the Swiss Alps and in the polar regions, and which can be described as a very dark variety of the winter form of P. napi. This bryonicz is in all probability the ancestral form of P. napi, whilst the winter form, and * The Entomologist, 1896, p. 240. f "Transmutationen der Schmetterlinge infolge Temperaturan- derungen," Berlin, 1894. $ Trans. Ent. Soc. 1893, p. 72. Trans. Ent. Soc. 1894, p. 425. | " Organic Evolution," p. 125. 240 THE EFFECT TEMPERATURE subsequently the summer form, of the common P. napi, have probably arisen gradually from it through the in- fluence of a warmer climate. Now Weismann found that he was unable to convert Iryonice into napi by the action of warmth, though he could by the application of cold readily change the summer form of napi into the winter form. Similarly, also, the progeny of V. levana are readily converted by cold into levana, but only ex- ceptionally can the progeny of V. prorsa be converted into prorsa. Now, as already mentioned, levana is probably the ancestral form, and so, in both this case and that of P. napi, cold readily produces what is probably a phylogenetically older form, whilst warmth generally has no effect. Certain observations by Merrifield * also afford some support to Eimer's view, for he found that " the capability of being turned during the pupal period from one type partially into the direction of the other exists in both the summer and the winter type, but is much greater in the former than in the latter." With regard to the critical period at which tempera- ture especially exerts its influence, there is a general consensus of opinion that it is confined to the pupal stage, and in most cases also to the first part of this stage. Dorfmeister f concluded that temperature exerted its greatest influence during the change from the larval into the pupal stage, or shortly afterwards. Weismann J noticed that in V. prorsa-levana it acted only at the beginning of the pupal stage. Standfuss, * Trans. Ent. Soc. 1892, p. 53. f Vide Eimer's " Organic Evolution," p. 131. j " Germ-Plasm," p. 402. AND OF LIGHT. 241 in almost all the observations above referred to, exposed his pupae to warmth for about three days, and then kept them at the room temperature until they emerged, this generally occurring 4 to 10 days later. The exposure to cold generally extended to about 30 days, and emer- gence took place after about 11 days more at room tem- perature. As the effects obtained by him are just as great, if not greater, than those obtained by other ob- servers, it would seem quite clear that in the forms he employed the critical time for temperature is cer- tainly during the first portion of the pupal period. However, Merrifield,* in his observations on the sum- mer and winter forms of P. napi, found the critical time to be in the last days of the pupal period, a directly opposite result to that of Weismann for the same insect. Weismann f explains the apparent contradiction by supposing that in P. napi adaptive and direct sea- sonal dimorphism are mixed. The species may have adapted itself to the seasons of the year by a double protective colouring, and the critical period for the de- termination of the adaptive form may be at the begin- ning of the pupal period. The direct reaction of the species to temperature may, however, as Merrifield found, be determined only at the end of the pupal period. In his experiments with P. phl&as, Merrifield found that pupae kept at 0.5 C. for ten weeks, and then ex- posed to a temperature of 32 for six days, gave speci- mens with features very similar to those obtained from pupse kept throughout at a temperature of 27 to 32. F * Trans. Ent. Soc., 1893, p. 55. t The Entomologist, 1896, p. 240. 242 THE EFFECT OF TEMPERATURE The reason of this is probably that a temperature of 0.5 is so low that it paralyses all tissue changes in the pupae, and at the end of ten weeks the stage of develop- ment is no further advanced than at the beginning. Thus the time of emergence of these pupae, after trans- ference to a temperature of 32, was just as long as for those kept only at this temperature. Arguing from his experiments on two moths, Selenia illustraria and Ennomos autumnaria, Merrifield * came to the conclusion that, in their case at least, the markings were chiefly affected by the temperature ex- perienced during the earlier part of the pupal period, whilst the colouring was " chiefly affected during the penultimate pupal stage, i. e., before the colouring of the imago begins to show." A low temperature dur- ing this latter period causes darkness, and a high tem- perature the opposite effect. Thus, by difference of treatment, it was found possible to obtain from the same brood individuals showing (1) summer markings with summer colouring; (2) summer markings with an ap- proach toward spring colouring; (3) spring markings with summer colouring, and (4) spring markings with almost spring colouring. We see, then, that in some cases seasonal dimorphism is a direct response to temperature, or is a somatic modification, whilst in other and perhaps the majority of cases it is only indirect, the temperature acting as a stimulus to arouse a blastogenic variation. When the response is direct, low temperature generally induces a darkening of colour, as, for instance, in V. urticcZj Las- ciocampa quercus (and callunce), Arctia caja and E. * Trans. Ent. Soc. 1891, p. 55. AND OF LIGHT. 243 autumnaria. In these forms, the darkening is caused either by the general colour being obscured, or by the size and general intensity of the dark markings being increased, or by both conditions.* In P. phlceas, as we have seen, low temperature causes a lightening of colour. When the response to temperature is indirect, the effect is as often as not in one direction as in the other, and there are generally more considerable changes in the markings, as well as in the general colouring. Upon the higher animals temperature probably acts but seldom as a direct cause of variation. The white coat which many quadrupeds develop on the approach of winter in northern and arctic climates is probably in great part a seasonally adaptive change, but it may also be to a certain extent the immediate, though perhaps only indirect, response to cold. This seems to be proved by an observation of Sir J. Ross on a Hudson's Bay Lemming. f This animal was protected from the low temperature by keeping it in the cabin, and had in con- sequence retained its summer coat through the winter. On exposing it in a cage on deck, where the temperature was 30 below zero, the fur on the cheeks and a patch on each shoulder became perfectly white during the first night. After another day's exposure " the patches on each shoulder had extended considerably, and the posterior part of the body and the flanks had turned a dirty white. . . At the end of a week it was entirely white, with the exception of a dark band across the shoulders, prolonged posteriorly down the middle of * Vide Merrifield. Trans. Ent. Soc. 1892, p. 33. f Appendix to Second Voyage. Nat. Hist. , p. xiv. /1835. Quoted from Poulton's " Colours of Animals," ed. i. p. 94. 244 THE EFFECT OF TEMPERATURE the back." No further change took place, and the ani- mal died of the cold a few days later. Examination of the fur showed that only the tips of the hairs had be- come white, so that on cutting these off, the coat re- gained its original dark colour. The observations of F. H. "Welch * on the American Hare (Lepus Americanus) throw further light on the nature of the change. Early in October the whiskers and a few of the longer hairs on the back were observed to become white at the tip or throughout. During No- vember a new and rapid growth of stiff white hairs ap- peared on the sides and back, these hairs being easily distinguishable from the autumnal hairs which were gradually turning more and more white, in that they were invariably white throughout. We have in this animal, therefore, a new white crop of hairs of gradual growth, or a blastogenic variation, stimulated to de- velop under stress of cold, and a rapid and direct trans- mutation of parts of the dark hairs to white; i. e., a somatic modification. Professor Poulton f explains this latter change as an indirect influence of cold upon the nervous system which presides over the nutritive and chemical changes involved in the growth of the hair. This probably leads to the production of large numbers of gas bubbles in the hairs, and thereby induces an ap- parent whiteness, in spite of the fact that the pigment is still present. In that the tips of the hairs are first affected, however, rather than the bases, it seems to me possible that the cold acts directly on the hairs them- selves, and not indirectly through the nervous system. *Proc. Zo5l. Soc. 1869, p. 228. f " Colours of Animals," p. 100. AND OF LIGHT. 245 It should be pointed out that some animals, such as the sable, musk-sheep, and raven, retain their dark colour throughout the Arctic winter, so that the reaction of the above-men- tioned animals to cold, whether direct or indirect, is a special and not a general phenomenon. Light. The effect of light upon growth, especially in plants, is well known to be very considera- ble. One might infer, therefore, that differences in the intensity of the light to which an organism is subjected would form a potent cause of variation. Such is actu- ally the case among members of the Vegetable Kingdom, though only exceptionally so among those of the Animal Kingdom. If plants be allowed to grow in absolute darkness, they, as a rule, become very much elongated in form, whilst their leaves are small and ill-shaped. The accompanying figure shows the relative growth of two seedlings of Sinapis alba of the same age, one of them reared in the dark ; and the other in ordi- nary daylight.* Sachs found that potato tubers grown in darkness for 53 days produced sprouts from 150 to 200 mm. high, whilst similar ones grown in day- * From Strasburger, Noll, Schenck, and Schimper'a " Textbook of Botany." Quoted from Davenport's "Experimental Morphology," p. 418. FIG. 23. Seedlings of Sinapis alba. E, reared in the dark. N, reared in ordinary daylight. 246 THE EFFECT OF TEMPERATURE light were only 10 to 13 mm. high. Again, he found that the hypocotyl of the buckwheat (Fagopyrum) reached a height of 35 to 40 cm. in the dark, whilst it grew only to 2 or 3 cm. when freely exposed to light. K. Goebel* has shown that if cactuses are cultivated in darkness, their form changes completely. The young shoots are rounded, and fail to show the angular irregu- larities of form which increase the surface capable of effecting assimilation under the influence of light. Darkness conduces to increased growth, therefore, or conversely, light tends to retard growth. That this is the case is well shown by an observation of Wiesner.f This observer exposed seedlings of the vetch (Vicia sativa) under a glass globe to sunlight for YJ hours. When placed horizontally, so as to get the full force of the sun's rays, no growth whatever occurred, but when placed vertically, so that the growing part of the seed- ling was more or less protected by its leaves, there was an increase in height of about .8 mm. On the other hand, a control seedling kept in a darkened globe grew about 2.8 mm. in the same period. This retarding effect of light is not universal, however. It is practically ab- sent in some cases, as of the yam and of a wild gourd (Bryonia), and in those plants whose rapidly growing parts are sheltered from the sun's rays by protecting coverings it is but little evident. Still Sachs' conclu- sion as to the effect of daylight on growth probably ap- plies with greater or less force to the majority of plants. Thus he found J that during the night the growth * Flora, Ixxxi. p. 96. f Davenport's "Experimental Morphology," p. 41. \ Arb. aus der Bot., Inst. Wiirzburg, i. p. 99. AND OF LIGHT. 24? gradually increases, and reaches a maximum at day- break. It then diminishes to a minimum a little before sunset, after which it rises again. It is not to be imagined that because daylight retards growth it is unfavourable to the proper develop- ment of a plant. For instance, Karsten * found that whilst a kidney bean reared in the dark for a month or two weighed 20 per cent, more than one reared in the light, yet the leaves did not weigh a fifth as much. Again, Clayton t allowed six bean plants to grow in a spot where they would catch all the sunshine of the day, whilst six other similar plants were protected by a boarding, which effectu- ally screened off the sun. When freshly gathered in October, the weight of the beans and pods of the exposed plants was to that of the protected as 99 : 29, whilst the weight of the dry beans was as 16 : 5. The next year, the weight of the fresh beans and pods ob- tained from the sunshine-grown seed of the previous year was half as much again as in the case of the plants from shade-grown seeds, in spite of the fact that all of the plants were now grown in sunshine and under pre- cisely similar conditions. " In the fourth year plants with an exclusively shady ancestry produced flowers, but failed to mature fruit." The intensity of the light to which a plant is exposed may considerably affect its form and structure, as well as its rate of growth. Thus the effect of direct sun- light, as compared with diffused light, on the absolute *Landw. Versuchs-Stat. , xiii. p. 176. Quoted from Davenport's " Experimental Morphology," p. 419. f Nat. Science, xi. p. 12. THE EFFECT OF TEMPERATURE size of leaves has been shown by Stahl* to consist chiefly in a reduction of the leaf surface. Accompanying this there is usually an increase in the thickness of the leaf. In addition to the reduction of size, Scott Elliott f has shown that there may be a considerable change in the form of the leaves, owing to the reduction in the length of the exposed leaves being much greater than the re- duction in the breadth. The accompanying table shows the average ratio of length to breadth in from 50 to 100 leaves of various grasses and plants, which were collected in the one case from sheltered and shady places, and in the other from the most exposed and driest spots known: NAME OP SPECIES. SHELTERED SPECIMENS. EXPOSED SPECIMENS. PER CENT. REDUCTION. Stenotaphrum glabrum, Paspalum distichum, . . . J 15.8 23.0 6.1 11.0 61.4 52.2 Cynadon dactylon, . Eragrostis ciliaris, . 7.3 32.4 4.5 17.5 38.4 46.0 Cenchrus echinatus, 22.5 19.8 12.0 Microrhyncus sarmentosus, 6.8 4.7 30.9 Lobelia Scaevola, . 2.2 1.6 27.3 Psiadia dodonsefolia, . 10.2 11.4 +11.2 Helichrysum emirnese, . 9.4 5.6 40.4 Spermacoce globosa, 6.0 4.2 30.0 Lycium capense, . * 1.9 1.9 0.0 Brexia madagascariensis, 1.8 1.5 16.7 Camptocarpus, sp., 4.1 3.6 12.2 Periploca ovata, 1.6 1.5 6.2 Commelina nodiflora, 3.2 2.9 9.4 Tanghinia venenifera, 4.7 3.6 23.4 Brachystephanus cuspid atus, 1.7 1.7 0.0 Monimia, sp., 2.0 1.8 10.0 Sida carpinifolia, . Vinca rosea, ..... 2.6 2.3 2.2 2.3 15.4 0.0 . * Jenaisch. Zeit., Bd. xvi. p. 102, 1882, and Bot. Zeit., 1880. fProc. Linn. Soc., vol. xxviii. (Botany), p. 375, 1891. AND OF LIGHT. 249 Of these twenty different species of grasses and plants (collected in Madagascar), we see that the ex- posed specimens had a reduced leaf length ratio in 16 cases, whilst in only one was the length actually in- creased. The average reduction for the whole series amounts to 21.0 per cent., or is very considerable. Similar results to these have also been obtained by Sorauer.* Upon members of the Animal Kingdom the direct effect of light is not nearly so considerable. Yung t found that tadpoles exposed to daylight during the first 25 to 60 days of development were about 16 per cent, larger than those kept in absolute darkness. He found also that eggs of the sea-trout, if reared in the light, hatched a day earlier than if reared in the dark, whilst pond snails (Lymruza stagnalis) hatched in 27 days in the light, as against 33 days when in the dark.J It is possible, however, that these effects were due rather to the presence or absence of heat rays than those of light. The most important influence of light in the produc- tion of variations in animals lies in its connection with the phenomena of pigmentation. Absence of light leads to diminution or even total abolition of pigmenta- tion, whilst its presence leads to an increase in some degree proportionate to the intensity of the light. This, at least, is the more or less direct action of light. The indirect action, through the intermediation of the nervous system, is, as a rule, exactly the reverse. A * Wollny's '* Forschungen a. d. Geb. Agricultur," Bd. ix. f Arch. ZoOl. Exper. et Gen., vii. p. 251. \ Davenport's " Experimental Morphology," p. 426. \ 250 THE EFFECT OF TEMPERATURE well-known instance of the direct action of light is found in the bronzing of the human skin following on undue exposure to the sun; but to what extent are we entitled to refer the black skin of inhabitants of the tropics to a similar, but more pronounced, action? Eimer * is of the opinion that the effect is the direct result of the more intense light and heat. Thus he found that in passing down the Nile valley from the Delta to the Soudan, the natives gradually became more and more dark-skinned, the further south they lived. The increased light and warmth, according to Eimer, lead to a greater flow of blood to the skin, and the con- sequent deposition of pigment. This effect is inher- ited, and has become a constant character. There is, of course, no warrant for laying down the law with such assurance as this, for one can easily imagine several other equally possible and plausible explanations to ac- count for the facts. For instance, pigmentation may be correlated with a greater resistance to the climate of hot countries, or with greater physical strength, and may have been increased by sexual selection. Still, Eimer's explanation may contain a distinct modicum of truth, and I hope to prove in a subsequent chapter that the heritableness of acquired characters such as increased or decreased pigmentation may be deduced without assuming anything further than the present state of knowledge legitimately warrants us in doing. The diminution or disappearance of pigmentation fol- lowing upon withdrawal of light is best illustrated by reference to the well-known cave animals. Of these, * " Organic Evolution," p. 87. AND OF LIGHT. 251 one of the most interesting is Proteus anguineus, which is found in the subterranean caves of the Karst Moun- tains about Adelsberg. This amphibian is almost white, but if kept for some time in the light, it gradu- ally becomes pigmented. Pigment cells are, in fact, still present in its skin,* and in all probability these are directly stimulated to exert their function by the action of the light. A similar effect of exposure to light has been demonstrated by Cunningham f for the under sur- face of the flounder (Pleuronectes flesus). This surface is normally quite white, but by keeping young flounders for nearly four months in a glass dish illuminated from beneath by a mirror placed at a proper angle, Cunningham found that 10 out of the 13 specimens ex- perimented with developed black and yellow chromato- phores. Three of the specimens showed well-developed bands of pigment, similar to those of the upper side, over the area occupied by the muscles of the longi- tudinal fins. Subsequently, Cunningham and Mac- Munn J succeeded in keeping flounders alive under these conditions of illumination for from 9J months to nearly two years. They found that the amount of pig- ment steadily increased with the duration of the ex- posure, so that ultimately almost the whole of the lower side might become pigmented. This colouration was (histologically) of exactly the same kind as that of the upper side in normal specimens, though it was never by any means so marked. Its production is more re- * Vide Poulton's " Colours of Animals," p. 91. f Zool. Anzeiger, xiv. p. 27, 1891. \ Phil. Trans. 1893, B. p. 765. 252 THE EFFECT OF TEMPERATURE markable than in the case of Proteus, in that pigment cells are entirely absent from the skin of the lower side of the normal Flounder.* The observations of List f upon certain Lamelli- branch Molluscs afford evidence as to the effects both of decrease and increase of illumination. List noticed that various species of Mytilus (gallo-provincialis and minimus), which had been collected in caves, were dis- tinctly less pigmented than usual. In fact, those ob- tained from the extreme ends of the two grottoes underneath the ruined Palace of Donn'Anna in the Bay of Naples were all of them pale or colourless. In- dividuals of the same species were also found in the dark underground tanks of the Zoological Station, and here again they were little, if at all, pigmented. Speci- mens of M. minimus were even found in the pipe through which the water is pumped from the sea into the Aquarium, and these were characterised by an ab- solute want of pigmentation. The converse observa- tions were made upon LitJiodomus dactylus. These molluscs, which are usually concealed in borings in the sand of the sea bottom half a metre deep, are pigmented only at the tip of the foot and the edge of the anal siphon, these being the only parts at all exposed to light. After keeping specimens for a year in a glass vessel exposed to daylight, however, the whole surface of the anal siphon became coloured an intense red brown, whilst the imperfect branchial siphon, the border of the mantle, the whole of the foot, and the other ex- posed parts, were pigmented also. *IUd., p. 767. f Arch, f . Entwickelungsmechanik, Bd. viii. p. 618, 1899. AND OF LIGHT. 253 Again, Vire * has obtained somewhat similar results in his observations on the Fauna of subterranean caves and streamlets in France. As regards the Crustacea, he found that Niphargus virei, which is of a rose colour, after a few weeks' exposure to light becomes covered with brown spots, and thus undergoes a rapid return to its ancestral form. On the other hand Gam- marus puteanus, when kept for ten months in the tanks of an underground laboratory, began to lose its gray- green pigment, and after twenty months most speci- mens had entirely lost it. Again, the common Gam- marus fluviatilis, when kept for fifteen months under- ground, developed organs of touch and smell which at- tained nearly half the size of those exhibited by the true cave Niphargus. The observations which have been made on Amblyop- sis (a fish found in the caves of the Mississippi Valley) do not agree with the above results. Thus Eigen- mann f states that the pigment is very abundant when the young fish are two months old, but even when these fish are kept in light during growth, they show a de- crease and not any increase of pigmentation, so that a ten-months' fish was found to have taken on the exact pigmentless condition of the adult. Both the pig- mented condition and the subsequent depigmentation are hereditarily transmitted, therefore, and seem prac- tically unaffected by changes of environment acting through a single generation. *"La Faune souterraine de France," Paris, 1900; vide Abstract by P. Kropotkin in Nineteenth Century, September, 1901, from which this reference is taken. f Biological Lectures, Wood's Holl, 1899, p. 124, 254 THE EFFECT OF TEMPERATURE The direct dependence of pigmentation on light seems to be proved by the generality of the reverse phenomenon as observed in cave animals. Whenever light is totally excluded, the pigmentation appears to become diminished or abolished, whatever class of the Animal Kingdom the individuals belong to. Thus there have been found more or less unpigmented Coe- lentera, Worms, Crustacea, Myriapoda, Arachnida, Coleoptera, Fish, and other animals in the various sub- terranean caves of Europe and North America.* However, the abyssal fauna of the ocean, occurring at depths such that (presumably) no light can penetrate, includes numerous species which are just as much pig- mented as those exposed to light. Thus Faxon f di- vides deep sea Crustacea into two types; (1) those living in the bottom mud, which are mostly pale in colour, and often blind; (2) those which swim freely, have well-de- veloped eyes, and are coloured bright red. He con- siders that this red colour is due to the absence of light at these profound depths, for S. Jourdain $ has shown that two different species of Crustacea, which are brown when exposed to light or partial darkness, become red when placed in total darkness. MacCulloch and Cold- stream have suggested a " theory of Abyssal Light " to account for the existence at profound depths of these Crustacea, and of the Fish, Mollusca, Crabs, and other animals with well-developed eyes. This hypothesis msists essentially in the idea that light diffused by * Packard, Memoirs of National Academy of Sciences, iv. p. 3, 1888. f Mem. Mus. Harvard, xviii. p. 251. \ Comptes Rendus, Ixxxvii. p. 302, 1878. AND OF LIGHT. 255 phosphorescent creatures is capable of taking the place of sunlight in those depths which the rays of the sun cannot penetrate." * The more striking and considerable effects produced by light in members of the Animal Kingdom are mostly confined to cases of so-called " Variable Protective and Aggressive Resemblance/' or reaction to the colour of the surroundings which either protects the animals from their enemies, or assists them to capture their prey.f Such a reaction is rarely, if ever, a direct re- sponse to light of the superficial tissue cells as a whole, or even of the sensitive pigment cells in the skin which have been gradually formed in the course of evolution through the agency of Natural Selection and other processes. It is an indirect effect produced by the in- termediation of the nervous system. This was first proved to be the case by Briicke J for the chameleon, and by von Wittich for the frog. The latter observer regarded the variations in colour as probably reflex in their nature, he attributing them to a peripheral gan- glionic apparatus in the skin itself. A few years later Lord Lister | took up the problem and correctly solved it, he concluding that in Rana temporaria " the eyes are the only channels through which the rays of light gain access to the nervous system so as to induce changes of colour in the skin." The very conspicuous changes which can be produced in this manner may be illus- * Quoted from Semper's " Animal Life," p. 85. f Vide Poulton's " Colours of Animals," pp. 81 to 158. JIV. Bd, d. mathemat. naturwiss., Classed. Kaiserl. Acad. d. Wissenschaft, Wien, 1852. Muller's Archiv. 1854, p. 41. | Phil. Trans. 1858, p. 627. 256 THE EFFECT OF TEMPERATURE trated by another quotation from Lord Lister's paper: " A frog caught in a recess in a black rock was itself al- most black; but after it had been kept for about an hour on white flagstones in the sun, was found to be dusky yellow, with dark spots here and there. It was then placed in the hollow of the rock, and in a quarter of an hour had resumed its former darkness. These effects are independent of changes of temperature." These changes of colour have been shown by Briicke, von Wit- tich, Lister, and others to be due to the pigment gran- ules of certain stellate cells in the skin varying in their degree of concentration towards the centre of the cell, and in their diffusion peripherally through the branch- ing processes. These pigment cells are often of differ- ent colours and are arranged in layers, so that widely different effects may be produced by varying degrees of concentration in them. That the reflex mechanism takes its origin in the eye, which is stimulated by the light reflected from the ani- mals' surroundings, was proved by Lord Lister in the following manner: He found that a frog with its eyes removed was totally unaffected by the colour of its surroundings. The nervous system still re- tained the capacity for acting on the pigment cells, however, as the frog, originally dark, became ex- tremely pale after struggling violently to escape. It was then placed in a bright light, but within half an hour became almost coal black again. Occa- sionally protectively coloured animals are found in nature showing a total want of adjustment to the colour of their surroundings. Thus Pouchet* noticed that one * Quoted by Semper, " Animal Life," p. 95. AND OF LIGHT. 257 single plaice out of a large number upon a bright sandy surface was dark-coloured, and Mcoll * noticed that in addition to the light-coloured trout usually seen in a chalk stream in Hampshire, very dark individuals occa- sionally appeared. In both instances, however, it was proved that the fish were blind, and therefore unable to respond to the stimulus of reflected light. Besides the amphibia, fish, and reptile mentioned, many other animals belonging to the same groups ex- hibit a similar power of rapidly adapting their colour to that of their surroundings. The power is also pos- sessed by many invertebrate animals. It is probably very common among Crustacea, and some cuttle fish can modify their colours with extreme rapidity. In Octopus vulgaris the protective resemblance is very striking, and so completely is it under the control of the nervous system that I have seen an individual change its colour from a dirty white to a dark brown in less than a second. It is amongst the Lepidoptera, how- ever, that our knowledge has been furthest advanced. The power of adaption has so far been proved to exist in this group during the larval and pupal stages only, though it is probable that a relatively small num- ber of perfect insects also possess it.f Again, it is present only in such pupae as are exposed, and has been found wanting in those of moths which are as a rule either buried in the earth or concealed in opaque cocoons. Professor Poulton J has shown, however, that the pupa of the Swallow-tailed moth forms an ex- * " Colours of Animals," p. 86. f " Colours of Animals," p. 110. $ Ibid., p. 111. 258 THE EFFECT OF TEMPERATURE ception to this rule. Also he has found that the cocoons themselves may undergo protective colour- ation. The first recorded instance of variable protective re- semblance in Lepidoptera is due to T. W. Wood,* who in 1867 demonstrated it for the pupae of the Large and Small White butterflies (Pieris brassicce and P. rapes). A few experiments were made from time to time by other observers, but it was not until 1886 that they were undertaken systematically on a large scale. This was done by Professor Poulton, who obtained most striking results.f Working upon over 700 chrysalides of Vanessa urticce (Small Tortoiseshell), he found that pupae placed against black surroundings became as a rule extremely dark, whilst against white surroundings " not only was the black colouring matter as a rule ab- sent, so that the pupae were light-coloured, but there was often an immense development of the golden spots, so that in many cases the whole surface of the pupae glittered with an apparent metallic lustre." Against a gilt background a much higher percentage of gilded chrysalides was obtained, and this led Professor Poulton to suggest that in its original habitat the larvae pupated either against glittering micaceous rocks which had a somewhat metallic appearance, or against dark rough weathered rocks, and that they had acquired the power of protectively resembling either of these surfaces. In that such metallic looking rocks occur over a compara- tively limited area, whilst the species has a consider- * Proc. Ent. Soc. , vol. xiii. , 3d series, p. xcix. 1867. fPhil. Trans. 1887, B. p. 811, also " Colours of Animals," p. 119 et seq. AND OF LIGHT. 259 able range, Dr. A. R. Wallace * considers Professor Poulton's suggestion rather improbable; still it should be noted that the Small Tortoiseshell almost invariably seeks mineral surroundings for the pupal period, and very rarely becomes a chrysalis on its food plant. The time at which the colours are determined was found to be especially during the resting stage of the caterpillar, just before pupation, and to a less degree during the onset of the pupal stage, when the caterpillar hangs head downwards, suspended by its last pair of claspers. The former stage lasts about 15 hours, and the latter about 18, and at the end of it the skin splits along the back of the head, and the chrysalis becomes exposed. The reaction of the skin of the larva to the colour of its surroundings was proved by some ingenious experiments to be an indirect one, effected probably through the medium of the nervous system. Thus when a larva, during the onset of the pupal stage, was so placed that part of it was illuminated by a gilded background, and part by a black one, parti-coloured pupae were never obtained. The effective results were produced by that colour to which the larger area of skin had been exposed. We see, then, that in the development of certain Lepi- doptera there is a period, lasting only a day or two, dur- ing which an extreme sensitiveness to the colour of the surroundings is present, and we have also seen that dur- ing the pupal period there may be a great sensitiveness to the temperature of the surroundings. These cases therefore form exceptions to the conclusion arrived at in the last chapter, viz., that reaction to environment * " Darwinism," p. 198. 260 THE EFFECT OF TEMPERATURE diminishes regularly with progress of development. It is obvious, however, that both these capacities for re- action are quite unusual, and have been specially ac- quired for a special purpose. In all probability the organisms are not more sensitive to environmental con- ditions in general at these periods than they are at the earlier ones; in fact, they are probably very much less so, in that the growth has almost ceased. In certain caterpillars the existence of a variable pro- tective resemblance has long been recognised, several instances of the phenomenon being collected by Mel- dola * in 1873. For example, the larvae of the genera SmerintTius and Sphinx, which are green when feeding on their respective food plants, become brown previous to pupation, when the caterpillars are crawling over the ground to find a suitable burying place. Again, the Geometer Acidalia degeneraria is greenish brown in the summer, but changes to a rusty brown in the autumn, at the period preparatory to hibernation. Some years later, Meldola recorded an observation by Mr. E. Boscher,f relative to the larvae of Smerinthus ocellatus (Eyed Hawk Moth). These larvae were noticed to be of a whitish green colour when feeding on one species of willow, and of a bright yellowish green when feeding on another species, these colours being, on the whole, protective. It was generally believed that such .varia- bility in the colour of caterpillars is due to the direct chemical effect of the different kinds of leaves eaten, but Professor Poulton,J by his experiments on the *Proc. Zool. Soc. 1873, p. 153. f Weismann's " Studies in the Theory of Descent," 1882, p. 241. i" Colours of Animals," p. 149. AND OF LIGHT. 261 larvae of Smerinthus, proved that it was the colour of the leaves, and not their food quality, which provoked the change. Thus he sewed leaves together, " so that the caterpillars were exposed to the colour of the upper or of the under side alone, although they ate the same leaf in both cases. In other instances the bloom was rubbed off the under sides of some leaves, whilst others were left normal." More striking cases of protective resemblance were obtained by Professor Poulton for various Geometra larvae. Larvae surrounded by the leaves on which they fed, became, in the majority of species, light brown or light gray in colour. If, however, an abundance of twigs had been mixed with the leaves of the food plant, they became dark in colour. The larvae of the Pep- pered moth (Amphidasys betularia) afforded the most striking result of all, for when reared amongst green leaves and shoots they became bright green without ex- ception, whilst in the presence of dark brown twigs they nearly all assumed a corresponding colour. The influence of the surroundings acts only very slowly upon the colour of the caterpillars, the coloured part being " actually built up of the appropriate tint." Probably this is the result of light stimuli acting on the surface of the skin, and not reflexly through the eye. Thus painting the eyes (ocelli) with opaque varnish led to no diminution of reaction. CHAPTER VIII. THE EFFECT OF MOISTURE AND OF SALINITY. Effect of humidity of soil on plant growth Effect of dry and moist surroundings on characters of plants Desert plants and Aquatic plants Effect of moisture on Lepidoptera and on Molluscs Characters of maritime plants probably due to saline environment Conversion of A. saUna into A. milhausenii and into BrancMpus Effect of increased salinity on characters of the cockle Influence of salinity on rate of growth of Tubularians, and on size of sea- urchin larvae. IN that the presence of water is absolutely necessary to enable living organisms to exhibit activity, and very probably, indeed, to enable them to retain vitality at all for even spores contain a small percentage of water so we should conclude that differences in the amount of water in the environment of the organisms would form a fertile source of variation. Such is, in fact, the case in the Vegetable Kingdom. For example, the amount of water in the soil has a considerable influence on the rate of growth, as is shown in the table given below. These figures, which were obtained by Hellriegel,* represent the amount of dry substance contained in the grain and chaff of barley which had been reared in soils containing various percentages of the saturation quan- tity of moisture. We see that the rate of growth varies but little until the humidity falls below 30 per cent., and then it diminishes so rapidly that with a hu- * Quoted from Davenport's " Experimental Morphology," p. 353. THE EFFECT OF MOISTURE 263 midity of 10 per cent, it has almost ceased. In the observations made by Gain,* the fresh weight of the en- tire plant was determined. Seeds of various species were planted in soil containing either from 3 to 6 per cent, of water, or from 12 to 16 per cent. Growth was more rapid in the damp than in the dry soil, so that the weight of the full-grown plant was 1.12 times greater in the radish, 2.23 times in the bean, and 2.7 times in the flax. PRODUCTION IN DRY SUBSTANCE. HUMIDITY OF BOIL. GRAIN. CHAFF. per cent. 80 8.8 9.5 60 10.0 11.0 40 10.5 9.6 30 9.7 8.2 20 7.7 5.5 10 .7 1.8 5 .0 .1 The effect of a dry soil and atmosphere is well shown by the characters of desert plants. These are stunted in growth, and are of a nearly uniform gray colour, ow- ing to their intense hairiness. The leaves are more fleshy, and there is a great tendency to the formation of spines. That these characters are in part at least the direct result of want of water is shown by the fact that they may disappear if an abundance of water is supplied. Thus Ononis spinosa. L., if grown in a rich, well-watered soil, or in a moist atmosphere, gradu- ally loses its spines, those first formed under the new * Ann. Sci. Nat. Bot. (7), xx. p. 63, 1895. 