/B E R K E L t Y 
 
 [ LIBRARY 
 
 I UNJVERSiTY OF 
 \CAUFORN!A 
 
 EARTH 
 
 SCIENCES 
 
 LIBRARY 
 
THE LIBRARY 
 
 OF 
 
 THE UNIVERSITY 
 OF CALIFORNIA 
 
 PRESENTED BY 
 
 PROF. CHARLES A. KOFOID AND 
 MRS. PRUDENCE W. KOFOID 
 
 KARTO 
 SCIENCES 
 
CONVERGENCE IN EVOLUTION 
 
FIG. 7 (see p. 90). HEAD OF INDIAN MUNGOOSE (Hcrpcstes 
 smithii\ showing horizontal pupil. 
 
CONVERGENCE IN 
 EVOLUTION 
 
 BY ARTHUR WILLEY 
 
 D.Sc., LOND.; HON. M.A., CANTAB.; F.R.S. 
 
 SOMETIME BALFOUR STUDENT OF THE UNIVERSITY OF CAMBRIDGE ; LATE 
 
 DIRECTOR OF THE COLOMBO MUSEUM, CEYLON; STRATHCONA 
 
 PROFESSOR OF ZOOLOGY IN THE M'GILL UNIVERSITY, 
 
 MONTREAL, CANADA; AUTHOR OF " AMPHIOXUS 
 
 AND THE ANCESTRY OF VERTEBRATES " 
 
 WITH ILLUSTRATIONS 
 
 LONDON 
 
 JOHN MURRAY, ALBEMARLE STREET, W. 
 
 1911 
 
Gift of C. A. Kofoid 
 
 EARTH 
 
 SCIENCES 
 
 LIBRARY 
 
DEDICATION 
 
 To SIR E. RAY LANKESTER, K.C.B., LL.D., F.R.S. 
 
 DEAR SIR RAY, 
 
 Twenty years have elapsed since I 
 published my first preliminary note under your 
 guidance, and sixteen since I endeavoured to make 
 a scanty return by dedicating a volume to you. 
 
 If there was a reason then, there must be a two- 
 fold reason now for associating your name with a 
 book of this nature, the one to balance the other. 
 
 To you, with your intimate and apostolic know- 
 ledge of the situation, and mastery of the facts which 
 govern it, it may seem rash to attempt anything 
 like a revision of first principles ; but having regard 
 to the actual state of zoological speculation, yesterday 
 or to-day, it is clear that much remains to be done 
 before the ground can be freed from many of the 
 qualities which favour the growth of error. 
 
 Yours sincerely, 
 
 A. WILLEY. 
 
 a 2 
 
Kj jdb r~ 
 
 EARTH 
 
 SCIENCES 
 
 LFBRARY 
 
 PREFACE 
 
 IN these pages I have attempted to place on 
 record in a connected form some facts of natural 
 history which have, directly or indirectly, passed 
 under my own observation during a number 
 of years, together with others which I have 
 collected from various sources. Many of them 
 are known facts brought into fresh conjunctions, 
 others are somewhat less familiar, and some are 
 practically new. I have endeavoured to render 
 the language intelligible to those who have 
 an inkling of biological knowledge; and terms 
 which may appear difficult in the text are 
 usually explained in the context; but I have 
 found it impossible to avoid using the current 
 phraseology of the subject. 
 
 Morphology as comprehending the analysis of 
 organic form is distinguished from art which 
 concerns itself with the portrayal of form ; and 
 it appeals to a much more limited circle, its 
 applications being obscure and, as a general 
 rule, of interest only to specialists. Its relation- 
 
 vn 
 
 M335738 
 
viii PREFACE 
 
 ship to art is obvious in many of its aspects, 
 not least so in its bearing towards truth. The 
 flamboyant manner in which Napoleon crosses 
 the Alps on canvas is not farther from the 
 truth than many a morphological theory which 
 has received the last touches of a master hand. 
 Yet the theory and the picture though wrong 
 are presumably good, that is to say, the technique 
 is competent, and the moment selected for inter- 
 pretation and presentation interesting ; and if 
 there is a preference it must be for the picture 
 which makes no pretension to scientific accuracy. 
 
 Beyond a certain point morphology becomes 
 an art, and as such its scope is infinite ; as an 
 expression of simple truth its horizon is strictly 
 limited, and consequently there have not been 
 wanting those who have depreciated its value, 
 perhaps through forgetfulness of the fact that 
 the complexity of biological questions is rarely 
 realised even by specialists, because the origins 
 of morphological types are rooted in remote 
 and intangible obscurity. It should be added 
 that my remarks refer almost exclusively to 
 the zoological side of morphology, although 
 many of the principles involved apply with 
 equal, if not with greater force to its botanical 
 aspect. 
 
 The history of zoological classification since 
 
PREFACE ix 
 
 the time of Linnaeus has consisted largely of 
 rectifications in the systematic positions origin- 
 ally assigned to animals which were thought 
 to resemble others in certain superficial traits 
 of their organisation, and hence were classed 
 together, e.g., the Foraminifera and the Nauti- 
 loidea, the Cephalopoda and Pteropoda, the 
 Cirripedia, Tunicata and Mollusca. It is only 
 after such groups are relegated to their proper 
 position in the zoological system that the 
 phenomena of convergence begin to appear in 
 their true light with an independent interest 
 of their own. The limitations of convergence 
 coincide with those of homology, and the criteria 
 of the one are inversely those of the other. 
 Its importance in morphology is therefore clear 
 enough, and no apology would be required for 
 the repeated discussion of it, were the treatment 
 adequate. 
 
 Up to a certain point the gist of what I have 
 set down in these pages may be regarded as 
 an attempt at a reply to a recent " earthquake 
 hypothesis " concerning the origin of Vertebrates, 
 which has been published in a remarkable volume 
 by a very eminent physiologist, Dr W. H. 
 Gaskell. I can only hope that Dr Gaskell and 
 others will accept it in the spirit in which it is 
 offered. Besides this I have endeavoured to 
 
x PREFACE 
 
 suggest in outline a system of convergence 
 which is capable of further expansion ; and I 
 have made a point, so far as my purview ex- 
 tended, of acknowledging my indebtedness to 
 the work of predecessors and contemporaries. 
 
 The technical expressions which are scattered 
 through the text will be found not to interfere 
 with the comprehension of the whole. They 
 constitute part of the scheme of arrangement 
 proposed for the phenomena of convergence. 
 Phenomena require to be classified after the 
 same manner as animals, rocks and plants ; 
 and the pregnant technical terms are intended 
 to carry this out for future use and reference, 
 not for the bewilderment of the reader. 
 
 This opuscule was planned and the bulk of 
 the manuscript written in Ceylon. After my 
 return to England in the summer of this year, 
 and after the manuscript had been sent to the 
 printer, my attention was drawn by a friend 
 to an important paper by Professor Henry 
 Fairfield Osborn on " Homoplasy as a Law of 
 Latent or Potential Homology " in the American 
 Naturalist, vol. xxxvi., 1902, pp. 259-271. This 
 paper is illustrated by examples of convergence 
 in tooth-structure (dental convergence), and con- 
 tains a valuable exposition of homoplasy from the 
 palaeontological standpoint. The only point in 
 
PREFACE xi 
 
 which I find myself, unfortunately, somewhat at 
 variance with Professor Osborn is in the matter 
 of terminology. 
 
 I have used the word convergence in a wide 
 sense, embracing habits, functions, structure, and 
 physiognomy ; and I venture to hope that, after 
 due consideration, this extended application will 
 prove acceptable. Under the heading "Analogy 
 in Evolution," Professor Osborn tabulates four 
 categories, namely, Analogous Variation, Paral- 
 lelism, Convergence, and Homoplasy, each of 
 which he defines. Convergence is there defined 
 as " independent similar development of un- 
 related animals, bringing them apparently closer 
 together " ; and Parallelism is " independent 
 similar development of related animals, plants, 
 and organs." 
 
 These definitions leave us in the dark as to 
 what degrees of relationship would entitle a 
 given case to be classed as one of parallelism 
 or of convergence. The four classes represent 
 degrees of one general phenomenon, and the 
 term Analogy seems to be rather too vague 
 as applied to it. For the rest I refer the 
 reader to what I have said on the subject of 
 degrees of convergence and to the diagrams 
 of convergence and parallelism. 
 
 Professor Gilbert C. Bourne (in the Quarterly 
 
xii PREFACE 
 
 Journal of Microscopical Science, vol. lv., 1910, 
 see pp. 36-43), also finds difficulty in differenti- 
 ating between Osborn's categories, and adds 
 that " Lankester's term, homoplasy, as origin- 
 ally defined [1870], covers all the cases." 
 
 In the following pages I have tried to show 
 that homoplasy does not cover all the cases 
 which are included under convergence in the 
 wider acceptation of the term. 
 
 A. WILLEY. 
 
 LONDON, 15^ August 1910. 
 
CONTENTS 
 
 CHAP. PAGE 
 
 PREFACE vii 
 
 I. THE ART OF MORPHOLOGY, BEING A DISCOURSE 
 
 UPON ORGANIC FORM I 
 
 II. PHYSIOLOGICAL CLASSIFICATION 13 
 
 III. EXPOSED AND CONCEALED ANIMALS (PHANEROZOA 
 
 AND CRYPTOZOA) 23 
 
 IV. FREE AND FIXED ANIMALS (ELEUTHEROZOA AND 
 
 STATOZOA) 35 
 
 V. MIMICRY AND HOMOPLASY $2 
 
 VI. DIVERGENCE AND PARALLELISM 69 
 
 VII. SPECIAL CONVERGENCE 87 
 
 VIII. HABITUDES AND ATTITUDES (BIONOMICAL CON- 
 VERGENCE) 113 
 
 IX. THE WAYS OF BREATHING (RESPIRATORY CON- 
 VERGENCE) 138 
 
 X. CONVERGENCE IN MINUTE STRUCTURE (HISTO- 
 
 GENETIC CONVERGENCE) 153 
 
 INDEX 173 
 
 111 
 
LIST OF ILLUSTRATIONS 
 
 HEAD OF INDIAN MUNGOOSE (HERPESTES SMITHII\ 
 
 SHOWING HORIZONTAL PUPIL . . . Frontispiece 
 
 PAGE 
 LEAF-FISH 66 
 
 DIAGRAM OF PHYLETIC DIVERGENCE AND PARALLELISM . 70 
 PARALLEL DIVISIONS OF ECHINODERMATA . . . .71 
 
 PARALLEL DIVISIONS OF VERTEBRATA 72 
 
 PARALLELISM WITH CONVERGENCE AMONGST MAMMALS . 80 
 HEAD OF DRYOPHIS PULVERULENTUS IN SIDE VIEW . 93 
 
 EYE OF DUSSUMIERIA, SHOWING PERFORATED ADIPOSE 
 
 EYELID 95 
 
 INTESTINAL TRACT OF DUSSUMIERIA . . . . log 
 
 DISSECTION OF THE ALIMENTARY TRACT OF CHATOESSUS 
 
 NASUS FROM THE LEFT SIDE IIO 
 
 PARALLELISM WITH CONVERGENCE BETWEEN TWO FAMILIES 
 
 OF FISHES, MUGILIDjE AND CLUPEID^ . . .112 
 
 HEAD AND FORE-BODY OF ICHTHYOPHIS GLVTINOSUS, TO 
 
 SHOW THE COLLAR-LIKE RESPIRATORY REGION . 143 
 
CONVERGENCE IN EVOLUTION 
 CHAPTER I 
 
 THE ART OF MORPHOLOGY, BEING A DISCOURSE 
 UPON ORGANIC FORM 
 
 MORPHOLOGY, in the modern sense, usually con- 
 veys a genetic meaning, implying morphogeny 
 or the origin of forms. Its purpose is to follow 
 the clue of affinity which connects related but 
 divergent classes of animals and systems of 
 organs, and to recognise primary forms when 
 disguised beneath secondary facies. It is the 
 handmaiden of comparative anatomy which is 
 a much older discipline, the offspring of human 
 anatomy and physiology, born in the schools 
 of Vesalius, Fabricius ab Aquapedente, William 
 Harvey and Malpighi. Morphology, on the 
 other hand, is the child of evolution, reared 
 under the tutelage of Cuvier and Lamarck, 
 Von Baer and Haeckel, Darwin and Huxley, 
 and taken into the service of comparative 
 anatomy and embryology. It is a branch of 
 philosophy dealing partly with positive data, 
 
2 THE ART OF MORPHOLOGY 
 
 partly also with fanciful abstractions culled from 
 the contemplation of concrete phenomena, but 
 quite distinct from the soul of mankind, although 
 dependent for its existence upon the human 
 understanding. 
 
 Human life is guarded and controlled by laws, 
 the principles and spirit of which have been 
 evolved out of chaos by the mind of man 
 stimulated thereto by the necessities of the 
 fundamental claims of property, of exogamy, 
 and of fear. The existence of non-human and 
 undomesticated organisms is governed by 
 natural laws which regard not individuals except 
 in so far as they are subservient to the needs 
 of the race to which they belong. In civilisation 
 the individual is sacred and inviolable, but in 
 nature obviously at a discount. 
 
 The geological record, made up as it is of 
 organic remains preserved in a more or less 
 fragmentary condition, must always continue to 
 be incomplete precisely in regard to those 
 organisms about whose past history information 
 would be peculiarly acceptable, namely, the soft- 
 bodied animals ; but imperfect though it be, it 
 suffices to show, what the historical record of 
 seven thousand years fails to prove, that there 
 is a specific as well as an individual longevity ; 
 and that when the form manifested by a given 
 species, the resultant of a particular combination 
 of heritable characters, becomes antiquated and 
 
PURPORT OF LIFE 3 
 
 incapable of prolonging the struggle against 
 the influences, whether aggressively hostile or 
 relentlessly dominating, which possess it from 
 within and from without, it must give place, by 
 substitution, to a new combination represented 
 by a new form, which may be an outsider or a 
 scion of the old stock. 
 
 Perhaps it may be said without undue pre- 
 sumption that it is one of the vanities of the 
 imagination to consider that everything exists 
 for his special behoof. This is one of those 
 pious beliefs which are common to men and 
 dogs. One of the elementary truths revealed 
 by biology is this : that whatever use man may 
 make of his contemporaries, and however success- 
 ful he may be in bending them to his will or in 
 exterminating them, the fact remains that they 
 are not here to do his bidding but to find their 
 own sustenance, to secure their own safety, and 
 to propagate their own kind. They have an 
 objective existence guided by instincts of which 
 the superficial reactions alone can be observed 
 and named. Biology is based upon a detached 
 investigation of the properties of living matter, 
 and is thus distinguished from teleology which 
 depends upon an anthropocentric interpretation 
 of the organic environment, whereas mythology 
 may be said to proceed from an anthropomorphic 
 conception of the inorganic environment. 
 
 Anatomy, defined as the science of the structure 
 
4 THE ART OF MORPHOLOGY 
 
 of organised bodies, is usually comprehended in 
 the three faculties of human anatomy, zoologi- 
 cal anatomy (zoology), and vegetable anatomy 
 (botany). Each of these primary disciplines 
 introduces the student to several paths of learn- 
 ing, e.g., topographical anatomy (anatomy sensu 
 stricto), physiological anatomy (physiology or the 
 treatment of structure in relation to function), 
 histological anatomy (histology or microscopic 
 anatomy), ontogenetic anatomy (embryology), 
 and philosophical anatomy (morphology). These 
 are what Owen (1866) called the ways of 
 anatomy. 
 
 Histology, the anatomy of the constituent 
 parts or tissues of organs, is investigated by 
 microscopical and microchemical methods which 
 were pioneered by Xavier Bichat whose work, 
 " Anatomic ge"nerale appliquee a la Physiologic " 
 (1801), preceded by more than a quarter of a 
 century the enunciation of the cell-theory. This 
 theory laid the foundation of a great part of 
 modern biological research, and it is a significant 
 circumstance that it was published in the decade 
 which witnessed the death of Cuvier (1832). 
 It teaches that the ultimate units of all the 
 tissues of an organism consist of nucleated proto- 
 plasmic bodies, conventionally termed cells, which 
 are derived by repeated cell - division from a 
 primordial fertilised egg-cell ; and it also forms 
 the basis of the subdivision of the animal 
 
ECONOMIC ZOOLOGY 5 
 
 kingdom into protozoa or unicellular animals, 
 and metazoa or multicellular animals. 
 
 Having now briefly defined the objects and 
 the position of our enquiry, let us turn for a 
 moment to consider it from the economic or 
 cui bono standpoint. Economic zoology is to 
 morphology what the industrial arts are to fine 
 art with a difference upon which it is not 
 necessary to insist. Many of the aforesaid 
 objects of enquiry are so rarely seen that their 
 disappearance from the face of the earth would 
 not produce any perceptible effect upon the 
 economy or balance of nature. Most living 
 things are popularly conceived to be of no use 
 in the wild state ; and their extermination, so 
 long as they remain indomitable and refuse, as 
 it were, to take the oath of allegiance to man, 
 appears to be inseparable from the progress 
 of civilisation. In many cases their specific 
 longevity seems to be approaching the end, in 
 precisely the same manner as is the racial 
 longevity of the primitive and fast vanishing 
 races of mankind. 
 
 The Rev. Gregor Mendel, Abbot of Brtinn, 
 whose experiments in practical heredity are now 
 appraised by some of his followers at an even 
 higher rate than those of Charles Darwin, was 
 coeval with his posthumous rival, though appar- 
 ently unknown to him even by repute. They 
 
 were contemporaries working independently 
 
 A 2 
 
6 THE ART OF MORPHOLOGY 
 
 along the same general lines though with differ- 
 ent perspectives and methods, the one exer- 
 cising a profound substantive influence upon 
 his day and generation, and occasioning a great 
 public disputation, the personal influence being 
 permanent, the public disputation ending in 
 platitudes ; the other, in the quietude of the 
 sanctuary, unseen and unknown except to a few 
 intimates, pursuing far-reaching numerical re- 
 searches which have been rescued from oblivion 
 years after their good author had been laid to 
 rest, secure alike from excommunication and 
 from hair-splitting controversy. Students of 
 practical questions of heredity confess to the 
 hope that their investigations may lead to the 
 regeneration not only of cultivated and domesti- 
 cated races of plants and animals but also of the 
 human race, by the selection and perpetuation of 
 desirable characters and the weeding out of un- 
 desirables. From this point of view their work 
 has undoubtedly a direct economic significance. 
 
 In the same way as benefiting humanity, the 
 discoveries in protozoology and parasitology 
 which have rendered the infancy of this century 
 so illustrious, of the relations of protozoa to 
 certain epidemic and epizootic diseases, the life- 
 histories and nuclear changes of the parasites, 
 and their transmission by intermediate hosts ; 
 as well as such discoveries as that of the 
 conveyance of plague bacilli by rat-fleas, of the 
 
SUMMUM BONUM 7 
 
 micrococci of Malta fever by goats, and Looss's 
 discovery of the stereotropism or contact require- 
 ments of the larva of the tunnel worm (Anchylo- 
 stomum) all of these are of primary hygienic 
 importance. 
 
 So much however cannot be said convincingly 
 of purely morphological researches. The origin 
 of the human race, the ancestry of vertebrates, 
 the correct systematic position of fleas, 1 and 
 other cognate problems perennially awaiting solu- 
 tion, are matters of no importance in domestic 
 life, and it is hard to believe that they ever 
 will possess any such value. Although it may 
 be true that to the individual worker it is the 
 summum bonum when his science overtakes a 
 human need, it does not follow that we must 
 accept the theory that every scientific fact will 
 sooner or later directly and materially benefit 
 mankind at large, and that this consideration 
 should be the guiding motive of scientific 
 investigation. Many precious facts have been 
 
 1 For an account of a group of wingless Diptera to which fleas 
 do not appear to be related except by convergence, and the 
 amusing controversy to which it gave rise, the following papers may 
 be consulted : I. F. Dahl, " Puliciphora, eine neue, flohahnliche 
 Fliegengattung," Zool. Anz., 1897, pp. 409-412 ; translated by E. E. 
 Austen in Ann. Nat. Hist, (vii.) I. pp. 99-101, 1898 (Puliciphora, 
 a new flea-like genus of Diptera.) 2. B. Wandolleck, "1st die 
 Phylogenese der Aphanipteren entdeckt?" Zool. Anz., 1898, pp. 
 180-182. 3. F. Dahl, " Uber Puliciphora lucifera," Zool. Anz., 1898, 
 p. 308. 4. B. Wandolleck, " Die Stethopathidae eine neue fliigel 
 und schwingerlose Familie der Diptera," Zool. Jahrb. Syst., xi., 
 1898, pp. 412-441. 
 
8 THE ART OF MORPHOLOGY 
 
 lost for ever in the cataclysms of the past 
 without damaging the prospects of the human 
 race. It is certain, however, that morphological 
 researches, and the speculations which proceed 
 from them, contribute to the precision of our 
 rational conceptions, and on this account they 
 cannot be neglected. Their intellectual value is 
 an indication of their art. But I will not run 
 the risk of sapping the strength of morphology 
 by enquiring too closely where it lies. 
 
 Biometrical researches, using the expression 
 in an extended sense to include all statistical 
 studies of variation and cell-lineage, have, on 
 the whole, as little in common with morphology 
 as with parasitology. But the point of view of 
 biologists undergoes periodical change, as indeed 
 it must do, for unless the methods and the 
 outlook are varied the art will be lost. The 
 scope of any method is limited, more narrowly 
 in some cases than in others ; and after it has 
 been tested by a crowd of workers its limita- 
 tions become uncomfortably patent, and one 
 turns with relief to something altogether new. 
 At the same time it should be remembered that 
 the application of what may be termed ad hoc 
 methods to the elucidation of morphological 
 problems may only serve to demonstrate the 
 inadequacy of such methods in respect of that 
 particular application. What yields good results 
 in heredity may not be equally fruitful when 
 
TYPOLOGY 9 
 
 applied to phylogeny, the origin of the leading 
 branches of the tree of life. 
 
 Morphology in general treats of the evolution 
 of forms and organs within the compass of a 
 given type. Phylogeny does this too, and 
 more besides, since it is concerned with the 
 almost impossible task of tracing the pedigrees 
 of the primary types themselves. A little 
 explanation may be necessary at this point to 
 render my meaning intelligible. The word 
 " type " has been used in many different senses, 
 and is consequently a dangerous equivocal word 
 to trifle with, though very convenient. Examples 
 of types of the first order of magnitude are 
 the platyhelminth type, the annelid type, the 
 arthropod type, the molluscan type, the verte- 
 brate type. Probably in consequence of a 
 reaction against the old dogmatic use of the 
 term, these types have, at one time or another, 
 been connected together by imaginary links in 
 a manner which in many cases, is now recog- 
 nised as being quite inadmissible or, as we may 
 otherwise express it, in a manner contrary to 
 the canons of the art of morphology. Hopeless 
 pictures and impossible pedigrees have resulted 
 from this confusion of ideas. The criteria of 
 homology are indeed hard to define, being partly 
 intuitive, but we may find some indications which 
 may assist towards their future formulation. 
 The fish type, the bird type, and the mammalian 
 
io THE ART OF MORPHOLOGY 
 
 type may be quoted as instances of the second 
 order of magnitude ; the hoofed mammal or 
 ungulate type and the primate type are of the 
 third order ; the odd-toed ungulate of the fourth 
 order ; the equine type and the human type are 
 examples of the fifth order of magnitude. Lower 
 categories are exemplified by the type genus of 
 a family, the type species of a genus, and the 
 type individual of a species, the last being purely 
 conventional, referring to the actual specimen 
 upon which the original specific description was 
 based. The types of cultivated races are deter- 
 mined in a somewhat different manner. 
 
 From the fourth grade downwards it is 
 frequently possible to determine the pedigree 
 of a specialised type from a more generalised 
 ancestry, with considerable probability of accuracy 
 and, in some cases, as with the pig, horse, and 
 elephant, with a great deal of precision. In the 
 superior grades it becomes increasingly difficult 
 to arrive at an understanding. It is something 
 to be able to say that the vertebrate type can 
 be traced with certainty to a chordate type ; 
 but it cannot be asserted without contradiction 
 that the craniate vertebrate is derived from an 
 acraniate chordate ; because the gap between 
 these two grades of organisation is too great, 
 and many of the intervening stages of substitu- 
 tion and of cephalisation (head - formation) are 
 lost. 
 
ORGANIC SYMMETRY n 
 
 Phylogeny and heredity illustrate different 
 phases of the theory of descent with modifica- 
 tion, and have to be approached by appropriate 
 methods. The test of breeding, which is crucial 
 in heredity, is obviously inapplicable in a phylo- 
 genetic investigation where the last positive links 
 in the chain of evidence can, as a rule, only be 
 afforded, if at all, by palaeontology. 
 
 Morphology, in its aesthetic aspect, is the 
 perception of symmetry and organic beauty. Its 
 interest and importance lie less in its ultimate 
 truth than in its relative completeness, by which 
 I mean its capacity for embracing the widest 
 possible range of facts without ignoring those 
 which may not fall into harmony with a parti- 
 cular theory. 
 
 The distinction between convergent and normal 
 morphogeny has long pervaded biological litera- 
 ture ; but I believe it to be a fact that the 
 equality of interest which attaches to the two 
 branches of the subject has not been recognised 
 in an adequate manner hitherto. It will appear 
 in the course of this essay that convergence is 
 neither identical with homoplasy l nor with ceno- 
 genesis, 2 but that it includes these and some- 
 
 1 Homoplasy signifies similarity of form unaccompanied by 
 community of pedigree. As explained below, we owe the term 
 to Sir Ray Lankester. 
 
 2 Cenogenesis implies the origin of structural features by 
 relatively recent adaptation, in contrast with palingenesis or 
 primordial adaptation. The terms are due to Professor Haeckel. 
 
12 THE ART OF MORPHOLOGY 
 
 thing else besides. All homoplasy is convergence, 
 but all convergence is not homoplasy ; and the 
 same dictum may be repeated, mutatis miitandis, 
 for cenogenesis. 
 
 Professor H. F. Osborn, in the paper to which 
 I have referred in the Preface, finds homoplasy 
 to be of very great importance in the evolution 
 of mammalian teeth, " because it seems to coin- 
 cide with the principle of definite or determinate 
 evolution " which has proceeded " independently 
 in a great many different families of mammals." 
 He discusses the special value of the evolution 
 of teeth as bearing upon homoplastic mutations, 
 inasmuch as teeth are preformed beneath the 
 gum so that "new cusps, folds, crests, and styles 
 are invariably congenital." 
 
CHAPTER II 
 
 PHYSIOLOGICAL CLASSIFICATION 
 
 THE animal kingdom, according to the teaching 
 which was customary in Professor Huxley's time, 
 was arranged under the two great divisions of 
 the vertebrata or animals provided with a jointed 
 axial skeleton, the backbone, and the invertebrata 
 or animals destitute of a backbone; whereby a 
 single phylum or primary branch of the tree of 
 life was contrasted and placed upon an apparent 
 equality with a whole complex of other phyla. 
 This method of classification denoted one of 
 the great generalisations in zoology at the birth 
 of the nineteenth century, enduring from the 
 beginning to the end of that century. Those 
 who attend to these matters no longer continue 
 to oppose the homogeneous group of the verte- 
 brates to the negative association of invertebrates. 
 The former term, introduced by Lamarck and 
 Cuvier, retains its full significance, but the word 
 invertebrate, though still useful as an adjective, 
 is comparatively barren as an idea. 
 
 The classification now adopted is one based 
 
 13 
 
i 4 PHYSIOLOGICAL CLASSIFICATION 
 
 upon a profounder analysis of the animated 
 frame, namely, upon the cellular constitution of 
 the body. This second great generalisation, as 
 we have seen above, had already been achieved 
 during the first half of the preceding century, 
 but the fate of the backbone in classification 
 was not sealed until the end. In the light of 
 recent researches we may find a further reason, 
 besides the fallacy of contrasting non-equivalent 
 groups, for discarding the older method. The 
 presence of a backbone is a character which 
 does not pair with the absence of one, but it 
 does pair with the presence of a notochord. 
 
 We may venture to lay down, as a funda- 
 mental principle in morphology, that only those 
 characters pair which are connected together by 
 the bond of substitution. By way of extending 
 this principle one step further whilst reference 
 is being made to the vertebrata, we may add 
 that the presence of a bony endoskeleton is a 
 character which pairs morphologically with a 
 cartilaginous endoskeleton, but not with a horny 
 exoskeleton. On the other hand the vertebrate 
 endoskeleton, in its capacity of affording attach- 
 ment to muscles, can be compared physiologically 
 with an arthropod exoskeleton or a molluscan 
 shell. 
 
 A morphological classification must be physio- 
 logically true, i.e., it must conform to physio- 
 logical laws ; but a physiological classification is 
 
DUAL CLASSIFICATION 15 
 
 under no corresponding obligation, and it is 
 partly through losing sight of this circumstance 
 of the differential purport of morphology and 
 physiology, partly also through lack of know- 
 ledge of palaeozoic physiology, that much con- 
 fusion has arisen. 
 
 Returning now to the consideration of the 
 dual classification of the animal kingdom : the 
 current system, which has received so much 
 additional stability by the recent progress of 
 protozoology, recognises the two great sub- 
 divisions of the protozoa and the metazoa. It 
 is sometimes considered an advantage to unite 
 all unicellular organisms under the Haeckelian 
 term protista from which are derived the parallel 
 stems of the metazoa and the metaphyta (multi- 
 cellular plants). In the vegetable kingdom the 
 transition from unicellular to multicellular forms 
 is graduated even amongst existing natural 
 orders ; and the algae include both kinds. 
 
 The protozoa are commonly treated as a 
 phylum of the animal kingdom equivalent in 
 classification to the other phyla, but this pro- 
 cedure is not in strict accordance with facts. 
 The protozoa occupy a position which is unique, 
 and they are rightly contrasted with the poly- 
 phyletic metazoa. There is no such abstraction 
 as a protozoan type in the same sense as there 
 is a molluscan type or ,a vertebrate type, 
 because the combination of characters which is 
 
16 PHYSIOLOGICAL CLASSIFICATION 
 
 common to all protozoa is also that which is 
 common to all metazoan cells, except their in- 
 dependence. The positive distinction between 
 protozoa and metazoa is thus justified alike on 
 morphological and on physiological grounds. 
 
 It remains to be seen whether there are not 
 some other properties of animals that could be 
 utilised temporarily for the purpose of obtaining 
 a comprehensive survey, beyond those which 
 are afforded by their bodily composition. A 
 grouping of physiologically comparable forms 
 may indeed be very instructive, and it has 
 this additional advantage that it enables us to 
 visualise, in a single perspective, a multitude 
 of facts which would otherwise lie beyond our 
 range for the time being. It will thus be seen 
 that physiology can render great service in 
 marshalling together scattered morphological 
 units. In support of the soundness of these 
 assertions I may quote the following statement 
 from Prof. A. A. W. Hubrecht's recent mono- 
 graphic essay on the " Early Ontogenetic 
 Phenomena in Mammals." 1 Referring to the 
 relation between trophoblast 2 and amnion of 
 which he is the exponent, he says : 
 
 "A. Willey ('98) has speculated upon similar 
 relations between arthropod embryos and their 
 larval envelope, also designated as amnion, in 
 
 1 Quart. Journ. Micro. So., vol. liii., 1908, p. 76. 
 
 2 The trophoblast is the outer absorbent germ-layer attached to 
 the embryos of many viviparous animals. 
 
LAND AND WATER 17 
 
 Peripatus, Lepisma, Gryllus, Forficula, and 
 others ; has rightly interpreted the direct com- 
 parability with the vertebrate trophoblast, and 
 has looked upon it (as I have done ['95 B] for 
 vertebrates) as an adaptation to a viviparous 
 habit acquired by the terrestrial descendant of 
 an aquatic ancestor." . 
 
 An example of a simple and obvious physio- 
 logical system of classification, based upon dis- 
 tribution and covering the entire animal kingdom, 
 is that which distinguishes between aquatic and 
 land animals. There is a peculiar propriety in 
 this disposition inasmuch as all the groups of 
 land animals have had a more or less remote 
 aquatic ancestry, some inconceivably remote, 
 others comparatively recent ; whilst others again 
 have reverted to an aqueous medium in which 
 their lives are spent. Land mollusca, land 
 leeches, land nemertines, land planarians, and 
 land crabs, are examples of terrestrial groups 
 which have had a recent aquatic ancestry, their 
 immediate congeners being still aquatic animals. 
 
