GENETIC STUDIES ON A CAVY SPECIES cross BY J. A, DETLEFSEN Assistant Professor of Genetics, U' niversity of Illinois WITH A PREFATORY NOTE BY W. E. CASTLE WASHINGTON, D. C. PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1914 Ti Carnecig INSTITUTION OF WASHINGTON, PuBiicaTion No. 205 Paver No. 23 or Tom StaTION ror EXPERIMENTAL EVOLUTION At Cotp Sprinc Harsor, New York From THe Lasporatory or GENETICS, oF THE BussEy InstrTuTION, Fortst Hitt, MASSACHUSETTS Copies of this Book were first issued DEC 311914 PRESS OF GIBSON BROTHERS, INC. WASHINGTON, D. C. PREFATORY NOTE BY W. E. CASTLE. In July 1903 I received from Mr. Adolph Hempel, of Campinas, Brazil, three wild cavies, a male and two females, of a species supposed at the time to be Cavia aperea, but now referred to Cavia rufescens. The male and one of the females bred in captivity and produced a considerable number of descendants, certain of which (together with the original male) were employed in crosses with ordinary guinea-pigs. The hybrids thus obtained proved completely sterile in the male sex, but the females were entirely fertile. Further propagation of the hybrid race was thus restricted to crossing the female hybrids with males of one of the parent species. In December 1909 I turned over to my assistant, J. A. Detlefsen, for further study, the stock of hybrid animals, together with the pedi- gree records and notes of such observations as I had been able to make upon the hybrid race. The present paper will indicate how successful he has been in propagating the hybrid race and what conclusions may be drawn concerning the inheritance of various characters in these hybrids. The long series of experiments upon which a partial report is here made was rendered possible by a grant from the Carnegie Institution of Washington and by the provision of special facilities on the part of Harvard University. Grateful acknowledgment is made of my obliga- tion to both institutions. 3 1. 2. mo o on x © 10. 11. 12. 13. CONTENTS. GENERAL INTRODUCTION. 4 Systematic position of the parent races Materials and methods............... The wild race................. One-half wild hybrids.......... One-quarter wild hybrids....... Three-quarter wild hybrids...... Fertile males in matings........ Accumulation of data One-eighth wild hybrids, and later generations Part I. Cotor anp Coat CHARACTERS. Introductory discussion........,.... The agouti character in the wild race and in hybrids Homozygous agoutis in crosses. . Heterozygous agoutis mated inter The wild agouti and tame agouti contrasted Modification of the wild agouti Modified wild agouti in crosses “Presence and absence’’ hypothesis applied Non-agoutis mated inter se...... . Black and brown.................55 Homozygous blacks in crosses... Browns mated inter se.......... Extension and restriction............. Homozygous condition of extension in crosses Heterozygous agoutis mated to non-agoutis SOs fe Git AS gana Gactnaeabenes ani ye QI nate Gia aperera Heterozygous blacks mated with brown Heterozygous blacks mated inter se Heterozygous condition of extension crossed with restriction..........-....-- Heterozygotes for extension mated inter 8€....... 0.2 c eee eens Reds mated inter se............. Color and albinism................. Homozygous condition of the color factor in crosses........... 00. ce eee eens Heterozygous colored animals in crosses with albinos Heterozygous colored animals mated inter se........0.. 00. cece eee tenet AlbINGS MAE ALE 86 is, sroiniekiniey a RES wa ie 4 AGW ARALE NS pH Rime du Ele PeslMwa aes Roughness and smoothness........... Homozygous rough animals in crosses... 1... 0... cee eee eee teens Heterozygous rough animals crossed with smooth animals.................... Smooth animals mated inter se... . Other color and coat characters...... Uniformity and spotting. 0.6 i.e. ee cee ened tenes aa G tawew Sead ones Intensity and dilution.......... Long hair and short hair... . 0.0.0... cette eee eee The fertile hybrid males in color crosses... 1.2... 0... cece eee e nee eee General conclusions as to color and coat character. .......... 0... c eee ee eee eee Part II. GrowTH AND MoRPHOLOGICAL CHARACTERS. Introductory discussion. ...... 0... ccc cece cece eee eee tener eee nee eeeteeee . Growth: sisssieses oi car ehaweene oud THe data cos o4 ceheee a UROL 204 THA SUMA TT Hen eRE EERE RANG ES Gonder ae Comparison of growth curves..... Th AVELALCS co Sssscisueasiesd Saunsia dd oes hag Reunite Hake sak A gee The coefficients of variability 13 14 14 17 19 20 22 25 28 30 30 30 31 32 33 33 34 35 35 35 35 36 37 38 38 38 39 41 41 41 42 43 43 45 6 CONTENTS. Pace 14, Skeletal dimensions) «i: sicncien anc hee Sas Re SE OLE ad FEN EU ERNE DLE 63 The data on skeletal dimensions. ......6.500cs sce ct ven eeeeesaeenneesebeuee 63 Comparison of skeletal dimensions. ........... 0... e cece ete ence nen eens 65 "THE AVELAZE CIM ENSIOUS! sooo 55 esses ee esaveier de acoanain seh diene wim nanan Hea Maqueies oe 65 Coefficients of variability of dimensions.............0.. 000.0 cee ee eee 69 TOE Be skull suture sis cuseesce saa cncecscanei so % dot aeschca venenatis wee Leah th do abe phan adeno NONNCR ee Wed Le 72 16... Miseellancous morphological characters... cs0ccswx ss cca cunwcaes soe awenea dee eens 74 The interparietal. bones iss yego. eeegunpa carded don gii's a @ acerave toaweoneaiwinn iene wade ae wes 74 The siape-at tie sells. cs ovens oe 4e e424 G24485- 24 BY KHOR ROE ASRRE LEO eH NS HS 74 The efieet of aterility in the males; «, vcs caceny sass. ahuw eae ene werenn bene 75 Anomalies occurring in the hybrids. .......... 0.0... ccc eee 76 17. General conclusions as to growth and morphological characters.................4. 77 Part III, Tue Ferriviry of THE PARENT SPECIES AND HYBRIDS. 1S. Tntrodnetery Cace sO yew su eawk se 26d es aoe goer e en lgee pwede bo ey eee woseeue 79 10. "Te Tertigio’ of thewale bvitids, as o5 4 nceu anny s veces eaves ex daie Del een Tees 85 Moatetials and miethOds 1x44 es42005 4450945 5044S FoR ROE ERLE SRO e EEN OEE EO 85 The results of the simple breeding tests alone ............. 00sec ee ene eee 87 The results of all microscopic test8.......... 00. ccc eect eeeteeee 88 The results of a combined microscopic and breeding test..................00. 90 ‘THE TRRENLSHES GE BICTIY 5 oe ac dekh eee ned eemanees hb ee8s eke bireeemns 92 The male offspring of fertile male hybrids............ 00.0.0. cece ene eee 97 The secondary sexual characters... 0.00.00... ccc cette tne nee 99 20. The fecundity of the female hybrids............. 0.00 nee nen 100 21. "Phe: sex-ratio im the by rid Ske we cici.sin iyiersicatd, 4.0, 808 6 54 il wep: x aie ual Oar ods ig 'h moe eeddaneae janie 101 22. Sumenary and general coneluslone.. vs ceca en vo hs 6224006 Ee eee ARE EX ORDEWROERES 103 LABORS 5.6 e445 SURGE WER SHF EE EE EEE ESE RGNGHEIGERAAESA EAS E004 05 ERATE SERS 104 BTU TG AD YF scsi ass ado ga ight ed NESS Gosden aguante at doth eae Ae aS 129 DRACRIPUION OF PHAR Soc adn cctcanriesnn peaod ss aba de embed chaodmu puasesoeea we 133 GENETIC STUDIES ON A CAVY SPECIES CROSS. GENERAL INTRODUCTION. The genetic studies herewith presented were made possible for the author, by the reception of the foundation stock, in December 1909, from Dr. W. E. Castle. The first crosses had been made in 1903, and about 200 of the wild and intense wild-blooded hybrid animals had been born when the stock was received. The birth records, the weights, and such skeletons as had been saved, as well as the living hybrids, were made available to the author, who here expresses his gratitude for the privilege of using this material and for generous assistance, which was never withheld. He also wishes to acknowledge the valuable aid of Mr. Elmer Roberts, in the preparation of the manuscript. Most of the manuscript was written and most of the data were analyzed at the College of Agriculture of the University of Ilinois, to which the author is deeply indebted for liberal use of time and facilities. 1. THE SYSTEMATIC POSITION OF THE PARENT RACES. This paper is based on a study of the wild Brazilian guinea-pig, (Cavia rufescens Lund), the common domestic guinea-pig (Cavia por- cellus Linn.), hybrids between these, and subsequent progeny obtained inthe next eight generations by various matings. About 1,800 animals, wild or hybrid, enter in one way or another into experiments on color, growth, size, and fertility. Besides these, approximately 600 guinea- pigs, living under the same conditions in collateral experiments, serve as a basis for necessary comparisons. That the hybrids are the result of a species cross rather than a variety cross can hardly be doubted, since the 3 wild and + wild males are entirely sterile. In order to meet any doubt or criticism at the outset, I may briefly give my reasons for assigning the parent stocks to such diverse and distantly related species. In the summer of 1903 Dr. W. E. Castle received one wild male and two wild females from Mr. Adolph Hempel, Campinas, Sao Paulo, Brazil. These and their progeny were kept for some time at the Harvard Zoological Laboratory, and were removed later to the Laboratory of Genetics, Bussey Insti- tution, Harvard University. In the summer of 1911, three years after the last animal of pure wild pedigree had died, we again received from Mr. Hempel one wild male and one wild female. At first it was thought 7 8 GENETIC STUDIES ON A CAVY SPECIES CROSS. that these wild cavies belonged to the commonly described Cavia aperea Erxleben, but a more careful investigation showed later that they belonged to the less well-known Cavia rufescens Lund (Lund 1841, Waterhouse 1848, Thomas 1901). This cavy is considerably smaller than Cavia aperea or Cavia porcellus, both in total size and in the individual bone measurements. Thomas asserts that Cavia rufescens never reaches the size of Cavia aperea. The color is agouti or “‘ticked,”’ as in most wild rodents, but somewhat darker than the agouti of Cavia porcellus, because more black shows in the individual hairs and less yellow on their subapical bands. The belly varies from a light yellow to a slightly ticked condition. The systematists lay great stress on the formation of the last upper molar, in which a deep, narrow inden- tation on the outer surface almost separates the small third lobe from the body of the tooth. Lund describes his specimen from Minas Geraes, Brazil. In all essential points the wild animals in this experi- ment agree with the descriptions, plates, and general locality given by the above-mentioned authors. A report of the experimental work does not necessitate an argument on the number of differential characters which would infallibly place two types in those more or less arbitrary categories—‘‘species.”” It is sufficient for the purposes of this problem to find that the wild cavies used belong to a species more distantly related to the tame guinea-pig than are Cavia aperea or Cavia cutlert, according to the methods of most taxonomists. The taxonomists differ much among themselves. For instance, Waterhouse held that Cavia porcellus, Cavia aperea, and Cavia cutlert might all be placed in the same species. He found forms bridging typical differences. Darwin (1876) held that Cavia aperea was not the ancestor of the guinea-pig, basing his views on the fact that a distinct genus of lice infested each form. As far as his evidence goes, it might be considered decisive, for entomologists have reported that closely related mammals are infested by closely related lice (Osborn 1908). Giebel (1855) placed a number of cavy forms in the species aperea, and held that Cavia rufescens was only a variety of the larger Cavia aperea. Nehring (1889) considered Cavia cutleri to be the direct ancestor of our tame guinea-pig, being inclined to such a view on both historical and morphological grounds. He later showed (Nehring 1893, 1894) that Cavia aperea may be reciprocally crossed with the guinea- pig and give perfectly fertile offspring—fertile inter se or when mated back to either parent. Thomas (1901) is in doubt as to which of the two wild forms, Cavia aperea or Cavia rufescens, is the real ancestor of the guinea-pig. It would appear, from a comparison of Nehring’s experiments and the experiments described in this paper, that Cavia aperea must be more nearly related to the guinea-pig than Cavia rufescens is, for the latter gives sterile male offspring in a cross with the tame guinea-pig, whereas Cavia aperea does not. GENERAL INTRODUCTION. 9 It would be equally difficult to formulate any rule by which we could determine how great must be the difference in color, shape, size, tooth formation, and the like, between genuine ‘‘species,” but for our present purposes this, fortunately, is unnecessary. The reasons for considering the wild stock used in these experiments to be specifically distinct from the guinea-pig are as follows: (1) The skull characters, size, and color of our wild stock undoubt- edly place it in the species rufescens. I am indebted to Dr. G. M. Allen for a corroboration of this classification. (2) Hybrids between our wild stock and the guinea-pig are sterile in the male sex, regularly through two blood dilutions and in many cases through more blood dilutions. The other parent species, the common domestic guinea-pig, Cavia porcellus (also called Cavia cobaya),is too well known to require identi- fication or description. The peculiarities of the stock used in these experiments, if there be such, are described in the detailed discussion of their inheritance. The ancestors of the guinea-pigs, in these experi- ments, were obtained by purchase from dealers and fanciers, but the animals which were used were of known zygotic color formule, size variability, and fertility. 2. MATERIALS AND METHODS. THE WILD RACE. The original wild #1 was mated to wild ? 2 2 and 3, to increase the stock. (See fig.1.) He was likewise mated to his daughters, as were his sons, 6124 and 0338, and his grandson 0°55. The young of 92 died prematurely, and so do not figure in any of the later crosses; hence all the wild stock came from two original parents, 71 and 93. The pure wild line eventually died out, for, even with the greatest care and experience in handling domestic cavies, it was not possible to carry the wild stock more than 5 years in captivity. The animals were prone to fight. Only one female could be penned with one male at the same time. The total number born in captivity was 46, but of these only 4 females and 3 males reached sexual maturity. Our experi- ence does not agree with that of Nehring (1894), who realized little difficulty with Cavia aperea in captivity. This fact again distinguishes the two stocks and experiments. The two wild cavies received in 1911 have not bred up to the time of writing. ONE-HALF WILD HYBRIDS. The original wild male, #1, and his sons, 724 and 733, and his grandson, o'55, were used to obtain hybrids between the pure wild stock and the tame guinea-pig. (See fig. 2.) The reciprocal cross (tame male X wild female) was not obtained or even attempted, for it was feared that such small females might die in pregnancy when 10 GENETIC STUDIES ON A CAVY SPECIES CROSS. impregnated by the larger-sized guinea-pig male. The matings were obtained with much difficulty, for the wild sire at first harassed and bit the tame females almost beyond recognition; but, by keeping him in solitary confinement for some time, and then placing him with a female which had just given birth to young, copulation was success- fully brought about. The young appeared in due time (63 to 67 days) and in the usual guinea-pig number, showing that such wild males, producing an abundance of sperm, are wholly fertile with tame females. Our stock of tame females used as the mothers of the hybrids con- sisted of large healthy animals of known color varieties (except the dams in two cases of young not used in further experiments). The offspring were all agouti-colored like the wild father; 39 such 4 wild offspring were obtained, but of these only 10 females were successfully used for breeding purposes. The males were all sterile. I have used the terms 4 wild, 3 wild, 3 wild, +4, wild, etc., but wish to state here that these terms are used only for convenience, without implying blending inheritance. They simply denote the generation to which a hybrid belongs. ONE-QUARTER WILD HYBRIDS. Since the 4 wild males were sterile, the } wild females were mated to both parent stocks. When mated to the guinea-pig they produced 1 wild rufescens hybrids; but when mated to the wild Cavia rufescens they produced 2 wild rufescens hybrids. Of the + wild young 83 were obtained, sired by pedigreed male guinea-pigs. In this blood the males were again sterile; therefore the females were mated back to guinea- pig males. The numbers of sexually mature females increased with each generation; hence there was no difficulty in procuring sufficiently large numbers of the more dilute-blooded hybrid animals. THREE-QUARTERS WILD HYBRIDS. Only one wild male (#24) and one 3 wild female (950) were used for this part of the experiment, and they produced four young, of which two, a male and a female, reached maturity but proved to be sterile. (See fig. 3.) The wild males died out soon after this, and effectually put an end to this class of matings. ONE-EIGHTH WILD HYBRIDS, AND LATER GENERATIONS. Proceeding in the same manner used to obtain the previous genera- tions, the females of one blood were continually mated back to guinea- pig males to produce animals of the next blood-dilution. Thus, from our } wild females we obtained § wild, and from the + wild females we obtained ;/, wild. Up to the time of writing, the blood has been reduced to z+, wild, with s+, wild young in utero; 1. e., the Fy genera- tion. Naturally most of the animals now living are not so far removed GENERAL INTRODUCTION. 11 as this from the original stock. At present most of our pens contain gz wild hybrids. The numbers of hybrids obtained up to October 1911 were as follows: 2 wild, 39; 3 wild, 4; 4 wild, 83; } wild, 217; +1, wild, 312; 4, wild, 344; = wild, 122; +4, wild, 37; +4, wild, 2; total, 1,160. Since that time 600 more hybrids have been born. Unfortunately, for comparisons, mammalian species crosses are not common. When they have been made the number of offspring has been small, thus affording small basis for generalization. The most reliable data are drawn from species crosses among the ungulates, but ungulates are not adapted to laboratory experiments in large numbers. Species crosses are unknown among the Monotremata, Edentata, Insec- tivora, Chiroptera, Sirenia, Proboscidea, and Hyracoidea (Przibram 1910). The species crosses among ungulates, like horse and ass, or cow and bison, involve the question of sterility and fertility. The similar sterility in the cross of the wild and tame guinea-pig affords excellent material for comparison with these larger economic forms. FERTILE MALES IN MATINGS. The most interesting part of the whole problem is the origin of fertile hybrid males and crosses of such males with females of the different blood-dilutions and with guinea-pig females. Seven fertile males appeared among our 4 wild hybrids. The number increased rapidly in the ;', wild, 5‘, wild, and later generations. The importance of these males is apparent; for it gives opportunity to study sterility and fertility, and to test whether any segregation of characters in the direc- tion of Caviarufescens is possible. Previously, any segregation possible was in the direction of the guinea-pig, Cavia porcellus. The fact is realized that a large number of characters is involved, and it will there- fore require the observation of many individuals before we can reason- ably expect to observe complete segregation of either the guinea-pig or the rufescens characters as a group. Fortunately many of the characters are so unmistakable and definite as to allow of no doubt or uncertainty in their case. The detailed result of the matings of the fertile male hybrids is given in Part III. The young from such mat- ings have not reached maturity and consequently their bone measure- ments and growth curves can not be given at this time. ACCUMULATION OF DATA. It has been stated that a number of differential characters mark the wild guinea-pig in distinction from the tame. Records of the expres- sion of these characters and new characters which appeared have been made. Color.—Cavia rufescens is not known to occur in any color, except agouti of a rather distinct and specific type. Color records of each hybrid were made at birth. 12 GENETIC STUDIES ON A CAVY SPECIES CROSS. Growth.—The wild species was observed to grow more slowly than the tame guinea-pig. Hence weights of the wild, the hybrids, and the tame were taken at frequent intervals and recorded. The animals were weighed at birth, or shortly after, and then each week until the nature of the individual growth curve was established. After that they were weighed at less frequent intervals until they died or were killed. This method also afforded an opportunity to keep strict watch on the health of each animal, for a sharp, unexpected drop in weight indicated sickness, fighting, or some other disturbing cause. Sexually mature females were weighed immediately after parturition, in order to eliminate the error due to a varying number of fetuses. Skeletal dimensions.—Just as the wild animal is smaller in total size, so its individual bones were observed to be shorter and more slender than those of thetame. The skull, lower jaw, right scapula, right fore- leg, and right hind-leg of such adults as died were saved for further observations. Whenever a hybrid reached maturity and could no longer be used for other purposes, it was killed and the bones similarly saved. A careful examination of the growth curve and the bone sutures showed that guinea-pigs and hybrids are of full adult size when 15 months old. Measurements, of which a detailed account is given later, were made and tabulated. Fertility. The fertility of the wild, hybrid, and tame females was not uniform. Records of the size of each litter were kept, from which averages could be calculated. The wild males were fertile in captivity, but their 4 wild hybrid sons and their + wild grandsons were sterile. The problem immediately suggested itself: how great must be the blood dilution, or for how many generations must the hybrid females be crossed back to the guinea-pig, before producing fertile males? The numbers of males to be tested increased to such an extent that facilities were lacking to test their fertility by mating them to females. Furthermore, it is well known that a male may be potentially fertile, but fail to show it because of some physiological state, such as extreme emaciation from sickness, or through the sluggishness of obesity. Another method was devised. It was observed from many cases that the breeding test was negative whenever a male lacked spermatozoa in the epididymis or when these spermatozoa were few, degenerate, or non- motile. On the other hand it was found that fertile males invariably have many motile spermatozoa in the epididymis. Examination of the sperm content of the epididymis therefore affords a clear index of fertility. The examination is readily accomplished by placing a drop of the contents of the epididymis in normal salt solution at a bodily temperature and examining it under the microscope. An operation of this sort performed on one side of the body only does not preclude subsequent breeding of the animal operated upon. PART I. COLOR AND COAT CHARACTERS. 3. INTRODUCTORY DISCUSSION. Instances of alternative or Mendelian inheritance have been rapidly accumulating since the rediscovery of Mendel’s law in 1900, but most of the cases known among mammals are based on relatively simple and easily executed crosses, namely, crossing varieties of a species. Hence the criticism has been offered that this form of inheritance does not occur in species crosses or in nature. It has been maintained that Mendelian phenomena are the result of laboratory methods, in which we deal with man’s domestic varieties. No contention is offered that this or any other wild cavy mates with the guinea-pig in nature. We have no evidence for or against such an hypothesis. In fact, it is more probable that such crosses do not occur, for the repulsion which one species of mammal usually shows to mating with another was evident even in this experiment. When, however, a species cross is actually made, whether it is in the laboratory or elsewhere, the data accruing from the experiment may be legitimately offered to bear on the mode of color inheritance in a species cross. The papers of Castle (1905, 1905a, 1907, 1907a, 1908, 1909) and Sollas (1909) deal with the subject of color inheritance in guinea-pigs in a summary manner, and so much has been written upon this subject in other forms that I should feel most apologetic in offering more data upon alternative inheritance of color in plants or animals were it not for the fact that my observations cover a very definite category of cases which have received little attention up to the present time, and which may be of some general interest to students of heredity because of the nature of the cross which gave rise to them. The symbols used to designate the color and coat factors are, briefly, as follows: C, a factor necessary to the production of color inanimals. Albinos lack this factor; the allelomorphic condition is represented by c. A, a factor restricting black or brown in the individual hairs, pro- ducing the ticked or agouti type of coloration. This factor may restrict differently in different parts of the coat. Black and brown are restricted in the yellow subapical band on the dorsal surface. They may be completely restricted on the belly, giving yellow belly, as in the domestic guinea-pig; or they may be partially restricted, and so allow a ticked or barred appearance on the belly as well as on the back. The latter is the condition in some wild Cavia rufescens and some hybrids. The allelomorphic condition is designated by a. 13 14 GENETIC STUDIES ON A CAVY SPECIES CROSS. B, a factor for black. Black is usually considered as the most com- plete oxidation product of the yellow-brown-black series. Animals lacking this factor to produce black are brown, or can transmit only brown. The latter condition is indicated by b. E, a factor for the extended condition of black and brown pigmenta- tion, in distinction from the restricted condition. This factor produces self-colored black or brown animals, whereas its absence, designated by e, is characteristic of the black-eyed or brown-eyed red or yellow coat. Rf, a factor for the rough or rosetted coat character. Smooth-coated animals lack this factor, and the condition is represented by rf. Each color table deals with a single allelomorphic pair, thus keeping the ratios as simple as possible. A number of years ago it was necessary to explain the various kinds of chance ratios, but such discussion may now be advantageously omitted. Likewise it has been shown that the ratios obtained by dealing with two, three, or more pairs of allelomorphs without coupling are the squares, cubes, or higher powers of the simple 3:1 ratio. Hence, it is obvious that the more complex ratios may be obtained from the simple and we need not deal with all the color char- acters of each animal at one time, but just deal with a single character and its allelomorph in each case. The tables deal with zygotic consti- tution rather than somatic appearance; for instance, an albino may transmit agouti, and therefore be entered in a table in which all the animals entered transmit this factor, irrespective of the somatic colors, or lack of color. 