264 THE EFFECT OF MOISTURE conditions being much reduced in size and rigidity.* Lothelier f has made numerous observations in which individuals of the same species were placed side by side, some exposed freely to the air, and others kept moist under a glass shade by a vessel of water. He found that, for instance, Berberis vulgaris bore non-spinescent leaves in a moist atmosphere, but spines and spines alone in a perfectly dry one. Again, the shoots which in Lycium barbarum, Ulex europceus, etc., would nor- mally have formed thorns by arrest in development and sclerosis, in a very damp atmosphere continued to grow, and elongated into leafy branches. Microscopical ex- amination showed that in the moist atmosphere the parenchyma was only imperfectly differentiated into spongiform and palisade tissue, whilst in dry air there was a great arrest in the area of parenchymatous tissue, but the palisade cells were well developed, and there was a special consolidation of fibrous tissues. Again, the common water-reed, Phragmites communis, when growing in the unirrigated areas of the Mle Valley, forms a stunted growth, with very short and sharp- pointed leaves. " Close to the Mle, however, in Ehoda Island, it grows nine or ten feet high, with long leaves almost exactly like the plants in English rivers." $ The effect of drought upon Dioscorea batatus (Yam) has been carefully studied by Duchartre. Though not allowed to have any water, some tubers of this plant *Rev. G. Henslow, " The Origin of Plant Structures," p. 40. f See also " Recherches anatomiques sur les epines et les aiguillons des plantes," Lille, 1893. \ " Origin of Plant Structures," p. 41. Bull, de la Soc. Bot. de France, 1885, p. 156. AND OF SALINITY. 265 produced long shoots. The stem was more slender than usual, but excessively rigid, owing to the reduction of the parenchymatous tissues, and the predominance of the elements of consolidation. The leaves were small and undifferentiated, and the stomata undeveloped. Many other instances showing the relations between floral structures and arid surroundings have been col- lected by Henslow in his book on the " Origin of Plant Structures," where the subject is dealt with in extenso. The effects of a very great increase in the humidity of the surroundings, such as is experienced by plants which actually live in water, lie, as might be expected, in a very different direction. That the peculiar char- acters of aquatic plants are in considerable measure the direct effects of their peculiar environment, is proved by the fact that plants, normally terrestrial, often de- velop such characters when grown in water. Thus Costantin found that under such conditions a diminu- tion in the number of the vessels of the fibro-vascular system of the stem invariably occurred. For instance, in Vicia sativa (Vetch) the middle of the stem of the aquatic form of the plant had only 38 vessels, whilst the aerial form had 47. In Ricinus communis (Castor-oil plant) the aquatic form had 10, as against 21; in Faba vulgaris it had 2 at the sides and 15 at the angles, as against respectively 5 and 36. The pseudo-aquatic forms thus tended towards true aquatic plants, in which the fibro-vascular system is always more or less de- generate. Again, Costantin found that there was an increase in the lacunae, when stems normally aerial are kept submerged, just as the aquatic form of amphibious plants is found to have more of such lacunae than the 266 THE EFFECT OF MOISTURE aerial. As regards the leaves, it is well known that when aerial and floating leaves are present on the same aquatic plant, they differ greatly in structure, and as a rule also in form, from the submerged leaves. In Ranunculus heteropJiyllus and Cabomba aquatica, for instance, the floating leaves are more or less rounded, whilst the submerged ones have dissected and filiform segments. In Hippuris (Mare's tail) the aerial and floating leaves are short, and in CallitricJie rounded, but the submerged leaves of both are linear or ribbon- like. In all cases the submerged leaves are of a more delicate texture, more or less translucent, and of a brighter green colour than the others. They show degradation of anatomical structure in every part, the cuticle and stomata disappearing, whilst the chlorophyll grains and the mesophyll are greatly reduced in quan- tity.* Even better evidence of the direct relation be- tween environment and character is afforded by certain other observations of Costantin. Thus he found that he could change the form of Hippuris at will, " by transplanting an aquatic plant on to land, and vice versa; all the leaves produced under water were long, undulated, and delicate; whereas those in air were short, erect, and firm." Again, he found that the leaves of Sagittaria (Arrowhead), when deeply submerged, are soft and flexible, and may reach a length of over six feet, but when developed in air they are short and erect. When a leaf is full grown, sudden change of environ- ment kills it, aerial leaves perishing under water, and *F^Henslow's "Origin of Plant Structures," chap, viii., from which the greater part of this paragraph is drawn. AND OF SALINITY. 267 aquatic ones perishing in air; but if it is only in the course of its development, it can adapt itself to a changed environment. Thus, if a half -formed floating leaf of Ranunculus heterophyllus or of Sagittaria is submerged, "it is at once arrested, and begins to re- adapt itself to water." There are some species, how- ever, such as many algae, which show no power of adap- tation, and can only live entirely under water. The converse experiments of growing aquatic plants on land afforded equally striking results. For instance, it was pointed out by Godron,* as long ago as 1839, that whilst Ranunculus aquatilis (Water crowfoot), when wholly submerged, has all its leaves delicately lacini- ated, yet " if the plant is able to send some of its leaves to the surface, they float and assume a very different form, being kidney-shaped and lobed. The same plant when growing entirely out of water presents a very dif- ferent appearance; the stem is short, much divided into branches, which bear a large number of small leaves, cylindrical, much divided, and somewhat thick. If it were not for the floral organs, one would certainly be- lieve in two or three species." Again, Costantin grew a plant of the aquatic form of Peplis Portula on land, and found that the internodes were changed from their elongated form to a short one. The septa of the cor- tical parenchyma of the stem remained homogeneous, instead of being hollowed out into secondary lacunae, and also the number of vessels was increased. Thus there were 53 vessels in the land form of Peplis Por- tula, instead of 25; 12 instead of 4 in Callitriche, and 57 instead of 18 in Nasturtium. * Quoted from De Varigny's "Experimental Evolution," p. 97. 268 THE EFFECT OF MOISTURE It may have been noticed that in speaking of these adaptations of terrestrial plants to water, and of aquatic plants to land, it has been more or less tacitly assumed that the effects observed were due to the direct influence of the surroundings on the tissues. It is of course possible that they are partly or even largely in- direct, and that the change of habitat merely calls up latent characters long since possessed by the ancestral plants which lived in similar surroundings. Upon members of the Animal Kingdom, observa- tions as to the effect of moisture are exceedingly meagre. This is probably attributable to the fact that in most cases a direct effect is either slight or wanting. Thus Merrifield * could not observe any influence upon the pupae of certain Lepidoptera (E. autumnaria and 8. illustraria), nor could Standfuss upon those of cer- tain other species. Koch,f however, came to the con- clusion that a long period of dry or moist weather might exercise a considerable influence on the size of the suc- ceeding generation. Immediately after a continuously dry summer, butterflies are always smaller than after a moist one. Likewise also the second generation of Argynnis selene, which takes flight in the height of sum- mer, is always smaller than the spring generation; but it seems to me highly probable that these effects are of an indirect nature, dependent, perhaps, on changes effected by the moisture in the vegetation on which the larvae feed. Leydig $ has endeavoured to trace a connection be- * Trans. Ent. Soc. 1891, p. 163. f Quoted fromEimer's " Organic Evolution," p. 152. i " Organic Evolution," p. 97. AND OF SALINITY. 269 tween the moisture in the environment and the dark- ness of colouring of certain animals. Thus he observed that Molluscs such as Arion empiricorum (common slug), Helix arbustorum, Succinea Pfeifferi, and Helix circinata became darker than usual in moist localities. He observed a similar condition also in certain Am- phibia and in Lacerla vivipara. However, Eimer ob- served just the reverse condition in Arion, finding it darker upon the heights, where there was little water, than in well-watered valleys. In any case, the effect is probably an indirect one, acting through the vegetation. The Effect of Salinity. The effect of salinity upon members of the Vegetable Kingdom is well illustrated by the peculiarities of form and structure possessed by maritime plants. That these characters are at least in part the direct effect of the salinity of the soil and at- mosphere, is proved by comparison of plants growing near the sea-shore with individuals of the same species growing inland. Thus Lesage * has investigated no less than 85 different species. He found that in 54 of them the leaves were thicker in the maritime indi- viduals than in the inland ones, they being about four times as thick in Cakile maritima and Silene mari- tima; in 27 there was no apparent difference, and in 4 they were thicker in the inland individuals. With re- gard to the mesophyll, there was no noticeable change in 11 species, but in all the other shore plants the pali- sade cells were more numerous or attained greater thickness, and at the same time the interspaces under- lying the stratum of palisade cells were much reduced. *" Influence du Bord de la Mer sur la Structure des Feuilles," Rennes, Oberthur, 1890. Also Rev. Gen. de Bot., torn. ii. p. 54. 270 THE EFFECT OF MOISTURE Changes in the epidermis were much less frequent, there being no appreciable difference in 31 plants. In 23 of the shore plants the cells were larger, however, they being three to five times as large in Beta vulgaris and Silene maritima. In four instances these cells were larger in the inland plants. With regard to the chlorophyll, there was no difference in some cases, but in others it was marked. Thus in Thesium humifusum and CaJcile maritima the grains were much smaller in the maritime plants, and in other species the number of grains was reduced. Even more conclusive evidence of the direct effect of salinity in producing these peculiarities of structure has been afforded by experiment. Lesage cultivated various plants under similar conditions except that some of them were watered with water containing com- mon salt, and he found that characters were developed similar to those exhibited by maritime plants. In Pisum sativum the leaves increased in thickness, pali- sade cells became larger and more numerous, whilst the intercellular spaces and the chlorophyll diminished. Lepidium sativum (Garden Cress) gave even more marked results. The palisade tissue was more devel- oped and possessed an extra layer; the lacunae were less pronounced, and the chlorophyll less abundant. On sowing the seeds of this plant a second year, moreover, and again treating the plants with salt water, a still more marked result was obtained, it appearing as if the alteration in the tissues of the second generation was carried on more or less from the point gained in the first. The salted water might even affect physiological processes. Thus radishes usually contain no starch, AND OF SALINITY. 271 but, after treatment with salted water (.3 to 1 per cent, in strength), might contain a great deal. On the other hand, watering cress with 1 per cent, solution caused the starch normally present to disappear, either wholly or in part. Upon the growth and even on the actual structure of animals, changes of salinity may in some instances exert a marked action. Animals accustomed to develop in fresh water have their growth retarded by the addition of salt. Thus Yung * reared frog's embryos in solu- tions containing respectively 0, .2, .4, .6, and .8 per cent, of salt, and found that except in the .2 per cent, solution, which had no influence, there was a retarda- tion in development. This increased with the concen- tration of the solution, so that in the .8 per cent, solu- tion the larvae took 17 days longer to hatch than in pure water. Again, Sargeant f found that the rate of reproduction by fission of the naid Dero vaga becomes slower and slower according to the concentration of the solution it is reared in. Taking the rate in pure water as 11.3, it becomes reduced to 8.5 in .05 per cent, solu- tion of salt, to 7.7 in .1 per cent, solution, 4.1 in .2 per cent solution, and .3 in .3 per cent, solution. Still stronger solutions stop reproduction altogether, and kill off some of the worms. The more considerable effects which change of sa- linity may produce are well illustrated by the interest- ing and widely known observations of Schmanke- witsch J upon Artemia salina and A. milkausenii. *Arch. des Sci. Phys. et Nat., xiv. p. 502, 1885. f Davenport's " Experimental Morphology," p. 365. JZeit. f. wiss. Zool., xxv. p. 103, 1875, andxxix. p. 429, 1877. 272 THE EFFECT OF MOISTURE The history of these observations is as follows: Through the breaking of a dam across a salt lake (Kuyalink), a number of individuals of A. salina were washed from the upper less saline waters into the lower more concentrated waters, and the Specific Gravity of these was at the same time reduced to 1.058. After the dam was repaired the concentration gradually in- creased again through evaporation, the Specific Gravity rising to 1.105 the year after; to 1.135 the next year, and to 1.205 the year after that. Accompanying this concentration, the generations of Artemia progressively degenerated, till they finally attained the characters of A. milhausenii. Schmankewitsch also succeeded in con- verting a brood of A. salina into A. milhausenii by the artificial process of gradually increasing the percentage of salt in the water in which they were living (the Specific Gravity being raised from 1.028 to 1.205). These two forms of Artemia have been held to be distinct species, in that milhausenii shows an absence of fins and bristles on the lobes of the tail, and has much smaller tail lobes, but larger branchial appendages to the legs, than salina. Schmankewitsch himself did not hold this opinion, however, and Bateson,* who has recently studied the question afresh, thinks similarly. Bateson collected samples of Artemia from a number of different salt lakes in Western Central Asia and West- ern Siberia, and, consonant with Schmankewitseh's statement, he found that, on the whole, the number of bristles on the caudal fins, and likewise the size of the {ins, was smallest in specimens collected from waters of high Specific Gravity. The accompanying table shows *" Materials for the Study of Variation," p. 96. AND OF SALINITY. 273 the range in the number of bristles on a single fin, only adult females bearing eggs in the ovisac having been reckoned. Only the Specific Gravities of the various waters are given, but their chemical composition varied between even wider limits, and this may be responsible for some of the irregularities observed: SI'. 0. BRISTLES. SP. 0. BRISTLES. SP. 0. BRISTLBS 1.030 1024 1.100 410 1.160 1619 1.050 1113 1.105 59 1.165 13 1.056 917 1.105 4-8 1.165 15 1.065 27 1.115 16 1.170 68 1.075 813 1.115 59 1.175 1-5 1.075 57 1.130 1216 1.179 49 1.085 1315 1.140 37 1.204 25 1.095 2028 1.150 410 1.215 2-4 1.100 814 1.150 7-8 1.215 27 1.100 812 1.150 01 As this table and also Schmankewitsch's results show, there is no true differentiation between A. salina and A. milhausenii, in that these extreme forms are con- nected by a continuous series of naturally occurring intermediate forms. As Bateson remarks, " it has never been shown that there is a male A. milhausenii, with distinctive sexual characters, and among the Branchiopoda the various sexual characters of the second antennae in the male are most strikingly dis- tinctive of the several forms." In addition to the particular character in question, Bateson found that there was a great variation in other characters as well. Thus he says : " Almost each locality has its own pattern of Artemia, which differs from those of other localities in shades of colour, in average size, or in robustness, and in the average num- 274 THE EFFECT OF MOISTURE ber of spines on the swimming feet, but none of these differences seem to be especially connected with the de- gree of salinity." Probably the Artemia recently found by R. T. Giinther * inhabiting Lake Urmi in Persia in such enormous numbers is only another local variety of A. salina. It differs from this species in possessing an incompletely segmented abdomen, in the claspers of the male being larger, and in other char- acters, but Giinther says he is nevertheless inclined to agree with Packard that there is only one well-defined Old World species of Artemia, viz., A. salina. Schmankewitsch also changed the salinity in the re- verse direction, and gradually diluted the salt water containing some A. salina till it finally became per- fectly fresh. The Crustaceans, which had gone through several generations during the process, had meanwhile so changed their character that in Schman- kewitsch's opinion they now resembled the form of the genus Branchipus. Thus the last segment of the post- abdomen became divided into two segments, and Schmankewitsch maintained that this division of the segment is the only structural character really differ- entiating the genus Branchipus from Artemia. How- ever, Glaus f has shown that there are many other points of difference, and that the division in question is not a structural character of great importance. Also Branchipus is distinguished by the sexual characters of its males, which possess no structure in any way similar to the great leaf-like second antennae shown by the male Artemia. "We must conclude, therefore, that * Journ. Linn. Soc. (Zool.), xxvii. p. 395. f Auz. Ak. Wiss. Wien., p. 43, 1886. AND OF SALINITY. 275 though decrease of salinity does produce distinct struct- ural changes, yet Schmankewitsch considerably exag- gerated their importance, and deduced from them more than he had any justification for. In addition to Artemia, Schmankewitsch * studied the effect of salinity on several other Crustaceans such as Daphnia rectirostris, DapJinia magna, and Branchi- pus ferox. He found that in their case also consider- able structural and physiological changes were brought about, the fresh- and salt-water forms differing, in his opinion, by characters usually held to be specific. Equally interesting evidence as to the effect of grad- ual increase of salinity has been obtained by Bateson f in the case of the common cockle, Cardium edule. This mollusc, together with several others, is present in enormous numbers in the brackish waters of the Aral Sea. The waters of this closed basin have been gradu- ally drying up and receding, but the area left exposed " is not a level tract, but contains three considerable depressions, called respectively Shumish Kul, Jaksi Klich, and Jaman Klich. . . These depressions re- mained, for a time, as isolated lakes, each containing a separate sample of the fauna of the sea living in it." As they gradually dried up, becoming salter and salter, the character of the shells progressively changed. To determine this change, samples were collected at vari- ous levels in the lake areas, and were carefully com- pared. On the western shore of Shumish Kul there were seven very definite terraces of muddy salt, show- ing the position of the water at various periods during the gradual drying up. The changes produced con- * Ibid. f Phil Trans. 1889, B. p. 297. 276 THE EFFECT OF MOISTURE sisted in (1) a diminution in the thickness of the shells, this being first apparent in the shells of the third ter- race. So marked was this change that the shells of the seventh or lowest terrace were almost horny and semi- transparent, and their weight was not a third that of shells from the first two terraces; (2) a diminution in the size of the beak; (3) a high colouration in the shells. This change occurred almost uniformly, the shells of each terrace being very nearly alike in texture, thickness, and degree of colouration; (4) grooves be- tween the ribs appearing on the inside of the shell as ridges with rectangular faces; (5) a great diminution in the absolute size of the shells on the lowest terrace; (6) an increase in the length (greatest antero-posterior dimension) of the shells in proportion to their breadth, this ranging from the average ratio of 1 : .80 in the shells from the first terrace to 1:.725 in shells from the seventh. In Jaksi Klich lake the shells from the lowest and most saline deposit were even more elon- gated, the ratio of length to breadth being as 1: .68 for samples of smaller shells, and 1:.66 for samples of larger ones. Those from the lowest deposit of Jaman Klich showed about the same degree of elongation as those from the lowest terrace of Shumish Kul. It was very noticeable that the shells of each sample, whether from a separate lake or only from a particular terrace, resembled each other more closely than they did shells from one of the other lakes, or those from another terrace in the same lake, as at Shumish Kul. In each of the three lakes mentioned (and also in an entirely distinct locality, the lagoon of Abu Kir near to Alexandria), it was thus found that shells which had AND OF SALINITY. 277 lived in very salt water had become like each other in possessing the characters of thinness, high colour, small beaks, ribbing on the inside, and great relative length. " In view of these four instances of similar variations occurring under similar conditions," says Bateson, " it seems almost certain that these condi- tions are in some way the cause of the variations." In that the variations in the quality, texture, and colour of the shell are found developed to nearly the same degree in all the individuals of successive terraces, Bateson considers they may be fairly supposed to be the direct result of environmental change; but the quality of in- creased proportional length is not found in all the in- dividuals, and hence may have arisen in some other way, as by Natural Selection of the type best fitted to live in the altered state. A further proof of Bateson's view is afforded by the fact that when the salinity was altered in the direction of diminution, the characters, of the shells were similarly changed in a reverse direction. Thus, as already men- tioned, the cockles from the very saline lake of Abu Kir resembled those from the lakes of the Aral Sea, but close to this lake are three small areas of water, the Kamleh lakes, of which the water is now quite fresh (owing to their receiving waste water from the irriga- tions). One of these lakes contains living cockles, and another the shells of extinct ones. Now in both in- stances the shells are thick and coarse in texture, and comparatively light-coloured. However, the feature of great proportional length still remains. Other evidence as to the relation between salinity and structure in molluscs has been obtained by Gib- 278 THE EFFECT OF MOISTURE bons * for certain tropical and sub-tropical species of Littorina. These organisms are confined to more or less brackish waters, and seem incapable of living in pure salt water. Gibbons says he has " met with three of these species, and in each case they have been dis- tinguished from the truly marine species by the ex- treme (comparative) thinness of their shells, and by their colouring being richer and more varied; they are also usually more elaborately marked." Thus diminu- tion of salinity seemed to have produced thinness of shell in the species as a whole, but within their own limits it was found that the reverse relation held, and that, as in Bateson's observations, the shells became thinner as the water they lived in became more salt. In marine animals, as in fresh-water ones, increase of salinity probably tends to diminish the rate of growth. Decrease of salinity, on the other hand, may have the reverse effect, and within certain limits actu- ally increase the growth rate. Thus Loeb t determined the rate of regeneration of decapitated hydroid polyps (Tubularia mesembryanthemum) placed in sea-water of various degrees of dilution and concentration, seven to nine individuals being measured at each concentration. His results are reproduced in the figure given below. Here the Specific Gravity of the water is represented along the abscissa, and the amount of regeneration by the height of the ordinates. We see that the maximum rate of regeneration took place in water of Sp. G. 1.025, or in very considerably diluted sea-water (the Sp. G. of this being about 1.038). At this dilution the re- * Quart. Journ. Conch., i. p. 339. t " Biological Lectures delivered at Wood's Holl," 1893, p. 46. AND OF SALINITY. 279 generation was more than twice as fast as in normal water, but with further dilution it rapidly diminished, and ceased altogether in water of Sp. Gr. 1.013. Somewhat similar results to these were obtained by the author for sea-urchin plutei. As we saw in the last chapter, the actual size of an organism is probably affected by environment in a similar manner to the growth rate, and the author found that these plutei, i.fe ife" .If nb Concentrated Sea Water 01 1.02 Dilute Sea Water Specific Gravity FIG. 24. Effect of salinity on growth of Tubularia. allowed to develop in sea-water of various concentra- tions, attained a greater size than the normal when kept in moderately diluted water, and probably a slightly smaller size when kept in concentrated water. The results obtained are indicated in the subjoined figure, where the abscissae represent the salinity of the water, and the ordinates the average percent- age variation in the size of the larvae after eight days' growth, as compared with that of larvae grown in normal sea-water. The salinity of normal water was taken as 1000, and the less saline waters were *Phil. Trans. 1895, B. p. 586. THE EFFECT OF MOISTURE. obtained by diluting respectively 950, 900 c. c., etc., of water to a litre; the more saline by concentrat- ing 1050, 1100 c. c., etc., to a litre. The maximum effect on size was produced by a solution containing 50 c. c. of fresh water per litre, the increase amount- ing to 15 per cent. With further dilution, the favour- able influence became less and less, till, with water con- +20 415 -1: 850 900 950 1000 1050 Salinity o^ gea-water 1100 1150 1200 FIG. 25. Effect of salinity on size of sea-urchin larvae. taming 150 c. c. per litre, it was negative. Thus the optimum salinity is for a much less diluted water than in the case of the tubularians. The present results also differ from these latter in that more concentrated waters have exceedingly little effect on the size of the larvae. CHAPTER IX. THE EFFECT OF FOOD AND OF PRODUCTS OF METABOLISM. Effect of artificial manures on growth of crops Effect of nutrition on plant variation Development of bees and of aphides in rela- tion to food Influence of nature of food on wing markings of certain Lepidoptera Dependence of colour of larvae on plant pig- ments Influence of food on growth of tadpoles Plumage of cer- tain birds altered by abnormal diet Quality of food influences organs of digestion Every organism probably has specific metab- olism, which has especially adverse action on its own growth Products of metabolism may stimulate growth Effects of small quantities of urea, uric acid, and ammonium salts Influence of volume and of surface area of water on growth of pond snail Influence of surface area on growth of tadpole Effects of increas- ing quantities of metabolic products on characters of a snail, and of a Crustacean. DARWIN records * that Andrew Knight was of the opinion that " of all the causes which induce variability, excess of food, whether or not changed in nature, is probably the most powerful." Darwin himself, more- over, was inclined to accept this view of the potency of food as probable. That changes in the amount and the quality of the food available for an organism during its growth must of necessity exert an important influence on the course of that growth, and presumably, there- fore, on the final limits of its attainment, is sufficiently obvious both from one's own everyday experience, and * " Animals and Plants," vol. ii. p. 244. THE EFFECT OF FOOD from a simple recognition of the relation between cause and effect. Growth can only take place at the expense of food material, and unless this is always more than sufficient for the needs of the organism, the rate of growth must be dependent upon it. In spite of the importance of changes in feeding as a source of variation, the number of direct and une- quivocal experiments made upon the subject is compara- tively small, for most of them are complicated by simul- taneous changes in other conditions as well. Upon members of the vegetable kingdom, the experiments made by Lawes and Gilbert * at Kothampsted during the last fifty years afford most valuable evidence. These concern the effect of various manures on the growth of barley, wheat, and various leguminous plants. In the accompanying table are given the average 200 LB. AM- 275 LB. SO- NO NITRO- MONIUM DIUM NI- 1000 LB. ADDITIONS TO SOIL. GEN OTIS MANURE. SALTS, = 43 LB. NI- TRATE, = 43 LB. NI- RAPE CAKE = 49 LB. TROGEN. TROGEN. NITROGEN. Without mineral manure, 16.5 29 32.7 41.2 Superphosphate, Potassium, sodium and magne- sium sulphates, Superphosphate and K. Na. and 21.7 18 42.7 31.4 45.7 33.5 43.4 39.5 Mg. sulphates, 22.4 43.5 45.5 43.2 amounts of barley grain (in bushels per acre) obtained each year from soils treated in various ways. These observations were carried on for forty years in succes- sion (1852-91) upon the same land, and so represent strictly average results, from which errors due to varia- *"The Rothampsted Experiments," p. 78, Edinburgh and Lon- don, 1895. AND OF PRODUCTS OF METABOLISM. 283 tions of season, and other causes, are practically elimi- nated. From this table we see that the average yield from land left entirely without manure was 16.5 bushels of grain. On adding various manures, all of which con- tained about the same weight of combined nitrogen, the yield of grain was doubled, or, in the case of rape-cake manure, increased to two and a half times the amount. The addition of various inorganic salts to the soil also had a favourable effect, though to nothing like the same degree as that of the nitrogenous manures. Thus we see that, when no nitrogenous manure whatever is present, the addition of superphosphates increases the yield by 32 per cent.; of potassium, sodium and magnesium sul- phates by 9 per cent. ; and of both superphosphates and these sulphates, by 36 per cent. When the nitrogen is added as ammonium salts or nitrates, then combinations of nitrogenous and mineral manures give a very much better yield than the nitrogenous manure alone, but when it is added as rape cake, the growth of the crop has already been so much increased that the further addition of mineral salts effects but little. Yet even the highest of the numbers in this table does not represent the maximum amount of growth of which the barley is capable, for a soil treated with farm-house manure, and no additional mineral salts, yielded on an average 48.6 bushels per acre. In all these experiments the yield of straw was increased in more or less similar proportions to the yield of grain, and hence we may conclude that the growth of a plant in normal soil can be very nearly trebled if only favourable enough conditions are afforded it. 284 THE EFFECT OF FOOD Somewhat similar results to these were obtained by Lawes and Gilbert for wheat, bean, clover, and other crops, but it is deemed unnecessary to reproduce them here. Most striking evidence as to the influence of nutri- tion on variations has been obtained by De Tries.* When carrying out his artificial selection experiments on five-leaved clover, he found that in one series of ob- servations seeds from some plants grown in a poor soil yielded 39 per cent, of 3-leaved, and 48 per cent, of 5- to 7-leaved clover. Those from plants of the same stock which had been grown in a rich soil, however, gave only 14 per cent, of 3-leaved, and 73 per cent, of 5- to 7-leaved clover. The effect of nutrition on Ranunculus bulbosus (Crowfoot) was almost as striking. f Wild flowers col- lected near Hilversum were found by De Vries to have the following frequencies of distribution in the numbers of their petals : Number of petals, 5 6 7 8 9 10 11 657 41 11 2 4 2 In the autumn of 1887 De Vries planted some of these plants in his culture garden, where they bloomed the following year. Owing, presumably, to the better nutrition, the proportion of flowers with more than five petals was considerably increased: Number of petals, 5 6 7 8 9 10 133 55 23 7 2 2 The seed from the many-petalled flowers was collected *" Die Mutationstheorie," p. 448. fBer. d. deutsch. Bot. Ges., xii. p. 197, 1894. AND OF PRODUCTS OF METABOLISM 285 and sown through two seasons, and from the seed then obtained 372 plants were grown. Some of these ger- minated early, and so developed under less favourable conditions than the others. As will be seen from the accompanying figures, these early plants had 9 petalled flowers occurring the most frequently, whilst the later ones had 10 petalled flowers; i. e., flowers with twice the original number of petals: Number of petals, 5 6 7 8 9 10 11 12 13 14 15 1681 Early plants, 409 532 638 690 764 599 414 212 80 29 18 20 Late plants, 40 52 126 165 204 215 177 104 35 8 4 The somewhat unexpected results obtained by Mac- Leod * with Ficaria ranunculoides may also be attrib- uted, at least in some degree, to the effects of nutrition. MacLeod determined the numbers of stamens and of pistils in the flowers borne by a number of plants at the beginning of the flowering season, and again in the flowers borne by the same plants at the end of the sea- son. The early flowers had on an average 26.73 stamens and 17.45 pistils, whilst the late ones had only 17.86 stamens, and 12.15 pistils. Also the variability in the number of stamens and of pistils was very dif- ferent in the two cases, the coefficients of variation being respectively 14.1 and 22.3 per cent, in the early flowers, and 18.5 and 27.9 per cent, in the late ones. The method adopted by MacLeod for estimating the cor- relation between the numbers of stamens and of pistils is erroneous, so Professor Weldon has recalculated the constants.! He finds that MacLeod's figures indicate the correlation to be much less in the early than in the * Botanisch. Jaarboek., xi., 1897. fBiometrika, i. p. 125,1901. 