 The general structural differences which are 
 associated with the radical change of environment 
 are chiefly connected with the organs of respira- 
 tion and, to a less extent, those of locomotion. 
 In some cases, as with the operculate land 
 molluscs, the change from water to land has 
 only affected the organ of respiration. In others 
 no striking alteration is manifested in essential 
 organs, although accessory structures sometimes 
 
 B 
 
i8 PHYSIOLOGICAL CLASSIFICATION 
 
 appear, as in Birgus latro, the coconut robber 
 crab, where there is an accessory pulmonary 
 chamber above the gills, and in Ampullaria, a 
 tropical freshwater operculate snail living in tanks 
 and swamps exposed to drought, where there 
 is an analogous arrangement. 1 An interesting 
 and rather subtle change has taken place in the 
 habits of typical land leeches in that they have 
 lost the power of swimming, having become 
 entirely stereotropic. 
 
 Purely aquatic creatures are those which 
 breathe by absorbing the oxygen dissolved in 
 the water, and they are sometimes called " water- 
 breathing " animals on that account. If kept in 
 a limited volume of water the oxygen would 
 become exhausted and the animals asphyxiated 
 unless the water was frequently renewed as it 
 is in an aquarium. Asphyxiation sometimes 
 takes place under natural conditions in low-lying 
 swamps when the water in the deeper places 
 becomes stagnant after a prolonged drought. 
 The mud-dwelling fishes at the bottom may then 
 be seen swimming spasmodically towards the 
 surface to gulp in atmospheric air, often too 
 late to recover from the poisonous effect of the 
 noxious gases below. 
 
 There are many partially or completely aquatic 
 animals which find their food under water and 
 
 1 Birgus and Ampullaria are figured in Semper's " Animal Life, 5 ' 
 London, 1890. 
 
AMPHIBIA 19 
 
 whose appendages are adapted for swimming, 
 although they breathe atmospheric air and are 
 therefore true air-breathers. The cetacea and 
 the sirenia are completely aquatic in so far that 
 if they are driven ashore by storm or accident 
 they are irretrievably lost ; at the same time 
 they can only breathe air. The distinction 
 between air - breathers and water - breathers is 
 therefore more fundamental than that between 
 aquatic and terrestrial animals. The inter- 
 mediaries between these two groups are the 
 amphibious creatures which are by no means 
 confined to the class Batrachia or Amphibia 
 sensu stricto, but include all land crabs, some 
 molluscs, and even certain fishes, both dipnoan 
 and teleostean, as well as some reptiles. 
 
 The essential organs which perform the 
 principal functions of the body might each in 
 turn be made the basis of a physiological classifi- 
 cation which would convey a certain amount of 
 information to those for whom such knowledge 
 possesses any value. More interesting for our 
 present purpose are such groupings as depend 
 upon the relations of animals to their immediate 
 surroundings irrespective of the actual nature of 
 the environment, and it will be seen that the 
 pairs of characters illustrated by the dual sub- 
 divisions which will be discussed below, virtually 
 coincide in their nature with some of the primary 
 differences between animals and plants. 
 
20 PHYSIOLOGICAL CLASSIFICATION 
 
 The various tactics to which animals and plants 
 resort when driven by necessity or threatened by 
 danger, and the reactions which they exhibit 
 towards external stimuli such as heat, light, 
 drought, moisture, contact, proximity of food 
 and of hostile influences, have long engaged 
 the attention of biologists for the reason, amongst 
 others, that the differently constituted nervous 
 system of invertebrate animals and the lack of 
 this system in plants, in protozoa, and in sponges, 
 impart a special character to their vital reactions, 
 in contrast with the more or less intelligent 
 responses of craniate vertebrates. 1 All the 
 tendencies or "tropisms" to which protoplasm 
 is liable are to be observed, either positively or 
 negatively, in every organism, and frequently, 
 at different times, both positively and negatively 
 in the same species. 
 
 One of the most commonly observed reactions 
 is that known as phototaxis or heliotropism, these 
 terms being, however, not quite synonymous 
 inasmuch as the reaction to sunlight (i.e., helio- 
 tropism in the strict sense) is not the same 
 
 1 Cf. J. von Uexkiill, " Leitfaden in das Studium der experi- 
 mentellen Biologic der Wassertiere." Wiesbaden, 1905. 
 
 M. Verworn, "Allgemeine Physiologic," 4th edit., Jena, 1903. 
 English translation " General Physiology " from the 2nd edit, by 
 F. S. Lee, London, 1899. 
 
 H. S. Jennings, " Behaviour of the Lower Organisms." New 
 York, 1906. 
 
 J. E. Duerden, " On the Habits and Reactions of Crabs bearing 
 Actinians in their Claws." P. Zool. Soc. London, ii. 1905 (1906). 
 
ATTRACTION OF LIGHT 21 
 
 as that to artificial light (i.e., phototaxis in 
 the narrower sense). Professor Max Verworn, 
 whose work on General Physiology has done 
 so much to give precision to the interpretation 
 of elementary vital phenomena, apparently wishes 
 to discard the term heliotropism altogether in 
 favour of phototaxis ; but there are certain cases 
 which oppose this blending of terms. 
 
 The nocturnal lepidoptera are negatively 
 heliotropic but positively phototactic, being im- 
 pelled by an irresistible attraction to a bright 
 lamp. Amphioxus, the lancelet, is passively 
 heliotropic, but actively phototactic, evincing 
 extreme agitation upon the approach of a lighted 
 candle at night. In the same way fishes are 
 attracted by torchlight at sea, and lighted lamps 
 are placed at the head of fish-traps, where they 
 are kept burning all night to attract prawns 
 which are required for bait in large quantities. 
 Lighted prawn-traps and moth-traps represent 
 practical applications of the knowledge of the 
 phototactic properties of these animals. The 
 phosphorescent organs of deep-sea fishes no 
 doubt also act as luminous lures. 
 
 An intimate acquaintance with the habits 
 and reactions of animals is not confined to the 
 human race. Many reptiles, birds, and mammals 
 are accomplished field entomologists, know- 
 ing where to look for their food, and how to 
 distinguish edible from inedible kinds. The 
 
 B 2 
 
22 PHYSIOLOGICAL CLASSIFICATION 
 
 tropical tree lizard Calotes will take its stand 
 beside a beehive to catch the bees as they issue, 
 swallowing them entire ; * or beside an incipient 
 termite mound picking off the workers. Bulbuls 
 will gather round a swarming termite nest and 
 decimate the winged emigrants. Bats will fly to 
 a bright light, not because they are themselves 
 attracted by it, but because they know where to 
 find insects, especially winged termites, which are. 
 Geckoes and toads similarly approach artificial 
 light in search of insect food. In the three 
 last-mentioned examples of nocturnal animals 
 seeking their food by artificial light we find 
 once more (as in the case of moths) positive 
 phototaxis in apparent conflict with negative 
 heliotropism. 
 
 It seems important to take notice of this 
 association of contrary reflexes, of which more 
 instances will be found below. Meanwhile it 
 may be useful to distinguish habits from re- 
 actions : a habit is the behaviour of an animal 
 in the open ; a reaction is its behaviour in 
 the laboratory. Finally, whilst dealing with 
 definitions we may add that the word tropism 
 means the tendency to react in a definite manner 
 towards external stimuli. 
 
 1 Mr Francis Buckland has recorded a similar observation of the 
 common toad eating bees in Oxfordshire. 
 
CHAPTER III 
 
 EXPOSED AND CONCEALED ANIMALS 
 (PHANEROZOA AND CRYPTOZOA) 
 
 PHANEROZOIC or diurnal animals are positively 
 heliotropic ; cryptozoic animals including crepu- 
 scular, nocturnal, and subterranean forms, in 
 fact all that avoid the light of day, are negatively 
 heliotropic. Flights of butterflies at high noon 
 are followed by swarms of hawk-moths and 
 noctuids at twilight. The contrast is particularly 
 well illustrated on a broad and spectacular scale 
 by the interchange of amenities between such 
 gregarious creatures as crows and the so-called 
 flying foxes or fruit-eating bats, due to their 
 homing in the same place, as may be observed 
 daily in certain favourable localities in the mari- 
 time districts of Ceylon. I recorded these facts 
 in " Spolia Zeylanica," after mentioning them 
 in a letter to the late Professor Alfred Newton 
 who expressed his appreciation of them. These 
 animals keep remarkably regular hours of work 
 and sleep, the birds foraging by day, the bats 
 by night. 
 
 2 3 
 
24 EXPOSED AND CONCEALED ANIMALS 
 
 At one place in particular, a small lighthouse 
 islet off the south-west coast named Barberyn, 
 which is covered by a coconut plantation, they 
 congregate in the palm trees alternately by night 
 and by day. If one crosses over to the island 
 in the heat of the day all is quiet and nothing 
 out of the common is to be noted ; but about 
 the time of sunset a great commotion of crows l 
 among the tree tops bursts upon the ear, and 
 gradually subsides in the dusk of the evening. 
 This signalises the arrival home of the colony 
 of crows after their day's work is over. The 
 approach of sunrise, on the other hand, is 
 announced by the chattering and squabbling 
 of numerous flying foxes 2 overhead. 
 
 At sundown the passage of immense flocks 
 of crows and flying foxes in opposite directions 
 across the strait which divides the island from 
 the mainland can be witnessed, the former 
 bound for the island to rest for the night, the 
 latter speeding their way to the mainland intent 
 upon their nocturnal forage. The flying foxes 
 travel on the average distinctly higher than 
 the crows, starting singly and increasing to 
 large flocks of twenty-five and upwards, finally 
 becoming a continuous stream. The crows 
 obviously outnumber the bats, although weight 
 for weight they probably represent an equiva- 
 
 1 The grey-necked Indian crow, Coruus splendens 
 
 2 Pteropus medius. 
 
CROWS AND FLYING FOXES 25 
 
 lent bulk of living matter. In the evening the 
 
 crows begin to arrive in small numbers before 
 
 the vanguard of the bats has started, increasing 
 
 in their turn to large battalions until a period of 
 
 maximum migration is reached, and then troops 
 
 of bats are to be seen passing over still larger 
 
 columns of crows in the opposite direction, the 
 
 whole of the cross-migration occupying about 
 
 half an hour. The reverse passage, namely, 
 
 the matutinal flight, takes place towards sunrise, 
 
 the bats returning from the mainland to rest 
 
 for the day suspended in rows from the midribs 
 
 of the palm leaves, the crows crossing over on 
 
 their daily quest for garbage. This instance 
 
 may be classed as one of convergent homing, 
 
 the same trees affording hospitality in regular 
 
 alternation to day-flying birds and night-flying 
 
 mammals. 
 
 Phanerozoa and cryptozoa are two well-marked 
 physiological groups, comprising between them 
 all the forms of animal life ; and the distinction 
 is paralleled by the following more fundamental 
 contrast between animals and plants. In general 
 habit and mechanism of growth the vegetable 
 kingdom as a whole exhibits positive phototaxis 
 in the wide sense, whereas the animal kingdom 
 as a whole seems to show negative phototaxis ; 
 the former proposition can be accepted without 
 question, but the latter requires substantiation. 
 Passing over the vast array of strictly cryptozoic 
 
26 EXPOSED AND CONCEALED ANIMALS 
 
 forms, it may be noted in the first place that 
 while the green colouring matter of plants 
 effects its characteristic reaction, the chlorophyll 
 reaction, by reason of its exposure to sunlight, 
 the respiratory pigments of animals which pro- 
 mote another kind of gaseous interchange, are 
 enclosed in blood-vessels and sinuses, generally 
 concealed from the light by an opaque skin, and 
 always independent of the light, even though the 
 skin should happen to be transparent. 
 
 As a second example to illustrate broadly the 
 essential cryptotaxis (tendency to concealment) 
 of animals, may be mentioned the well-known 
 fact that even among the higher vertebrates 
 many birds and most mammals are, in the wild 
 state, concealed from view by their protective 
 coloration. Of course this does not in itself 
 shelter them from the sun ; the environment 
 provides the cover ; but it protects them from 
 the visual acuity of their enemies, many of 
 which are notoriously cryptotactic. A common 
 method of shooting wild animals, for sport or 
 for food, is to lie in wait at night concealed 
 behind an ambush within range of a water-hole 
 where they will come to drink. The largest 
 mammals, whether carnivorous or herbivorous, 
 such as the tiger, elephant, and giraffe, are 
 known to be almost invisible when at rest in 
 the forest, jungle, and high grass which they 
 frequent in the daytime. In Ceylon, the rounded 
 
CHROMATISM 37 
 
 bosses of gneiss with their dark smoothly 
 weathered surfaces, which are such prominent 
 features in the landscape in many parts of the 
 country, often present an uncommon resemblance 
 to the form of the elephant ; and the converse 
 is equally true as may be verified when a herd 
 of elephants is seen against a background of 
 gneiss, the likeness being based upon the posses- 
 sion of a common ground colour superimposed 
 upon a ponderous bulk. 
 
 Gorgeously coloured butterflies and beetles, 
 small flower-haunting birds, such as the humming- 
 birds of the New World and the sun-birds of the 
 Old World, whose plumage reflects a wonderful 
 metallic lustre, larger birds with majestic males, 
 the powerful diurnal birds of prey, and others 
 afford, to the general rule of concealment, 
 exceptions which do not stultify the rule but 
 are explicable on special grounds of adaptation, 
 selection, season, locality, and dominance. 
 
 Butterflies exhibit themselves in their seasonal 
 flights sometimes in vast numbers, but their 
 abundance at certain times only serves to 
 accentuate their absence or rarity at others ; the 
 rapidity with which an immense noon-day swarm 
 can efface itself is marvellous to behold ; while 
 the pupae and even the larvae are often quite 
 concealed by various devices of shape, position, 
 and colour. So with diurnal birds, the nests of 
 the commonest are often very hard to find and 
 rarely seen. 
 
28 EXPOSED AND CONCEALED ANIMALS 
 
 Land planarians offer instances of cryptozoic 
 animals which are brightly coloured ; the species 
 are differentiated according to their colour mark- 
 ings, and richly tinted individuals are to be found 
 in rotten logs in a dormant state by day, with the 
 body looped in a characteristic manner; if left un- 
 observed for a short time exposed to light, they 
 glide rapidly away to a place of concealment. 
 
 A journey through any tropical forest or 
 jungle, or even a little reflection will, I think, 
 suffice to convince one that while the vegetation 
 is luxuriantly phanerotactic, animal life is pre- 
 dominantly cryptotactic. The jungle is like the 
 desert and the ocean, to all superficial appear- 
 ances frequently devoid of animal life. This is 
 possibly not the impression which one would 
 receive from the perusal of faunistic works ; but 
 it is certainly that which is produced by observa- 
 tions in the open, and I regard it as one of the 
 radical bionomical or habitual differences between 
 animals and plants. 
 
 Vegetation is always exposed, it is permanently 
 phanerotactic, but the most conspicuous animals 
 in the natural state are only partially or periodi- 
 cally so. In many instances, especially among 
 the hoofed mammals (Ungulata), the cryptotaxis 
 can be properly appreciated only under natural 
 conditions. Arboreal mammals, such as squirrels 
 and monkeys, are amongst the most clearly phan- 
 erotactic members of the class Mammalia, but 
 
DAY AND NIGHT 29 
 
 many of the arboreal Lemurs, the most primitive 
 of the order Primates, are strictly nocturnal and 
 hence cryptotactic. Burrowing mammals are 
 generally nocturnal as well, including the egg- 
 laying mammals of Australia, the scaly ant-eaters, 
 ground-porcupines, and others too numerous to 
 mention ; the combination of fossorial with diurnal 
 habits is rare, the common rabbit affording an 
 instance of it. Others resort during the daytime 
 to ready-made holes, hollow trees and caves, as 
 the dasyures, mungooses, and sloth-bears. Some 
 species exhibit mixed habits ; but the cryptotactic 
 tendency is universal amongst mammals. 
 
 It is not possible to tabulate lists of phaner- 
 ozoa and cryptozoa, because all animals have 
 cryptozoic potentialities. Those that are not 
 definitely cryptotactic have the power of retreat- 
 ing to cover at the sign of danger ; or of causing 
 the aggressor to vanish from the scene by reason 
 of their weapons of offence. The scent-glands 
 of the mephitic skunk, of the naked bat (Chiro- 
 meles], and in a lesser degree of the civet cats, 
 are cryptotactic organs of a virulent type not- 
 withstanding that they effect their object through 
 the olfactory sense ; besides being protected by 
 their secretions, these animals are nocturnal. 
 Scorpions, centipedes, and poisonous snakes are 
 cryptozoic as well as armed for attack, so that 
 they also are doubly protected and doubly 
 dreaded by their enemies. 
 
30 EXPOSED AND CONCEALED ANIMALS 
 
 A special example of the cryptozoic tendency 
 in poisonous snakes is the behaviour of the cobra 
 in captivity in the presence of its food. Many 
 Eastern jugglers and gipsies carry cobras about 
 with them coiled up in round flat baskets, which 
 are kept covered except during a performance. 
 These cobras are fed upon eggs, which they 
 swallow whole without breaking the shell. If 
 they are preparing to cast their skin, and the 
 eyes are glazed, they do not take food. But 
 if they are in a normally hungry condition they 
 will not touch a proffered egg so long as the 
 basket is uncovered. If the cover is removed 
 again two or three minutes after it has been 
 replaced with the egg in the basket, the egg 
 will be found to have been swallowed, and can 
 be perceived travelling down the gullet by the 
 protuberance which it causes ; a second egg 
 may be taken in like manner shortly afterwards. 
 
 Mr Frank Buckland (" Curiosities of Natural 
 History," second series, reprinted, London 1903, 
 p. 131) wanted to see whether a hedgehog would 
 eat a common harmless snake. He caught a 
 snake near Harrow, and bought a hedgehog 
 in St Giles's. For several mornings he placed 
 them together on the grass ; but they took no 
 notice of each other. At last, one evening, he 
 shut them both up in a box together. During 
 the night the hedgehog attacked and devoured 
 half the snake, beginning at the tail. In a 
 
PHOTOPHOBIA 31 
 
 few more hours the rest of the snake was con- 
 sumed. 
 
 Blood - sucking leeches prefer to feed under 
 cover and, when gorged, retreat to cover as 
 quickly as may be. A Ceylon land leech which 
 I was keeping under observation remained on 
 my arm for ten minutes without attempting to 
 puncture the skin, in a contracted, flushed, 
 quiescent posture. I then placed a silk handker- 
 chief over the waiting worm, with the result that 
 two minutes later a puncture was felt, and the 
 attachment continued for thirty - eight minutes. 
 When the leech was placed upon a dead leaf after 
 the meal it immediately crawled underneath. 
 
 Apart from parasitism, the purest expression 
 of cryptotaxis is that which is manifested by 
 genuine cave-dwellers, and by animals which not 
 only burrow underground, but also find their 
 food underground, such as the mole - crickets 
 (Gryllotalpa 1 }, the earthworms and the creatures 
 which feed upon them, namely, moles, worm-eat- 
 ing slugs (Testacella), earth-snakes (Uropeltid&\ 
 and apodous Batrachia (Cceciliidcz). I have kept 
 an Ichthyophis alive for ten months in one half 
 of a coconut shell covered over by the other 
 half, feeding it exclusively upon earthworms. 
 
 The most primitive members of the phylum 
 Arthropoda are to a great extent cryptozoic ; in 
 Ceylon there is a species of the Thysanuran 
 
 1 Gryllotalpa flies to artificial light. 
 
32 EXPOSED AND CONCEALED ANIMALS 
 
 genus Machilis which is found living on the 
 moist and shaded surfaces of lichen-covered tree- 
 trunks and rocks, the mottled appearance of 
 which the insects closely resemble. It is doubt- 
 less in virtue of this singular property of con- 
 cealment that so many of the primitive forms 
 have survived to the present day to be at once 
 the delight and bewilderment of the systematist. 
 The noxious arthropods, which are only too 
 familiar in certain quarters, are also cryptozoic, 
 but it may be satisfactory to note that they are 
 not primitive ; they belong irreclaimably to the 
 things of darkness. 
 
 Earthworms are well known to every horti- 
 culturist and tiller of the soil, and to every one 
 who has read Darwin's famous book about them, 
 but the other cryptozoa (sensu stricto) are hardly 
 known outside the ranks of specialists. Darwin 
 estimated that in many parts of England more 
 than ten tons of dry earth annually passes through 
 the bodies of earthworms on each acre of land. 
 Equally impressive figures could be given to 
 show the amount of sand-casting performed by 
 the lugworms (Arenicolidce) on some temperate 
 shores, and by the Enteropneusta (Balano- 
 glossus) on some tropical shores. 
 
 So far I have referred almost exclusively to 
 land animals because they are most in evidence 
 by the effects which they produce in relation to 
 husbandry. Predatory aquatic animals, especially 
 
CRYPTOTAXIS 33 
 
 the marine mammals, sea snakes, the larger 
 fishes, free-swimming molluscs and Crustacea 
 (e.g., Portunidse, the swimming crabs), may be 
 described as phanerozoic ; but we may confidently 
 assert that in all cases a definite cryptotactic bias 
 of varying intensity could be demonstrated. The 
 large Cephalopod molluscs, including the squids, 
 cuttlefishes, and octopus, actually have a crypto- 
 tactic mechanism, the ink-sac, by the compression 
 of which a black fluid is discharged, effectually 
 covering the retreat of the individuals. 
 
 Nautilus, with its great external chambered 
 shell, has no ink-sac, but lives in deep water 
 and presents a scheme of pigmentation which 
 secures its partial invisibility ; the upper part 
 of the body has the form of a fleshy hood from 
 the dark-brown surface of which arise whitish, 
 wart-like prominences giving a mottled colora- 
 tion to the exposed part of the body, which 
 harmonises well with the zebra-like markings on 
 the shell. These simulate the play of light upon 
 the surface ripples of the sea, whence I called 
 them "ripple- markings." On one occasion I 
 accidentally dropped a healthy Nautilus overboard 
 in four or five fathoms of very clear water in 
 Sandal Bay, Lifu (Loyalty Islands), and, not- 
 withstanding the translucency of the water, it 
 disappeared instantaneously from view, and baffled 
 all the efforts of an expert native diver, who was 
 
 with me on the raft, to recover it, the alternate 
 
 C 
 
34 EXPOSED AND CONCEALED ANIMALS 
 
 light and dark bands on the shell constituting 
 a most effective cryptic device. 
 
 The multitudes of floating organisms which 
 compose what is known as the plankton, the 
 primary food - supply of sea, river, and lake, 
 are often rendered invisible by their extreme 
 transparency. The animals which feed upon 
 them do not pursue them by the aid of sight, 
 but simply engulf them. Conversely, many of 
 them have what may be called a phanerotactic 
 mechanism, rendering them phosphorescent and 
 visible at night. The utility of this mechanism 
 to the owners of it is not very clear ; it may 
 be a product of metabolism, neither useful nor 
 dangerous. 
 
 The basic quality underlying all animal life 
 is the cryptic, the fear of the sun. Basking in 
 the sun is a dangerous pastime. 
 
CHAPTER IV 
 
 FREE AND FIXED ANIMALS 
 (ELEUTHEROZOA AND STATOZOA) 
 
 WE may now consider briefly another couple 
 of essential properties of plants and animals, and 
 note once more how the contrasted qualities are 
 paralleled within the limits of the animal kingdom 
 alone. The power of automatic locomotion, of 
 executing free movements of translation from 
 place to place, is one of the distinctive privileges 
 of animals. The property of becoming rooted 
 to the soil, and of undergoing perennial growth 
 and regeneration, is the no less picturesque 
 attribute of plants. 
 
 So inseparable from animal life did mobility 
 appear, that the sponges, hydroids, corals, and 
 bryozoa were formerly classed together as 
 zoophytes, and were either believed to be plants 
 of a peculiar kind or, as the name implies, to 
 partake of the nature of both plants and animals. 
 The purely animal nature of the coral polyps 
 was established in the middle of the eighteenth 
 century, that of sponges towards the middle of 
 
 35 
 
36 FREE AND FIXED ANIMALS 
 
 the nineteenth. Sponges were distinguished from 
 the other zoophytes by the absence of polyps, but 
 at an early stage of modern investigation the 
 discovery was made that currents of water enter 
 the body of the sponge through small pores in 
 its outer wall, and leave it through larger apertures 
 called oscula ; subsequently the free-swimming 
 ciliated larva was made known. In the same 
 way the structure and life-history of other fixed 
 or sessile animals were described, and it became 
 an axiom that all such animals have been derived 
 from free-living ancestors. 
 
 True sessile animals have been defined as 
 those which, during the greater portion of their 
 lives, are unable to transport themselves from 
 the spot to which they have become attached, 
 but, unlike parasites, are able to secure their 
 own sustenance. Each phylum, with the excep- 
 tion of the craniate Vertebrata, has sedentary 
 members ; all sponges, most actinians (corals and 
 sea anemones), and hydroids are sessile ; com- 
 paratively few echinoderms are fixed at the 
 present age, but "the combined evidence of 
 comparative anatomy, embryology, and palaeon- 
 tology indicates that the Echinoderma owe most 
 of their obvious characters ... to their having 
 passed through a pelmatozoic (stalked) stage, 
 i.e., a stage in which the animal was attached 
 by a part of its body wall." 1 The recent and 
 
 1 F. A. Bather, "Echinoderma in Lankester's Treatise on 
 Zoology," part iii., 1900, p. 3. 
 
SEDENTARY HABITS 37 
 
 extinct echinoderms are accordingly divided into 
 two grades : A. Pelmatozoa ; B. Eleutherozoa 
 (s. str.). 
 
 The remarkable prevalence of the sedentary 
 habit amongst the lower (i.e. invertebrate) animals 
 seems to indicate that something peculiarly prim- 
 ordial lies at the back of the phenomenon, and 
 makes it desirable that it should be expressed in 
 physiological classification. I venture to assume 
 the liberty to employ two contrasting terms in a 
 wide physiological sense, which have been used 
 already in a narrow morphological sense, namely, 
 Eleutherozoa or free animals, and Statozoa or 
 sedentary animals. 1 
 
 The effects of the sedentary habit upon 
 the growth and organisation of the animals 
 accustomed to it have been amply discussed by 
 Semper, Lang, and others, and need not be 
 described here ; but it would be instructive if 
 we could establish the existence of a primary 
 function which formed the physiological basis 
 of all this varied fixation occurring in widely 
 separated morphological groups. There is an 
 extra value in this inasmuch as we might by 
 its means obtain some light upon a number of 
 perplexing problems to which no direct answer 
 is possible, e.g., as to whether the fixed hydroid- 
 phase is more primitive than the free medusoid- 
 phase in Hydromedusse ; or again as to whether 
 
 1 In morphological nomenclature Statozoa is synonymous with 
 Pelmatozoa (vide Bather, op. tit.) 
 
 C 2 
 
38 FREE AND FIXED ANIMALS 
 
 the free Appendiculariae (tailed, pelagic Ascidians 
 or Tunicata) are to be derived from fixed or from 
 free-swimming ancestors. These questions may 
 be resolved into one, that of the statozoic origin 
 of Medusae and Appendiculariae respectively. 
 Analogy affords no safe guide in solving such 
 a riddle as this. There must be some common 
 principle lying at the root of such a widely 
 spread phenomenon as that of the fixation of 
 animals ; and in my judgment the principle is 
 expressed in the term stereotropism invented 
 by J. Loeb. 1 
 
 As a typical example of the positive working 
 of the stereotropic reaction Loeb quotes Dewitz's 
 account (1885-1886) of the fertilisation of the 
 egg of the cockroach. The spermatozoa are 
 attracted by any surface ; when examined in a 
 drop of salt solution under the microscope they 
 are seen to be adhering to the slide and cover- 
 glass, none being free in the middle of the drop. 
 If a glass bead is introduced into the drop they 
 adhere to that, never leaving it. "When once 
 on the surface of the egg, the spermatozoa can 
 no more leave it, but must move on its surface 
 incessantly. In this way one spermatozoon finally 
 reaches the micropyle and gets into the egg. 
 The impregnation of the egg is therefore in 
 
 1 Jacques Loeb, "Dynamics of Living Matter," New York, 1906, 
 p. 156. Stereotropism implies contact requirements or the reaction 
 to hard surfaces. 
 
STEREOTROPISM 39 
 
 this case a function of the stereotropism of the 
 spermatozoa." [Loeb., loc. cit.~\. 
 
 As an instance of the negative working of 
 the same reaction, Loeb quotes the case of the 
 nauplii of Balanus. This does not appear to 
 be so successful ; it means either very little, or 
 something which I do not understand. Referring 
 to positive stereotropism, which I regard as a 
 factor of great potentiality, Loeb explains that it 
 is no real tropism inasmuch as lines of force do 
 not exist. Much less then must negative stereo- 
 tropism be responsive to force of any descrip- 
 tion ; not only is it a negative quality, it is, as 
 we shall see, practically nothing beyond words. 
 
 The acorn barnacles (Balanidae) make up the 
 familiar crowded communities of white shelly 
 bodies which are firmly attached to rocks 
 between tide marks, and to other suitable sur- 
 faces. All barnacles are fixed, but these are 
 doubly statozoic, cemented by their flat base 
 without a stalk, in contrast with the goose 
 barnacles (Lepadidse, pedunculate Cirripedes) 
 which hang by a stalk attached to logs and 
 other floating objects, sometimes to fishes, sea 
 snakes, sea birds, and Crustacea. 
 
 The stalked barnacles, although fixed, chiefly 
 affect what may be termed a planozoic (vagrant) 
 habit analogous, up to a certain point, with 
 that of ectoparasites. Commonly associated 
 with floating communities of Lepas, are two 
 
40 FREE AND FIXED ANIMALS 
 
 planozoic Annelid worms, both belonging to 
 the family Amphinomidae, Hipponoe gaudichaudi 
 and Amphinome rostrata. The former is some- 
 times found between the valves of Lepas and 
 is coloured pinkish orange like the egg ribbons 
 of the barnacle. This association of forms has 
 been met with on the Atlantic coast of North 
 America, 1 and I have obtained them from the 
 coast of Ceylon. Other planozoic forms with 
 similar distribution associated with Sargassum 
 or Gulf weed are the prawn Leander natator 
 Bate, and the mollusc, Scyllaa pelagica, both 
 of which simulate the colour and foliations of 
 the brown seaweed upon which they live. 
 
 Lepas is specially interesting, as it combines 
 the statozoic with the planozoic habit. As 
 mentioned above, statozoic adults usually have 
 free-swimming or pelagic, i.e., pleotropic larvae, 
 and the nauplius larva of barnacles, at least in 
 its last phase when it is called the metanauplius, 
 already possesses the primordium of the static 
 mechanism, namely, the cement organs in the 
 antennules. 2 Therefore unless there is some 
 special manifestation which Professor Loeb does 
 not explain, it would seem desirable to describe 
 the nauplius larvae as positively pleotropic, or 
 
 1 J. P. Moore, "Some Pelagic Polychseta new to the Wood's 
 Hole Fauna," P. Ac. Philad.^ vol. Iv. part iii., 1903 (1904), pp. 
 793-8oi. 
 
 2 Cf. " Korschelt und Heider. Lehrb. d. vergl. Entwicklungsge- 
 schichte der wirbellose Thiere," Jena, 1890, p. 405. 
 
PLEOTROPISM 41 
 
 simply pelagic, rather than as negatively stereo- 
 tropic. On the contrary, it is the intrinsic and 
 latent stereotropism of the larvse which leads 
 to the statozoic habit of the adult. 
 
 Furthermore the nauplii of Cirripedes are well 
 known to be markedly heliotropic, 1 and this 
 character might temporarily mask their potential 
 stereotropism, leading them out into the open 
 sea. Groom obtained them in large numbers 
 in the tow-net at Plymouth during February 
 and March, 1893, belonging to a species which 
 lives on the limestone rocks below the labora- 
 tory, in places considerably above high-tide 
 mark, where they can only be moistened by 
 the passing foam ; but he was unable to rear 
 them in confinement. 
 
 The Gastropod mollusca (snails, slugs, lim- 
 pets), with their creeping adhesive foot, are 
 pre-eminently stereotropic ; yet it is not doubted 
 that the typically pelagic or pleotropic Hetero- 
 poda and Pteropoda have been derived from 
 them along different lines of descent. The 
 free-swimming veliger larva of mollusca might 
 be described equally well with the nauplius of 
 Balanus, as showing negative stereotropism, and 
 as indicating a free-swimming ancestry. It 
 seems more just to suppose however that larvae 
 
 1 i. T. T. Groom. " On the Early Development of Cirripedes." 
 Phil. Trans., London, vol. clxxxv., 1894, pp. 119-232. 2. A. Gruvel, 
 " Monographic des Cirrhipedes," Paris, 1905, p. 445. 
 