4. THE AGOUTI CHARACTER IN THE WILD RACE AND IN HYBRIDS. HOMOZYGOUS AGOUTIS IN CROSSES. Agouti, the factor which restricts black or brown from the sub-apical portion of the hair and gives a barred appearance, is characteristic of Cavia rufescens. The character is common to all wild rodents. A number of investigations on rats, mice, and rabbits (Cuénot 1908, 1904, 1911; Castle 1905, 1905a, 1907, 1907a, 1908, 1909; Hurst 1905; Sollas 1909; Morgan 1911) give sufficient evidence that it acts as a unit character, dominant to the non-agouti condition, and segregating in the F, generation, according to Mendel’s law. The agouti of Cavia rufescens is of somewhat different appearance from that of Cavia porcellus or Cavia cutlert. It is darker than either, showing a narrower yellow subapical band and more black. There is some variation in this character in the wild rufescens, which accounts for slight differences in systematic descriptions. The belly-hairs of Cavia rufescens vary from yellow to slightly ticked, but in Cavia porcellus the variation is from complete yellow to yellow with a small amount of black at the base. In both species there is a constant relation COLOR AND COAT CHARACTERS. 15 between dorsal and ventral pigmentation; for the darker the dorsal surface, the darker is the ventral surface, and in any one animal of either species the dorsal surface is always darker than the ventral. Two facts may here be emphasized: (1) The agouti of the wild rufescens has less power to exclude black and brown from the hair than has the agouti of the tame. (2) Agouti, from whatever source derived, produces a more striking effect on the belly than on the back, restricting black or brown more completely in the former region. It is one and the same agouti factor, but it causes a different reaction in these two parts of the coat. Breed- ing many agoutis has shown that there are not two factors, one for restriction of black on the ventral side and one for the restriction of black on the back. If this were true, the two factors could be dissoci- ated and transmitted independently, but this has never been accom- plished. It may be objected that, even with this evidence, we can not be sure that two or more factors do not exist in complete coupling. The objection, in a measure, answers the argument, for if the coupling of factors is complete, we can only deal with them as one unit character. The surmise that the wild race would be found homozygous in agouti proved true (see table 1). The original wild male was father of 27 young, all agouti, like himself, while his sons and grandson sired 19 other agoutis. Had either one of the original parents, 71 or 23, been heterozygous, it would have been possible to extract recessive non- agouti individuals from the matings of their offspring inter se, for half of their offspring would likewise have been heterozygous. The con- viction that the wild race must be homozygous in agouti is furthermore strengthened beyond a reasonable doubt by the matings of 4 wild males with 10 different non-agouti guinea-pig females (table 2); 37 agouti young were thus produced. If the wild parent in any of these matings had been heterozygous, these matings must surely have produced some non-agouti offspring, but such was not the case. Our point is therefore well established by these 83 offspring. A second point, in a measure dependent on the first, may be stated thus: the wild agouti character dominates its absence, even though the absent condition is presented by the tame female parent.' In both particulars the results agree with similar matings among guinea-pigs. Such was Nehring’s experience, also, with Cavia aperea; for, though he did not understand alternative inheritance at the time of his experiments in 1893, he nevertheless gives data which show conclusively that Cavia aperea, a different wild Brazilian species, is likewise homozygous in agouti and dominant to its absence in a mating with Cavia porcellus. It is rather surprising that no one has studied Nehring’s data and referred to them or pre- 1Castle (1905) has reported on the dominance of the wild agouti when mated to non-agouti. The wild stock at that time was supposed to be the common Brazilian Cavia aperea. The fact of dominance reported was correct; the error of classification is corrected in this paper. 16 GENETIC STUDIES ON A CAVY SPECIES CROSS. sented them as evidence of Mendelism in a species cross, when the criticism on alternative inheritance in species crosses was first made. Possibly it was the lack of numbers in his experiments, but surely, as far as they go, the results are quite conclusive on this, as well as some other points. All 4 wild hybrids recorded in table 2 were heterozygous in agouti, for they were the result of matings between wild agouti males and non-agouti females. The agouti which they bore came from one defi- nite source, the wild strain. Tables 1 to 12 deal with both tame and wild agouti as one. This method of procedure is followed because both wild and tame agouti have many common characteristics. The discussion of their differences is reserved for tables 13 to 15. It has been proven that agouti obtained from the wild is dominant over the non-agouti condition in the tame. Therefore a number of matings were made to investigate the reciprocal cross, in which tame agouti guinea-pig males were mated to wild hybrid females. Two homozygous agouti males (971961 and #2157) were mated to 10 dif- ferent + wild females (table 3); 3 of these females were heterozygous in wild agouti, and the rest were non-agouti animals; the 27 young obtained were all agouti, like the father. These young should be of two zygotic classes; those produced by the 3 agouti females should half of them be homozygous and the remainder heterozygous agouti animals, whereas all the young from the 7 non-agouti females should be heterozygous. Both zygotic classes were produced; for in testing the offspring of the 3 heterozygous females, one female (580) was found to be heterozygous, and one female and one fertile male (? 485 and 506) were found to be homozygous. But the offspring of the 7 non-agouti females used were invariably heterozygous. The result of these matings shows that agouti obtained from either wild or tame is dominant to non-agouti, whether this latter condition is derived from tame females (table 2) or from hybrids (table 3). The matings indicated in table 4 corroborate this view. In this experiment 5 different ;1, wild hybrids were used. The hybrids were the result of matings calculated to produce homozygous agouti by crossing females of the wild agouti type with males of the tame agouti type. These 5 agouti hybrids showed their homozygous character by producing 21 offspring, all agouti. Their gametes evidently carry agouti in all cases, although this agouti was derived from two very different sources, the wild and the tame. When such gametes are formed they are presumably of two types, one bearing wild agouti and one bearing tame agouti; and when they meet gametes without agouti, the zygote formed produces an agouti animal, the agouti being theoretically in one case like the wild and in the other like the tame. The numbers are small, but quite conclusive; for not only were all the offspring agouti, but among them occurred agouti individuals of two COLOR AND COAT CHARACTERS. 17 different sorts, one sort resembling the agouti of C. rufescens, the other that of the guinea-pig. If we designate the tame agouti as A and the wild agouti as A’, then these five parents had a zygotic formula of AA’. It is evident, then, that they must have produced certain gametes which bore A, the powerful tame agouti factor, and others which bore A’, the weak wild agouti factor. The young accordingly were of two sorts, wild and tame agouti. This subject will receive consideration in a later part of this paper. For the present, all kinds of agouti will be considered as one, irrespective of their source. Summary.—tThe wild Cavia rufescens is homozygous in agouti. This condition is epistatic to the non-agouti condition of the tame guinea- pig. The agouti of the tame is likewise epistatic to the non-agouti condition of the hybrids. Hybrids may be produced which are homo- zygous in agouti. In table 5 the summary of tables 2, 3, and 4 shows that 85 agouti offspring were produced from matings of pure agouti animals. Therefore the agouti factor is epistatic, whether found in the wild, the tame, or the hybrid. This agrees with Nehring’s results on Cavia aperea, though his interpretation was different. To make data plain and not suppress any facts, it should be stated that a few albinos enter into some of the tables. Such albinos, we know, carry all color factors in the same proportions as their colored brothers and sisters, with the exception of the basic color factor itself. It may therefore be understood that albinos have been omitted from the tables, unless a thorough breeding test has demonstrated to what color class each albino belongs, in which case it has been included in the corresponding colored class. HETEROZYGOUS AGOUTIS MATED TO NON-AGOUTIS. All the 3 wild hybrids derived from the cross (table 2) of a Cavia rufescens male with female guinea-pigs were supposedly heterozygous in agouti; 9 female 4 wild hybrids were mated with male guinea-pigs. The sterility of the male hybrids prevented a breeding test in their case. The female 4 wild all bore the agouti coat and had received the agouti factor from the wild parent. A priori, they should have been hetero- zygous in this factor, having received it from one parent only. Such they proved themselves to be in their matings with the non-agouti guinea-pig males. They gave offspring of two sorts, agouti and non- agouti (in this case all were black) in approximately equal numbers; 83 such offspring (table 6) were obtained, of which 47 were agouti (also heterozygous) and 36 were non-agouti. To strengthen the case, it may be pointed out that each female 3 wild should prove her hetero- zygous condition by giving both sorts of young, provided the numbers are large enough; 7 of the 9 females gave both sorts of young. One female (975) gave 4 agoutis only, and another female (9 72) gave 4 non- agoutis. Presumably these last two females would have yielded both 18 GENETIC STUDIES ON A CAVY SPECIES CROSS. classes of young had they been more prolific. The law of probable error would account for the occasional occurrence of these ratios of 0 : 4 and 4 : 0, where we expect equality as an average result. It is, therefore, clear that when a wild species of cavy known to be homozygous in agouti is mated with a tame race lacking agouti, and hybrid females are thus produced, these are heterozygous in the agouti factor. Let us follow the 4 wild agouti offspring of the heterozygous 4 wild females. Since they were produced by matings in which only one parent (the 4 wild) carried agouti, they too should be heterozygous; 20 females of the 47 agouti + wild individuals were mated to non-agouti guinea-pig males (table 7). Just as in the matings of table 6, each female should in this case produce both agouti and non-agouti young. Females 95, 97, and 98 produced young of only one kind as far as we know, but since the total young of these 3 females is only 4, we may legitimately neglect them. The total number of offspring of all the females in this experiment (table 7) was 55 agouti and 59 non-agouti, a close approximation to the expected equality. It is interesting to note that, whereas the 4 wild females gave a slight preponderance of agouti young, the 4 wild agouti females gave the reverse. Adding the matings of tables 6 and 7, we see that our intense wild-blooded hybrids acted just as the guinea-pig does in matings of this description, and produced an approximate equality of agouti and non-agouti young, in this case 102 agouti to 95 non-agouti. The most probable expectation is either 98 or 99 of either sort. We have traced (in tables 8, 9, 10) the matings of all the rest of our heterozygous agouti females with non-agouti males. Since, in the intense wild-blooded hybrids, the color inheritance for agouti has been shown to be the same as that described by Castle (1905) and Sollas (1909) in the guinea-pig, we had no reason to expect our dilute-blooded hybrids to behave differently, for they surely are still more like guinea- pigs than the earlier generations of hybrids. In table 8 are summarized the matings of heterozygous 4 wild females with recessive guinea-pig males and in the lower division of the table matings reciprocal to those just described. Since the reciprocal matings gave like results they may be combined. ‘The offspring, all told, are 50 agouti and 37 non-agouti. In tables 9 and 10.are summarized matings in which the females alone bore agouti. They evidently produced gametes of two sorts in equal numbers, those bearing agouti and those without it. Tables 6 to 10 deal with similar matings, namely, the heterozygous agouti mated to non-agouti, in the different blood dilutions. The summary of these experiments constitutes table 11. It is noteworthy that some of these agouti hybrids received their agouti character from the original wild parent and some others (after the 4 wild of table 8) received the agouti from the tame. The two are distinguishable. COLOR AND COAT CHARACTERS. 19 The tables show that the wild agouti has been kept in a heterozygous condition up to the 34, wild females. Matings made since these tables were constructed prove the same up through the ;4, wild, i. e., for Seven generations. In other words, one dose of agouti was received from a wild race, and this one dose was handed on for seven generations; and each female that received it passed it on to one-half of her offspring in the next more dilute generation. Also, one dose of agouti derived from tame guinea-pigs was given to some zs Wild hybrids, and this was similarly inherited for three generations. In all these cases, agouti may be said to act as a unit character, just as in the well-known tame crosses. Summarizing all the matings of all generations of hybrids (table 11), in which one parent is heterozygous in agouti and the other is recessive, such matings have produced 226 agoutis and 214 non-agoutis. The most probable expectation is 220 of each sort. A departure of 6 individuals is explicable by the law of chance. HETEROZYGOUS AGOUTIS MATED INTER SE. The matings of female hybrids, heterozygous in agouti, to male guinea-pigs, likewise heterozygous in agouti, are of very limited number, but more are in progress at the present time. Eight female hybrids, known to be heterozygous, were mated to 5 different male guinea-pigs, also heterozygous. The results of these 8 matings (table 12) are 36 off- spring, of which 32 are agouti and 4 non-agouti. The most probable expectation is 27 agoutis to 9 non-agoutis. In these matings, 71436, &2196, and 2002 did not produce any recessives, yet table 8 shows that #2196 and #2002 were heterozygous. Male 1436 is known to be heterozygous from pedigree, so that his 4 agouti young (table 12) do not indicate any error. Male 1917 (table 12) produced 9 agoutis and only 1 non-agouti. The ratio 32:4 shows a considerable excess of agoutis over the usual 3:1. Such deviations are usually explained by the Law of Error, according to which any ratio might be obtained in place of a 3:1; but the wide departures from such a ratio must occur with minimum frequency. Possibly the deviations observed in this case are due to chance. In mating heterozygotes inter se we expect two visible classes, but three actual zygotic classes. One-third of the agouti individuals should breed true; two-thirds should be heterozygous; the recessives should breed true. To test the validity of the ratio, the breeding records of the agouti animals produced by the experiment of table 12 have been studied. It was possible to mate 12 agouti females and 1 fertile agouti male to non-agouti guinea-pigs. The rest of the 32 agouti animals 1Since these records were made, 103 young have been born in crosses similar to those above. Of these young, 46 were agoutis and 57 were non-agoutis. Adding these to those previously obtained, we have aratio of 272 agoutis to 271 non-agoutis—actually the most probable expectation. 20 GENETIC STUDIES ON A CAVY SPECIES CROSS. either died or were sterile males; 8 of the 13 animals tested have proved to be heterozygous, 3 homozygous, and 2 are questionable, for the last produced only agouti young, but in such small numbers that no con- clusions can be drawn in regard to their zygotic formula. It is apparent that both the expected classes of agouti individuals were produced, and that the heterozygotes (8) occur approximately twice as frequently as the homozygotes (3 to 5). These homozygous animals are interest- ing particularly because the agouti came from two sources, the wild and the tame, and they produced agouti young of two sorts. It seems paradoxical to assert that a homozygous agouti animal produces two sorts of agouti, yet, as we have already observed, the agouti of C. rufescens is distinguishable from that of C. porcellus. There is a sharp distinction between a factor and its allelomorph. No matter how much variation there may be in the tame agouti pattern, it always segregates clearly from its absence. The same has been shown for the wild agouti in tables 6 to 10. There is a certain amount of variability to all unit characters. This is especially true of the wild agouti pattern in a heterozygous condition in hybrid animals. Where the wild agouti pattern has been so modified in the hybrid animals that it can be distinctly discriminated from the tame, it offers splendid material for a cross with tame agouti. Although the wild has been described as somewhat darker than the tame agouti, hybrids arose which were nearly black, so weak was the wild agouti factor (see figs. 4 to 9). Without further preliminaries, the variability of the wild agouti and its action in crosses with the tame may be appropriately discussed. THE WILD AGOUTI AND TAME AGOUTI CONTRASTED. In the preceding discussion all agouti individuals have been classed together, irrespective of the differences which have been indicated as distinguishing wild from tame animals. Such is the usual method of procedure in genetic studies. For instance, in crosses of English-pat- terned rabbits, bearing a dominant restricting factor, with self-colored rabbits, the English pattern is held to act as a unit. The differences between various animals, possessing the same unit character, are explained by postulating either variability in this one unit character or a number of similar or dissimilar genes for this one character, or other modifying unit characters, such as intensity, dilution, and the like. Black, in crosses, is dealt with in much the same way, and differences, easily discernible or seen with difficulty in different individuals, are similarly explained. A clearer example of this is shown in the crosses of hooded and self-patterned rats. The hooded pattern shows a very wide range of variability, yet any hooded pattern acts as a unit in crosses with self. Pure genotypical races as regards color in animals have not been isolated. Even agouti itself, in variety crosses, has been COLOR AND COAT CHARACTERS. 21 treated as a unit; yet, in the guinea-pig, differences in the agouti factor can be seen. Recently Morgan (1911) has reported on a cross between gray-bellied agouti mice and light-bellied agouti mice, and although the numbers given are small, it is quite clear that each form acts as a unit, and that the gray-bellied agouti is recessive. Although Morgan does not state it, it would appear that the difference between gray-bellied agoutis and light-bellied agoutis is not a difference in separable belly- ticking factors. The difference is probably a difference between two kinds of agouti, in which the peculiarity of one agouti is a weakened restricting power and the consequent appearance of black on belly hairs, whereas the other agouti is a more powerful restrictor and therefore gives yellow or light belly, without the usual black in the belly hair. It has been found expedient to treat all kinds of agouti as one, whether found in the wild, tame, or hybrids. This treatment of the ticking factor has been adhered to, because all forms of agouti have some qualities in common, and whatever the agents may be that cause the exclusion of black or brown from a part of the hair, the qualitative effect of the agents appears the same, but the quantitative effect varies. To be concrete, all the agouti animals have a factor which restricts black or brown in the subapical band of the ticked hair, but the amount of this restriction differs, particularly when a wild agouti or a hybrid- bearing wild agouti is contrasted with the tame. The common qualities of all agoutis are as follows: (1) All restrict black or brown on the individual hairs in the sub- apical band, giving each dorsal hair a barred appearance. (2) Any agouti expresses itself more powerfully on the belly than on the back, restricting black more in this region. (3) Any agouti is epistatic to the non-agouti condition, and allelo- morphic to the absence of agouti. But to class all agoutis together, without a thorough consideration of their differences, would be a superficial method of treatment. From an examination of many tame agoutis the conclusion is reached that these never show the condition which the wild agouti presents in some pure wild animals and in some hybrids. These differences are briefly as follows: (a) The very weak restricting power, which some wild individuals and some hybrids show, is unknown in tame guinea-pigs. This differ- ence in the restricting power may be readily seen from measurements of the yellow subapical band, for the greater the power to restrict black or brown, the broader the yellow band from which these pigments are excluded. The narrowest yellow band on a mid-dorsal hair of a tame agouti animal measures about 2mm. The yellow band of a hybrid or wild agouti may measure as small as 1 mm. In a number of cases the wild agouti was so powerless to restrict black in young hybrids that yellow was not visible at all in the dorsal hairs, and only very slightly 22 GENETIC STUDIES ON A CAVY SPECIES CROSS. visible on the belly. Such animals show an extremely slight sprinkling of agouti hairs when they become adult (figs. 6 to 9). (b) No tame agouti guinea-pig, to my knowledge, has ever shown a ticked belly, by which term I understand a condition in which the individual hairs are barred with yellow and have black tips and bases (figs. 4 and 5). I do not mean that all wild C. rufescens individuals and all wild hybrids are a very black agouti with ticked bellies. Such is not the case. The agouti pattern in the wild, and in hybrids receiving agouti from the wild, varies from a form very closely com- parable to the tame to forms almost indistinguishable from black, the latter occurring only in the hybrids. MopiricaTION OF THE WILD AGOUTI. To leave comparisons and return to the wild agouti pattern, it may be said at the outset that we do not know how the different shades of wild agouti are inherited when the wild C. rufescens individuals are mated inter se. The wild were animals that would not bear much handling, and so our records simply state that they were of the agouti pattern, with some additional data such as “‘dark”’ or “‘light.”” They could not be classified as so many distinct forms, for their range was great. However, it would have been desirable to know if the darker forms were hypostatic and whether any forms could have been gotten which breed true to one shade as far as could be detected by our crude methods of classifying by visual inspection. The apparent confusion and contradictions were only increased when the wild were mated to non-agoutis to produce 4 wild hybrids, hetero- zygous in agouti. Although these animals were heterozygous in the agouti factor (each one having received its share of agouti from one gamete, coming from the wild sire), they produced both dark and light agoutis of various shades in addition to recessive non-agouti offspring. All of the female wild hybrids were mated to non-agouti males up through the matings of the + wild; hence we are sure of the source of the agouti in every case, and no admixture of tame agouti could have occurred. The 4 wild females also produced both dark and light agoutis, irrespective of whether the mother was dark or light. As the wild agouti was being passed from one generation of hybrids to the next more dilute generation of hybrids, one fact stood out very clearly. Weaker agoutis gradually made their appearance; in fact, so weak was the agouti becoming that it failed to restrict black altogether dorsally and only very slightly on the belly in some cases (see figs. 6 to 9). This weakening of the power of the agouti factor can not be attributed to the fact that the wild agouti is always less potent to restrict black, which comes wholly or partly from the guinea-pig source; for, as has been stated, some hybrid females with strong agouti produced young with weak agouti, and vice versa. COLOR AND COAT CHARACTERS. 