286 THE EFFECT OF FOOD late flowers (the r being respectively .51 and .75 in the two cases). As Professor Weldon remarks, these re- sults " provide a most valuable lesson as to the possible danger of asserting that such differences are significant of local races." By observations upon the growth of seedlings placed in various solutions, it has long been known that normal growth is possible only if various inorganic salts are present. There must be nitrogen in the form of nitrates or ammonium salts, sulphur in the form of sul- phates, phosphorus as phosphates, chlorine as chlorides, and the metals sodium, potassium, magnesium, calcium, and iron in solution as salts. The absence of any one of these substances speedily inhibits normal growth; as soon, in fact, as the seedling has exhausted the small quantity of it stored up within itself. For instance, plants grown in solutions containing no iron soon show a sickly appearance; "the leaves are no longer green, but white, and microscopic examination of them shows that abnormal chlorophyll bodies, or none at all, are present in their cells. If we add to the food solution a few drops of dilute ferric chloride solution, the pre- viously white leaves become green in two or three days, and the growth of the plants now proceeds normally." * It follows, therefore, that if the absence of these vari- ous substances stops growth altogether, a deficiency in them must produce diminished or abnormal growth, and so lead to the production of variations. With members of the Animal Kingdom, variations in the inorganic salts of the food may also be a source of * Quoted from Detmer's " Practical Plant Physiology," p. 84. AND OF PRODUCTS OF METABOLISM. 287 variation. Thus Cooke * states that " a deficiency of lime in the composition of the soil of any particular locality produces very marked effects upon the Mol- lusca which inhabit it; they become small and very thin, occasionally almost transparent. The well-known var. tennis of Helix aspersa occurs on downs in the Channel Islands where calcareous material is scarce. For simi- lar reasons, II. arbustorum develops a var. fusca, which is depressed, very thin, and transparent, at Scilly and also at Lunna I., E. Zetland." However, in animal development the supply of in- organic salts is almost always more than sufficient for the needs of the organism, and such variations as are produced are due chiefly to the organic constituents of the food. Among invertebrate animals, our knowl- edge of the direct influence of food is almost confined to certain of the Insecta. In the case of bees, it has been known for a very long time that the quality and quan- tity of the food supplied to the larvae determines whether the reproductive organs shall undergo their full development, and produce fertile queens, or remain undeveloped, and so produce non-fertile working fe- males. According to A. von Planta, the diet of the queen larvae contains twice as much fatty material as that of the worker s.f Again, Eimer has pointed out that in the case of the humble bee, the first brood of ova, laid in the spring, get only a scanty supply of nutri- ment, and develop into small females, which are fertile though they can only produce drones. The next brood * Vol. iii., " Cambridge Natural History," p. 89. f Quoted from Creddes and Thomson's " Evolution of Sex," p. 43, 288 THE EFFECT OF FOOD born obtain more nourishment, and develop into larger females, which are capable of occasionally producing females, as well as drones. Finally the future queens, which obtain a still richer diet, are born. The deter- mination of sex seems to be dependent on nutrition also in aphides or plant-lice. Thus " during the summer months, with favourable temperature and abundant food, the aphides produce parthenogenetically genera- tion after generation of females. The advent of au- tumn, however, with its attendant cold and scarcity of food, brings about the birth of males, and the conse- quent recurrence of strictly sexual reproduction." * In this instance, therefore, the effect of nutrition is bound up with that of temperature, and there are no data to show whether either of these conditions could produce the effect if acting alone. Upon the Lepidoptera the effects of various foods have been tested in a considerable number of instances. Observations were made by Gr. Koch f in Germany as long ago as 1832. By feeding the caterpillars of Che- Ionia Jiebe with different plants, he obtained specimens which were either fiery or dull red on the under wings, and which varied in the extent of black marking and white ground. In the case of Euprepia caja (Common tiger moth) it is known, Koch says, " that when the caterpillars are fed from their hatching to their meta- morphosis with leaves of lettuce or deadly nightshade, not one of the imagines produced resembles the origi- nal form; when the insects have been fed on lettuce, the white ground-colour of the wings predominates; *Ibid. t p. 46. fEimer, " Organic Evolution," p. 149. AND OF PRODUCTS OF METABOLISM. 289 when fed on deadly nightshade the brown markings of the upper wings often coalesce and the white vanishes; in like manner the blue markings on the lower wings fuse together and displace the orange-yellow ground colour." * A careful series of observations upon various moths, extending over some ten years, has been made by Greg- son, f His results may be tabulated as follows: Pygcera bucepTiala (Buff Tip) is finer and darker when fed upon sycamore. Xylophasia polyodon (Dark Arches) is dark, sometimes black, when fed upon heather. Hadena adusta (Dark Brocade) is darker when fed upon heath. Acronycta menanthydis (Light Knot-Grass) when fed on sallow, often produces var. A. salycis; fed on heath, produces light specimens. Hybernia defoliaria (Mottled Umber) is beautifully marked when fed upon birch; but on elm gives dull-coloured forms, almost without markings. Eupithecia venosaria (Netted Pug) fed on inflated catchfly is almost white; on shore catchfly is much larger and almost lead colour. Noctua f estiva (Engrailed Clay) fed on thorn is rich red and well marked: on grasses is light yellowish, and rarely well marked. Noctua triangulum (Double-Square Spot) fed on thorn is dark: fed on low plants is light. Abraxas grossulariata (Magpie) fed on red currant is light; on black- thorn is darker; on bullace or wild plum is darker still, the white sometimes becoming yellow. The following case, recorded by the late Mr. New- man^ is of especial interest in that it occurred under natural conditions. The larva of V. polychloros (Large ., p. 151. fThe Zoologist, p. 7903, 1862. j The Entomologist, vi. p. 88, 1872. 290 THE EFFECT OF FOOD Tortoiseshell) usually feeds upon elm, but that of V. urticce (Small Tortoiseshell) upon nettles. Some larvae were found by Mr. J. A. Tawell feeding upon nettles, and so were considered to be those of V. urtiwz. They were accordingly kept on nettles, but to his sur- prise developed into V. polychloros imagines. " These specimens," records Newman, to whom they were shown, " have a wonderful similarity to urticce, which they do not at all exceed in size; still the colour is nearer to that of polychloros than that of urtictz" The effect of abnormal food on Melitcea artemis (Greasy Fritillary) has been noticed by H. Gross.* By feeding the larvae on honeysuckle, a series of very dark imagines was obtained, which differed both in size and colouring from all other specimens known to him, though these had been derived from very varied localities in Eng- land, Ireland, and Scotland. Again, the quantity of the food supplied may have as considerable an effect as the quality. By mistake some V. io (Peacock butter- fly) larvae, captured by Mr. R. Cox,f were left for sev- eral days without fresh food, and all the dead leaves and stalks were devoured. Nearly all the imagines ob- tained from them were rather small, but they also varied much in the intensity of their colouring, two specimens being very much darker than usual, with the yellow in the costal spot and ocellus much reduced. It seems to me, however, that probably these changes were due rather to the abnormal food devoured by the larvae than to the actual lack of food. As regards the larvae of Lepidoptera, the obvious re- *The Entomologist, vii. p. 203, 1874, f The Entomologist, ix. p. 58, AND OF PRODUCTS OF METABOLISM. 291 lation between their colour and that of their food seems to show that the one is directly dependent on the other. Meldola * accounted for it by sup- posing that the larvae had been rendered transparent by Natural Selection, whereby the colour of the vege- table food eaten was itself enabled to give the colour to the larvae. Poulton f has shown that the colours of the larvae are due partly to the pigments proper to the larva, and partly to the pigments derived from the food plants. These pigments undergo some modification in the tissues, but Poulton states that as far as he has in- vestigated the subject " all green colouration without exception is due to chlorophyll ; while nearly all yellows are due to xanthophyll." The chlorophyll, or some modification of it, tinges the blood of the larvae, the green colour of which is often due to this cause alone. From these observations, therefore, it follows that a change of food may also effect a change of colouration. That this is so is strikingly shown by some other obser- vations by Poulton. J Obtaining a large number of larvae of Tryphcena pronuba (Common yellow under- wing) from the same batch of eggs, he split them up into three groups. One he fed on the white midribs of the cabbage, from which the yellow blade had been carefully removed with scissors. These larvae remained almost white at first, and afterwards showed a moderate amount of black pigmentation. The other two groups of larvae he fed respectively on the yellow etiolated leaves from the heart of the cabbage, and upon the deep * Proc. ZoSl. Soc. 1873, p. 155. fProc. Hoy. Soc., xxxviii. p. 269, 1885. jProc. Roy. Soc., liv. p. 417, 1893. 292 THE EFFECT OF FOOD green external leaves. These larvae, however, were all of a bright green or brown colour. Hence it would seem that both etiolin and chlorophyll are capable of being transformed into a larval colouring matter, which may be either green or brown. As regards the effects of feeding among vertebrate animals, a careful series of experiments upon tadpoles (Eana esculenta) has been made by Yung.* The tad- poles were all derived from the same batch of eggs, and were placed, in groups of fifty, in six similar jars of water. All the conditions of development such as light, temperature, and frequency of change of water, were identical, the food alone being varied. The kinds of food supplied, and the average size attained by the tadpoles after 42 days' development (three being meas- ured in each case), are given in the accompanying table: NATURE OF POOD. W ii o S'fc l! go* Sgw* 1 B 1 lr' go Length of tadpole Breadth of tadpole 18.3 4.2 23.2 5.0 26.0 5.8 33.0 6.6 38.0 8.8 43.5 9.2 Per cent, of frogs after 58 days. 14 20 48 66 Here we see that the purely vegetable diet acted least favourably, and the beef diet the most favourably. Egg yolk did not answer so well as coagulated egg al- bumen, but better than uncoagulated albumen. From * Arch, de ZoSlogie Exper., 1883, p. 3. AND OF PRODUCTS OF METABOLISM. 293 the bottom line of the table we see that, 58 days after the beginning of the experiment, none of the 50 tad- poles fed on plants and on liquid egg albumen were sur- viving; but of those fed on fish and on beef, respect- ively 48 and 66 per cent, were alive, and had under- gone their metamorphosis into frogs. The effects of certain foods on the plumage of birds is well known to bird fanciers. Thus hemp seed causes bull-finches and certain other birds to become black. Cayenne pepper, mixed with the food,, changes the yel- low colour to an orange red. This colour change can only be effected by feeding the very young birds; with adults there is no effect whatever. Sauermann * found that all races are not equally susceptible to the abnor- mal diet, some being changed to a crimson, others to a beautiful orange, whilst others remain absolutely un- affected. He found also that canaries are not alone in their susceptibility, for on feeding some white Italian fowls, eight weeks old, with the pepper, orange stripes appeared on the breast feathers of one of them after ten days. Later on, the whole body was covered with mixed white and orange feathers, and the breast had be- come red. One other fowl also developed a red breast, but the remaining ten showed no change whatever. The doses of Cayenne pepper given were enormous (50 gm. daily), so that the conditions were absolutely un- natural. More remarkable than these observations are the facts ascertained by A. R. Wallace, and communicated by him to Darwin, f Thus he states that " the natives * Archiv f. Anatomic u. physiol. Physiol. Abtheil., p. 543, 1889. f " Animals and Plants," ii. p. 269. 294 THE EFFECT OF FOOD of the Amazonian region feed the common green par- rot (Chrysotis f estiva) with the fat of large Siluroid fishes, and the birds thus treated become beautifully variegated with red and yellow feathers. In the Ma- layan archipelago, the natives of Gilolo alter in an analogous manner the colours of another parrot, namely, the Lorius garrulus, and thus produce the Lori rajah or King Lory." As regards mammals, it is asserted by Nathusius * that if rich and abundant food be supplied to young pigs, it has the direct effect of producing a broader and shorter head. Poor food, on the contrary, produces a longer and narrower head, or a tendency towards the characters of the wild boar. Again, Krockerf has shown that the amount of wool yielded by sheep is greatly influenced by the quantity of food. The fol- lowing are the weights of wool yielded per day by sheep weighing in aggregate 1000 kilograms : DIET. KILOGRAMS OP WOOL. Scanty winter food, 69 Plenty of hay, 87 Good pasture, ....... .96 Fattening process, 1.081.24 The quality of the food may considerably affect the organs of digestion. Thus Cuvier $ found that in the wild boar the length of the intestines is to that of the body as 9 to 1, but in the common domestic boar it is as 13.5 to 1. It is, of course, impossible to say for certain whether this increased length was the direct result of a more vegetable diet, but it seems highly probable that * " Schweineschadel," p. 99; also " Animals and Plants," i. p. 75. f De Varigny's "Experimental Evolution," p. 90. \ " Animals and Plants," i. p. 77. AND OF PRODUCTS OF METABOLISM. 295 this was so, at least in part. The observations which have been made from time to time as to the effects of various kinds of food on the thickness of the stomach wall, are, however, free from all such doubt. The change produced must evidently be the direct result of the altered diet. Thus John Hunter observed a most marked thickening and hardening in the stomach of a gull (Larus tridadylus) which had been fed for a year on grain. It is stated by Dr. Edmondston that a similar change takes place under natural conditions every year in the stomach of the common Herring gull (Larus argentatus). Thus in the Shetland Islands this bird feeds in the winter on fish, but in the summer fre- quents the cornfields and feeds on grain. Dr. Edmond- ston has also noticed a somewhat similar change in the stomach of a raven which had been fed for a long time on vegetable food. Again, Menetries found that in an owl (Strix grallaria) the effect of vegetable diet was to change the form of the stomach, and make the inner coat leathery.* The converse experiment of feeding graminivorous birds on a flesh diet has been made by Dr. Holmgren. By feeding pigeons on meat for a considerable time, he found that the gizzard gradually acquired the qualities of a carnivorous stomach. Again, Delage f fed a fowl for three years on meat, and found that the muscular substance of its gizzard was considerably decreased. All these results, though apparently so unequivocal, have not passed unchallenged; for G. Brandes,J who * Vide " Animals and Plants," ii. p. 292. f L'Annee Biologique, 1896, p. 468. j Biol. Centralblatt, xvi. p. 825. 296 THE EFFECT OF FOOD fed both flesh-feeding birds on grain, and grain-feeders on flesh, states that he was unable to trace any adapta- tion to the altered conditions in either case. The Effects of Products of Metabolism. That organ- isms react on each othe*r has long been recognised. The interdependence is especially obvious in the case of parasite and host ; but reflection will show, I think, that the interaction is of much wider scope than is included in such self-evident cases as these. In any given volume of water, or any given area of land, every ani- mal and every vegetable organism may to some extent affect the well-being of every other organism, both ani- mal and vegetable. The animal does this largely through the agency of its own specific metabolism, or through the specific products of excretion which, com- ing into contact with the other organisms, in turn affect them. That every species of animal does possess a specific metabolism is, perhaps, scarcely what one would on a priori grounds expect; but the observations made by the author * tend to prove that such is actually the case. These observations chiefly concern Echinoids, both adult forms and plutei, but more especially the al- ready so frequently mentioned plutei of Strongylo- centrotus. On allowing the fertilised ova of Strongylocentrotus or of Echinus microtuberculatus to develop in water in which another batcfi of larvae (Strongylocentrotus, Spharechinus or Echinus) had already been developing for 8 to 12 days, but from which they had been re- moved by filtration, it was found that in every case they * Vide Mittheilungen a. d. Zool. Stat. z. Neapel., Bd. xiii. p. 389 et seq. AND OF PRODUCTS OF METABOLISM. 297 were diminished in size. In ten experiments the aver- age diminution was 7.1 per cent. It was concluded, therefore, that the first batch of larvae had excreted some products of metabolism into the water which had adversely affected the growth of the second batch. Other observations* showed that the growth of larvae may be affected by their own metabolic products. Thus it was found that the arm lengths of the larvse became smaller and smaller the larger the number of larvae al- lowed to develop together in a given volume of water. In the accompanying table are given the mean results of 159 sets of measurements, each on the anal and oral arm lengths of 50 larvae. NUMBER OP MEAN MEAN DIFFERENT NUMBER OF LARV^J PER LITRE. LENGTH Of LENGTH OF OBSERVATIONS. ANAL ARM. ORAL ARM. 37 Under 1500 121.2 118.4 32 1500 to 3500 114.0 110.5 21 3500 to 6000 105.8 101.0 34 6000 to 11,000 102.9 99.4 27 11,000 to 20,000 95.7 94.2 6 20,000 to 30,000 85.5 86.3 2 Over 30,000 56.6 68.5 Here we see that when less than 1500 larvae were devel- oping together, their relative anal and oral arm lengths were respectively 121.2 and 118.4. As the number in- creased, the lengths steadily dwindled down, till with over 30,000 per litre they became reduced to respect- ively 56.6 and 68.5, or about half their original amount. Now it was found that the body lengths of the larvae, *Phil. Trans. 1895, B. p. 603. 298 THE EFFECT OF FOOD or the dimension measured in all the observations on larvae hitherto described, was practically unaffected by the " concentration " of the larvae. This apparent con- tradiction is easily accounted for by the fact that the times of development of the body and of the arms of the larvae is not the same. At moderate temperatures, the body attains about 80 per cent, of its full length by the end of the second day, and 90 per cent, by the end of the third. The arms are practically non-existent at the end of the second day, however, and attain only 65 per cent, of their full length by the end of the third. As, therefore, the products of metabolism in the water are practically nil during the first day or two, and only gradually accumulate with progress of time, it follows that the growth of the body tissues is unaffected by them, whilst that of the arm tissues is restrained. The influence of the excreta of adult Echinoids upon larval growth was then tested. Echinoids of known weight were kept for a known time in a known volume of water, so that, on determining the absolute effect produced on larvae grown in this water, it was possible to calculate the relative effect produced by unit weight of Echinoid kept for unit time in unit volume of water. On growing larvae in water previously fouled by adult Echinoids of their own species, it was found that, as a mean of five observations, they were diminished in rela- tive size by 2.6 per cent., whilst only 41 per cent, of the ova employed reached the larval stage. On growing them in water fouled by Echinoids of other than their own species, the larvae, as a mean of five observations, were diminished by only 1.9 per cent., whilst 54 per cent, of the ova reached the larval stage. That is to AND OF PEODUCTS OF METABOLISM. 299 say, the products of excretion of an Echinoid act more adversely both on the death rate and on the growth of embryos if these belong to its own species, than if they belong to another species. At least this is the case with Strongylocentrotus, Sphcer echinus, and Echinus. With two other (physiologically) less closely related species, viz., Arbacia pustulosa and Dorocidaris papil- lata, it was even found that the products of excretion, so far from acting adversely on growth, actually fav- oured it. Thus Strongylocentrotus larvae grown in water fouled by these two species were increased in size by respectively 4.3 and 1.7 per cent., whilst respect- ively 81 and 50 per cent, of the ova employed reached the pluteus stage. It will probably be thought that this last result is erroneous; but other observations showed that it was not so. Thus Strongylocentrotus larvae were grown in water fouled by various other animals, and it was found that in this case also there was generally a distinct in- crease in size. We see in the accompanying table that ABSOLUTE ABSOLUTE AN I M A LB USED FOR POU LING PER CENT. VARIATION IN SIZE OP RELATIVE PER CENT. VARIATION ANIMALS USED FOR FOULING WATER. PER CENT. VARIATION IN SIZE OF RELATIVE PER CENT. VARIATION. WATER. LARVAE. LARVAE. 1 Fish + 1.4 + 1.8 3 Holothurians +50 +2.0 2 Fish + 8.3 +12.8 3 Holothurians +4.7 + -9 1 Crab +3 Anem- 4 Crabs + 1.6 + 2.1 ones 1.5 1.0 30 Molluscs + 2.8 + 1.3 3 Anemones .6 .5 48 Molluscs + 4.8 + 1.1 1 Medusa 2.2 -1.9 of the ten observations made, a positive effect (averag- ing 4.1 per cent.) was produced in seven instances, 300 THE EFFECT OF FOOD whilst a much slighter negative effect (averaging 1.4 per cent.) was produced in only three. The relative variation in size produced by 100 grams of animal foul- ing 1 litre of water for 1 hour is also given. These values are somewhat more variable than those repre- senting the absolute variation, but they to some extent corresponded to the amount of nitrogenous matter actually excreted into the water, as was proved by chemical analysis of the various samples. We may conclude, therefore, that under certain con- ditions products of metabolism may stimulate an organ- ism to increased growth, whilst under certain others they may retard growth. What is the nature of these excretory products which exert so potent an effect? Observations made on the influence of various simple substances on larval growth seem to throw some light on the question. The results obtained with uric acid and urea are given in the accompanying table: SUBSTANOB PRESENT, AND AMOUNT. PER CENT. VARIATION IN SIZE OF LARVAE. Uric acid, 1 in 154,000 + 5.3 " " 1 in 70,400 -j-12.2 " " 1 in 58,000 -f 5.8 " " lin 28,000 - 2.1 Urea lin 65,000 + 2.3 lin 59,000 + 3.7 lin 44,000 + 2.2 Here we see that uric acid in moderate amounts exerts a very favourable influence on the size of larvae. It is only when the proportion is raised to 1 in 28,000 (a more than half saturated solution), that an unfav- AND OF PRODUCTS OF METABOLISM. 301 curable effect shows itself. Urea also acts favourably, though not to the same extent as uric acid. If, there- fore, these two simple bodies are capable of stimulating the tissues to increased growth, it is possible that the effects produced by animal excreta may be due to minute quantities of other but more complex nitro- genous bodies. That they are not due to simple urea and uric acid was proved by the chemical analyses of the fouled waters, for the amount of nitrogen found to be present was never half sufficient, and as a rule was very much less. As to the substances producing an ad- verse influence on growth, no definite evidence was obtained, but it seemed possible that they might be derivatives of ammonia, perhaps amines or amido- bodies. Thus ammonium salts themselves exert an ex- ceedingly poisonous action, as may be gathered from the following data: WEIGHT OF AMMONIUM CHLOBIDB PER LITRE. EFFECT PRODUCED. .0258 gin. Larvae diminished 7.3 per cent, in size. .0394 " " 19.0 per cent. " .1075 59 per cent, blastulse formed. Larvae lived 3 days. .3745 37 per cent. " " No larvae. . 7890 Most of the ova had disintegrated after 24 hours . That the effect produced by nitrogenous bodies depends almost entirely upon the form in which the nitrogen is combined, is shown by the fact that nitrates and nitrites have no influence on larval growth unless the propor- tions added be over 1 gram and .3 gram per litre re- spectively. The products which every organism excretes prob- ably consist, therefore, of various complex nitrogenous 302 THE EFFECT OF FOOD bodies, which differ in different organisms. If they come into contact again with the tissues from which they have been expelled, they retard the growth of these tissues, but if with other tissues with which they have no direct chemical relation or association, they may under certain circumstances stimulate them to in- creased growth. The effects of products of metabolism upon growth have been tested at considerable length in the case of certain Molluscs. At least it is to this influence that the results obtained by Karl Semper and by De Varigny in their experiments on Limncea stagnalis, the common pond snail, ought, in my opinion, to be ascribed. Semper * found that if various numbers of the small snails were placed in equal volumes of water immedi- ately after hatching, and were kept there under other- wise equal conditions as to food, temperature, etc., for about two months, then the size to which they attained was by no means equal, but varied in more or less in- verse proportion to the number of snails present. In four very consistent experiments, the numbers of snails placed in volumes of 2000 cc. of water were in each case respectively 1, 5, 10, and 20, or each snail obtained respectively 2000, 400, 200, and 100 cc. of water. The lengths attained by the snails after two months' growth are given in the table below. Here we see that snails allowed to grow singly in the 2000 cc. vessels of water attained to more than three times the size of those grown in twenties. This was not merely a question of nutrition, as the amount of food * Arb. a. d. Zool. Inst. in Wurzburg, i. p. 137, 1874; also " Animal Life," ed. 4, p. 51. AND OF PRODUCTS OF METABOLISM. 303 supplied was always at an optimum. It was evidently in some way the result of the volume of water available for each snail's needs. Other experiments in which the number of snails was constant, but the volumes of water unequal, gave a similar result. The manner in which the volume of water affected the snail's growth, Semper confessed himself unable to determine; but he supposed that the water must contain some substance, as yet unknown, which is essential for stimulating the growth of the snails. The less of this hypothetical body available, therefore, the more retarded their growth. NUMBER OP SNAILS IN VOLUME OF WATER LENGTH IN MILLIMETRES. AVERAGE 2000 CC. PER SNAIL. LENGTH. 1 2000 cc. 17.5 19.7 18.5 17.0 18.2mm. 5 400 cc. 11.7 10.1 10.8 10.5 10.8 10 200 cc. 8.8 7.5 6.8 8.6 7.9 20 100 cc. 6.2 6.2 4.6 5.0 5.5 tr' Within recent years De Yarigny * has re-studied the unsolved problem, and has extended Semper's methods in several directions. He used both Limn&a stagnalis and what he termed L. auricularis, though this form was probably L. pereger, judging from his figures.f He confirmed Semper's conclusion that the size is in- fluenced by the number of individuals in the vessel, but he did not find the snails nearly so sensitive to differ- ences in the volume of the water as had Semper. Dif- ferences in the superficial area of the water exposed to * Journ. de 1'Anat. et de la Physiol., p. 147, 1894. fNat. Sci. v. p. 168. 304 THE EFFECT OF FOOD the atmosphere he found to be much more important than differences of volume. Thus a snail kept five months in a litre of water having a surface of 18 cm. in diameter attained to nearly twice the length of one kept in an equal volume of water which had a surface of only 2 cm. diameter. In order to test Semper's hypothesis of the essential substance in the water, De Yarigny suspended a glass tube 2 to 3 cm. in diameter in various sized vessels of water. A piece of muslin was tied over the bottom of the tube, so as to permit of interchange of water, but prevent the snails placed in the tube and in the outer vessel of water from inter- migrating. After two to five months' growth it was found that the snail placed in an outer vessel of 4200 cc. capacity sometimes attained to more than twice the length of that placed in an inner one of 250 cc. capacity. Again, snails were placed in two tubes of the same size, one of which was suspended in a vessel containing 100 cc. of water, and the other in a vessel containing 1150 cc.; in another similar experiment tl^e external volumes of water were respectively 50 and 500 cc. In each case, however, the snails in the two inner vessels attained to practically the same size. Still again, two similar tubes, holding 50 to 70 cc. of water, were placed in a vessel containing 4200 cc. of water. One tube was closed with muslin, and the other with a tight-fitting cork, which of course prevented all inter- change with the outer vessel of water. Nevertheless the snail in this tube, after two months' growth, was only very slightly smaller than that in the other tube, but both of them were only about three-fifths the size of the shell grown in the external vessel. It should AND OF PRODUCTS OF METABOLISM. 305 be mentioned that in all these experiments De Varigny lifted each tube out of its vessel of water and replaced it two or three times a day, in order to mix the water in it with that in the external vessel. He concluded, therefore, that Semper's hypothesis is not tenable, and that the size of the snails actually depends in some way on the volume of water containing them, and on the superficial area of this water. His explanation of the phenomenon is that in small vessels the snail would need to move about less in order to obtain food, for this would always be near at hand. With less exercise, the growth rate might accordingly be diminished. This explanation does not account for some of the principal results obtained by Semper and by De Yarigny himself, however. Thus in vessels of equal volume, but con- taining various numbers of snails, the amount of move- ment and exercise necessary would be just the same in each case, and yet, as we have seen, the growth rate varies enormously. In all probability, the results obtained both by Sem- per and by De Yarigny can be most simply accounted for in the manner already suggested. Thus De Ya- rigny actually found that snails grown in water in which other snails had already been growing several months were distinctly smaller than those grown in fresh water, and if the excreta of snails had been added as well, they were smaller still. If, then, the observed differences in growth are due to the accumulation of various quan- tities of products of metabolism, how can we account for the results obtained by De Yarigny in his muslin- bottomed tube experiments? We must imagine that the mixing of the internal and external waters two or 306 THE EFFECT OF FOOD three times daily, and the constant slow interchange through the muslin, were insufficient to equalise the proportions of metabolic products in the two vessels for more than a short time, so that on an average the water in the inner vessel was more foul than that in the outer. This fouling would probably be much increased by particles of decomposing vegetable matter and of animal excreta collecting in the fibres of the muslin and on the inner walls of the glass tube, and constantly poisoning the water. The outer water would also be fouled in this manner, but to a very much slighter ex- tent, for the " fouling area " of muslin and walls of ves- sels would be proportionately very much less. That the metabolic products from unhealthy or decomposing vegetable matter can exert a most harmful influence on growth is shown by some of my own experiments with plutei. Thus ova allowed to develop in water which had previously contained 1 or 2 gm. per litre of (pre- sumably unhealthy) seaweed, were diminished in size by as much as 13.2 and 18.1 per cent.* De Varigny's experiments on the influence of super- ficial area of water must be considered in conjunction with some observations by Yung f on tadpoles. Yung put twenty-five freshly hatched tadpoles in each of three vessels which contained equal volumes of water (1200 cc.), but of which the diameters were respect- ively 7 cm., 11 cm., and 14.5 cm. Thus the surface of water exposed to the air varied in the proportions of 1 : 2.5 : 4.3 After a month and a half the tadpoles were found to have attained the following average sizes : * Mittheihmgen a. d. Zool. Stat. z. Neapel., Bd. xiii. p. 348. t Arch, des Sci. Phys. et Nat., xiv. p. 502, 1885. AND OF PRODUCTS OF METABOLISM. 307 SUPERFICIAL AREA OP WATER. 1 2.5 4.3 Length of tadpoles Breadth" " ... Date of first metamorphosis 26.2 mm. 6.1 Aug. 4 34.2 7.8 July 22 41.2 8.8 June 8 The greater size of the tadpoles bred in the more ex- posed water Yung attributed to this water absorbing a larger proportion of oxygen from the air. This is in all probability the correct explanation both of these obser- vations and of the similar ones of De Yarigny on snails. The greater supply of oxygen would not only stimulate the rate of growth of the tadpoles and of the snails, but would also hasten the oxidation of the harmful products of metabolism. It is true that De Varigny found that a snail kept for eight months in a corked vessel contain- ing about 550 cc. of water and 500 cc. of air attained to only slightly less a size than another snail kept in a similar but unstoppered vessel, but this may have been due to the fact that green plants were flourishing healthily in each vessel throughout the whole period, and these may have been sufficient to remove most of the metabolic products excreted by the snails. Further evidence as to the influence of volume of water on the growth of molluscs has been obtained by Whitfield.* This observer kept a Limncea megasoma in a small aquarium, and after some months it depos- ited eggs. These hatched out, grew in size, and in due course themselves deposited eggs. This process con- tinued for four generations in all, the shells of each generation being smaller than those of the one before. * Bull. Amer. Mus. Nat. Hist., vol. i. p. 29; and Amer. Naturalist, xiv. p. 51. 308 THE EFFECT OF FOOD Those of the last generation had altered so much that a conchologist of experience was of the opinion that they could bear no possible specific relation to those of the first. Thus in addition to the diminution in size, the spire had become very slender. In a second experi- ment of a similar kind, the shells of the third genera- tion were only 4-7ths as long as those of the parent stock, and, still more remarkable, the male organs had disappeared, whilst the liver had become considerably reduced in size. These extraordinary effects were probably due to the cumulative action of the increasing quantities of meta- bolic products in the water in which the molluscs were living. Still another series of observations on the effect of a confined volume of water was described by Warren * only a year or two ago. These were made upon Daphnia magna (Water-flea). Four adult individuals were placed separately in covered glass vessels contain- ing 200 cc. of water, together with some Conferva and some mud containing algSB, etc. Four others were placed in similar, but uncovered vessels, and four more in still other vessels, of which the water was changed about once a day. The water in the former vessels was never changed, but the loss due to evaporation in the uncovered vessels was compensated for by the occasional addition of rain water. The Daphnias produced broods of four or five offspring each after about 15 days, and these offspring were allowed to grow in the vessels, and after a time produced offspring in their turn. It was interesting to note, however, that in the vessels in *Q. J. Microsc. Sci., vol. 43, p. 212, 1900. AND OF PRODUCTS OF METABOLISM. 