4 FREE AND FIXED ANIMALS 
 
 in general can be regarded with greater propriety 
 as the vehicles of future adult types than as 
 recapitulations 1 of past ancestors. The larva, 
 looked at synthetically in its entirety as an 
 independent organism, apart from structural 
 details, is essentially the vehicle of the adult, 
 not that of an ancestral form. Thus the frog 
 doubtless had a fish-like ancestor, but the 
 tadpole, which is the fish-like larva of the frog, 
 does not recapitulate the fish-like ancestor of 
 the frog. The external facies of the tadpole is 
 common to the larval or postlarval stages of 
 many fishes, and represents a fundamental form 
 likely to recur by convergence. 
 
 The oft-quoted Pilidium larva of Nemertine 
 worms is an actual vehicle within human ex- 
 perience, a vehicle from which the worm alights 
 when it has reached a certain stage on its 
 journey through life. The Tornaria larva of 
 Enteropneusta and the larval forms of Echino- 
 derms which are related to it are also obviously 
 nothing else but larval, i.e., not ancestral forms 
 in themselves. We may therefore repeat our 
 dictum that the nature of larval forms is that 
 of vehicles of the future rather than relics of the 
 past, true larval characters never having been 
 adult characters. 
 
 1 For a recent discussion of the theory of recapitulation see the 
 following: Adam Sedgvvick, "The Influence of Darwin on the 
 Study of Animal Embryology" in "Darwin and Modern Science," 
 Cambridge, 1909. 
 
SWIMMING ANNELIDS 43 
 
 It is sometimes customary to speak simply of 
 positive and negative reactions, but it should not 
 be forgotten that there are all degrees of stereo- 
 tropism, from an occasional or facultative mani- 
 festation to a chronic condition, leading on to a 
 definite sedentary habit. Many cases amongst 
 the Annelid worms could be instanced where 
 stereotropism and pleotropism (the free swimming 
 habit) exist side by side, the latter frequently 
 only manifesting itself at the breeding or swarm- 
 ing season, as with epigamous Nereids and the 
 celebrated Palolo worm (Eunicidse) of Samoa 
 and elsewhere in the South Seas. 
 
 Professor Hugo Eisig l has made exact observa- 
 tions upon a number of Annelida Polychseta, re- 
 presenting ten families, some of which exhibit 
 a high degree of stereotropism together with an 
 equally high degree of pleotropism, e.g., Lepid- 
 asthenia and Ophiodromus. Others again like 
 Psammofyce and Diopatra, never advance beyond 
 what is described as a swimming gait, i.e., crawl- 
 ing along the bottom with a horizontal undulating 
 movement of the body producing alternate arcs 
 or waves of progression. When the amplitude of 
 these undulations surpasses a certain magnitude, 
 the rapidity of movement is thereby increased to 
 such an extent that the animal rises from the 
 bottom and swims through the water. It is a 
 
 1 H. Eisig, Ichthyotomus sanguinarius. Monograph 28 in "Fauna 
 und Flora des Golfes von Neapel," Berlin, 1906, see pp. 190-267. 
 
44 FREE AND FIXED ANIMALS 
 
 remarkable and suggestive fact that the typically 
 pelagic Alciopidae, freed entirely from their stereo- 
 tropic bonds, are usually very poor swimmers. 
 
 The case of Nephthys is worthy of special 
 mention. This is one of the best swimmers 
 known amongst Annelids, although when left 
 alone it passes most of its existence buried in 
 the sand. Eisig found that, unlike Psammolyce 
 which paddles along the bottom but never 
 swims, Nephthys on the contrary undulates 
 rapidly through the water, but never creeps along 
 the bottom ; it is either at rest or swimming, 
 passing abruptly from the one condition to the 
 other without an intervening reptant phase. 
 
 An analogous instance of spasmodic pleo- 
 tropism on the part of a normally stereotropic 
 and partly statozoic organism is the case of 
 the bivalve mollusc, Pecten the scallop. "If 
 disturbed the attached scallop can break or 
 cast off its byssal threads and swim by clapping 
 the shell " ; the gaping ventral lips of the shell 
 being directed forwards in locomotion. 1 
 
 The fact, elucidated by Dr Eisig, that the 
 posture of the Annelid appendages or podia on 
 opposite sides of the arcs is the same when the 
 worm is progressing slowly along the bottom 
 as when swimming rapidly through the water, 
 has led him to the conclusion that the stereo- 
 tropic movement (on the bottom) is secondary 
 
 1 Cf. W. J. Dakin, "Memoir on Pecten, 3 ' L.M.B.C. Memoir 
 XVII., Liverpool, 1909. 
 
LEECHES 45 
 
 as compared with the pleotropic movement 
 (through the water), and is derived by inheri- 
 tance from swimmers. 
 
 In the same strain, referring to Amphioxus, 
 Dr Eisig (loc. cit., p. 276) says that however 
 primitive one may hold its organisation to be, 
 yet nobody will assert that its limbless body and 
 its predominantly cryptoid locomotion represent 
 archaic features. I am unable to follow these 
 conclusions myself, as I approach this matter 
 independently from a cryptozoic and stereo- 
 tropic standpoint ; and in this, as in many 
 another affair, everything depends upon the 
 point of view. The facts are undeniable, but 
 the way of dealing with them cannot be other 
 than arbitrary. 
 
 Another instructive example of the association 
 of stereotropism of a highly specialised type with 
 well - marked pleotropism is afforded by some 
 leeches. The stereotropism of the medicinal 
 leech is of such a nature that it has passed into 
 a proverb. Closely connected with this habit 
 is its method of obtaining its food by sucking 
 the blood of vertebrates. It progresses along a 
 hard surface by the looping gait, and can also 
 swim rapidly by vertical undulations of the body. 
 The land leeches of Ceylon and Japan belong to 
 genus Hczmadipsa. They also, as mentioned 
 above, feed upon the blood of vertebrates, for 
 which they exhibit a remarkable propensity 
 
46 FREE AND FIXED ANIMALS 
 
 which might almost be raised to the rank of 
 a tropism. Professor C. O. Whitman has 
 published observations on the Japanese land 
 leech, 1 some of which I have confirmed for the 
 Ceylon land - leech, including the fact that it 
 cannot swim ; it is entirely stereotropic. It 
 has the power of repairing or closing severe 
 wounds, but cannot regenerate the head or the 
 posterior sucker. The lack of the capacity for 
 regeneration in leeches is paralleled by the 
 similar behaviour of Amphioxus. Leeches offer 
 as strong a contrast to other Annelids in this 
 respect as Amphioxus does to the Entero- 
 pneusta. 
 
 The loss of the posterior sucker puts an end 
 to the looping gait but not to the effort to 
 achieve it, and the lack of adhesion behind is 
 compensated by peristalsis on a level surface, 
 and by the great power of extension possessed 
 by the forebody when climbing up a vertical 
 surface. On placing two leeches on a stone in 
 the middle of a large dish of water they remained 
 motionless in a sub - erect attitude, resting upon 
 the hind-sucker for some time, and then descended 
 into the water, looping along the bottom to the 
 side of the vessel. When watching them by 
 candle-light they would not venture across until 
 the light was removed. 
 
 1 Quoted in the Cambridge Natural History -, vol. ii., 1896, 
 "Earthworms and Leeches," by F. E. Beddard, p. 408. 
 
CEYLON LAND LEECH 47 
 
 Whilst sucking, abundant watery fluid issues 
 from the epidermal glands of the Ceylon land 
 leech, so copious that it may flow off the arm, 
 upon which it happens to be feeding, like water. 
 The glands occur on epidermal papillae which 
 are retractile, and can be protruded simultaneously 
 like a flush passing over the body. No spilling 
 of blood takes place normally ; but if some be 
 taken and mounted in a drop of the epidermal 
 fluid, the red corpuscles become aggregated 
 together to form the characteristic rouleaux, 
 whilst retaining their shape and healthy appear- 
 ance. The fluid is perfectly clear, like the 
 aqueous humour of the vertebrate eye, and forms 
 an excellent medium for the examination of fresh 
 blood. 
 
 In the Japanese land leech a similar fluid is 
 stated to issue from the nephridial pores. I have 
 not seen this in Ceylon ; but when a feeding leech 
 is turgid with blood, the nephridiopores become 
 clearly visible, fourteen pairs situated in the lower 
 half of the lateral pale bands at the posterior 
 borders of the segments, the space of five annuli 
 intervening between successive pores. There 
 are three segments or fifteen annuli in front of 
 the first pore ; the posterior, nuchal eyes appear 
 to lie on the second annulus. The erectile 
 papillae are arranged in a single row across each 
 annulus. Certain of these papillae have a special 
 character and occur in longitudinal rows, lying 
 
48 FREE AND FIXED ANIMALS 
 
 singly in the first annulus of each segment (i.e., 
 in the annulus following that which carries the 
 segmental nephridiopores) ; each shows a dark 
 centre surrounded by a pale border on the top 
 of a retractile papilla. These segmental ocelli- 
 form organs commence in the segment in front 
 of the first nephridiopore, and there are twelve 
 of them in each segment, i.e., six on each side 
 disposed as follows : two lateral rows, one above 
 and one below the lateral band, a dorso-lateral 
 and a ventro - lateral row, a submedian dorsal 
 and a submedian ventral row. 
 
 The preceding notes serve to illustrate, in 
 respect of external characters alone, the highly 
 specialised organisation of the land leech, which 
 is already indicated by the concentration of its 
 stereotropic mechanism, by its lack of regenera- 
 tion, and by its passion for blood which amounts 
 to a hsematotropism, as no one can doubt who 
 has read Sir Emerson Tennent's account of it, 
 or who has experienced it in the flesh. Hardly 
 anything proclaims a finished organisation, the 
 culmination of a phyletic career, so plainly as 
 an exclusive diet. In point of fact the leeches 
 represent the culmination at the present age of 
 the Annelid branch in a definite direction. 
 Branches of the tree of life often terminate in 
 a brush of points, not necessarily in a single 
 point ; and leeches constitute one of these 
 terminations, 
 
TREE SNAKES 49 
 
 Few free-living animals except terrestrial 
 snakes, are so exquisitely stereotropic as are 
 leeches ; and snakes offer a further analogy 
 with Annelids as a whole in the combination 
 of stereotropism and pleotropism in some fresh- 
 water species, stereotropism alone in burrowing 
 and in arboreal species, pleotropism alone in 
 many marine species. In some tree snakes, e.g., 
 Dendrophis, Dendrelaphis, and Chrysopelea, there 
 is a lateral keel or suture on each side of the 
 ventral shields. It has been suggested by Mr 
 R. Shelford 1 that this carination or hinging 01 
 the ventral shields, by allowing the body to 
 flatten itself out and become concave below, 
 enables the snake, when springing from a height, 
 to descend gently like a parachute. Against this 
 it has to be noted that burrowing snakes like 
 Cylindrophis, which have very small ventral 
 shields, are able to flatten the body in a 
 marvellous manner. I rather think that the 
 lateral hinge-lines are in the first instance an 
 accessory stereotropic contrivance enabling the 
 snakes to climb up rough tree-trunks with greater 
 facility; moreover, a lateral angulation of the 
 ventral shields is found in the terrestrial genera, 
 Lycodon and Hydrophobus. Nevertheless, the 
 possibility is not excluded that they might also 
 serve on occasion for the parachute flight ; but 
 
 1 R. Shelford. A note on "Flying" Snakes, P. Zool. Soc.> 
 London, 1906, i., pp. 227-230. 
 
 P 
 
50 FREE AND FIXED ANIMALS 
 
 that the latter ever really takes place is a matter 
 which requires confirmation. 
 
 Other reptiles which take refuge in trees, e.g., 
 Calotes and Varanus, when alarmed, will project 
 themselves from a great height to the ground 
 without elaborating a mechanism for aerial flight. 
 The case of Draco, the flying lizard of Malaya, 
 where such a mechanism is developed (dermal 
 membrane supported by ribs), belongs to a 
 different category ; and the facts seem to show 
 clearly that it is not merely the habit of taking 
 flying leaps, like monkeys, for example, that has 
 led to the formation of organs of flight. 
 
 We may now submit the conclusion that just 
 as all divisions of the animal kingdom display 
 a cryptozoic bias, so do they all show a statozoic 
 tendency. The sedentary habit is referable to 
 a stereotropic basis, and pelagic or pleotropic 
 forms belonging to groups which are pre- 
 dominantly sedentary, have had a probable 
 statozoic origin. All land animals have had 
 aquatic ancestors, or, more precisely, all air- 
 breathers have descended from water-breathers 
 as defined above. The latter are primarily 
 aquatic, and are either stereotropic at the present 
 time, or it may be argued that they have had 
 a more or less remote stereotropic ancestry. 
 
 Thus the permanent fixation of so many 
 aquatic animals is not such a bizarre phenomenon 
 as it appeared to be to the older naturalists 
 
COMMON PROPERTIES 51 
 
 of the eighteenth century, but it is merely 
 an extreme manifestation of a very primitive 
 property of animal life. My coupling of the 
 cryptozoic habit with the stereotropic reaction 
 is incidentally justified by a statement of Eisig 
 to the effect that stereotropism signifies rest 
 and concealment. The purpose of this and the 
 preceding chapter has been to insist upon the 
 fundamental nature of cryptotaxis and stereo- 
 tropism in the convergent evolution of animals. 
 These are some of the primary properties of 
 living matter upon which the moulding forces 
 of nature have been at work for untold ages. 
 
 A somewhat analogous case, in so far as it is 
 an extreme manifestation of a common property, 
 is the electric power of the three typical kinds 
 of electric fishes (ray, eel of South American 
 rivers, and catfish of African rivers). As shown 
 by E. du Bois-Reymond, 1 these fishes owe their 
 capacity for imparting severe shocks, in different 
 directions, to an intensification of the common 
 electromotive properties of nerve and muscle. 
 In each case, as pointed out by Dr D. S. Jordan, 2 
 closely - related species show no trace of the 
 electric endowment. 
 
 1 See "Biological Memoirs," Oxford, 1887, vol. i. 
 
 2 "Guide to the Study of Fishes," 1905. 
 
CHAPTER V 
 
 MIMICRY AND HOMOPLASY 
 
 THE term convergence is applied to resemblances 
 amongst animals which are not due to direct 
 relationship or genetic affinity ; in other words, 
 which are not derived by inheritance from 
 common ancestors, but which result from inde- 
 pendent functional adaptation to similar ends : 
 e.g., the exuviation and regeneration of newts 
 as compared with the same phenomena in crabs 
 and lobsters, or again the coiled prehensile tail 
 and great rolling eyes, moving independently, 
 of Hippocampus and Chameleons. 1 
 
 It has long been recognised as a dominant 
 factor in comparative morphogeny. Common 
 characters of adaptation in different animals 
 may be due to common inheritance, or they 
 may be due to convergence. Hence it follows, 
 in Semper's words, that the problem of the 
 morphologist is to learn to distinguish such 
 
 1 These two striking instances of double convergence were 
 presented in respective juxtaposition by Mr Frank Buckland. 
 ("Curiosities of Natural History." Reprinted 1903, Macmillan, 
 First and Second Series.) Compare the case of the egg-laying, 
 duck-billed Platypus of Australia. 
 
 5 2 
 
WAYS OF CONVERGENCE 53 
 
 characters as have been developed by adapta- 
 tion, quite independently of the affinities of 
 animals, from such as have been transmitted by 
 inheritance through a long series of generations. 1 
 
 Morphology therefore naturally falls into two 
 divisions : convergent morphogeny and homo- 
 genetic or normal morphogeny. The former is 
 not generally admitted as a legitimate branch 
 of positive morphology ; the latter is the way 
 usually followed, and common characters which 
 are not homologous are sometimes attributed 
 to "mere convergence." It appears that there 
 is more joy amongst morphologists over one 
 attempt at genealogy than over ninety and nine 
 demonstrations of convergence. Personally, I 
 take the view that both of the above-named 
 divisions of morphology are equally important. 
 
 There are several ways of convergence and 
 each is expressed in various degrees of intensity. 
 The two most widely diffused ways of con- 
 vergence are known respectively as homoplasy 2 
 and mimicry. 
 
 True mimicry is defined by Mr Alfred Russel 
 Wallace 3 as "a form of protective resemblance 
 in which one species so closely resembles 
 another in external form and colouring as to 
 
 1 Karl Semper, " The Natural Conditions of Existence as they 
 affect Animal Life." Fourth edition, London, 1890. 
 
 2 The term and idea of homoplasy were introduced by Sir E. 
 Ray Lankester : " On the Use of the Term Homology in Modern 
 Zoology," Ann. Nat. Hist., 1870. 
 
 3 A. R. Wallace, " Darwinism." Second edition, London, 1890. 
 
 D 2 
 
54 MIMICRY AND HOMOPLASY 
 
 be mistaken for it, although the two may not 
 be really allied and often belong to distinct 
 families or orders." The first constant condition 
 under which true mimicry occurs is "that the 
 imitative species occur in the same area and 
 occupy the very same station as the imitated." 
 The last essential condition of mimicry, as 
 explained by Mr Wallace, is "that the imita- 
 tion, however minute, is external and visible 
 only, never extending to internal characters or 
 to such as do not affect the external appearance." 
 Furthermore, it is explained that warning colora- 
 tion is the basis of the phenomenon of mimicry, 
 but it is not stated what lies at the root of 
 warning coloration. Of course it is fostered by 
 natural selection, but we do not know the funda- 
 mental reaction that determines it. Professor 
 Poulton 1 adds as another attribute of true 
 mimicry that it is a sham, that is to say, a 
 false alarm ; true warning colours, occurring in 
 unpalatable or dangerous bodies, are genuine. 
 
 Warning coloration, mimicry, protective re- 
 semblance, and the death feint, are facts in 
 nature which are known to have been doubted 
 in the past by "armchair philosophers." This 
 has doubtless been due to the fact that some 
 delicate shade of meaning has often been over- 
 looked in the original descriptions. Even if 
 
 1 E. B. Poulton, "Colours of Animals," Encyc. Brit., ninth 
 edition, Supplement. 
 
MOMENTS OF CONVERGENCE 55 
 
 both form and colour of a mimic harmonise 
 with its model, there must still be some slight 
 movement, some particular action or habit to 
 complete the illusion. Thus, to take one example 
 only as the type of a long series, there occurs 
 in Ceylon a longicorn beetle, Callichroma chryso- 
 gaster, sufficiently rare to satisfy Wallace's con- 
 dition that the imitators are always less numerous 
 than the imitated, which shows characteristic 
 waspish marks, orange yellow antennae and legs 
 set upon a body uniformly bluish black above ; 
 but it only resembles a wasp when it is in flight 
 and at the moment of alighting on a bush in full 
 sunlight; when pinned in a box it is a "mere 
 beetle." 
 
 Amongst insects especially there are many 
 examples of forms belonging to different orders, 
 protected by different noxious secretions, and 
 yet possessing common warning colours, which 
 do not occupy identical stations. Thus a Reduviid 
 bug living under dry bark may resemble a Bom- 
 bardier beetle living under damp logs. This 
 belongs to Poulton's category of synaposematic 
 resemblances "such as obtain between animals 
 of all kinds adopting sematic [signalling] methods 
 of defence in the same country " ; this is some- 
 times known as Miillerian mimicry, as dis- 
 tinguished from true or Batesian mimicry. The 
 common cryptic or protective coloration of many 
 animals (oceanic, deserticolous, lichenicolous) in 
 
56 MIMICRY AND HOMOPLASY 
 
 a special environment produces likenesses among 
 different animals which are incidental and ana- 
 logous to those convergent resemblances "caused 
 by functional adaptation, such as the mole-like 
 forms produced in the burrowing Insectivora, 
 Rodentia, and Marsupialia" [Poulton]. 
 
 We may safely claim that the possession by 
 noxious animals of common warning coloration is 
 as much due to convergence as is the possession 
 by harmless animals of a common protective 
 coloration ; and both of these colour - schemes 
 are referable to conceivable though indefinite 
 reactions. On the other hand, the resemblances 
 and associations between palatable and unpalat- 
 able insects are hard to explain on the tropism 
 theory, unless we suppose that they arose by 
 ordinary convergence before advantage was taken 
 of them by natural selection. We may there- 
 fore differentiate usefully between passive or con- 
 vergent mimicry (including Poulton's syncryptic 
 and synaposematic categories) and active or 
 selective mimicry (including true mimetic re- 
 semblances) ; and we may add that while in 
 typical cases the categories are sharply divided, 
 in less obvious cases it is often as difficult to 
 separate pure convergence from either form of 
 mimicry as it is to distinguish it from homology ; 
 the temptation being to describe the more or 
 less sensational instances and to ignore those 
 which are less convincing, thereby imparting a 
 
PHYSIOGNOMY 57 
 
 one-sided and exaggerated impression to a very 
 common natural phenomenon. 
 
 Mimicry, in the wide sense of the term, in- 
 volves the entire superficial aspect of the body ; 
 the moment the details are analysed the like- 
 ness disappears ; it is absolutely essential to its 
 success that structural details should be disre- 
 garded. We may define mimicry broadly as a 
 physiognomical convergence between two or 
 more species of animals, in this way distinguish- 
 ing it from homoplasy, which depends upon a 
 more deep - seated organic or structural con- 
 vergence. Such coincidences of form and 
 function as those presented by the mole - like 
 burrowing mammals referred to above, by other 
 marsupial and placental mammals, and by some 
 arboreal mammals, e.g., Tupaiidse (tree-shrews) 
 and Sciuridae (squirrels), the flying Insectivore 
 Galeopithecus and the flying squirrel Pteromys, 
 in so far as the whole bodily appearance is 
 involved, afford examples of a third way of 
 convergence known as parallel evolution ; it will 
 soon be clear how convergence is not incom- 
 patible with parallelism. They resemble each 
 other because of their approximation to funda- 
 mental forms, e.g., the burrowing form, the 
 arboreal form, and so on. A carnivorous Mar- 
 supial must necessarily bear some resemblance 
 to a true Carnivore, the fact being that it is* 
 an extraordinarily close one. 
 
58 MIMICRY AND HOMOPLASY 
 
 There are very many examples, chiefly amongst 
 insects and other arthropods, at all stages of 
 growth and metamorphosis, of resemblances to 
 fragments of vegetation, excrescences on bark, 
 droppings, etc. It is usually not difficult to dis- 
 criminate between effective resemblances which 
 are objective, and fanciful resemblances which 
 are subjective. With the latter I have nothing 
 to do here, although they are sometimes of such 
 a nature as to have given rise to enduring popular 
 traditions, and on that account are entitled to 
 respect within their own scope. 
 
 The resemblance to a common fundamental 
 form is well illustrated by the leaf-mimics, of 
 which the best known are the Leaf Butterflies 
 and the Leaf Insects. Mr Wallace remarks 
 that " many butterflies, in all parts of the world, 
 resemble dead leaves on their under side, but 
 those in which this form of protection is carried 
 to the greatest perfection are the species of the 
 Eastern genus Kallima" This genus of butter- 
 flies is noted for the extreme amount of indi- 
 vidual variation in the markings on the under 
 side of the wings, simulating all degrees of decay 
 and discoloration and fungus-attack ; to this must 
 be added the perfect leaf-shape and veining of 
 the closed wings. Besides all this, its method 
 of alighting is such as to complete the illusion. 
 Referring in particular to the Sumatran species 
 Wallace says that "this is effected by the butter- 
 
LEAF BUTTERFLIES 59 
 
 fly always settling on a twig, with the short tail 
 of the hind-wings just touching it and forming 
 the leaf -stalk." In Sumatra he has often seen 
 one enter a bush and then disappear like magic. 
 In this account one may wonder why the butter- 
 fly resembles a dead leaf seeing that it rests upon 
 a twig in the attitude of a normal leaf; perhaps 
 it may be explained by the circumstance that 
 in tropical vegetation a regular leaf - fall is the 
 exception instead of being the rule as it is in 
 temperate climates, and it is a very common thing 
 to see a few conspicuously discoloured leaves still 
 attached to the branches in the midst of green 
 foliage. 
 
 The Kallima philarchus of Ceylon seems to 
 behave rather differently since, as Mr E. E. 
 Green has noticed and recorded, it " more usually 
 settles head downwards on the trunk of a tree, 
 . . . swaying gently from side to side. It might 
 then be mistaken very easily for a detached 
 leaf that in its fall has hitched up in a cobweb 
 and is being shaken by the breeze." 1 Another 
 observer, Mr W. A. Cave, 2 reports that he had 
 exceptional opportunities for watching one of 
 these butterflies at close quarters, as it frequented 
 a certain spot on the trunk of a tree for two days ; 
 as soon as it settled, " which it did in the usual 
 way, it immediately turned round so that its 
 head pointed downwards." Mr Cave was much 
 
 1 E. E. Green, "Mimicry in Insect Life, as exemplified by 
 Ceylon Insects," Spolia Zeylanica, vol. v., 1908, see p. 89. 
 * W. A. Cave, Note on Kallima philarchus , t. <:., p. 142. 
 
60 MIMICRY AND HOMOPLASY 
 
 puzzled over this proceeding, which occurred every 
 time the butterfly settled, until he discovered the 
 reason: "the butterfly turned round so that the 
 tail of its two hind wings would almost come into 
 contact with the trunk of the tree, thus representing 
 a stalk, and the apparently dead leaf would hang 
 in a perfectly natural way, drooping downwards." 
 
 A Mexican leaf butterfly (Tagetis mermeria), 
 according to Mr C. W. Beebe, behaves in such 
 a manner as to cause itself to be mistaken 
 at first glance, not only for a fallen but for a 
 falling leaf as well Here also the individual 
 variation is very great, in correspondence with 
 the "variety of hues and mottlings which exist 
 among dead and withered leaves. . . . When 
 the insect took to wing it shot almost straight 
 upward, and instantly attained the highest point 
 of its flight. From here to its place of alighting 
 its course was a gradual descent this living leaf 
 unconsciously reflecting every detail of the fall 
 of the withered bits of vegetation. And further, 
 when the butterfly alighted, it was not with a 
 fluttering and a few moments of hovering, but 
 as a leaf comes to rest, so the insect a sudden 
 drop to the very ground, wings snapped together, 
 and the apparently dried, worm-eaten leaf leaned 
 far over to one side and swayed with every 
 breath of air." 1 
 
 It is interesting to remark that a notion exists 
 to the effect that the constant repetition of such 
 
 1 C. William Beebe, "Two Bird- Lovers in Mexico," Boston 
 and New York, 1905 ; see pp. 241-243. 
 
LEAF INSECTS 61 
 
 considerable variations as are met with in leaf- 
 like Lepidoptera and Orthoptera, from genera- 
 tion to generation, is a standing witness against 
 the truth of " Darwinism," inasmuch as, accord- 
 ing to the Darwinian theory, such variations 
 ought either to become fixed by natural selection 
 or swamped by interbreeding. On the face of 
 it there would seem to be some force in this 
 argument. Whatever the answer may be, it is 
 not that these varieties may become fixed in 
 course of time ; on the other hand, it may be that 
 natural selection is interested in keeping alive the 
 variations for the benefit of the species, not for 
 the production of new species. In any case it is 
 a good point, worthy of special consideration. 
 
 Butterflies owe their leaf-like appearance to 
 the effect of bilateral compression when the 
 wings are closed. Leaf insects of the genus 
 Phyllium, which belong to the order Orthoptera 
 (containing also the cockroaches, grasshoppers, 
 and stick-insects), owe it to the dorsiventral 
 flattening of the body. A similar antagonistic 
 flattening of the body occurs between two 
 human ectoparasites where the compression 
 also achieves a cryptic object indirectly, viz. : 
 the bilaterally flattened Pulex and the dorsiven- 
 t rally flattened Cimex. 
 
 An excellent description and plate of Phy Ilium 
 crurifolium, a wonderful leaf insect which is 
 found in the Seychelles islands as well as in 
 Ceylon, has been published recently by Mr 
 
62 MIMICRY AND HOMOPLASY 
 
 H. S. Leigh, 1 whose account, based on material 
 obtained from the Seychelles and bred in 
 England, I can confirm, having reared them in 
 Ceylon for several years. The body of the 
 males is generally green with a pair of clear 
 ocellations on the fourth abdominal segment ; 
 they vary very little in colour ; they have small 
 tegmina and large wings, and are capable of 
 flight ; their legs break off by autotomy with 
 the greatest ease. The flightless females, on 
 the contrary, appear in two principal varieties, 
 green and red (ferruginous), the green indi- 
 viduals predominating over the russet, but I 
 have been unable to determine any constant 
 numerical relation between them. The russet 
 forms are more variegated and show more 
 individual variation than the green forms. Out 
 of one hundred and eight individuals reared 
 from eggs laid by green females, six russet 
 forms appeared after the later moults ; out of 
 seventy-five individuals reared from eggs laid 
 by two russet females, one russet form appeared. 
 The female leaf insects therefore resemble 
 both green and sere leaves in about the same 
 proportion as the latter occur on green trees in 
 the tropics. Not only this, but when immature 
 they can fold up their abdomen like a curled 
 leaf; and at all stages they sometimes sway like 
 leaves in a breeze, and at other times remain 
 
 1 H. S. Leigh, " Preliminary Account of the Life-History of the 
 Leaf Insect, Phyllium crurifolium Serville," P. Zool. Soc.^ London, 
 1909, pp. 103-113. 
 
AUTOTOMY 63 
 
 rigid with fore-limbs extended like leaves in a 
 calm. The adult females, as duly noted by 
 Leigh, are more sluggish than the males ; and 
 perhaps connected with this circumstance is 
 their much lower capacity for autotomy of the 
 appendages. A female can be held by the legs 
 without discarding them, but if a male be so 
 held mutilation ensues. The eggs of PhylKum, 
 as of all Phasmidae, resemble plant seeds. 1 
 
 I may mention here an observation bearing 
 upon the autotomy of the wings of termites. 
 When they alight upon the ground after swarm- 
 ing out of the nest, it is well known that they 
 throw off their wings by a special act of autotomy; 
 but if seized by the wings before the performance 
 of this act, they cannot accomplish it so long as 
 the wings are held. It is the same when only one 
 wing is taken between the fingers ; the autotomy 
 is inhibited. The casting of the wings by termites 
 and by ants is of course not followed by a 
 corresponding act of regeneration, such as follows 
 upon the autotomy of the legs of immature 
 phasmids ; mature male phasmids, as we have seen 
 in the case of the leaf insect, cast their legs but 
 will not live long enough to regenerate them. 
 
 Amongst fishes the Australian " Sea-horse " 
 
 1 The newly-hatched young of Phyllium crurifolium present a 
 warning scheme of coloration which disappears with the first 
 moult, blackish head and legs and a scarlet abdomen. I brought 
 a quantity of eggs recently from Ceylon to England, some of 
 which hatched out during the voyage, others later in the Zoological 
 Gardens. On showing them fresh out of the egg to my friend Mr 
 R. I. Pocock, he remarked instantly upon their striking resemblance 
 to certain plant bugs. 
 
64 MIMICRY AND HOMOPLASY 
 
 Phyllopteryx, of which there is a figure in Dr 
 Giinther's " Study of Fishes " (1880), is furnished 
 with cutaneous processes which resemble trailing 
 seaweed. Another fish, during a limited period 
 of its life-history, namely, about the yearling 
 stage, effectively resembles a dead water-logged 
 leaf. This is the young of the so-called sea- 
 bat, Plat ax vesper tilio? The mature fish is 
 nearly uniformly dark-coloured ; it attains a 
 moderate size and is very high in the body, 
 which is strongly compressed. It swims about 
 in small shoals and has the habit, possessed by 
 many other similarly constituted fishes, of turning 
 over on one side momentarily whilst swimming. 
 In the young stage, when the total height of 
 body and fins is about two or three inches, 
 the colour is pale yellowish or brownish with 
 variegated, irregular and variable markings. 
 When seen living in the water, its resemblance 
 to a sere leaf is extraordinary ; when taken out 
 of water it loses something of the strangeness 
 of its appearance, and when preserved, the fins 
 collapse, thereby destroying the living aspect of 
 the animal. 
 
 In February 1904 I had the opportunity of 
 seeing the peculiar movements of a young Platax 
 under normal conditions, off the west coast of 
 Ceylon. Whilst a fisherman was attempting to 
 catch it with a pole-net, it suddenly toppled over 
 
 1 A. Willey, " Note on Leaf-Mimicry," Spolia Zeylanica, vol. ii., 
 1904, pp* 51-55. See also Nature^ vol. Ixxx,, 1909, p. 247. 
 
LEAF FISHES 65 
 
 and sank gently and inertly to the bottom like 
 a yellow leaf, for which I mistook it at first. As 
 I was about to turn away from such a common- 
 place object as a drenched leaf, it righted itself 
 once more and darted away. It was subsequently 
 captured and sketched. 
 