23 To put this matter in concrete form, table 13 has been drawn up. In this table all mothers and young are classified as ticked-bellied, dark-bellied, or light-bellied animals. The correlation existing between the ventral and the dorsal sides allows the inference that the shade of agouti on the back of animals classified as ticked is the darkest, whereas the back of animals marked light is the lightest, and an intermediate category, dark, falls in between these two. The animals which had the hair on the belly barred with yellow, but with hair-tip and base dark, were called ticked-bellied, and these animals were the darkest hybrids, both dorsally and ventrally. A few animals were called dark- bellied which had hair on the belly that was yellow at the tip but had much black at the base. Those animals in which the hair on the belly was entirely yellow or yellow with little black at the base were classified as light-bellied. These last were the lightest animals dorsally and ventrally and resemble the domestic guinea-pig closely. All the mothers were heterozygous in agouti, having received their agouti factor from one parent, the wild, or the wild hybrid. They were mated to non- agoutis and produced equal numbers of agouti and non-agouti offspring, and have been discussed in this light under tables 6 to 11. Now, table 13 shows that these same agouti offspring were of variable character. The recessive non-agouti offspring are here disregarded. The Cavia rufescens had been mated with guinea-pig females, and yielded all agouti offspring. The records show that 11 were very dark with ticked bellies, 1 dark with dark belly, and 2 light with light bellies. Just what the rest were can not be told, for they died young or were aborted. The 4 wild used as mothers of the } wild had ticked bellies, and are entered on the first line of table 13. In spite of their dark color they produced only 18 like themselves (43 per cent), 5 intermedi- ates, and 19 light agoutis. The } wild with ticked belly transmitted their character to a large proportion (90 per cent) of their offspring, producing 19 ticked bellies and 2 light. The § wild with ticked bellies, and all hybrids thereafter, produced only ticked-belly offspring (100 per cent). Since the construction of the table, new experiments with fertile hybrid 74; and z+, wild males show that these also transmit the very dark agouti with ticked belly to their offspring, irrespective of whether they are mated to } wild non-agoutis or to guinea-pig non-agouti or whether they are fathers of +'y'y wild, or +3, wild, or 23> wild. The dark-bellied females used were only two in number, both 3 wild animals; one produced a dark-bellied young one and the other a light- bellied one. They evidently do not always transmit agouti just like their own, but nothing can be said further than that. The light-bellied females also fail to transmit in all cases agouti which acts just as their own; for the ¢ wild mothers with light bellies gave 7 ticked-bellied young (41 per cent) and 17 light bellied young. The 24 GENETIC STUDIES ON A CAVY SPECIES CROSS. 3 wild, a single individual (9 140), though light-bellied, produced only young with ticked belly. Thus it is seen that light-bellied may produce the darkest shade and vice versa. It may be objected that the difference between the lightest and darkest is a small one, and renders close analysis and tabulation impossible. Such an objection is hardly valid when one considers that the darkest forms are often almost indistinguishable from black, whereas the lightest form is almost as yellow as an ordinary golden-agouti guinea-pig (figs. 4 and 5). Whether or not light agouti females would gradually or quickly be replaced by dark ones upon continued crossing with the guinea-pig can not be said, for further crossing of the light- bellied females was omitted at the time and no light-bellied females occurred after the 3 wild generation, but a few light-bellied 4 wild are still alive and, with these, it is hoped to investigate the question further. It is most perplexing to assign reasons for these various expressions of the agouti factor. One can hardly suppose that the very darkest agouti, which is almost black, possesses precisely the same thing which was contributed by the wild. In some cases (275) the 4 wild was very dark. In others (as through the series, 1991, 39723, 7’. 71082) the change was carefully watched and the transition was noted, but it did not take place in one generation. It might be supposed that the C. rufescens agouti factor has inherently less restricting power in the hybrids than in its own species, but this explanation obviously will not apply to those hybrids which are light-bellied, nor to those cases in which a gradual loss of restricting power was observed to occur in a series of generations. Furthermore, it does not explain why light forms gave both light and dark, just as the dark forms gave dark and light progeny. No matings of any description among tame guinea-pigs have yet made it necessary to postulate a number of similar agouti factors which are coupled. If wild agouti is held to be made up of Aj, As, Az, . . . . Aj, then it could be supposed that one or a number of these factors dropped out and gave a weaker and darker agouti. This would explain how 963, 968, 969, and other 4 wild animals happened to be very dark, because of a weak agouti with less genes; but it would never explain how some of the F, offspring and all of the F; offspring of particular females could possibly acquire these lost genes again and become light yellow agoutis with an agouti factor that is more powerful to restrict black. No admixture of tame agouti can be considered a causal agency in the change, since tame agouti hybrids were not pro- duced until the F, generation. In analyzing the case, it must be remembered that the Cavia rufescens agouti factor has been acting on Cavia rufescens black for centuries. Whatever agouti is, it is something which determines physiologically a rhythmical deposition of pigments in the growing hair. It is not sur- prising that such an activator, or whatever it is that is contributed COLOR AND COAT CHARACTERS. 25 by the sperm of the wild male, upon entering the egg of a tame female, should show many strange and unaccustomed reactions, disturbances, and possibly modifications. No one was surprised that Hertwig (1910) could cause crippled embryos to appear by treating frogs’ eggs with radium rays, and no one need postulate that such treatment eliminated some of the genes necessary to the normal development of certain organs. And so, the series of reactions which take place in a fairly stereotyped manner, when wild agouti develops in the pure wild race, may well be upset when one or several materials, necessary for this series of reactions, are carried by the wild sperm to such an unaccus- tomed environment as the egg of another species. This modification of agouti does not vitiate the Mendelizing inheritance shown in tables 6 to 11, for the material body which carries the agouti factor originally contributed by the wild sperm separates from its homologue, contrib- uted by the egg. The material bodies or carriers (possibly chromo- somes) separate. The activator of the rhythmic deposition of pigment in the hair, the agouti factor, residing in one of these carriers may have been modified or unmodified; yet, modified or unmodified, it separated from its allelomorph. Summarizing the facts observed: (1) Each 4 wild hybrid received a single dose of agouti from a wild male; 11 of the 14 } wild were dark with ticked bellies, and varied from forms much darker than the wild to forms like the darkest wild. (2) This modification shown by some 3 wild females was present in their offspring for the next successive six generations. In some cases the agouti gradually became darker, but in others the change took place more quickly. (3) The modification shown by other 3 wild females did not persist in all cases, for they produced light individuals as well as very dark ones. When light + or + wild forms were thus produced, these gave rise either to very dark forms again or to light forms. When dark } wild were produced they also gave dark and light offspring. Disturbances which quite baffle the cut and dried Mendelian inter- pretation are not unknown in wide crosses. Not only do we find meta- bolic disturbances, as in the echinoderms and insects, but in cases where adults have been raised there often occur gynandromorphs, hermaph- rodites, and the like (Standfuss, 1895). Up to the present time the mitoses of the hybrid germ-cells in these crosses have not been given the study which they deserve, and consequently an intimate acquaintance with internal mitotic phenomena of hybrids has not been formed. Mopiriep WiLp AGouTI IN CROSSES. Irrespective of the uncertain manner in which the agouti character expressed itself in the first three hybrid generations, there were some families which consistently gave dark forms for a number of generations, 26 GENETIC STUDIES ON A CAVY SPECIES CROSS. and since these were easily distinguishable from the light tame agouti, several crosses were made, into which they both entered, and many more are in progress. Ten different § wild females and one +, wild female were used in the following crosses: Cross 1: 9247, 9248, and 9311 were crossed with guinea-pig males homozygous in agouti. Cross 2: 9108, 9131, 9166, 9172, 9198, 2203, 9219, and 9536 were crossed with male guinea-pigs heterozygous in agouti. The result of the first cross was a complete dominance of the tame, light, and powerful agouti over the wild, dark, and weak agouti; hence all the young were light yellow agoutis with light bellies. If the wild heterozygous agouti is designated by A’a, and the tame homozygous agouti by AA, then the gametes formed and zygotes resulting from their union in this cross were: ACES ecns Ge eiieses fos Gales yee vet sede gametes of hybrid. A SbrApccckuans hey ase meewenwied oak tan gametes of tame. DAA DAD o's sca srcace. dee ccceush aha tia wna Sc zygotes. It is evident that in half the zygotes produced are found both kinds of agouti, while in the other half only tame agouti occurs. Since the tame agouti is dominant, all zygotes look alike, but the heterozygous animals should give only tame agoutis and non-agoutis when they are bred to non-agouti animals. Their gametes should be A-+a, and combined with those of a recessive, a--a, should give zygotes 2Aa+2aa. On the other hand, the animals homozygous in agouti should produce gametes A and A’; and when such animals are mated to recessive non- agoutis, with gametes a-+-a, they can give only young of the two sorts expressed by the formula Aa+..... F, zygotes. aaBbCe aabbCc 13D) + 3(2D) + 3(D) +1(d)......... F, distribution of classes, It is apparent that with three allelomorphic pairs the coefficients of the classes are derived from the expansion of (1+1)’, and that each class has the dominant-size factor represented one less time than the preceding class, the first class having it three times. The total number of individuals is 2* or 8. Hence, for ‘‘n” allelomorphic pairs we would theoretically expect a series as follows: UnD)+al (n—1)D] + 22SY (n—2yD] + PETVE—) (nay) +....+n{{n—M—1)}D] +1m—n)D. This means that the coefficients for the classes are derived from the expansion of (1+1)" and the dominant factors are represented in the GROWTH AND MORPHOLOGICAL CHARACTERS. 51 classes in an arithmetical progression derived from the exponents of the first term of the binomial, 7. ¢., n, n—1, n—2, n—3, ...n—n. The total individuals in the series would be 2°. Had the same heterozygous F, hybrids been mated to the larger parent instead of the smaller, the distribution of classes in the resulting F, generation would appear as follows: ABC + AbC + ABe + Abe + aBC + aBe + abC + abc...... F, gametes. BC cA DO cictscis saben see ect th eins Aa oy baghdad. coe larger parent gametes. AABBCC + AABbCC + AABbCce + AaBbCe AABBCe AaBBCe ... F, zygotes. AaBBCe AaBbCC 16D) + 36D) + 38(4D) +1(8D).... F, distribution of classes It is apparent here that the coefficients of the classes are again derived from the expansion of (1+1)’, but, unlike the previous illus- tration, we find the dominant factor represented in the classes in an arithmetical progression, the first term of which is equal to twice the ely) number of allelomorphic pairs. Hence, for ‘“‘n” allelomorphic pairs we would theoretically derive a series as follows: 1(2nD)-+n{(2n—1)D] + 2S) jean—ayp] + BOTVO—2) +... . tn[{2n—@—1)}Dj+12n—n)D. [2n—3)D] The gist of all this is that the F, generations of which we are speaking would theoretically show a range from the larger to the smaller parent with the mode in center when the F; has been produced by mating the F, individuals inter se. An F, generation produced by mating the F, to the smaller parent shows a range from the F, to the smaller parent, with the mode half way between these. An F; generation produced by mating the F; to the larger parent shows a range from the F, to the larger parent with the mode half way between. This is, briefly, the theory of multiple factors as applied to size- inheritance. If, after sufficiently numerous experiments with plants and animals, it is found to be applicable to such complex cases, it will show that segregation into apparently continuous classes is really dis- continuous, or, in other words, Mendelian. At present we know of no adequate hypothesis, other than the Mendelian, by which to explain the uniform F,; generation, the more variable F, generation, the recovery of parental types, and the tendency for certain recombinations to breed true while others split up again. There is a small number of cases of size-inheritance in which a Men- delian explanation seems well justified. It is logically defensible to resort to this explanation when possible, since it fits a large number of cases involving qualitative characters. However, it is too early to insist that size-inheritance is universally Mendelian, for the number of crucial experiments is few. 52 GENETIC STUDIES ON A CAVY SPECIES CROSS. In actual breeding experiments one would undoubtedly meet with much deviation from a perfect blend of quantitative characters in the F, generation, or from such a distribution of F2 classes according to the formula (1+1)”", as in the hypothetical case used above as an illustration. This is particularly true of size-characters in which the theory of multiple factors, incompletely dominant, is most often invoked; for external conditions affect growth and size very easily. Further- more, there are many other misleading circumstances in such a complex that render analysis difficult. How often could we be sure that a parent race possessed, or was homozygous in, each one of the multiple factors affecting a character; or how often would we find them so, especially in animals? Different individuals in the parent strains might appear alike in a certain character and yet carry different sets of genes for this character. Hayes (1912) had a case in tobacco which could be inter- preted in this way. He crossed two varieties of Nicotiana tabacum, both having about the same mode, mean, and low coefficient of vari- ability with regard to number of leaves. The F, was like the parents, but the F, showed such a marked increase in variability that he was led to believe there had been a recombination of several factors for leaf-number. The argument involved in his explanation is essentially as follows: one parent might have a formula AABBecdd and the other parent aabbCCDD. They would be of the same leaf-number, since each had the cumulative effect of a double dose of two factors, and they would breed true because each was homozygous. ‘The F; genera- tion, AaBbCcDd, would also be of the same leaf-number, having the cumulative effect of four factors. But when the F, plants were crossed, the F, generation could have recombinations ranging from AABBCCDD to aabbeedd. The frequency distribution of the classes would be obtained by expanding the binomial (1+1)*®. Hence, plants occurred with much larger and with much smaller leaf-numbers than in the parental forms or the F, generation. Thus, in actual breeding experiments, one might use parent plants which were of identical appearance but of different zygotic formule. In the simple illustrations of the theory, we suppose that one dose of each factor, such as A,, lends an effect about equal to that of any other factor, such as Ag, Az, Ay. . . . A,. But we do not really know for how much influence each factor might be responsible, or whether any one factor always causes the same result under all conditions. Factors in a heterozygous condition may act more vigorously (East and Hayes 1912), or the vigor due to heterozygosis might raise the size of certain classes only. Sterility or partial sterility of one sex might also impair any sort of an analysis on the theoretical scheme suggested. Environmental influence might affect certain individuals subject by chance, or they might regularly affect individuals of a particular zygotic formula. GROWTH AND MORPHOLOGICAL CHARACTERS. 53 Physiological correlation is not always explained by gametic coupling. It is not difficult to understand how a whole organism or parts of an organism are permanently influenced by even normal conditions. For example, we should hardly expect a normal but very small rabbit to have as large ears as a large rabbit, although both might have theo- retically the same set of genes for ear-size. In fact, if one carried out the whole scheme of independent size-factors without reference to physiological correlation it would lead to an absurdity. If a guinea- pig had genes for a small radius and a large ulna, or for a large tibia and a small fibula, would the animal be a cripple? In dealing with the inheritance of size of certain bones of the body, one can not over- look the influence which other parts, or even the whole body itself, may have upon the development of particular characters studied, irrespective of the hypothetical genes. It is well known that certain color characters in plants and animals develop only through the interaction of two or more independently transmitted factors. Thus, the factor for the agouti pattern in the hair of rodents acts only when black or brown is present in the zygote; but black or brown pigments in turn are restricted to the eyes and extremities unless the extension factor is present. It may be added that the basic color factor must also be present in order to activate the development of color. Therefore, to obtain the agouti pattern it is necessary to have at least four independently heritable factors, viz, the color factor, the extension factor, the brown or black factors, and the agouti factor. When we recall such facts as these, and realize that several or many factors may interact in the production of size- characters, we see how difficult it is to attempt or rather attain a satisfactory solution. Considering briefly the evidence which tends to show that a number of factors may exist for one and the same visible character, we find comparatively few experiments. Most of these arein plants. Nillson- Ehle (1909,1911) paved the way by showing how some apparently continuous variations might be interpreted as discontinuous variations. The black glumes of oats, he showed, might be due to two factors, either of which alone could cause the development of black in the glume. If a plant homozygous for both kinds of black (B:B:B,B.) was crossed with a plant lacking black (bib:bebe), the heterozygotes were black and hold a formula B,b,Bybz. Crossing the heterozygotes inter se gave an average of one entirely recessive individual in every 16. It proved to be a simple dihybrid cross, in which 15 of every 16 F; individuals carried at least one dose of a dominant factor and were black. The F, generation should theoretically consist of 9 B,Bz : 3 Bibs : 3 biBe : 1 bibe. Since either factor B; or B, caused a development of black in the glume, the first three classes were alike black. Subsequent a 54 GENETIC STUDIES ON A CAVY SPECIES CROSS. self-fertilization proved that the F, individuals were of the formula demanded by such an explanation, viz: { 1 ae iibcabhae ea Aate bred true. 2 11D 2D2.... see ec ete ee oe 9 BiB | 2 B,B,Byb2 Pe ee ee a " 4 BibiBebo.... 6... ee eee gave 15 black, 1 white. 1 BiBibsbe.......2- 6-2 ee bred true. 8 Bibs { } Eibitses ds oe gave 3 black, 1 white. 3b,B J 0 bibiBeBo: ec ess cece eens bred true. : PED bibiBabi was dore aeey see gave 3 black, 1 white. lL bib: 1 bybibsbe..........2.05- bred true. B, was not allelomorphic to By, but each was allelomorphic to its own absence; both B,; and B, caused development of black in the glume independently. Carrying out similar work on other characters, Nillson-Ehle found that the presence of red in the pericarp, presence of brown in the ears, presence of ligules, internodal length, rust resistance, and the like were due to more ‘‘present mutually independent, separable factors than might be concluded from external appearances.’’ In any such case involving n allelomorphic pairs, the ultimate recessive would appear in 1 out of 4° individuals. In a trihybrid or tetrahybrid cross, the ratios would be 63:1 and 255 : 1 respectively—subject, of course, to the law of error. It is true that the dominant classes may often show whether they contain a smaller or larger quota of the dominant factors. Environmental conditions may also prevent the complete somatic development of the characters which a plant may transmit to its progeny. Emerson (1910) gave a short, concise interpretation, in Mendelian terms, of the inheritance of shape and size in three species of plants. His data (on size and shape of the fruits in gourds and summer squashes, size and shape of bean seeds, and size of seeds and height of the stalk in corn) show a blend in the F, generation and a marked increase of variability in the F, generation over the parents or F; generation. The difference between the F, and F, plants is great enough to leave no doubt that this increased variability has been delayed until the second generation after the cross. Shull (1910) reports a similar increase in the F, generation in the variability of the number of rows per ear in corn. East and Hayes (1911) likewise demonstrated that yellow in the endosperm of maize may be due to two factors, Y; and Yo, each allelo- morphic to its own absence. Hence, they obtained in a cross between a homozygous yellow race (Y1Y:Y2Y2) with a white race (y1y1yey2) a ratio of 15 yellows:1 white. In crossing types of maize, differential characters in the number of rows per ear, length of plant, length of ear, and weight of seed were studied. By crossing the dominant with the recessive type of each character, an increased coefficient of variability GROWTH AND MORPHOLOGICAL CHARACTERS, 55 was obtained in the F, generation. This they held due to a rearrange- ment of a number of separable factors for the character involved. Tammes (1911) has likewise thought it possible to ascertain a number of separate, independent factors for characters in species and varieties of flax (Linum). She has calculated the approximate number of factors for each character, such as length and breadth of the seeds, length and breadth of the petals, color of the flowers, and dehiscence of the capsules. The proportionate number of individuals in the F, generation, which show the pure parental character, was taken as an index of the number of factors for that character. Phillips (1912) has recently crossed two races of ducks, differing in size, and obtained an increase in variability in the F, generation. MacDowell (unpublished) had similar experience with rabbits. An increase in variability in the F, generation can not in itself be considered a final criterion of Mendelizing inheritance, for the F, individuals should be tested in order to show that all do not regress to the mean, but some pure recombinations have been formed. Very little has been done with F; generations in such crosses. East and Emerson (1913) have continued their researches in maize on the inheritance of number of rows per ear, length of ears, diameter of ears, weight of seeds, breadth of seeds, height of plants, number of nodes per stalk, internodal length, number of stalks per plant, total length of stalks per plant, and duration of growth, and have given evidence that the F, generation is in general more variable than the F, or either parent. Furthermore, the F; generations indicated that the parental types recovered in the F, might breed true, that inter- mediate types with new modes had been obtained, and that some F; individuals gave F; progeny just as variable as the F;. They conclude “that the results secured in the experiments with maize were what might well be expected if quantitative differences were due to numer- ous factors inherited in a strictly Mendelian manner.” The striking similarity between these crosses and some of the well- known color crosses makes it seem probable that both forms of inherit- ance may be Mendelian; for in both the segregation is delayed until the F, generation. Nevertheless, the clearness shown in color-inheritance does not stand out in size-inheritance. Interaction of many factors and environmental effects may play a greater part. Whether or not the general size of mammals will lend itself to such a solution is difficult to say. There is much correlation in the size of parts, although we do find that partially uncorrelated individual parts, such as short legs, tails, or ears, may exist in mammals. It would be theoretically and practically desirable to know whether the inheritance of the general body size is Mendelian when mammals of the same proportions but of different size are crossed. 56 GENETIC STUDIES ON A CAVY SPECIES CROSS. The two parent species, C. porcellus and C. rufescens, and their hybrids of various blood dilutions, which formed the material for Part I, are also used as the basis for Part II. Each parent species is of very distinct and specific size, such that environment does not obliterate the difference. Unfortunately, the cross involved sterility and necessitated crossing back to males of the parent species. A careful examination of growth curves and skeletal dimensions was made to study size- inheritance in such a mammalian species cross and to compare it with the work already cited. 13. GROWTH. THE DATA. Cavia rufescens is a smaller species than Cavia porcellus. The average healthy adult weighed about 425 grams; the females were a trifle lighter, or about 420 grams. One male (1) alone reached the 500-gram mark in any of his weights; but he was fat, and his weight was above normal. His son, though slightly larger in skeletal dimen- sions and in good condition, was about the average weight. The average weight of guinea-pigs is twice that of the wild; and since the average is so much larger, it follows that many guinea-pigs are more than twice as heavy. I have never seen a guinea-pig of either sex, with a normal healthy growth curve, maintain such a low weight as the wild. This statement is made on the basis of an intimate acquaint- ance with the growth curves of several hundred guinea-pigs. In order to study the growth curves of the parent species and hybrids, the weights were taken about once a week until the curve was well established. After that, observations were made at less frequent intervals. The weights of pregnant females were taken during the period of gestation, but not used on account of the varying number of fetuses. Having obtained the weights, the growth curve of each animal was plotted on coordinate charts by placing the days on the abscissas and the grams on the ordinates. Any individual curve would naturally show a depression when external conditions were poor and an elevation when conditions were conducive to fatness. Since no growth curve is in itself an infallible expression of the general growth tendencies of an animal, a second set of curves was drawn, in which the irregularities were smoothed to show as nearly as possible the normal growth of each individual. This may seem arbitrary; but in reality it does not signify any bias, for in all cases the smoothed curve was determined by the majority of points in the actual curve. Minot (1891) has shown ‘‘that any irregu- larity in the growth of an individual tends to be followed by an opposite, compensating irregularity; and that variability decreases with age.” To be concrete, all animals in these experiments showed a decrease GROWTH AND MORPHOLOGICAL CHARACTERS. 