309 which the water remained unchanged, the rate of repro- duction and the number of offspring in a brood was con- siderably diminished. The third generation was pro- duced about 22 days after the second, and the fourth about 25 days after that, and then the breeding stopped. In the vessels with frequently changed water, the third to seventh generations were produced at intervals of respectively 18, 14, 15, 16, and 26 days, and then for some unknown reason the families died out. The confined volume of water had another and even more remarkable effect, however, as it caused a distinct shortening in the length of the spine formed by the posterior prolongation of the cara- pace. Thus in one series of observations it was reduced from a length of 241 (relative to the carapace length taken as 1000) in the parents to one of 171 in the off- spring; and in another series from a length of 276 in the parents to one of 249 in the children, and 185 in the grandchildren. In this latter case, therefore, it would seem as if the acquired character of shortened spine was inherited. Warren attributes these remarkable effects to the ex- cretory matter thrown off by the Daphnias into the water. Also he concluded that this matter " may feasibly be supposed to be particularly injurious to Daphnia; for when the Daphnia are fast disappearing, there may be a swarm of Ostracods or Copepods (still living healthily in the water)." In fact Warren in- clined to the view, already suggested by the author in the case of Echinoids and other marine animals, that water fouled by Daphnia " becomes specifically injuri- ous to Daphnia." CHAPTEE X. THE EFFECTS OF CONDITIONS OF LIFE IN GENERAL. Local conditions of life perhaps the cause of local races, but proof of this is as a rule impossible American and European trees compared Alpine and Arctic plants Effects of cultivation Local races of oysters and of snails Lepidoptera in Malay Archi- pelago Local races of shrimps, of mackerel, and of herring North American birds and mammals Action of climate on goats and on rabbits Effect of domestication on rabbits, pigeons, fowls, and ducks. IN the three preceding chapters we have examined numerous cases of variation produced wholly or in great part by a change in some one condition of environment. In the present chapter no such attempt is made to trace an effect to any single cause, but we shall examine the effects of all conditions of life together, such as climate, nutrition, moisture, and sunlight, in the production oi variations. The variations more particularly to be studied are those which are common to whole groups of organisms, and which go to form what are known as local races. Unfortunately in the majority of cases it is impossible to prove that such races are the direct or indirect results of the differences of environment, even when there is a high probability that such is the case. Hence, when local races are referred to, it is not intended to imply that the distinguishing characters exhibited are definitely due to the action of the environ- 310 THE EFFECTS OF CONDITIONS OF LIFE. 311 ment. The inference is only that they may be, if not wholly, then in part. Among plants, a striking instance of the apparently direct action of conditions of life in the production of variations has been afforded by Meehan.* This ob- server " has compared twenty-nine species of American trees with their nearest European allies, all grown in close proximity and under as nearly as possible the same conditions. In the American species he finds, with the rarest exceptions, that the leaves fall earlier in the sea- son, and assume before they fall a brighter tint; that they are less deeply toothed or serrated; that the buds are smaller; that the trees are more diffuse in growth and have fewer branchlets; and lastly, that the seeds are smaller all in comparison with the corresponding European species." f The trees compared belong to several distinct orders, and are adapted to widely dif- ferent stations, hence Darwin considers that the ob- served differences should be " attributed to the long continued action of a different climate." More conclusive evidence of the direct effect of en- vironment has been obtained in the case of Alpine plants. The especially characteristic features of these plants, as compared with similar or allied plants growing at lower levels, are a dwarfing in size and compactness of growth sometimes giving rise to a moss-like appear- ance; a more intense green colour in the leaves, and greater brilliancy and size in the flowers; an increased hairiness of the leaves, and occasionally a certain degree of fleshiness of the tissues. Now by growing lowland * Proc. Acad. Nat. Sci. of Philadelphia, January 28, 1862. f Quoted from "Animals and Plants," ii. p. 271. 312 THE EFFECTS OF CONDITIONS plants at high altitudes, Bonnier,* Flahault,t and others have shown that such characters as these may be rapidly acquired. For instance, Bonnier made observations on Teucrium Scorodonia for no less than eight years, and he found that this plant, when sown at a high situation in the Pyrenees, produced very short aerial stems, with more hairy and darker green leaves, and more compact inflorescence. On the other hand, seeds gathered from plants growing at high altitudes, and sown in Paris, after three years produced elongated stems, with less hairy and brighter green leaves, or plants very similar to those from seeds obtained in the neighbourhood of Paris. The modifications acquired during a given time by a lowland plant grown at a high level, or a highland plant grown at a low level, took about the same amount of time to disappear, on returning the plants to their primitive climates. Again, Bonnier found that plants of Lotus corniculatus from Alpine situations had a very- thick epidermis, a collenchymatous cortex, and a rela- tive reduction of the wood. Those cultivated in lower altitudes had, on the other hand, a thinner epidermis, a cortex without collenchyma, and an increased devel- opment of wood. With reference to the inflorescence, " Dr. Schubeler sowed seeds of various plants in different latitudes in Norway, and proved that the brilliancy of the flowers increased with the latitude. So great were the differ- ences that it was difficult to conceive that they were produced from the same batch of seeds. The differ- Ann. Sci. Nat. Bot., vii serie, xx. p. 217, 1894. t Ann. Sci. Nat. Bot., p. 159, 1879, and Rev. Gen. de Bot., ii. p. 513, 1891. OF LIFE IN GENERAL. 313 ences appeared in the first year. Similarly, seeds from Germany exhibited analogous differences." * Dr. Schubeler also observed an increased greenness of the foliage. The Arctic climate, though similar in many respects to the Alpine, yet differs considerably in others. By comparing plants from the Islands of Spitzbergen and Jan Mayen, with specimens of the same species col- lected in the Alps and the Pyrenees, Bonnier f has shown that there are modifications of structure cor- responding to these differences of environment. The Arctic plants have more rounded cells and more con- siderable intercellular spaces in their tissues, whilst the cuticle is diminished in thickness, and the vessels are diminished in number and in calibre. These changes towards an incipiently aquatic type are probably due to the greater humidity of the air. The fleshiness of the leaves Bonnier attributes to the continuous solar illumi- nation, though it may perhaps be due to the neighbour- hood of the plants to the sea. The effect of cultivation on the variation of plants is well known to be in many cases exceedingly great; but in hardly any of the recorded cases is any mention made of the extent to which artificial selection was practiced. One cannot tell, therefore, how much ought to be at- tributed to the direct action of the environment, and how much to selection. The following instance, how- ever, seems to be the direct result of cultivation. It concerns the spiderwort, Tradescantia virginica. * Quoted from Henslow's "Origin of Plant Structures," p. 118. fRev. Gen. Bot., vi. p. 505, 1894. 314 THE EFFECTS OF CONDITIONS G. A. Brennan * records that he set out this plant in 1872, giving it very rich treatment. " In 1874 it be- gan to deviate from the original trimerous type and to assume the tetramerous one, by developing another petal, and instead of doing this at the expense of the pistil or stamens, it added another sepal, another carpel with style, and two stamens, thus making a typically tetramerous flower." In 1876 a pentamerous plant was evolved; in 1879 a hexamerous; in 1882 a dimerous; and in 1884 a heptamerous. Thus as the result of thir- teen years of cultivation, "a monocotyledonous plant has in bloom, at the same time, flowers of dimerous, trimerous, tetramerous, pentamerous, hexamerous, and heptamerous types respectively, each flower having twice as many stamens as sepals, petals, or carpels of ovary." To turn to the Animal Kingdom, an interesting in- stance of variation following directly on change of en- vironment is that noticed by Costa f in the oyster. Thus, on transferring young oysters from English shores to the Mediterranean, it was found that their manner of growth at once altered, and prominent diverging rays were formed, like those on the shells of the native Mediterranean oyster. The variations noticed by Ley- dig $ in the snail Helix nemoralis are attributed by him to the direct influence of environment, and this may be actually the case, but there is no evidence to prove it. He noticed that at Mainz the shell of this snail exhibits a fine citron yellow. This hue disappears further down the *Amer. Naturalist, vol. xx. 551, 1886. f Quoted from " Animals and Plants," ii. p. 270. J Eimer's " Organic Evolution," p. 137. OF LIFE IN GENERAL. 315 Rhine, and at Bonn and in the still lower reaches the red of the snail deepens to a chocolate brown. Cock- erell* also has noticed how sensitive is this species of snail to a changed environment. Thus it was intro- duced from Europe into Lexington, Virginia, a few years ago, and has since then varied extraordinarily. Already, in 1890, 125 varieties had been discovered in this locality. Of these no less than 67 were new, and unknown in Europe. The variations noticed by Gu- lick t in the land snails of the Sandwich Islands may also be due partly to the effects of environment. In a small forest region about forty miles by six miles in area, in the Island of Oahu, Gulick found about 175 different species, represented by 700 or 800 varieties. Successive valleys often showed allied species belong- ing to the same genus, and Gulick noticed that in every case, " the valleys that are nearest to each other fur- nish the most nearly allied forms; and a full set of the varieties of each species presents a minute gradation of forms between the more divergent types found in the more widely separated localities." Only a very few of the species ranged over the whole Island, most of them extending over only five or six miles, and a few over only one or two square miles. These variations did not appear to be due to the action of the environment, as the food, climate, and enemies in the different valleys seemed to be the same. Also the snails on the rainy side of the mountains did not differ any more from those on the dry side than they did from those inhabit- ing a neighbouring wet valley an equal distance away. * Nature, vol. li. p. 393. f Journ. Linn. Soc. (Zool)., vol. xi. p. 496. 316 THE EFFECTS OF CONDITIONS As Wallace points out,* however, " it is an error to assume that what seem to us identical conditions are really identical to such small and delicate organisms as these land molluscs." Upon Lepidoptera, we have seen in a previous chap- ter that the effect of particular conditions of environ- ment, such as temperature and nutrition, may be considerable. One would imagine, therefore, that changes in the conditions of life as a whole might form an even more potent source of variation. Conclusive evidence upon this point is, unfortunately, almost un- obtainable, though of the inconclusive kind which forms the larger part of this chapter there is plenty. For in- stance, Wallace f came to the conclusion that, with reference to the local forms occurring in the Indian and Malayan regions, " larger or smaller districts, or even single islands give a special character to the majority of their Papilionidse. For instance : The species of the Indian region (Sumatra, Java, and Borneo) are almost invariably smaller than the allied species inhabiting Celebes and the Moluccas. The species of New Guinea and Australia are also, though in a less degree, smaller than the nearest species or varieties of the Moluc- cas. . . The species and varieties of Celebes possess a striking character in the form of the anterior wings, different from that of the allied species and varieties of all the surrounding islands. Tailed species of India or the Indian region become tailless as they spread east- ward through the Archipelago; in Amboyna and Ceram the females of several species are dull-coloured, while *" Darwinism," p. 148. t" Contributions to Natural Selection," p. 167, 1870. OF LIFE IN GENERAL. 317 in the adjacent islands they are more brilliant." By actual measurement, Wallace found that " no less than fourteen Papilionidse inhabiting Celebes and the Moluc- cas are from one-third to one-half greater in extent of wing than the allied species representing them in Java, Sumatra, and Borneo. Six species inhabiting Amboyna are larger than the closely allied forms of the northern Moluccas and New Guinea by about pne-sixth. These include almost every case in which closely allied species can be compared." There are equally distinct local variations of form and colour. For instance, almost every Papilio inhabiting Celebes has wings of a pe- culiar shape, which distinguish it from the allied species of every other island. Thus the upper wings are more elongate and falcate, and the anterior margin is much more curved. A remarkable instance of the direct effects of food, or perhaps of conditions of life in general, is mentioned by Darwin, who was himself informed of it by Moritz Wagner. " A number of pupae were brought in 1870 to Switzerland from Texas of a species of Saturnia widely different from European species. In May, 1871, the moths developed out of the cocoons, and resembled entirely the Texan species. Their young were fed on leaves of Juglans regia (the Texan form feeding on Juglans nigra), and they changed into moths so differ- ent, not only in colour, but also in form, from their parents, that they were reckoned by entomologists as a distinct species." * -* Reference has already been made in a previous chap- * Quoted from Beddard's " Animal Colouration," p. 51. 318 THE EFFECTS OF CONDITIONS ter to the observations made by Weldon * on the corre- lation between certain dimensions in local races of the shrimp. The degree of correlation was found by him to be practically constant, but the mean measurements themselves show distinct differences in the various local races. The variability or range of variation of the measurements about their mean shows much greater LOCALITY. lit . !,. M O MSB a H* il ii|| 1000 shrimps 800 from Plymouth " Southport 249.63 248.31 4.55 3.96 177.53 180.29 3.50 3.65 500 Roscoff 251.51 3.32 178.00 3.02 380 " Sheerness 247.33 3.29 179.68 2.91 300 14 Helder 251.38 4.36 181.67 4.02 differences still. As we see in the accompanying table, the total carapace length (expressed in terms of the body length, taken as 1000) varies in different localities from 247.33 to 251.38, or by 1.6 per cent. However, the probable error of variation of this dimen- sion varies from 3.29 to 4.55, or by no less than 38.3 per cent. The mean post-spinous carapace length varies in its extreme limits by 2.06 per cent., and its probable error by 38.1 per cent. Now in that the shrimps differ so little in their average dimensions, they cannot be very divergent races, and hence one must conclude supposing, of course, that the samples meas- ured were fair ones, collected under similar conditions that the differences in the variability of the shrimps obtained from the various regions are due chiefly to the *Proc. Roy. Soc., vol. li. p 2, 1892. OF LIFE IN GENERAL. 319 action of a more or less correspondingly variable en- vironment. It follows, therefore, that the environ- ment at Plymouth and at Helder is more variable than that at Roscoff and at Sheerness. To turn from marine Invertebrates to marine Verte- brates, the local races of the Mackerel have recently been studied in considerable detail by Garstang.* Some of the chief of the important results obtained by him are embodied in the accompanying table : MEAN NUMBER OF LOCALITY Sde ill 1 1? a S.9 ^ ii s w s " S PH" P^ c3 j 3 \S''*7 t ' o gaa S3 2 &jq II ca SSB lill p^ S ( Lowestoft -< Ramsgate f Plymouth 300 100 300 26.75 26.88 26.79 18 28 20 1-12.14 12.00 j-84.3 81.3 [94.5 92.3 J Scilly 1 Brest 74 100 26.82 26.85 18 26 j- 12.16 j-82.2 j-93.0 j Kinsale 410 27.15 19 12.14 85 4 94.4 1 Kerry Newport, U. S. A. 245 100 27.27 27.38 10 66 12.33 11.88 85.3 63.0 94.3 97.0 The number of black transverse bars or stripes across the sides of the fish was found to vary from 23 to 33, 27 being in almost every instance the most frequently occurring number. The differences in the numbers of bars occurring in the various samples do not seem very great, but it is noticeable that all the samples from the North Sea and English Channel had invariably less than 27 bars (on an average), whilst those from the coasts of Ireland and America had more than 27. On Journ. Marine Biol. Soc., vol. v. p. 235, 1898. 320 THE EFFECTS OF CONDITIONS classifying the fish according to the number of bars, it was found that, of those in the English Channel and North Sea (including Brest and Scilly), 20 to 22 per cent, had 28 or more bars; of those on the Irish coasts, 34 to 38 per cent.; and of those on the American coast, no less than 44 per cent. The proportions of fish hav- ing one or more round black dorso-lateral intermediate spots situated between the transverse bars, showed even more distinct differences. Thus 21 per cent, of the fish from the North Sea and English Channel were spotty; 22 percent, of those from Brest and Scilly; only 15 per cent, of those from Ireland; but no less than 66 per cent, of those from America. It was found that the number of fin-rays in the first dorsal fin varied somewhat according to the size of the fish, it being, for instance, 12.33 in Irish fish under 13 inches long, and 11.92 in those of 15 or more inches. To get rid of this variable factor, only fish 13 inches long were compared. Here again the American fish showed the greatest divergence from the general mean, whilst the Brest and Scilly fish were more or less mid- way between the North Sea and Channel fish on the one hand, and the Irish on the other. As regards the second dorsal fin, the variation in the number of fin- rays is much slighter than for the first dorsal fin, it being practically only from 11 to 13 (as against 10 to 15). The American fish showed a much wider varia- tion than any of the others, only 63 per cent, of them having the modal number of 12 fin-rays, whilst the two samples of Irish fish showed least variation, 85.3 per cent, and 85.4 per cent, of them respectively having 12 fin-rays. In the number of dorsal finlets the American OF LIFE IN GENERAL. 321 fish again showed the widest variation, only 79 per cent, of them having 5 finlets. The Irish fish again showed the smallest variation, 94.4 per cent, of them having 5 finlets, as against 93.6 per cent, in North Sea and Plymouth samples, and 93.0 per cent, in those from Brest and Scilly. It is obvious, therefore, that the American mackerel constitute a distinct variety or race, the most notice- able characteristic of which is the high degree of spot- tiness. Garstang is of the opinion, also, that the mackerel which frequent the British coasts should be subdivided into two principal races: an Irish race, and an English Channel and North Sea race. The chief differences between these two subdivisions lie in the number of transverse bars and of spots, and to a lesser degree, of dorsal fin-rays and finlets. It is a striking fact, also, that " these peculiarities are greatest between the races of localities which are geographically remote, and least between those which occupy areas that are geographically contiguous. Between the mackerel of the North Sea and English Channel there are no dif- ferences at all; but the Irish race is distinctly divisible into two stocks, one of which is restricted to the west coast, and the other to the south." Into the causes of the variations shown by these local races Garstang does not enter. It is highly improba- ble that all of the observed differences are the direct or indirect result of differences of environment, but it is possible that some of them, such as the bars and spots, and size of the fish (and with_ this the number of dorsal fin-rays), may be considerably influenced thereby. 322 THE EFFECTS OF CONDITIONS The local races of the herring have been studied by Dr. Friedrich Heincke * with even greater minuteness than those of the mackerel by Garstang. Samples of herring from no less than a hundred different locali- ties were examined, most of them in respect of about 25 different characters, and some in respect of over 50 characters. Heincke came to the conclusion that the various local races of herring examined by him differed from each other in the very characters which are used to differentiate the species of the genus Clupea from each other, though, as a rule, the differences were smaller. For instance, the most extreme variations noticed in the average number of vertebrae ranged from 57.6 in the Norwegian spring herring to 53.6 in the White Sea herring, or a difference of four vertebrae. The average number of vertebrae in the species Her- ring can be taken as 56, or eight more than in the Sprat, which can be taken as having 48. On the other hand the species Clupea pilchardus has, on an average, about 52 vertebrae, or does not differ any more from the sprat on the one hand, and the herring on the other, than do the most widely divergent local races of the herring. Heincke found that, as a rule, the more widely the races are separated from each other geographically, or rather, the more their environmental conditions differ, the more do they differ in respect of certain characters. For instance, the number of vertebrae, and of scales between the ventral fin and the anus, showed the fol- lowing mean variations: * " Naturgeschichte des Herings," Abhandl. d. Deutsch. Seefisch- erei-Vereins, Bd. ii. Heft i. u, ii. OF LIFE IN GENERAL. 323 LOCAL HACK. VERTKBIUE. SCALES. s P( Au ring her umu ring, Norway, Schley, Great Belt, . Rllgen, 57.6 55.5 55.8 56.0 53.6 55.3 56.5 56.4 56.6 55.7 14.0 13.7 14.4 13.9 12.4 14.3 14.8 15.0 14.5 14.5 White Sea, . Zuidersee, . E. Coast, Scotlan North Sea (S. E.) Jutland Bank, Baltic Sea (W.), i, Heincke seems to be of the opinion that these differ- ences are largely the direct result of the environment, for he says that all the local races of herring are sub- jected to a very complex combination of environmental conditions, and that these react upon them to produce their especial characters. The White Sea herring is the most divergent from the general mean in respect of other characters besides the number of vertebra) and of scales. Thus it has only 2 to 10, or, on an average, 6 vomerine teeth, mostly in a single series, whilst other races have, on an average, 10 to 20 teeth, arranged in several series. Heincke finds that the spring herring and the au- tumn herring are two more or less distinct races, not only in the Baltic Sea, but in other localities as well (West and East coasts of Scotland, North Sea, etc.). The spring herring differs from the autumn herring in that it is, as a rule, considerably larger; it has longer anal fins, and often a larger number of vertebrae. It always has a smaller number of keeled ventral scales, and a narrower skull, and very seldom has less than 9 rays in its ventral fins. Autumn herring with 8 ven- tral fin-rays occur fairly frequently, however (20 to 30 824 THE EFFECTS OF CONDITIONS per cent, of all individuals). These differences of character Heincke attributes very largely to the differ- ent conditions of development. Thus, as regards the Western Baltic herring, the larvae of the spring brood, developing in the warm and less saline waters of the Schley, reach the young herring stage within three or four months. Those of the autumn brood, on the other hand, which hatch in the more saline waters of the open sea, need the whole winter and spring, or 7 or 8 months, to reach the same stage. The fish of the Atlantic and Pacific slopes have been studied and compared by Eigenmann.* In the eight families compared, the number of species on the At- lantic slope was more than twice as great as on the Pacific, but, nevertheless, the variation in the number of fin-rays in the Pacific species was greater in all but two of the families. The author considers that this may be due to the fauna being of diverse origin, and to its being comparatively new, and not yet in a state of equilibrium. The fish Leuciscus balteatus was studied in detail, and it was found that every locality in which it was examined had a variety peculiar to itself. As a rule, the lower the elevation of the locality from which the fish were obtained, the greater the number of fin- rays. The following are the mean values in support of this statement: NUMBER OF AVERAGE ELEVATION. SPECIMENS EX- NUMBER OF AMINED. RATS. 1750 feet, 189 18.4 1078 2000 feet, . . . . . 234 16.6 2001 3100 feet, 388 17.5 5000 feet or more, 10 16.0 * Amer. Naturalist, xxix. p. 10, 1895. OF LIFE IN GENERAL. 325 From his extensive researches on the variation and distribution of mammals and birds in North America, J. A. Allen * has been able to arrive at several general- ised conclusions concerning the geographical distribu- tion of local races. Thus he finds that, as a rule, the mammals and birds of North America increase in size as we pass from the southern towards the northern regions. In the accompanying table are given the mean values obtained by him for the length of body, PER CENT. DIFFERENCE IN DIMENSIONS OP SOUTH- CORRESPONDING DIMEN- ERN SPECIMENS. SIONS OF NORTHERN SPECIMENS. SPECIES. X w CO 3 A ^J s I |I 1 1 3 II 1 Pipilo trythrophthdlmus (townee) $ 7.88 9.88 3.56 +3.9 +14.6 5.6 Ageloeus phaniceus (red- winged blackbird) $ 9.03 14.41 3.61 +1.6 +2.1 + .6 Sturnetta ludoviciana (meadow lark) $ 9.81 15.70 2.85 +6.3 +3.8 +10.9 Sturnella ludoviciana (meadow lark) Quiscalus purpureus (purple grackle) Qtuscalus purpureus (purple grackle) Cyanura cristata (blue 9 $ 9 8.96 12.19 11.12 14.09 16.64 14.86 2.57 5.22 4.55 +6.6 +3.6 +3.0 +2;4 +6.6 +6.1 +9.7 +1.5 1.3 jay) ___ 10.98 15.11 5.00 +6.6 +11.6 2.2 Colaptes auratus (golden- winged woodpecker) _ 11.66 18.82 4.40 +6.8 +6.0 1.1 Ortyx virginlanus (com- mon quail) Ortyx virginiamis (com- mon quail) $ 9 9.46 9.37 14.16 14.02 2.52 2.54 +7.6 +4.9 +9.0 +7.7 +11.9 +5.1 stretch of wings, and length of tail of seven different species of birds. In the middle portion of the table are given the actual values (in inches) for the Southern races, from Florida, whilst in the right half are given * Bull. Mus. Comp. Zool. Harvard, vol. ii. p. 161-490; also Cope's "Factors of Organic Evolution," p. 45. 326 THE EFFECTS OF CONDITIONS the percentage variations, on these values, of the cor- responding values for the Northern races (from North- ern States, Massachusetts, and Southern New Eng- land). On an average fifteen specimens were measured in each case, the extreme numbers varying between 6 and 40. As regards body length, we see that the North- ern forms invariably exceeded the Southern, the aver- age difference amounting to 5.1 per cent. In alar extent they were likewise invariably greater, the average ex- cess being 7.0 per cent. In tail measurement, how- ever, the difference was not nearly so constant, it being greater in the Southern races than in the Northern in four out of the ten sets of measurements, whilst the average excess amounted to only 2.9 per cent. Accompanying the increase in size of the Northern forms, Allen finds that, as a rule, there is an apprecia- ble decrease in colour. In the South, dark-coloured birds, such as the red-winged blackbird, become blacker. The slaty and olive tints of other birds, and the various shades of red and yellow, become far more intense as one proceeds south, and the pigmentation of the bill and feet also increases. Allen says " the dif- ference in colour between the extremely Northern and extremely Southern representatives of a given species is often so great that, taken in connection with other differences, as in the general size and the size and form of the bill, the two extremes might excusably be taken for distinct species." The size of the bill varies, as a rule, in the inverse ratio to the size of the body, and " in many species there is not only a marked relative increase in the size of the bill to the southward, but in some an absolute increase, OF LIFE IN GENERAL. 327 especially in its length." This increase is quite marked in the genera Quiscalus, Agel&us, Geothlypis, Troglo- dytes, Seiurus, etc. As to the causes of these geographical variations, it is of course impossible to ascribe them with any cer- tainty even to the indirect effects of change of environ- ment, much less to the direct. Still, as Allen points out, there is often a somewhat close correlation be- tween geographical varieties and the meteorological peculiarities of the regions in which they occur, which suggests a connection of some sort between the two. The increase in colour towards the south coincides with the increase in the intensity of the sun's rays, and in the humidity of the climate. The increase in colour observed in birds on passing from East to West seems also to coincide with an increase of humidity, " the darker representatives of any species occurring where the annual rainfall is greatest, and the palest where it is least." This coincidence occurs not only in the birds of the United States, to such a degree that Allen says he knows of no exception, but in Europe also. Thus birds from the Scandinavian coast are very much darker than in central Europe, where the rainfall is only half as great. Allen says that this cor- relation of brighter and deeper tint with increased humidity is exhibited by the mammals of these dis- tricts, as well as by the birds. The differences in the local races of certain Mam- mals are even more striking than in those of the birds. The Canidse, for instance, are represented in North America by six species, viz., gray wolf, common fox, gray fox, coyote, arctic fox, and kit fox, of which the THE EFFECTS OF CONDITIONS first three are the widest ranging species. Allen found that, in respect of skull measurement, the com- mon wolf is fully a fifth larger in the northern parts of British America and Alaska than it is in northern Mexico, the southern limit of its habitat, whilst, as we see in the accompanying table, specimens from inter- mediate regions show a gradual intergradation between these extremes.* The common fox from Alaska is about 10 per cent, larger than that in New England, ogo I 8 j ggg ^ II 8PECIKS. LOCALITY. * en :S SB! 3o HO I 8 Forts Simpson, Yukon and Gray *if Rae, Forts Benton and Union, 9 16 10.38 in. 9.45 5.40 in. 5.07 wolf Forts Kearney and Harker, 9 9.69 5.18 Rio Grande and Sonora, 3 8.37 4.31 Alaska, 9 5.98 3.20 Common Mackenzie River District, 18 5.80 3.02 fox Upper Missouri, Essex County, New York, 9 12 5.78 5.40 2.90 2.80 whilst the gray fox probably varies considerably more in size with locality, but the number of skulls obtained for measurement (15 in all) is insufficient to warrant any generalisation in its case. This increase in size on passing from South to North is not universal, however. Thus lynxes and wild cats, though series of skulls were obtained from such widely separated localities as Alaska, California, and Northern Mexico, revealed no appreciable variation of size with * U. S. Geol. and Geographical Survey, vol. ii. p. 309, 1876. OF LIFE IN GENERAL. 329 locality. Panthers and ocelots, indeed, showed a very considerable increase in size on passing southward. Still the increase in size of Carnivorous Mammals on passing from South to North may be taken to be a very general rule. In addition to the Felidse men- tioned, this relation is well shown in the badger, mar- ten, wolverine, and ermine. Of other Mammals, the relationship between locality and size is well shown by members of the deer family, the Virginia deer affording an especially striking in- stance. The Glires (squirrels, marmots, mice) also increase, as a rule, towards the North. For instance, the northern race of flying squirrels is half as large again as the southern, but these two extremes are con- nected by a complete chain of intermediate forms. As in the case of birds, mere size of body is not the only characteristic which varies with locality. The ears and the feet may undergo considerable changes in addition. Thus in mammals with large ears, such as wolves, foxes, some of the deer, and especially the hares, there is often a striking increase in the size of these appendages on passing from North to South. The ears of the little wood hare (Lepus sylvaticus), found in Western Arizona, are nearly twice the size they attain in the variety found in more easterly and northerly regions. Again, in Lepus callotis the ear is one-third to one-fourth larger in examples obtained from Mexico than in those from Wyoming, whilst the little brown hare (L. trowbridgei) shows a similar in- crease in the size of the ear in the south. Darwin * has collected several cases in which climate * " Animals and Plants," ii. p. 268. 330 THE EFFECTS OF CONDITIONS had an influence on the hairy covering of animals. Thus he says " Dr. Falconer states that the Thibet mastiff and goat, when brought down from the Him- alaya to Kashmir, lose their fine wool. At Angora not only goats, but shepherd-dogs and cats, have fine fleecy hair, and Mr. Ainsworth attributes the thickness of the fleece to the severe winters, and its silky lustre to the hot summers. Burnes states positively that the Kara- kool sheep lole" their peculiar black curled fleeces when removed into any other country." What may be termed, perhaps, the classical instance of the formation of a local race through changed con- ditions of life, is that of the Porto Santo rabbit.* A female rabbit and her young were turned loose on the Island in 1418, and they increased so rapidly as to be- come a nuisance, and finally caused the abandonment of the settlement. The present-day form of these rab- bits shows very considerable differences from the original form. Thus the two measured by Darwin were only 14r| and 15 inches in length, instead of the 17 or 18 inches of the English rabbit. The weight of one of them which had, however, become somewhat thin from living in captivity was only 1 pound 9 ounces, four English wild rabbits averaging 3 pounds 5 ounces. The limb bones were smaller in the proportion of rather less than five to nine. In colour, the Porto Santo rabbits have a redder upper surface, rarely interspersed with any black or black- tipped hairs, and in none of the seven specimens exam- ined by Darwin had the upper surface of the tail and the tips of the ears any of the blackish gray fur which * " Animals and Plants," i. p. 118. OF LIFE IN GENERAL. 331 is generally regarded as one of the specific characters of the rabbit. Finally, two male Porto Santo rabbits, when kept in captivity, never lost their extreme wild- ness, and would never associate or breed- with the females of various breeds placed with them. To what extent these remarkable changes were the direct result of a changed environment, it is, of course, impossible to say; but it was proved th^^t least the colouring was a direct effect. Thus o^^of the feral rabbits, after being kept for four years in captivity, was found by Darwin to have acquired a^jlackish gray fur on the upper surface of the tail and the edges of the ears, whilst the whole body was much less red; i. e., it had recovered the proper colour of its fur after four years of English climate. The influence of domestication combined with arti- ficial selection is well known to everyone, but what shares of the changes produced are to be assigned to each of these agencies is, as a rule, quite indeterminable. However, one may with some reservation ascribe to domestication changes effected in characters which have never been the subject of selection. For instance, the weight of the rabbit was found by Darwin to in- crease on domestication, the result, probably, both of more ample feeding and of artificial selection. The skull capacity, on the other hand, by no means propor- tionately increased in size; and as this is scarcely a character on which selection would be practised, we may consider the relative diminution as probably due to the direct influence of domestication. The reason why we cannot say with absolute certainty that it is a direct effect, lies in the fact that the character of 332 THE EFFECTS OF CONDITIONS "skull capacity" may be more or less closely corre- lated with some other character which has been the ob- ject of selection, and so have been thereby uninten- tionally modified. In the accompanying table are given the mean values of Darwin's measurements:* 4 BREED OP RABBIT. WEIGHT. i LENGTH OP BODY IN INCHES. LENGTH OP SKULL IN INCHES. CAPACITY OP SKULL. RATIO OP BRAIN CAPACITY TO LENGTH. 