 Fallen leaves of various kinds are commonly 
 seen drifting about in the sea, sometimes at a 
 considerable distance offshore. The most familiar 
 examples of such drifting leaves are those of the 
 mangrove trees which fringe the borders of 
 estuaries and backwaters in the tropics. This 
 singular Leaf Fish, besides having the requisite 
 shape and colour, has also the power of assuming 
 the inertia of a dead leaf, or of feigning death 
 in the sea, thus completing the disguise. The 
 contour of its body is strengthened behind by 
 a line of dark pigment, which passes along the 
 hinder margins of the dorsal and anal fins, but 
 through the base of the caudal fin ; the latter is 
 unpigmented, hyaline and invisible under water. 
 The pectoral fins are also transparent, but the 
 elongated ventral fins are opaque, showing the 
 yellow ground colour and a dense outer border 
 conterminous with the border of the anal fin. 
 There can be no mistake as to the effective- 
 ness of this case of leaf-mimicry under water; 
 moreover, the surface of the body shows lines 
 of pigment and small spots such as are seen 
 in a decaying leaf (Fig. i). 
 
 The three different types Phyllium, Kallima, 
 
 E 
 
66 
 
 MIMICRY AND HOMOPLASY 
 
 and Platax, which have so little in common 
 otherwise, agree in exhibiting the shape of a 
 
 FiG. I. The Leaf Fish or yearling stage of Platax 
 vespertilio. Original from Spolia Zeylanica. 
 
 simple bifacial leaf, one of the commonest 
 objects, dead or alive, ashore or afloat, in nature. 
 Yet they do not effectively resemble each other 
 except in an abstract manner by conforming to 
 a fundamental pattern, i.e., by convergence. 
 The death feint is a common and very widely 
 
COINCIDENCE AND CORRELATION 67 
 
 spread method of protection from predatory 
 enemies. It is an extreme manifestation of the 
 virtue of immobility which is equally well known 
 to hunters and to their quarry. When it is 
 associated with singularities of form and colora- 
 tion, it produces a combination which is so 
 remarkable that it is generally received with a 
 proper amount of scepticism. The resemblance 
 which these animals bear to a leaf is a selective 
 one ; that which they bear incidentally to each 
 other is a convergent one, due to their inde- 
 pendent acquirement of a common fundamental 
 form. In these cases it may be supposed that 
 natural selection has taken advantage of a 
 previously existing aptitude, the essence of the 
 leaf-form being present as a generic character 
 before the actual assumption of the intimate 
 resemblance became fixed. 
 
 Other striking convergences depending upon 
 form and coloration are those which obtain 
 between some harmless and some poisonous 
 snakes inhabiting the tropical regions of both 
 hemispheres. Darwin described the examples 
 from South America. The mutual resemblances 
 proceed largely from the character of the trans- 
 verse banding of the body which they have in 
 common ; and the colour coincidence may be 
 so close that in a rapid sorting the species 
 would be thrown together and confounded one 
 with another. 
 
 Lastly, reference should be made to the well- 
 
68 MIMICRY AND HOMOPLASY 
 
 known cases of plumage resemblances amongst 
 birds and to the longitudinal striping of some 
 mammals, e.g., young wild pig, young tapir, 
 mouse-deer, striped squirrels, etc. In all these 
 cases convergence is independent of selection. 
 
 In order to guard against conveying a wrong 
 impression, it may be well to add that in cases 
 of double or multiple convergence such as are 
 mentioned on p. 52, there is no a priori pre- 
 sumption of correlation between the several 
 characters which are repeated in the different 
 types. Thus there is no correlation between 
 the separately moving eyes of Hippocampus and 
 its prehensile tail, because other Lophobranchiate 
 fishes (Pipe-fishes), which have a straight tail, roll 
 their eyes in the same way, as can be seen very 
 prettily in the Naples Aquarium. The property 
 which these fishes present in common with 
 chameleons is the possession of exceptionally 
 sluggish habits ; and the extreme mobility of 
 the eyes compensates for the relative inflexibility 
 of the body. Nor is there any correlation 
 between the oviposition of the Platypus (Ornitho- 
 rhynchus) and the shape of its bill ; the latter is 
 correlated with its aquatic and feeding habits. 
 The conjunction of characters may be a coin- 
 cidence ; the repetition of the coincidence is an 
 act of convergence. 1 
 
 1 For a discussion of correlation between teeth and limbs in 
 the evolution of Mammalia see Professor H. F. Osborn's paper 
 on "The Law of Adaptive Radiation," American Naturalist^ 
 vol. xxxvi., 1902, pp. 353-363. 
 
CHAPTER VI 
 
 DIVERGENCE AND PARALLELISM 
 
 FEW things are more astonishing at first acquaint- 
 ance than examples of close convergence, whether 
 mimetic or homoplastic, because they appear para- 
 doxical and contrary to expectation. In fact the 
 ways of convergence are devious, many-hued, 
 illuminating, full of surprises, inviting away from 
 the narrow path of homology which, at least in 
 extreme cases, begins and ends in obscurity. It 
 must, however, be acknowledged that the pioneer 
 work which effected the discovery of the pelagic 
 larvae of so many marine animals about the 
 middle of the nineteenth century, was thoroughly 
 satisfactory, and seemed to indicate to a later 
 generation that the pursuit of homology as an 
 aid to phylogeny was not wholly elusive. If 
 there has been any mistake it has been that of 
 not defining the different degrees of homology. 
 But there is no necessity to dwell upon mistakes 
 nor to regret them. 
 
 As a seeming paradox there comes the pre- 
 liminary axiom that convergence depends firstly 
 on divergence and secondly on parallelism. As 
 
 69 E 2 
 
DIVERGENCE AND PARALLELISM 
 
 we ascend higher in the scale of nature we find 
 that the co-ordinating mechanism of the animal 
 body keeps pace with the rest of the organisation, 
 and the leading members of the several phyla or 
 branches of the animal kingdom become ever 
 farther removed from each other and from what 
 
 FlG. 2. Diagram of phyletic divergence and parallelism. 
 
 we may designate vaguely as their common 
 starting-point. Thus the highest Mollusca, the 
 Cephalopoda (squids, octopods, cuttle-fishes, and 
 Nautilus), which are characterised by a high 
 degree of cephalisation or concentration and 
 
PARALLEL DIVISIONS 71 
 
 amalgamation of parts to form the head, repre- 
 sent the maximum organic development within 
 the Molluscan phylum, and are therefore farther 
 removed from the highest mammals than is the 
 lowly Chiton, because the Mollusca and the 
 Vertebrata have been advancing along inde- 
 pendent parallel lines, and the cephalisation, with 
 its implied brain-power, is an act of convergence. 
 
 FIG. 3. Parallel divisions of Echinodermata. 
 
 The Molluscan and Vertebrate phyla, and 
 all the phyla, although their actual origins are 
 lost in Silurian darkness, may or must be sup- 
 posed to have diverged from a common base 
 in the first instance. This point is illustrated 
 graphically in Fig. 2 ; and the same principle of 
 primary divergence and secondary parallelism will 
 apply to each phylum (Figs. 3 and 4). Nothing 
 
72 DIVERGENCE AND PARALLELISM 
 
 in phylogeny seems more firmly established 
 than the pelmatozoic ancestry of the Echino- 
 derms (star-fishes, sea-urchins, etc.), nor than 
 the bilaterally symmetrical ancestry of the 
 Pelmatozoa, the former deduction largely result- 
 ing from their paleontology, the latter confirmed 
 by their embryology. 1 
 
 As for the Vertebrata, not all of them possess 
 vertebrae, but all have a notochord. The term 
 
 /n. 
 
 FIG. 4. Parallel divisions of Vertebrata. 
 
 Protochordata, with the subsidiary terms Uro- 
 chordata (for the Tunicata or Ascidians) and 
 Cephalochordata (for the Acrania or Amphioxus), 
 was introduced by Balfour in his "Comparative 
 Embryology" in 1882. Subsequently in 1884 
 Bateson suggested the additional term Hemi- 
 chordata for the Enteropneusta (" Balano- 
 
 1 C/.E.W. MacBride, " Echinodermata " in Cambridge Natural 
 History, vol. i., 1906. 
 
MECHANISM OF NUTRITION 73 
 
 glossus "). Dr Gaskell l has reversed the position 
 by the recent statement that a large number 
 of zoologists, who consider the notochord as a 
 great characteristic of the vertebrate organisa- 
 tion, have followed Bateson in the matter of 
 terminology. 
 
 Taken as a whole the habits of animals are 
 cast into certain well-defined moulds, and it is 
 quite easy to understand why this should be 
 so ; but the much more apparent than real 
 simplicity of the matter is no reason why it 
 should not be stated at length. Animals differ 
 from plants in their nutrition, mobility, and 
 sensibility. The manner in which the nutrition 
 of animals differs from that of green plants is 
 sharp and unequivocal, so much so that it can 
 be expressed in two contrasting terms, holozoic 
 and holophytic, between which extremes, how- 
 ever, all grades of what is known as mixotrophism 
 (mixed dietary) are to be found, chiefly but 
 not entirely amongst the lower orders. 
 
 Green plants alone can elaborate their own 
 food, with the help of the sun, from inorganic 
 substances ; animals require their food ready 
 made before they can begin to assimilate it. 
 Plants have only to assimilate, and that in one 
 way ; animals have first to procure their food, 
 and that in many ways. And yet there is as 
 
 1 W. H. Gaskell, "Origin of Vertebrates," London, 1908, set 
 p. 1 6. 
 
74 DIVERGENCE AND PARALLELISM 
 
 much relative diversity amongst plants as 
 amongst animals, showing that differentiation 
 proceeds from within though it may be modified 
 from without. The methods adopted for securing 
 food determine in great measure the habits of 
 animals. Their instincts and intelligence are 
 concentrated upon their food-supply, reproduc- 
 tion being automatic and all besides incidental. 
 It is small wonder then that many animals 
 follow similar methods ; and in this way habitual 
 convergence is brought about. Thus the car- 
 nivorous, insectivorous, herbivorous, and fru- 
 givorous habits are widely disseminated ; and 
 community of habit and experience is often 
 accompanied by greater or less similarity of 
 structure of certain parts, giving rise to a two- 
 fold exhibition of convergence, bionomical and 
 homoplastic. 
 
 Next to the function of nutrition we have to 
 consider the organs which perform that function, 
 i.e., the mechanism of nutrition. For the sake of 
 brevity we may concentrate our attention upon 
 the higher or leaf-bearing plants and the higher 
 or gut-bearing animals. Now some parasitic 
 plants, the so-called root-parasites (Balanophora, 
 Rafflesia), can manage without green leaves, 
 and some parasitic animals, both endoparasites 
 (Cestoda and Acanthocephala) and ectoparasites 
 (Sacculina), can dispense with a gut. When, 
 however, as in the great majority of vascular 
 
HOMOLOGY OF THE GUT 75 
 
 plants, true foliage leaves occur, they are homo- 
 logous throughout. Sometimes other structures 
 (cladodes and phyllodes) can function as leaves, 
 i.e., as organs of assimilation ; but the special 
 homologies of certain leaf-like structures do not 
 affect the general homology of leaves, which is 
 one of the cardinal points in vegetable mor- 
 phology. The distinction between the special 
 homology of parts and the general homology 
 of the whole requires to be emphasised. 
 
 The special morphology of the alimentary tract 
 of animals is a large subject which demands 
 exhaustive analysis ; but it is distinct from the 
 question of the general homology of the gut. It 
 is probable that no single division of an insect's 
 gut, for example, can be strictly homologised 
 with a corresponding part of a fish's gut, but 
 looked at in their entirety, the general homology 
 of the one with the other is still a question 
 apart, to be answered by an appeal to first prin- 
 ciples. That it must be answered one way or 
 the other is evident from the fact that objections 
 have been raised against the alleged view "that 
 the one organ which is homologous throughout 
 the animal kingdom is the gut," [Gaskell, /..]. 
 This is perhaps an extreme way of putting it, 
 but we may examine the point very briefly, 
 always bearing in mind the distinction between 
 special and general homology. 
 
 Dr Gaskell regards the nervous system as 
 
76 DIVERGENCE AND PARALLELISM 
 
 constituting the master tissue of the animal 
 body, and there is no doubt that he is right 
 up to a certain point. The behaviour of 
 Protozoa and Sponges in this regard, however, 
 teaches us that in the course of evolution the 
 nervous mechanism is secondary as compared 
 with the mechanism of locomotion on the one 
 hand, and nutrition on the other. An obvious 
 rejoinder to this argumentation might be that 
 the organisation of Protozoa and Sponges also 
 teaches us that the gut, as an organ, is secondary 
 to nutrition as a function. That this is so is 
 borne out by the fact, referred to above, that 
 just as in plants certain structures which are not 
 leaves have secondarily acquired the function 
 of leaves, so amongst animals the parasitic 
 Cestode and Acanthocephalous worms, which 
 have secondarily lost the use and presence of 
 a gut as a consequence of their mode of life, 
 absorb fluid nutriment, prepared by the host, 
 through the outer surface of the body. Here 
 therefore the body-wall does duty for the gut- 
 wall, combining the somatic and splanchnic 
 functions into one. 
 
 We are therefore driven to enquire whether 
 the general homology of the gut of triploblastic 
 (three-germ-layered) animals still retains its 
 dominance, like that of the foliage leaf, in spite 
 of exceptional cases and special homologies. 
 The only reply to this imaginary question that 
 
CONVERGENT METAMERISM 77 
 
 presents itself to my mind is this : The 
 integrity of the gut throughout the Triplo- 
 blastica cannot be assailed without invalidat- 
 ing the continuity of the archenteric cavity or 
 primitive gut throughout the Metazoa ; but this 
 is to strike at the root of the entire fabric of 
 comparative morphology, since the archenteron 
 is as much the palladium of animal morphology 
 as the primitive leaf-member is the root- 
 conception of botany. 
 
 The Triploblastica are composed of two great 
 series representing distinct phyletic types, namely, 
 the Platyhelminthes and the Coelomata. The 
 highest flatworms, the Nemertina, show points 
 of convergence towards Coelomata just as the 
 highest Mollusca, the Cephalopoda, show points 
 of convergence towards Vertebrata. 
 
 Inseparable from the question of the general 
 homology of the gut in coelomate animals is 
 that of the general homology of the coelom or 
 secondary body-cavity itself. It must suffice here 
 to affirm this general homology, again on first 
 principles ; but just as the divisions of the gut are 
 obviously not homologous throughout the coelo- 
 mate series, neither are the subdivisions of the 
 coelom, i.e., the somites, necessarily homologous. 
 A phenomenon analogous to metameric segmenta- 
 tion is that of strobilation, which is definitely 
 known to occur independently in different phyla, 
 e.g., Coelenterata, Platyhelminthes, Annelida. In 
 
78 DIVERGENCE AND PARALLELISM 
 
 some Annelid worms, e.g., Syllidae, strobilation is 
 superadded to the normal segmentation. As I 
 have indicated in the diagram (Fig. 2), the 
 Appendiculata are here regarded as a parallel 
 stem to the Vertebrata, having a common 
 Coelenterate origin. This scheme is in accord- 
 ance with the conviction that the somites of 
 Annelids and Arthropods have been evolved 
 along their own line of descent, independently 
 of the somites of Vertebrata. 
 
 The metamerism of Appendiculata and Verte- 
 brata is therefore, on this view, a conspicuous 
 example of convergence, and as such I regard 
 it, taking into account the facts which have been 
 made known concerning it. If this view is 
 correct, as I believe it to be, there can be no 
 question of comparing the regional differentia- 
 tion of Arthropods, such as Limulus and the 
 Scorpion, with that of Vertebrata. The whole 
 comparison, except by way of a possible con- 
 vergence here and there, is ruled out of court. 
 The existence of convergent strobilation is in 
 itself presumptive evidence of the possibility of 
 convergent metamerism, and there are other facts 
 which render it extremely probable and, to my 
 mind, practically certain. 
 
 Instances of parallel convergence are so 
 numerous and so common that we begin to 
 realise that convergence is a regular and not 
 an exceptional phenomenon. The most striking 
 
MAMMALIAN CONVERGENCE 79 
 
 example of the three principles of divergence, 
 convergence and parallelism, at one and the 
 same time, is of course that which is afforded by 
 the parallel series presented by the Marsupial 
 Mammals or Metatheria, on the one hand, and 
 the ordinary Placental Mammals or Eutheria, on 
 the other. I have compiled a table and give 
 a graphic representation of it in Fig. 5.* The 
 inward bends of the parallel waved lines indicate 
 particular features of convergence selected for 
 the purpose of the diagram. A similar diagram 
 could be constructed for comparing the series of 
 Insectivora and Rodentia, the spiny armature of 
 the hedgehogs approximating to that of the 
 porcupines, the arboreal habit of tree-shrews 
 (Tupaiidae) to that of squirrels (Sciuridae), the 
 terrestrial, nocturnal, and semi-domesticated habit 
 of land-shrews to that of mice and rats, while 
 the aquatic habit and the parachute flight are 
 also met with in both orders. The musk-shrew, 
 Crocidura murina, is very rat -like in general 
 deportment, although its eyes are small and its 
 dentition that of Insectivora. 
 
 Parallel evolution accompanied by convergence 
 is the expression of analogous formations in two 
 or more animals belonging to different subdivi- 
 sions, which may have acquired a similar differ- 
 entiation of outward appearance or internal 
 
 1 The items are gathered from Flower and Lydekker's textbook 
 gn Mammals, with reference also to Max Weber's " Saiigethiere," 
 
8o 
 
 DIVERGENCE AND PARALLELISM 
 
 organisation independently along different lines 
 of descent, the points in which they resemble 
 each other giving no indication of genetic affinity 
 nor even of bionomical association. 
 
 The general resemblances cited above relate 
 
 FIG. 5. Parallelism with convergence amongst mammals. 
 
 M = Marsupials ; P = Placentals. 
 
 This phenomenon was observed by Cuvier and by Owen. The 
 habits which lie at the base of the convergence in each case are 
 indicated by the numbers : 
 
 1. Carnivorous. 
 
 2. Ant-eating (myrmecophagous). 
 
 3. Flying. 
 
 4. Swimming. 
 
 5. Burrowing (large-eyed forms). 
 
 6. Burrowing (small-eyed forms). 
 
APPENDICULATES AND VERTEBRATES 81 
 
 to the external configuration of the body, and 
 some of them at least are sufficiently remark- 
 able, not to say astonishing. Perhaps one of 
 the closest degrees of functional convergence is 
 that which is manifested between the appendi- 
 culate (annelid and arthropod) and vertebrate 
 central nervous systems, upon which attention 
 has been focussed by Dr Gaskell. The central 
 nervous apparatus is the dominant system in 
 the organisation of the higher animals, and he 
 thinks that it has been a dominant factor in 
 evolution. The comparison tabulated below from 
 Dr Gaskell's recently published volume on the 
 "Origin of Vertebrates "( London, 1908) appears 
 to me to serve as a most beautiful example of 
 true physiological convergence, but likely to be 
 confounded with the spurious convergence pre- 
 sented by the so-called cephalic stomach of the 
 arthropod in comparison with the ventricles of 
 the brain in the craniate vertebrates. Now, it 
 may appear something like an impertinence to 
 describe as a spurious convergence what Dr 
 Gaskell regards as a true homology ; it is, how- 
 ever, nothing of the kind, and is no more than 
 an expression of the opposite point of view. On 
 that understanding we may leave it for the 
 present, perhaps returning to it later, either 
 directly or by implication. 
 
82 
 
 DIVERGENCE AND PARALLELISM 
 
 TABLE OF CONVERGENCE IN THE STRUCTURE OF 
 THE CENTRAL NERVOUS SYSTEM (NEURAL 
 CONVERGENCE) 
 
 APPENDICULATE. 
 (Segmented invertebrate). 
 
 1. Supra-oesophageal ganglia, 
 giving origin to the nerves of 
 the eyes and antennules, z.e,, 
 the optic and olfactory nerves. 
 
 2. Circum-oesophageal com- 
 missures connecting the cerebral 
 ganglia with the infra-cesopha- 
 geal ganglia and the ventral 
 chain. 
 
 3. Infra-cesophageal ganglia 
 and the paired ventral chain of 
 segmental ganglia. Each pair 
 of ganglia gives rise to the 
 nerves of its own segment, 
 motor and sensory; by the 
 agency of which food is in- 
 gested, respiration and loco- 
 motion effected. 
 
 VERTEBRATE. 
 (Craniate vertebrate}. 
 
 1. The brain proper from 
 which arise only the olfactory 
 and optic nerves. 
 
 2. Crura cerebri^ strands of 
 fibres on each side of the in- 
 fundibulum, connecting the 
 higher brain region proper 
 with the lower region of the 
 medulla oblongata and spinal 
 cord. 
 
 3. Region of the mid-brain, 
 medulla oblongata, and spinal 
 cord ; from these arises a series 
 of nerves segmentally arranged, 
 which, as in the invertebrate, 
 give origin to the nerves 
 governing mastication, respira- 
 tion, and locomotion. 
 
 With reference to the first sub-heading in the 
 above table it is to be remarked that Dr Gaskell 
 duly points out that the crustacean antennules 
 are olfactory in function, but he does not add 
 that they are also the carriers of the auditory 
 organs and that the latter are innervated from 
 the cerebral ganglia as they are in Mollusca. At 
 the same time a physiologist may be pardoned 
 for regarding the close correspondence as a mani- 
 festation of homology rather than as a case of 
 
APPENDICULATES AND VERTEBRATES 83 
 
 convergence ; and there are other facts which 
 serve still more to complicate the issue and to 
 baffle the most competent judgment. Chief 
 amongst these is the apparent topographical 
 coincidence between the infundibulum in the 
 floor of the third ventricle of the vertebrate 
 brain and the appendiculate oesophagus. 
 
 If we admit the convergence as tabulated 
 above, why should we not regard this further 
 coincidence as another case of convergence ? 
 There are at least two reasons why I think we 
 should not : firstly, because the structural con- 
 vergence in the two types of central nervous 
 system is accompanied by a certain degree of 
 functional equivalence, whereas the topographical 
 coincidence of infundibulum and oesophagus is 
 not accompanied by any functional equivalence ; 
 secondly, because the infundibulum can be brought 
 into interesting topographical correlation with 
 a primitive feature quite different from the 
 appendiculate oesophagus, namely, the anterior 
 neuropore or anterior neurenteric canal of pro- 
 tochordates. If the striking correspondence in 
 plan of composition of the central nervous 
 system in appendiculates and vertebrates had 
 been recognised as an instance of general 
 structural convergence thirty - five years ago, 
 when the late Dr Anton Dohrn, the founder of 
 the Stazione Zoologica at Naples, explained his 
 application of the principle of change of function 
 
84 DIVERGENCE AND PARALLELISM 
 
 in relation to the origin of vertebrates (1875), 
 it would not have led to what many zoologists 
 are constrained to regard as a series of false 
 homologies ; and per contra it would not have 
 led to many interesting discoveries. 
 
 There are two factors which conspire to 
 beguile a morphologist into spurious relations 
 and vain imaginings : these are the potency of 
 convergence and the weight of coincidence. A 
 very eminent French zoologist, who enjoyed the 
 respect and veneration of all who came directly 
 or indirectly within the sphere of his influence, 
 the late Professor Henri de Lacaze-Duthiers, 
 thought from the beginning to the end that 
 the Ascidians (Tunicata) are related to the 
 Mollusca. In apparent confirmation of this belief 
 he discovered the case of a simple Ascidian, a 
 species of Molgula, which develops without the 
 intervention of the usual tailed larva ; and then 
 he discovered a form in the Mediterranean which 
 he called Chevreulius, but which had already 
 been named Rhodosoma from the Pacific, where 
 the atrial and buccal siphons lie within a hinged 
 valve which can be opened and closed like the 
 valves of a lamellibranchiate mollusc. 
 
 It is indeed extraordinary to relate, in Dr 
 Gaskell's words, that the infundibulum "lies just 
 anteriorly to the exits of the third cranial or 
 oculomotor nerves ; in other words, it marks the 
 termination of the series of spinal and cranial 
 
VERTEBRATE DESCENT 85 
 
 segmental nerves. . . . Not only, then, are the 
 nerve-masses in the two systems [appendiculate 
 and vertebrate] exactly comparable, but in the 
 very place where the cesophageal tube is found 
 in the invertebrate, the infundibular tube exists 
 in the vertebrate. ..." All this is true, but 
 what Dr Gaskell does not relate at this crucial 
 point of his narrative (/.., p. 14) is that the 
 infundibulum also occurs essentially at the same 
 horizon, with reference to the segmental nerves, 
 as the anterior neuropore and olfactory pit of 
 Amphioxus, 1 where there are likewise two 
 anterior pairs of sensory cranial nerves ter- 
 minating in peripheral ganglion cells on the 
 prseoral lobe. 
 
 It is not particularly edifying to pursue one 
 phantom rather than another ; and the spirit of 
 morphology may be said to reside not so much 
 in a desire for ultimate truth, that is to say, to 
 attain the unattainable, as in the will to accept 
 nothing but the truth. Applying this aphorism 
 to the special case of the two opposing theories 
 of vertebrate descent, namely, the Appendiculate 
 theory and the Protochordate theory, we may 
 say that if the latter fails to satisfy, so at least 
 does the former. It is necessary to codify the 
 facts ; but this can be done for the present upon 
 a broad basis of convergence better than by 
 an unrestrained use of homology. 
 
 1 Dr Gaskell's references to the anterior neuropore of Amphioxus 
 occur on pp. 220 and 457 of his work, to which I refer the reader. 
 
 F 2 
 
86 DIVERGENCE AND PARALLELISM 
 
 Meanwhile we may agree unreservedly with 
 Dr Gaskell's opinion that nothing but good can 
 result from the incursion of the physiologist into 
 the realm of the morphologist. Then, and then 
 only, do we begin to realise how we stand in 
 regard to the facts of comparative anatomy, 
 and to perceive the paramount importance of 
 convergence. 
 
 With regard to the special case of the nervous 
 system we must add that there is no ground for 
 imagining that it is the one organic system 
 throughout the animal kingdom which offers no 
 instances of convergence. 
 
CHAPTER VII 
 
 SPECIAL CONVERGENCE 
 
 SPECIAL convergence depends upon functional 
 equivalence, which may or may not be accom- 
 panied by partial homology. The different 
 degrees may be expressed in various ways at 
 convenience, according as it is desired to focus 
 attention upon a function or an organ, or upon 
 the systematic position. In the systematic 
 method of arrangement one makes use of the 
 terms employed in classification. Individual or 
 racial convergence refers to those resemblances 
 between varieties of a species such as are 
 familiar in the human race as the so-called 
 " doubles." 1 Specific, generic, family, ordinal, 
 class, and phyletic convergence refer respectively 
 to the species of a polytypic genus, to the 
 
 1 Since the above was written Professor Adam Sedgwick has 
 kindly drawn my attention to a paper by Professor Arthur Keith on 
 " The Position of the Negro and Pygmy amongst Human Races" in 
 Nature^ vol. Ixxxiv., 1910, p.54. Professor Keith thinks that pygmies, 
 who are widely distributed, are "modifications produced locally from 
 the larger negro. . . . The Congo pygmies share all the physical 
 features of the Bantu except size ; the Bushman has the characters 
 of the Hottentot, while the pygmies of the Far East find their nearest 
 representatives in the negroes of Oceania." 
 
88 SPECIAL CONVERGENCE 
 
 genera of a family, to the families of an order, 
 to the orders of a class, to the classes of a 
 phylum and to the phyla of the animal kingdom. 
 One of the best examples of special family 
 convergence is afforded by the pectoral fins of 
 flying fishes. Semper (pp. cit.) dealt with the 
 organs of flight in Vertebrata, and figured a 
 flying " herring," Exoccetus, belonging to the 
 Teleostean family Scombresocidse, in juxta- 
 position with a bat ; but he did not refer to 
 the case of the flying gurnard, Dactylopterus, 
 which belongs to another family, the Triglidse 
 (sometimes called Cottidae and Cataphracti). 
 Both of these genera occur in the Mediter- 
 ranean as well as in the Indian Ocean, and 
 are totally different from each other, not only 
 in systematic position but in external form. 
 Dactylopterus has a broad, depressed head, 
 armed with powerful spines, on account of 
 which the head is commonly fractured by a 
 blow when the fish is caught by native fisher- 
 men off the coast of Ceylon ; the scales are 
 hard, keeled scutes ; the tail - fin is truncated, 
 the abdomen flattened, and the body coloured 
 red. Exoc&tus has a normal herring-like head 
 and body, unarmed, with smooth scales, a deeply- 
 forked tail-fin, a convex abdomen, and a silvery 
 ground colour. Both kinds of flying-fishes owe 
 their limited power of flight above the surface 
 of the sea to the secondary elongation and 
 
WING FLIGHT 89 
 
 expansion of the pectoral fins, an exceptional 
 modification which has been acquired inde- 
 pendently within the limits of two very distinct 
 families. 
 
 It is remarkable to find such strictly homo- 
 logous organs as the pectoral fins of Teleostean 
 fishes modified in a virtually identical manner 
 to perform a special and exceptional function, 
 whose transformation is nevertheless not homo- 
 genetic but homoplastic. This typical example 
 serves to illustrate one of the most interesting 
 manifestations of convergence, namely, the homo- 
 plastic modification of homogenetic structures. 
 
 The comparison between the wings of birds 
 and of bats offers different conditions. They are 
 equally efficient as organs of sustained aerial 
 flight ; they show complete functional equiva- 
 lence ; the fore-limbs are homologous ; but the 
 transformation has proceeded along divergent 
 lines, an expanded wing-membrane or patagium 
 with reduced epidermal appendages (hairs) in the 
 one case ; a rudimentary patagium and enlarged 
 epidermal appendages (feathers) in the other; with 
 corresponding differences affecting the skeleton of 
 the wing-supporting limbs, hypertrophied digits 
 in the bat, atrophied digits in the bird. Unlike 
 the flying fishes, the modification has not taken 
 place in an identical manner, but in an almost 
 diametrically opposite manner ; it cannot there- 
 fore be described as homoplastic in the strict 
 
90 SPECIAL CONVERGENCE 
 
 sense of the term, although it is frequently quoted 
 as a stereotyped example of homoplasy. Perhaps 
 it would be more logical to express it in other 
 language for which a suitable technical term does 
 not seem to be available, unless we may borrow 
 a word from another division of the subject and 
 call it a heterotypic modification of homogenetic 
 structures, the common starting-point being the 
 pentadactyle limb of terrestrial vertebrates. 
 
 Birds and bats are directly comparable in terms 
 of the theory of convergence, not only because 
 they are alone amongst vertebrate animals in 
 their capacity for rapid and changing flight, but 
 also because they are both warm-blooded. Fly- 
 ing fishes bear the same relation to birds as they 
 do to bats, whatever that relation may be held 
 to be. In my opinion they are in no way com- 
 parable, inasmuch as the flight of fishes is of the 
 nature of a parachute flight and so belongs to a 
 different category. Although well enough twenty 
 years ago, it would not be admissible to-day to 
 figure a flying fish and a bat upon the same page 
 for the purpose of illustrating the principle of 
 homoplastic wing-formation, except in respect of 
 the one point of the hypertrophy of the pectoral 
 rays and digits. It would be more appropriate 
 to present a flying fish in conjunction with a 
 flying lizard, or even a flying squirrel, the nature 
 of the relation between them being one of general 
 functional convergence, neither associated with 
 
PARACHUTE FLIGHT 91 
 
 homology nor with general structural convergence 
 or homoplasy. 
 
 The case of Pteromys and Galeopithecus, 
 quoted above, may be regarded as an example 
 of the homoplastic modification of homogenetic 
 structures in different orders, i.e., ordinal con- 
 vergence ; but the homoplasy, in the literal sense, 
 is not complete, since there are important differ- 
 ences of structural detail, the digits of Galeopithecus 
 being webbed, 1 and the wrist of Pteromys carry- 
 ing an accessory cartilaginous rod which gives 
 material support to the parachute. 
 
 The degrees of convergence are endless and 
 can only be classified in a broad way ; it is, we 
 may repeat, a fundamental phenomenon and con- 
 sequently of frequent occurrence, though not 
 always equally demonstrative. Very striking are 
 those cases where special features which are 
 exceptional within a certain limited systematic 
 range are repeated independently, as the phos- 
 phorescent and electric organs of fishes. The 
 late Professor Howes, in one of his addresses, 
 quoted as a critical example of convergence the 
 fact that the Mesozoic reptiles yielded terrestrial, 
 aquatic, and aerial forms, just as did the Tertiary 
 mammals which replaced the former in order of 
 dominance. 
 