57 in weight about the end of April, when the quality of beets and turnips was poor and the supply was low; but toward the middle of May they recovered completely through copious feedings with fresh green grass and winter rye. The irregularities in growth were caused by external conditions, and observations of these conditions were made. It requires no stretch of the imagination or undue speculation to smooth such irregularities and thus procure a curve which more truly shows the general growth tendency in an animal. Smoothed curves were made for parent and hybrid individuals as follows: OeTUPesCens ss wees cae crams yas maw oes 9 P wild wesc suw es vaiicea ides oh Meee ss 15 BWA sso. diced dapat ig docedadallecdns aubeaiet ase abe 37 Re Wich iiss sxavona\osavovene ca docs, avenans, 2:58 seo aioe 138 C. porcellus (small inbred strain)...... 45 (mormal strain)........... 109 Total jie skevosiven deasesass sas 353 All sick animals and those whose curves were not well established because of early death or present immaturity, were neglected. Diseased animals show an irregular curve with a large final loss in weight and were therefore neglected. Many other guinea-pigs and hybrids have been studied, and can be added, when their growth is sufficiently com- plete. It is quite significant that a duplicate set of smoothed curves was made for about 75 animals. This set did not vary much from the first set. We thus have a check on errors in judgment, for the duplicate set was made a number of months after the first set and with- out any reference to the same. I am, therefore, led to believe that the average of the smoothed curves is correct within +25 grams. Having once obtained the smoothed curves, composite curves for the males and females of different classes were calculated. The method is simply to find the arithmetic average of the weights in the smoothed curves at 15 different intervals (tables 59 and 60). For example, if we average the weights of all } wild hybrid males at the age of 100 days, as given in their smoothed curves, we obtain an average of 555 grams. This gives one point from which to plot the average of the curves of animals in that group. The other points were similarly calculated, and a composite curve or average of the smoothed curves was plotted. The composite curves of the wild and tame species and three classes of hybrids are shown in text-figures 1 and 2. Needless to say, all animals were kept in healthful quarters, with an abundant supply of food and water. The food at all times con- sisted of oats. In the winter this was supplemented by daily feedings of beets or turnips; in the summer, by fresh green grass and clover. 58 GENETIC STUDIES ON A CAVY SPECIES CROSS. COMPARISON OF GROWTH CURVES. Tue AVERAGES. Minot (1891) has shown, in the case of the guinea-pig, that growth is rapid at first; and as the animal grows older a smaller daily incre- ment is added. Stating it differently—as an animal grows older it requires a constantly increasing span of time to add successive, equal increments of weight, until finally growth ceases. This means that the growth curve is steep at first, and that the early growth is the greatest. Gradually the curve approaches a straight line, the adult weight. The composite curves for both sexes (text-figures 1 and 2) show this in the wild, the tame, and the hybrids. At the end of a year practically all of the animals were full-sized adults; but in nearly all cases an extra 3 months was given to each animal to follow a full compensation for any possible early retard. At the end of a number of curves a slight unexpected increase appears. This is due to the fat condition of a number of the older animals, as the individual records show. One can easily follow any curve to its logical conclusion. From an examination of tables 59 and 60 and their graphic repre- sentation in text-figures 1 and 2, a number of salient facts, concerning the average weights of the parent species and the hybrids, may be recorded: (1) The average weights, and consequently the composite growth curves of the wild, are well below the tame guinea-pig at all ages and in both sexes. These do not show, however, that this is not completely true for all individual weights of each species. For example, the indi- vidual records reveal that some male guinea-pigs at the age of 10 days were lighter than 95 grams. Although there was some overlapping of the early individual weights of the wild and tame, as time progressed the wild showed their specific character, and it required only a few weeks before all the wild were well below all the tame. Weights were obtained for more than 4 wild males and 5 wild females. Originally, composite curves were made including these. They only served to augment the difference between the wild and the tame, for they were sickly, did not thrive well in captivity, and died prematurely. Those animals which entered into the tables and composite curves represented in a fair way the natural growth of the wild Cavia rufescens. (2) The 4 wild hybrids of both sexes were remarkably vigorous animals. The males attained an average which exceeded their larger parent, the guinea-pig. They were also larger than all succeeding hybrids. The females were likewise very vigorous. Curiously enough, the middle portion of the composite curve of the females is below the guinea-pig and the 34 wild. But, if an anticipation is permitted, it will be shown that the bones of the $ wild hybrids are larger than those GROWTH AND MORPHOLOGICAL CHARACTERS. 59 Grams Days TEXT-FIGURE 1.—Composite growth curves of the males in the parent species and hybrids. Trxt-FiGuRE 2.—Composite growth curves of the females in the parent species and hybrids. 60 GENETIC STUDIES ON A CAVY SPECIES CROSS. of the guinea-pig or of other classes of hybrids. The depression in the composite curve of the 3 wild females from the 120th day to the 340th day was due largely to our eager haste to breed these unusual hybrids as soon and as frequently as possible. Furthermore, I should not consider the composite curve as trustworthy as the skeletal dimen- sions; because adult weights are more variable than adult skeletal dimensions; and because possible errors in judgment arise, especially when one subconsciously tries to avoid a bias in favor of ‘‘too much heterozygosis”’ in smoothing the individual growth curves from which the composite curves were calculated. The species cross between the horse and ass gives the well-known vigor for which the mule is so highly valued. Darwin (1876) pointed out that cross-bred plants were often more vigorous than the inbred parents. East and Hayes (1912) have concluded that the vigor is in a measure proportional to the number of factors in a heterozygous condition. Our 4 wild hybrids were undoubtedly heterozygous in many factors, but we can not be sure that the more vigorous were heterozygous in a greater number of factors. What part sterility may play is also unknown. (3) The 4 wild of both sexes clearly lacked the vigor which charac- terized the 4 wild. The composite curves of the males and females lie entirely below those of the } wild. The greater part of both also lies below the guinea-pig and the 3 wild. Although these + wild were produced by mating the vigorous } wild back to the larger of the original two parent species, it is obvious that both the males and females were smaller at all ages than the 4 wild, and also smaller than the guinea-pig during the larger part of their growth curve. If we regard the sexes separately, it will be seen that the 4 wild males averaged less than the guinea-pig throughout the greater part of their growth curve, for they lie distinctly below these up to the age of 360 days. Their curves take an unexpected rise at the age of 340 days, but from personal experience with these animals I am led to believe that this was due to the obesity of anumber of older males which were kept alone to prevent fighting. The difference between the + wild males and the 3 wild is quite apparent, for they are separated by an average of about 150 grams during a large part of their growth. It is difficult to ascertain how much significance to attach to the aver- age difference between the } wild males and their smaller parent, the guinea-pig. They were consistently lower at all ages than the smaller race of guinea-pig males until 360 days, although the difference was not great. The + wild females resembled their brothers in many respects. They likewise lie below the guinea-pig during the greater part of the growth curve, for they were smaller up to the age of 260 days. Their growth GROWTH AND MORPHOLOGICAL CHARACTERS. 61 curve rises above the smaller race of guinea-pigs at this date, and this is not due to an abrupt change in their curve, as was the case with their brothers. Like their brothers, they averaged less than their 3 wild parent at all ages and the difference is also well defined. Summarizing, we may say that the 1 wild of both sexes lacked the vigor of the } wild, although the 3 wild females were used as one parent. The ¢ wild males were in general smaller than the guinea-pig parent; but the } wild females did not agree perfectly with their brothers, for they did not average less than the guinea-pig as constantly. (4) The § wild showed a complete return to the parental guinea-pig average and any possible indication of the loss of vigor shown by the ¢ Wild parent was absent. The } wild males have a composite growth curve which is actually higher than the larger guinea-pig race after the 140th day. The 3 wild females agree closely with the larger guinea- pig race. It is possible that the composite curve of this hybrid class of males is higher than it should be, for on account of sterility they were unmated and often kept alone to prevent fighting. On the whole, we may consider the § wild of both sexes the equal of the larger race of guinea-pigs. The 4 wild males averaged larger than the 4 wild males throughout their whole life. Their sisters were larger than the 4+ wild females up to the age of 340 days, or, in other words, until about that time when the adult size was reached. The 4 wild, however, did not equal the vigor of the } wild. The data on skeletal dimensions will corroborate all these facts in a general way. (5) Two composite curves are given for each sex in the case of the guinea-pig. One curve represents the average growth curve of a healthy, vigorous strain of guinea-pigs. The other curve is taken from the records of a closely inbred strain which was not so vigorous; hence the latter lies below the former at all points. The stock used as the guinea-pig parent in these experiments corresponded closely to the larger strain. The difference between the two curves shows the possibilities with the species C’. porcellus itself. (6) The average weights of the females, and hence their composite growth curves, were below those of the males at all ages. This was true of both parent species and the three classes of hybrids given. It was equally true of the other classes of hybrids subsequently produced. Summarizing the general results obtained as shown by all the dif- ferent averages of weights, it was obvious that (1) the } wild were more vigorous than either parent species; (2) the ¢ wild lacked this vigor; (3) the + wild regained the size of the original larger parent species by the continued crossing back to this species. These facts will be consid- ered later in connection with the discussion of the averages of the skeletal dimensions. 62 GENETIC STUDIES ON A CAVY SPECIES CROSS. Tue COEFFICIENTS OF VARIABILITY. It is indicated on pages 48-55 that a number of recent papers on size-inheritance postulated multiple factors for size with incomplete dominance. According to this theory, a cross between a pure large race and a pure small one would result in a blend, in the absence of disturbing influences such as the vigor of heterozygosis, environment, and the like. If the F, generation were then crossed inter se, one should obtain an increased coefficient of variability and, with sufficient numbers, recover the parental forms. If, however, the F, generation were crossed back to either parent, one should obtain a range from the F, to that parent with the mode in between. The usual method of procedure would be to mate the Fi generation inter se in order to obtain a maximum coefficient of variability as the best evidence of segregation and recombination of size factors. But this was impossible in these crosses, for the males were entirely sterile. Two alternatives remained, either to cross the F; females back to the guinea-pig or to the small wild C. rufescens. The latter would have been preferable, but not enough cases were successful to give data of value, hence all results were based on crossing the F; females back to the guinea-pig. The F, males were likewise sterile and consequently the F. females had to be crossed back to the guinea-pig. This meant that conditions made it necessary to resort to the class of matings least advantageous for a study of size-characters. The study of the average weights at different ages is quite insufficient to show the complete relation between the size of parents and hybrids, for they do not indicate the dispersion of the individuals from the average of the group considered; or, in other words, averages do not give evidence of segregation and recombination of possible unit factors for size. Therefore, the coefficients of variability of the weights of the parents and hybrids were calculated from the individual smoothed curves for six different ages ranging from 100 days to 380 days (see tables 61 and 62). It must be stated at the outset that the data and results are very unreliable, for the numbers are small, although breeders of mammals must be content with such; and environment affects growth and weights greatly. The coefficients of variability for weights of the males and females (tables 61 and 62) give no clear, pronounced evidence that the hybrids of the second generation were more variable than those of the first or than the guinea-pig parent, and hence there is no evidence of segre- gation and recombination of factors. It is true that some classes of hybrids were very slightly more variable thar either original parent species, but it is difficult to know whether this was due to real inherent variability or to experimental error. Furthermore, such differences as do obtain are not wholly consistent with an explanation that postulates GROWTH AND MORPHOLOGICAL CHARACTERS. 63 multiple factors for size with incomplete dominance. For example, the 3 wild males were as variable as the 3 wild males. Had the parent races and the F, hybrids shown a comparatively small degree of varia- bility and the F, hybrids a decided increase in variability, then we might have concluded that there were indications of a recombination of factors for size. The results by no means disprove that the size- difference between the guinea-pig and the wild species may not be due to a difference in size factors, but the various crosses actually made failed to give evidence to that effect. One could conclude more logically that (1) the guinea-pig was dominant, or very nearly so, to the wild species in respect to size; (2) the immediate hybrids, the 3 wild, were very vigorous because of heterozygosis; and (3) therefore, repeated crossing back to the dominant form would not increase the variability. In deciding what the normal growth curve of any individual is, in order to obtain the smoothed curves and calculate the averages and coefficients of variability, errors in judgment may occur. In this particular case the number of individuals was small and experimental errors may have been large; hence no probable errors were calculated for the average weights or coefficients of variability. The adult skeletal dimensions offered material with less objections than did the growth curves. The results of both can be compared. In passing, it may be pointed out that all classes of individuals in both sexes appeared to be less variable as they grew older. 14. SKELETAL DIMENSIONS. THE DATA ON SKELETAL DIMENSIONS. It was shown that the adult weight of C. rufescens was much less than that of C. porcellus. The bones of the wild are likewise shorter and more slender than those of the tame guinea-pig. In order to make a more extended study of the size relation between the two parent species and their hybrids, measurements of bones were taken from prepared adult skeletons. The materials available were as shown in the accompanying table. Class. Male. | Fem. Wild.......... 3 1 4 wild......... 5 8 A wild......... 16 20 4 wild......... 60 65 Guinea-pig.... 78 63 Total..... 162 157 64 GENETIC STUDIES ON A CAVY SPECIES CROSS. It was found that the skeletons had completed growth at the end of 15 months. Osseous nodules and ridges, to be sure, are laid down at alaterdate; but they do not influence the measurements considered. Care was taken to see that sutures between the epiphyses and diaphyses were closed. Furthermore, the suture between the basioccipital bone and the basisphenoid bone is one of the last to fuse in mammalian skulls, and this was completely fused at the age of 15 months. The bone measurements were, therefore, taken from fully adult animals whose bones had reached their maximum size. In preparing the skeletons all individuals were boiled separately in soap and water. The flesh was brushed away and the bones were dried, properly labeled, and filed in separate boxes. Errors were thus avoided. In all cases the skull, lower jaw, scapula, right front leg, and right hind leg were saved. Whenever possible the entire skeleton of the wild and early hybrids wassaved. Unfortunately, a number of adult wild and adult 4 wild were discarded by a laboratory helper when they died. Sixteen different measurements were taken on all skeletons. In addition to these, 13 more measurements were taken in the case of the wild, the 4 wild, and the + wild. The results, given in tables 63 to 66, were calculated from these measurements. In deciding upon the different possible measurements to be used, those actually used were chosen for the following reasons: (1) They could be taken accu- rately without any slipping of the calipers; (2) they were the largest measurements, thus diminishing the effect of any experimental errors; (3) they took into account those dimensions in which the wild and tame parents differed in the most marked degree. All the dimensions were taken with sliding vernier calipers and recorded in terms of 0.1 mm. The averages, however, are given in millimeters. For example, the average skull length of 78 male guinea-pigs was 68.48 mm. The use of skeletal dimensions in a study of size-inheritance has advantages which the weights lack. In the case of the growth curves, two observers might arrive at different conclusions with regard to an adult weight; or even the same observer has slightly different views at different times. The measurements of the adult skeleton, however, were so exact that a remeasurement gave the same result at all times within + 0.2 mm. In repeating many bone measurements, it was found that the second observation tallied completely with the first in almost all cases. When a difference did occur, it was so small as to be negligible. Furthermore, the adult skeletal dimensions were far less variable than the adult weights, meaning that the environment prob- ably affects the weights more. Of course, no claim is made that the adult skeletal dimensions represent the precise genetic possibility of an animal, but under normal conditions they probably approximate it more closely than do the weights. All of these considerations made the skeletal dimensions a better basis for study than weights. GROWTH AND MORPHOLOGICAL CHARACTERS. 65 The measurements considered in the tables are as follows: Skull measurements: 1. Median sagittal length, from cranial edge of fused premaxillary bones to lambdoidal ridge of occipital bone. 2. From the same cranial edge to the ventrocranial edge of the foramen magnum. 3. Length of the zygomatic arch from the laterocaudal margin of i infraorbital foramen tothe caudal margin of the mandibular ossa. . From the laterocaudal margin of the infraorbital foramen to the exoccipital bone immediately dorsad of the jugular process. . From the premaxillary bone to the medial lachrymal sulcus. . From the premaxillary bone to the medial caudal margin of the palatine bone. . From the caudal edge of the foramen incisivum to the ventro- cranial edge of the foramen magnum. . Width immediately craniad of the external acoustic pore. . Width at caudal portion of zygomatic arch, where skull is broadest. 10. Width at cranial edge or point of the zygomatic bone. 11. Width at laterocaudal margin of infraorbital foramen. Mandibular measurements: 12. Extreme length from angular process to laterocaudal margin of incisor alveolus. 13. From concave edge between condyloid process and angular process to cranial edge of first molar alveolus. Ooo N Ho aa Humerus: 14, Length, from fossa between greater and lesser tuberosity to fossa between capitulum and trochlea. Femur: 15. Length, from trochanteric fossa to intercondyloid fossa. Tibia: 16. Length, from fossa between spine of tibia and lateral tuberosity to lateral concavity at distal end. COMPARISON OF SKELETAL DIMENSIONS. Tue AVERAGE DIMENSIONS. The wild C. rufescens has long been known and recorded by taxono- mists as a small cavy species, smaller than the guinea-pig, C. porcellus. Hence, the average skeletal dimensions given in tables 63 and 64 were not taken from individual, small specimens that may have been wide variates. Other wild skeletons were examined and measured, but were omitted for the sake of accuracy in these averages because a few sutures were not closed, although they were sexually mature. They were in reality smaller than the average recorded. A number of fully adult living specimens were carefully examined both in our own laboratory and in European collections and were clearly much smaller than the guinea-pig. The wild, which enter into the averages in tables 63 and 64, were the two original wild parents used to propagate the wild stock in captivity, and all of their sons (24 and 33) who, with their 66 GENETIC STUDIES ON A CAVY SPECIES CROSS. father (1) were used as the wild parent in the crosses that produced the hybrids. They were fully adult, healthy animals, and in all prob- ability as large or larger than most members of their species. The tables indicate that C. rufescens is smaller than the tame parent species in all measurements considered. This was also found to be true of the scapula, radius, ulna, innominate bone, fibula, and the different verte- bre. (See figs. 10, 11, 15, 16, 20, 21, 25, 26, 30, 31, and 34 to 41.) The long bones of the wild were likewise more slender than those of the tame. The average skeletal dimensions of the tame were found to be higher than those of the wild in every case in both sexes. It is appropriate to say, briefly, at this point that all the figures of the skulls and bones given in the plates are of natural size and represent as nearly as possible the averages given in tables 63 and 64. The skulls and bones shown in these plates were chosen because each one represents the average of its class. In all cases the figures are not visibly different from the computed average and any actual difference is generally much less than 1 mm. It may seem that the differences between the averages of individual measurements are too small to separate the two species distinctly; but if, for example, an average guinea-pig skull is compared with a wild skull (figs. 10, 11, 15, and 16), it will be seen readily that the total effect of all these differences in the eleven skull dimensions is enough to separate the wild from the tame distinctly. Furthermore, there is a minimum amount overlapping between individuals of the two species. Although the 2,250 individual measurements for 78 male and 63 female guinea-pigs are not presented, there were very few cases in which any guinea-pig was found to be as small in any of its dimensions as the longest wild. The exact number of guinea-pigs overlapping the wild is as shown in the accompanying table. Measure- | Male, | Fem. No. 3....| 0 3 6....4 4 0 8....) 49 5 9....1 5 1 12....) 3 0 Lbecas) % 0 Therefore, out of a total of 2,250 guinea-pig measurements, there were only 77 which overlapped the corresponding wild measurements. This means that no guinea-pig of either sex was as small as the wild in the case of 10 of the 16 dimensions. In the 6 dimensions given above, there were a few cases in which some guinea-pigs equaled or were smaller than the wild, but when they were smaller it rarely GROWTH AND MORPHOLOGICAL CHARACTERS. 