7 various wild rabbits . 3 Porto Santo rabbits . 4 various domestic rabbits 7 large lop-eared rabbits . 2 Ib. 15 oz. (1 Ib. 9 oz.) 3 Ib. 4 oz. 7 Ib. 4 oz. 17.1 (14.75) 19.75 34.62 3.09 2.88 3.47 4.11 950 828 864 1136 100 101.1 78.8 83.1 The weight of most breeds of domestic rabbit is not much greater than that of wild ones, but that of the lop-eared variety is more than twice as great. The length of body was measured from incisors to anus, whilst the capacity of the skull was determined by weighing the small shot taken to fill it (the numbers given in the table being the weight in grains). Tak- ing the relation of capacity of skull to length of body in the wild rabbit as 100, we see that, on an average, the skull capacity of the domestic rabbit is about 20 per cent. less. That of the Porto Santo rabbit is, on the other hand, very slightly greater. The diminution in the size of the rabbit's brain is at- tributed by Darwin, to the effects of disuse, and * " Animals and Plants," i. p. 133. OF LIFE IN GENERAL. 333 he ascribes certain of the changes in other domesticated animals to a similar cause. Thus he found the length of the sternum in eleven different breeds of domestic pigeon to be on an average 13.0 per cent, shorter than in the wild rock pigeon.* The crest of sternum, scapulae, and furculum were also reduced in size, but the wings were slightly increased, owing to the greater length of the wing feathers. Again, in eight out of the eleven breeds of fowl examined, the weight of humerus and ulna, relative to that of femur and tibia, was, on an average, 24.2 per cent, less than in the wild Gallus bankiva, and in all eleven breeds the depth of the crest of the sternum (to which the pectoral muscles are attached), relative to its length, was diminished, the average diminution being 17.5 per cent.f In the case of the duck, Darwin weighed the entire skeleton, as well as individual parts, and he found that whilst in the four breeds of domestic duck examined the weight of the femur, tibia, and tarsus, relative to that of the body, was, on an average, 28.5 per cent, greater than in the wild mallard, that of the humerus, radius, and meta- carpus was 9.0 per cent. less.$ That this decrease in the weight of the wing bones is the direct result of dis- use was proved by the fact that in a domestic call duck which was in the habit of constantly flying about for miles, the relative weight of the wing bones was actu- ally 12.1 per cent, greater than in the wild mallard. * "Animals and Plants," i. p. 184. \Ibid. i. p. 285. J/W&, i. p. 301. PART III. VAKIATION IN ITS KELATIOISr TO EVOLUTION. CHAPTER XI. THE ACTION OF NATURAL SELECTION ON VARIATIONS. "Selection does nothing without variability, and this depends in some manner on the action of the surrounding circumstances on the organism" (Darwin, " Animals and Plants" i. p. 7). ' ' The foundation of the Darwinian theory is the variability of species " (Wallace, "Darwinism," p. 41). " What forms the basis of the constant 'individual variations 1 which, after the precedent of Darwin and Wallace, we regard as the foun- dation of all pi'ocesses of natural selection ? " ( Weismann, " Oerm- Plasm,"p. 410.) Proof of Natural Selection in the crab and in the sparrow Selection in man Evolution of the mouse Inheritance of acquired char- acters seems to be shown by cumulative effects of conditions of life, as European climate acting on American maize; domestica- tion acting on wild turkeys and ducks ; changed climate acting on sheep and dogs Environment may act on germ- plasm through specific excretions and secretions Cases of inherited effects of use and disuse, and of epilepsy, accounted for Somatic variations may increase variability, and so afford Natural Selection a better handle to work upon. THE contents of this chapter are well summarised in the three quotations given at its head. It deals with variations in their relation to Natural Selection, and 336 ACTION OF NATURAL SELECTION with the gradual evolution thereby brought about. The fundamental importance of variations in the evo- lutionary process has been dwelt on again and again by Darwin, by Wallace, and by most of the subsequent writers on the subject, and as this doctrine is so uni- versally admitted, it is unnecessary to discuss it any further here. At the present day, however, there ap- pears to be a considerable amount of scepticism among some men of science as to the extreme importance which has been generally attached to the agency of Natural Selection. Some, such as Driesch, have even denied its existence altogether, whilst many others hold that its existence has never been demonstrated. They hold with Lord Salisbury * that " no man, as far as we know, has ever seen it at work." The evidence to be adduced will show, I believe, that this statement is erroneous, but even if it be correct, it cannot dis- prove the theory, the validity of which seems to me a logical necessity. Thus, granted the geometrical rate of increase possessed by all organisms, and the severe struggle for existence thereby entailed; granted that all organisms show individual variations, and, to a con- siderable extent, hereditary transmission of such varia- tions, then it must follow that, on an average, more of the organisms possessing favourable variations better adapted to their environment will survive than of those possessing less favourable ones. That is to say, the species will become gradually modified by the action of Natural Selection. Numerical evidence in support of the theory of Natural Selection has been obtained only quite * "Presidential Address, British Association," 1894. ON VARIATIONS. 337 recently, and this is not to be wondered at, considering the numerous and extended observations it, as a rule, entails. In the case of the small shore crab, Carcinus mcznas, however, Professor Weldon * has succeeded in overcoming most of the inevitable difficulties and pit- falls, and has obtained evidence which, though at present not absolutely convincing, yet has a very high degree of probability of truth. In 1893 Mr. H. Thompson carefully determined the relation of the mean frontal breadth to the carapace length in male crabs collected at a particular patch of beach in Plymouth Sound. The mean breadth was found to vary very rapidly with the length of the crab, hence its value was determined separately in small groups of crabs, differing from each other by not more than .2 mm. Twenty-five such groups, for crabs between 10 and 15 mm. long, were measured in respect of fron- tal breadth. A similar series of measurements was carried out by Thompson on crabs collected at the same spot in 1895, and another by Weldon on crabs collected in 1898. On comparing the three series of data thus obtained, it was evident that the mean breadth of crabs of a given carapace length had steadily decreased. For instance, in crabs with a carapace length of 11.5 mm., the frontal breadth had a percentage length of 79.72 in 1893, 78.88 in 1895, and 78.40 in 1898. Again, in 14 mm. crabs, it had a length of 76.26 in 1893, 75.44 in 1895, and 74.44 in 1898. It would seem, therefore, that the frontal breadth of these crabs is diminishing, year by year, at a very rapid rate. This Professor Weldon attributes to a selective * Report of Brit. Assn., 1898, p. 887. 338 ACTION OF NATURAL SELECTION destruction, caused by certain rapidly changing condi- tions in Plymouth Sound. Owing to the building of a huge breakwater, the scour of the tide has been dimin- ished, and the large quantities of china clay carried down by the rivers from Dartmoor into the Sound therefore settle in increasing quantities in the Sound itself. Also the quantity of sewage and refuse finding its way into the Sound is steadily increasing, owing to the increase in the size of the contiguous towns and dockyards. " It is well known," says Professor Wei- don, " that these changes in the physical condition of the Sound have been accompanied by the disappearance of animals which used to live in it 2 but which are now found only outside the area affected by the break- water." In order to test his supposition of selective destruction, Professor Weldon placed a number of crabs in a large vessel of sea water, in which a consider- able quantity of very fine china clay was suspended. The clay was prevented from settling by a slowly mov- ing automatic agitator. After a time, the dead crabs were separated from the living, and both were meas- ured. In the figure given below is shown the result obtained. Here the upper curve shows the distribution of fron- tal breadths of the 248 male crabs experimented on, and the dotted curve the distribution of frontal breadths of the 94 survivors. The line represents the mean frontal breadth of all the crabs, the dotted line S the mean of the survivors, and the dotted line D the mean of the dead crabs. The crabs which survived thus had a distinctly smaller frontal breadth than those which were killed, just as the 1898 crabs had a smaller ON VARIATIONS. 339 breadth than the 1895 ones, and these than the 1893 ones. There seems no reason to believe that the action of the mud upon the beach is different from that in an experimental aquarium, and hence, in Professor Wei- don's opinion, there is " no escape from the conclusion that we have here a case of Natural Selection acting with great rapidity because of the rapidity with which the conditions of life are changing." The selective de- struction seems to depend on the nitration of the water into the gill chambers of the crabs. To quote Pro- fessor Weldon, " The gills of a crab which has died during an experiment with china clay are covered with fine white mud, which is not found in the gills of the s D 30 20 10 A A ~2_ ^ - ^> i ^ X, --A ^M FIG. 26. Distribution of the frontal breadths of 248 male crabs, and of the 94 survivors. survivors. In at least 90 per cent, of the cases this difference is very striking." Professor Weldon thinks it can be shown that a narrow frontal breadth renders one part of the process of filtration of water more effi- cient than it is in crabs of greater frontal breadth. Such, then, is Professor Weldon's demonstration of a particular instance of Natural Selection. In order to strengthen the proof of its existence, further meas- urements of crabs collected at the same spot a few years 340 ACTION OF NATURAL SELECTION hence ought to be made, as Professor Weldon him- self well recognises, in order to see whether the de- structive process is still continuing. If this is the case, and if crabs measured in, say, 1903 and 1908 show a further diminution of frontal breadth, then the evi- dence in favour of selection would amount to a very high degree of probability indeed. Owing to the changing relation of its parts with growth, the crab is a somewhat unsatisfactory organism to work with, and, indeed, the apparent change observed between 1893 and 1898 might possibly, though not probably, owe its origin to quite another cause than Selection. For in- stance, the conditions of environment such as tempera- ture, nutrition, and purity of the water may have acted directly on the crabs so as to retard their growth. Now Professor "VYeldon assumes that all crabs of, say, 12 mm. length are approximately the same age, but ob- viously this need not be so from year to year. Under less favourable conditions, the moulting may have gone on as usual, but the rate of growth have been reduced. Now we have seen that the frontal breadth diminishes very rapidly with growth, and hence it might happen that the narrower fronted 12 mm. crabs of 1898 are narrower simply because they are older than were the more favourably situated 12 mm. crabs of 1893. Mr. J. T. Cunningham * has pointed out that in 1893 the temperature of the Channel waters was abnormally high, and he considers that this produced a more rapid growth of the crabs, and hence, for a given size of crab, an apparent increase of frontal breadth. However, Professor Weldon f does not believe that the tempera- * Nature, vol. Iviii. p. 593. \Ibid., p. 595. ON VARIATIONS. 341 ture of the beach where his crabs were collected, in that it looks due south and is uncovered for hours daily, was any lower in 1898 than in 1893, and also he found that crabs gathered in January were no narrower fronted than those gathered in August, as they ought to have been on Cunningham's hypothesis. The proof of the existence of Natural Selection really centres upon the proof of a selective destruc- tion or death rate. If among any group of organisms the eliminated individuals can be measured and exam- ined, as well as the survivors, and if it be found that these two divisions differ in their mean characters, then Natural Selection must have been at work. Very likely the parts or organs measured do not represent the characters upon which the selective process had been acting, but are merely correlated with them. But that is no matter. The offspring of the survivors will have different average qualities from those of the previous unselected generation, or the race will be- come modified by Natural Selection. Unfortunately in the majority of cases, as in Pro- fessor Weldon's crabs, it is impossible to get hold of the eliminated individuals, and hence the proof of Natural Selection is rendered much more laborious, and at the same time more open to possible source of error. In the case of the (introduced) English spar- row (Passer domesticus), however, Bumpus * has suc- ceeded in obtaining the desired material. One hun- dred and thirty-six of these sparrows were collected after a very severe storm of snow, rain, and sleet in North America, and of these 72 revived, whilst 64 * Biol. Lectures, Wood's Holl, 1898, p. 211. 342 ACTION OF NATURAL SELECTION perished. On comparing the survivors with the elimi- nated individuals, very appreciable differences were found to exhibit themselves. The means of the values obtained with all the birds, both male and female, are given in the accompanying table : MEAN VALUES. ARITHMETIC MEAN ERROR. i i . H GQ g i H HH H H W 1 g o PS E ^ B Q g fc B H tf ^ i HI a H Total length, Alar extent, 158 mm. 245mm. 160 mm. 245mm. +1.27 0.0 2.51 4.20 3.48 4.60 +38.6 + 9.5 Weight, Length of beak & head Length of humerus, Length of femur, Length of tibio-tarsus 25.2 gm. 31.6 mm. .736 inch .716 " 1.138 " 25.8 gm. 31.5 mm. .728 inch .709 " 1.128 " +2.38 .32 -1.09 - .98 .88 10.9 5.51 .016 .014 .0294 12.6 5.64 .0201 .020 .0338 +15.6 + 2.4 +25.6 +42.9 +15.0 Width of skull, .603 .601 " .33 .010 .012 +20-0 Length of sternum, .845 " .834 * 1.30 .032 .038 - 3.1 Here we see that the average characters differ but little. The eliminated individuals are 1.27 per cent, greater in length, and 2.38 per cent, greater in weight, whilst they are about 1 per cent, smaller than the survivors in respect of most of the other characters measured. The variability, or range of variation of the eliminated birds about their mean, is, however, very much greater than in the case of the survivors. Of the nine characters measured, the varia- bility is greater in eight, the average excess being no less than 18.8 per cent. The variability was less in re- spect of the sternum alone, and then only by 3.1 per cent. In the accompanying figure are given curves of distribution of the lengths of the surviving and of the eliminated birds. It is obvious that the dotted line ON VARIATIONS. 343 curve, which represents the eliminated individuals, is, on the whole, much more flat-topped than the other curve. The very long individuals seem especially handicapped in the struggle for existence, as of the 18 birds obtained in which the length was 164 mm. and upwards, no less than 14 perished. Also the two shortest birds obtained perished. The conclusion which Bumpus draws from these most interesting ob- servations is as follows : " Natural Selection is most destructive of those birds which have departed most from the ideal type, and its activity raises the gen- eral standard of excellence by favouring those birds which approach the structural ideal." The observa- tions really show more than this, however. It is 15 40 \ 3M 256 B8 150 153 154 - i'56 J58 JI6Q _ Length of birds in millimeters. FIG. 27. Distribution of the lengths of surviving and of eliminated sparrows. true that, as a rule, the most extreme individuals in either direction are eliminated, but if the distributions of the various characters be plotted out as in the above figure, it will be seen that in the case of some of the other characters, as in that of length, the elimi- nating process acts much more on the extreme in- dividuals in one direction than on those in the other. 344 ACTION OF NATURAL SELECTION For instance, the accompanying figure shows the distri- bution of the weight values of the birds. The curves are very irregular, but it is obvious that the dotted line curve is shifted distinctly to the right, indicating that the eliminated birds were, on an average, heavier. This conclusion has already been obtained by the simple process of taking averages; but the curves show in ad- dition that it is the very heavy birds which were more 115 lo 8 5 V 24 25 26 27 28 29 Weight of birds in grams 30 31 32 FIG. 28. Distribution of the weights of surviving and of eliminated sparrows. especially eliminated. Thus of the 14 birds of 27.3 gms. and upwards obtained, only three survived. Similarly also in respect of the femur measurements, it was found that of the 19 birds obtained with a femur length of .685 inch or less, only 7 survived whilst 12 were eliminated. The next generation of birds collected in tEe storm- swept area would accordingly be shorter in length, weigh less, have longer legs, have a longer sternum and a greater brain capacity than the former generation; supposing, of course, that the variations existing in ON VARIATIONS. 345 these characters were partly of blastogenic, and not wholly of somatogenic origin; and this could scarcely fail to be the case. Several of the changes in char- acters, especially of the total length, weight, and femur length, might possibly be present, on an average, to just as marked an extent as in their parents (the survivors of the previous generation); for though the characters would tend to undergo reduction by^ virtue of their " regression towards mediocrity," yet they would tend to be enhanced by reason of the fact that more of the extreme individuals (which would be of propor- tionally greater weight in determining the characters of the next generation) had been weeded out than of the mediocre ones. Bumpus does not give any details as to the way in which the sparrows were collected, and whether the sample obtained was repre- sentative of all the sparrows in the area in question. Supposing it were not, and it included only spar- rows which were exposed to the force of the storm through failing to get adequate shelter, then, of course, the average change produced in the characters of the next generation would be much less than that suggested by the above figures. Professor Weldon * has adopted a very ingenious method for determining the presence or absence of Nat- ural Selection in the case of a certain terrestrial mol- lusc, Clausilia laminata. The outer layer of the shell in this and other pulmonates is secreted by the growing edge of the mantle once and for all, and it undergoes practically no subsequent change. The upper whorls of an adult shell therefore afford an unaltered record *Biometrika, i. p. 109, 1901 346 ACTION OF NATURAL SELECTION of the condition of the young shell, from which this adult was formed by the subsequent deposition of new material. By measuring the upper whorls of the adult shells, one is accordingly able to determine the char- acters possessed, not by all young shells, but by the young shells which were successful in attaining the adult condition. How would measurements on such adult shells compare with those on young and growing shells, some of which would almost certainly undergo destruction before attaining their full development? To answer this question, Professor Weldon measured with great exactness the radius of the spiral at various (angular) distances from the apex of the shell in 100 adult individuals, and also in 100 young individuals of less than half their length. The means of the values so obtained were practically identical in the two classes of shells, so it seems to follow that the mean spiral of young shells is not altered during growth by any process of selective destruction. On the other hand, the varia- bility of the radial spiral measurements was consider- ably greater in the young shells than in the adult ones (on an average, in the proportion of 120 to 100 for the first whorl and a half.) Hence we may conclude that during the growth of this mollusc some processes are at work which effectually eliminate the abnormal shells more rapidly than the others, and so diminish the variability of the survivors. As the average character of the race does not undergo any change, it follows that the abnormalities eliminated are evenly distributed about the mean. Such a process of selection has been termed by Professor Pearson * periodic. *" Grammar of Science," p. 413. ON VAEIATIONS, 347 The proof given by Professor Pearson * of the exist- ence of a selective death rate in the case of man seems to me scarcely to entitle him to speak of it as a case of " Natural Selection." Thus he shows that there is a fairly close correlation (r = .26) between the ages at death of brother and brother, and a less close one be- tween those of fathers and sons (r .12 to .14). For instance, the mean age at death of men not dying as minors is 61 years. If, however, one brother of a pair dies at 25, then the other will, on an average, die at 51.6, or 9.4 years earlier than the mean; if one dies at 85, then the other will, on an average, die at 67.2, or 6.2 years later. There is something in the constitution of a man, therefore, which to a large extent determines when he shall die, or undergo elimination. His death is not at all a matter of chance. Further analyses of data by Miss Beeton and Professor Pearson f indicate the same thing, though they also lead to other and some- what unexpected conclusions. Thus, from the pedigree records of members of the society of Friends, the au- thors found that elder (adult) brothers and sisters on an average live distinctly longer than younger (adult) brothers and sisters, and that the greater the difference in age, the greater the difference in expectation of life. For instance, a man born 6 years after his elder brother will probably live 4 years less than he will; one born 10 years after, 7 years less, and one born 17 years after, as much as 12 years less. Put in other words, the eldest children of a family have the best chance of life, *" Grammar of Science," p. 497; also Beeton and Pearson, Proc. Roy. Soc., Ixv. p. 290, 1899. fBiometrika, i. p. 50, 1901. 348 ACTION OF NATUEAL SELECTION and the youngest the worst. We may perhaps look upon this decrease of vitality as the direct result of the diminished vigour of the parents at the time of con- ception, and of the mother during the period of intra- uterine development. If this is actually the case, then we ought to find that the expectation of life is more closely correlated with the age of the mother at the time of conception than with that of the father. Again Miss Beeton, in conjunction with Mr. G. U. Yule and Professor Pearson,* have found that there is a direct correlation between the duration of life in parents, and the number of children borne by them. It was found that fertility was correlated with longevity even in parents of 50 years and upwards, when the fecund period is passed, though the correla- tion is not by any means so close as in parents under 50. For instance, American mothers dying at 25 had on an average 2.2 children; those at 35, 4 children; and those at 50, 5.7 children: but mothers dying at TO had on an average 6.8 children, and those at 90, no less than 7.6 children. "With English mothers dying at 50 years and over, the increased fertility is not so marked, and it becomes slightly diminished in those living over 75 years. Similarly also with English fathers the re- lation of fertility to longevity is less marked than in the case of American fathers. All these data may undoubtedly be taken to indi- cate, therefore, that longevity is inherited in man, and as long life means a healthier and stronger constitu- tion, it is natural to find that it also betokens increased *Proc. Roy. Soc., Ixvii. p. 159, 1900. ON VARIATIONS. 349 power of procreating offspring. It does not neces- sarily follow, however, that Natural Selection, in the ordinary sense of the term, is at work. The time of death may be quite uncorrelated with any particular structural characters of the body, but be dependent only on the so-called vigour or vitality of the organism. Each subsequent generation may therefore be more " vigorous " than the one before it, owing to the elimi- nation of a portion of the less vigorous individuals, but as, in all probability, there is always a tendency to the production in each generation of a certain number of unfit individuals, or a slight diminution in the average vitality of the whole group, it would follow that a cer- tain amount of elimination is always necessary, to en- able a race to maintain its average vitality from one generation to the next. Certainly, in the case of the human race, there is no evidence that the average vigour and vitality are increasing. Everything goes to prove rather that they are on the wane. Man is therefore an unsatisfactory organism in which to determine either the existence or the non-existence of Natural Selection. His conditions of death are as unnatural as his conditions of life. The usual cause of his death, disease, counts for little or nothing amongst the lower animals, whilst the usual causes of death amongst them, namely, want of food and natural enemies, count for little or nothing with man. To prove the existence of Natural Selection, one must choose for observation an organism living under nat- ural conditions. A very interesting case of the formation of a local race through the probable agency of Natural Selection 350 ACTION OF NATUEAL SELECTION has recently been described by H. L. Jameson.* On the north side of Dublin Bay there is a tract of sand- hills, running along the coast for about three miles. It is separated from the mainland by a tidal channel about a quarter of a mile wide at high water, but only 20 yards or so at low water. These sandhills are thickly populated with mice, which were noticed by Jameson to harmonise strikingly in colour with the sand. Traps were set, and altogether 36 mice were caught. The specimens varied considerably in the shade of their fur, showing every gradation from the typical Mus musculus of the farmhouses in Ireland and England to individuals with extremely pale dorsal fur usually of a rufous or fulvous gray colour pale buff ventral sur- face, and pale and fulvous appearance of the hairs on the ears, tail, and other parts of the body. Also the feet were white or pale buff, instead of the smoky gray or white of the ordinary House-mouse, whilst the claws were flesh-coloured. Of the 36 specimens, only five were of more or less the typical colour, four were slightly paler, and the remaining 27 markedly palles- cent. These mice differ in other characters also. Thus, if the adult specimens be split up into three groups, ac- cording to their colouration, and means taken of the measurements made by Jameson, the values given in the table below are obtained. Though the number of measurements is so small, there can be little doubt that the tail of the palest individuals is distinctly longer than that of the typical ones. Per- haps also the head and body and the hind foot are * Journ. Linn. Soc. (Zool), vol. xxvi. p. 465, 1898. slightly shorter, though one cannot speak with any cer- tainty. The habits of the mice had changed in addi- tion, as they were found to burrow their own holes, no holes burrowed by other animals being available, as in the case of the typical wild mouse. The development of the protective colouration and habits probably owes its origin to the short-eared owls and hawks which were noticed to frequent the sand- hills, and which would more readily perceive and cap- ture the darker mice. These would gradually be weeded out, therefore, whilst the light-coloured indi- viduals would survive and propagate their more favour- able characteristics. LENGTH IN MILLIMETRES OF NUMBER OF MEASURES. Head and Body. Tail. Hind Foot. 7 Typical or slightly pale 81.3 76.5 18.1 9 20 Markedly pale Still more pale 76.3 80.2 77.8 80.0 16.9 17.2 Perhaps the most interesting point of all about these observations is that it has been found possible to fix a time limit for the duration of the evolutionary process. The sandbanks are known to be gradually increasing in area, and, by a careful study of old maps, Jameson found that previous to 1780 they did not exist at all. In 1810 the island was only a quarter of a mile long, so we may conclude that the pale race of mice has had not more than about a hundred years for its evolution. Are Acquired Characters Inherited? We see that 352 ACTION OF NATURAL SELECTION Evolution is brought about by the action of Natural Selection on variations, it selecting some and rejecting others, and so gradually altering the average char- acters of the race; but are blastogenic or germinal variations alone of value to such a selective agency, and are somatic variations, or so-called acquired char- acters, valueless in this respect ? As is well known, the question of the heritableness of acquired characters has been one of the most hotly debated of all biological problems, and is one which even now separates most biologists into two opposite and apparently irreconcil- able camps. It behoves us, therefore, to see if we can- not find some via media, which, though unable to ad- mit of the heritableness of localised tissue changes such as injuries and mutilations, is yet able to adopt reasonable evidence, both experimental and theoretical, in favour of a partial inheritance of certain general- ised tissue changes. The chief experimental evidence in support of the apparent heritableness of acquired characters lies in the numerous and undoubted proofs of the cumulative action of conditions of life. Of such proofs, one of the most striking is that recorded by Darwin * with refer- ence to the effects of a European climate on the Ameri- can varieties of maize. Thus Metzger cultivated in Germany a tall kind of maize, Zea altissima, brought from the warmer parts of America, and, " During the first year the plants were twelve feet high, and a few seeds were perfected. . . In the second generation the plants were from nine to ten feet in height, and ripened their seed better. . . Some of the seeds had * " Animals and Plants," i. p. 340. ON VARIATIONS. 353 even become yellow, and in their now rounded form they approached common European maize. In the third generation nearly all resemblance to the original and very distinct American parent-form was lost. In the sixth generation this maize perfectly resembled a European variety." Other instances of the cumulative effects of condi- tions of life on plants have already been recorded in former chapters. Thus Lesage found that if Garden cress were treated with salted water, a much more marked effect was produced in the second year than in the first, the alteration effected in the tissues of the second generation seeming to be carried on from the point gained in the first. Bonnier found that seeds of Teucrium scorodonia gathered from plants growing at high altitudes, and sown in Paris, only produced plants showing nearly similar characters to the local forms after three years' exposure to the new environment. Among members of the Animal Kingdom the evi- dence is no less conclusive. Thus Darwin * records that " Dr. Bachman states that he has seen turkeys raised from the eggs of the wild species lose their metallic tints and become spotted with white in the third generation." Again, Mr. Hewitt, who often reared ducks from the eggs of the wild bird, and who was careful that there should be no crossing with do- mestic breeds, " found that he could not breed these wild ducks true for more than five or six generations, as they proved so much less beautiful. The white col- lar round the neck of the mallard became much broader and more irregular, and white feathers ap- , ii. p. 250. 354 ACTION OF NATURAL SELECTION peared in the ducklings' wings. They increased also in size of body; their legs became less fine, and they lost their elegant carriage. Fresh eggs were then pro- cured from wild birds, but the same result followed." Again, Darwin* records that " according to Pallas, and more recently according to Erman, the fat-tailed Jirghisian sheep, when bred for a few generations in Russia, degenerate, and the mass of fat dwindles away, the scanty and bitter herbage of the steppes seems so essential to their development." The fleece of sheep imported from Europe to the West Indies is much af- fected, and " after the third generation, the wool dis- appears from the whole body, except over the loins; and the animal then appears like a goat with a dirty door-mat on its back. A similar change is said to take place on the West Coast of Africa." Another in- stance of the effect of climate on sheep is recorded by Brewer.f Sheep taken from southeastern Ohio, a dis- trict noted for its excellent wool, and pastured on the alkaline soil of a certain portion of Texas, had the texture of their wool much altered, and its reaction to dyes showed obvious differences. Brewer states that " the change in the character of the wool begins imme- diately, but is more marked in the succeeding fleeces than in the first. It is also alleged that the harshness increases with succeeding generations, and that the flocks which have inhabited such regions several gener- ations produce naturally a harsher wool than did their ancestors, or do the newcomers." The deteriorating effect of an Indian climate on *Ibid.,i. p. 102. t Vide Cope's " Factors of Organic Evolution," p. 435. ON VARIATIONS. 355 dogs is well known. Hounds seem to degenerate most rapidly of all, whilst greyhounds and pointers also de- cline rapidly. Darwin * was informed by Dr. Falconer that bull-dogs " not only fall off after two or three generations in pluck and ferocity, but lose the under- hung character of their lower jaws; their muzzles be- come finer and their bodies lighter." He also men- tions a case of a pair of setters, born in India, which perfectly resembled their Scotch parents. Several litters were raised from them in Delhi, but none of the young dogs obtained resembled their parents in size or make, their nostrils being more contracted, their noses more pointed, their limbs more slender, and their size inferior. On the coast of Guinea, " dogs, according to Bosnian, alter strangely; their ears grow long and stiff like those of foxes, to which colour they also incline, so that in three or four years they degenerate into very ugly creatures ; and in three or four broods their bark- ing turns into a howl." Darwin considers this tendency to rapid deteriora- tion in European dogs may be largely attributed to re- version. It is of course possible that this may be the case, but it seems to me more probable that it is due to the direct and cumulative effects of changed conditions of life. The cumulative effect of conditions of life is ad- mitted, even ]py Weismann, in the case of the butterfly Polyommaius phlcuas. As already mentioned in Chap- ter VII., this occurs as a reddish gold variety in Ger- many and other countries of similar latitude, and as a much darker variety in more southerly countries, as *Ibid.,\. p. 39. 356 ACTION OF NATUEAL SELECTION Italy and Greece. Though these forms can be more or less transformed into each other, by suitable exposure of the pupae to warmth or cold, yet Weismann found that from German pupae he could never obtain butter- flies so dark as the darkest forms of the southern variety, whilst from Neapolitan pupae he could never get them so light as the ordinary German variety.