 Amongst animals which are more or less 
 
 1 Flower and Lydekker, "Mammals Living and Extinct," 
 London, 1891, Fig. on p. 615. 
 
92 SPECIAL CONVERGENCE 
 
 nocturnal in their habits, the pupil of the eye 
 becomes contracted in a bright light to a narrow 
 aperture which may retain its normal and primitive 
 circular shape or may become a vertical ellipse 
 or a narrow vertical slit or a horizontal ellipse or 
 slit. The pupil of the cat's eye is round at night 
 time, vertically contracted during the day ; the 
 leopard's eye and the jaguar's, on the contrary, 
 like the lion and tiger, contract in bright sun- 
 light to a circular pinhole aperture. Mr Frank 
 Buckland has noted that the pupil of the fox's 
 eye behaves, not like the dog's, as one would 
 expect, but like the cat's, and this linear pupil is, 
 by some writers, 1 advanced as a generic character 
 of the genus Vulpes, as distinguished from Cam's. 
 Amongst tailless Batrachia (frogs and toads) 
 the most extensive genera, Rana, Bufo, Rhaco- 
 phorus and Ixalus, have a horizontal pupil, a 
 vertical pupil being the exception in this division. 
 Geckoes (Geckonidse), the lizards of tropical and 
 subtropical countries which can run along upright 
 and over - hanging surfaces by virtue of the 
 adhesive pads on their digits, have the vertical 
 form of pupil ; when partially contracted the 
 margin of the iris in some species is crenulated ; 
 and when quite contracted so that the anterior 
 and posterior borders are in contact, leaving only 
 a vertical pupillary suture between them, the 
 
 .%.) E. T. Seton, " Life-Histories of Northern Animals." New 
 York, 1909, 2 vols. 
 
PUPIL OF THE EYE 93 
 
 latter has a beaded appearance, the beading 
 corresponding with the previously noted crenula- 
 tions. Crocodiles also show the narrow vertical 
 pupil. 
 
 Among snakes, round and vertical pupils occur 
 with nearly equal frequency, the former rather 
 more frequently than the latter; but Dryophis, 
 called the whip-snake, has a very pronounced 
 horizontal pupil (Fig. 6). 
 
 FIG. 6. 
 
 Head of Dry aphis pulverulentus in side view. 
 
 In the class Mammalia the round pupil pre- 
 vails, whereas the vertical pupil, as mentioned 
 above, characterises certain groups. The oblong 
 horizontal pupil of the horse, camel, ox, and sheep 
 was referred to by Owen in his " Anatomy of 
 Vertebrates" (1866); it is very sharply defined 
 in goats that have a gilt-edged iris. Owen also 
 stated that in a dead Agouti (Rodentia) the 
 pupil was a horizontal ellipse ; but he did not 
 allude to the horizontal pupil of the mungoose, 
 nor have I seen it mentioned in any other work 
 so far as I can remember. In the Civet family 
 (Viverridse), which is closely allied to the Felidae, 
 the vertical pupil appears to be the rule, but 
 
94 SPECIAL CONVERGENCE 
 
 Herpestes, the Indian mungoose, has a well- 
 defined, horizontal pupil in the daytime, becoming 
 round at night (Fig. 7, Frontispiece). The allied 
 genus Suricata, the South African meerkat, has 
 a similarly shaped pupil, shown very clearly in 
 the specimens at the London Zoological Gardens. 
 
 Hubrecht (I.e., 1908) has given reasons for 
 believing that the cartilaginous fishes occupy a 
 side branch of the vertebrate stem, and he 
 goes so far as to suggest a radical change in 
 the classification of vertebrates, dividing them 
 into four super-classes : Cephalochordata (Amphi- 
 oxus\ Cyclostomata (lampreys, etc.), Chondrophora 
 (sharks, etc.), and Osteophora (bony fishes and 
 higher vertebrates). In view of this weighty 
 opinion, the a priori probability that the nicti- 
 tating membrane of some sharks (Carchariidse), 
 which is an accessory protection for the eye, 
 sweeping over the surface of the eyeball, has 
 no genetic connection with the functionally 
 equivalent membrane of many of the higher 
 vertebrates, is greatly strengthened. 
 
 Bony fishes have no movable eyelids, but 
 transparent adipose eyelids are sometimes present, 
 composed of an anterior and posterior membrane 
 or of a continuous circular membrane perforated 
 at or about the centre. Both the Grey Mullet 
 family (Mugilidse) and the Herring family 
 (Clupeidae) are respectively differentiated into 
 groups by the presence or absence of adipose 
 
EYELIDS 95 
 
 eyelids which must have arisen independently 
 in those families. Where there is a continuous 
 perforated membrane, as in some Clupeidse 
 (e.g., Dussumieria, Fig. 8), the condition pre- 
 sented calls to mind that of the Oegopsid cuttle- 
 fishes, where the outer covering of the eye is 
 likewise perforated. No doubt the two condi- 
 tions are physiologically comparable. 
 
 FIG. 8. Eye of Dussumieria showing perforated adipose eyelid. 
 The outline of the pupil is seen below the transparent [dotted] 
 adipose membrane. 
 
 Raptorial or prehensile appendages amongst 
 invertebrate animals offer further analogies. 
 Chelate extremities occur alike in Crustacea 
 and Arachnida ; and the characteristic clasp 
 knife appendages of the Stomatopod Crustacea 
 (Squilla) are found again in Orthopterous insects 
 (Mantodea) and in Neuropterous insects (Mantis- 
 pidse). In Chiromantis, a tropical African genus 
 of frogs allied to Rhacophorus, the two inner 
 fingers of the hand are opposite and opposable 
 to the two outer fingers, so that its structure re- 
 sembles that of the grasping hand of chameleons. 1 
 All these cases belong to the great category 
 of opposable extremities, some of which are 
 
 1 G. A. Boulenger, Catal. Batr. Sal, 1882, p. 92, pi. x., Figs. I 
 and 2. 
 
96 SPECIAL CONVERGENCE 
 
 reciprocally homologous, whilst others, like the 
 above, are homoplastic. 
 
 Another case worth quoting, as well known to 
 students of osteology as it is remarkable, is that 
 of convergent opisthocoely already mentioned by 
 Owen l as occurring in the North American fresh- 
 water Ganoid fish, Lepidosteus the Garpike, in 
 the Surinam toad, Pipa americana, and in the 
 Salamanders. In Batrachia and Reptilia the 
 vertebrae articulate together by a cup-and-ball 
 or ball-and-socket arrangement, and, as a general 
 rule, the cup lies at the front end of the centrum 
 of each vertebra to receive the condyle which 
 projects from the hinder end of the preceding 
 vertebra. Such vertebrae with the concavities in 
 front are called procoelous. In rare cases the 
 reversed condition obtains, where the condyles 
 project at the front ends of the vertebrae and fit 
 into the cups at the hinder ends of the preceding 
 vertebrae. Such vertebrae with the concavities 
 behind are called opisthocoelous. 
 
 The vertebrae of fishes are joined together by 
 ligament, and the centra do not possess articu- 
 lar surfaces ; both in front and behind there is 
 a deep cup, whence the vertebrae are designated 
 amphicoelous or biconcave. Lepidosteus is unique 
 amongst existing fishes in having opisthocoelous 
 vertebrae. The same description of vertebrae 
 only occurs otherwise in certain Batrachia, in- 
 
 1 R. Owen, "Anatomy of Vertebrates," London, 1866, 1. pp. 33-34. 
 
OPISTHOCOELOUS VERTEBRA 97 
 
 eluding the newts and salamanders among the 
 Urodela, and a number of genera of Anura of 
 peculiar habits belonging to separate families 
 and each containing a single species, namely, 
 Bombinator the European Fire-bellied toad, Pipa 
 the Surinam toad, Alytes the Midwife toad, and 
 Discoglossus, the Painted frog, a circum-Medit- 
 erranean batrachian. The opisthocoely of Pipa 
 is probably independent of that of the other 
 named Anura ; that of the named Anura collec- 
 tively is probably independent of salamandrine 
 opisthocoely ; and Batrachian opisthocoely as a 
 whole is independent of that of Lepidosteus. 
 
 Somewhat analogous to the reversed articu- 
 lation of vertebrae described in the preceding 
 paragraph is the reversed spire of the shell in 
 Molluscs and the reversed pose of the body in 
 Flatfishes. These two latter cases of situs inver- 
 sus were quoted by Bateson as examples of 
 discontinuity in substantive variation, but they 
 must be mentioned again here as examples of 
 convergent variation. " In both of the groups 
 named, some species are normally right-handed, 
 others being normally left-handed, while as indi- 
 vidual variations reversed examples are found. 
 . . . The fact that the reversed condition may 
 become a character of an established race is 
 familiar in the case of Fiisus antiqmts. This 
 shell is found in abundance as a fossil of the 
 Norwich Crag, such specimens being normally 
 
 G 
 
98 SPECIAL CONVERGENCE 
 
 left-handed, though the same species at the 
 present day is a right-handed one." 1 In India 
 and Ceylon, where the chank shell (Turbinella 
 pyruni) plays such a large part in religious 
 ceremonies, the rare apparition of a sinistral 
 form is regarded as a favourable portent, and 
 such a shell is very highly esteemed. There 
 would appear to be some common causes 
 operating during the early development to 
 procure such cases of reversal. A spiral shell 
 can only be twisted in one of two directions ; 
 and a Pleuronectid fish can only lie on one of 
 two sides. These alternative conditions, one of 
 which is frequently the rule, the other the 
 exception, afford an interesting sidelight upon 
 the incidence of convergence. 
 
 The classification of reversed molluscs 2 is the 
 same in principle as that for reversed flatfishes, 
 namely : 
 
 (1) Cases in which the genus is normally 
 
 sinistral. 
 
 (2) Cases in which the genus is normally 
 
 dextral but certain species are nor- 
 mally sinistral. 
 
 (3) Cases in which the species is indifferently 
 
 dextral or sinistral. 
 
 (4) Cases in which a sinistral form is an 
 
 abnormal monstrosity. 
 
 1 W. Bateson, " Materials for the Study of Variation," London, 
 
 1894, p. 54. 
 
 2 Cf. A. H. Cooke "Molluscs," Cambridge Natural History, 
 
 1895, p. 249. 
 
TEN-LEGGED PANTOPODS 99 
 
 A striking example of generic convergence is 
 exhibited in the ten-legged Pantopods or "sea- 
 spiders " which have been obtained from the 
 South Polar region by the English and Scottish 
 National Antarctic Expeditions. They belong 
 to two quite distinct genera, Pentanymphon 
 Hodgson, 1904, and Decolopoda Eights, 1837; 
 the former only differing from Nymphon by its 
 possession of an extra pair of legs ; the latter 
 most nearly related to Colossendeis, from which 
 it differs likewise by its possession of an extra 
 pair of legs, and also by the presence of a 
 pair of well-developed, three-jointed mandibles. 
 These two Antarctic genera are thus nearly 
 related to two other representative Pantopod 
 genera but not to each other. Professor D'Arcy 
 Thompson 1 places them in widely separated 
 families : Decolopodidae and Nymphonidse ; and 
 Schimkewitsch 2 has arrived at the same con- 
 clusion, namely, that the ten-legged Pantopods 
 do not constitute a distinct group in contrast 
 with the more usual eight-legged forms, but 
 that they arose independently from eight-legged 
 forms, the number of legs giving no clue to 
 their affinities. 
 
 The following table, compiled from the data 
 furnished by Hodgson, 3 who collected and 
 
 1 D'Arcy W. Thompson, " Pycnogonida," Camb. Nat. Hist., iv., 
 1909, pp. 531 and 537. 
 
 2 W. Schimkewitsch, " Uber die Periodicitat in dem System der 
 Pantopoden," Zool. Anz.^ xxx., 1906, p. 3. 
 
 3 Besides the Reports of the Voyages, see the following papers 
 
ioo SPECIAL CONVERGENCE 
 
 described Pentanymphon during the National 
 Antarctic Expedition in the Discovery, and 
 redescribed Decolopoda from material procured 
 by Mr W. S. Bruce during the Scottish National 
 Antarctic Expedition in the Scotia, shows the 
 radical differences which exist between the two 
 genera simultaneously with the numerical con- 
 vergence of the ambulatory appendages. The 
 differential characters indicate the respective 
 places of the genera in the system of Pantopoda ; 
 the convergent character gives no indication of 
 their affinities to any existing forms. 
 
 TABLE OF CONVERGENCE IN PANTOPODA. 
 
 PENTANYMPHON. DECOLOPODA. 
 
 1. Ambulatory appendages, Ambulatory appendages, five 
 five pairs. pairs. 
 
 2. Body elongate, clearly Body short, obscurely seg- 
 segmented, very slender, with mented, broadly elliptical in- 
 lateral processes widely separ- eluding the lateral processes in 
 ated. the contour of the ellipse. 
 
 3. Proboscis cylindrical, Proboscis clavate, bent down- 
 straight, much shorter than the wards, longer than the body, 
 body. 
 
 4. Mandibles chelate, 2- Mandibles chelate, 3-jointed. 
 jointed. 
 
 5. Palps 5-jointed. Palps lo-jointed. 
 
 6. Ovigerous legs, lo-jointed, Ovigerous legs, 10- jointed, 
 formed plainly as in Nymphon. looped as in Colossendeis. 
 
 7. Abdomen small, ovoid, Abdomen long, slender, sub- 
 directed obliquely upwards. clavate. 
 
 by T. V. Hodgson : I. " On a New Pycnogonid from the South 
 Polar Regions," Ann. Nat. Hist. (7), xiv., 1904, pp. 458-462, pi. 
 xiv. 2. "Scotia Collections. On Decalopoda australis Eights, 
 an old Pycnogonid rediscovered," Proc. Roy. Phys. Soc., Edin- 
 burgh, vol. xvi., 1905, pp. 35-42, pi. iii. 3. " Decalopoda and 
 Colossendeis," Zool. Anz., xxix., 1905, pp. 254-256. [The accepted 
 spelling of the former name is now Decolopoda.] 
 
TEN-LEGGED PANTOPODS 101 
 
 This table shows two genera belonging to 
 different families of a comparatively small order, 
 with one very prominent character in common 
 which cannot be referred back to a common 
 origin. 
 
 Quite recently a letter from Dr W. T. Caiman, 
 containing further references to Antarctic Panto- 
 poda, appeared in Nature (vol. Ixxxiv., 28th July 
 1910, p. 104), which adds to the importance of 
 this case. A third genus of ten-legged Panto- 
 pods, named Pentapycnon by Professor E. L. 
 Bouvier, has been obtained by Dr Charcot's 
 expedition in the Pourquoi Pas, and "strange 
 to say, it appears to be quite unrelated to the 
 other two. Pentapycnon charcoti is a near relative 
 of Pycnogonum, hitherto regarded as the most 
 highly-specialised of all Pycnogonida. Further, 
 just as Pentanymphon is accompanied by a species 
 of Nymphon, and Decolopoda by Colossendeis, so 
 Professor Bouvier finds that Pycnogonum> hitherto 
 unknown from Antarctic seas, is represented by 
 a new species alongside of Pentapycnon at the 
 South Shetlands." 
 
 We may now consider the case of two genera 
 of toads belonging to one family (Bufonidae) of 
 a large order (Batrachia-Anura). These show 
 several exceptional characters in common, side 
 by side with a number of strongly divergent 
 features. The divergences in structure and 
 distribution are so great as to make it doubtful 
 
 G 2 
 
102 SPECIAL CONVERGENCE 
 
 whether any of the other characters are trace- 
 able to direct genetic affinity ; and Gadow 1 
 suggests that their community of habits with 
 concomitant modifications of structure .may be 
 due to convergent evolution. The structural 
 features in the table given below are taken from 
 Boulenger ; 2 the feeding habits are gathered 
 from Gadow. Both of these authors note their 
 resemblance to the firmisternal family of Engy- 
 stomatidae, on the one hand, and their systematic 
 position under the arciferous Bufonidae, on the 
 other. The vertical pupil of the eye is an 
 exceptional character among the Bufonidae. 
 
 TABLE OF CONVERGENCE IN BUFONID^. 
 MYOBATRACHUS. RHINOPHRYNUS. 
 
 1. Distribution. Australia. Mexico. 
 
 2. Shoulder girdle. Epicor- The same, except sternum 
 acoid cartilages narrow and rudimentary. 
 
 scarcely overlapping; no omo- 
 sternum ; sternum (or meta- 
 sternum) cartilaginous, ossified 
 or calcified on the median line, 
 
 3. Eye. Pupil vertical. Pupil vertical. 
 
 4. Ear. Tympanum distinct. Tympanum absent. 
 
 5. Tongue small, elliptical, Tongue elongate, subtrian- 
 entire and free behind. gular, free in front. 
 
 6. Fingers and toes free. Fingers free, toes webbed. 
 
 7. Limbs very short, adapted Limbs very short, 
 for burrowing. 
 
 8. Male with a subgular vocal Male with two lateral vocal 
 sac. sacs internal behind the angle 
 
 of the mouth. 
 
 9. Habits. Feeding on ter- Feeding on termites and ants, 
 mites and ants. 
 
 1 H. Gadow, "Amphibia and Reptiles," Camb. Nat. Hist., 
 1901, pp. 166 and 227. 
 
 2 G. A. Boulenger, Catalogue Batr. Sal, 1882, pp. 328-329. 
 
PLACENTATION 103 
 
 The phenomena of placentation, according to 
 Hubrecht, are " intimately related to the higher 
 development which characterises the mammalia 
 as against the lower Vertebrates." This means 
 that the intensive nutrition of the embryo is a 
 prior condition to the development of brain 
 power, and it should be kept in mind as a 
 counterpoise to Gaskell's opinion that the brain 
 power is the primary directive influence in evolu- 
 tion. We may indeed state the case outright 
 by saying that cerebral concentration is the end 
 and not the beginning of evolution. 
 
 The history of the mammalian placenta affords 
 many examples of convergence. Diffuse placenta- 
 tion, in which the surface of the maternal uterine 
 mucous membrane is thrown into a dense net- 
 work of folds and crypts into which correspond- 
 ing folds or villi, diffused over the entire surface 
 of the blastocyst, fit without any strong adhesion 
 to the uterine wall, has usually been regarded as 
 a primitive condition ; and the arrangement seen 
 in the horse and the pig has always been looked 
 upon as the prototype of the diffuse placenta. 
 Hubrecht, however, gives weighty reasons for 
 thinking that it is due to a secondarily simplified 
 process, or descent with simplification, in the 
 course of which the intense phagocytic activity 
 of the trophoblast or outer foetal envelope had 
 subsided and had given place to a diffused osmotic 
 
104 SPECIAL CONVERGENCE 
 
 activity usually accompanied by great increase 
 in the size of the blastocyst. 
 
 Similar simplification and change of function 
 leading to parallel results have occurred in other 
 orders of mammals, e.g., in certain Edentates 
 (Manis the pangolin, or scaly ant-eater) and in 
 the Lemurs. Hubrecht expresses the hope that 
 his discussion of the facts will prevent any future 
 attempt to place Ungulates and Lemurs on the 
 same level of so-called primitive placentation. 
 
 In the same way the old hypothesis which 
 maintained that the early villiferous state of the 
 human blastocyst, in the phase called " Reichert's 
 ovum," recapitulates ontogenetically a diffuse 
 phase to which the discoid stage succeeded 
 later, ought to be discarded, because, inter alia, 
 Reichert's ovum is enclosed by a decidua reflexa 
 sen capsularis, and is not freely suspended in the 
 uterine cavity as are the blastocysts which show 
 diffuse placentation. Furthermore, Hubrecht 
 adds that this "phenomenon of encapsulisation 
 inside the mucosa has appeared independently 
 in more than one order of mammals." The 
 arrangement in question, by which the develop- 
 ing blastocyst is withdrawn from the uterine 
 lumen and enclosed by a decidua capsularis, has 
 been realised in man and anthropoid apes, in 
 different genera of rodents, in the spiny hedge- 
 hog Erinaceus, in the spineless Malayan hedge- 
 
PLACENTATION 105 
 
 hog Gymnura, and rather less completely in the 
 Chiroptera (bats). 
 
 Similarly the discoidal form of the placenta 
 gives no indication of affinity between the widely 
 different mammals which present it. It is, as 
 Hubrecht says, a temporary production, the dis- 
 coidal shape being of no value when considering 
 questions of affinity. He instances the discoid 
 placenta of the mole out of which the allantoic 
 villi are withdrawn at birth like the fingers 
 out of a glove, the young when born being 
 " enveloped in the allantois with the fully extracted 
 villi forming a woolly covering to that foetal in- 
 volucrum" ; the discoid placenta of Galeopithecus, 
 containing lacunae filled with maternal blood and 
 imbedded in the uterine wall ; " the discoid 
 placenta of the rabbit and of Tarsius which, 
 when full grown, is attached to the mother by a 
 stalk of much smaller diameter than the placenta 
 itself; the discoid placenta of the hedgehog and 
 of man, the latter with its loose and floating 
 villi as against the dense trellis -work of villi 
 and trophoblast in the former." 1 
 
 Before leaving the subject of the nutrition of 
 the young of vertebrates, one more matter may 
 receive attention, namely, the secretion of milk 
 that takes place in the mammary glands from 
 which the mammalia take their name. According 
 
 1 A. A. W. Hubrecht, "Early Ontogenetic Phenomena in 
 Mammals," Quart. Journ. Micr. St., vol. liii., 1908, see pp. 70, 113, 
 129, 142, etc. 
 
io6 SPECIAL CONVERGENCE 
 
 to Owen (op. cit ., vol. ii. p. 160), John Hunter 
 "made many interesting observations on the crop 
 of pigeons, which takes on a secreting function 
 during the breeding season, for the purpose of 
 supplying the young pigeons in the callow state 
 with a diet suitable to their tender condition. 
 An abundant secretion of a milky fluid of an 
 ash-grey colour, which coagulates with acids and 
 forms curd, is poured out into the crop and mixed 
 with the macerating grains. This phenomenon 
 is the nearest approach in the class of birds to 
 the characteristic mammary function of a higher 
 class ; and the analogy of the * pigeon's milk ' 
 to the lacteal secretion of the mammalia has 
 not escaped popular notice. . . . The secretion 
 consists of proteine with oil, but contains no 
 sugar of milk nor fluid caseine." 
 
 An analogous secretion has been discovered 
 by Alcock to take place in the uterus of vivi- 
 parous rays. In pregnant females of sting-rays 
 (Trygon\ of the eagle-ray (Myliobatis\ and in 
 the bat-ray (Pteroplataa), Alcock found that the 
 young are nourished before birth by a milky 
 secretion that exudes from glandular filaments 
 or villi on the inner surface of the uterine wall, 
 a large bundle of the filaments passing through 
 each spiracle into the pharynx of the immature 
 fish. The milk-secreting filaments are penetrated 
 by a capillary network in the meshes of which 
 the milk-glands are imbedded. Each filament 
 
SECRETION OF MILK 107 
 
 is provided with superficial muscles whose con- 
 traction must serve to squeeze the milk out. 
 These vascular villi give a shaggy appearance 
 to the mucous membrane of the oviduct, and 
 Alcock describes them as " dripping with milk" 
 and discharging a creamy, albuminous fluid. 1 This 
 method of intra-uterine nutrition of the young by 
 a special secretion of the uterine glands offers at 
 once a great contrast and a remarkable parallel 
 to the corresponding phenomena in mammals. 
 
 Every system of organs throughout the animal 
 kingdom will be found to yield abundant instances 
 of convergence. An excellent example of what 
 may be called gastral convergence is exhibited 
 in the structure of the gizzard of some Teleostean 
 fishes. In the first place, as between the pyloric 
 gizzard of the few fishes which possess one and 
 the pyloric gizzard which is so characteristic of 
 birds, it is to be noted that the most absolute 
 comparability prevails, differences in detail not- 
 withstanding. They are homoplastic modifica- 
 tions of a homologous structure, namely, the 
 pyloric division of the vertebrate stomach. The 
 exceptional occurrence of a gizzard in fishes 2 and 
 the regular presence of one in birds give no more 
 indications of affinity than does the gizzard of 
 
 1 A. Alcock, "A Naturalist in Indian Seas," London, 1902, 
 see pp. 71, 159, 210. This valuable book contains a bibliography 
 of the published work of the Indian Marine Survey ship In- 
 vestigator. 
 
 2 Compare also the gizzard of the toothless Ant-eaters. 
 
io8 SPECIAL CONVERGENCE 
 
 the earthworm or that of some molluscs. The 
 latter are examples of phyletic convergence not 
 based directly on special homology ; the gizzard 
 in fishes and birds is a case of class convergence 
 based, as we have seen, upon a homogenetic 
 foundation. We may thus once more distinguish 
 clearly between convergence with homology and 
 that without it. This may seem rather para- 
 doxical, but it is true and sufficiently significant 
 to bear repetition : the pyloric division of the 
 stomach is homologous in fishes and birds ; but 
 the gizzard of fishes, which is a direct modifica- 
 tion of the pylorus, is not homologous with the 
 gizzard of birds, which is likewise a direct, though 
 independent, modification of the pylorus. 
 
 In that rich storehouse of facts, Owen's 
 "Anatomy of Vertebrates" (1866), the gizzard 
 of the Grey Mullet (Mugil) is thus described 
 (vol. i. p. 418) : "The cardiac portion here forms 
 a long cul-de-sac ; the pyloric part is continued 
 from the cardiac end of this at right angles and 
 is of a conical figure externally ; but the cavity 
 within is reduced almost to a linear fissure by 
 the great development of the muscular parietes, 
 which are an inch thick at the base of the cone ; 
 and this part is lined by a thick horny epithelium." 
 In other cases the pylorus will be found to be 
 hard and muscular, as in the Red Mullet ( Upene- 
 oides\ In the Clupeoid genus Megalops, the 
 pylorus is a long, straight, porrect ascending tube, 
 
GIZZARD IN FISHES 
 
 109 
 
 rather hard at its distal extremity ; in another 
 Clupeoid genus Dussumieria, the pylorus appears 
 as an arched muscular tube lying across the body 
 cavity from left to right, the concavity of the arch 
 being directed forwards (Fig. 9). 
 In a species of Grey Mullet 
 from Ceylon which I identified 
 as Mugil olivaceus Day, I found 
 the gastric ccecum rudimentary 
 and the cardiac division of the 
 stomach was followed by a round, 
 white, muscular, bulbous gizzard ; 
 the intestine was full of fine sand ; 
 six pyloric cceca embraced the 
 gizzard; another specimen had 
 only five pyloric cceca. What, 
 however, surprised me very 
 much at the time and I have 
 not found it mentioned in any 
 work which I have consulted 
 since was the discovery of an 
 almost precisely similar gizzard 
 in a Clupeoid fish, Chatoessus 
 nasus, which frequents the same g-c- Gastric [cardiac] 
 
 ccecum. 
 
 back-waters as the grey mullet. p, c . Pyloric cceca. 
 In this species (C. nasus} the * Destine, 
 stomach is frankly destitute of a cardiac ccecum, 
 and terminates in a round, muscular, red, bulbous 
 gizzard. In one of the specimens dissected, the 
 gizzard and intestine were full of fine sand as in 
 
 FIG. 9. 
 
 Intestinal tract of 
 Dussutnieria. 
 p. Pylorus. 
 
no SPECIAL CONVERGENCE 
 
 the mullet ; and the presence of sand was noted 
 in the pyloric cceca. The latter arise in tufts 
 along the length of the duodenum (Fig. 10). 
 
 FIG. 10. Dissection of the alimentary tract of Chatoessus nasus 
 from the left side. The dotted line indicates the cut edge 
 of the body- wall. 
 
 1. Gall-bladder. 
 
 2. Visceral nerve lying upon the cardiac tube of the stomach. 
 
 3. Air-bladder. 
 
 4. Ductus pneumaticus. 
 
 5. Pyloric gizzard with a tuft of cceca projecting in front of it. 
 
 6. Duodenum (green-tinted when fresh). 
 
 7. Rectum. 
 
 8. Hind-portion of left testis. 
 
 9. Vas deferens. 
 
 Here, then, we have two fishes belonging to 
 widely separated families, though pursuing similar 
 habits, and presenting independently an identical 
 modification of the pyloric division of the stomach. 
 I confess that at first acquaintance with this case 
 I began to distrust my own eyes ; perhaps it will 
 not strike the reader of these pages so forcibly ; 
 but I submit it as one of the most stringent 
 object lessons in homoplasy imaginable. 
 
 It may be admitted that wherever there is a 
 
GIZZARD IN FISHES in 
 
 distinct pyloric division of the stomach, the 
 material for a hard muscular gizzard exists. It 
 is obvious that the constitution of the alimen- 
 tary canal depends upon the nature of the food. 
 Comparative anatomy and physiology teach us 
 that the digestive tract reacts to different condi- 
 tions of nutrition; but not many such unequivocal 
 demonstrations of the independent acquisition, 
 within the limits of an order, of a compact 
 specialised structure as that described above, 
 can be pointed to. 
 
 Regarding the families of the Mugilidae and 
 Clupeidae in their entirety, we have now con- 
 sidered two characters with respect to which 
 they converge, namely, the adipose eyelids (above, 
 p. 94) and the pyloric gizzard. In both of these 
 cases, as well as in that of the pectoral fins of 
 the flying fishes, we have anatomically identical 
 structures arising independently from a common 
 origin. Facts of this nature apparently take the 
 ground away from any intelligible conception of 
 homology but only apparently. The relations 
 can be illustrated graphically, as shown in the 
 diagram (Fig. n). 
 
 The gizzard of the mullet was known to John 
 Hunter, 1 who tells us that of all the fish seen 
 
 1 John Hunter, " Observations on the Gillaroo Trout, commonly 
 called in Ireland the Gizzard Trout," Phil. Trans.) 1774; reprinted 
 in his "Animal Oeconomy," which was republished in vol. iv. of 
 Hunter's Works, edited by James F. Palmer, 1835-1837. Here 
 also is to be found Hunter's paper, " On a Secretion in the Crop 
 of Breeding Pigeons, for the Nourishment of their Young." 
 
ii2 SPECIAL CONVERGENCE 
 
 by him the mullet offered the most complete 
 instance of this structure, its strong muscular 
 stomach being evidently adapted, like the gizzard 
 of birds, to the two offices of mastication and 
 digestion. The stomach of the gillaroo trout, in 
 
 FIG. ii. Parallelism with convergence between two families 
 of fishes, Mugilidas and Clupeidas. 
 
 Hunter's experience, held the second place. But 
 the gizzard of Chatoessus, described above, holds 
 a position of equality with that of the mullet. 
 
 It should be added that Darwin ("Origin of 
 Species," p. 235) referred to the luminous organs 
 of insects of distinct families and diverse topo- 
 graphy as a phenomenon parallel with that of 
 the electric organs of fishes. 
 
CHAPTER VIII 
 
 HABITUDES AND ATTITUDES 
 (BIONOMICAL CONVERGENCE) 
 
 IDENTITY of habitat occasions convergence of 
 habits on the part of diverse animals. A great 
 deal of this has been implied in the preceding 
 chapters, but more remains to be said. The three 
 principal functions of animal life, at least above 
 the level of the sponge, are metabolism, 1 repro- 
 duction, and neuration, the last being at once the 
 master and the servant of the others, and leading 
 on from a lower to a more advanced cerebration. 
 The number of phyla or leading types of the 
 animal kingdom is considerable, and there are 
 only these three primary functions upon which 
 to ring the changes. 
 
 If there is one aspect of convergence more 
 widespread than another, it is that of cerebral 
 convergence, which is associated with the pheno- 
 menon of cephalisation or head-formation ; and 
 we can infer from this circumstance how hope- 
 less must be the search for neural homology as 
 
 1 Metabolism comprises nutrition, respiration, and excretion. 
 113 H 
 
ii4 HABITUDES AND ATTITUDES 
 
 between parallel types of the animal kingdom. 
 The cerebration of the ant is comparable, in 
 several of its manifestations, with that of the 
 higher mammals, but its evolution is distinct 
 and its mechanism different. The comparison 
 between the two forms is most profitable in the 
 abstract ; but it points away from homology, not 
 towards it. Perhaps either direction leads round 
 in a vicious circle to the same goal, namely, the 
 point from which we started ; but we may gather 
 much by the way if we do not stray too far from 
 the track and lose ourselves in a jungle of facts 
 and speculations. 
 