67 amounted to more than 0.3 mm. In measurement 8, the width of the skull immediately craniad to the external acoustic pore, the males of both species were more nearly equal, and 49 out of 78 male guinea- pigs were actually as small as or smaller than the largest wild. The reason the wild are so large in this measurement is due to the large bulla, possibly associated with the organs of hearing. Many other guinea-pig skeletons were examined at a later date, but none could be mistaken for the wild species. The wild C. rufescens in these experiments were, therefore, distinctly smaller than the tame C. porcellus. The skeletal dimensions corrobo- rate the data presented in the composite growth curves. The number of wild in tables 63 and 64 is too small to give significant averages; but the known facts regarding C. rufescens and our own observations on immature animals indicate clearly that it is specifically smaller than the tame species. Furthermore, since the number of tame is large enough to be significant, it is noteworthy that their lower extremes rarely overlapped with the measurements of our largest, healthy, adult wild animals. The one-half wild hybrids, obtained by crossing the wild males to guinea-pig females, were larger and more vigorous than either parent species. The males averaged larger in all measurements taken, and the females averaged larger in all but two (see figs. 12, 17, 22, 27, 32, and 34to41). In these latter two exceptional cases (measurements 10 and 13) the females were really as large as the guinea-pig, for the difference was hardly significant, considering the probable errors. This increased size and vigor was not only true of the 4} wild as a whole, but every individual male and female was larger than the average guinea-pig in all its measurements, except two 3 wild hybrids. These two exceptions (#117 and 9118), a brother and sister, were fully as large as the average guinea-pig. The individual measurements and the averages of the progeny in this first cross thus attested the remark- able vigor of the + wild hybrids. The skeletal dimensions, therefore, corroborate the data presented in the composite growth curves. This was not only true of size but also of endurance; for, although they were very wild in disposition and difficult to keep in captivity, when successfully reared they showed their physical strength. They lived through winters when ordinary guinea-pigs succumbed to disease. One female had 15 litters of young and is still breeding at the age of 7 years. Alezais (1903), quoting Metschnikoff, states that this age would be remarkable for a guinea-pig. None of the several thousand guinea-pigs in this laboratory have ever been as long-lived; neverthe- less, it must be stated that there has been no close study of their longevity. Other } wild females were equally vigorous and fertile, but were killed for the purpose of study. 68 GENETIC STUDIES ON A CAVY SPECIES CROSS. We can not dispatch the whole situation by a simple statement that the guinea-pig is dominant in size. Possibly it is somewhat so, but we do not know how much of this vigor and size was due to heterozy- gosis. Furthermore, since the female was the large parent, it may be that the reciprocal cross, with C. rufescens as the female parent, would not have given the hybrids such a good start. It is conceivable that two fetuses in the guinea-pig uterus would have a greater chance for initial development than the same two in the uterus of C. rufescens. That the guinea-pig is in all probability not completely dominant one can conclude from the size of the next generation. The one-quarter wild hybrids were produced by mating the } wild females back to guinea-pig males (see figs. 18, 18, 28, 28, 33, and 34 to 41). They showed a striking loss of the vigor which characterized the 3 wild, for both sexes averaged smaller than these in all dimensions, except measurement 3 in table 64. The single exception was the length of the zygomatic arch in the female sex, in which dimension the 3 wild and } wild females averaged exactly the same. The } wild not only averaged less than the 4 wild, but no one of the 36 individuals was as large in any measurement as the largest } wild, and very few were as large as the smallest 4 wild. Comparing the average of the + wild males with their male parent, the guinea-pig, it was found that there was a general tendency for the hybrids to be smaller, in which respect the growth curves and skeletal dimensions again agree. The averages of the male 4 wild were less in all measurements except 8 and 9. The female } wild averaged smaller in all measurements except 3, 5, 6, 8, and 9. Although the growth curves and skeletal dimensions of the + wild were in general consistently lower than those of the guinea-pig. the differences were not great. What seems to be a general tendency must be cautiously considered, in view of the small differences, which were often not much larger than the probable error of the averages. The one-eighth wild hybrids, or F3 generation, were produced by mating the { wild females back to the guinea-pig males (see figs. 14, 19, 24, 29, and 34 to 41). The males of this generation were larger than the 4 wild in 14 of the 16 dimensions; and the females were larger in 7 dimensions, and exactly equal in 3. Comparing the + wild males with the guinea-pig, it was found that they were slightly larger in 13 of the 16 averages, whereas the females were slightly smaller in 15 of the 16. Here again, the differences must be cautiously interpreted, for they were small in comparison with the probable errors and especially in comparison with four times the probable error. The differences between the ; wild and the guinea-pig were extremely small, and not apparent to the naked eye, as the figures of average dimensions show. Irrespective of whether or not we consider the + wild smaller than the guinea-pig, it is quite certain that two back-crosses made the § wild the equal of the guinea-pig in size. GROWTH AND MORPHOLOGICAL CHARACTERS. 69 Summing up the data bearing on average skeletal dimensions in C. rufescens, C'. porcellus, and three generations of hybrids, we may say that: (1) C. rufescens is smaller than C. porcellus. (2) The 3 wild hybrids were larger and more vigorous than either parent species. (8) The 4 wild were smaller than the 4 wild and possibly showed a general tendency to be smaller than the guinea-pig, particularly in the male sex. (4) The 4 wild and the guinea-pig were of the same size and practi- cally indistinguishable. CoEFFICIENTS OF VARIABILITY OF DIMENSIONS. C. rufescens is specifically smaller than C. porcellus. We do not know whether the smaller species lacks factors for size, or whether it has factors inhibiting growth, or whether there are any ‘‘factors” involved at all. If we suppose that the difference in size is due to one or many completely dominant factors, then the F; should be like the dominant parent; and crossing the F, and F, generations back to this parent should give only the dominant form. But if we suppose the difference to be due to multiple, incompletely dominant factors, then the F, generation should be a blend, and the F, should show an increased variability, as was shown on pages 50-51. It has been pointed out by East (1910) that ‘‘as dominance becomes less and less complete, the Mendelian classes vary more and more from the formula (3+1)" and approach the normal curve, with a regular gradation of individuals on each side of the mode.” In order to ascertain whether the hybrids were more variable than the parents, the coefficients of variability were calculated (see tables 67 and 68). The variability of C. rufescens is unknown. The number of adult skeletons available in our own experiments was far too small to use as data. If we analogize with the tame C. porcellus, it is probable that the wild is not very variable. The coefficients of variability of the guinea-pig were extremely small. The highest coefficient of any dimension in either sex was only 3.73 per cent +0.20 (measurement 9, table 67). Only 6 of the 32 coefficients were 3 per cent or more. Furthermore, they were very uniform, for the lowest was exactly 2.00 per cent +0.12; and they range, therefore, from 2 per cent to 3.73 per cent. Compared with the parent stock used in experiments on maize (Shull 1910, East and Hayes 1911), or with the stock used in experiments on gourds and beans (Emerson 1910), these coefficients in the guinea-pig are very small. In the case of maize, the coefficients of variability of the parents were sometimes as large as 14 per cent. Emerson gave a coefficient of variability as 26.9 per cent for the shape of one parent (scallop) in summer squashes. This in no way reflects on the results and interpretations of these 70 GENETIC STUDIES ON A CAVY SPECIES CROSS. investigators; but the comparison is interesting and shows how uni- form the skeletal dimensions of adult guinea-pigs really are. It is probable that the wild cavy species is likewise very uniform. The coefficients of variability of the one-half wild hybrids were calcu- lated from such small numbers (5 males and 8 females) that their value is doubtful. Such coefficients are most valuable and accurate when the number of variates is large. When the total number of variates is small, a few wide deviates greatly increase the standard deviations, and therefore increase the coefficients of variability also. According to the theoretical scheme involving multiple, independent size factors, incompletely dominant, the F, generation should be a blend and no more variable than the parents, if the parents were practically pure. As a matter of fact, the 4 wild were larger than either parent. We say that such phenomena accompany the heterozygous condition, but we can not adequately explain it. Taking the coefficients as they stand, the variability of the 3 wild females was slightly greater than that of the guinea-pig parent, but the male hybrids were on the whole no more variable than their parent. In view of the fact that the chances of error are great, no conclusions can be drawn. The one-quarter wild hybrids, or F, generation, showed no great increase in variability, as one would expect on the hypothesis of many interchangeable factors without dominance. The males were no more variable than the guinea-pig, and the females were only slightly so. Here, again, the numbers were small (16 males and 20 females) and the results are subject to a serious objection. The one-eighth wild hybrids, or F; generation, were on the whole only slightly more variable in both sexes than the guinea-pig. It can be readily seen that all the coefficients of variability are small and form no series consistent with the hypothesis advanced, according to which the F; generation should be no more variable than the parents, but the F, generation should show an increased variability, while the F; should be less variable than the F, generation. The whole 128 coefficients in tables 67 and 68 are very small and close together. Moreover, if one considers the probable errors, the chances are small that the differences in variability are not due to random sampling. Practically every coefficient in any particular dimension would over- lap every other one in that dimension if the probable error is multiplied by four. Therefore, from the standpoint of pure random sampling, the chances are large that a repetition of these experiments, under similar conditions and involving the same numbers, might easily give results with no significant differences between the coefficients of varia- bility. It must be stated that probable errors for the + and 3 wild are very unreliable, since the numbers are so small. Examining the data as they stand, to ascertain which dimensions .show a series of coefficients most variable in the + wild and grow less GROWTH AND MORPHOLOGICAL CHARACTERS. 71 variable as they approach the guinea-pig, we find such to be the case for the males in measurements 3, 6, and 10, and for the females in measurements 3, 5, and 8 to 16. (The 3 wild are not considered on account of the small numbers.) Now, if we had by chance chosen only measurements 3 and 10 as the basis for our comparisons, then we would have been led to the conclusion that there was consistent evidence of segregation and a recombination of size factors in both sexes. But had we chosen other measurements we might have arrived at different conclusions. The question naturally arises, are the series of coefficients in any one dimension more significant than those in any other? Are we justified in selecting particular series which conform to the results presented by other investigators, and thus indicate a recombination of factors? As far as we can tell, we are not; for at present we know of no reason why special emphasis should be attached to the results obtained in certain measurements in preference to others. There is another method of approach by which it is possible to avoid attaching questionable weight to a few dimensions. We may average all the coefficients of variability in each of the different classes to see, for example, if the + wild were on the whole more variable than the guinea-pig. Table 69 gives the averages of the different coefficients of variability in the guinea-pig and hybrids, the purpose being to ascertain what the general tendencies of any class might be and to see whether on the whole the hybrids showed a general tendency to greater average variability than the parent guinea-pig. We also wished to see if, on the whole, the 4 wild were more variable than the } wild and the guinea-pig. However, the male } wild averaged no more variable than the guinea-pig; but the female } wild were more variable. All the different classes of males were of equal average variability except the 4 wild. All the female classes were statistically of equal average variability except the female guinea-pigs. The males do not show a series indicating that the } wild average most variable and that this variability decreases as we approach the guinea-pig; but the females do. In other words, there is little, if indeed any, evidence of segregation and recombination of factors for size in these crosses. It is interesting to note that the F; hybrids (4 wild) of both sexes averaged more variable than the guinea-pig. These expressions of average variability were based upon 16 different coefficients of varia- bility. Back of each coefficient of variability there were from 60 to 78 variates. If one interprets the data from a purely statistical point of view, then the 4 wild hybrids were inherently more variable than the parent guinea-pig and the chances are enormous that this difference is not due to random sampling. However, in interpreting biological data, other considerations are of importance. It is shown that all the coefficients of variability in the 4 wild and the guinea-pig are extremely small. The averages of the guinea-pigs and 3 wild, in tables 63 and 64, 42 GENETIC STUDIES ON A CAVY SPECIES CROSS. were shown to be practically the same; and, hence, a difference of less than 1 mm. in the standard deviation of any measurement would com- pletely obliterate the differences in the coefficients of variability. Although I have undertaken no experiments to ascertain the effect of environment on skeletal dimensions, experience with many hundreds of guinea-pigs and hybrids leads me to believe it would be decidedly strange if environment could not effect a difference of less than 1 mm. in the standard deviation of the guinea-pigs and hybrids. Summarizing the facts concerning variability in the guinea-pigs and hybrids, we may say that— (1) The variability of all the classes of hybrids and the guinea-pig was very small. (2) There were no great differences in variability in the back crosses of hybrids to guinea-pigs which would indicate segregation and recom- bination of factors for size. This is true for the individual measure- ments and for the general average variability of each class. (3) The results in no way controvert the possibility that size may be due to factors which are inherited in Mendelian fashion; but segrega- tion was not apparent in these classes of matings in this species cross. The dominance of the guinea-pig may well be very nearly complete. Since the hybrids were mated back to the guinea-pig each time, it is simply a case of dominance with little or no evidence of segregation. According to this explanation, the vigorous growth of the first, or 4 wild, hybrids was due to their heterozygosity, but without the effect of heterozygosis they would have been a little smaller than the 4 wild. Mating the $ wild to the guinea-pig raised the mean of the } wild nearly to that of the guinea-pig and a second back-cross raised the mean of 4 wild right up to the guinea-pig. If the guinea-pig is dominant, or almost so, one would expect little or no evidence of segregation. (4) It would be interesting to know whether the small C. rufescens was derived from a larger species such as C. aperea, C. cutleri, or C. porcellus by the loss of size factors, or whether the larger species arose by progressive variations from this small wild species. 15. THE SKULL SUTURES. Among other characters which differentiate the wild C. rufescens from the guinea-pig, the nasal-frontal suture and frontal-parietal suture appear to be prominent. In the wild, the suture between the nasal and premaxillary bones and the frontal bones forms an M. The caudal margin of the nasal bones forms a V, and with the premaxil- laries the whole suture is more or less M-shaped. In the tame, this suture is approximately truncate. The suture between the frontal and parietal bones in the wild is practically a straight line; but in the tame this same suture dips distinctly backward (see figs. 10, 11, 15, 16, and 31). GROWTH AND MORPHOLOGICAL CHARACTERS. 73 No satisfactory measure of the sutures could be found and, therefore, camera-lucida tracings were made of the nasal-frontal suture in all available skulls. The original data are pre- sented directly in figures 42 to 47. Draw- ings were made as shown in the table here- C. rufescens........ 6 with. C. porcellus........ 53 < : ‘ ‘ 3 wild hybrids.... 13 Fifty-three camera lucida drawings of this 1 eid hybridisiec a 24 suture in the guinea-pig are given in figure 4 wild hybrids.... 133 42. Several hundred skulls were examined, dy wild hybrids.... 189 but no cases were found which could be con- Total......... 438 fused with the wild (fig. 48). There is a range of variability in the tame; but in general the suture may be described as forming nearly a transverse line. Only 6 C. rufescens sutures are shown. We do not know whether the wild is very variable or not. Nor do we know that the wild males used in the crosses were pure for such a character. When the wild males were mated with tame females, the 4 wild (fig. 44) showed the effect of the wild parent. None of the 13 4 wild were truncate, but all were mM- shaped. The 4 wild females were mated to guinea-pig males. Their 4 wild offspring were very variable. Forty-four of these showed a range of forms from those like the 4 wild to forms just like the tame (see fig. 45). It may mean that there was a rearrangement of factors, and the tame form segregated out in this F, generation, as one might expect on the basis of several incompletely dominant factors. The + wild females were mated with guinea-pigs to produce the 4 wild, and these in turn were mated to guinea-pigs to produce the sty wild. The 3 wild (fig. 46) and +’, wild (fig. 47) presented a wide range of forms. This was to be expected, for the hybrid females used as dams were of many very different types. No series of guinea-pigs, to my knowledge, ever showed such a range as these hybrids. If the wild form is regarded as dominant, then the perfectly truncate forms which segregated out in the F, (or + wild) might be expected to breed true when mated back to the recessive guinea-pig. This was not found to be the case; for some of these female hybrids with per- fectly truncate sutures had offspring showing M-shaped sutures. In other words, those F, individuals which appeared to be recessive often gave M-shaped sutures in the F; generation. It is difficult to say whether or not this was due to the interaction of complementary factors. 'The number of offspring from each F, female was necessarily small. Some bred true to the recessive truncate form, others did not. The frontal-parietal suture of the wild was also apparently dominant in the F,. The F, generation was variable, giving some segregates like the tame (see figs. 12, 13, 14, 17, 18, 19, 32, and 33). 74 GENETIC STUDIES ON A CAVY SPECIES CROSS. 16. MISCELLANEOUS MORPHOLOGICAL CHARACTERS. THE INTERPARIETAL BONE. An interparietal bone occurs in young guinea-pigs, but after a few weeks it generally becomes fused with the parietals and can not be detected. We do not know whether it ever occurs in the adult wild. Table 70 shows its occurrence in the wild, tame, and hybrid guinea-pigs which were available for study. Figures 18, 18, 19, and 33 show its form, usually a very distinct triangular bone. Its occurrence in guinea- pigs is infrequent. It occurred in 9 out of 141 guinea-pigs, or 6.4 per cent. None of these guinea-pigs were used in matings with the wild or hybrids. Among the 3 wild it was found in two cases, or 15.4 per cent. These two cases were a brother and sister, but none of the subsequent hybrids showing an interparietal bone were descendants of these two. The interparietal was present in 15 out of 46 3 wild hybrids, or 32.6 per cent. Eight of the 9 } wild females showing it were mated with the guinea-pig, and 5 of them had some offspring which also showed it. But other + wild females had offspring which showed the same anomaly. In other words, some of the 23 4 wild hybrids showing an interparietal bone were descended from females which had it, while others were descended from females showing absolutely no trace of it. The inter- parietal seemed to be most frequent (32.6 per cent) in the } wild, and when these were mated to guinea-pigs the + wild showed it in 18.4 per cent. One would expect it to decrease in frequency, for continually mating back to the guinea-pig should eventually establish the zygotic constitution of guinea-pigs in most dilute hybrids, and thus reduce the frequency of an interparietal bone to that of a race of guinea-pigs. THE SHAPE OF THE SKULLS. The skull of the wild C. rufescens is specifically much more pointed than that of C. porcellus (see figs. 10, 11, 15, 16, and 31). In crossing these two species, the Fi, or $ wild, was an apparent blend (see figs. 12, 17, and 32). Crossing the F, generation back to the guinea-pig gave some forms just like the guinea-pig, although most of them showed traces of the wild influence (see figs. 18, 18, and 33). The next back- cross, giving the } wild or F; generation, were in general similar to the guinea-pig, but possibly showed a little wider range. To ascertain the magnitude of pointedness or triangularity of a skull is difficult. If one takes the ratio of the greatest width of a skull to the width at the laterocaudal margin of the infraorbital foramen, one obtains an idea of the triangularity; but the quotients thus obtained can not be regarded as more than approximations. Measurements 9 and 11 were the widths of the skulls at these two levels. Dividing measurement 9 by measurement 11 gives an index of the triangularity; GROWTH AND MORPHOLOGICAL CHARACTERS. 15 for, the more pointed a skull is, the greater will be the quotient, pro- vided the distance between these two transverse measurements remains the same. The sagittal length of the skull between measurements 9 and 11 is in reality the altitude of a trapezoid, of which these widths are the bases. There are two ways of dealing with the pointedness of these skulls. One can take the ratio of the averages of measure- ments 9 and 11, given in tables 63 and 64, or, one can take the average of the ratios of measurement 9 to 11 in the individual skulls. Ratios of averages and the average of ratios are not necessarily the same, to be sure. The first case would mean the pointedness of an average or ideal skull in a given class, and the second case would mean the average pointedness of an array of many skulls in this class. Both sets of quotients were calculated and are given in table 71. They are practi- cally the same, and this is probably due to the high degree of corre- lation between measurements 9 and 11 in any given class. The indications are that: (1) The wild was more pointed than the tame. (2) The 3 wild were an apparent blend. (3) The ¢ wild, according to the table, were the same as the 3 wild; but as a matter of fact they were less pointed. The 4 wild skulls were very large, and since the distance between the two widths (the altitude of a trapezoid) was longer, the same ratio must mean that the 4 wild were more pointed than the + wild. (4) The 3 wild were approaching the guinea-pig-skull shape. The coefficients of variability for the ratios of measurement 9 to 11 were calculated; but like the coefficients of variability for the linear dimensions, they were small and showed no significant differences. However, there is no doubt but that individuals were obtained in the F, and F; generations which were identical with guinea-pigs. Possibly we are justified in regarding these as segregates, due to the recombina- tions of factors. EFFECT OF STERILITY IN THE MALES, Throughout the discussion the sterility of the males has been neglected. In the case of non-functioning testicles it has been shown that ossification is delayed, particularly in the long bones. Recently Geddes (1910) has shown this to be the case in pathological conditions as well as in castration. The measurements of all hybrid males in table 63 were taken from fully sterile animals (except two males). By sterile we mean that they lacked motile spermatozoa and were incapable of fertilizing an egg. In many cases they showed no sperma- tozoa at all in the epididymis. The averages and variability of these sterile + wild males are so close to the guinea-pig that it may be safely concluded there was no effect from such sterility. The number (60) of instances is large enough to make the average significant. That 76 GENETIC STUDIES ON A CAVY SPECIES CROSS. the testicles were entirely non-functional can not be maintained, for the cells of Sertoli, the interstitial cells, and spermatocytes may have been present. These may exercise some normal functions. A cyto- logical study will be undertaken later. The § wild females were fertile and also equal to the guinea-pig in size. Therefore, the 3 wild of both sexes average the same as the guinea-pig, and the peculiar sterility of the males has no effect, similar to that reported by Geddes. The sterile 1 wild males are actually smaller than the guinea-pig. The difference between these sterile males and those in Geddes’s experiments is that, in the former, the testicles were present and may have functioned in secreting hormones; whereas in the latter case they were really entirely non-functional. That the testes of sterile male hybrids were partially functional we are quite certain, for the secondary sexual characters were all present. The prostate glands and seminal vesicles were perfectly well developed. Sixteen hybrids were castrated at the age of 3 weeks, for the sake of comparison. Their seminal vesicles were greatly atrophied and they showed no sexual instinct throughout life. All these facts lead us to believe that the sterility of the male hybrids is not comparable at all to that sterility due to pathological conditions, kryptorchism, and castration. It is not surprising then, that the long bones of the sterile $ wild male hybrids and the male guinea-pigs were of equal length. ANOMALIES OCCURRING IN THE HYBRIDS. In addition to the frequent occurrence of the interparietal bone, peculiar to the hybrids, there were a number of other anomalies which should be mentioned. (1) The wild C. rufescens and the guinea-pig have 4 toes on the front feet and 3 on the hind feet. By selection, Castle (1906) was able to produce a race of guinea-pigs having 4 toes on the hind feet. There occurred among the 4 wild a male (<*202) with 5 well-developed functional toes on the left front foot and left hind foot. Like most males of this blood, he was sterile. The anomaly was never repeated. This may have been a reversion to the ancestral pentadactylous con- dition, brought about by recombining factors. It is interesting to note that the extra toes occurred on the left side, for Castle found that the extra toe in his polydactylous race was more frequent on the left side also. (2) There occurred some monstrosities in the hybrids which I have never seen in guinea-pigs, although many hundreds have been care- fully studied. In one of the hybrids the first cervical vertebra, the atlas, was completely fused with the skull. In another hybrid both scapule were bent so as to form a sharp angle, whereas normally they should be flat. In two female hybrids (? 