* It seemed to him, therefore, " that the two varieties may have originated owing to a gradual cumulative in- fluence of the climate, the slight effects of one summer or winter having been transmitted and added to from generation to generation." Weismann explains this case of apparent transmission of acquired characters by supposing that the temperature slightly affects the de- terminants of the wing scales contained in the germ- plasm, as well as more markedly influencing the deter- minants of the rudimentary wings in the chrysalis. Moreover he suggests that " in many other animals and plants influences of temperature and environment may very possibly produce permanent hereditary variations in a similar manner." This suggestion of Weismann's contains in it, it seems to me, the germ of an idea which further obser- vation and experiment may prove to be of fundamental importance in evolution. The idea itself is no new one, and has probably occurred independently to many writers. As far as I am aware, it was first suggested by Galton,f when propounding the theory of heredity to which that of Weismann bears so striking a resem- *" Germ-Plasm," p. 399. fProc. Roy. Soc., xx. p. 394, 1872; Contemporary Review, December, 1875; Journ. Anthropol. Inst., 1875, p. 346. ON VARIATIONS. 357 blance. Thus he concluded that we are almost justified in reserving our belief that the body cells can react on the sexual elements, i. e., that acquired characters can be inherited; but he himself proposed to accept the supposition of their being faintly heritable. More recently, Cope* has embodied the idea in his " Theory of Diplogenesis." Thus he says, " Now, since these somatic cells develop the modifications which constitute evolution in their subsequent growth into organs, there is no reason why the reproductive cells which experi- enced similar influences should not develop similar characters, so soon as they also are prepared to grow into organs. . . The effects of use and disuse are two- fold, viz. : the effect on the soma, and the effect on the germ-plasm. . . The character must be potentially ac- quired by the germ-plasma, as well as actually by the soma." However, when Cope begins to briefly expand his theory, he seems to me to drift into improbable and unverifiable speculations. Thus he imagines that the transmission of external influences is primarily through the nervous system perhaps through the organisation of some peculiar mode of motion and secondarily through nutrition. In order to account for the numerous instances of the cumulative effects of changed conditions of life, it seems, therefore, that we may assume with consider- able probability and reason that the germ-plasm is di- rectly affected as well as the body tissues. These ap- parent instances of the inheritance of acquired char- acters are in reality, therefore, nothing of the kind, but are due to the germ-plasm reacting to change of en- * Amer. Naturalist, xxiii. p. 1058, 1889. 358 ACTION OF NATURAL SELECTION vironment simultaneously with the body tissues. As Weismann points out, a necessary corollary to this view is " the assumption of material determinants which exist in the germ-plasm and are passed on from one generation to another." If change of environment acts cumulatively on the fleece of the sheep, or the structural characters of a dog, it follows that it must in each of the first few generations act also on the " de- terminants " in the germ-plasm specifically represent- ing such specific characters. The effect produced on such determinants in the first generation must serve more or less as a starting point for the environment to work upon still further in the next generation, and so on. Through what agency is the environment enabled to act on the germ-plasm? To me the only conceivable one is a chemical influence, through products of metab- olism and specific internal secretions. We have seen in a previous chapter that the products of metabolism of an organism may exert a retarding effect on its own growth, and in some cases a stimulating effect on the growth of other organisms. Physiological re- search of the last few years has shown that most of the organs and tissues of the body have specific internal secretions, which, passing into the general cir- culation, may exert an influence of vital importance on the general metabolism of the organism. Thus extir- pation of the thyroid gland produces symptoms which in many animals end fatally, but which may be dimin- ished or suppressed by feeding on the gland substance, or injection of extracts of it. Extirpation of the suprarenal glands results in much more speedy death, ON VARIATIONS. 359 and here again the injection of extracts may delay the fatal issue. Extirpation of the pancreas causes the production of severe diabetes, and ultimately death, but such an effect may be avoided by the grafting of a portion of excised gland in the peritoneal cavity or the tissues. In such a case it cannot, of course, exercise its digestive function, but its internal secretion pre- vents the onset of the fatal diabetes. Again, extirpa- tion of the total kidney substance of a dog leads, not to a diminished secretion of urine, but to a largely in- creased secretion, accompanied by a rapid wasting away which soon ends fatally. Hence the kidneys may pos- sess an influence on the metabolism of the whole body, as well as their obvious secretory function. The spleen appears to have an internal secretion which is of influence in setting free the pancreatic ferment. Finally, extracts of various nervous tissues, brain, spinal cord, and sciatic nerve, have been found when intravenously injected to produce a distinct fall of blood pressure, whilst those of the pituitary body pro- duce a marked rise. Does it not seem distinctly probable, therefore, that every tissue in the body to some extent affects every other tissue? Each may have its own specific products of metabolism, and perhaps specific internal secre- tions, which, passing into the general circulation, may in turn stimulate or depress, or otherwise affect, every other tissue in the body. Whenever a changed environment acts upon the organism, therefore, it to some extent affects the normal excretions and secre- tions of some or all of the various tissues, and these react not only on the tissues themselves, but also to a 360 ACTION OF NATURAL SELECTION lesser degree upon the " determinants " representing them in the germ-plasm. It should be mentioned that the influence of somatic variations on the germ-plasm through the agency of various secretions has already been suggested by De- lage.* Though he does not admit Weismann's doctrine of determinants, he thinks that the ovum may contain specific substances of an identical nature to those con- tained in the cells of the principal classes of tissues, such as the nervous, muscular, and perhaps glandular. Conditions of life such as climate and food, which through the intermediation of the blood influence the constituents of certain of the body tissue cells, will therefore influence the same substance in the ovum, or produce hereditary variations. The hypothesis of specific secretions is of distinct help in accounting for certain apparent instances of the inherited effects of use and disuse. As we have seen in a former chapter, Darwin found that the rela- tive size of the brain of the domestic rabbit has con- siderably diminished. Possibly this may have been the result of more ample food, and of artificial selection of individuals with large bodies and small heads, and of panmixia (cessation of Natural Selection), but it seems almost more probable that it is due, at least in part, to the inherited effects of disuse. Thus a rabbit, when kept in captivity, would need to use its brain but little, and hence the excretions and secretions of the nervous tissues would be diminished. The " determi- nants " in the germ-plasm corresponding to these would be less stimulated than in wild rabbits, and * " Heredite," pp. 806 to 812, Paris, 1895. ON VARIATIONS. 361 hence in the next generation the development of the brain (and probably the other nervous tissues) would take place somewhat less vigorously, and the adult brain be in consequence somewhat diminished in size. In the next generation the diminution would be greater still, and so on. Again, we have seen that in man, for instance, the degree of (hereditary) pigmentation of the skin seems to vary closely with the intensity of the heat and light experienced. It is possible that the specific excretory products of the pigment deposited in the skin, as a direct response to the action of the environment, may stimulate the pigment " determinants " in the germ- plasm to increased vigour, so that in the next genera- tion the organism will tend to become slightly more pigmented than it had been in the previous one. Sup- posing, on the other hand, the pigment cells of the skin received no light rays whatsoever, as in animals which had wandered into a subterranean cave, their metab- olism would be reduced almost to nil, and so the pig- ment " determinants " in the germ-plasm would diminish in vigour, and the offspring of the animals would be (at birth) somewhat less pigmented than they had been in previous generations. It is obvious thajb on our specific secretion hypothesis only a certain class of acquired characters can be in any degree heritable; only those, in fact, of which the corresponding tissues possess a specific secretion or ex- cretion, capable of acting specifically on the " deter- minants " of such tissues in the germ-plasm. For in- stance, the blacksmith cannot transmit his brawny arm in any degree to his descendants, as it is scarcely pos- 362 ACTION OF NATURAL SELECTION sible that the arm muscles can have a secretion differ- ent from that of the other muscles of the body. The greater muscular development of the man as a whole, however, may lead to the production of slightly more muscular children than the average. On our hypothesis, the heritableness of mutilations and injuries is not admissible. It is almost inconceiv- able that each spot of skin on the body, or each finger, should have a specific secretion, and that an injury to it, by changing its secretion, should so affect the germ- plasm as to produce a similar change in the correspond- ing area of skin or the finger of the offspring. How, then, is it possible to account for the various apparent instances of inherited injuries, such as are quoted by Eimer,* Cope,f and others who believe in the transmis- sibility of such characters? There certainly seem to be a small number of thoroughly well authenticated cases, but the number is so small that we may perhaps attribute them to mere coincidence. The millions of instances of injuries which show no trace of any trans- mission provoke no remark, as it is only what we are led by common experience to expect. Supposing, on the other hand, a child exhibits any birth mark or de- formity bearing some similarity to an injury or mutila- tion in a parent, it is at once hailed as a remarkable case of inheritance of an acquired character. There are, however, certain cases of the apparent in- heritance of acquired characters which require more detailed criticism. These are the well-known experi- ments and observations of Brown-Sequard on injuries * " Organic Evolution/' p. 173. f " Factors of Organic Evolution," p. 431. ON VARIATIONS. 363 of the nervous system in guinea-pigs. As Brown- Sequard experimented over a period of thirty years on thousands of guinea-pigs, it might be thought that we could accept his results as absolutely conclusive. Yet a repetition of some of his experiments by Romanes and by Hill seems to show that they may be very largely erroneous. Thus, like Brown-Sequard, Ro- manes * found that some of the progeny of parents in which an injury to the restiform body had produced protrusion of the eyeball, showed a protrusion likewise, though this was less marked, and always affected both eyes; but it seemed that this might be an accidental occurrence, in that normal guinea-pigs are sometimes to a certain extent exophthalmic. Again, Romanes found that some of the progeny of animals in which hsematoma and dry gangrene of the ears had super- vened after injuring the restiform body, also became affected. However, the morbid state seemed to arise at any time in the life history of the individual, and the process not only affected a much less quantity of the ear, but also a different part of it. One therefore might imagine it to be due to mere coincidence, or to transmitted microbes; but Romanes does not think this can be the case, as, on the one hand, he has never seen the peculiar morbid process of the ears in other guinea- pigs, and, on the other hand, he was unable to inoculate the ears of healthy animals with matter from the ears of mutilated guinea-pigs. Romanes repeated Brown-Sequard's experiments on the section of the cervical sympathetic nerve, but he never observed in their progeny any change in the shape * " Darwin and after Darwin," vol. ii. p. 104 et seq. 364 ACTION OF NATURAL SELECTION of the ear or partial closure of the eyelids. Dr. Leonard Hill* has also repeated them with some thoroughness. The operation was performed on six guinea-pigs, and these animals were allowed to interbreed. It was again performed on twelve of their offspring, and these were also allowed to interbreed, but none of the young of either the first or the second generation showed any persistent droop of the eyelid. Hill found, however, that many of the young guinea-pigs exhibited a partial closure of the eye for some time after birth, but this phenomenon was due entirely to conjunctivitis, the re- sult of dirt getting into the eyes. It affected both eyes equally often, and when it terminated the droop disap- peared also. One is strongly tempted to conclude that the partial closure of the eyelids observed by Brown- Sequard was due to a similar cause, and was no more hereditary than in Hill's guinea-pigs. Certain of Brown-Sequard's experiments have, however, been cor- roborated by subsequent observers, and must therefore be accepted. Thus he found that animals which had been rendered epileptic by injury to the spinal cord, or section of the sciatic nerve, might transmit this epilepsy to their offspring. These results have been confirmed by Obersteiner,f and Westphal has even succeeded in producing epilepsy, which was transmitted to the off- spring, by striking guinea-pigs on the head with a ham- mer. It has been suggested by "Weismann that the transmission might be due to the introduction of some microbe into the operative wound, which both caused epilepsy in the parent, and, by invading the germ cells, * Proc. Zool. Soc. 1896, p. 785. f Oesterreiohiscbe medicinische Jahrbiicher, 1875, p. 179. ON VARIATIONS. 365 produced it in the offspring also. This could not have been the case in Westphal's experiments, however, as in them no wound at all was made. How, then, can this apparent transmission of ac- quired characters be accounted for? Our hypothesis of internal secretions supplies a very simple explana- tion. Thus the secretions from the brain of an epi- leptic guinea-pig, no matter how this epilepsy had been produced, would almost certainly be abnormal. Even supposing that they were without effect on the " de- terminants " of the nervous tissues in the germ-plasm, it is a very probable supposition that they might so affect the growth of the nervous tissues of the offspring, during intra-uterine development, as to provoke a similar abnormal condition in them. In mammals and other viviparous animals, it is prob- able that changed conditions of life produce part of their cumulative action during the period in which the embryo is under the influence of the maternal fluids. It is of course possible that all of the cumulative effect is then produced, though in such a case we should have to find some other explanation than that given above of the cumulative effects noticed in oviparous animals as Polyommatus phlceas. In the case of the gradual de- generation of the pure bred dog under an Indian climate, for instance, the environment may so act upon the maternal parent as to produce slight changes in the body tissues, and also to alter the character of the secretions and excretions. These, acting on the off- spring during their embryonic development when, as we have seen in a previous chapter, the tissues are extraordinarily sensitive to their environment 366 ACTION OF NATURAL SELECTION may produce more obvious degenerative changes, which will of course continue and be increased during extra-uterine growth. This second genera- tion of dogs, besides being modified in external characters, will therefore have the nature of their internal secretions more altered than had the first generation. These changes will react still further on their offspring during intra-uterine development, and BO on. Our conclusions as to the reaction of the germ-plasm to the external conditions of environment place a much higher value on somatic variations as a factor in Evolu- tion than that accepted by Weismann and his followers. It is for this reason that the effects of environment in the production of variations have been dealt with at such length in the preceding chapters. Every obvious effect produced in an organism by the direct action of the environment, may, in my opinion, be accompanied by a more or less corresponding, though much slighter, effect upon the determinants in the germ-plasm, and express itself in the next generation as an apparently cumulative effect of the changed environment. How often this possible influence on the germ-plasm actually shows itself, and what may be the numerical measure of its extent, can only be determined by long continued observation and experiment. As we shall see in the next chapter, somatic varia- tions may be of very great importance in evolution by reason of their adaptiveness to sudden changes of en- vironment; but, quite apart from any question of adap- tation, it is probable that they may be of value in affording Natural Selection a better chance of exert- ON VARIATIONS. 367 ing its influence. How this is so, is best explained by means of a diagram. XX XX XXX X xxxx XX X X xxxx XX X X xxxx X X X X X xxxxx xxxxx XX XX XX X X xxxx XX X X xxxx xxxx x xxxx xxxx xxxx XX xxxxx X 9 xxxxxxxx 8G42024.681D x x x x*xxxxxx xxxxxxxxxxxXxxx x xl xxxxxxxxx*xxxxxxxxxxx| x xxxxxxxxx x xxxxxxxxxxxxxxxxxxxxxxxl xxxxxxxxxxlxxxxxxx XXXXX X_X XXXXXXXXXXXXXXX X XX X ' x \99 16 14 12 10 6 6 .4 t 4 6 8 10 OS 16 FIG. 29. Effect of variable environment on variability of organisms. 368 ACTION OF NATURAL SELECTION The upper of the two accompanying figures represents a (roughly) normal curve of distribution of 324 meas- urements (represented by dots and crosses). As is in- dicated on the base line, these vary in size from the general mean by from 1 to 10 per cent. Sup- posing that the larger individuals (the crosses) were better adapted to their environment, and were being gradually selected by the agency of Natural Selection, whilst the shorter ones were gradually being weeded out, then it is probable that the selective process would act only very slowly, as the range of variation is so slight. Thus differences of 2 or 4 per cent, from the average are so small as to be almost inappreciable, whilst greater differences, as of 7 or 8 per cent., are exhibited by only a very small proportion of the whole (only 12 out of the 324 measurements being 7 per cent, greater than the average, and only 6 of them 8 per cent, greater). Now let us suppose that this group of 324 individuals is acted on by a variable environment, so that the slight range of blastogenic variations is en- hanced by the superposition of somatic modifications. Out of every ten organisms of any particular size, let one be increased by 2 per cent., another by 4 per cent., another by 6 per cent., another by 8 per cent., and an- other by 10 per cent., owing to the action of a favour- able environment, whilst the other five of the ten organisms are diminished by similar amounts, owing to the action of an unfavourable environment. The lower figure shows the new distribution of the organisms ac- cording to their altered magnitudes. For instance, of the 40 crosses representing individuals 1 per cent, larger than the average, 4 are increased by 2 per cent., ON VARIATIONS. 369 and so are placed in the -p3 per cent, column, and 4 are diminished by 2 per cent., and so are placed in the 1 per cent, column. Four more are placed in the + 5 per cent, column, and 4 in the 3 per cent, column, and so on. A similar process was ap- plied to all the other measurements, as far as possible, the 6 individuals 7 per cent, larger than the average being, for instance, increased and diminished by 4, 6, and 8 per cent. The curve of distribution of the 324 individuals now takes the form of the lower figure. "We see that it is much more flat-topped, indicating that the range of variation is much greater than before. This is, in fact, more than doubled, the arithmetical mean error being increased from 3.2 per cent, to 6.8 per cent. The individuals now vary in size by 17 per cent., so Natural Selection can act with much greater celerity and certainty than before. Thus no less than 32 of the 324 individuals are now 11 per cent, or more larger than the average, and so offer a very ap- preciable handle for Selection to work upon. The dis- tribution of the dots and crosses shows us, also, that all the extremely large individuals are also individuals which were larger than the normal before the variable environment was brought to bear on them. Many of the larger individuals were rendered smaller than the average by the action of an unfavourable environment, and many of the smaller rendered larger by a favour- able environment i. e., there has been a good deal of mixing of the individuals as originally distributed but the fact remains that the extremely large individuals, which Natural Selection would be especially likely to favour, and the extremely small ones, which it would be 370 ACTION OF NATURAL SELECTION especially likely to eliminate, are still those which were originally, as the result of blastogenic variation, re- spectively larger and smaller than the average. The selected individuals are therefore not only larger in themselves, but in that their " largeness " is to some extent a blastogenic variation, their offspring will also be, on an average, larger than the normal. It is not intended to imply that increased variability is by any means always an advantage. In a stable form, upon which Selection is acting but little, it might be a distinct disadvantage, as the more variable indi- viduals might be less adapted to their environment than the less variable. In the case of the sparrow, for instance, we saw that Bumpus found that the extreme individuals in either direction tended to be weeded out, though there was a much greater elimination of the ex- treme individuals in one direction than of those in the other. What is true for one form, however, is by no means necessarily true for another, and in a rapidly evolving organism, such as the pale-coloured mouse found on the sand-banks off Dublin, it is probable that the eliminated individuals would be chiefly confined to the darkest specimens, and include but few of the palest ones. CHAPTER XII. ADAPTIVE VAKIATIONS. Adaptability a fundamental property of protoplasm Instances of adaptive variation in plants Acclimatisation of Protozoa to high temperature, to poisons, to mechanical stimuli, to saline solutions Acclimatisation of fresh-water Mollusca to salt water, and of various marine animals to fresh water Acclimatisation of Mam- mals to vegetable poisons, and to toxins Sum total of somatic variations always in direction of adaptation Somatic varia- tions of importance in evolution, but they can effect little without Natural .Selection Germinal Selection. THE question of the definiteness or indefiniteness of variations has been frequently and hotly debated, but there has been a singular absence of exact definitions of the views actually held by the supporters of the rival theories. Had such definitions been forthcoming, I doubt if any fundamental differences of opinion would have been found to exist at all. Take, for instance, Darwin's definition of definite variations, viz. : " The effects of (conditions of life). . . may be considered as definite when all or nearly all the offspring of individuals, exposed to certain conditions during several generations, are modified in the same man- ner." * There is surely nothing in this definition which would not be generally admitted. As has been shown at some length in several of the preced- ing chapters, change in one or many of the conditions * " Origin of Species," p. 6. 371 372 ADAPTIVE VARIATIONS. of life may lead to very considerable changes in the form and structure of all or most of the organisms exposed, even in one generation. Hence, if Darwin's definition be accepted as it stands, we are compelled to admit that variations may be definite. Suppos- ing, however, it be taken to imply the cumulative, and so hereditary, action of conditions of life acting for several generations, then those who refuse to admit the validity of the instances of such cumulative action ad- duced in the last chapter might also refuse to admit the existence of definite variations. But assuming the former interpretation as the correct one, are we to agree with lienslow * that "in nature variations are always definite," or are we to follow Darwin in believ- ing that variations are, as a rule, indefinite, and only exceptionally definite ? Here, it seems to me, we are in want of more exact definitions. Probably Henslow would admit that the variations in the number of car- pels in the common daisy, or of veins in the leaf of the beech tree, or of stigmatic bands on the seed capsules of the poppy, are governed by the laws of chance, or if he did not, how could he account for the fact that the frequencies of distribution of the respective numbers are in accordance with the Law of Error? Clearly, in such cases the variations must be indefinite. Suppos- ing, however, that the distribution of the variations in the length of the leaves of a plant grown upon land occur according to the laws of chance, whilst that of the leaves of the same species of plant grown in water also follows these laws, but supposing also, that the average length of the aquatic plant leaves is considerably *" Original of Plant Structures," p. ix. ADAPTIVE VARIATIONS. 373 greater than that of the land plant leaves, then obvi- ously we should have here a case of both definite varia- tion and indefinite variation. The leaves of the aquatic plant would have varied in the direction of greater length, or would have varied definitely in adaptation to their new environment, but the distribution of their variations about their mean would still be in accord- ance with the laws of chance, or would be indefinite. The term " definite," as applied to variations, seems to be generally regarded as more or less synonymous with " adaptive." Thus Lloyd Morgan * defines defi- nite or determinate variations as " variations along special or particular lines of adaptation," while Hen- slow f says " Definite variations are always in the direc- tion of adaptation to the environment itself." Hence it seems to me that the discussion of the definiteness or indefiniteness of variations may, for practical purposes, be narrowed down to the following questions: (1) Have conditions of life an appreciable influence on organisms, and if so, (2) Is this influence in any case cumulative, i. e., partly inherited, and (3) How far are the effects produced adaptive? The first two questions I have already endeavoured to answer in the preceding chapters. The third we will now proceed briefly to inquire into. As far as the limited number of observations avail- able can show, adaptability would seem to be a funda- mental property of protoplasm. Whenever an organ- ism is exposed to changed conditions of life, then it is found that the original want of adaptation becomes gradually and progressively diminished with increase * " Habit and Instinct," p. 311. f Ibid., p. viii. 374 ADAPTIVE VARIATIONS. in the duration of the exposure. In most animals, the change in the direction of adaptation is slight, but it is probably always there, if only it be carefully looked for. In plants it is, as a rule, greater, and may be ob- vious to the most cursory observation. Instances of it have already been described at some length in pre- ceding chapters, and hence it is unnecessary to do more than briefly recall these here. We saw that Karsten found that a kidney bean reared in the dark for a month or two weighed 20 per cent, more than one reared in the light, yet, owing to the absence of the stimulus of light, its leaves did not weigh a fifth as much. Lothelier found that plants such as Berberis vulgaris bore non-spinescent leaves in a moist atmos- phere, but spines and spines alone in a perfectly dry one. Costantin found that he could change the form of Hippuris at will, by growing the aquatic form of the plant on land, and the terrestrial form in water. All the leaves produced under water were long, undulated, and delicate, whilst those in air were short, erect, and firm. Costantin, and also Godron, obtained very simi- lar results by growing other aquatic plants on land, and terrestrial ones in water, the change being always in the direction of adaptation to the new surroundings. Again, Lesage found that by watering various plants with water containing salt they developed characters similar to those exhibited by maritime plants, viz., in- creased thickness of leaves, larger and more numer- ous palisade cells, and diminution of. the intercellular spaces and of the chlorophyll. Bonnier found that plants of Teucrium scorodonia, when grown at a high situation in the Pyrenees, exhibited features character- ADAPTIVE VARIATIONS. 375 istic of alpine plants, viz. : very short aerial stems, with hairy and dark green leaves, and compact inflorescence. Seeds gathered from these plants and sown in Paris after three years produced elongated stems, with less hairy and brighter green leaves, or plants very similar to those from seeds obtained in the neighbourhood of Paris. In addition to changes of climate and soil, plants can adapt themselves also to mechanical stresses and strains. Thus Ray * sowed a mould (Sterigmatocystis) in two vessels, one of which was fixed, and the other subjected for two months to a rapid oscillatory move- ment. Instead of a thick feltwork of mycelium, this latter vessel contained small, perfectly spherical, elas- tic masses consisting of entangled filaments. The sup- , porting tissues of the plant were strengthened in re- sponse to the violent mechanical strains, the mem- [j- branes being twice or three times as thick, and the fila- ments having many more partition walls. Again, R. Hegler f found that " the hypocotyl of a seedling sun- flower, which would have been ruptured by a weight of 160 gm., bore a weight of 250 gm. after having been subjected for two days to a strain of a weight of 150 gm. The weight was subsequently increased to 400 gm. without injury. . . Leaf stalks of Helleborus niger, which broke with a weight of 400 gm., were able to resist one of 35 kgm. after having been subjected to a strain for about five days." Thus protoplasm has the power of responding to, and counteracting the action *C. R. Acad. 8ci. f cxxiii. p. 907. fBer. Verkandl. K. Sachs. Gesell. Wiss., v. p. 638, 1892 (quoted from Henslow's " Origin of Plant Structures," p. 204). 376 ADAPTIVE VARIATIONS. of, external mechanical forces by the formation of sup- portive tissues. It is by reason of this power that plants grow vertically upwards in opposition to the force of gravity. Some of the most remarkable instances of adaptation in plants are those relating to the interchange between roots and stems. In many cases, at least, it would seem that when a subterranean root becomes aerial, its characters tend to approach to those of a stem, whilst a normally aerial stem, grown underground, develops the characters of a root. For example, " An old acacia with a decaying trunk sent down an aerial root from the living part, about six feet from the ground. When it had been rooted in the soil for some time, it became detached by the wind; the root then became a ' stem,' the upper part putting out foliage." * Again, Dr. Lindley records that " a young willow tree had its crown bent down to the ground; this was covered with earth, and soon emitted an abundance of roots. The true roots were then removed from the soil, and the stem inverted. The roots now became branches and emitted an abundance of buds, and the tree ever after- wards grew upside down." Accompanying such changes of function are found corresponding changes of histological structure. Costantin t determined the effect of growing stems of brambles underground, and he found that the number and volume of the cortical cells increased, the collenchyma disappeared, the liber fibres diminished or disappeared, and starch could be formed and stored up in the parenchymatous tissues. * Quoted from Henslow, ibid., p. 179. f Bull, de la Soc. Bot. de Fr., p. 230, 1883. ADAPTIVE VAEIATIONS. 377 Moreover these modifications were uniform, affected all the tissues, and were rapidly produced, a week or two sufficing. There is no evidence, as far as I am aware, to show how widespread is this phenomenon of interchange be- tween roots and stems, and hence one cannot accept it as a generalised property of plants. In any case one must bear in mind that it may not, after all, be a case of direct adaptation to surroundings in the ordinary acceptation of the term, but may be the calling up, in response to one of two stimuli, of one of two groups of characters long since acquired by the plant protoplasm. A case of adaptation which appeared to be to some extent hereditary has recently been recorded by Errera.* Conidia of the mould Aspergillus niger were cultivated by Dr. Hunger for two generations in Rau- lin's nutritive solution, to which 6 per cent, of common salt had been added, and when placed in a similar salt solution they were found to produce spores in 3f days. Conidia which had been cultivated in the salt Raulin solution for only one generation took 4 days to produce spores, however, whilst those which had been culti- vated in Raulin solution containing no additional salt took 5 days. On the other hand, when some of the conidia cultivated under the three sets of conditions were placed in ordinary Raulin solution, those kept two generations in salt solution showed only slight sporifi- cation in 5 days, those kept one generation showed more marked sporification, whilst those kept through- out in ordinary Raulin solution spored in 4 days. Spores from these three last cultures in normal Raulin *Bull. Acad. Roy. Beligque, p. 81, 1899. 378 ADAPTIVE VAEIATIONS. solution were then sowed in a solution to which 18.4 per cent, of salt had been added. After 5 days the original normal Raulin culture showed no germination; that originally kept one generation in salt Raulin solu- tion showed slight germination, and that originally kept two generations distinct germination. Thus the adap- tation to a concentrated salt solution was not entirely lost even after rearing in a normal medium, or was in some degree inherited, especially in the case of the greater degree of adaptation produced by the growth of two generations in salt solution. Doubtless this " inheritance of acquired characters " was due to the salt solution influencing the germ cells at the same time as the body cells. The same explanation may be used to account for the somewhat similar results ob- tained by Ray * with Sterigmatocystis alba. Oonidia of this mould were sown in a solution of dextrose, the development taking place but slowly, owing to the want of adaptation to the new environment. On continuing the culture in the sugar solution, however, the rate of development gradually increased from generation to generation, till, finally, the sixth generation showed a more abundant development after 8 days than the first one had after 15 days. The morphological characters were progressively modified in addition, so that the mould came finally to resemble a penicillium. In the members of the Animal Kingdom, the power of adaptation is, as a rule, far less marked than in those of the Vegetable Kingdom, but probably it is present to a greater or less extent in all organisms, from the lowest to the highest. In certain Flagellata, for in- *Rev. Gen. de Bot., ix. p. 193, 1897. ADAPTIVE VARIATIONS. 