 National life is chiefly controlled by the desire 
 to capture markets. Animal life is chiefly con- 
 cerned with the occupation of feeding grounds. 
 In general, any given feeding area can support 
 a very mixed population, and the association of 
 forms which batten upon any particular source 
 of supply constitutes what has been termed the 
 bioccenosis [Dahl] of that centre. This com- 
 munity of interests is a phase of symbiosis or 
 commensalism, not based upon mutual advantage 
 nor even upon mutual toleration, in the first 
 instance, but upon strict independence and self- 
 help. 
 
 All the species which take part in such associa- 
 tions may be said to converge towards a common 
 centre of sustenance, and it is generally, though 
 by no means invariably noted that the assemblage 
 
ECTOPARASITES 115 
 
 is a heterogeneous one, the species composing it 
 belonging to so many genera, and the genera 
 usually belonging to different families or orders. 
 How far this community of habitat leads to 
 structural convergence is not clear, because the 
 anatomical characters of the associated forms 
 have not been worked out in sufficient detail 
 from this point of view. A case in point is the 
 overlapping of head and pronotum in some 
 ectoparasitic insects to which attention has been 
 drawn recently by Dr K. Jordan. 1 In Arixenia, 
 an apterous earwig found in the nursing pouch of 
 the naked bat of the Sunda Islands, the occipital 
 margin of the head is slightly concave, without 
 a sharp edge, and is not closely applied to the 
 pronotum. In Hemimerus, another apterous ear- 
 wig bearing a superficial resemblance to some 
 Blattidae, from the African murine genus Crice- 
 tomys, the hind edge of the head projects back- 
 wards, overlapping the pronotum to a slight 
 extent. Dr Jordan adds that this overlapping, 
 which is exceptional among insects, is best 
 known in fleas and in some Hemiptera parasitic 
 on bats. In the beaver parasite, Platypsyllus 
 castoris, the head and pronotum fit well together, 
 and there is a comb of spines extending from the 
 edge of the head on to the thorax, bridging over 
 the gap which might be formed when the head 
 
 1 K. Jordan, " New Apterous Earwig (Arixenia)J Novitat. 
 y xvi., 1909, p. 318. 
 
u6 HABITUDES AND ATTITUDES 
 
 is bent down. The overlapping between head 
 and pronotum of parasites which live in the fur 
 of mammals renders the surface uniform and 
 more suitable for gliding through the fur, and 
 is a secondary development which has taken 
 place independently in these not nearly related 
 insects [Jordan, loc. cit.~\. 
 
 The identification of the species of arthropods, 
 land-planarians, earthworms, and molluscs found 
 in or about a single fallen log is a matter of 
 almost unspeakable difficulty under the present 
 conditions of zoological nomenclature and publica- 
 tion. By far the greatest amount of biological 
 interest connected with a particular species may 
 be due to the association under which it was 
 found, but this is a matter of the least possible 
 systematic importance, except in the case of 
 parasites. The mere commingling of different 
 forms, apart from structural details, constitutes 
 a phase of our subject which is worthy of notice, 
 and may be termed bioccenotic convergence. 1 
 
 It is known and will be found recorded in 
 Darwin's "Monograph of the Cirripedia," that 
 barnacles are frequently found attached by their 
 peduncles to the skin of sea-snakes (Hydrophidse). 
 Occasionally a considerable cluster of barnacles 
 may be fixed upon the flattened tail of a slender 
 snake, as in the case of a Hydrus platurus 
 
 1 Such expressions as this are introduced merely for con- 
 venience of reference and classification of the phenomena. 
 
ASSOCIATIONS 117 
 
 from the coast of Ceylon, now preserved in the 
 Colombo Museum, whose tail was beset with a 
 group of barnacles composed of two species, 
 Lepas anserifera and Cone ho derm a hunteri. The 
 barnacles are not ectoparasitic, since they do not 
 feed upon the skin of the snake nor do they 
 assist the snake in any way ; on the contrary, 
 their presence must have seriously impeded its 
 movements. This snake is sufficiently protected 
 from larger enemies by its warning coloration 
 (black and yellow) and by its possession of poison 
 fangs. Moreover, the barnacles thrive equally 
 well when attached to floating bottles and drift- 
 ing spars, and the sea-snake in question was 
 merely their facultative vehicle. 
 
 The relation of barnacles to the skin of sea- 
 snakes is somewhat analogous to a remarkable 
 case of association between certain Hydroid 
 polyps (Stylactis minoi) and a small rock perch, 
 Minous inermis, which has been found in several 
 places off the west coast of India at depths of 
 45-150 fathoms. The skin of the fish is beset 
 with the commensal polyps which have never 
 been found elsewhere, and Colonel Alcock (op. 
 cit., 1902) thinks that they help to conceal the 
 fish from its enemies, in that they play the 
 same part which is, in other cases, performed by 
 frond -like, cutaneous filaments. In both cases 
 feeding is carried on independently of the verte- 
 brate host. The barnacles, named above, are 
 
 H 2 
 
n8 HABITUDES AND ATTITUDES 
 
 sedentary animals that require to be kept in 
 motion at the surface of the sea ; while the 
 hydroid is a sedentary species which requires an 
 adventitious mobility at the bottom of the sea. 
 
 The Tubificidae are a family of small fresh- 
 water Annelid worms, allied to earthworms, which 
 live in the mud of water - courses, sometimes 
 occurring in almost pure cultures of innumer- 
 able individuals. They keep the head and 
 fore-body buried in the mud, whilst the hinder 
 portion of the body, through which respiration 
 is effected, is kept constantly waving as near 
 the surface of the shallow water as possible. 1 
 When alarmed, an entire colony will instantly 
 withdraw out of sight into the mud as with one 
 consent. Living in the same environment and 
 sometimes in company with the worms are to 
 be found the larvae of midge flies (Chironomus]. 
 They have the same habit of waving the body 
 in the water and the respiratory processes occur 
 at the hinder end ; and, most singular coincidence 
 of all, the blood of the insect larvae is coloured 
 red with haemoglobin like that of the worms. 
 
 In his description of the larva of Chironomus, 
 commonly called the " Blood-worm," Professor 
 Miall (" Natural History of Aquatic Insects," 
 London, 1895) says: "When undisturbed, they 
 
 1 The habit of waving the posterior end of the body in the water 
 has been noted for Tubifex rivulorum by L. Atheston, Anat. Anz^ 
 xvi., p. 497 ; I have observed it in the case of an unnamed species 
 of Limnodrilm from Ceylon, as described in the text. 
 
SOCIAL INSECTS II9 
 
 may often be seen to push the head-end well 
 out of the burrow for purposes of feeding ; at 
 other times the tail - end is pushed out and 
 waved to and fro in the water, as a help to 
 respiration." This may be compared with 
 Atheston on Tubifex\ "It forms flexible tubes 
 open at both ends, of particles of soil cemented 
 by mucus. When contracted, the worm lies 
 wholly within the tube. When active, the 
 posterior end (third or half of the body) pro- 
 jects into the water, waving about in a rapid, 
 undulatory manner, while from the other end of 
 the tube is protruded the anterior end of the 
 worm which is thrust about through the mud 
 in search of food." Limnodrilus, as observed 
 by me in Ceylon, forms dense aggregates of 
 individuals surrounded by mud, but does not 
 form definite tubes which can be isolated from 
 the clumps. 
 
 No example of bionomical convergence is 
 more remarkable than that which is presented 
 by the social insects, ants, bees, wasps, and 
 termites. The termites so much resemble ants 
 in their mode of life and social organisation 
 that they are commonly known in the tropics 
 as white ants, though they are not ants and 
 are not always white. Not only have they 
 no direct genetic relationship with the true 
 ants (Formicidae, order Hymenoptera), but the 
 latter are among their most formidable enemies. 
 
120 HABITUDES AND ATTITUDES 
 
 A minute ant will overpower a termite twice 
 or three times its own size, seizing it from 
 behind round the middle, causing it to twist 
 and writhe, but never loosening its hold upon 
 the doomed victim. The social differentiation 
 of workers, soldiers, kings, and queens is the 
 same in the ant family and in the termite 
 family ; but the feeding-habits are quite differ- 
 ent, and termites do not keep slaves. The 
 nearer relationships of the termites are with 
 cockroaches and earwigs. Froggatt thinks that 
 the discovery of the giant termite from Port 
 Darwin, Mastotermes darwiniensis, "brings them 
 almost into touch with the family Blattidae." l 
 
 Although termites do not employ slaves as 
 some ants do, yet they entertain guests which 
 are known as termitophilous insects, as dis- 
 tinguished from the myrmecophilous insects 
 which frequent ants' nests. Interesting exhibi- 
 tions of mimicry and other special adaptations 
 result from these associations. The phase of 
 mimicry which involves the assumption of the 
 ant-facies is called myrmecoidism by Wasmann. 
 He distinguishes passive mimicry, or the decep- 
 tive resemblance of the outward form, from active 
 mimicry, which consists in an imitation of the 
 behaviour of the hosts. The mimicry which has 
 
 1 Walter W. Froggatt, "White Ants." Department of Agriculture, 
 New South Wales. Miscellaneous Publication No. 874, 1905 ; 
 containing a bibliography of papers dealing with Australian 
 termites. 
 
MYRMECOIDISM 121 
 
 for its end the deception of the hosts is further 
 distinguished from other forms of myrmecoidism 
 amongst Arthropoda ; one phase of the latter 
 is merely a morphological family resemblance 
 without apparent biological significance. 
 
 Some guests are welcome, others are hostile. 
 An example of the latter is a small beetle, 
 Myrmedonia, some species of which are myrme- 
 cophilous, whilst others are termitophilous. They 
 lie in wait in obscure corners of the nest, whence 
 they fall upon isolated ants and tear them to 
 pieces ; on the other hand, they have much 
 ado on their part to escape the vigilance of the 
 warrior ants. The myrmecophilous species of 
 Myrmedonia resemble the ants with which they 
 associate both in colour and in form ; so much 
 so that Wasmann himself has been repeatedly 
 deceived at a first glance. The object of this 
 mimicry is demonstrated by Wasmann to be 
 the optical delusion of the ants. 
 
 Among the numerous Oriental species of 
 termitophilous Myrmedonia there is no such 
 mimicry, although the Termes soldiers are as 
 capable of defence as are the ant soldiers, but 
 with this difference, that they are completely 
 blind. In the nests of the termite genus Hodo- 
 termes, whose soldiers possess well - developed, 
 facetted eyes, no Myrmedonia has been found. 1 
 
 1 E. Wasmann, "Die psychischen Fahigkeiten der Ameisen," 
 second edition. Published in Zoologica^ Bd. xi., Heft 26. Stuttgart, 
 
122 HABITUDES AND ATTITUDES 
 
 Another phase of bionomical convergence is that 
 contained in the phenomena of direct and indirect 
 development amongst many invertebrate animals. 
 Direct development is straightforward, gradual, 
 without striking metamorphosis, and must be 
 regarded, in a certain sense, as normal. Works 
 which deal with direct development are apt to 
 become standard and classical ; they give the 
 clue to fundamental homologies and generalisa- 
 tions. The theory of the coelom could hardly 
 have been established upon an indirect course 
 of development ; and on the other hand, the 
 relationship between the Enteropneusta and 
 Echinodermata would not have been recognised 
 so easily from the direct development, although 
 the larva of Asterina gibbosa would doubtless 
 have called for comparison with Bateson's larva 
 of Balanoglossus. 1 The differences between the 
 larvae of certain marine organisms consist in the 
 littoral or stereotropic habit of the direct forms 
 as contrasted with the pelagic or free-swimming 
 (pleotropic) habit of the indirect forms ; and they 
 are anticipated by the size of the egg and the 
 amount of food yolk it contains. A very small 
 egg will develop into a pelagic larva ; a relatively 
 large egg will develop directly into a larva in 
 
 1909, pp. 190, 5 plates. Contains bibliography of Wasmann's 
 numerous contributions to the knowledge of myrmecophily and 
 termitophily. 
 
 1 1 employ this name in the old sense, just as Amphioxus is still 
 extensively used, without trenching upon questions of nomenclature. 
 
TORN ARIA 123 
 
 which the form of the adult can be more or less 
 clearly discerned, unless, as in the case of the 
 above-named starfish, the adult form has suffered 
 a peculiar change from a bilaterally symmetrical 
 type. 
 
 The general name of the pelagic larval form 
 of those Enteropneusta (Balanoglossida), which 
 produce small ova not exceeding 0.15 mm. in 
 diameter, is Tornaria. All members of the 
 families Ptychoderidae and Glandicipitidae (or 
 Spengelidse) produce such eggs, The Harri- 
 manidae produce large yolky eggs from 0.4 mm. 
 in Dolichoglossus kowalevskii to as much as 
 1.5 mm. in Stereobalanus kupfferi. Tornaria 
 was discovered in 1848 by the famous German 
 physiologist Johannes Miiller of Berlin, who 
 thought that it was an Echinoderm larva ; its 
 real nature was determined in 1869 by Professor 
 Elias Metschnikoff. 
 
 It is well known how much the quantity of 
 food-yolk in the egg-cell affects the course not 
 only of the segmentation-stages, but also of the 
 subsequent embryonic and larval development; 
 and that, under normal conditions of nutrition, 
 the size of the egg is a fairly constant specific 
 character. 
 
 The dimensions of the eggs of different animals 
 do not always vary in proportion to the bulk of 
 the progenitors. Amongst existing birds, where 
 oviparity is the rule without any exceptions, the 
 
124 HABITUDES AND ATTITUDES 
 
 smallest species yield the smallest eggs and the 
 largest birds lay the largest eggs, but the inter- 
 vening sizes show great fluctuations, as may be 
 seen by comparing the egg of a Megapode with 
 that of a Jungle Fowl. 
 
 Amongst reptiles there are many exceptions 
 to the rule of oviparity, but there is no known 
 correlation between lecithality, or the relative 
 quantity of yolk in the egg, and viviparity ; this 
 latter habit not leading to a reduction of the 
 vitellus, although it may affect the shell-formation, 
 causing the complete absence of calcareous de- 
 posits, so that the embryo shows clearly through 
 its thin transparent envelopes. 1 The intra-uterine 
 development of a yolk - laden egg is a special 
 phase of incubation or brood - nursing, and has 
 been assumed independently by some arboreal 
 and deserticolous lizards of diverse families in 
 distant parts of the world. 
 
 A fundamental rule with regard to methods 
 of propagation is that oviposition preceded vivi- 
 parity, but this rule does not assist in determining 
 whether the ancestors of a particular class were 
 egg-layers or vivipars, and consequently whether 
 the egg-laying members of a group where vivi- 
 parity predominates are primitive in that respect. 
 Lecithality, oviparity, and viviparity are or may 
 be independent phenomena. It is probable, for 
 
 1 Cf. A. Willey, 1906, "Viviparity of Cophotis ceylanica and 
 Oviparity of Ceratophora stoddartii" Spolia Zeylanica^ vol. iii., 
 pp. 235-237 and figure. 
 
PERIPATUS 125 
 
 example, that the egg -laying habit of the 
 Australian monotremes is due to early regression 
 or descent with simplification in a primitive type, 
 while the lecithality of their eggs is not a palin- 
 genetic or high ancestral quality, but adaptive 
 or cenogenetic. 1 The abundance and scarcity 
 of yolk may be alike primitive and secondary 
 features within certain defined limits, in a manner 
 analogous with the plurality and duality of teats 
 in mammals. 2 
 
 It is not only among mammals that we find 
 large eggs associated with oviparity and small 
 eggs with viviparity and placentation ; and not 
 only among marine organisms do we find large 
 eggs associated with direct development and 
 small eggs with indirect development. 
 
 Peripatus is a soft-bodied, cryptozoic animal, 
 living in the tropics under logs and leaves, 
 whose body has the consistency of a caterpillar, 
 and whose feet are numerous like those of a 
 centipede but not jointed. It is known by no 
 other name except that of its own division of 
 the Appendiculata, namely, the Onychophora. 
 The class name Prototracheata was applied by 
 Moseley to this group, but since that time it 
 has been recognised that tracheate animals do 
 not form a homogeneous series, as will be ex- 
 plained more fully below. 
 
 1 This is Professor Hubrecht's view (/.., 1908). 
 
 2 Cf. A. Willey, "The Lacteal Tract of Loris gratilis? Spolia 
 Zeylanica, vol. iii., 1905 (1906), pp. 160-162 and figure. 
 
is6 HABITUDES AND ATTITUDES 
 
 The Onychophora furnish instructive gradations 
 from placentation, at one extreme, to oviposition 
 at the other, the sizes of the eggs increasing part 
 passu from minimal to maximal dimensions. The 
 species show a peculiar correlation between their 
 geographical distribution and their morphological 
 differentiation ; l and to some extent, though not 
 entirely, the modes of development vary with 
 the distribution. The placental condition of the 
 small - egged South American species is highly 
 specialised ; the oviparous habit of some Austra- 
 lasian species 2 is, to my mind, clearly secondary. 
 As for the species themselves, one form is not 
 appreciably more primitive than another. 
 
 In this group, therefore, the two extreme 
 methods of nutrition of the embryo are equally 
 cenogenetic. There still remain three inter- 
 mediate states of the embryo, namely, the tropho- 
 blastic development of the small - egged New 
 Britain species, and also of a South African 
 species ; the normal or direct development of 
 the Cape species ; and the yolky or meroblastic 
 development of some Australian and Malayan 
 species. 
 
 Just as among aquatic organisms which dis- 
 charge their spawn freely into the water, a 
 
 1 Cf. Adam Sedgwick, "The Distribution and Classification of 
 the Onychophora," Quart. Journ. Micr. Sc., vol. lii., 1908, pp. 
 379-406. 
 
 2 Cf. A. Dendy, " On the Oviparous Species of Onychophora, 5 ' 
 Quart. Journ. Micr. Sc.> vol. xlv., 1902, pp. 363-414. 
 
OPTIMUM CONDITIONS 127 
 
 distinction has to be made between direct and 
 indirect development, so in the intra - uterine 
 development of Peripatus a similar contrast 
 occurs. What is meant by a normal course of 
 development is one that is not disturbed by 
 adventitious factors. We must assume certain 
 optimum conditions which need by no means 
 be the most primitive and indeed on a priori 
 considerations are not likely to have been the 
 earliest conditions encountered by the amphibious 
 palaeozoic prototracheate ancestors of our terres- 
 trial arthropods. Optimum conditions may be 
 brought about by the acquisition of a little yolk 
 in the egg or by the loss of a little yolk ; and 
 this may make a considerable difference in the 
 behaviour of the embryos or larvae, as the case 
 may be. 
 
 All methods of development may be regarded 
 as more or less cenogenetic, and the selection of 
 a particular type as primitive may be purely 
 arbitrary ; but there can hardly be two opinions 
 as to a normal or direct course which can be 
 established as a legitimate standard or constant. 
 Any life-history can recapitulate certain ancestral 
 or palingenetic features, some more recent, others 
 more remote. Different species may provide 
 seemingly identical conditions for the maturation 
 of the ova and the development of the young ; 
 but it does not follow that the reactions or adap- 
 tations to these conditions will be identical. 
 
u8 HABITUDES AND ATTITUDES 
 
 When both conditions and reactions in different 
 specific groups are identical, convergence based on 
 homology ensues. Thus the trophoblastic vesicle 
 of Peripatus novcz-britannice is doubtless homo- 
 logous with that of the South African P. sedg- 
 wickii\ but the adaptation which produced the 
 structure out of the primitive matrix may have 
 been evolved independently, i.e., along closely 
 parallel lines in the two species. If this is so it 
 is an example of close specific convergence. 
 
 In accordance with the Trophoblast Theory, the 
 trophoblast, wherever it occurs, is to be defined 
 as an outer larval, or extra-embryonic envelope 
 or appendage which has acquired phagocytic 
 properties in order to provide for the intra-uterine 
 nutrition of viviparous, terrestrial animals which 
 have descended from oviparous, aquatic ancestors. 
 In accordance with the theory of Larval Forms 
 amongst aquatic organisms other than Arthro- 
 poda, what may be termed by contrast the 
 kinetoblast or outer ciliated investment of the 
 larva has acquired special locomotor properties, 
 frequently concentrated along definite tracts, in 
 order to provide for the pelagic life and inde- 
 pendent nutrition of the young. 
 
 From whatever point of view the matter may 
 be regarded, the parallelism between direct and 
 indirect development in Enteropneusta and 
 Onychophora, not to mention other cases, is a 
 very suggestive one. 
 
PALOLO 129 
 
 The swarming habit of some marine Annelids 
 for the purpose of breeding is worth mentioning 
 as affording a particularly interesting example 
 of bionomical convergence. The celebrated 
 palolo worm of Samoa and neighbouring parts 
 of the South Pacific is the occasion, in Samoa, 
 of an annual national festival owing to the 
 regularity and abundance of its swarms and to 
 its edible properties. It is eaten raw or baked 
 in leaves of the bread-fruit tree, and the natives 
 send presents of it to distant friends and to 
 the chiefs. 1 The most remarkable feature in 
 the biology of the palolo (Eunice viridis) is its 
 perfectly constant appearance in the months of 
 October and November when the moon is in 
 its last quarter. 
 
 S. J. Whitmee, who first reported upon this 
 phenomenon in 1875 (P- Zool. Soc., London, pp. 
 496-502), found that the palolo keeps astro- 
 nomical time ; as an indication of this it will 
 suffice to quote the recorded fact that in 1874 
 it reappeared, after an interval of thirteen lunar 
 months, on 3ist October and ist November. 
 Nineteen years later, in 1893, Dr Kramer col- 
 lected the material upon which A. Collin reported, 
 again on 3ist October and ist November. 
 
 Subsequently an Atlantic palolo manifestation, 
 occurring in the Tortugas during June and July, 
 
 1 Ant. Collin, 1897, " Bemerkungen iiber den essbaren Palolo- 
 wurm." In Kramer's " Bau der Korallenriffe." Kiel und Leipzig. 
 
 I 
 
1 30 HABITUDES AND ATTITUDES 
 
 on the part of another species, Eunice fucata, 
 was described by A. G. Mayer (1900); and 
 later still a Japanese palolo, belonging to a 
 different family (Nereidae), was described under 
 the name Ceratocephale osawai by Izuka. The 
 last-named author 1 says that the swarming of 
 the Japanese worm takes place in October and 
 November during the nights closely following 
 the new and the full moon, and it resembles 
 that of the Atlantic and Pacific species in its 
 general course, but differs in the circumstance 
 that whereas in the species of Eunice the sexual 
 segments occupy the posterior portion of the 
 body which becomes detached from the head- 
 end at the time of swarming, in the Japanese 
 worm the sexual segments are confined to the 
 anterior portion which, at the swarming season, 
 sheds the posterior, shrunken, asexual segments. 
 
 Amongst the numerous methods of brood- 
 nursing or care of eggs and young, we meet 
 with some extraordinary cases of convergence 
 of varying degree. Frequently they are obvious 
 enough, but that very fact only serves to 
 strengthen the argument which I am endeavour- 
 ing to develop, that convergence is not a sub- 
 ordinate but a dominant factor in morphology. 
 Firstly, let us take the phenomenon of buccal 
 
 1 Akira Izuka, " Observations on the Japanese Palolo," Journ. 
 Coll. Sci. Imp. Univ,, Tokyo, 1903, vol. xvii., article n, pp. i-37> 2 
 plates. 
 
BUCCAL INCUBATION 131 
 
 incubation, well known in point of time, but 
 not too familiar in personal experience. In 
 Ceylon it was first described by the Rev. B. 
 Boake (1867-1870) as occurring in the Siluroid 
 fish Arius falcarius, which has been determined 
 by Dr Francis Day to be synonymous with 
 Arius boakei. This fish frequents estuarine 
 waters and is very common in an extensive 
 backwater on the west coast of Ceylon called 
 the Panadure River. It has very large yolk- 
 laden eggs more than half an inch in diameter. 
 After being laid by the female they are found 
 nowhere except in the mouth of the male, where 
 they remain until they are hatched and until 
 the young have completely absorbed the yolk- 
 sac. I have observed that the intestine of 
 brood-nursing males is generally contracted to 
 narrow dimensions and empty, a fact which 
 was also noted by Day ; and as generally, the 
 palatine teeth of ovigerous males are greatly 
 reduced. In one case, however, which came 
 under my notice, where there were fifteen eggs 
 in the mouth, each containing an advanced 
 embryo, the palatine teeth were not appreciably 
 reduced and the hind-gut contained shell-dk#rw. 
 The opening of the oesophagus is constricted and 
 virtually closed, while the mouth and pharynx 
 are expanded to form a spacious brood-pouch, 
 where the eggs and young are exposed to a 
 constant current of water which passes over 
 
I 3 2 HABITUDES AND ATTITUDES 
 
 them and through the gill-clefts of the foster- 
 parent. Day observed the same incubation in 
 the allied genus Osteogeniosus, and Dr Gunther 
 and others had recorded them for South American 
 species of Arius. 1 
 
 Over against the buccal incubation of fishes 
 we may place that of the Chilian toad - like 
 Batrachian, Rhinoderma datwini, which was 
 regarded as one of the most interesting finds 
 of the voyage of the Beagle, although it is not 
 mentioned in Darwin's " Naturalist's Voyage." 
 The male of this toad possesses a median gular 
 sac, representing an extension of the buccal 
 cavity, which opens by two apertures on the 
 floor of the mouth. At the breeding season it 
 becomes a large brood-pouch, lying freely in the 
 ventral lymph-sinus and reaching back to the 
 pubic region. Jiminez de la Espada, who first 
 brought this fact to light in 1872, found as many 
 as fifteen metamorphosing larvae in the brood- 
 pouch. Howes 2 dissected a male in 1888 which 
 contained eleven larvae ; unlike the eggs and fry 
 of Arius, these larvae were unequally advanced, 
 only five of them being still provided with a tail. 
 
 Espada noted a shrinkage of the viscera as if 
 the foster-parent had ceased to feed during the 
 
 1 In the Japan-British Exhibition (1910) specimens of a Nile fish, 
 Tilapia nilotica^ are exhibted by Mr C. L. Boulenger, showing 
 females carrying their eggs and fry in the mouth. The eggs are 
 small. 
 
 2 G. B. Howes, " Notes on the Gular Brood-Pouch of Rhino- 
 derma dariuini" P. ZooL Soc. y London, 1888, pp. 231-237. 
 
CUTANEOUS INCUBATION 133 
 
 period of incubation as during that of hiberna- 
 tion; but Howes found the small intestine normal 
 and full of food-material in an assimilable condi- 
 tion, the large intestine fully charged with excreta, 
 and the stomach distended with small beetles and 
 diptera ; and he adds that the alimentary viscera 
 in general were those of a healthy animal in full 
 diet. Howes thought that Espada was mistaken, 
 but it seems possible that both conditions, hunger 
 diet and full diet, may occur. 
 
 In contrast with the* foregoing examples of 
 buccal incubation on the part of the male, we 
 may quote cases of cutaneous incubation on the 
 part of the female, as also in the male, both in 
 fishes and batrachians. In the Lophobranchii 
 or Pipe-fishes the cutaneous incubation of the 
 eggs by the male attains a high degree of per- 
 fection within the family Syngnathidae ; in the 
 allied family Solenostomidae the ventral fins 
 (absent in the Syngnathidae) are enlarged and 
 combine together to form a brood-chamber, with- 
 in which the eggs are borne upon cutaneous 
 discs, but in this case the female performs the 
 parental office. In the Siluroid genus Aspredo, 
 which occurs in Guiana, Dr Giinther described 
 the remarkable mode in which the female takes 
 care of her ova, carrying them, after oviposition, 
 attached to the spongy integument of the belly, 
 as the Surinam toad Pipa carries her ova on the 
 back. But a closer analogy than the Surinam 
 
 I 2 
 
i 3 4 HABITUDES AND ATTITUDES 
 
 toad is afforded by a Ceylon frog Rhacophorus 
 (Polypedates] reticulatus as described by Giinther. 1 
 Here the ova, about twenty in number, were 
 found attached to the abdomen of the female, 
 and when detached they adhered firmly together 
 so as to form a flat disc. In the European Mid- 
 wife toad, Alytes obstetricans, the mode of nursing 
 is analogous to that of Rhacophorus reticulatus, 
 except that in Alytes it is the male that takes 
 care of the spawn. 
 
 In illustration of the widespread nature of the 
 phenomenon of convergence may be mentioned 
 a case of brood-nursing on the part of some 
 insects. The family of amphibious water-bugs, 
 Belostomatidae, is noted for the possession of 
 the habit of carrying the eggs in the form of 
 a disc cemented upon the back of the male, to 
 which they are attached by the female. 2 A 
 species of this family where the habit can be 
 observed with comparative ease is Sphcerodema 
 rusticum Fabr., which occurs in the tanks of 
 Ceylon amongst the weeds at the margin. The 
 eggs remain on the back until the young hatch 
 out into the water. 
 
 Amongst fresh - water fishes many cases of 
 convergence are found in respect of the manner 
 in which they prepare their nests and deposit 
 their eggs. One pair of examples will suffice, 
 
 1 Ann. Nat. Hist., May, 1876, p. 379. 
 
 2 Cf. W. L. Distant, "Rhynchota," vol. iii., Fauna Brit. 
 1906, p. 34. 
 
FLOATING EGGS 135 
 
 namely, the Indian murral, Ophiocephalus striatus? 
 as compared with the American bowfin, Amia 
 calva, species which are far enough apart in the 
 systematic scale, the former a Teleostean, the 
 latter a Ganoid. Both make circular clearings 
 in the shallow water of lakes or irrigation tanks, 
 biting away the surrounding weeds ; and in both 
 cases the male parent tends the nest, lying in 
 wait in special runways. But whereas the eggs 
 of Amia are strewn over the bottom of the nest, 
 those of the murral float in a single layer at 
 the surface, in contact with one another, but not 
 adhering together. This exceptional behaviour 
 of the eggs of a fresh-water fish is shared by 
 those of another species of the same genus, 
 Ophiocephalus punctatus. 
 
 These floating eggs of Ophiocephalus owe their 
 buoyancy to the presence of a single large oil- 
 globule, which occupies the greater part of the 
 bulk of the ovum and is immersed in the amber- 
 coloured yolk adjacent to the uppermost pole 
 of the egg. The eggs thus come to lie immedi- 
 ately below the surface film of water, exposed 
 to the quickening influence of air and sun, and 
 protected thereby from the attacks of fungi, to 
 which they are extremely liable as soon as 
 the conditions of existence fall below a certain 
 
 1 A. Willey, " Observations on the Nests, Eggs, and Larvae of 
 O. striatus? Spolia Zeylanica, vol. vi., 1909] (1910), pp. 108-123. 
 The comparison with Amia is based on the description given by 
 Professor Bashford Dean. 
 
136 HABITUDES AND ATTITUDES 
 
 optimum. For three days after hatching the 
 larvse remain at the surface, floating on one side 
 with yolk - sac well up. Before hatching, the 
 body of the embryo encircles about two-thirds 
 of the equatorial region of the yolk like a belt ; 
 the tail then twitches, the vitelline membrane is 
 ruptured, and the larva with its yolk-sac is set 
 free. 
 
 The floating eggs of these tropical fresh-water 
 fishes are not comparable with the hyaline 
 pelagic eggs of many marine fishes because they 
 do not move from the nest ; their buoyancy is part 
 of the method of nidification and brooding, not 
 a means of dispersal. I do not know any other 
 instances of eggs of fresh-water fishes floating at 
 the surface of the water by their own buoyancy ; 
 but the same advantages proximity to atmos- 
 pheric air and to sunlight are secured in other 
 ways, as by attachment to aquatic plants, or by 
 deposition in very shallow water, or by special 
 methods of nidification, as in the floating nests 
 of Gymnarchus or the foam nests of Sarcodaces, 
 which were described by the late J. S. Budgett. 1 
 There would seem, however, to be points of closer 
 comparison between the eggs of the murral and 
 those of some marine fishes, namely, the weevers 
 (Trachinidse), according to the observations of 
 J. Boeke (1903). These eggs are maintained 
 
 1 Cf. The Budgett Memorial Volume, edited by J. Graham 
 Kerr, Cambridge Univ. Press, 1907. 
 
YOLK-SACS 137 
 
 at the surface of the sea by means of a single 
 large oil-globule ; four or five days after oviposi- 
 tion the embryos are hatched ; four or five days 
 after hatching the yolk has become absorbed. 
 The buoyancy of the yolk-sac causes the larvae 
 to float helplessly for some time after hatching, 
 with the yolk-sac uppermost. 
 
 The accumulation of yolk in eggs, or the 
 lecithality of the ovum, also offers numerous 
 examples of convergence as between cartilaginous 
 fishes, bony fishes, and sauropsida (birds and 
 reptiles) as well as amongst invertebrate animals. 
 The large yolk-sacs of Arius and of Gymnarchus 
 have no genetic relation to each other nor to 
 those of sharks and rays. The yolk - sac of 
 cephalopod molluscs is a structure sui generis^ 
 and it is particularly noteworthy that the most 
 ancient existing genus, Nautilus, has the most 
 macrolecithal egg of all whose eggs are known. 
 