263 and 9393) the clitoris GROWTH AND MORPHOLOGICAL CHARACTERS. 77 was greatly enlarged and possessed the two lateral horns at the distal end which characterize the penis. Their sexual propensities are dis- cussed in Part III. A female +4, wild hybrid had large caudal vertebree which, although normal in number and shape, formed a small tail about half an inch in length. In the absence of more data relating to these and other anomalies, one can only speculate as to their cause and significance. 17. GENERAL CONCLUSIONS AS TO GROWTH AND MORPHOLOGICAL CHARACTERS. (1) The wild C. rufescens used in these crosses were about half as large as the guinea-pig, C’. porcellus. ‘They were not only less in weight, but their bones were also shorter and more slender. The 3 wild hybrids were usually heavier at all ages, had larger skeletal dimensions, and gave every indication of being more vigorous than either parent species. The + wild hybrids lacked this vigor, for they were smaller than the 3 wild hybrids in every way. They were very nearly the equal of the guinea-pig in average size and skeletal dimensions. Possibly the males were a little smaller than the guinea-pig. The § wild hybrids averaged about the same as the guinea-pig in weight and skeletal dimensions. Two back-crosses were sufficient to render the F; hybrids and guinea- pigs practically indistinguishable in size and skeletal dimensions. (2) The number of adult wild available was too small to give a satisfactory index of their variability. The same was true of the 3 wild hybrids. The guinea-pigs were remarkably uniform. The variability of all hybrids in both sexes was very low and gave no clear indication of segregation. (3) The M-shaped nasal-frontal suture of the wild appeared to be dominant. Crossing back to the tame species gave a wide range of variability in the F:, F;, and Fy, generations. The truncate nasal- frontal suture of the tame species was recovered in the F, generation or 3 wild, but did not breed true. (4) The differences in skull-shape between the wild and tame were blended in the F, generation. In later generations all traces of the pointed, wild skull-shape were gradually lost. ‘The deep, narrow inden- tation on the outer:surface of the last upper molar, almost separating the small third lobe from the body of the tooth, was reduced in the F, generation; and all traces of it were lost in later generations. The taxonomists lay great stress on this character. (5) There was no apparent effect of sterility on size in the male hybrids. : (6) The unusual frequency of an interparietal bone, the occurrence of a 5-toed individual, and other anomalies were observed in the hybrids but not in the guinea-pig. PART III. THE FERTILITY OF THE PARENT SPECIES AND HYBRIDS. 18. INTRODUCTORY DISCUSSION. When the wild Brazilian male cavy, Cavia rufescens, was crossed with the tame domestic female guinea-pig, Cavia porcellus, the hybrids were fertile females and sterile males. At least three problems were immediately self-apparent: for how many generations would the hybrid females have to be crossed back to the parent males before producing fertile hybrid males; what proportion of sterile males would the more dilute wild hybrid females produce; and when fertile hybrid males were produced, would their offspring be fertile in both sexes if these males were mated with their hybrid sisters or with guinea-pig females. Sterility is a common phenomenon in the hybrids obtained by cross- ing individuals belonging to distantly related groups or types, both in animals and in plants. In fact, there is a tacit understanding among biologists that members of the same species produce fertile offspring; but a successful cross between members of different species or genera may result in sterility of the hybrids, in one or both sexes. In case both sexes in a species cross are sterile, a continuation of the genetic investigation becomes impossible. If one sex alone is sterile, then the fertile sex can be crossed back to either parent species, and it becomes possible to study the inheritance of various other characters as well as their fertility and sterility. In the experiments recorded in this paper, wild C. rufescens males were mated with the tame guinea-pig females and produced fertile female and sterile male hybrids. The fertile hybrid females were crossed back to the males of both parent species. The back-cross to the wild C’. rufescens males succeeded in so few cases (four offspring were produced) that this class of matings had to be abandoned. The back-cross to the guinea-pig males was entirely suc- cessful. The 1 wild females alone were fertile, and a second back-cross to the guinea-pig produced the $ wild. In this manner there were produced ten generations of hybrids, by repeatedly crossing female hybrids of one generation back to guinea-pigs to obtain the next more dilute wild-blooded generation. The results of these crosses have been studied with regard to coat, color, growth, size, and morphological characters and recorded in Parts I and II of this paper. The same animals were used in studies on fertility and sterility. Bateson (1913), in his review of ‘‘ Mendelian segregation and species,”’ is inclined to the view ‘‘that successful investigation of the nature even of sterility consequent on crossing, the most obscure of all genetic phenomena, may become one of the possibilities of Mendelian research.” The material presented in this part of the series of studies in a mam- 79 80 GENETIC STUDIES ON A CAVY SPECIES CROSS. malian species cross deals mainly with sterility in the male sex, conse- quent on crossing. That such complicated physiological phenomena as fertility and sterility in all kinds of crosses and under all conditions can be discussed or treated solely as problems in heredity is out of the question. Prob- ably no one would insist that fertility or degrees of fertility always depend upon “factors” or ‘‘germinal determiners.” However, it does not follow that in certain crosses factors may not be transmitted in Mendelian fashion which influence the fertility of the hybrids. Ona priori grounds we have no reason to suppose that all cases of varying fertility and sterility are due to environmental conditions; for, although environment undoubtedly influences fertility, there are unquestionable instances in which the results may be ascribed to other causes. There seems to be little doubt that environmental conditions may affect the fertility of one or both sexes, and this should be carefully con- sidered when we are dealing with the inheritance of the same. Marshal (1910) states: “‘it is well known that wild animals, when removed from their natural conditions and brought into captivity, often become partly or completely sterile.’ He cites cases from different groups of mammals and birds. Darwin (1876) also drew attention to this fact. Both of these investigators recognized that animals differ widely in this respect. The Indian elephant, chetahs, some carnivores, some rodents, monkeys, hawks, finches, parrots, and many other cases show sterility; but one can not generalize hastily and infer that all changes from a wild state to captivity result in a lowered fertility, for it is also known that certain gallinaceous birds, ostriches, pigeons, ducks, geese, and gulls, and some mammals like the skunk, ferret, mink, and Cavia aperea will breed readily in captivity. It is often asserted that wild animals in captivity are sterile because of change in diet, temperature, surroundings, lack of exercise, and the like; but none of these factors necessarily causes sterility, for one can always cite contradictory evidence. It is no easy task to differentiate between the effect of environmental factors and hereditary factors, particularly when the influence of the different factors is small and their number is large. In any comparison between the fertility of the wild C. rufescens, the domestic guinea-pig, and the various hybrids, a number of environmental factors should be given careful consideration, since it may be supposed that the wild species underwent a great change when transferred from its native habitat in Brazil to the laboratory of the Bussey Institution. All of the causes which are cited as disturbing fertility appeared to be of little or no consequence in these crosses, for it will be shown that the wild were apparently quite fertile inter se; and the wild males were surely fertile in crosses on tame females. The change from a wild habitat with the concomitant changes in diet, temperature, surround- FERTILITY OF PARENT SPECIES AND HYBRIDS. 81 ings, and the like did not prevent the wild females from breeding. The wild males, as previously stated, could only be mated to tame females with difficulty; and yet, when successful matings were secured, these tame females bore the usual average per litter characteristic of the guinea-pig. This shows that the wild males produced an abundance of spermatozoa and fertilized the usual number of eggs, exactly the same as a tame male would have done. A study of the fecundity of the wild, tame, and hybrid females will show whether or not we are justified in concluding that environment has played little or no part. No attempt is being made to underrate the effect of environment upon fertility, for it is recognized that nutri- tion, age, change of surroundings, temperature, drugs, disease, and the like may exercise profound effects. However, since the wild breed in captivity and the wild males are fertile in crosses with guinea-pigs, captivity itself may be eliminated as a factor causing sterility in the less wild hybrid sons. The original wild male (*1) lived and bred in captivity from 1903 to 1908—a period of almost 5 years. The great difficulty with these wild in captivity was not that their wildness pre- vented fertility, but that their nervous, excitable disposition made them difficult to handle and led to injuries in one way or another. Nehring experienced little or no trouble with wild C. aperea in captivity and they remained fertile at the same time. We do not know what the exact fertility of the wild C. rufescens may be in its native habitat, nor have we any basis upon which to compare its fertility in the wild state with its fertility in the laboratory pens. There are some observations by naturalists upon the fertility of C’. aperea in the wild state, but they are meager and contradictory. Nehring found that this species was more prolific in captivity than it was reported to be in the wild state. The wild C. rufescens, which were bred in captivity, aborted their young in a few cases. Abortion is, of course, not infrequent in the domestic guinea-pig, but I am inclined to believe that these abortions were more frequent in the wild cavy. The abortions may possibly be supposed to indicate a degree of disturbance in the sexual functions and signify a tendency toward sterility. If this is true it is the only evidence of any lessened fertility in the wild due to captivity. The abortions ceased in the hybrid females, and there were no other signs of any sexual disturbances in the later, more dilute wild hybrids, other than the sterile males previously mentioned. The pure wild were very easily frightened, and when disturbed would run about frantically. It is not impossible that the abortions were caused by these violent paroxysms of fear and the subsequent effects on foetal nutrition and other functions. The fertility of the other parent species, the tame guinea-pig, is well known. Under the excellent conditions of housing, food, and care in our laboratory, a sterile guinea-pig is very uncommon. Of all males 82 GENETIC STUDIES ON A CAVY SPECIES CROSS. which came under my observation, there were only two which failed to breed. When the contents of the epididymis were examined it was found that they had an abundance of live, motile spermatozoa. Their impotence may have been due to sluggishness and a failure to copulate rather than to innate sterility. Female guinea-pigs in good condition are rarely sterile. In view of the foregoing facts it would seem that the problem of sterility in the male hybrids in these crosses was fundamentally a problem of physiology and heredity, and not one of environment. The facts may be summarized as follows: (1) The wild cavy species was fertile in both sexes in captivity. (2) The tame domestic species was likewise fertile under the same conditions. (3) The hybrids resulting from a cross between these two species were not like either parent, for they were sterile males and fertile females. Nevertheless these hybrids were very vigorous, as was shown in Part II. (4) The peculiar sterility of the males persisted in later, more dilute wild generations in a manner which will be described subsequently. These later hybrids, however, could not be distinguished from the tame guinea-pig in shape, size, growth, mental traits, or any other characters, except their peculiar sterility. Therefore, since the wild were difficult to raise in captivity, but were fertile, and since their less- wild hybrid sons were easily raised in captivity but were sterile, it would appear that their sterility is not due to captivity or environment. If the facts have been correctly interpreted, some sort of consistent explanation should be found, based on heredity. The cross resulted in a definite disturbance in fertility such as did not obtain in either parent species when kept under the same conditions. Many species crosses have been made in both plants and animals. In most cases the crosses were made by those who were merely inter- ested in the sheer possibility of a cross, but not for the purpose of an extended genetic study. Much of the literature deals with the subject of sterility from a taxonomic point of view, for the fertility or sterility of the hybrids is considered a criterion of the close or distant relation- ship between the parents. From time to time compilers have given lists of species crosses with brief mention of the partial or complete dominance of one parent and the fertility of the hybrids when known. As in most other genetic studies, the botanists have led the way, and the studies of the early plant hybridists include many accounts of species crosses, or at least what were regarded as “‘species” crosses. Very complete summaries of species crosses in plants were made by Gartner (1849) and Focke (1881). Numerous crosses have been made since, but in all the crosses between varieties or between species but few of them deal with the inheritance of fertility and sterility. FERTILITY OF PARENT SPECIES AND HYBRIDS. 83 Bateson and Punnett (Bateson 1913) have reported a case of simple Mendelian inheritance of sterility in sweet peas, in which normal anthers were dominant to sterile anthers. The case is complicated by coupling with a color factor. Biffen (1905) crossed species of barley having well-defined grades Of fertility. His results showed that the hooded barleys, Hordeum trifur- catum and H. hexasticofurcatum, which are more fertile than the normal- awned barleys, were dominant to four different species of the latter kind. Segregation took place and it was inferred that only one allelo- morphic pair of characters was involved. In other crosses between well-defined types of barley he found various kinds of sterility dominant over the normal perfectly developed floret. ‘‘In these cases the various degrees of sterility, ranging from complete suppression of the repro- ductive organs in the lateral florets to reduction in size only, are clearly dominant over the perfectly developed floret.”’ Here, again, the classes obtained in the F, generation gave evidence of a simple segregation. Brainerd (1907), in his résumé of the interesting behavior of certain hybrids between violet species, reports that pronounced degrees of sterility occurred in some of the crosses. When the hybrids were mated inter se he recovered plants of normal fertility in the F, genera- tion. In discussing the phenomenon of this segregation of normally fertile strains from an almost sterile hybrid F, generation, he says: “With this diminution or entire loss of hybridity, we should expect a partial or total recovery from the impairment of fertility produced in the first cross. At any rate, it is an observed fact that many violet seedlings whose hybrid parents produced seed from only about one-tenth of their ovules, are themselves normally fertile.” We are still at a loss to know whether the fertility returned because there were recombinations of definite factors for fertility or because the simple recovery of parental types gave fertility like the parents. In the latter case the sterility of the F, hybrids might be thought to be due to disturbances arising from the admixture of widely diverse germinal elements, and a subsequent segregation of the parental types would mean a combination of factors and characters from one source, and with these the fertility of this parental type. But if fertility and sterility are due to independent factors, one should be able to combine the characters of either parent with fertility or sterility, or degrees of either. DeVries (1909) found that Oenothera lata produced no fertile pollen, although it was normally pistillate. The anthers showed all conditions, from the absence of grains to normally developed pollen, but they were always sterile. He was able to fertilize O. lata with pollen from O. lamarckiana. The anther sterility was transmitted through the ovules of O. lata, but was coupled with other O. lata characters, for it segre- gated out associated with them. 84 GENETIC STUDIES ON A CAVY SPECIES CROSS. Bauer (1911) studied a cross between the self-fertile Antirrhinum majus with the self-sterile A. molle and obtained dominance of self- fertility. The F, generation split up into self-fertile and self-sterile forms, the majority being self-fertile, but the exact ratios were not determined. Since A. molle is never self-fertile, Bauer interpreted the phenomenon as physiological rather than mechanical. This case is a peculiar kind of sterility, inasmuch as the gametes are not sterile except in certain kinds of crosses. The inheritance of this peculiarity, never- theless, follows Mendel’s laws in its essentials. Bauer also reported a cross between A. siculum and A. majus which gave sterile ovules and fertile pollen. The pollen of these hybrids was capable of fertilizing A. majus, segregation taking place subsequently. In plants, as in animals, the sterility following wide crosses is not of the same sort always, for sometimes both sexes are sterile or partly sterile, while in other cases one sex alone may be sterile or partly sterile. The literature on species crosses in mammals is meager, particularly relatively to the inheritance of sterility. Compilations of species crosses in animals by Ackermann (1897, 1898), Rérig (1903), and Przibram (1910) give a fairly comprehensive conception of the amount of work done. One is reminded of Bauer’s (1911) statement: “Noch weniger als tiber Bastarde zwischen Pflanzen-species, sind wir tiber Artbastarde bei Tieren unterrichtet. Es sind zwar auch hier zahllose Art- bastarde gelengentlich beobachtet oder auch kiinstlich erzeugt worden, aber eine auch nur einigermassen genitigende F,—Analyse est nie durchgefithrt, ja tiberhaupt nie versucht worden.” Since so little is known of the inheritance of any characters in species crosses in animals, it is not surprising that nothing is known of the inheritance of sterility subsequent to such crosses. Sterility, to be sure, often accompanies wide crosses in animals. In the Lepidoptera the classical experiments of Standfuss (1895) have shown that such crosses may give partial or complete sterility in either sex, gynandromorphs, hermaphrodites, and even the complete suppression or elimination of one sex. Recently Goldschmidt (1912) has attempted, on a Mendelian basis, to explain gynandromorphism in the cross between Lymantria dispar with L. japonica, upon the assumption that the factors for the secondary sexual characters of the two parent species are of various grades of potency. For our purposes it is not necessary to enumerate all the species crosses resulting in sterility. These have been fully recapitulated, summarized, and described by other investigators (Poll 1910, 1911; Przibram 1910). A few bovine crosses have yielded results somewhat similar to the cavy crosses in this paper. Kiihn began a series of crosses, using the genera Bibos, Bison, and Bos. The original papers were not accessible, but a summary is given by Nathusius (1912). The yak, Bibos grun- niens, has been crossed with the domestic cow, Bos taurus, and pro- FERTILITY OF PARENT SPECIES AND HYBRIDS. 85 duced sterile male but fertile female hybrids. The female hybrids were crossed back to males of both parent types; but the male hybrids remained sterile, although 19 were tested and included 3, 3, 3, and 3 domestic-blooded males. The gayal, Bibos frontalis, has been crossed with the domestic cow and likewise produced fertile female but sterile male hybrids. At least 6 3-gayal bulls were tested and found to be sterile, but 3 out of 9 3- gayal bulls were fertile. The gaur, Bibos gaurus, considered a close relative to the gayal, was crossed with the domestic cow. A male hybrid was sterile to cows (although he covered 19), but, strangely enough, he was fertile with his own sisters. The banteng, Bibos sondaicus, was crossed with the zebu, Bos indicus, and produced a sterile male. I have been told that the female hybrids are fertile, and regard the sources of information as reliable. The bison, Bison americanus, has been reciprocally crossed with domestic cattle, but most successfully when a domestic bull is used. The hybrids, frequently called cattaloes, are sterile males and fertile females. The female hybrids have been crossed back to males of both parent species, thus producing } and ? bison (Boyd 1908; Iwanoff 1911). The + bison females are fertile, as may be expected. The 3 bison females have not been fully tested, but are presumably also fertile. The + bison males are not always fertile, for Boyd reports the appearance of but 1 out of 4 tested males. Iwanoff reports a fertile 2 bison male and supposes, on purely theoretical grounds, that a mating of such a fertile male with a 4+ bison female would result in fertile 2 bison of both sexes. Boyd has more recently reported other fertile hybrid males (Boyd 1914). 19. THE FERTILITY OF THE MALE HYBRIDS. MATERIALS AND METHODS. The first two generations of male hybrids (the 5 and the { wild) were few in number and could be tested thoroughly by mating them to guinea-pigs or to their fertile hybrid sisters. But since the number of hybrids to be tested increased so rapidly in the succeeding generations (see table 72) that facilities were lacking to mate all of them, it became necessary to resort to another method, if any knowledge of their fertility was to be acquired. In testing the fertility of hybrid males by breed- ing, it was necessary to keep them with four of five females for at least 4months. Furthermore, it was found that much time and space were being wasted in trying to prove animals sterile or fertile by a breeding test, when a simple examination of the contents of the epididymis would show immediately whether it was useless to attempt to breed the hybrid. Therefore I decided to test each animal microscopically 86 GENETIC STUDIES ON A CAVY SPECIES CROSS. to ascertain whether or not a further breeding test should be applied. The value of the test is apparent, for out of 102 males tested micro- scopically 43 were found to have very few or no motile spermatozoa present and every one of them failed to breed after the most rigid breeding test. On the other hand, 44 males which proved to be fertile in breeding had an abundance of motile spermatozoa in every case. The microscopic test was simple and expedient. A male tested in this manner was anesthetized by etherization; the scrotum was thor- oughly washed with 75 per cent alcohol, and dried; and the animal was stretched on his back. A small incision, or a cut made with scissors, about 32-inch long, at the posterior end of the scrotum, exposed the edipidymis. Several of the tubules were then transected with a very small, sharp scalpel, and the liquid contents which collected were placed on a cover-glass. The cover-glass was transferred to a slide, on which a drop of physiological salt solution had been placed. The cover-glass, slide, and salt solution were, to be sure, always kept at bodily temperature. The slide was then examined under the microscope and a careful record of observations was made. There were 433 males of the different hybrid generations tested in this manner. In all cases a record was kept, showing which testicle has been used for operation. For the sake of convenience the left testicle was always used. Bilateral tests were made in enough cases to show that either testicle would give the same result; but such tests were made only after a thorough breed- ing test or with surplus animals, for transection of the epididymis on both testicles might make an animal sterile in breeding, although potentially fertile. The wound was covered with iodoform and healed completely in a week. In order to exclude any possibility of varying tests on one and the same animal under different conditions, over 100 males were retested, both on the left side and on the right, in summer and in winter, and in good condition as well as in very poor condition. The second and third tests always gave the same results as the first, with the following exceptions: the cellular contents of the epididymis were always of the same character; but it must be stated that 3 males showed immotile sperm on the first test, but motile sperm on a second test some months later.