379 stance, Dallinger * has demonstrated a most remarkable and extreme adaptability to high temperature. Start- ing at 15.6 C., he gradually raised the temperature of the water containing these monads up to 70.0 C., when the experiment was ended by an accident. The exact times in the course of the experiment at which a given temperature was reached are not mentioned, but from the description afforded they are gathered to be roughly those given in the accompanying table : Original temperature, After 4 months, 15.6 C. 21.1 A1 < fter 36 mo 39 nths, < 34.4" 38.9 6 22.8 < 41 i 41.7 1 8 23.3 48 68.3 ' 15 25.6 60 i 58.9 ' 23 26.7 1 61 61.1 1 32 33.9 (Several months more) 70.0 From this we see that the experiment, as far as it was carried, lasted about six years. The raising of the temperature was not by any means even, the organisms frequently reaching stages at which for months at a time an increase of temperature of half a degree or less was immediately followed by adverse effects, and in some instances by the death of many of the organisms. For instance, when a temperature of 25.6 had been reached, it was found that for a space of five months the temperature could not be raised even .3 of a degree without a distinctly evil effect being produced. In fact, it was found that the progress of acclimatisation at lower temperatures was, as a rule, much slower than at high ones. Thus, within a space of seven months, it * Journ. Roy. Microsc. Soc., vol. vii. p. 191, 1887. 380 ADAPTIVE VARIATIONS. was found possible to raise the temperature of the or- ganisms from 41.7 to 58.3. Also to raise the tem- perature from 61.1 to 70.0 took only a few months (number not stated). As to the absolute upper limit of temperature these infusoria can withstand, it is of course impossible to judge, but there seems no reason to suppose that it might not be considerably higher than that reached by Dallinger. A striking proof of the altered condition of the organisms was furnished by the fact that some of those acclimatised to 70.0, died off when placed in a suitable nutritive solution at 15.6. This acclimatisation was probably for the most part a direct adaptation of the protoplasm to its new en- vironment, but it must also have been in part due to natural selection. Dallinger noticed on more than one occasion that a good many of the organisms were killed off, and these would doubtless have been the less adaptable ones, the more adaptable surviving. Still, as far as one can judge from the brief account given, the temperature was often raised over considerable in- tervals without any such fatalities. Dallinger's results, in addition to their intrinsic value, are of great interest in that they enable us to account for the presence of various Protophyta, such as Oscillatorice and Nostocacece in hot springs. The tem- perature of many of these springs is considerably above 60.0 C., and that of the California geysers, in which Nostocacecz (possibly Protococcus) are found, reaches the remarkable temperature of 93. Certain metazoa, also, are stated to live at temperatures considerably above 45, or temperatures which prove fatal to their ADAPTIVE VARIATIONS. 381 allies. All such instances as these * are probably due to gradual acclimatisation, accompanied by a variable amount of selective destruction of the less adaptable organisms. Other observations on acclimatisation among Pro- tozoa have been made by Davenport and Neal.f The acclimatisation of Stentor cceruleus to weak corrosive sublimate solution was tested. Stentors kept for two days in .00005 per cent, solution were found, on immer- sion in .001 per cent, solution of sublimate, to be killed off after (on an average) 304 seconds' exposure. Sten- tors kept in pure water, on the other hand, were killed after only 83 seconds 7 immersion. Similar results were obtained in other experiments, it appearing that within certain limits the resistance period varied directly with the strength of the solution in which the protoplasm had been cultivated. If, however, the culture solution were too strong, (above .0001 per cent.), the organism became so weakened that it was less resistant to the killing solution than those reared in pure water. As no deaths occurred in the culture solutions, the adapta- tion must have been a direct one, and in no way de- pendent on natural selection. Stentor can also become acclimatised to mechanical stimuli, for Castle $ has observed a colony of Stentors in an aquarium being constantly struck by Tubifex moving backwards and forwards, and yet showing no contraction as they usually do when struck. *For a detailed account see a paper by Davenport and Castle, Arch, f . Entwick. d. Organismen, Bd. ii. p. 227, 1895. t Arch, f . Entwick. d. Organismen, Bd. ii. p. 564, 1896. \ Vide Davenport's " Experimental Morphology," p. 109. 382 ADAPTIVE VARIATIONS. Upon acclimatisation to saline solutions a consider- able number of observations has been made, especially in the case of Protozoa. As long ago as 1869 Czerny * experimented on amoebae, and found that by the very gradual addition of salt, he could acclimatise them to a 4 per cent, solution. With the unacclimatised organ- isms, the sudden addition of .33 per cent, of salt had in many cases a fatal effect, though some were able to stand even a 1 per cent, solution. None could resist a 2 per cent, solution, however. More recently Massart f has made quantitative determinations of the acclimati- sation of certain ciliated infusoria to solutions of potas- sium nitrate. Unacclimatised cysts of Vorticella nebu- lifera first began to show plasmolysis when the strength of the potassium nitrate solution in which they were placed amounted to 1.2 per cent. On the other hand, cysts previously kept 22 hours in a 1.8 per cent, solution did not show any plasmolysis until the concentration was raised to 2.5 per cent. Observations on Colpoda cucul- lus gave similar results. The degree of effect produced by a .8 per cent, solution in unacclimatised organisms required a 2.5 per cent, solution in organisms previ- ously kept 22 hours in 1.8 per cent, solution. The capacity for acclimatisation varies greatly in different organisms, for Bichter $ succeeded in acclimatising Tetraspora to 16 per cent, sodium chloride solution, whilst Spirogyra, similarly treated, was unable to resist even a .5 per cent, solution. The acclimatisation of certain of the metazoa to * Arch. f. mik. Anat., Bd. v. p. 158, 1869. t Arch, de Biol., ix. p. 515, 1899. t Flora, 1. p. 4, 1892. ADAPTIVE VARIATIONS. 383 changes of salinity appears to have been first studied by Beudant * more than eighty years ago. He placed a number of fresh water molluscs, such as Lymnaa, Planorbis, Physa, Ancylus, and Paludina in a vessel of water, and added a small quantity of salt every day. After a few months the water contained 4 per cent, of salt, and 170 of the original 400 molluscs were still sur- viving. Of another 400 kept under otherwise similar conditions in fresh water, 184 were surviving. All species are not equally adaptable, however, as Unio and Anodonta, though they throve well in fresh water, all died in salt. Beudant also performed the converse ex- periment of acclimatising marine molluscs to fresh water. He made observations on 38 different species of the genera Haliotis, Cerithium, Buccinum, Tellina, Venus, Ostrea, Pecten, and Mytilus. He added fresh water every day, so that after five months the animals came to live in absolutely fresh water. Out of the 38 species experimented with, 20 withstood the change perfectly well. The experiment was started with 610 individuals, and of these 375 survived. Of a similar number kept for the same length of time in normal sea water, 401 survived, or only 4.2 per cent. more. How- ever, all the other 18 species experimented with died during the course of the experiment. In still another series of experiments, Beudant succeeded in acclimatis- ing marine molluscs to a solution containing no less than 31 per cent, of salts. These consisted chiefly of sodium chloride, but contained also calcium and magnesium chlorides. Numerous observations on the acclimatisation of * Journal de Phys., Ixxxiii. p. 268, 1816. 384 ADAPTIVE VAKIATIONS. other organisms such as Myxomycetes, Actinos- pherium, Crustacea, and tadpoles have been made by other observers,* but it is unnecessary to mention more than a single experiment, one made by De Yarigny t upon a number of different species of marine animals. Some Carcinus mcenas, Pagurus Prideauxii, Dromia vulgaris, Anthea cereus, Sagartia parasitica, Portunus puber, Doris tuberculata, Venus, Actinia mesembryan- themum, and HolotJiuria tubulosa were placed in an aquarium supplied with a constant flow of water. This water was gradually diluted more and more with fresh water, with the following results : ft, . fc WC3 & |nH o W!H ANIMALS KILLED. H H ANIMALS KILLED. "1 H g | K > ft gc> 1 22.2 nil 22 68.7 3 6 11 33.3 44.4 nil 1 C. mcenas, 1 Pagurus 1 C. mcenas, 1 Dromia 25 29 77.8 1 Portunus, 3 Anthea 2 Doris and 2 Venus 13 55.6 and rest of Pagurus 32 35 88.9 1 Portunus, 1 Anthea 17 3 Sagartia, 2 Holo- thuria. On the 38th day, when the experiment was ended, there were still living all of the original eight Actinia mesembryanthemum and one Carcinus mcenas. De Yarigny suggests that the greater resistance of A. mesenibryantJiemum is probably connected with the fact that these organisms are attached to rocks near the sea *For literature vide Davenport's "Experimental Morphology," p. 86. tCentralb. f. Physiol., i. p. 566. ADAPTIVE VAEIATIONS. 385 surface, and so are frequently uncovered. They must therefore be exposed to sea-water diluted by rivers, and to rain water. The less resistant Anthea is found fur- ther below the surface, however, whereas the still less resistant Sagartia lives in water several metres deep. Observations on acclimatisation to saline solutions are, perhaps, less important and less interesting than those on acclimatisation to other conditions, in that, within certain limits, the phenomenon is probably a purely physical one, dependent on differences of osmo- sis, and the pressures and strains thereby set up. There is little doubt that if sufficient care and time be employed, any marine organism could be acclimatised to fresh water, and any fresh water form to salt water, or solutions of even greater density. If it be remem- bered that the osmotic pressure of a 1 per cent, solution of sodium chloride is over seven atmospheres, then it is obvious that the strain upon the tissues of an organism suddenly transferred from one solution to another of considerably greater or less salinity may easily be suffi- cient to rupture and kill them. Direct observations on the acclimatisation of the vertebrata are extremely few, except in the case of cer- tain mammals experimented on in connection with serum therapeutics. Davenport and Castle * have made some interesting observations on the acclimatisa- tion of tadpoles to heat, however. Recently laid eggs of Bufo lentiginosus were divided into two lots, one of which was allowed to develop in a warm oven at a tem- perature of 24 to 25, and the other kept at 15. * Loc. cit. 386 ADAPTIVE VARIATIONS. After four weeks, the temperature of heat rigor was determined by gradually heating the water containing the tadpoles. Whilst all the tadpoles kept at 15 went into heat rigor at or below 41, those reared at 25 did not in any case die at a temperature below 43, the average increase of resistance amounting to 3.2. This adaptation to higher temperature gradually disappears on returning the tadpoles to water at ordinary tempera- atures, more than half of the 3.2 increase being lost after keeping them for 17 days at 15. Probably the capacity for acclimatisation is present to a greater or less degree in every organism. In some observations carried out at Naples,* I found that the death temperatures of a Medusa (Rhizostoma), a salp (Salpa africana\ and of Amphioxus were, on an aver- age, respectively 1.3, .6, and 1.5 higher in August than they had been in April, when of course the tem- perature of the sea was several degrees lower. The adaptability of the highest organisms to changes of environment does not afford so much support to our thesis viz., that adaptability is a fundamental prop- erty of protoplasm as does that of the lowest organ- isms, because the adaptation is, as a rule, indirect and complex. Still the intrinsic interest of the subject is so great as to warrant a brief reference to it. Almost all of the exact observations deal with acclimatisation to chemical agents, especially the toxins secreted by bacteria. Upon mice Ehrlich * has made some very exact observations on adaptation to a vegetable poison, ricin. The mice were fed on food cakes soaked in * J. Physiol., xxv. p. 181, 1899. t Deutsche med. Wockenschr., 1891, p. 976. ADAPTIVE VARIATIONS. 387 solutions of the poison of increasing strengths, and after feeding for various lengths of time, the maximum amount of poison the animals could withstand was de- termined. This amount rapidly increased after the first day, so that after three weeks' feeding it was found to be no less than 200 to 800 times the original dose. Some of these mice were then kept on normal food for over six months, and at the end of that time could still withstand considerably more than fifty times the origi- nal amount of poison. Even more remarkable results have been obtained in the preparation of diphtheria antitoxin. For this pur- pose, Roux * uses the filtrate from diphtheria bacillus cultures, it being at first mixed with an iodine solution to reduce its virulency. One-quarter cc. of the iodised toxin is injected on the first day, and this is increased to 1 cc. on the 13th day. On the 17th day \ cc. of the pure toxin is injected, and this is gradually in- creased in amount till on the 41st day 10 cc. is in- jected, and on the 80th day no less than 250 cc. The virulency of the last dose must have been some 5000 to 10,000 times greater than that of the first dose, and, supposing the effect produced on the horse was more or less the same after each injection, its acclimati- sation to the toxin must have increased in similar pro- portion. As is well known, animals can be acclimatised to toxins produced by other bacteria, such as those of anthrax, tetanus, cholera, typhoid, plague, and likewise also to snake venom; but it is unnecessary to refer to these here. Upon acclimatisation in man there are probably no exact observations, but the inexact and un- * Vide Crookshank's " Text-Book of Bacteriology," 1896, p. 58. 388 ADAPTIVE VARIATIONS. scientific are matters of common personal experience. A hot day following suddenly on a long spell of cold weather, or a cold one on a long spell of hot weather, is felt much more keenly than days of considerably higher or lower temperature which are led up to by the gradual change of the seasons. Likewise also, weather which appears very hot to one's self will be looked upon as temperate by a native Indian, or even an Anglo-Indian. Acclimatisation is often experienced by those who indulge in excessive amounts of alcohol, opium, or tobacco. For instance, De Quincey was at one time in the habit of taking 8000 drops of laudanum daily, this enormous quantity probably producing no greater effect than a dose of 30 to 50 drops in an ordi- nary man. Again, arsenic eaters are able to swallow as much as A gm. without injury, or about four times the ordinary lethal dose. These various observations made upon members of all classes of the Animal and Vegetable Kingdoms will, I believe, be held sufficient proof of the contention that adaptability is present in all organisms, and is therefore a fundamental property of protoplasm. Whether every variation produced by change of en- vironment is in the direction of adaptation to the change, it is of course impossible to say; but probably this is not the case, as, by reason of the close correla- tion existing between many of the characters of an organism, the change may produce a want of adapta- tion in some of them, but an increased adaptation in others. Supposing, however, it were possible to esti- mate the change produced in every character in the body, it seems to me almost certain that the sum total ADAPTIVE VARIATIONS. 389 of all the changes would be rather in the direction of adaptation to the new surroundings, than in that of non-adaptation. It is not to be supposed for a moment that every one of a group of organisms exposed to new conditions of life will become better adapted to them than any one of the group had been originally; but merely that the characters of the group will, on an average, become better adapted than they had been be- fore. Doubtless many instances can be thought of in which the effect produced by a change of environment has no appearance of being in the least adaptive, but this may be due to our ignorance of what constitutes an adaptation. For instance, it may be asked in what way a starved animal is better adapted to semi-starva- tion than a well-nourished one? It is, of course, less adapted in that it has, stored up in its body, less food material such as fat and glycogen on which it can live, but it is obviously better adapted in that its metab- olism is considerably smaller than that of a well- nourished animal; i. e., it actually lives on considerably less food. Again, it may be asked in what way a dusky coloured Polyommatus phlceas is better adapted to a warm climate than a copper-coloured one, and vice versa with reference to a cold climate ? Possibly there is nothing adaptive about the colour of the wing scales, but doubtless it would be found that, on an average, the dusky butterflies could withstand a greater degree of heat than the coppery ones, and the coppery ones a greater degree of cold. Hence the change would, on the whole, be in the direction of adaptation. It is probable that somatic variations, by reason of their adaptation to changed surroundings, are of very 390 ADAPTIVE VARIATIONS. great importance in the evolution of more adaptive forms; in some cases, perhaps, of greater importance than genetic variations. Supposing, for instance, a number of organisms are more or less suddenly exposed to a considerable change of environment, whereby the majority of them are killed off. The survivors will be those which had the greatest power of adaptation to the new surroundings, and though the somatic varia- tions will not be, as such, inherited, yet the survivors will be, on the whole, those organisms which originally possessed the largest proportion of the particular char- acters which have appeared as adaptive somatic varia- tions. That is to say, adaptive somatic variations are, on an average, a magnified image of similar, but much more minute genetic variations, and hence the average hereditary characters of the survivors are in the direc- tion of adaptation. Again, the survivors will be those individuals possessing the largest degree of innate adaptability to the particular environment in question. Hence their offspring will also possess this adaptability, and in that they will have been exposed to the changed environment throughout the whole period of develop- ment, they will show much more marked somatic varia- tions than those shown by their parents. Finally, if it be admitted that the effects of conditions of life may be in some degree cumulative, then the adaptation of the second generation to the environment will be from this cause still further increased. Views somewhat similar to these as to the importance of somatic varia- tions have been set forth, at considerable length and with admirable lucidity, by Professor Lloyd Morgan in his work on " Habit and Instinct " (p. 316), and to this ADAPTIVE VARIATIONS. 391 the reader who desires more detailed discussion is re- ferred. Admitting that somatic variations are, on the whole, adaptive, and admitting also to a very limited extent the cumulative influence of changed conditions of life, are we to agree with Henslow * that the close adapta- tion of plants to their environment is due entirely to the responsive power of protoplasm to the external en- vironmental forces, and that it is absolutely unneces- sary to call in the aid of Natural Selection? By no means. Adaptive variation may be responsible for a good deal of the adaptation observed in plants, and for a very small part of that observed in animals, but prob- ably in each case by far the larger portion must be as- cribed to the ever present and ever acting agency of Natural Selection. For instance, Henslow argues very plausibly that inasmuch as certain plants when kept in a dry atmosphere develop spines and other characters similar to those possessed by desert plants, it is valid to conclude that these desert plants owe their peculiar characters to the direct action of the dry hot climate, and to that alone. Supposing this explanation to be correct, however, we ought, as Wallace points out,f to find plants with spines and the other characteristics of desert plants abounding in all dry countries, but very rare or wanting in moist and fertile districts. But this is by no means the case. Wallace states that many of the peculiarities of desert plants are present in the flora of the Brazilian Campos, and in that of the Galapagos and the Sandwich Islands, but very few of *Ibid., pp. 14 and 32. f Nat. Science, vol. v. p. 177, 1894. 392 ADAPTIVE VAEIATIONS. the plants indeed show any spines. Again spiny plants are exceedingly rare in the Canaries, though much of the surface, owing to long periods of drought, presents the conditions which elsewhere are supposed to produce spines. Though not prepared to deny that, other conditions equal, aridity may favour and humidity check the growth of spines, yet Wallace considers that a more important condition lies in the presence or ab- sence of herbivorous mammals, against whose ravages the spines afford protection. Thus he mentions sev- eral countries which are not particularly arid, but in which spiny plants, and also these destructive mam- mals, both abound. The development of the spines is chiefly dependent, therefore, on the action of Natural Selection, and is not a direct adaptation. In other cases also Wallace believes that the " direct action of the environment can have produced only a very small portion of the modifications and adaptations that actu- ally exist. In by far the larger number of cases no such explanation is possible, and no other adequate ex- planation has been suggested except variation and Natural Selection." Though it seems to me that Wallace, by excluding all other agencies, is inclined somewhat to exaggerate the importance of Natural Selection, yet his explana- tion of the evolution of adaptive forms seems much more rational, and in much better agreement with facts, than that given by Henslow. The view to which the present state of our knowledge seems to me to af- ford best support is one which lies more or less between these two extreme explanations. It is most con- veniently indicated by a diagram. Let us consider, ADAPTIVE VARIATIONS. for instance, the evolution of a typical aquatic plant from a typical terrestrial one. Supposing it were pos- sible to estimate the extent to which characters useful to aquatic life were present in a group of terrestrial plants, and supposing we were to plot out the fre- quency of their distribution, then this might take the form of the curve given in the extreme left of the upper portion of the accompanying diagram. Here Distribution of characters in plants. alter exposure for after i typical one. generation to for many terrestial plant: *quoous environment: generations. aquatic pU plant. Distributior of eha -acters animals. 10 20 80 90 1.08 80 40 50 60 7.0 Percentage of aquatic characters. FIG. 30. Evolution of the aquatic plant and the aquatic animal. we see that the most frequently occurring plant had 10 per cent, of " aquatic " characters, the extremes ranging from to 20 per cent. Supposing now this group of plants were exposed for one generation to an aqueous environment. It would be found at the end of that time that the proportion of aquatic characters had considerably increased, say to 26 per cent., but the fre- quency of distribution of the characters about the mean would still be symmetrical as it was before, the extremes now varying from 14 to 38 per cent. Some of the plants, therefore, would still possess fewer aquatic 394 ADAPTIVE VARIATIONS. characters than were possessed by a small number of the original group of plants, in accordance with Dar- win's dictum concerning plants that " whether the station (they inhabited) was unusually dry or humid, variations adapting them in a slight degree for directly opposite habits would occasionally arise." It will be noticed that in the diagram the curve of distribution of the characters is made slightly more flat topped than the other curves, indicating that the variability of a group of plants suddenly exposed to a changed environment is increased. Supposing that this group of plants is exposed to the aqueous environment for a number of generations, then, through the cumulative action of con- ditions of life, the adaptation will become considerably increased, and the plants will now show, on an average, say 40 per cent, of the aquatic characters of a typical aquatic plant. This increase of adaptation from the stage reached after one generation is supposed to be more or less permanent and hereditary, or would still be present if the plants were returned to their original dry land environment. But, however many genera- tions the plants be kept in their watery surroundings, it is supposed that they will never become adapted to it like typical aquatic plants. In order to evolve such plants, Natural Selection must be present in addition, and in this case the distribution of the plants, in respect of aquatic characters, will ultimately arrive at that in- dicated in the curve on the extreme right of the diagram. The lower half of the diagram is meant to represent the evolution of an aquatic animal, such as a mammal, from a land animal. Such an animal would in the first ADAPTIVE VARIATIONS. 395 place have very few characters adapting it to an aquatic existence, and so the curve of distribution of such char- acters will be a more steeply sloped one than that for plants. Also the direct effect of environment in the direction of adaptation will be very much less than in the case of plants, even after exposure for a large num- ber of generations. In fact, to effect any real and con- siderable change it will be essential to call in the aid of Natural Selection, and this, by acting constantly for a very large number of generations^ will gradually evolve a typical aquatic mammal such as the seal, dol- phin, or whale. In spite of all that has been written to account for the almost universally present adaptation which we see in animate nature, there is still a lingering doubt in the minds of many men as to the entire adequacy of the explanations hitherto offered. It is a feeling such as this which prompted Weismann to formulate an ad- ditional principle in explanation of adaptation, and of other phenomena, as the degeneration of disused organs, viz., his theory of Germinal Selection.* This theory supposes that, similar to the struggle for existence ex- perienced by individual organisms, so there is a strug- gle among the determinants of the germ-plasm of each single individual to obtain as great a supply of nutri- ment as possible, and so flourish at the expense of weaker determinants. Supposing, for instance, that parts of the body, such as the hinder extremities of the quadruped ancestors of our common whales, are ren- dered useless. As selection ceases, individuals with *"Ueber Germinal Selection," Jena, 1896 (English translation, Chicago, 1896). 396 ADAPTIVE VAEIATIONS. small hind legs, represented (Weismann supposes) by weaker determinants in the germ, are as favourably placed in the struggle for existence as those with large hind legs, represented by stronger germ determinants. The weaker determinants, in their struggle with the other determinants which represent useful organs in the body, will be worsted, and gradually become more and more enfeebled, the hind legs which they represent be- coming correspondingly smaller and smaller till they finally disappear altogether. Supposing, on the other hand, that the individuals showing a greater develop- ment of any particular characters than the average are for this reason favoured by Selection, then the determi- nants representing these characters in the germ-plasm will also be more powerful than the average, and by ab- sorbing more nutriment will become still more robust, and produce descendants exhibiting the characters in an increased degree. That is to say, the descendants will, by Germinal Selection, become more and more adapted to the conditions in respect of which they were originally favoured by Natural Selection. This theory, though plausible enough, is absolutely opposed to fact in so far as it relates to the evolution of more adaptive forms. As we have seen in Chapter IV., so far from the individuals selected in respect of any character tending to transmit that character in in- creased strength to their descendants, they almost in- variably transmit less of it, or the offspring show, on an average, a greater or less degree of regression towards mediocrity, according to the amount of the character present in their more remote ancestors. The degeneration of disused organs is, it must be ad- ADAPTIVE VARIATIONS. 397 mitted, a difficulty which has never been hitherto ade- quately accounted for, and hence, in lieu of something better, Weismann's hypothesis may in this respect be provisionally accepted. Still it is always to be remem- bered that it is no more than an hypothesis, which has not, and never can have, any experimental evidence to support it. AUTHOR'S INDEX. A Ainsworth, 330 Allen, J. A., 4, 7, 325, 326, 327, 328 B Bachman, 353 Baldwin, M., 223 Bateson, W., 37, 41, 51, 52, 53, 54, 55, 57, 59, 64, 91, 92, 137, 160, 272, 273, 275, 277, 278 Baxter, 93 Beddard, 317 Beddoe, 85, 93 Beeton, 347, 348 Bennett, 69, 163 Bertillon, 118 Beudant, 383 Blankinship, 41, 42, 44 Bonnier, 312, 313, 353, 374 Boscher, E., 260 Bosnian, 355 Bowditch, 202, 206 Bramley-Moore, L., 87 Brandes, G., 295 Bravais, 74 Breman, G. A., 314 Brewer, 354 Brewster, 24, 28 Britton, 60 Broca, 71 Brooks, W. K., 114 Brown, 60 Browne, E. T., 30 Brown-Sequard, 362, 363, 364 Brucke, 255, 256 Bullard, 25 Bumpus, 27, 206, 212, 214, 215, 216, 341, 343, 345, 370 Burnes, 330 Castle, 381, 385 Claus, 274 Clayton, 247 Cockerell, T. A. D.,315 Coldstream, 254 Cooke, 232, 287 Cope, 325, 354, 357, 362 Correns, 155, 157, 159, 160 Costa, 314 Costantin, 265, 266, 267, 374, 376 Cox, R, 290 Crookshank, 387 Cunningham, J. T., 251, 340, 341 Cuvier, 294 Czerny, 382 Dallinger, 379, 380 Dareste, 173, 174 Darwin, C., 2, 53, 58, 59, 65, 66, 68, 69, 70, 85, 96, 97, 114, 139, 140, 141, 148, 161, 166, 172, 185, 186, 187, 197, 219, 221, 222, 281, 293, 311, 317, 329, 330, 331, 332, 333, 335, 336, 352, 353, 354, 355, 360, 371, 372, 394 Davenport, C. B., 25, 36, 41, 42, 44, 90, 213, 245, 246, 247, 381, 384, 385 Delage, Y., 295, 360 Detmer, 286 Dixey, 167, 230, 238, 239 Dorfmeister, 233, 240 Driesch, H., 336 Duchartre, 264 Duncker, G., 14, 26, 27, 28, 34, 36, 83, 84, 215 400 AUTHOR'S INDEX. E Edmondston, 295 Ehrlich, 386 Eigenmann, 253, 324 Eimer, 233, 234, 239, 240, 250, 268, 269, 287, 288, 314, 362 Elliott, S., 248 Engleheart, 163 Erman, 354 Errera, 377 Ewart, J. C., 113, 141, 148, 152, 168, 169, 176 Falconer, 330, 355 Faxon, 254 Filon, 84 Finn, F., 176 Fischel, 207 Fischer, 232, 238, 239 Flahault, 312 Fletcher, W. H. B., 92 Focke, 161, 163, 166 Haacke, 143 Heape, W., 119, 120, 121 Hefferan, 35 Hegler, R., 375 Heincke, F., 322, 323, 324 Hellriegel, 262 Hennig, 201 Hensen, V., 115, 118 Henslow, G., 264, 265, 266, 372, 376, 391, 392 Herbert, 161, 163 Herbst, C., 55 Heron, R., 64, 148 Hertwig, O., 227 Hewitt, 353 Higginbottom, 225 Hill, L., 363, 364 His, 201 Holmgren, 295 Humphreys, 58 Hunger, 377 Hunter, J., 295 Hurst, 165, 166 G Gain, 263 Galton, F., 7, 13, 17, 19, 20 22 52, 65, 74, 75, 76, 84, 116, 118,' 122, 123, 124, 125, 126, 127 128, 129, 130, 132, 133, 134, 149, 151, 152, 170, 171, 172, 208, 356 Garstang, 25, 319, 321, 322 Gartner, 69, 159 Gauss, 10 Geddes, P.,104, 287 Giard, 40, 41 Gibbons, 278 Gilbert, H., 282, 284 Godron, 267, 374 Goebel, K. , 246 Goss, H., 290 Gregson, 289 v. Guiata, 142 Gulick, 315 GUnther, R. T., 274 Jameson, H. L., 350, 351 Jordan, 70 Jourdain, S., 254 Karsten, 247, 374 Kerner, 164, 165 Knight, A., 281 Knowlton, 225, 227 Koch, G., 268, 288 Kohlbriigge, 221 Kolreuter, 159, 161, 162 Koppen, 228 Kraatz, 54 Kriechbaumer, 54 Krocker, 294 Kropotkin, P., 253 Lawes, J., 282, 284 Lecoq, 161 Lee, 82, 87 AUTHOR'S INDEX. 401 Lesage, 269, 270, 353, 374 Leydig, 268, 314 Lillie, 225 Lindley, 376 Linton, 164 List, 252 Lister, Lord, 255, 256 Loeb, J., 278 Lothelier, 264, 374 Ludwig, 15, 46, 48 M MacCulloch, 254 MacLeod, 285 MacMunn, 251 Massart, 382 Mayer, A. G., 90 Meehan, 311 Meldola, 260, 291 Mendel, G., 155, 156, 157, 158, 159, 160, 169 Menetries, 295 Merrifield, 234, 235, 239, 240, 241, 242, 243, 268 Metzger, 352 Milardet, 162 Millais, E., 123, 176 Milne-Edwards, 54 Minot, 24, 201, 203, 204, 205, 208 Mobius, 231 Montgomery, 219, 220 Moreau, J., 116 Morgan, L., 223, 373, 390 Moulton, F., 67 N Nathusius, 294 Neal, 381 Newman, 289, 290 Nicoll, 257 Noll, 245 O Obersteiner, 364 Packard, 254, 274 Pallas, 354 Pearson, K, 26, 30, 32, 33, 39, 80. 82, 84, 87, 88, 89, 90, 124, 125, 127, 128, 133, 135, 136, 137, 150, 151, 153, 176, 180, 181, 182, 185, 205, 210, 211, 346, 347, 348 Petersen, 27 Pfitzner, 221 Planta, A. von, 287 Pledge, J. H., 16, 30 Pouchet, 256 Poulton, E. B., 243, 244, 251, 255, 257, 258, 259, 260, 261, 291 Preyer, 201 Q Quetelet, 12, 13 Quincey, De, 388 R Ray, 375, 378 Richter, 382 Ridgway, 219 Rimpau, 166 Rolfe, 163, 164 Romanes, G. J., 66, 67, 70, 120, 121, 363 Ross, J., 243 Roux, 387 3 Sachs, 245, 246 Salisbury, Lord, 336 Sargeant, 271 Sauermann, 293 Saunders, E. A., 64, 160 Schimper, 245 Schmankewitsch, 271, 272, 273, 274, 275 Schubeler, 312, 313 Schwalbe, 221, 222 Scott, 70 Sedgwick, A., 184, 210, 211 Semper, K., 255, 256, 302, 303, 304, 305 Smith, M., 70 Sorauer, 249 402 AUTHOR'S INDEX. Stahl, 248 Standfuss, 167, 171, 172, 175, 230, 236, 238, 240, 268 Steinert, 92 Strasburger, 245 Struthers, 58, 172 Tawell, J. A., 290 Tegetmier, 85 Thompson, H., 14, 82,337 Thomson, A., 104,287 Tschermak, 155, 159 Varigny, H. de, 267, 294, 302, 303, 304, 305, 306, 307, 384 Verschaeffelt, 28 Vines, S. H., 228 Vire, 252 Vochting, H.,15 Vries, H. de, 15, 29, 33, 50, 59, 60, 61, 62, 63, 114, 155, 159, 162, 171, 199, 200, 228, 284 W Wagner, M., 317 Wallace, A. R, 7, 66, 67, 259, 293, 316, 317, 335, 336, 391, 392 Wallace, 149 Walsh, 96 Warren, E., 8, 14, 26, 34, 81, 83, 84, 97, 178, 179, 180, 308, 309 Weisbach, 24 Weismann, A., 101, 102, 103, 114, 115, 116, 118, 134, 142, 169, 177, 178, 180, 197, 223, 233, 235, 236, 238, 239, 240, 241, 335, 355, 356, 358, 360, 364, 366, 395, 396, 397 Welch, F. H., 244 Weldon, W. F. R., 13, 21, 23, 26, 39, 40, 77, 79, 80, 81, 82, 97, 160, 206, 285, 286, 318, 337, 338, 339, 340, 341, 345, 346 Westphal, 364, 365 Whitfleld, 307 Wichura, 164 Wiesner, 246 Wilson, J. H., 163 Windle, B., 174 Wittich, 255, 256 Wood, T. W., 258 Yule, G. U., 84, 348 Yung, 249, 271, 292, 306, 307 SUBJECT INDEX. Aberrations, among Lepidoptera, 237 Abnormalities, spontaneous ori- gin of, 57, 59 Abyssal, light, theory of, 254; fauna, 254 Acacia, interchange of root and stem in, 376 Acceleration of growth, persist- ence of, 204, 207 Acclimatisation, to high tem- perature, 379; to saline solu- tions, 382; in tadpoles, 385; in mice, 386; in horse, 387; in man, 387 Acerina cernua, measurements of, 14 Acquired Characters, in lowland and highland plants, 811; in spiderwort, 313 Acquired Characters, heritable- ness of, 352; in maize, 352; in cress, 353; in ducks, 353; in sheep, 354; in dogs, 355; in butterfly, 356; in guinea-pigs, 363; in mould, 377, 378 Adaptability, a property of pro- toplasm, 373, 387; innate, 390 Adaptation, want of, as a cause of variability, 217, 218; sea- sonal, in Lepidoptera, 234, 241 ; in mammals, 243; in aquatic and terrestrial plants, 266; in maritime plants, 270; of di- gestive organs to food, 294, 295; of plants to changed climate, 311, 312; and definite varia- tions, 373 Adaptive variations, 371; in plants, 374; in moulds, 375, 377, 378; in sunflower, 375; in acacia, 376; in willow, 376; in brambles, 376; in Flagellata, 379; in Stentor, 381; in other Protozoa, 382; in Mollusca, 383; in Actinia, 384; in tad- poles, 385; in mice, 386; in horse, 387; in man, 387; con- cern average variations, 389; do not render natural selection unnecessary, 391; explained by germinal selection, 395 Aerial and aquatic plants, 265 Aggressive resemblance, 255 Ainos, correlation in skeletons of, 82 Alpine plants, compared with lowland, 311 Alpine climate, effect of, on plants, 311 Alternative heritage, 132, 156 Ammonia, effect of, on larvae, 301 AmcEbse, acclimatisation of, to saline solutions, 382 Amphibious plants, 265 Amphidasys betularia, variation in, 91 Amphimixis, 103, 114 Analogous Variation, 96 Ancestors, impress individuality on sex-cells, 134 Ancestral Heredity, Law of, 122 Ancon sheep, origin of, 58, 172 Animals, hybrids among, 167, 168 Anthropometric data, analyses of, 122 403 404 SUBJECT INDEX. Anthropometric measurements, obey Laws of Chance, 12, 13; variability in, 24 Aphides, measurements upon, 179 Aquatic plants, 265; evolution of, 393 Aral Sea, cockle present in, 275 Arctic climate, effect of, on mam- mals, 243; effect of, on plants, 313 Arithmetical mean error, deter- mination of, 23 Artemia, the effect of salinity on, 271, 274 Asexual reproduction, in an ostracod, 177; in daphnia, 178; in aphis, 179; in plants, 183; in potato, 184; in apple, 185 Assertive mating, increases cor- relation, 151 Asymmetrical curves of distribu- tion, 29; obtained by expand- ing binomials, 31 Asymmetrical series, fitted with calculated curves, 30; types of, 33 Asymmetry, index of, 33 Atavism, 139 Aurelia aurita, variations in, 30 B Babies, variability of, 205 Basset hounds, records of, 122 Bees, effect of nutrition on, 287 Bertillon's system, useless for identical twins, 118 Binomial, expansion of, 31 Binomial curve, approximation of, to probability curve, 12; types of, 32 Biophors, 102 Birds, effects of foods on, 293; geographical races of, 325 JttscuteUa Icevigata, discontinu- ous variation in, 64 Bisexual descent, controlled by Law of Heredity, 134 Blastogenic variations, 101, 242, 244 Blended heritage, 132 Brain, diminished by disuse, 332 Branchipus, in relation to Ar- temia, 274 Brothers, relation between, 128, 135; hereditary resemblance be- tween, 151; correlation be- tween, 347 Bud- Variation, in nectarine, 185; in Chrysanthemum, 186; in moss-rose, 186; causes of, 186; probably, discontinuous varia- tion, 187 Buttercup, variation in petals of, 29 Canidse, variation in, 327 Carcinus mosnas, asymmetry in measurements of, 39; correla- tion in, 80; natural selection in, 337 Cardium edule, effect of salinity on, 275 Castration, effect of, on animals, 95 Caterpillars, protective resem- blance in, 260 Cave animals, 251, 253, 254 Centroid vertical, defined, 17 Cervical sympathetic nerve, in- herited effects of section of, 363 Characters, new, produced by crossing, 161 ; acquired, herit- ableness of, 352 et seq.; in guinea-pigs, 363 Chlorophyll, colouring matter of larvae, 291 Chrysalides, resemblance of, to surroundings, 258 Chrysanthemum leucanthemum, variation in, 46 Chrysanthemum tsegetum, varia- tion in, 15; origin of many- rayed form of, 63 Civilisation, effect of, on correla- tion, 82 Climate, producing local races, 311, 327; effect of, on hairy covering of animals, 329 SUBJECT INDEX. 