CHAPTER IX 
 
 THE WAYS OF BREATHING 
 (RESPIRATORY CONVERGENCE) 
 
 THE factors which combine to produce a structural 
 or organic unit in the animal body which will be 
 fixed by inheritance are inconceivably complex. 
 Nevertheless, we have seen that essentially the 
 same combination can be repeated independently 
 in different cases. The present state of know- 
 ledge justifies the provisional assertion that the 
 higher combination which leads to the establish- 
 ment of an animal form possessing the essential 
 component elements of a definite morphological 
 type, cannot be repeated. The theory of con- 
 vergence is therefore not calculated to precipitate 
 us into morphological chaos, howsoever startling 
 its manifestations may be. 
 
 A great deal might be written upon the large 
 subject of respiratory convergence, but I will 
 touch upon it briefly. Respiration is one of 
 the primary properties of living matter, and the 
 general principle which governs the mechanism 
 of respiration, namely, the diffusion of gases 
 (oxygen and carbon dioxide) through moist 
 
 138 
 
MECHANISM OF RESPIRATION 139 
 
 membranes or across moist surfaces, is common 
 to all animals and even to plants as well ; but 
 the special mechanism of respiration exhibits 
 great phyletic diversity. There are four or five 
 principal methods of breathing : cutaneous, by 
 the entire surface of the body, as with earth- 
 worms, leeches, and planarians ; branchial, by 
 specialised cutaneous processes called gills or 
 branchiae, as with many marine Annelid worms, 
 some leeches (Branchellion and Ozobranchus], 
 molluscs, and Crustacea ; tracheal, by air-tubes 
 traversing the body and surrounding the viscera, 
 as with insects and spiders ; trematic, by gill- 
 clefts or visceral clefts piercing the body-wall 
 and leading from the cavity of the pharynx or 
 anterior part of the alimentary canal to the 
 exterior ; pulmonary, by lungs or air-chambers, 
 as with pulmonate molluscs (Gastropods) and 
 air-breathing vertebrates. 
 
 These different methods of respiration are not 
 merely adaptations to the environment, but they 
 are adaptations which keep pace with evolution 
 independently of the environment. The super- 
 ficial resemblance in shape of body, swimming 
 movements, and burrowing habits, between the 
 Annelid worm Ophelia, which breathes by simple 
 filiform, cutaneous gills, and Amphioxus, which 
 breathes by gill-clefts, has been remarked by 
 Cav. Lo Bianco and by myself; they are often 
 taken together in the same sandy bottom, in the 
 
i 4 o THE WAYS OF BREATHING 
 
 Mediterranean and in the Indo-Pacific. The 
 environment is identical ; the adaptations are 
 different according to their respective grades of 
 evolution and lines of descent. Again, in many 
 ways the Annelida are more highly organised 
 than the Enteropneusta, though the latter possess 
 gill-clefts. 
 
 An equal degree of resemblance based on con- 
 vergence between members of distinct though 
 allied families is that of the Jumping Blenny, 
 Salarias, and the Jumping Goby, Periophthalmus. 
 These fishes habitually come out of the water, 
 the former to lounge and skip about rocks, the 
 latter upon mud and mangrove roots ; they both 
 have large goggle-eyes and both can leap about 
 as a normal mode of progression out of water, 
 while the jumping goby can also ricochet over 
 the surface of the water. As described by 
 Moseley 1 and again by Hickson, 2 Periophtkalmus 
 jumps both out of water and on the surface by 
 means of the bent, muscular, pectoral fins, of 
 course assisted by the tail. Salarias, which is 
 common on certain parts of the coast of Ceylon, 
 performs surprising leaps by the action of its 
 tail, which is kept curved when on the ground 
 ready for a spring. Any one familiar with Peri- 
 ophthalmus, which is one of the everyday sights 
 in suitable localities in the Eastern Tropics, 
 
 1 H. N. Moseley, " Notes by a Naturalist on H.M.S. Challenger? 
 second edit. London, 1892. 
 
 S. J. Hickson, " A Naturalist in North Celebes." London, 1889. 
 
AIR-BREATHING FISHES 141 
 
 would be certain to mistake Salarias for it 
 unless otherwise instructed. When out of water 
 the opercular membrane is kept closely pressed 
 against the body behind the gill-opening, so that 
 the gill - cavity is temporarily converted into a 
 virtual lung-chamber. 
 
 Other fishes which can progress out of water 
 on suitable ground without falling on to one side 
 areAnabas, Clarias, Saccobranchus, Ophiocephalus, 
 and others. These may all be designated walk- 
 ing fishes. Anabas helps itself along by means 
 of its opercular spines, Clarias and Saccobranchus 
 by their pectoral spines, Ophiocephalus by move- 
 ments of its flattened head, assisted by flexions 
 of body and tail. All of these fishes can and 
 must breathe air by means of special growths or 
 diverticula connected with the upper division of 
 the gill-cavity above the gill-clefts. Clarias and 
 Saccobranchus are both Siluroids, but the coral- 
 like dendritic appendages in the suprapharyngeal 
 chambers in Clarias are much more like the 
 lamelliform labyrinthine organs in Anabas than 
 the diverticula of Saccobranchus. If these air- 
 breathing Teleostean fishes are prevented from 
 reaching the surface in order to take the air, they 
 become drowned ; and this habitual aeropneustic 
 function of the accessory branchial organs was, 
 when first demonstrated, regarded by Professor 
 Huxley as a great fact. 1 Saccobranchus will live 
 
 1 Rev. Barcroft Boake, " On the Air-Breathing Fish of Ceylon," 
 Journ. Ceylon Branch Roy. Asiat. Soc. t iv., 1867-1870. 
 
I 4 2 THE WAYS OF BREATHING 
 
 longer out of water than it will when kept in 
 water and deprived of access to the surface, 
 but it cannot withstand desiccation, nor will it 
 voluntarily leave the water when the latter is 
 polluted ; if its excursions to the surface are not 
 sufficient to counteract the poisonous effects of 
 the gaseous exhalations from the muddy bottom 
 of a stagnant pool, it will perish miserably, some- 
 times coincidently with its last gulp of air, as 
 I have witnessed. 
 
 The above-named fishes are amphipneustic 
 without being amphibious, and in this respect 
 are comparable, by way of convergence, with the 
 fresh- water pulmonate Mollusca (Limnseidse), and 
 with the Dipnoan fishes or Dipnoi ; whilst the 
 latter are comparable, by way of homology, with 
 the Amphibia. The respiratory movements of 
 the legless batrachian, Ichthyophis, which, with 
 the loss of appendages, has still retained many 
 primitive features, involve the entire branchial 
 or pharyngeal region including the throat, rapid 
 contractions shimmering over from the ventral to 
 the dorsal surface, while the nasal orifices are 
 kept perpetually open. The original branchial 
 region is clearly marked off like a collar from 
 the rest of the body during life (Fig. 12), and 
 the impression conveyed, which is in accord- 
 ance with its known anatomical structure, 1 is 
 
 1 Cf. the elaborate monograph by Drs P. and F. Sarasin. The 
 incubation of its eggs, the parent coiling round the egg-clump, 
 resembles that of a Python. 
 
ICHTHYOPHIS 143 
 
 that while the essential organs of respiration 
 have changed from gill - clefts to lungs, the 
 muscular and skeletal mechanism of respiration 
 has remained practically unchanged. It may be 
 added that Ichthyophis is a good swimmer but 
 tries to get out of water as soon as possible, 
 creeping with difficulty on a free surface by 
 serpentine jerks, but moving easily through 
 narrow crevices where its contact requirements 
 (i.e., stereotropism) are satisfied. Similar obser- 
 vations with regard to the respiratory movements 
 have been made by my friend Professor Graham 
 Kerr (in Hit.} on the Dipnoan fish Lefidosiren. 1 
 
 FIG. 1 2. Head and fore-body of Ichthyophis glutinosus from above, 
 to show the collar-like respiratory region. At the sides of 
 the head in front of the eyes is the pair of peculiar retractile 
 tentacles. 
 
 The fact that in the course of the substitu- 
 tion of lungs for gills the respiratory movements 
 
 1 For a discussion of the nerve-supply see J. Graham Kerr, 
 "Note on Swim-Bladder and Lungs," Proc. Roy. Physical Soc. t 
 Edinburgh, vol. xvii., 1908, pp. 170-174. 
 
144 THE WAYS OF BREATHING 
 
 are continued without interruption from the one 
 to the other system, shows how the transition 
 could be effected, but it does not, in my opinion, 
 necessarily support Spengel's and Goette's theory 
 of the origin of vertebrate lungs and air-bladders 
 from gill-pouches. 1 This is an interesting theory, 
 and adds weight to the conception of the gill- 
 cleft as an autonomous "morphon," but there are 
 difficulties in the way of its acceptance. In the 
 first place, from the analogy of hydrostatic organs 
 in other phyla, it does not seem necessary to 
 invent a factitious explanation of the air-bladder 
 as a gill-pouch derivative ; the accumulation of 
 air or gas in air-chambers is a widely distributed 
 phenomenon, and in some fishes (e.g., the Globe 
 Fishes) the oesophagus itself can be converted 
 into an air-chamber at the will of the animal. 
 
 There seems to be no particular difficulty in 
 supposing that the air-bladder arose as a single 
 or paired diverticulum of the fore-gut for the 
 storage of air; and whether the first function 
 was respiratory or hydrostatic or both, is another 
 question. 2 Spengel lays great stress upon the 
 
 1 J. W. Spengel, " Ueber Schwimmblasen, Lungen und Kiemen- 
 taschen der Wirbelthiere," Zool. Jahrb. Suppt., 1904, pp. 727-749. 
 In this paper the term "morphon" is suggested; it means a 
 morphological unit or element. 
 
 1 Facts in support of the view that the original function of the 
 air-bladder was respiratory were brought forward in a paper by 
 Charles Morris on " The Origin of Lungs, a Chapter in Evolution," 
 in The American Naturalist, vol. xxvi., 1892, pp. 975-986. This 
 opinion is tentatively supported in principle by Hubrecht (loc. "/., 
 1908), who refers to Assheton's paper on " Gymnarchus " in the 
 Budgett Memorial Volume. 
 
MEGALOPS 145 
 
 fact that the arteries which supply the lungs of 
 Dipnoi and the air-bladder of some other fishes 
 (e.g., Polypterus) and the lungs of all air-breathing 
 vertebrates from the Amphibia onwards, arise as 
 branches of the fourth afferent branchial arteries. 
 But Burne x has shown that the accessory branchial 
 diverticula or air-pouches of Saccodranckus, which 
 penetrate the body musculature throughout the 
 greater part of the length of the trunk from the 
 gill-cavity to the tail, lying above the transverse 
 processes on either side of the vertebral column, 
 are also vascularised from the fourth afferent 
 branchial artery on each side ; but these air- 
 pouches co-exist with a true air-bladder. We 
 have, in fact, in this case a very delicate example 
 of vascular convergence. 
 
 The air-bladder of the estuarine Clupeoid fish 
 Megalops cyprinoides, common in Ceylon, is pro- 
 vided on its inner surface with an abundance 
 of spongy, vascular, alveolar proliferations which 
 are especially dense just behind the large orifice 
 by which it communicates directly, without the 
 intermediation of a pneumatic duct, with the 
 oesophagus immediately behind the gill - clefts 
 dorsally ; it is not vascularised from the branchial 
 arches, but this species can live for a long 
 time after capture, an uncommon feature in the 
 
 1 R. H. Burne, " On the Aortic Arch System of Saccobranchus 
 fossilis? Journ. Linn. Soc. Zool^ xxv., 1894, pp. 48-55. 
 
 K 
 
146 THE WAYS OF BREATHING 
 
 herring family and contrary to the experience 
 of the common herring. 
 
 The Siluroid fish Arius falcarius, of which 
 mention has been made above, exhibits great 
 viability out of water. A large female, 14^ 
 inches long, which had been caught by a hook 
 but had not been injured, lived for nearly five 
 hours out of water, breathing regularly by 
 mouth and gills, without accessory structures, 
 closing the gill - cavity behind by the opercular 
 membrane. After it had been out of water for 
 about an hour I counted 80 buccal respirations 
 to the minute, the opercular membrane beating 
 time with the mouth but remaining quite closed 
 behind. After two hours there were 70 respira- 
 tions to the minute ; and after three hours 60 
 feebler respirations, the opercular membrane now 
 commencing to gape behind. At the end of 
 the experiment the heart was removed and con- 
 tinued beating outside the body. This example 
 shows that the phenomenon of viability out of 
 water is something apart from the possession 
 of accessory branchial organs. 
 
 Another Siluroid fish, Plotosus canius, exhibits 
 similar viability, living from early morning until 
 afternoon (over six hours) out of water, the gills 
 remaining quite fresh to the end. In this case 
 periodical expirations took place through the 
 gill-opening ; every now and then the opercular 
 membrane was raised several times in succession, 
 
PLOTOSUS 147 
 
 and this was followed by a number of buccal in- 
 spirations without the gill-covers being reflected. 
 The raising of the gill-covers coincides with the 
 taking of a deep breath. I counted first 32 in- 
 spirations and 13 expirations in a minute, then 
 32 inspirations to 10 expirations ; a third count- 
 ing gave 30 and n, a fourth 29 and 12. The 
 expirations were not evenly distributed, some- 
 times 5 or 6 occurring together. These observa- 
 tions show the great physiological importance of 
 the opercular membrane whose systematic worth 
 is sometimes regarded as nil. 
 
 The typical Molluscan gill was named the 
 ctenidium by Lankester 1 in order to fix its 
 morphological independence and to distinguish 
 it from other gills of invertebrates such as the 
 Annelid parapodial gill, the Crustacean arthro- 
 podial gill, the Limuloid gill-book, and others, 
 as well as from secondary Molluscan pallial gills, 
 as in the Prosobranchiate genus Patella (limpet) 
 and in the Opisthobranchiate genera Phyllidia 
 and Pleurophyllidia. Thus the organs of re- 
 spiration which take the form of cutaneous gills 
 are not homologous throughout ; in other words, 
 they are not monophyletic but polyphyletic in 
 nature and origin ; but the particular structure 
 known as the ctenidium with which a special 
 sensory apparatus, the osphradium and its nerve- 
 supply, is associated, is homologous throughout 
 
 1 Zoological Articles (Encyc. Brit.\ Edinburgh, 1891. 
 
148 THE WAYS OF BREATHING 
 
 the Molluscan phylum, that is to say, the cten- 
 idium is monophyletic. 
 
 The spirally thickened air - tubes or tracheae 
 of Myriopods, Insects, and Spiders bear so close 
 a correspondence in structure and function, rami- 
 fying through the body like blood-vessels, but 
 effecting the circulation of air instead of blood, 
 that all the air - breathing Arthropods were 
 formerly classed together as Tracheata in con- 
 tradistinction to the aquatic Arthropods which 
 were called Branchiata. 
 
 For more than a quarter of a century it has 
 been recognised that the Tracheate Arthropods 
 could not be reduced to a common standard, and 
 it has also come to be realised that the tracheae of 
 Insects and Arachnids have had separate origins, 
 and are therefore different morphologically though 
 similar histologically and physiologically. 
 
 This fundamental example of tracheal con- 
 vergence is rendered more remarkable by the fact 
 that even within the limits of the Arachnoid sub- 
 phylum, the tracheae have had at least a twofold 
 origin, namely, from lung-books and from ecto- 
 dermal tendons; so that "similarity of structure 
 in the fully developed tracheae does not mean 
 similarity of origin " [Purcell]. Pureell has shown 
 that the tendinal or medial tracheal trunks in 
 Dipneumonous spiders are equivalent in their 
 entirety to metamorphosed entapophyses (ecto- 
 dermal tendons) ; the lateral tracheal trunks, on 
 
TRACHEAL CONVERGENCE 149 
 
 the contrary, are serially homologous with the 
 pulmonary sacs of the preceding somite, and 
 actually homologous with the second pair of 
 pulmonary sacs in Tetrapneumonous spiders. 1 
 
 Purcell affirms that the arguments in favour of 
 the branchial origin of the lung-books of spiders, 
 advocated by Lankester in 1881, appear over- 
 whelming. An interesting analogy may there- 
 fore be drawn between the Pulmonate Arachnida 
 and the Pulmonate Mollusca, where a lung- 
 chamber has likewise been substituted for a 
 gill-chamber ; and just as in Arachnida, accord- 
 ing to Pocock (1893), there is reason to believe 
 that tracheal tubes have replaced lung-books at 
 least twice, namely, in the Dipneumones and in 
 the Pseudoscorpiones ; so in the Land Mollusca 
 a pulmonary chamber has twice replaced a gill- 
 chamber, namely, in the Land Operculates, which 
 belong to the order Prosobranchiata, and in the 
 inoperculate Pulmonata, which are related to the 
 Opisthobranchiata. 
 
 The internal tracheae (of cutaneous origin) 
 are not homologous throughout the Arthropod 
 phylum they are polyphyletic like the external 
 cutaneous branchiae of soft-bodied invertebrates. 
 
 Now with regard to the trematic mode of 
 aquatic respiration, or breathing by means of 
 gill-clefts, the question naturally arises whether 
 
 1 W. F. Purcell, " Development and Origin of the Respiratory 
 Organs in Araneae," Quart. Journ. Micr. Sc.> vol. liv., 1909, pp. 
 l-no ; contains full bibliography. 
 
 K 2 
 
i S o THE WAYS OF BREATHING 
 
 the latter are another generalised form of re- 
 spiratory organ which may have arisen de novo 
 in different divisions of the animal kingdom, or 
 whether gill-clefts are not rather a specialised 
 formation, homologous throughout, like the 
 Molluscan ctenidium. The analogies of the 
 cutaneous branchiae and tracheae render the 
 question a legitimate one for discussion and 
 an extremely difficult one to settle. Many 
 zoologists, of whom I happen to be one, think 
 that the gill -cleft is a monophyletic structure, 
 and in order to give effect to this point of 
 view I suggested, as a memoria technica, the 
 phylogenetic term Branchiotrema, 1 to include all 
 animals possessing gill-clefts at any period of 
 their life-history. 
 
 The extraordinary persistence of gill -clefts 
 and gill-pouches in the embryos of the higher 
 lung-breathing vertebrates, the constancy of their 
 innervation in all Craniate vertebrates, and the 
 combination of structures which accompany them 
 in all Chordates, speak for their homogeneity. 
 Whilst we may admit that the general homology 
 of gill-clefts is open to question, we must at the 
 same time assert that the burden of proof to the 
 
 1 Cf. A. Willey, " Enteropneusta from the South Pacific," Zoo- 
 logical Results^ part iii., Cambridge, 1899, pp. 223-334. See also 
 R. C. Punnett's memoir on the Enteropneusta from the Maldive 
 and Laccadive Islands in Stanley Gardiner's " Fauna and Flora 
 of the Maldive and Laccadive Islands," 1904-1905. 
 
GILL-CLEFTS 151 
 
 contrary rests with those who deny it. Up to 
 the present there has been no brilliant demonstra- 
 tion of the diphyletic origin of gill-clefts. The 
 assumed homology between the gill-clefts of the 
 Enteropneusta and those of Amphioxus has been 
 doubted because their histological composition is 
 not identical ; but we know now that histological 
 identity is no safer guide to morphology than is 
 anatomical identity. This is where the matter 
 rests at present with regard to trematic respira- 
 tion. The differences between the gill-clefts of 
 the Enteropneusta where the tongue-bar is the 
 principal component, and the gill-clefts of Amphi- 
 oxus where the tongue-bar is a secondary com- 
 ponent, are such as one would be prepared to 
 find in allied groups which present unequal grades 
 of organisation. 
 
 We have already mentioned several examples 
 of convergent pulmonate respiration, and it only 
 remains to be added, for the sake of complete- 
 ness, that the lungs of the higher vertebrates 
 are acknowledged, on grounds of comparative 
 anatomy, to be homologous with the air-bladder 
 of Teleostean fishes in spite of differences in the 
 vascularisation and topography of these organs. 
 
 According to Owen (Preface to Hunter's 
 Animal Oeconomy, 1837), Harvey was the first 
 to compare the abdominal air-sacs of the bird 
 with those of reptiles and fishes ; and in this he 
 was followed by John Hunter, who discovered the 
 
152 THE WAYS OF BREATHING 
 
 " air-cells " in the bones and muscular interstices 
 of birds in 1774, simultaneously with Camper. 
 The abdominal air-sacs of birds are appendages 
 of the lungs, and although there exists a general 
 homology between lungs and the air -sac of 
 fishes, there is no special homology between the 
 abdominal extensions of the lungs in birds and 
 the latter. 
 
CHAPTER X 
 
 CONVERGENCE IN MINUTE STRUCTURE 
 (HISTOGENETIC CONVERGENCE) 
 
 IN the preceding chapter we have seen that 
 histogenetic convergence gives no clue to affinity, 
 and that histogenetic divergence is no proof of 
 want of affinity. We may now go further than 
 this and add that where we do find actual histo- 
 logical identity, as between members of different 
 phyla, it seems certain that we are in the presence 
 of true convergence in the sense in which that 
 term is employed here ; and, in the light of facts 
 which are now available, it even begins to appear 
 strange, although only a matter of a few years 
 or months ago, that histological identity should 
 ever have been insisted upon as a criterion of 
 homology except within well-defined limits. 
 
 To my thinking, one of the most remarkable 
 examples of histogenetic convergence is that of 
 the excretory organs of Amphioxus as com- 
 pared with those of certain Annelid worms. 
 The excretory tubules of Amphioxus were dis- 
 covered independently by Weiss and Boveri 
 in 1890 and were described in detail by the 
 
154 CONVERGENCE IN MINUTE STRUCTURE 
 
 latter in 1892. Attached in clusters to the 
 walls of the tubules Boveri found long pin- 
 shaped cells, which he called " Fadenzellen " 
 (thread-cells), projecting into the dorsal coelom. 
 They appeared to consist of a small roundish 
 cell-body mounted upon a long thread-like stalk 
 which was inserted into the tubule. Ten years 
 later Goodrich, who had been making some 
 notable discoveries in regard to the structure of 
 the nephridia of Polychsete worms, turned his 
 attention to the tubules of Amphioxus and 
 promptly discovered that Boveri's thread-cells 
 are identical with the solenocytes which he had 
 found in Polychaeta. 
 
 Goodrich showed that Boveri's thread - like 
 stalk is really a slender hollow tube of great 
 length, carrying the cell - body floating in the 
 coelom, or adhering to the adjacent walls of 
 the coelom, at its distal end, and opening into 
 the renal tubule at the proximal end. " A long 
 flagellum attached at its base to the cell placed 
 at the end of the tube works rapidly down the 
 tube and far into the excretory canal." 1 Goodrich 
 observes that the excretory organs of Amphioxus 
 and the nephridia of Phyllodoce are in all essentials 
 identical, and he adds that " if two such excretory 
 organs as the solenocyte - bearing nephridia of 
 Phyllodoce, and the solenocyte-bearing kidneys 
 
 1 E. S. Goodrich, "On the Structure of the Excretory Organs of 
 Amphioxus," Part i., Quart. Jour. Micr. Sc.^ vol. xlv., 1902, pp- 
 493-501 ; Part ii. Ibid. vol. liv., 1909, pp. 185-205. 
 
SOLENOCYTES 155 
 
 of Amphioxus could be shown to have been inde- 
 pendently evolved, we should have to give up 
 structural resemblance as a guide to homology." 
 He says in a footnote that the only case which 
 seems to him at all comparable is that of the 
 nematocysts in Coelenterates, Planarians, and 
 Molluscs. To this case we may add the 
 myoepithelial cells in Coelenterates, Nematodes, 
 and Tunicata ; and, as Goodrich admits, the 
 flame-cells of flatworms, Rotifers, and Polyzoa 
 (Entoprocta) are probably of the same nature 
 as his solenocytes. 
 
 I venture to interpret Goodrich's discovery 
 as a brilliant demonstration of histogenetic con- 
 vergence and submit the following explanation, 
 basing the argument on personal conviction, on 
 an appeal to known facts, and on certain general 
 considerations : 
 
 Just as various offices are constituent parts 
 of the body politic, so various organs are con- 
 stituent parts of the animal economy. Pharynx, 
 oesophagus, crop, gizzard, stomach, liver and 
 other diverticula, intestine and rectum, are 
 constituent parts of the digestive tract of a 
 coelomate animal and are likely to appear when 
 called into requisition by the necessities of 
 adaptation and evolution ; and equally likely to 
 disappear when not specially required. I may 
 refer here to the remarkably simple digestive 
 tract of the Scombresocidse (e.g., B clone), where 
 
156 CONVERGENCE IN MINUTE STRUCTURE 
 
 the oesophagus is suppressed, the outward differ- 
 entiation of the stomach is lacking, and there 
 is no pylorus and no pyloric cceca. So again, 
 continuing our analysis, we find that flame-cells 
 or solenocytes, nematocysts, myoepithelial cells, 
 spicule-forming cells, ganglion cells, pigment cells, 
 striated muscle cells, etc., may be regarded as 
 belonging to the category or repertory of the 
 primary constituent elements of animal tissues, 
 and will appear when required to fit in with 
 a special set of physiological conditions. Let 
 us imagine a limited number of primary physio- 
 logical conditions of excretion. Some or all of 
 these conditions will recur independently in each 
 of the phyla of the animal kingdom. To meet 
 them, the necessary cellular elements, till then 
 held in abeyance, will be forthcoming. 
 
 Bewildering as may be the striking identity 
 which Goodrich has established between the 
 solenocytes of Amphioxus and of Polychaetes, it 
 is perhaps not more remarkable in principle than 
 other cases of identical anatomical differentiation, 
 some of which have been referred to in the 
 preceding chapters. 
 
 As might have been expected, this case of 
 nephridial convergence has been hailed with 
 delight by Dr Gaskell, who looks upon it as 
 an acceptable confirmation of his deductions. I 
 may be permitted to make the following quota- 
 tion from his book, whilst at the same time 
 
SOLENOCYTES 157 
 
 strongly recommending the reader, if he has 
 not done so already, to refer to the original 
 volume, which is a very remarkable document 
 of research. 1 The citation will serve to illustrate 
 Dr Gaskell's point of view in this matter, and 
 I make no further comment upon it, relying 
 upon what has gone before except to point 
 out that it is not quite correct to say that the 
 Polychaeta as a whole are the highest forms of 
 Annelida. It is precisely among the Polychaeta 
 that the nervous system is often in contact 
 with the epidermis, and it is here that the 
 cerebral ganglion retains its primitive position 
 in the prostomium. 
 
 This is what he says : " It is to me most 
 interesting to find that the very group of 
 Annelids, the Polychaeta, which possess soleno- 
 cytes so remarkably resembling those of the 
 excretory organs of Amphioxus, are the highest 
 and most developed of all the Annelida. I 
 have argued throughout that the law of evolu- 
 tion consists in the origination of successive 
 forms from the dominant group then alive, 
 dominance signifying the highest type of brain- 
 power achieved up to that time." The ways of 
 evolution are obscure and peculiar ; logically 
 one would think that they ought to keep 
 pace with the increase of brain-power ; naturally 
 
 1 W. H. Gaskell, "The Origin of Vertebrates," London, 1908, 
 see p. 395. 
 
158 CONVERGENCE IN MINUTE STRUCTURE 
 
 we find that the increase of brain - power is 
 merely an incident in evolution. 
 
 It will be appropriate at this point to consider 
 another example of what I hold to be histo- 
 genetic convergence as between the lateral sense- 
 organs of the Polychaete family of the Capitellidse, 
 the abdominal sense-organs of Lamellibranchiate 
 Molluscs, and the lateral line sense-organs of 
 Vertebrates, thus involving three distinct phyla. 
 The lateral organs of Capitellidae are absolutely 
 comparable to the lateral line organs of Verte- 
 brata, but only by way of convergence, not by 
 way of homology. 
 
 Let us first look at the abdominal sense-organs 
 of the bivalve molluscs, about which I am able to 
 speak with some authority inasmuch as, in one 
 instance, namely, that of the Windowpane Oyster, 
 Placuna placenta, I have found the true unpaired 
 abdominal sense - organ which had been over- 
 looked by previous investigators. The story of 
 our knowledge of these sense-organs is interest- 
 ing. The year 1881 saw a considerable advance 
 in the morphology of the Mollusca in consequence 
 of the publication of Spengel's paper on their so- 
 called olfactory organs, subsequently called the 
 osphradia by Lankester, and on the nervous 
 system. In dealing with the Lamellibranchiata 
 he was at first at a loss where to look for them, 
 but, thanks to a lucky chance (" einem gliicklichen 
 Zufalle "), on opening an Area noa he saw at once, 
 
ABDOMINAL SENSE-ORGANS 159 
 
 between the hinder end of the foot and the vent, 
 a transverse, undulating line of greenish-brown 
 pigment interrupted in the middle line by a 
 narrow interval. That line proved to mark the 
 position of the pigmented sense-organs for which 
 he was searching. In the figure which he gave 
 to elucidate the topography of these organs, 
 another pair of organs, on either side of the vent, 
 was indicated with great distinctness, although no 
 reference was made to them in the text. These 
 latter organs were, eight years later, shown by 
 J. Thiele 1 to represent a pair of abdominal 
 sensory papillae having a highly characteristic 
 structure and innervated by a fine nerve pro- 
 ceeding backwards on each side from the visceral 
 ganglion. 
 
 When examined in toto in a fresh preparation 
 the abdominal sense-organs are chiefly char- 
 acterised by the possession of a dense coating 
 of long, motionless, stiff cilia or sense-hairs. In 
 transverse section these cilia are seen to be 
 carried by a very high epithelium containing 
 numerous basal nuclei at different levels, an 
 intermediate layer of nuclei at one level, and a 
 clear peripheral zone. Thiele described the 
 organs in a number of bivalves, including a 
 species of scallop, Pectcn varius, where he found 
 a sensory ridge on the right side only, there 
 being no corresponding ridge on the left side 
 
 1 Zeitschr. wiss. Zool., xlviii., 1889, pp. 47-59. 
 
160 CONVERGENCE IN MINUTE STRUCTURE 
 
 in this species. Similar relations have been 
 described by Dakin 1 in another species of scallop, 
 Pectcn maximus, and the unpaired condition of 
 the organ is obviously a character of the genus 
 Pec ten. 
 
 In the Oriental disc-shaped bivalve, P^tna 
 placenta, belonging to the family Anomiidae, 
 only one sensory abdominal ridge occurs. It is 
 defined by the presence of brown pigment, and 
 is placed upon the edge of an adrectal ligament 
 or cutaneous fold which unites the rectal complex 
 with the right mantle lobe at about the level of 
 the posterior insertion of the suspensory ligament 
 of the right ctenidium. The pallial circulation 
 of Placuna is characterised by the presence of 
 a pair of ascending pallial vessels, which arise 
 from the circumpallial arteries at the postero- 
 ventral border of each mantle lobe and pass 
 obliquely upwards towards the posterior end of 
 the suprabranchial region. On approaching the 
 latter their walls are usually inflated, after the 
 valves have been separated, to form an elongated 
 oblong or elliptical body which is non-contractile. 
 I do not know the precise significance of this 
 congestion of the walls of the ascending pallial 
 vessels of Placuna, but it concerns the vascular 
 system and has nothing to do with the adrectal 
 sensory organ. 
 
 The abdominal sense-organ of Placuna con- 
 
 1 Memoir on Pecten. Liverpool, 1909. 
 
SENSE-ORGAN OF PLACUNA 161 
 
 forms in its general histological structure to those 
 previously described in other Lamellibranchs, and 
 I have had the opportunity of comparing it in 
 section with the paired, whitish, adrectal, semi- 
 lunar organs of Area rhombea, a species which 
 occurs in the same localities as Placuna in the 
 back-waters of the eastern province of Ceylon, 
 near Trincomalee. Transverse sections of the 
 sense-organ in Placuna, stained with alum car- 
 mine, show externally the dense layer of perfectly 
 preserved, stiff, sensory cilia, exceeding in length 
 the height of the sensory epithelium. At their 
 bases they perforate the stout, cuticular mem- 
 brane, which presents a double contour. Below 
 the cuticle there follows an outer prismatic layer, 
 then a middle layer of evenly placed elliptical, 
 nuclear bodies representing the layer of spindles ; 
 thirdly, an inner fibrillar layer, and finally a basal 
 plexus of nuclear bodies adjoining the basement 
 membrane which separates the epithelium from 
 the gelatinous conjunctive tissue upon which it 
 rests. 
 