* Iam fully satisfied that the difference was due to my own early inexperience. Thereversenever occurred, for when asecond test showed immotile sperm after a first test had shown motile sperm, I could always locate the difficulty and immediately produce a repetition of the first results. Hence, I am inclined to believe that these 3 aberrant animals originally had motile sperm, and had simply failed to show it because the temperature of the slide was too low or because evaporation had *The term, sperm, used to avoid frequent repetition of the cumbersome term, spermatozoa, will be clear from the context. FERTILITY OF PARENT SPECIES AND HYBRIDS. 87 concentrated the salt solution on the slide. The results showed that a careful microscopic test, at the age of 5 months or over, is a very reliable index of sterilty or fertility. In order to test a male by breeding, it is essential that he should be healthy, and kept with vigorous adult females for a number of months. Even then a male may be potentially fertile, but fail to impregnate a female because of sluggishness or other external causes having no obvious relation to the mere presence or absence of motile sperm. The ideal test of fertility is the combination of a breeding and a microscopic test. There were, in all, 50 males tested by breeding alone, and 102 males tested in both ways. Whenever the breeding test was used a male was given every opportunity to demonstrate his fertility. The unreliability of a simple breeding test, however, was evident to me during the early part of the experiment, for a few males having an abundance of motile sperm failed to impregnate females, although continually with these for many months. Two such males were about to be given up as practically sterile after a breeding test of almost a year; but on deciding to continue the test I was greatly surprised and repaid by several litters from them. One of these two (375) did not impregnate a female until after 18 months of continued breeding. I suspect that some fertile hybrid males were not always as successful breeders as normal guinea-pigs, even though it was absolutely impos- sible to detect any difference in the abundance or character of their spermatozoa. A total of 483 males was tested by one or both tests. The indi- viduals ranged from the F, through the Fs generation, most individuals (329) belonging to the F3, Fy, and F; generations. The results are put in tabular form as far as possible and recorded in tables 73 to 77. Table 72 shows how many hybrid males in each generation were tested by either one or both methods. THE RESULTS OF THE SIMPLE BREEDING TESTS ALONE. About one-tenth of all the hybrid males were tested by a simple breeding test. They ranged from the 4 wild to the -, wild, a total of 50 individuals (see table 73). The breeding test was thorough and there is no doubt that each of them, except one 7, wild male (7305), was sterile for all practical breeding purposes. To be sure, some of them may have had immotile sperm or even some motile sperm, but they failed to impregnate any females as a normal guinea-pig would have done under similar circumstances. We have no knowledge of their germ cells. In the light of the other tests, these breeding tests became more significant. The reason that so many hybrid males of the early generations were not tested microscopically was because the animals were scarce and valuable and it was feared that an operation upon the epididymis might 88 GENETIC STUDIES ON A CAVY SPECIES CROSS. destroy any even remote chance of successful breeding. Furthermore, at this period of investigation, facilities were available for mating the males, and the need of a more rapid and expedient test was not felt. The testes of some of these males were preserved for a later cytological study. THE RESULTS OF ALL MICROSCOPIC TESTS. Our knowledge of the fertility of about two-thirds of the hybrid males depends entirely on the examination of the contents of the epididymis (see table 74). Out of a total of 483 males, 331 were tested in this manner alone, and 102 males received both a breeding and microscopic test. The total number of microscopic tests was therefore 433 (see table 75). Theresults of the microscopic examination in those animals having both tests are given in table 76. For the sake of convenience, all microscopic tests will be discussed together, thus giving larger numbers from which to draw conclusions in table 75. The hybrids are divided into four categories: with no evidence of sperm; with evidence of any sort of sperm; with any motile sperm; and with many motile sperm. A careful search made the first three classes easy to differentiate, but one must admit that there are no sharp class lines between the relative numbers of motile sperm. The classification “many motile spermatozoa”’ means that the examination showed an abundance of cells, all or practically all of which were motile sperm, being the same condition which prevails in the guinea-pig (see table 75). (1) Hybrid males without spermatozoa.—Although the contents of the epididymis were taken from several tubules at different levels, and often from both testes, and at different times, some hybrids failed to reveal any spermatozoa or any evidence of such in the form of disin- tegrating flagella and the like. Such hybrid males, however, varied widely in the nature of their contents. The early hybrids without spermatozoa, such as the } wild, usually showed a thin, clear, colorless liquid in the epididymis almost devoid of all cells, but hybrids of late dilute wild-blooded generations usually showed a thick, creamy liquid rich in cells and cell detritus. The cells present were apparently spermatogonia or spermatocytes, prematurely proliferated. The uni- formity of the cells also differed, for some males had various kinds of cells, while in others all or most of the cells were apparently alike. In the later generations, the entire contents were often large, highly refractive cells, possibly spermatids, inasmuch as cells of this type were observed to have, occasionally, incipient tails. The tubules of the epididymis in the + wild hybrids were thin and pale, but this con- dition became less and less frequent in later generations. The pro- portion of males without spermatozoa also gradually decreased. In general, we may say that the 4 wild hybrids without spermatozoa showed a thin, clear liquid with a few small cells; but later generations FERTILITY OF PARENT SPECIES AND HYBRIDS. 89 showed increasing numbers of cells and more highly differentiated cells. The transition was gradual. It is probable that the cells were incom- pletely matured germ cells. The § wild male in tables 75 and 76 (70) was examined from a histological preparation of the testis made by Dr. W. E. Castle. (2) Hybrid males with spermatozoa.—All classes of hybrids, from the F, generation on, contained some individualsshowing spermatozoa. The difference between individuals was great, both in respect to quantity and character ofsperm. Two}jwild males showeda few very imperfect, non-motile sperm mixed with a few of the usual cells. Twenty-two % wild males likewise showed sperm, but in greater numbers and some- times motile. The percentage showing sperm gradually increased, as would naturally follow, since the percentage without sperm gradually decreased. In the F; generation (;1, wild) about 96 per cent showed sperm. The F, generation showed sperm in 87 per cent of the cases; but since the total number was only 15, the results are subject to a valid objection. JI am inclined to believe that larger numbers would have given a perfect series. When few sperm were present, only few other cells might be present also, as in the 3 wild. In the later genera- tions, if the sperm were infrequent, there usually was an abundance of other cells. Moreover, the sperm present varied in motility or might be misshapen or normal. If we simply consider the presence of any kind of sperm, table 75 shows that the percentage of males with sperm gradually increased as the wild blood became more dilute. The proportion with many sperm also gradually increased, while the pro- portion with few sperm decreased. (3) Hybrid males with motile spermatozoa.—Hybrids showing sperm did not necessarily show motile sperm. Rarely a hybrid would have practically nothing but sperm, yet all of themimmotile. Such animals would of course be sterile. In other cases hybrids showed only few sperm mixed with the usual cells, but all the sperm were motile. The variations between these two classes were continuous. The percent- ages showing any motile sperm whatever increased from 16.33 per cent in the 4 wild to 86.67 per cent in the ;4, wild; and conversely, the proportion with no motile sperm gradually decreased in each genera- tion after the 3 wild. (4) Hybrid males with many motile spermatozoa.—Males having but few motile sperm could not be bred successfully. This may have been due simply to the fact that there was less chance for a spermatozoon to reach an egg. I am inclined to believe, however, that mere abund- ance of motile sperm is not the only essential to fecundation, as will be shown later. It may well be that hybrids producing motile sperm sometimes fail to produce sperm qualitatively adequate. The greatest success in breeding was obtained with males showing an abundance of motile sperm. By abundance or ‘‘many motile sperm,” as used in 90 GENETIC STUDIES ON A CAVY SPECIES CROSS. the tables, I mean, as stated above, that the epididymis was full of motile sperm and showed very few or no other cells. Males showing many motile sperm first occurred in the F; or § wild generation. Although no previous hybrid generations had shown motile sperm, nevertheless, in this generation, 7 individuals showed a condition similar to that of any mormal male guinea-pig. The percentages showing many motile sperm increased from 14 per cent in the F; to 73 per cent in the F, generation. THE RESULTS OF A COMBINED MICROSCOPIC AND BREEDING TEST. The results of the microscopic tests have been discussed. About one-fourth of the animals tested in that way were also tested by breed- ing. Of the 433 males tested microscopically, 102 also had a breeding test (see table 76). The order of the test was not always the same, for about two-fifths of these males were bred first and then subjected to a microscopic test; but since the contents of the epididymis were the same under varying conditions, it should have had no effect on the results. For convenience, we may divide the animals into classes somewhat similar to those used in discussing the miscrocopic tests. (1) Hybrid males without spermatozoa.—Twenty-three males of this type had been mated to females before a microscopic test was made. As was to be expected, none of them were fertile in breeding. (2) Hybrid males with immotile spermatozoa.—Eleven hybrids withim- motile sperm proved sterile in breeding. The number of sperm varied from a few in some cases to many or practically all sperm in others, but since all were immotile, they were, to be sure, completely sterile in breeding. (3) Hybrid males with a few motile spermatozoa.—It is very difficult to classify males with motile sperm, since all grades existed, ranging from individuals with very few motile sperm to individuals with thousands of them. In all microscopic tests animals were recorded with reference to the number of sperm present and proportion of these that were motile. The relative number of sperm was described as “‘few,” “‘half,”’ “over half,” and ‘‘all;” and the standard for “‘all’’ was the normal guinea-pig male or a completely fertile hybrid male. For example, a male recorded as “‘half” had sperm and the usual cells in about equal numbers, or he might have none of the usual cells but a deficiency of sperm. The motility was described in the records as ‘1,” ‘2,” ‘*3,” and “4.” These signs had the following significance: ‘1’? meant a few of the sperm present were motile; ‘2’ meant half of the sperm present were motile; ““3”’ meant over half of the sperm present were motile; ‘‘4” meant that all of the sperm present were motile. Obviously, this divided continuous variates into 16 crude classes. A male recorded as “‘half 4” had about half the usual number of sperm, FETILTILY OF PARENT SPECIES AND HYBRIDS. 91 but all were motile. A priori, one might expect “over half, 4,” ‘all, 3,” and ‘“‘all, 4” individuals to be fertile in breeding. In the tables, all males recorded with ‘‘many motile sperm” were of the grade “all, 4.” Any manifest departure from this condition is recorded in tables as having “‘few motile sperm.” This will make clear that our records were more discriminating than our tables. Using the term ‘‘few motile sperm” in the tables to mean any con- dition of number or motility plainly below that of a normal guinea- pig, we may say that 9 individuals out of a total of 10 were sterile in breeding. The exception was a } wild male (0469) recorded as “‘over half, 2.” This male had ‘‘over half” the usual number of sperm; but only half of these were motile. He was bred continuously for 9 months and sired one male. Possibly all males with any motile sperm what- ever might have fertilized eggs had we increased their chances by using large numbers of females and long periods of mating. (4) Hybrid males with many motile spermatozoa.—As previously stated, hybrids classified this way in the tables were as nearly like a normal guinea-pig as one could judge by examination of the contents of the epididymis. I expected they would prove to be just as fertile in breeding; but this was not the case, for some of them sired no young after a thorough breeding test. There were 7 males of this class among the 3 wild; and all but one were successful sires. This exceptional male (07721), large and vigorous, produced no young, although con- tinually with fertile females for many months. Among the ;, wild there were 22 males with many motile sperm, but only 16 of these were successful sires. The reason why the remaining 6 individuals were impotent is not clear; their weights and growth curves gave every indication of vigor; 3 of the 6 males were bred for the mini- mum time reasonably required to show fertility, and it is barely pos- sible that the cause lay there; but this still fails to account for the remaining 3. Likewise among the 5 wild, 6 males out of 24 had many motile sperm but failed to breed. Here again no evident reason, such as lack of vigor or early death, could be assigned to at least one of these cases. Of the =, wild males, 2 sired young, while one failed to—in all proba- bility because of poor condition. Summarizing the results, there were 58 males with many motile spermatozoa, and 44 of these were successful sires. The remaining 14 individuals were sterile in breeding; of these 14 it is just barely possible that because of external causes 9 may have been sterile in spite of their abundance of motile sperm; but there was surely no patent cause for the sterility of the remaining 5 males. In other words, of 49 males (58 minus 9) which gave every indication of being fertile by a microscopic test and had opportunity to prove themselves so in breeding, there were only 44 which actually impreg- nated females. To state it differently, 89.8 per cent of the male 92 GENETIC STUDIES ON A CAVY SPECIES CROSS. hybrids with an abundance of motile sperm were actually fertile, while 10.2 per cent were sterile in breeding, a phenomenon which would not happen with normal guinea-pigs. From this I conclude that the number and motility of the sperm are not the only essentials for a real fertility, inasmuch as real fertility in the last analysis must mean the capacity to fertilize eggs and sire young. There are further reasons for concluding that the motile sperm of hybrid males may be physiologically different from those of a normal guinea-pig; for it often required much more time to obtain young from the hybrid males, and the litters were unexpectedly small. In 129 litters from hybrid males, there were 238 young—an average of 1.84 per litter. The normal guinea-pigs produce about 2.4 young per litter. Some hybrid males produced large, vigorous litters, and others produced but few young after long mating. It was of course impossible to tell what proportion of the motile sperm formed were qualitatively complete in all essentials to perfect fertility; but undoubtedly some male hybrids with many motile sperm lacked other indispensable qualities, partly or completely. In addition, it may be stated that sterility was not due to the absence of the secondary sex characters, since all sorts of males, sterile or fertile, copulated and appeared otherwise normal. THE INHERITANCE OF STERILITY. Two species, fertile under the same conditions, were crossed and gave rise to sterility in the male hybrids. Some condition subsequent to hybridization disturbed gametogenesis in the males, but did not affect the females. The disturbing elements were carried and trans- mitted by the females, however, for crossing these back to the male guinea-pig gave sterile males again. After continued back crosses to the guinea-pig, increasing signs of fertility appeared and eventually completely fertile males were produced. The cause of the disturbance had, to all appearances, segregated out. One can hardly refrain from the thought that these fertile males segregated out in a Mendelian sense, and that there were a number of physiological factors involved and transmitted alternatively, the different recombinations of which gave the various expressions of fertility and sterility. To be concrete, had the sterility of the $ wild males been due to one simple factor, or to a group of completely coupled factors, or to disturbances between one homologous pair of chromosomes at some stage of reduction, then we should have expected 50 per cent of the + wild males to be fertile. If the heterozygous condition of an allelomorphic pair, Aa, caused ster- ility in the 4 wild males, but did not affect their sisters, then mating these females back to the tame, aa, would give 50 per cent Aa + 50 FERTILITY OF PARENT SPECIES AND HYBRIDS. 93 per cent aa, or fertility in one-half of the F, males. In Mendelian nota- tion it would be: AP Avsoeesacues wild gametes. Bs Bienen. pod ank ds tame gametes. RE RG ccs oc } wild zygotes {Sterile males. Abia un feawe ay 4 wild eggs. BPA cose hina eww tame sperm. ; 50 p. ct. fertile males. Aa +taa........... 3 wild zygotes; 50 p. ct. sterile females. 100 p. ct. fertile females. Furthermore, although all } wild females would be fertile, half of them would transmit sterility in the next back-cross to guinea-pigs. If the two classes of females occurred in about equal frequency (as one would expect) then 75 per cent of the $ wild males would be fertile. Express- ing this mating in the usual terms, it would read as follows: Agiech a8. ces ciweaieas + wild female zygotes. At+a+a+ta...... 4 wild eggs. BHP Bk cea ewes ee4 a tame sperm ooo 75 p. ct. fertile males. Aa +aa-+aa+aa..... + wild zygotes; 25 p. ct. sterile males. 100 p. ct. fertile females. Now, if the numbers were large, and the different zygotic classes of 4 wild females were represented in the expected proportions, then seven-eighths or 87.5 per cent of the =’, wild males should be fertile. n—1 In any generation oar males should be fertile (n being the number of the hybrid generation). Table 77 gives the probable percentages of fertile males expected in each generation from the F; to the Fy inclusive, it being supposed that very large numbers are involved and that the females of any generation are distributed approximately in the expected proportions of the differ- ent zygotic classes. Our actual experimental data show that the case is far from being as simple as this, for the percentage of fertile males in each generation does not agree with the series expected on the basis of one factor as given in table 77. Furthermore, on the basis of one factor, the males would also be divided into two distinct classes: sterile (Aa) and fertile (aa). It was shown that this was not the case. The hypothesis, at least in this simple form, does not agree with the facts. Now, if the sterility of the males had been due to disturbances between either one or both of two Mendelian pairs of factors or pairs of homologous chromosomes, then we should have expected one-fourth of the F, 4 wild males to be fertile. If we represent the two factors from the wild as A and B, and the two from the tame as a and b,then the mating of the wild, AABB, with the tame, aabb, would give hetero- 94 GENETIC STUDIES ON A CAVY SPECIES CROSS. zygotes AaBb. The females would be unaffected, but the males would be sterile on account of the disturbances between A and a, and between B and b. Mating the fertile females, which likewise have the zygotic formula AaBb, back to the guinea-pig, aabb, would give the following: Aa Bb is co cadeke caleniaerentaes 3 wild females. AB +Ab+aB-+ab........... 4 wild eggs. ab seab si cases cacaves wim ona eee tame sperm. 25 p. ct. fertile males. AaBb + Aabb + aaBb + aabb... 4 wild zygotes;75 p. ct. sterile males. eae ms 100 p. ct. fertile females. This hypothesis would explain the absolute sterility of some { wild males (AaBb), but also admit of a further maturation or tendency to fertility in those individuals with less disturbing combinations, 1. e., with more factors from the tame (Aabb and aaBb). The ultimate reces- sive, aabb, would be fertile and would occur in 25 per cent of the cases. Now, if the numbers were large and the different zygotic classes of 1 wild females were represented in about the expected proportions given, then 56.25 per cent of the Fs, or § wild males, would be fertile. One could not distinguish the different classes of F, females by inspec- tion, but the random mating to guinea-pig males would be symbolized as follows: AaBb + Aabb + aaBb + aabb..... 4 wild females. AB + 3Ab + 3aB + Qab.......... 4 wild eggs. Abi bab: weap eaeawean ea ian a ewe 6 tame sperm. SE 56.25 p. ct. fertile males. AaBb + 3Aabb + 3aaBb + Qaabb.. 3 wild zygotes 44.75 p. ct. sterile males. 100 p. ct. fertile females. Here again, if the numbers were large and the different zygotic classes of females were represented in the expected proportions, then 76.56 per cent of the 4, wild males should be fertile; and, in any generation, Ls 2 (7 *) males should be fertile (n being the number of the hybrid generation). Table 77 likewise gives the most probable percentages of fertile males expected in each generation from the F, through the Fy on the basis of two factors, it being supposed that the females of any generation are distributed in the expected proportions of the different zygotic classes. The most probable percentages of fertile males, the ultimate recessives in the different generations on the basis of various numbers of factors, from 1 to 9, are given in table 77. The general scheme will be evident from an examination of this table, for, stated in simple manner, the percentages of fertile males would be as given in table A. FERTILITY OF PARENT SPECIES AND HYBRIDS. 95 TaBLe A, Hybrid | with 1 factor. | With 2f With 3 f ith p f generation. i actor. i actors. it’ actors. | With p factors. Beeson yates .50 fertile. .25 fertile. (.50)3 fertile. (.50)P fertile. Posie wagenata .75 fertile. .5625 fertile. (.75)3 fertile. (.75)P fertile. Fa siaeree caged 875, fertile. .7656 fertile. (. 875)? fertile. (.875)P fertile. et 5 eit g Ve “ Ph. * Sierras ane fertile. ae) fertile. waa) fertile. Ee) fertile. From these series we may say that in any given generation, F,, in which the degree of wildness is ! the number of fertile males should om grt—1 Pp be “oer): where n equals the number of the hybrid generation and p equals the number of factors. In actual breeding experiments the chances of error would be great. To realize such a series of segre- gates, the different classes of females of each generation would also have to occur in approximately the expected proportions in order to give the expected percentage of ultimate recessive males in the next generation. This could only be accomplished by raising very large numbers. It is quite impossible to determine from our data whether or not the percentage of fertile males in each generation corresponds in any measure to a theoretical percentage which is based on a definite number of factors; because, as tables 72 to 75 show, not all males with many motile sperm could be tested also by breeding. Furthermore, it is shown in table 76 that at least 10 per cent of the males whose micro- scopic test gave every promise of being fertile were actually sterile after a rigid breeding test. We may feel more confident of the propor- tions with many motile sperm than of the proportions really fertile. If we examine the percentage of males in each hybrid generation, the contents of whose epididymis could not be distinguished from that of a normal guinea-pig, we find (see table 75) the following series of per- centages from the F, to the F, inclusive: 0.00 0.00 14,29 33.33 60.67 69.39 73.33 If we take the percentage of males with many motile sperm in the total tested by all methods the series is about the same: 0.00 0.00 9.46 32.38 60.67 69.39 73.33 This latter series of percentages imputes that all males sterile in a thorough breeding test alone did not have many motile sperm. From table 76 we see that this is not completely true in about one-tenth of the cases. ‘The first series is probably more accurate, as it is the percentage of males with many motile sperm in the total of microscopic 96 GENETIC STUDIES ON A CAVY SPECIES CROSS. tests rather than in the total of all tests. The series of percentages of ultimate recessives expected on the basis of eight factors (see table 77) is: 0.00 0.39 10.01 34.36 59.67 77.57 88.16 One must admit that there is a remarkable similarity between these three series for the first 5 hybrid generations at least—such a close resemblance that one wonders whether it is chance coincidence or whether there actually were 8 allelomorphic pairs involved, such that the ultimate recessives in each generation segregate out with many motile sperm. One would be forced to conclude that further factors were necessary to give real fertility in addition to mere numbers and motility, for it was shown that males with many motile sperm were not necessarily fertile. The great range of possibilities between no sperm and all motile sperm would, on this hypothesis, be due to recom- binations of factors. Individuals homozygous in 6 or 7 recessive factors would be almost fertile, for they would have segregated out most of the disturbing ‘‘wild chromosomes” and have replaced them with homologous pairs entirely from the tame source. Such an hypothesis is suggestive and alluring, but other critical considerations are necessary. The probable errors for the percentages were calculated, but are not given. I am indebted to Dr. H. L. Rietz for valuable suggestions regarding these. They would be extremely difficult to handle and very misleading. The probable error of any generation would have to be calculated on the supposition that the females of the preceding generation were normally distributed, or else one would have to take the error of all preceding generations into account. It is logically impossible to suppose that the females of any generation (except F,) could have been normally distributed. On this hypothesis we would suppose that the wild and tame had 8 factors or chromosomes which were incompatible in the F,; males, and this led to disturbances in the maturation of the sperm, but did not affect the females. We might represent the factors from the wild as AABBCCDDEEFFGGHH, and those from the tame as aabbecdd- eeffgghh. The 4 wild would be Aa Bb Ce Dd Ee Ff Gg Hh. The fertile F, females should then produce 256 kinds of gametes, but only one of these, abcdefgh, would have segregated out the disturbing elements from the wild. Now, when this gamete met its mate from the tame, also abcdefgh, it should have given fertility in the F, males. But the expectation of this combination based on random sampling is 1 in 256. The number of F, males (22) actually procured was far too small to expect an ultimate fertile recessive male. One would, however, expect recombinations which had eliminated some of the disturbing elements. Such were actually obtained, for 2 F, males showed a few deformed, immotile sperm. (See tables 74, 75.) If the ultimate recessive, fertile males actually lacked all disturbing elements from the wild, then in FERTILITY OF PARENT SPECIES AND HYBRIDS. 