405 Clover, variation in blossoms of, 29; five-leaved, origin of, 62; five-leaved, effect of soil on, 284 Coat colour, inheritance of, 133 Cockle, effect of salinity on, 275 Coefficient of Regression, 127, 128, 132; relation of, to corre- lation, 128 Coefficient of variation, 26; is this a correct measure of varia- bility? 27 Cold, effect of, on Lepidoptera, 233 Collateral inheritance, 125, 135; relation of, to lineal, 128 Colour of skin, changes in, pro- duced by light, 251, 256, 257 Colouration, of flounder, effect of light on, 251; of pupae, 259; of caterpillars, 260 Columba lima, reversion to char- acters of, 140, 141 Compositse, variation in, 47 Conception, condition of mother at time of, as regards offspring, 209 Conditions of life, may produce local races, 311; effect of, on Lepidoptera, 316; cumulative action of, in maize, 352; in cress, 353; in ducks, 353; in sheep, 354; in dogs, 355; in butterfly, 356 Correlated variation, in Sciurus carolinensis, 6; in Linaria spuria, 16 Correlation, determination of, 74; improved method, 84; capri- ciousness of, 86; between ances- tors, 125; relation of, to regres- sion, 128; in man, 75, 83, 150, 347, 348; in shrimps, 79; in crabs, 80, 81; in prawn, 82; in Vegetable Kingdom, 135; in daphnia, 178; in aphis, 180; in Ficaria, 185; relation of, to va- riability, 210; homotypic, 211 Correlation constant, 74 Correlation, negative, 73, 79 Correlation values, diagram of, 78 Corrosive sublimate, acclimatisa- tion to, 381 Cowslip, hybrids of, 164 Crabs, measurements of, 13, 14; natural selection in, 337 Critical period of reaction, in Lepidoptera, 240 Crops, growth of, affected by manures, 282 Crosses, between Echinoderms, 110; fertility of, 165; among Lepidoptera, 172 Crossing of type and varietal forms, 92 Crowfoot, effect of nutrition on, 284 Crustacea, effect of light on, 253, 254; effect of salinity on; 271, 275 Cultivation, effect of, on spider- wort, 313 Cumulative action of conditions of life, in maize, 352; in cress, 353; in ducks, 353; in sheep, 354; in dogs, 355; in butterfly, 356; in general, 394 Cumulative action of products of metabolism, 308 Curve of Error, normal, 18; values of, compared with act- ual measurements, 22 Curves of plant variation, 43, 44, 45, 49, 50 Curves, symmetrical and asym- metrical, represented by gen- eralized expression, 32 Cypris reptans, parthenogenesis in, 177 Daffodil, crosses of, 163 Daphnia magna, measurements upon, 178; reaction of, to en- vironment, 179; effect of prod- ucts of metabolism on, 308 Darkness, effect of, on plant growth, 245; on molluscs, 252; on Crustacea, 253; on fish, 254 406 SUBJECT INDEX. Daylight, effect of, on plants, 246; on flounder, 251 Death, ages at, in man, 347 Death rate, selective, in crabs, 338; in sparrows, 341 Definite variations, 371 Degeneration, explained by germinal selection, 396 Desert plants, 263 Destruction, selective, in crabs, 337; in sparrows, 341; in mice, 351 Determinants, 102; affected by temperature, 236; necessity of existence of, 358; and germinal selection, 395 Determinate variations, 373 Development, of Echinoderm larvae, effect of temperature on, 190, 194; rate of, in man, 201; diminution of variability with, 204, 206; stages of, in relation to reaction to environment, 190 Digits, variation in, 56 Dimorphism, in earwig, 38; in crab, 39; in fish, 42, in marsh plant, 42; origin of, 64, 65.; seasonal, in Lepidoptera, 233, 239, 241, 242 Diplogenesis, theory of, 357 Discontinuous variations, origin of, 51 ; importance of, in evolu- tion, 52; as regard bud-varia- tions, 187 Disease, affects blonds more than brunets, 93 Diseases, affect identical twins similarly, 116, 117, 118 Disuse, effect of, on skull, 332; on pigeon bones, 333; on duck bones, 333; inheritance of ef- fects of, 360 Dogs, effect of Indian climate on, 355 Domestication, effect of, on vari- ability, 221 ; on rabbit, 331 Dominant characters in hy- brids, 156, 159 Dryness of soil, effect of, on plants, 263 Duck, effect of domestication on, 333 E Earwig, variation in forceps of, 37 Echinoids, crosses among, 168; specific metabolism of, 296, 298 Echinoid larvae, effect of tem- perature on size of, 229 Elimination, of sparrows in storm, 341; of gray mice, 350 Embryos, rate of growth of hu- man, 201; variability of, 206 Environment, diminishing ef- fect of, with development, 195; variable reaction of organisms to, 197; relation of, to develop- ing plants, 199; effect of, de- pendent on rate of growth, 203; permanent effect of, 207; un- favourable, as a cause of vari- ability, 218; saline, as a direct cause of variations, 277; may produce local races, 310; action of, through internal secretions, 358; may affect determinants of germ-plasm, 366; does little without natural selection, 392 Environment, effect of, on growth, 184, 194; on variabil- ity of sparrow, 214; on varia- bility of periwinkle, 216; on variability, 218, 220; on repro- ductive system, 221 ; on soma, 223; on colour of chrysalides, 258; on Lepidoptera, 316; on plants, 311; on oyster, 314; on snail, 314; on Porto Santo rabbit, 331 Epilepsy, inheritance of acquired, 364 Equable Variability, Law of, 96 Equipotency of parental herit- ages, 150 Error of Mean Square, 26 Evolution, importance of varia- tion in, 336; of aquatic plant, 393 SUBJECT INDEX. 407 Exclusive inheritance, 132; obeys Law of Heredity, 133 Excreta, effect of, on growth, of larvae, 298; of molluscs, 302, 307; of tadpoles, 306; of Daph- nia, 308 Exophthalmos, apparent inherit- ance of, 263 Expectation of life, 347 Exposure, effect of, on plant growth, 248 Eye-colour, transmission of, 132 Eyes, relation of, to change of skin colour, 255 F Fertilisation, premature, in rab- bits, 113; effect of tempera- ture at time of, 191 Fertilisations, artificial, method of effecting, 105 Fertility, correlation of, with structure, 87, 88, 89; inherit- ance of, in man and in horse, 87; maximum, of type forms, 90; of hybrids, 165; in man, 348 Fibonacci, series of, 48 Finger-print method, as regards twins, 118 Finger-prints, 170 Fin-rays, variability of, in cer- tain fishes, 27 Fit of curves, estimation of, 35 Flagellata, adaptation of, to high temperature, 379 Floral structures, relation of, to arid surroundings, 265 Flounder, effect of light on, 251 Food, as a cause of variability, 281; effect of, on bees, 287; on aphides, 288; on Lepidoptera, 288-290; on larvae, 291; on tad- poles, 292; on birds, 293, 295; on mammals, 294; ouSaturnia, 317 Forficula auricularia, variation in, 37 Fowls, production of malforma- tions in, 174; effect of meat on, 295; effect of domestication on, 333 Fox, variation in, 328 Fraternal correlation, 128, 135; in aphis and daphnia, 180 Frog, development of, in rela- tion to temperature, 225, 226, 227 Frogs, effects of feeding on, 292 G Gallus bankiva, reversion to characters of, 142 Galton's function, 74 Gangrene, apparent inheritance of, 363 Genetic Selection, 81 Genetic variations, 102 Germ, infection of, 175 Germinal Selection, 395 Germinal variations, 102 Germ-plasm, existence of definite units in, 135; influenced by somatic variations, 360 Geographical races, variability of species with, 220; of mack- erel, 319; of herring, 322; of birds, 325; of mammals, 327 Grades, method of, 20, 22 Grandparents, contribution of, to characters of offspring, 122 Growth, rate of, in man, 201; in guinea-pig, 203; relative rate of, 203; retardation and accel- eration of, 204; loss of varia- bility with, 204; curves of, 208; in offspring, affected by mother, 209; effect of salinity on, 278; in relation to prod- ucts of metabolism, 296, 298, 302, 303, 307, 308 Growth of organisms, reaction of, to environment, 189 Growth power, progressive loss of, 203 Gull, effect of food on, 295 Hsematoma, apparent inheritance of, 363 Hair, effect of cold on, 244 408 SUBJECT INDEX. Hare, effect of cold on, 244; vari- tion in, 329 Heredity, decided at time of fertilisation, 116, 119; Law of Ancestral, 122; solution of problem of, 125; a phase of homotyposis, 137; intensity of, in relation to variability, 211 Heritage, alternative, 156 Heritages, unequal from parents, 150; absolute amounts of, in offspring, 153 Herring, local races of, 322 Homoeosis, in Saw-fly, Humble bee, Crustacea, 54 Homologous organs, correlation of, 72 Homotypes, 136 Homotyposis, 136; relation of, to variability, 211 Horses, trotting, prepotency in, 149 Humidity, effect of, on plant growth, 262; producing local races, 327 Hybridisation, Mendel's Law of, 155; proof of, 156; scope of, 159, 160 Hybrids, 154; among Echino- derms, 110; variation of struc- ture of, with season, 111; reversion of, 143; dominant and recessive characters of, 156; resolved into pure pa- rental forms, 157; useless for breeding, 161 ; give rise to new characters, 161 ; relation of, to parent forms, 161; occurrence of, in nature, 162, 165; arti- ficial, 165; fertility of, 165; bigeneric, 166; binordinal, 166; among Lepidoptera, 167; among Echinoids, 168; with zebras, 168; effect of condi- tion of sex-cells on characters of, 169 Id, defined, 103 Identical human twins, 115; measurements upon, 117 Impregnation, effect of temper- ature at time of, 191 ; effect of salinity, 192 Indefinite variations, 371 Index of variability, 20 India, Lepidoptera of, 316 Indices of variability, relations between, 28 Infants, variability of, 205; curves of growth of, 208 Infection of Germ, 175 Infusoria, adaptation of, to high temperature, 379 Inheritance, direct and collateral, determination of, 125; blended and alternative, 132; in man, 347; of Acquired Characters, 352; of acquired epilepsy, 364; of adaptation, in mould, 377 Injuries, non-inheritance of, 362; to nervous system of guinea- pigs, 363 Internal secretions, and action of environment, 358 Iron salts, effect of, on plant growth, 286 Isolation, index of, 42 Isolation, physiological, in ani- mals and in plants, 67; origin of, 68 Japanese waltzing mice, 142 Japanned peacocks, 172 Jews, prepotency of, 148 Larvae, of Lepidoptera, effect of food on, 291 Larvae of sea-urchin, method of obtaining, 105; influence of sex-cells on size of, 107; in- fluence of season on size of, 109; reaction of, to environ- ment, 190; metabolism of, 297 Latency of characters, 156, 159 Latitude, effect of, on plants, 311 Law, of Ancestral Heredity, 122; of Frequency of Error, 10; of Hybridisation, Mendel's, 155; of Reversion, 133 SUBJECT INDEX. 409 Laws of Chance, dependence of variability upon, 11 Laws of Variation, 189 Leaf surface, action of light on, 248 Lemming, effect of cold on, 243 Lepidoptera, crosses among, 167; sports among, 171, 172, 175; effect of temperature on size of, 230; reversion in, 234, 239; seasonal dimorphism in, 233, 239, 241, 242; seasonal adapta- tion in, 234; seasonal forms of , produced by temperature, 236; variable protective resemblance in, 258; effect of moisture on, 268; effect of food on, 288 Leuciscus balteatus, distribution of fin rays in, 42; variation in, 324 Life, expectation of, 347 Light, effect of, on plants, 245; on tadpoles, 249; on snails, 249; on pigmentation, 249; on Proteus, 251 ; on flounder, 251 ; on molluscs, 252; on Crustacea, 253; on fish, 254; on frog, 255 Limncea, effect of products of metabolism on, 302, 303, 307 Linaria spuria, variation in, 15 Linaria vulgaris, origin of muta- tion in, 61 Local race of mice, 350 Local races, the result of environ- ment, 310; in trees, 311; in Alpine plants, 311 ; in shrimps, 318; in mackerel, 319; in herring, 322; in fish, 324; in birds, 325; in mammals, 327; in rabbit, 330 Longevity, 347; and fertility, 348 Lowland plants, compared with highland, 311 M Mackerel, local races in, 319 Maize, cumulative action of con- ditions of life on , 352 Malformations, artificial produc- tion of, 173 Man, rate of growth of embry- onic, 201; variability of, 205; permanent effect of environ- ment on growth of, 208; effect of light on skin of, 250 Manures, effect of, on growth of crops, 282 Mare, Lord Morton's, 175 Maritime plants, 269 Maturity of sex-cells, variation in, 110, 112 Maximum temperature, of plant growth, 228 Mean Squares, method of, 25 Measurements, deviations of, accord with Law of Error, 12 Mechanical strains, adaptation to, 375 Median, defined, 17 Mediocrity, regression towards, 126, 129; of children, compared with parents, 129 Melanism in Mollusca, 232 Mendel's Law of Hybridisation, 155 Meristic variations, 53 Metabolism, effects of products of, 296; specific, 296, 298, 309; of molluscs, 302, 307; of tad- poles, 306; of DapJmia, 308; products of, and action of environment, 358 Mice, waltzing, reversion in, 142; acclimatisation in, 386 Mid-parent, 127, 131 Migration, effect of, on vari- ability, 219 Mode, defined, 33 Modifications, somatic, 223, 242, 244 Moisture, effect of, on barley, 262; on water-reed, 264; on vetch, 265; on Mare's tail, 266; on water crowfoot, 267; on Lep- idoptera, 268; on molluscs, 269 Molluscs, acclimatisation of, to saline solutions, 383 Molluscs, effect on, of tempera- ture, 231; of light, 252; of lime, 287; of products of me- tabolism, 302, 303, 307 410 SUBJECT INDEX. Mongrels, reversion of, 143 Monsters, artificial production of, 173 Moths, effect of food plants on, 289 Moulds, adaptation in, 375, 377, 378 Mouse, evolution in, 350 Mutation, DeVries' theory of, 59 Mutations, produced by keeping seed, 114 Mutilations, non-inheritance of, 362 Mytilus, effect of light on, 252 N Naquada race, correlation in, 83 Natural Selection, action of on variations, 335; validity of, 336; in crab, 337; in sparrow, 341; in mollusc, 345; in man, 347; in mouse, 350; and so- matic variations, 367; necessary in spite of adaptation, 391; produces spines in plants, 391 Nature, hybrids occurring in, 162, 165 Negative correlation, 73, 79 Nereis limbata, variation in, 35 Nervous system, effect of, on skin colour, 256, 257, 259 Nigella Hispanica, variation in seed capsules of, 89 Nitrates, effect of, on larvae, 301 Nutrition of plants, effect of, on development, 200 Nutrition, effect of, on germ- plasm, 102, 104, 112; on crops, 282; on five-leaved clover, 284; on Crowfoot, 284; on Ficaria, 285; on molluscs, 287; on bees, 287; on aphides, 288; on Lepi- doptera, 288-290; on larvae, 291; on tadpoles, 292; on birds, 293, 295; on mammals, 294; on spiderwort, 313; on Satur- nia, 317 O (Enothera Lamarckiana, muta- tions in, 60, 61; effect of nutri- tion on, 200 Optimum temperature of growth, 227, 228, 230 Orchids, hybrids among, 166 Organic stability, position of, 170 Organisms, reaction of growth of, to surroundings, 189; vari- able reaction of, to environ- ment, 197; variability of grow- ing, 204; variability of, in relation to environment, 218; variable reaction of, to temper- ature, 230 Otter, variability of, 221 Ova, rabbits', transplantation of, 119; reaction of impregnated, to temperature, 191, 194 Oxalis, variation in, 118 Palcemonetes, variation in rostral teeth of, 27 Parallel variation, 96 Parasites, influence of, on crab and on earwig, 40 Parents, contribution of, to char- acters of offspring, 122; rela- tion of to offspring, 126 Parthenogenesis, in an ostracod, 177; indaphnia, 178; in aphis, 179; variability in, 180; affected by nutrition, 288 Peacocks, japanned, 172 Pedigree stock, in relation to Law of Heredity, 134 Peloric flowers, distribution of, in Linaria spuria, 15 Peloric race, origin of, in Linaria vulgaris, 62 Penycuik experiments, 140, 148, 168, 176 Periodic selection in mollusc, 345 Periwinkle, variability of, 206, 214 Phylogenetic forms, produced in Lepidoptera, 237 SUBJECT INDEX. 411 Physiological characters, subject to variation, 71 Physiological Selection, 66 Pieris napi, effect of temperature on, 239, 241 Pigeons, correlation in, 85; re- version in, 140; effect of meat on, 295; effect of domestication on, 333 Pigment cells in skin, of Proteus, 251; of flounder, 251; of frog, 256; of Octopus, 257 Pigments in larvae, derived from food, 291 Pigmentation, effect of light on, 249, 251, 252; in man, inherit- ance of, 361 Pisum sativum, varieties of, 155; cross-fertilisation of, 155, 159 Plants, variations in, do not agree with Law of Error, 15; de- termination of variations in, 46; hybrids among, 161; diminishing effect of environ- ment on developing, 199; desert, 263; aerial and aquatic, 265; adaptation of, to changed environment, 267 Plant structures, origin of, 265 Pleuronectes flems, measurements of, 14; variability in fin-rays of, 27, 34; correlation in, 84; effect of lighten, 251 Plymouth Sound, changing con- ditions in, 338 Polygon of Variation, 17 Polyommatus phlceas, reaction of, to temperature, 234, 241 Poppies, variation in seed cap- sules of, 89 Porto Santo rabbit, 330 Portunus depurator, variation of, 9 Position of organic stability, as regards sports, 170 Potato, asexual propagation of, 184 Prawns, measurements of, 14 Pregnancy, condition of mother during, 209 Prepotency, in dog, 148; in bull, 148; in horse, 148, 149; in man, 150; in Basset hounds, 151; in relation to sex, 151; produced by inbreeding, 152; conforms to Law of 'Heredity, 154 Primrose, variation in, 48; hy- brids of, 164 Primula officinalis, variation in, 48 Probability curve, 10 Probability, Laws of, govern variations, 12 Probable Error, 17; determina- tion of, 19 Products of metabolism, effect of, on growth, 296, 298, 302, 303, 307, 308 Protective resemblance, 255, 260 Proteus anguineus, effect of light on, 251 Protophyta, adaptation of, to high temperature, 380 Pseudoclytia pentata, variation in, 90 Pupae, effect of temperature on, 240; effect of lighten, 257 Pupal stage, in relation to tem- perature, 240; in relation to light, 257 R Rabbits, effect of nutrition of sex-cells in, 113; transplanta- tion of ova in, 119; local race of, 330 Race, variability of, compared with that of species, 182; local, of mice, 350 Races, local, of Chrysanthemum segetum, 50; constant correla- tion in, 79, 80; of Lepidoptera, 237; the result of environment, 310; in trees, 311; in Alpine plants, 311; in shrimps, 318; in mackerel, 319; in herring, 322; in fish, 324; in birds, 325; in mammals, 327; in rabbit, 330 Ranunculus repens, variation in, 16,30 412 SUBJECT INDEX. Reaction of organisms to environ- ment, variable, 197 Recessive characters in hybrids, 156, 159 Reciprocal crosses, 112 Regression, in sweet-peas, 126; in man, 127; coefficient of, 127, 128, 132; towards medi- ocrity, 129; weeded out by selection, 154; in Daphnia magna, 178 Relative Probable Error, an index of variability, 20, 21 Reproductive Selection, 87; in sparrow, 214 Reproductive system, correlation with, 86; effect of, on internal secretions, 94; effect of en- vironment on, 221 Resemblance, protective and aggressive, 255 Retardation of growth, persist- ence of, 204, 207, 208 Reversion, Law of, 133; in man, 139; in dog, 139; in calf, 140; in pigeons, 140; in fowls, 142; in mice, 142; produced by crossing, 143; attempted ex- planation of, 144; in Lepidop- tera, 234, 239 Ricin, acclimatisation to, 386 Ripeness of sex-cells, variation in, 110, 113 Rock-pigeon, reversion to charac- ters of, 140, 141 Roots, interchange of, with stems, 376 Rothampstead experiments, 282 S Saline solutions, acclimatisation to, 382, 383, 384 Salinity, effects of, on develop- ment, 192, 199; on plants, 269; on garden cress, 270; on frogs, 271; on worms, 271; on Arte- mia, 271, 274; on common cockle, 275; on Littorina, 278; on Tubularia, 278; on sea- urchin larvae, 279 Sandwich Islands, snails of, 315 Sciurus carolinensis, variation of, 4 Season, effect of, on size of larvae, 109; on structure of hybrids, 111 Seasonal, variation in size of larvae, 109; dimorphism in Lepidoptera, 233, 239, 241, 242; forms, among Lepidop- tera, produced by temperature, 236, 241, 242; adaptation, 234, 241, 243 Sea-urchin larvae, relation of growth of, to salinity, 279 Sea-urchins, specific metabolism of, 296, 298 Secretions, internal, affect repro- ductive system, 94; and en- vironment, 358 Seed capsules, effect of nutrition on size of, 200 Selection, effect of, on variability, 182; in sparrow, 213; effect of, on breeding true to type, 153; artificial, effect of, on vari- ability, 211, 222; Genetic, 87; Germinal, 395; Periodic, in mollusc, 345; Physiological, 66 Selection, Natural, effect of, on variability, 211; action of, on variations, 335; validity of, 336; in crab, 337; in sparrow, 341; in mollusc, 345; in man, 347; in mouse, 350 Sex, effect of, on correlation, 82, 83; and prepotency, 151; function of, in evolution, 180, 185 Sex-cells, effect of nutrition on, 104; staleness of, 105 Sheep, effect of feeding on, 294 ; effect of conditions of life on, 354 Shells, variability of, 206, 214; of -cockle, affected by salinity, 275 Shore plants, 269 Shrimps, measurements of local races of, 13 SUBJECT INDEX. 413 Sinapis alba, effect of light on growth of, 245 Skew curves of distribution, 29, 34 Skull capacity, of rabbit, 332 Snail, upon, relative growth of, 302, 303, 307; variation in, 314, 315; periodic selection in, 345 Soma, effect of environment on, 223 Somatic modifications, 223, 242, 244 Somatic variations, influence of, on germ-plasm, 360; impor- tance of, 366, 390; adaptiveness of, 389 Somatogenic variations, 101, 223, 242, 244; laws governing, 189 Sparrow, variability of English and American, 212; natural selection in, 341 Species, and varieties, differ- entiation of, 41; a precise criterion of, 44; origin of, by mutation, 59; produced by hybridisation, 164 Specific Gravity of waters, effect of, on Artemia, 272; on Tubu- laria, 278 Specific metabolism, in animals, 296, 298, 309 Sphcerechinus granularis, dimin- ished fertility between varie- ties of, 71 Spines, development of, in plants, 263; produced by natural selection, 391 Sports, arising in CEnothera Lamarckiana, 61; nature of, 170; stability of, 170; among Lepidoptera, 171; prepotency of, 171; among sheep, 172; among peacocks, 172; obey Law of Heredity, 172; pro- duction of, 173; artificial pro- duction of, in Lepidoptera, 175; among Lepidoptera, 237 Squirrels, variation in, 329 Stability, organic, position of 170 Staleness of sex-cells, effect of, on development, 105, 107 Standard Deviation, defined, 26 Statistical methods, 36 Stature, of parents and children, 130; of school Children, 202; human, variability of, 205 Stems, interchange of, with roots, 376 Sterility, between varieties, 69; between races of men, 70; of atypical forms, 89, 90; pro- duced by changed conditions of life, 94 Sterna hirundo, variation of, 7 Sternum, diminished by disuse, 333 Strongylocentrotus lividus larvae, measurements of, 14; figures of, 105; reaction of, to temper- ature, 193, 194; effect of en- vironment, on variability of, 217 Substantive variations, 53 Subterranean caves, animals in, 254 Sunlight, effect of, on growth of plants, 246 Surroundings, effect of, on vari- ability, 218, 220; effect of, on growth, 189; adaptation to changed, 390 Swine, variation in Mullerian glands of, 25 Tadpoles, development of, in relation to temperature, 225; effects of feeding on, 292; acclimatisation of, to tempera- ture, 385 Telegony, in mare, 175; negative results regarding, 176; statisti- cal examination of, 176 Temperature, effect of, on growth of Echinoderm larvae, 190, 194, 196; variable effect of, on growth, 196; diminishing effect of, with development, 197; fatal, on developing 414 SUBJECT INDEX. larvae, 198; effect of, in pro- ducing variations, 224; effect of, on development of frog, 225, 226, 227; optimum, 227; maximum, 228; effect of, on plant growth, 228; on size of Echinoid larvae, 229; on Lepi- doptera, 230, 233, 239, 242; on Mollusca, 231; on lemming, 243; on hare, 244; adaptation to high, in Flagellata, 379 Temperatures, death, of develop- ing larvae, 198 Torilis anthriscus, variation in, 15,46 Toxin, diphtheria, acclimatisa- tion to, 387 Transplantation of rabbits' ova, 119 Trotting horses, prepotency in, 149 Twins, identical, 115; measure- ments upon, 117 Type forms, fertility of, 90; replacement of, by variety, 91 Typha, measurements upon, 42 Umbelliferae, variation in, 48 Undifferentiated like organs, 135 Unpigmented cave animals, 254 Urea, effect of, on larvae, 300 Uric acid, effect of, on larvae, 300 Urmi, Lake, Artemia present in, 274 Use and disuse, inheritance of effects of, 360 Vanessa cardui, production of sports of, 175; effect of tem- perature on, 237 Vanessa io, production of sports of, 175; effect of temperature on, 237, 238 Vanessa levana, seasonal di- morphism in, 233 Vanessa prorsa, seasonal di- morphism in, 233 Vanessa urtic, effect of temper- ature on, 237; protective re- semblance in, 258 Variable protective resemblance, 255, 260 Variability, diagrammatically represented, 5; method of estimation of, 17; Law of Equable, 96; in parthenogene- tic forms, 177, 180; in poppies, 181, 182; individual, in relation to racial, 181; racial and spe- cific, 182; small, of a sexually reproduced forms, 184, 185; relations of, to external condi- tions, 199; of growing guinea- pigs, 204; diminution of, with growth, 204; of growing man, 205; of crabs, 206; of peri- winkle, 206, 215; of embryos, 207; diminution of , with evolu- tion, 210; diminution of, with selection, 210; of offspring and parents, 210; relation of, to homotyposis, 211; relation of, to intensity of heredity, 211; effect on, of artificial selection, 211; effect on, of Natural Se- lection, 211; effect of fertility of type forms on, 212; of Eng- lish and American sparrow, 212; of environment and of or- ganisms subjected to it, 219; in relation to Domestication, 221; of otter, 221 ; produced by food , 281; produced by season, 285; of sparrow, 342; increased by somatic variations, 368 Variants, defined, 4 Variation, defined, 1; numerical treatment of, 3; universally present, 6; obeys Laws of Chance, 11; importance of numerical treatment of, 98; hereditary cause of, 102; due to amphimixis, 103, 114, 126; of organisms, chiefly blasto- genic, 121 ; produced by cross- ing, 161; Laws of, 189; in periwinkle, 215, in birds, 219 SUBJECT INDEX. 415 Variations, swamped by inter- crossing, 57, 58; relation of, to variable environment, 197, 199; and modifications, 223; adaptive, 371; definite and indefinite, 371 Varieties, sterility between, 69 Variety, replacing type form, 91 Ventricosity, variability of, in periwinkle, 206, 215 Vetch, effect of water on, 265 Volume of water, effect of, on growth of snail, 302; of tad- pole, 306; of Daphnia, 308 W Waltzing mice, reversion in, 142 Warmth, effect of, on Lepidop- tera, 233 Weight, human, variability of, 205 Willow, interchange of root and stem in, 376 Wolf, variation in, 328 Woman, variability of, 205 Zebras, crosses with, 168 THB END. HISTORY, POLITICAL SCIENCE, SOCIOLOGY, LAW, Etc. REFERENCE- AND TEXT- BOOKS. PUBLISHED BY HENRY HOLT & COMPANY, 29 N ^ & RK ST - f marked * are chiefly for reference or supplementary use, and may in Henry Holt &> CoSs List of Books for School Libraries. Books irked are more fully described in the publishers' Descriptive Educa- Books be found not so marked are more fully described in the publishers' Descripti tional Catalogue^ and may also be had at tkeir Chicago branch, 378 Wabash A -venue. Either of these Catalogues or the publishers' 1 Catalogue of Works in General Literature sent free. Postage on NET books^ %% additional. PRICE Adams' The Science of Finance. By Prof. H. C. Adams of Univ. of Mich. American Science Series, xiv+syi pp. 8vo net $3 5 *Bazin n s The Italians of To-day. 247 pp. 12010 125 Holies 1 Elements of Commercial Law. With very full summaries and index. 344pp. 12010 net i oo Burke : Speeches on arrival at Bristol, at Conclusion of the Poll, Letters to the Marquis of Rockingham. to the Sheriffs of Bristol, and to a Noble Lord, and other Selections. Edited by Prof. Bliss Perry of Princeton, xxvi 4- 298 pp. i6mo net o 60 *Champlin's Young Folks' History of the War for the Union. Il- lustrated. 606 pp. 8vo a 50 *Chevrillon's In India. Descriptions of Hindu India to-day. 265 pp. 12010 150 *Cory's Guide to Modern English History. Part I. 1815-1830. 276pp. 8vo 200 Part II. 1830-1835. 576pp. 8vo.... 350 *Cox's Introduction to the Science of Comparative Mythology and Folk-lore. 396 pp. 12010 175 *Creasy's History of the Ottoman Turks. 568 pp. 12010 250 *Crowell's Logical Process of Social Development. 358 pp. iamo.. 175 *Dabney's Causes of the French Revolution. By Prof. R. H. Dab- ney of the University of Virginia. 307 pp. 12010 125 *Doyle's English Colonies in America. By J. A. Doyle, Fellow of All Souls' College, Oxford. 8vo. Vol. I. Virginia, Maryland, and the Carolinas. 420 pp 350 Vols. II. & III. The Puritan Colonies. 333 + 416 pp 700 *Durand's New Materials for the History of the American Revo- lution. Translated from documents in the French archives and edited by John Durand. 317 pp. 12010 175 Duruy's Middle Ages. Translated by E. H. and M. D. Whitney. Ed- ited by Prof. Geo. B. Adams of Yale. With 13 new colored maps. 603 pp. 12010 net i 60 Modern Times (1453-1789). Translated by Edwin A. Grosvenor. With 5 colored maps. 575 pp. 12010 net i 60 *Escott's England : Her People, Polity, and Pursuits. 625 pp. 8vo. 4 oo Henry Holt <5r Co.'s Works on History, etc. * Falke's Greece and Rome : Their Life and Art. With 400 illustra- tions. 365 pp. 4to $1000 Fleury's Ancient History, Told to Children. Translated and edited for schools as an exercise for translating from English into French by Susan M. Lane. 118 pp. 12010 net 70 Ford's Federalist. A definitive edition edited by Paul Leicester Ford. With foot-notes showing relations of subsequent events to the views in the text, an appendix of 148 pages, including the Consti- tution and references to court decisions regarding it, a full and reliable index (55 pages), and other special features. Large izmo. (Circular free,),.,,,,,..., net * 75 Fournier's Napoleon. A compact, impartial, and reliable life. Edited by Prof. Edward G. Bourne of Yale. ( In preparation.) FREEMAN'S HISTORICAL COURSE. Under the general editor- ship of the late Prof. Edward A. Freeman of Oxford. The editors of the various volumes have especially adapted them for America. i6mo. x. General Sketch of History. By Prof . E. A. Freeman. New edition, revised, with chronological table, maps, and index. 432 pp. .net i 10 9. England. By Edith Thompson. 400 pp net 88 3. Scotland. By Margaret Macarthur. 213 pp net 80 4. Italy. By William Hunt. 285 pp net 80 S.Germany. By James Sime. 282 pp net 80 6. The United States. By J. A. Doyle. With maps, illustrative of the acquisition of territory and the increase of population, by Francis A. Walker. 424 pp. net i oo 7. France. By Charlotte M. Yonge. 267 pp , net 80 Fyffe's Modern Europe. By the late C. A. Fyffe, M.A., Barrister-at- Law; Fellow of University College, Oxford; Vice-president of the Royal Historical Society. 8vo. Vol. I. From 1792 to 1814. With two maps. 549 pp 250 Vol. II. From 1814 to 1848. 525 pp 250 Vol. III. From 184810 1878. 580 pp 2 50 Modern Europe (1792-1878) unabridged. One volume edition. 1120 pp. lamo net a 75 Gallaudet's International Law. A Manual. By Pres. E. M. Gallaudet of College for Deaf-Mutes, Washington. 358 pp. i2mo net 130 Gardiner's English History for Schools. B.C. SS-A.D. 1880. By Prof. S. R. Gardiner of King's College, London. Edition re- vised for American students. 497 pp. i6mo net 80 Introduction to English History. 209 pp. 12010 net 80 English History for Students. Being the Introduction to Eng- lish History by Prof. S. R. Gardiner. With a critical and bio- graphical account of the authorities by J. Bass Mullinger, 448 pp. lamo net i 80 Gordy's Political Parties in the United States. Vol. I. 1787-1809. (In preparation^) *Ihering's The Evolution of the Aryan. 4t2pp. 8vo net 300 "Jackson's (Lady) Old Paris, Its Courts and Literary Salons. 537 pp. 12010 1.50 The Old Regime, Court, Salons, and Theatres. 532 pp. 12010 1.50 *Jenks' Law and Politics in the Middle Ages. xiU^raTpp. 8vo. net 275 Johnston's History of the United States for Schools. With an intro- ductory history of the discovery and English colonization of North America. By the late Prof. Alexander Johnston of Prince- ton. Re-vised and continued to 1897 by Prof. W. M. Daniels of Princeton. With maps, plans, illustrations, questions, a bibli- ography, etc. 499 pp. zamo. Half roan net i oo 2 Henry Holt <& Co. 's Works on History, etc. PRICE Johnston's Shorter History of the United States. With references to supplementary reading. 356 pp. lamo net $o 95 History of American Politics. Revised and enlarged by Prof. Wm. M. Sloane of Princeton. 366 pp. i6mo net 80 Lacombe's The Growth of a People. A translation of "Petite Histoire du Peuple Franchise," 232pp. i6mo net 80 The same in French net 60 *Langlois and Seignobos' Introduction to the Study of History. Treats of the search for documents, constructive reason- ing, exposition, critical classification of sources, etc. xxvii+35o pp. i2mo net, special 225 *Lossing's Life and Times of Major-General Philip Schuyler. vols. With portraits. 520 + 560 pp. 12010 350 *Maine's Ancient Law. Its Connection with the Early History of Society, and its Relation to Modern Ideas. By Sir Henry Sumner Maine. With an introduction by Theo. W. Dwight, LL.D. 469 pp. 8vo 350 *Lectures on the Early History of Institutions. A Sequel to " Ancient Law." 420 pp. 8vo 350 *Village Communities in the East and West. Six Oxford Lec- tures; to which are added other lectures, addresses, and essays. 425 pp. 8vo 350 *Early Law and Custom. Taken chiefly from Oxford Lectures. 408 pp. 8vo 3 50 *Popular Government. Four Essays. 273pp. 8vo 275 *International Law. Cambridge Lectures, 1887. 234 pp. 8vo..., 275 *Maine, Sir Henry. A Brief Memoir of his Life by Sir M. E. Grant Duff, with some of his Indian Speeches and Minutes. Edited by Whitley Stokes. With portrait. 451 pp. 8vo 350 *Mill's Considerations on Representative Government. By John Stuart Mill. 371 pp. 8vo 200 *On Liberty : The Subjection of Women. 394 pp. 8vo 200 *Morgan's Ancient Society ; or, Researches 9n the Lines of Human Progress through Savagery and Barbarism to Civilization. By Lewis H. Morgan, LL.D. 576 pp. 8vo 400 *Pollard's Political Pamphlets. Pamphlets by Burke, Steele, Saxby, Halifax, Arbuthnot, Swift, Bolingbroke, and "Junius," that have done much to shape English history. Bound in i vol. 345 pp. iamo i yij Porter's Outlines of the Constitutional History of the United States. By Luther Henry Porter. 326pp. 12010 net i so Ringwalt's Modern American Oratory. An essay on the art, with Speeches by Jeremiah Black, Phillips, Depew, Curtis, Hy. W. Grady, and Beecher. 334pp. i2mo net i oo *Rowan and Ramsay's The Island of Cuba. With bibliography, col- ored maps, points of International Law bearing on the Cuban Question, President Grant on Cuba (1875), etc. 279 pp. i6mo.. i 25 *Sumner 1 s History of American Currency. By Prof. Wm. Graham Sumner of Yale. With chapters on the English Bank Restriction and Austrian Paper Money. To which is appended " The Bullion Report." With diagrams. 391 pp. Large 12010 300 *Essays in Political and Social Science. 176 pp. 8vo 150 Protectionism. The " ism " which teaches that waste makes wealth. 181 pp. i6mo i oo Problems in Economics. Interleaved. 137 pp. i6mo net i oo 3 BHB8I