 Thiele emphasised the extraordinary resem- 
 blance between the epithelium of the Molluscan 
 abdominal organs and the Annelid lateral organs 
 as described by Eisig in the Capitellidae ; and 
 satisfied himself by actual comparison of prepara- 
 tions that in fact the similarity is very close. 
 
 Recently another pair of sense-organs, histo- 
 logically resembling the abdominal sense-organs 
 
162 CONVERGENCE IN MINUTE STRUCTURES 
 
 though quite different topographically, has been 
 described by M. Stenta 1 in a Protobranchiate 
 bivalve mollusc. These are the marginal pallial 
 sense-organs of Leda commutata (Nuculidae) which 
 occur right and left at the anterior junction of 
 the right and left mantle lobes, lying in a crypt 
 on each side between the inner and middle folds 
 of the pallial margin. If any one ventured to 
 argue that the posterior adrectal or abdominal 
 organs of Lamellibranchiata might be related by 
 way of homology to the segmental abdominal 
 organs of Capitellidae, the same argument could 
 not be applied to the anterior pallial organs of 
 Leda, although all these structures apparently 
 belong to one physiological category, namely, 
 rheostatic organs. We may confidently conclude 
 that they are related to each other only by way 
 of sensory convergence. In this connection we 
 may take note of the well-known case of retinal 
 convergence between the pallial eyes of the scallop 
 and the cerebral eyes of vertebrates. 
 
 Eisig 2 treated the lateral sense-organs of Capi 
 tellidae under two categories thoracic organs 
 appearing in surface view as open pores ; and 
 abdominal organs, appearing as retractile knobs or 
 elevations, the distal extremity of which is beset 
 with stiff sense - hairs radiating in all directions. 
 
 1 Mario Stenta, "Uber ein neues Mantelrandorgan bei Leda 
 commutata? ZooL Anz., xxxv., 1909, pp. 154-157. 
 
 2 Hugo Eisig, " Die Capitelliden." " Fauna und Flora des Golfes 
 von Neapel." Berlin, 1887. 
 
SENSE-ORGANS OF CAPITELLID^E 163 
 
 These sense - organs were closely compared by 
 Eisig with the lateral line sense-organs of aquatic 
 vertebrates (fishes and batrachians), the function 
 of which is partly equilibrating and partly the 
 perception of currents and wave - movements. 
 The most remarkable histological resemblance is 
 manifested between the lateral sense-organs of 
 the Capitellidae and the lateral line sense-organs 
 of Vertebrates. In both cases the essential organs 
 consist of small, solid, roundish, epidermal buds, 
 from which fine stiff sense - hairs project freely 
 into the surrounding medium ; and the resem- 
 blance is further enhanced by their segmental 
 arrangement. The correspondence could hardly 
 be greater, the convergence could hardly be 
 closer, the homology could not be more remote 
 than infinity. 
 
 Within the limits of the order of Polychseta, 
 Eisig has established the true homology between 
 the lateral sense-organs of Capitellidae and the 
 dorsal cirri of neural parapodia such as occur in 
 the family Glyceridae. Through this family it is 
 made clear that the lateral organs are homologous 
 with the dorsal cirri of Polychseta as a whole. 
 
 Most morphological arguments work both ways, 
 and it is not surprising that after having estab- 
 lished the homology between the endostyle of 
 Amphioxus and the Ascidians and the thyroid 
 gland of Ammocoetes, the larval form of the 
 
164 CONVERGENCE IN MINUTE STRUCTURES 
 
 lamprey, Dohrn, committed to the Annelid Theory 
 of vertebrate descent, should have concluded 
 that the endostyle was derived by descent with 
 simplification from the more elaborate organ in 
 Ammocoetes, in which opinion he has been sup- 
 ported by Gaskell. Regressive convergence by 
 reduction and loss of parts is indeed a very 
 common phenomenon, expressive of the level- 
 ling effects of degeneration, e.g., the loss of 
 limbs in snakes, some lizards, some batrachians, 
 and some fishes ; compare also the effects of 
 sedentary and parasitic habits. In less obvious 
 cases this principle has to be applied with the 
 utmost circumspection. 
 
 For Gaskell's views on the homology which 
 he supposes to exist between the thyroid gland 
 of Ammocoetes and the uterus of the scorpion, 
 the reader should consult the work quoted above. 
 For my present purpose the following passage 
 will suffice (op. cit., p. 205) : " The resemblance 
 between the structure of the thyroid of Ammo- 
 coetes and the uterus of the scorpion is most 
 striking, except in two respects, viz., the nature 
 of the lining of the non-glandular part of the 
 cavity in the one case ciliated, in the other 
 chitinous and the place of exit of the cavity, 
 the thyroid of Ammocoetes opening into the 
 respiratory chamber, while the uterus of Scorpio 
 opens direct to the exterior." It may be added 
 by way of explanation that the supposed homology 
 
MESOSOMA 165 
 
 is founded, in the first instance, upon the mutual 
 topographical proximity of the respiratory appar- 
 atus, and of the terminal portion of the genital 
 apparatus in the scorpion group and in Limulus 
 the King Crab, and the relations of these organs 
 to the segments of the mesosoma. 
 
 It is not easy to deal with this comparison 
 in a satisfactory manner. It seems to be some- 
 what gratuitous and to proceed from a fallacious 
 assumption of community of regional differentia- 
 tion. We have seen in several instances that 
 identity of structure and function may go for 
 nothing in determining homology, and that the 
 same identity has no relation to topography. If 
 one chooses to compare the respiratory region 
 of the scorpion or of Limulus with that of 
 Ammocoetes, and to dub it mesosoma in both 
 cases, all sorts of curious sequences will be 
 encountered. Without a preconceived bias, such 
 as Dr Gaskell does not conceal, namely, in 
 respect of his ideas upon dominance and brain- 
 power, one would not consider the two regions 
 comparable morphologically. It is a very good 
 thing to have a guiding idea in morphology and 
 to follow it out, but at the best it can only lead 
 to a subjective conclusion. There is no necessity 
 to confound such a conclusion with the truth, 
 and this is all we can ask, since the reconcilia- 
 tion of the truth with one's individual views 
 
 is not a matter which can be settled within a 
 
 L 2 
 
166 CONVERGENCE IN MINUTE STRUCTURE 
 
 generation. At the same time it would seem 
 proper to insist here that in the face of an 
 assertive logical system which can only be com- 
 bated at the expense of infinite weariness and 
 vexation, the best and only course is to hold 
 fast to what one adopts intuitively as first 
 principles ; and, with regard to the special case 
 before us, to regard the gill-cleft as an autonomous 
 morphon and the Limuloid gill-book, with its 
 derivative the lung-book of Arachnida, as another 
 autonomous morphon, having nothing in common 
 except their function. 
 
 The histological resemblance between the 
 glandular part of the thyroid of Ammocoetes 
 and the glandular part of the uterus of the 
 scorpion may be regarded as a case of glandular 
 convergence, and as such it possesses a peculiar 
 interest of its own. 
 
 Now if we reject the Limuloid Theory of 
 Vertebrate descent, why should we accept as 
 normal the theory of the Arachnoid affinities 
 of Limulus ? The answer in brief is that the 
 normal morphology of Arachnoids and of Verte- 
 brates respects phyletic boundaries ; and that 
 the convergent morphology of these groups tran- 
 scends those limits. Both are good so long as 
 they are kept apart. 
 
 It would, of course, be possible to multiply 
 examples of histological parallelism to an almost 
 unlimited extent by comparing distantly related 
 
CYTOLOGY 167 
 
 forms and making extracts from published 
 records. All animals are related together by 
 the continuity of the germ-plasm ; the remoter 
 the relationship, the closer the convergence may 
 be, and vice versa. An instance of nuclear con- 
 vergence has been noted recently by Minchin. 1 
 In the collar-cells of some calcareous sponges 
 (Clathrinidce) he found that the nucleus occupies 
 a position at the base of the cell, and the 
 flagellum arises independently from a granule 
 or blepharoplast situated at the surface of the 
 cell in the centre of the area enclosed by the 
 collar. In the Leucosoleniidae the nucleus 
 occupies an apical position and the flagellum 
 appears as a direct continuation of the pointed 
 end of the nucleus. 
 
 Minchin quotes a parallelism to these alter- 
 native positions of the nucleus in the case of two 
 species of Mastigina described by Goldschmidt 
 (1907). Such a character, adds Minchin, in the 
 case of sponges, can have but little importance 
 in the struggle for existence, and yet in his 
 opinion it indicates the deepest phylogenetic 
 divergence in the pedigree of the calcareous 
 sponges. 
 
 Cytological convergence, as between Metazoa 
 and Protozoa, yields many points of instructive 
 
 1 E. A. Minchin, "The Relation of the Flagellum to the Nucleus 
 in the Collar-Cells of Calcareous Sponges." ZooL Anz., xxxv., 1909, 
 p. 227. 
 
1 68 CONVERGENCE IN MINUTE STRUCTURE 
 
 comparison. The phenomena of intracellular 
 digestion, phagocytosis, and wandering cells come 
 under this category. An elaborate case is that 
 of the pulsating, multinucleated chromatophores 
 of eight-armed Cephalopoda. In Octopus, under 
 certain conditions of excitement, the play of 
 colour may be observed taking place with 
 marvellous regularity and rapidity. Each chrom- 
 atophore, as Chun 1 has demonstrated, is a single 
 disc-shaped cell with contractile processes radi- 
 ating from the periphery ; and these contractile 
 processes originate from the cell-body in the 
 same manner as the pseudopodia of a rhizopod 
 Protozoan. 
 
 The preceding examples bear witness to the 
 truth of the conclusion that convergence depends 
 equally upon the unity of plan of composition 
 of the body of animals, as indicated in the 
 incidence of planes of symmetry, and upon the 
 continuity of the system of functions from 
 Protozoa to Metazqa. 
 
 Now, if we stop for a moment to enquire what 
 is the bearing of all this, it may be said that I 
 have failed in my purpose unless it has been 
 made abundantly clear that the influence of 
 convergence in evolution has been widespread, 
 
 1 Carl Chun, " Uber dieNatur und die Entwicklung der Chrom- 
 atophoren bei den Cephalopoden." Verh. d. Deutschen Zool. Ges., 
 1902, pp. 162-182. 
 
MORPHONS AND TROPISMS 169 
 
 deep-seated, and intimate, more so than is 
 generally recognised. What may appear to be 
 a brilliant discovery of morphological affinity 
 may in reality be an equally brilliant demonstra- 
 tion of the no less important and interesting 
 phenomenon of morphological convergence ; the 
 closer the identity, as between forms belonging 
 to different phyla, the greater likelihood, or, as 
 I should prefer to say, the greater certainty that 
 it is due to convergence. 
 
 What is known as homoplasy in morphology 
 might be called homotaxis in bionomics. In 
 contrast with the phenomenon of change of 
 function we have that of substitution of organs, 
 as Kleinenberg expressed it, or change of 
 morphon in Spengel's phraseology. Similarly, 
 in contrast with the Tropism Theory we must 
 have the Morphon Theory, and we must dis- 
 tinguish between primary or general morphons 
 which are, with due reserves, the intrinsic 
 property of all animals, and secondary or 
 phyletic morphons which are the special char- 
 acteristics of distinct groups. The relation of 
 morphons to tropisms lies at the basis of all 
 orthogenetic morphology and all convergent 
 morphology. Under the one or the other head- 
 ing all pertinent facts can be ranged, and none 
 need be left out of the reckoning. 
 
 It will be observed that no attempt has been 
 made in the foregoing pages to formulate any laws 
 
170 CONVERGENCE IN MINUTE STRUCTURE 
 
 of orthogenesis and convergence. A greater 
 assemblage of facts than has been marshalled 
 here would be necessary, nothing short of a 
 new cyclopaedia of anatomy and physiology, and 
 probably the time is not yet ripe for that. Up 
 to the present no other work with which I 
 am acquainted has dealt with convergence as a 
 general and positive phenomenon of equal import- 
 ance with orthogenesis or normal morphology, 
 although isolated cases are referred to in most 
 text-books of zoology and are exhibited in most 
 natural history museums. 
 
 The fact that no laws of convergence are or 
 even can be laid down now is one which is 
 fraught with the greatest hope for the future 
 of morphology ; and the breaking down of the 
 former landmarks of homology, such as histo- 
 logical structure and metameric repetition, except 
 within narrow limits, offers a great opportunity 
 for emancipation from the trammels of specu- 
 lation. If we were to tabulate laws they would 
 not be natural, but merely dogmatic, at the mercy 
 of the first unbeliever. Hardly one universal 
 criterion of strict homology can be mentioned 
 which would pass muster in a critical examina- 
 tion. Then away with laws and away with 
 criteria until they cease to obscure the facts as 
 they are. 
 
 Hypotheses are one of the chief means of 
 progress in morphology. Without them the 
 
CONCLUSION 171 
 
 advances which have been made during the past 
 hundred years would not have been so consider- 
 able. If the hypothesis is constructed after the 
 work is done or before it is completed, the work 
 accomplished remains after the need for the 
 hypothesis, which is either the best clue that 
 one can give for the time being or the best 
 guide that one can follow, has passed away. 
 In morphology everything is important except 
 the hypothesis, although practically nothing could 
 be done without it, since it is often the only 
 means available for digesting an accumulation 
 of facts. It is something intangibly necessary, 
 often quite wrong, always hopelessly incom- 
 plete, but ever ready to give way by substitution 
 to another invisible vehicle. The progress of 
 morphology depends upon the substitution of 
 ideas rather than upon the promulgation of laws. 
 The tree of life is polyphyletic, and the branches 
 do not anastomose after their zigzag course has 
 been set. 
 
INDEX 
 
 ABDOMINAL sense-organs, 159 
 
 Accessory branchial organs, 141 
 
 Acorn barnacles, 39 
 
 Air-bladder, 144, 151 
 
 Air-breathers, 19, 141 
 
 Alciopidae, 44 
 
 Alcock, A., 106, 117 
 
 Alytes, 97, 134 
 
 Amia, 135 
 
 Ammocoetes, 163 
 
 Amnion, 16 
 
 Amphibia, 19, 142 
 
 Amphinome, 40 
 
 Amphioxus, 21, 45, 46, 72, 85, 139, 
 
 151, 153, 163 
 Ampullaria.) 18 
 Anabas, 141 
 
 Analogy (see Parallelism) 
 Annelid Theory, 164 
 Annelida, 43, 77, 118, 129, 139, 
 
 153, 157, 161 
 
 Anterior neuropore, 83, 85 
 Ants, 63, 114, 119 
 Appcndicularia, 38 
 Appendiculata, 70, 78, 82, 125 
 Aquatic animals, 17 
 Arachnoidea, 148, 166 
 Arboreal mammals, 28, 57 
 Area, 158, 161 
 Archenteron, 77 
 Arius, 131, 137, 146 
 Arixenia, 115 
 Arthropoda, 148 
 Ascidians (see Tunicata) 
 Asphyxiation, 18 
 Aspredo, 133 
 Assheton, R., 144 
 Atheston, L., 119 
 Auditory organs, 82 
 Autotomy, 63 
 
 BALANOGLOSSUS (see Entero- 
 
 pneusta) 
 Balanus, 39 
 
 Balfour, F. M., 72 
 Barnacles (see Cirripedes) 
 Batesian mimicry, 55 
 Bateson, W., 72, 97, 122 
 Bather, F. A., 36 
 Batrachia, 19, 92, 96 
 Bats, 23, 89, 105 
 Beddard, F. E., 46 
 Beebe, C. W., 60 
 Belonc, 155 
 Belostomatidge, 134 
 Biocoenosis, 114 
 Birds, 89, 123, 151 
 Birgus, 1 8 
 Blenny, 140 
 Boake, B., 131, 141 
 Boeke, J., 136 
 Bois-Reymond, E. du, 51 
 Boulenger, C. L., 132 
 
 , G. A., 95, 102 
 
 Boveri, Th., 153 
 Brain-power, 103, 157, 165 
 Branchiata, 148 
 Branchiotrema, 150 
 Brood-nursing (see Incubation) 
 Bruce, W. S., 100 
 Buckland, F. T., 22, 30, 52, 92 
 Budge tt, J. S., 136 
 Bufonidse, 102 
 Bulbuls, 22 
 Burne, R. H., 145 
 Burrowing mammals, 29, 56, 80 
 Butterflies, 27, 59 
 
 CALLICHROMA, 55 
 
 Calotes, 22, 50 
 
 Camper, P., 152 
 
 Capitellidse, 158, 162 
 
 Cartilaginous fishes, 94 
 
 Cat's eye, 92 
 
 Cave, W. A., 59 
 
 Cell-theory, 4 
 
 Cephalisation, 70, 113 
 
 Cephalopoda, 33, 70, 95, 137, 168 
 
 173 
 
INDEX 
 
 Cerebration, 114 
 
 Ceylon, 45, 55, 59, 6l, 64, 88, 
 
 109, 131, 134, HO, 145, 161 
 Chameleon, 52 
 Chank shell, 98 
 Chatoessus, no 
 Chironomus, 118 
 Chlorophyll, 26 
 Chordata, 72, 150 
 Chromatism, 27 
 Chromatophores, 1 68 
 Chun, C, 168 
 Cimex, 61 
 Cirripedes, 39, 116 
 Clarias, 141 
 
 Clupeidae, 95, 109, 112, 145 
 Cobra, 30 
 Ccelenterata, 70, 78 
 Coelom, 77, 122 
 Coelomata, 77 
 Collar-cells, 167 
 Collin, A., 129 
 Cooke, A. H.,98 
 Crabs (see Crustacea) 
 Crocidura^ 79 
 Crows, 25 
 Crustacea, 52, 95 
 Cryptotaxis, 33 
 Ctenidium, 147 
 Cutaneous incubation, 133 
 Cuttlefishes (see Cephalopoda) 
 Cuvier, Baron, 4, 13, 80 
 CylindrophiS) 49 
 
 DACTYLOPTERUS, 88 
 
 Dahl, F., 7, 114 
 
 Dakin, W. J., 44, 160 
 
 Darwin, C., I, 5, 67, 112, 116, 
 
 132 
 
 Day, F., 131 
 Dean, B., 135 
 Death-feint, 66 
 Dendy, A., 126 
 
 Dewitz, , 38 
 
 Diopatra, 43 
 Dipnoi, 19, 142 
 Direct development, 122 
 Distant, W. L., 134 
 Dohrn, A., 83, 164 
 Dominance, 157, 165 
 Dorsalcirri, 163 
 Draco, 50 
 Dryophis, 93 
 Duerden, J. E., 20 
 Dussumieria, 95, 109 
 
 ECHINODERMATA, 36, 71, 122 
 
 Ectoparasites, 115 
 
 Eisig, H., 43, 51, 162 
 
 Electric organs, 51, 91, 112 
 
 Elephant, 27 
 
 Endoskeleton, 14 
 
 Endostyle, 163 
 
 Enteropneusta, 32, 42, 72, 122, 123, 
 
 128, 140, 151 
 Entoprocta, 155 
 Espada, J. de la, 132 
 Excretory organs, 153-155 
 Exoccetus, 88 
 Exuviation, 52 
 Eyelids, 95, 112 
 
 FLAME-CELLS, 155 
 
 Flatfishes, 97 
 
 Fleas, 7, 61, 115 
 
 Flower, W., and Lydekker, R., 79, 
 
 91 
 Flying fishes, 88 
 
 foxes, 24 
 
 leaps, 50 
 
 mammals, 57, 89 
 
 Food-yolk (see Lecithality) 
 Fox's eye, 92 
 Frog, 42, 92 
 Froggart, W. W., 120 
 FusuSy 97 
 
 GADOW, H., 102 
 
 Galeopithecus, 57, 91 
 
 Gaskell, W. H., 73, 75, 81, 82, 85, 
 
 103, 156, 164 
 
 Gastral convergence, 107-112 
 Gastropoda, 41, 149 
 Geckoes, 22, 92 
 Gill-book, 1 66 
 
 Gill-clefts, 139, 144, 149-151, 1 66 
 Gills, cutaneous, 139 
 Gizzard, 107, no 
 Glandular convergence, 106, 166 
 Globe-fishes, 144 
 Glyceridse, 163 
 Goat's eye, 93 
 Goby, 140 
 
 Goldschmidt, R,, 167 
 Goodrich, E. S., 154 
 Green, E. E., 59 
 Grey mullet, 109 
 Groom, T. T., 41 
 Gruvel, A., 41 
 GUnther, A., 64, 133, 134 
 Gut, 75, 155 
 Gymnarchus, 137, 144 
 
 H^MADIPSA, 45 
 
INDEX 
 
 Hsematotropism, 48 
 
 Haemoglobin, 118 
 
 Harvey, W., 151 
 
 Hawk-moths, 23 
 
 Hedgehog, 30, 105 
 
 Heliotropism, 20, 41 
 
 Hemichordata, 72 
 
 HcmimeruS) 115 
 
 Herpestes, 94 
 
 Herring, 146 
 
 Heteropoda, 41 
 
 Heterotypic convergence, 90 
 
 Hickson, S. J., 140 
 
 Hippocampus, 52, 68 
 
 Hipponoe, 40 
 
 Hirudo, 45 
 
 Histology, 153 
 
 Hodgson, T. V., 99 
 
 Homing, 25 
 
 Homology, 53, 75, *5i *S 2 > I 55y 
 
 163 
 
 Homoplasy, u, 12, 53, 57, 90, 169 
 Homotaxis, 169 
 Howes, G. B., 91, 132 
 Hubrecht, A. A. W., 16, 94, 103- 
 
 105, 125, 144 
 Hunter, J., 106, in, 151 
 Huxley, T. H., 13, 141 
 Hydroid polyps, 37, 117 
 Hydromedusae, 37 
 Hydrostatic organs, 144 
 Hydrus platurus y 116 
 
 ICHTHYOPHIS, 31, 143 
 Incubation, 131-134, 142 
 Infundibulum, 82, 83, 85 
 Ink-sac, 33 
 Insectivora, 79 
 
 Insects, 58, 115, 118, 120, 139, 148 
 Invertebrates, 13 
 Izuka, A., 130 
 
 JAPANESE palolo, 130 
 Jennings, H. S., 20 
 Jordan, D. S., 51 
 
 , K., 115 
 
 Jumping fishes, 140 
 Jungle life, 28 
 
 KALLIMA, s 8 
 
 Keith, A., 87 
 
 Kerr, J. G., 136, 143 
 
 Kinetoblast, 128 
 
 King Crab (see Limulus) 
 
 Kleinenberg, N., 169 
 
 Korschelt und Heider, 40 
 
 Kramer, , 129 
 
 LABYRINTHINE organs, 141 
 Lacaze-Duthiers, H. de, 84 
 Lamarck, 13 
 Lamellibranchiata, 158 
 Land crabs, 17 
 
 leeches, 18, 31, 45 
 
 mollusca, 17, 149 
 
 nemertines, 17 
 
 planarians, 17, 28 
 
 Lang, A., 37 
 
 Lankester, Sir E. R., 11,53, !47 
 
 149, 158 
 
 Larval forms, 42, 122, 128 
 Lateral sense-organs, 163 
 Leaf-fish, 66 
 Leaf-mimicry, 58-66 
 Leandcr, 40 
 Lecithality, 124 
 Lcda, 162 
 Leigh, H. S., 62 
 Lcpas, 39, 117 
 Lepidastkenia, 43 
 Lepidosircn, 143 
 Lepidosteus, 96 
 Limnaeidae, 142 
 Limnodrilus (see Tubificidae) 
 Lo Bianco, S., 139 
 Loeb,J., 3 8 
 Longitudinal stripes, 68 
 Lophobranchii, 68, 133 
 Lungs, 139, 145 
 Lycodon, 49 
 Lydekker, R. (see Flower and 
 
 Lydekker) 
 
 MACBRIDE, E. W., 72 
 
 Mammals, 12, 16, 26, 29, 57, 68, 
 
 80, 89, 91-93, 103-106, 125 
 Mammary glands, 105 
 Afants, 104 
 Marsupials, 56, 79, 80 
 Mayer, A. G., 130 
 Medicinal leech (see Hirudo) 
 Meerkat, 94 
 Megalops, 108, 145 
 Mesosoma, 165 
 Mesozoic reptiles, 91 
 Metamerism, 77, 78 
 Metazoa, 15 
 Metschnikoff, E., 123 
 Miall, L. C, 118 
 Milk, 107 
 Mimicry, 52 
 Minchin, E. A., 167 
 Minous inermis, 117 
 Mole-like forms, 56 
 Molgula, 84 
 
i 7 6 
 
 INDEX 
 
 Mollusca, 17, 41, 77, 97, 139, 147, 
 
 149, 158 
 
 Monkeys, 28, 50 
 Monotrema, 52, 68, 125 
 Moore, J. P. , 40 
 Morphon, 144, 169 
 Morris, C, 144 
 Moseley, H. N., 125, 140 
 Mugilidse, 112 
 Muller, J., 123 
 Mullet, 108 
 Mungoose, 94 
 MyobatraehuS) 102 
 Myoepithelial cells, 155, 156 
 Myriopoda, 148 
 Myrmecoidism, 121 
 Myrtnedonta, 121 
 
 NAUPLIUS, 39-41 
 
 Nautilus, 33, 137 
 
 Nematocysts, 155 
 
 Nematodes, 70, 155 
 
 Nemertines, 42, 77 
 
 Nephridia, 154 
 
 NcphthyS) 44 
 
 Nervous system, 75 & 2 > 86, I][ 3 
 
 Newton, A., 23 
 
 Newts, 52 
 
 Nictitating membrane, 94 
 
 Notochord, 14, 72, 73 
 
 Nucleus, 167 
 
 Nuculidse, 162 
 
 Nutrition, 73, 103, 126 
 
 OESOPHAGUS, 83, 85, 156 
 Olfactory pit, 85 
 Onychophora, 125 
 Opercular membrane, 141, 147 
 Ophelia, 139 
 Ophiocephalus, 135, 141 
 Ophiodromus, 43 
 Opisthobranchiata, 147, 149 
 Opisthocoelous vertebrae, 97 
 Opposable extremities, 95 
 Osborn, H. F., 12, 68 
 Osphradium, 147, 158 
 Osteophora, 94 
 Oviparity, 124, 125 
 Owen, Sir R., 4, 93, 96, 106, 108, 
 I5i 
 
 PALLIAL sense-organs, 162 
 
 Palolo, 43, 129 
 
 Pantopoda, 99 
 
 Parachute flight, 49, 90, 91 
 
 Parallelism, 57, 70, 79, 128, 149 
 
 Parasitism, 31, 74, 76 
 
 Patella, 147 
 Pecten, 44, 159, 160 
 Pectoral fins, 88 
 Pelmatozoa, 37, 71 
 Pentanymphon, 100 
 Periophthalmus, 140 
 Peripatus, 17, 125 
 Phanerozoa, 23 
 Phasmidse, 63 
 
 Phosphorescence, 21, 34, 91, 1 12 
 Phototaxis, 20 
 Phyllidia, 147 
 Phy Ilium, 61 
 Phyllodoce, 154 
 Phyllopteryx, 64 
 Physiognomical convergence, 57 
 Pilidium, 42 
 Pipa, 96, 133 
 
 Pipe-fishes (see Lophobranchii) 
 Placentation, 103-105, 126 
 Placuna, 160 
 Planozoic, 39 
 Platax, 64, 66 
 Platyhelminthes, 70, 77 
 Platypus (see Monotrema) 
 Pleotropism, 41, 43, 45, 122 
 Plotosus, 147 
 Pocock, R. I., 63, 149 
 Polychaeta (see Annelida) 
 Polypedates, 134 
 Polypterus, 145 
 Poulton, E. B., 54 
 Prehensile tails, 52, 68 
 Protective coloration, 26 
 Protochordata, 72, 83 
 Protozoa, 15, 167 
 Psamtnolyce, 43 
 Pteromys, 57, 91 
 Pteroplatcea, 106 
 Pteropoda, 41 
 Pulmonate Arachnida, 149 
 
 Mollusca, 149 
 
 Punnett, R. C, 150 
 Pupil of the eye, 93, 102 
 Purcell, W. F., 148, 149 
 Pygmies, 87 
 Python, 142 
 
 RACIAL convergence, 87 
 Raptorial appendages, 95 
 Recapitulation, 42 
 Regeneration, 46, 52, 63 
 Regression, 125, 164 
 Reptiles, 21, $o, 124 
 Respiration, 139 
 Respiratory pigments, 26 
 Retina, 162 
 
INDEX 
 
 177 
 
 Reversed shells, 97 
 
 vertebrae, 97 
 
 Rhacophorus ; 92, 95, 134 
 Rheostatic organs, 162 
 Rhinoderma,) 132 
 Rhinophrynus , 102 
 Rhodosoma, 84 
 Rodentia, 56, 79 
 Rotifers, 155 
 
 SACCOBRANCHUS, 141, 145 
 
 Salarias, 140 
 
 Sarasin, P. and F., 142 
 
 Sarcodaces, 136 
 
 Sauropsida, 72, 137 
 
 Schimkewitsch, W., 99 
 
 Scombresocidae, 88, 155 
 
 Scorpion, 29, 164 
 
 Scylltza, 40 
 
 Seasonal flights, 27 
 
 Sedgwick, A., 42, 87, 126 
 
 Semper, K., 18, 37, 52, 88 
 
 Sensory convergence, 158-163 
 
 Sessile animals, 36 
 
 Seton, E. T., 92 
 
 Shelford, R. , 49 
 
 Siluridae, 131, 141 
 
 Sinistral shells, 98 
 
 Situs tnversus, 97 
 
 Snakes, 49, 67, 93, 116, 164 
 
 Social insects, 119 
 
 Solenocytes, 154-157 
 
 Somites, 77 
 
 Spengel, J. W., 144, 158, 169 
 
 Sphcerodema, 134 
 
 Spiders, 148 
 
 Sponges, 36, 167 
 
 Spurious convergence, 81 
 
 Statozoa, 37 
 
 Stenta, M., 162 
 
 Stereotropism, 38, 122, 143 
 
 Sting-rays, 1 06 
 
 Strobilation, 77, 78 
 
 Stylactis, 117 
 
 Substitution, 14, 143, 169 
 
 Surinam toad (see Pipa) 
 
 Swarming habits, 43, 129 
 
 Synaposematic, 55 
 
 TADPOLE, 42 
 
 Tagetis, 60 
 Teats, 125 
 Teeth, 12 
 Teleostei, 88, 112 
 Tennent, Sir E., 48 
 Termites, 63, 120 
 Testacella, 31 
 Thiele, J., 159, 161 
 Thompson, D'Arcy W., 99 
 Thyroid gland, 163, 164 
 Tilapia nilotica, 132 
 Toads, 22, 92, 101 
 Topography, 165 
 Tornaria, 42, 123 
 Tracheae, 148 
 Tracheata, 148 
 Triploblastica, 77 
 Trophoblast, 16, 103, 128 
 Tropism, 20, 22, 169 
 Tubificidae, 118 
 Tunicata, 72, 84, 155, 163 
 Types, 9 
 
 UEXKULL, J. von, 20 
 Ungulates, 104 
 
 VARANUS, 50 
 
 Variation, 58, 60, 62 
 
 Veliger, 41 
 
 Vertebrae, 96 
 
 Vertebrata, 13, 81, 85, 158 
 
 Verworn, M., 21 
 
 Viability, 146 
 
 Visceral clefts (see Gill-clefts) 
 
 Viviparity, 124 
 
 Viviparous rays, 106 
 
 WALLACE, A. R., 53, 58 
 
 Warning coloration, 54 
 
 Wasmann, E., 120, 121 
 
 Water-breathers, 19 
 
 Weiss, F. E., 153 
 
 Whitman, C. O., 46 
 
 Whitmee, S. J., 129 
 
 Willey, A., 16, 64, 124, 125, 135, 
 
 ISO 
 
 Windowpane-oyster, 158 
 
 Wings, 89 
 
 ZOOPHYTES, 35 
 
 M 
 
Printed at 
 
 The Edinburgh Press 
 9 and n Young Street 
 
RETURN EARTH SCIENCES LIBRARY 
 
 TO^ 230 Earth Sciences Bldg. 642-2997 
 
 LOAN PERIOD 1 
 
 ALL BOOKS MAY BE RECALLED AFTER 7 DAYS 
 
 Books needed for class reserve are subject to immediate recall 
 
 DUE AS STAMPED BELOW 
 
 UNIVERSITY OF CALIFORNIA, BERKELEY 
 FORM NO. DD8, 6m, 477 BERKELEY, CA 94720 
 
 'PS