97 mating them to the tame guinea-pigs we should expect them to breed true to fertility on this hypothesis. In spite of hypotheses, when fertile males occurred and were bred to guinea-pigs the male offspring were not all completely fertile, as will be shown. Hence we can not regard the fertile males as simple, ultimate recessives in a Mendelian sense. There is evidence of segregation of factors for fertility, but the case is more complicated than the strict hypothesis of 8 factors allows. What part interaction of factors plays, we do not know. Nor do we know that all guinea-pigs carry the absence of factors disturbing fertility in these crosses. It may be added that some definite characters from the wild were surely compatible with fertility, because males with the “‘wild agouti” were also fertile. THE MALE OFFSPRING OF FERTILE MALE HYBRIDS. Offspring of fertile male hybrids were also tested. They may be divided into two classes: the offspring of fertile male hybrids and female hybrids, and the offspring of fertile male hybrids and guinea- pigs. It seems that when male hybrids were fertile they could be bred to any sort of fertile female (see table 78). Male guinea-pigs have been bred to all classes of female hybrids from the F, to the Fy genera- tion inclusive. Male hybrids of every class from the F; through F, were bred successfully to guinea-pig females. Male hybrids in each generation from the F; to the F; inclusive were successful sires in matings with female hybrids of the same or different generations. In this last class of matings such diverse crosses as the following were possible: F; males were bred to F, ,F., F;, and Fs female hybrids, while F, males were bred to F:, Fs, Fs, and F; females. A 4 wild female, 5 years old, was impregnated by her great great grandson. A 7 wild female was successfully mated with a ;'; wild male. The different possible combinations of successful matings indicate that fertile male hybrids of any blood dilution can impregnate any sort of fertile female. Fermitt Mate Hysrins in Crosses with Femate Hysrips. In all, 39 offspring from this sort of mating were tested (see table 72); 36 received only a microscopic test, while 3 received both tests. Tables 74 and 76 show that all classes of males were produced, ranging 1The percentages of males with many motile sperm in the Fe and Fy generations were 69.4 per cent and 73.3 per cent respectively (table 75). As a matter of fact, these percentages do not correspond to the expectations based on 8 factors (table 77), but are nearer the results one would expect with 12 or 20 factors in the Fs, and F7 generations respectively. This can be readily computed from the formula given on page 95. Fass Sa P=.604 —(.96875)P=.604 — pp log .96875=log.694 pp = 11.51 Ee awecnege (.98438)P =.733 p log .98438 =log .733 p=19.72 98 GENETIC STUDIES ON A CAVY SPECIES CROSS. from those with no sperm to those fertile in breeding. The § wild males, bred to } wild females, gave one male with many motile sperm (fertile in breeding also) out of 7 tested. The +’, wild males, bred to their sisters in blood, gave 8 males with many motile sperm out of 14 tested. The F; males bred to F; females gave 5 males with many motile sperm out of 8 tested. One F, male bred to an F, female gave one male, and he had many motile sperm. The other 9 matings correspond to these, for irrespective of what generation the fertile male sires were they gave a preponderance of sterile male offspring when bred to intense wild-blooded female hybrids, but increasing signs of fertility in their sons when bred to females of later generations. For example, the F; and F, males, bred to F, and F, females respectively, gave entirely sterile sons; but one F; male gave sons with many motile sperm when mated to F, females, while two F, males, bred to F; females, gave sons with many motile sperm also. If the hypothesis advanced is correct, and a fertile male hybrid represented the same combination of factors for fertility as a guinea- pig male, then from mating fertile male hybrids with female hybrids we should expect about the same results that were obtained by mating guinea-pigs to similar female hybrids. We have already shown that. when guinea-pigs were mated to the different generations of female hybrids, increasing signs of fertile males came with each back-cross. The hypothesis implies that more and more females were being obtained which lacked the disturbing factors and failed to transmit such. The results in the sons of fertile male hybrids bred to female hybrids are consistent with this hypothesis, for the intense wild females gave more sterile sons than the dilute wild females in this class of matings, just as they did when mated to guinea-pigs. The two series of percentages of males with many motile sperm produced in these mat- ings are given in table 79. The number of sons from female hybrids and male hybrids is far too small for broad generalizations; but the results indicate that sterility is transmitted in the same manner by the female hybrids crossed with male hybrids as when crossed with guinea- pigs. The percentages of sons with many motile sperm in both sorts of crosses in the different generations from the F, to the F, are as follows: 0.0 0.0 14.3 58.8 63.6 .... 100.0 with fertile hybrid sires. 0.0 0.0 14.3 33.3 60.7 69.4 73.3 with guinea-pig sires. For further details and numbers involved, see table 79. FERTILE HysBrip Maes 1N Crosses WITH GUINEA-PIGS. A total of 22 sons from this sort of mating was tested, all having a microscopic test only. The fertile hybrid sires belonged to the Fs, F,, F;, and Fy generations (see tables 72 to 76). To test the hypothesis that the fertile hybrid sires, with many motile sperm, had segregated FERTILITY OF PARENT SPECIES AND HYBRIDS. 99 out as recessives and that we should expect the same results from such fertile hybrid males as with guinea-pig males, 14 of them were mated with guinea-pig females. The microscopic tests showed that 21 of their 22 sons were indistinguishable from a normal guinea-pig male. The one exception (071524) was the son of an F, male (7506) and a guinea-pig female (9186). The same sire and dam gave two other sons (15 and #16) with many motile sperm. The exceptional son had nothing except motile sperm in the epididymis, but they were extremely few in number. I am informed by Dr. W. E. Castle that other sonsof fertile hybrid males and guinea-pig females likewise showed signs of sterility. Fertility, however, appears to be obtained most frequently from this class of matings, as the records show that 95.5 per cent of the sons of fertile hybrid males and guinea-pigs had many motile sperm. In view of these facts, the hypothesis (that fertility in the hybrids simply means eliminating 8 disturbing factors) can not be maintained. There is strong evidence of segregation, but the case may be complicated by other conditions, such as interaction of factors and the like. THE SECONDARY SEXUAL CHARACTERS. Some observations on the secondary sexual characters were made. Although not taken as statistical data, they were numerous enough to be of value. Male hybrids of all classes showed the sex instinct. In fact, I have never seen a single healthy male hybrid, sterile or fertile, which did not attempt copulation. The hybrids fought with each other for the possession of the females. How successful they were in copu- lation is not known, but since the organs were morphologically similar to those of a normal guinea-pig, it is probable that there were no diffi- culties in this respect at least. There is good evidence that ejaculation took place and that normal uterine plugs were formed from the clotted mass, as in the case of any normal guinea-pig. It is well known that severing the spinal cord will often produce an ejaculation. ‘The method of killing the hybrids was to sever the skull and axis by holding the head, swinging the animal and suddenly arresting the motion. It was noticed that in all cases an ejaculation took place if one then pressed the groin, a clot forming almost immediately. This clotting or coagula- tion of the semen, supposed to be due to a ferment, vesiculase (Marshall 1910), is common to both the hybrid and the guinea-pig males and gives rise to the uterine plug in the female. More than 200 hybrid males: showed this peculiar reaction when properly stimulated. No hybrids failed to show it if they were killed when adult. Hence it is almost certain that they were physiologically potent in every respect, except. in the production of sperm. The accessory organs, including the seminal vesicles and prostate glands, were always apparently normal. The only differences noted were that sterile hybrids might have small testes and a pale, small epididymis. 100 GENETIC STUDIES ON A CAVY SPECIES CROSS. 20. THE FECUNDITY OF THE FEMALE HYBRIDS. Almost every female hybrid in each generation was fertile in breeding. The only exceptional generation was the ~ wild, in which the only female was sterile. Occasionally a female hybrid was sterile, but such cases were infrequent. Although no data were taken on sterility in guinea-pigs, I am of the opinion that sterility in the female hybrids was no more frequent than in these. There was at least one source of data which gave information on the degree of fecundity in the female hybrids—the average number of young per litter (see table 80). The wild C. rufescens, bred in captivity, gave 46 offspring in 34 litters, or an average of 1.35 per litter. We do not know what their average per litter is in the wild habitat. The tame guinea-pigs, used as dams in matings with wild sires to produce the 3 wild hybrids, gave 37 young in 16 litters, or an average of 2.31 per litter. This shows that the wild males impregnated the guinea-pigs just as successfully as a guinea-pig male would have; for the average per litter in our guinea-pigs was 2.34. Minot (1891) found an average of 2.56 in his experiments with tame guinea-pigs, but his numbers were smaller (see table 80). The F, hybrids were intermediate, for they produced 83 young in 52 litters, or an average of 1.60. In fact, they were a little less fecund than a theoretical midparental condition would demand, for this would be 1.845. They were about as ‘‘wild,” to all appearances, as the pure wild females, but were slightly more prolific. The F, hybrid females, the 1 wild, produced 217 young in 114 litters, or an average of 1.90 per litter. The F; hybrid females produced 312 offspring in 152 litters, or an average of 2.05. The subsequent hybrid generations did not show an increased average, although they were produced by successive back- crosses to the guinea-pig male. The analysis of these data is complicated by a number of conditions. The guinea-pigs raised in our laboratory gave larger litters in summer than in winter; for in summer they produced 218 young in 85 litters, or an average of 2.56, whereas in winter they produced 266 young in 122 litters, an average of 2.16 per litter. The young born from January 15 to July 15 were considered winter litters in these data, because the ovulations and conceptions corresponding to these births ranged from about November 8 to May 8. Minot (1891) found a similar condition in his experiments. Minot also found that the first litters were smaller than the average; but first litters are usually borne by young females and it may mean that the smallness of first litters is entirely an effect of age. This may account for the fact that our F,, F;, and Fs females failed to show an increased average per litter, since many of the female hybrids in these generations were young, and the records contain a large proportion of litters from such females. FERTILITY OF PARENT SPECIES AND HYBRIDS. 101 The results therefore show that in mating the wild C. rufescens to guinea-pigs, the litter average of the F, hybrids was about intermediate, and continued back-crosses raised this average gradually. It may be added that the proportion of females producing some fertile males or males with all evidences of fertility gradually increased in each generation. Certain females in the later generations produced only fertile males, but the number of young from one female was necessarily small and we can not be sure but that they would have given sterile sons had larger numbers been possible. However, one should eventually be able to produce female hybrids with the fecundity of the guinea-pig species and having only fertile sons. Combining these characters with wild characters, such as the peculiar wild agouti, should also be possible. Two abnormal females (9263 1 wild and 9393 +4, wild) should be recorded. The former, 9263, had an enlarged clitoris, resembling a penis, but also all the female characters, bore 2 young, and gave milk. It was difficult to keep her with a male, for as she grew older they fought continually. The latter, 9398, also had an enlarged clitoris, which was very nearly of the same form and size as a normal penis. The female external characters were all normal. She had no young to my knowledge, but upon one occasion she showed large and abrupt loss of weight, and gave milk at the same time. It is barely possible that she had aborted. When kept alone for some time, and subse- quently placed with a male, she allowed the male to attempt copulation. When placed with females she always attempted copulation, making the same sounds and going through the movements of a normal male. If she was penned with a male and females, she and the male fought continually for the possession of the females. She was killed at the age of 2 years, and the ovaries were examined. They were abnormally large, measuring about 13 inches in length and an inch in width. The follicles were greatly distended, some measuring 0.75 inch in diameter. Abnormal ovaries of this type were not uncommon in other female hybrids which bore young and were otherwise perfectly normal in all respects. The viscera of many female hybrids were examined, but no data were taken on the occurrence of this type of abnormality. 21. THE SEX RATIO IN THE HYBRIDS. The many recent experiments with sex-linked and sex-repelled char- acters have led to the current opinion that sex itself is a Mendelian character, and that one sex is homozygous while the other is hetero- zygous for sex-determining factors. One would expect an equality of the two sexes in the long run on this hypothesis; but when an excess of one sex occurs consistently, it is supposed that the heterozygous parent fails to produce the two kinds of gametes in equal numbers, or 102 GENETIC STUDIES ON A CAVY SPECIES CROSS. that selective fertilization takes place, or that unequal viability of the two sexes during early development accounts for the discrepancies. Guyer (1909), compiling the proportion of sexes in hybrid birds, stated: ““‘When due allowance is made for all errors, the facts still indicate that there is a marked tendency for the hybrids, especially those from widely separated parents, to be male.’”’ Since the female is supposed to be favored by increased nutrition, he thought the excess of males might be due to default in metabolic processes because of incompati- bilities between dissimilar germ plasms, such incompatibilities being especially inimical to the production of females, King (1911), tabulating the sex ratios of hybrids between wild and albino rats, stated: “It appears, therefore, that hybridizing alters the sex ratio by producing a marked increase in the relative proportion of males. This conclusion is in essential agreement with that reached by Buffon, by R. and M. Pearl, and by Guyer.” King found 231 males to 194 females in the totals of the first three hybrid generations, this being a ratio of 119.07 males to 100 females. Minot (1891) crossed guinea-pigs inter se and obtained 223 males to 187 females, or a ratio of 119.2 males to 100 females. The results in the hybrids between C. rufescens and C. porcellus did not show an excess of males, but, to the contrary, a significant excess of females (see table 81). The wild parent bred in captivity gave 20 males, 25 females, and 1 of unknown sex. The 3 wild hybrids gave 14 males and 23 females, or a ratio of 60.87 males to 100 females. There were 2 young of unknown sex, having died prematurely. If we call them males, the ratio is 69.57 males to 100 females. The Fs, or 2 wild, gave 31 males and 52 females, or a ratio of 59.62 males to 100 females. The F3, or § wild, gave 101 males to 116 females, or a ratio of 87.07 males to 100 females. It is apparent that, as the generations became less hybrid in nature, the sexes were gradually approaching equality. After the § wild, the sexes were more nearly equal, for the next four generations gave a total of 406 males to 409 females, practically an equality of sexes, for the ratio is 99.24 males to 100 females. This is strikingly different from the total of the first three generations, in which there were 146 males to 191 females, or a ratio of 76.44 males to 100 females. The total results of all hybrids were 552 males and 600 females, or a ratio of 92 males to 100 females. These ratios do not confirm the results shown by Guyer or King. Previous data have shown that sterility was common in the males of the early hybrid generations, for there were disturbances in sperma- togenesis. It is shown here that the early generations also gave a deficiency of males. May it not be possible that the same incompati- bilities between dissimilar germ plasms which gave rise to sterility in gametogenesis also caused disturbances in fertilization. Possibly male FERTILITY OF PARENT SPECIES AND HYBRIDS. 103 zygotes may have been formed less frequently, or may have been elimi- nated at an early stage. The only similar case in mammalian crosses which I have been able to find is that described by Boyd (1914), in which the bison and domestic cattle were crossed. Boyd found that his hybrids gave 60 females to 17 males, or a ratio of 28.33 males to 100 females. Boyd likewise found sterility common in the males, similar to that in my hybrids. 22. SUMMARY AND GENERAL CONCLUSIONS. (1) Crosses between C.rufescens males and C. porcellus females gave completely sterile male hybrids and fertile female hybrids. By cross- ing the female hybrids back to guinea-pig males, } wild hybrids were obtained, which were again sterile males and fertile females. A few males of this second hybrid generation, however, showed some degen- erate non-motile sperm. By repeated back-crosses of female hybrids to guinea-pigs, increasing signs of fertility appeared. Fertility seemed to act like a very complex recessive character; for the results obtained were what one would expect if a number of dominant factors for sterility were involved, the elimination of which would give a recessive fertile type. There was an enormous range of forms between hybrids with no sperm and fertile hybrids with many motile sperm. (2) The results indicated that a completely fertile hybrid male could be bred to female hybrids or to guinea-pigs, giving about the same results as a normal guinea-pig male in such matings. (3) The secondary sexual characters of all malehybrids were normally developed. (4) The wild C. rufescens has a smaller litter average than the guinea- pig. When the wild males were bred to guinea-pig females, the size of the litters was that of the guinea-pig. The female hybrids produced by this cross, however, gave a litter average intermediate between that of the wild and tame. By repeatedly crossing the hybrid females of one generation back to guinea-pig males to produce the next hybrid generation, the litter average was raised almost to that of the guinea- pig itself. This is all the more interesting since guinea-pig males were used to raise the litter average. (5) Two female hybrids showed some male secondary sexual char- acters. One of these with marked male instincts had abnormal ovaries. Abnormal ovaries were common in the female hybrids. (6) The sex ratio in the hybrids showed a marked preponderance of females, expecially in the early hybrid generations, 7. ¢., in those genera- tions which must have been most hybrid in constitution. TABLES, TaBLe 1, TABLE 2. Matings of wild females with wild males, all Matings of non-agouti guinea-pig females to the offspring like the parents showing the wild agoutt males, producing heterozygous agouti pattern. agouti young. Parents. Offspring. Parents. Offspring. QAAX SAA AA Qaa X PAA Aa 2 1 3 1125 1 11 3 1 7 1625 1 4 4 1 2 3204 1 1 15 33 3 9470 33 3 25 1 7 9473 33 3 46 1 6 9536 33 2 46 55 3 9586 33 2 #f3 ae i 8370 55 3 4 A 1 9586 55 5 2 A 1 55 1 3 24 1 4 1 Wild male 1 15 24 % 2 Total....... 37 3 4 15 are ats - i i TABLE 3. 3 Matings of % wild females, heterozygous in 4 1 be or lacking agouti, with guinea-pig males 15 24 < homozygous in agouti. Offspring, all 25 33 6 agouti. Total..... 46 Parents. Offspring. *In this and the subsequent matings an uncertainty exists as to the identity of one or 9 Aaor 9aa X GAA | AA or Aa both parents. The record here given indicates all the possibilities, based on the record of what animals were penned together. 135 1961 3 248 2157 3 TaBe 4, 311 1961 2 Matings of 3g wild, homozygous in agouti, ae Ih 2 with guinea-pigs lacking it. 1 at 1961 2 137 Parents. Offspring. oa 2157 4 129 247 2157 2 AA X= aa Aa 252 135 310 1961 5 9399 co 40 5 312 9448 166 3 129 9 485 215 3 137 9499 166 4 247 elar : o'506 9 186 6 252 Total........ 21 Total......... 27 TABLES. 105 TaBLe 5. Summary of Tables 2, 3, and 4, showing agouti always epistatic. Parents. Offspring. antes Females. | Color. | Males. Color. Agouti. 2 Guinea-pig..... aa Pure wild...... AA 37 3 4 wild......... { el \cuinea-pig. cay AA 27 4 dg wild........ AA Guinea-pig.... aa 15 4 Guinea-pig..... aa qe wild........ AA 6 Totaly .2i ss deavawwaed eee teee ease ssa dueeecewe se 85 TABLE 7. Matings of 4 wild females, heterozygous in agouti, with guinea-pig males lacking it. TABLE 6. Matings of 4 wild females, heterozygous in Parents. Offspring. agouti, with guinea-pig males lacking it. QAa X laa Aa + aa 87 170 1 1 Parents. Offspring. 90 1541 3 6 90 214 2 1 91 1541 4 3 QAa X Saa Aa + aa 91 214 2 1 92 9246 1 0 92 1541 1 1 63 11030 2 5 95 1541 1 0 96 12612 4 8 63 2193 3 1 96 170 1 1 68 11030 6 4 97 11030 0 1 98 12612 0 2 68 a19s ® - “105 1541 6 3 68 4 3 2 106 0 0 1 107 9758 2 1 69 11030 3 3 107 617 1 1 69 4 0 2 107 —25 1 2 107 199 1 0 me a) 8 122 12612 | 2 1 75 9246 2 0 147 1543 1 1 149 G 1 1 a a 150 1543 | 6 2 118 11030 3 1 150 201 0 1 160 617 1 1 208 Si, Se oe 160 ~35 | 1 38 253 617 0 1 550 199 1 1 69 606 201 1 2 11030 1 1 645 199 1 1 118 90 69 95 1541 2 2 i 2193 | 3 1 a 91 a. 9246 | 2 1 seer | 4 wild 98 lwild 9246 | 6 6 i) 12612 | 2 4 Awild 11030 | 8 2 a ea | ae Total..... 47-36 rie 141 | 2 3 Most probable ia 41 Total..... 55 59 tion...| )41 42 Most probable ee expectation...) 57 57 106 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 8. TABLE 9. Matings of } wild females with guinea-pig Matings of 2g wild females, heterozygous in males, in which one parent is heterozygous agouti, with guinea-pig males lacking it. in agouti and the other lacks it entirely. Parents. Offspring. Parents. Offspring. QAa X caa Aa + aa 9 Aa X caa Aa + aa 195 2132 2 4 140 1881 5 1 277 2132 3 1 166 85 0 1 277 72 1 4 238 2366 0 3 304 223 2 1 311 2278 2 0 317 163 0 1 470 2036 2 0 318 163 1 1 599 2278 2 3 340 12815 3 4 723 117 J 0 341 12815 4 3 356 223 2 2 e hes.| | hee cae 392 12815 30 8 rae? cs 393 64 3 3 141 2002 o 1 pos ai 3 3 143-2002 1 0 415 40; 0 1 144 2002 0 1 416 40 1 0 145-2196 1 2 421 54 2 6 170 —30 1 2 435 166 0 4 173-1917 5 0 436 166 o 3 177-1923 1 0 461 217 2 1 178 ~—-:1928 7 5 481 215 2 2 197 —98 6 1 519 144 2 3 205 —98 0 4 523 53 0 3 207 =. 2083 1 2 536 64 2 0 208 —- 2083 1 4 544 40 3 0 209 =. 2083 2 1 556 104 1 2 212 =. 2083 1 0 560 64 1 3 215 —30 0 1 565 64 2 3 234 —30 2 1 580 103 2 3 264 2002 4 1 601 103 1 0 361 2196 1 0 614 224 0 2 367-2196 1 1 2030 = 2006 2 2 208 2030 2132 1 1 209+ 20 3 2 304 29} 83 cpat Roa 4 2 Total...) 50 37 277 Most probable fe 44 os} 72 | 2 expectation.) |44 43 Total..... 57 72 Most probable ie 65 Tase 10. expectation. .| |65 64 Matings of }_ wild females, hetero- zygous in agoutt, ‘with guinea- pig males lacking it. TasBie 11. ; Summary of Tables 6-10, in which an equality of agouti Parents. Offspring. and non-agouti young is expected. QAa x dc'aa Aa + aa Parents. Offspring. Table. 403 2278 1 0 Females. Males. Aa + aa 403 12835 3 0 529 12835 6 5 6 4wild...... Guinea pig. 47 36 603 42 2 2 : : 687 115 2 1 7 dwild...... Guinea pig. 55 59 702 201 1 0 8 twild...... Guinea pig. 50 37 733 201 2 1 : : : 850 12835 0 1 9 fg Wild...... Guinea pig. 57 72 10 wild...... i ig. Total....... 17 10 = i ae Most probable { 13. «14 Ot al aici au. mentale naeoa pa nae 226080214 expectation...| \14 18 Most probable expectation........... 220 220 TABLES. 107 TaBie 12. TaBLe 13. Matings of § and #5 wild females with Offspring of female hybrids according to color of wild guinea-pig males, both heterozygous agoutt. im agouti. Offspring. Parents. Offspring. ae Mothers. A x Per cent. QO Aa X Aa AA or Auaa | Ticked. | Dark. | Light. ticked. TICKED. pa an ‘ ; 3 wild..... 18 5 19 43 108 1917 5 0 2 wild..... 19 0 2 90 pe ba a 9 3 wild... 8 0 0 100 172 1917 3 0 ds wild...... 29 0 0 100 a. ae : 6 gy wild...... 21 0 0 100 219 —98 6 3 DARK ze wild. lw 536 1917 1 ‘ i wild...... 0 1 1 0 LIGHT, Total. ssc cseens 32 4 do 7 Most probable ex- : mle ego ? 17 41 pectation........ 27 9 § wild...... 5 0 0 100 TaBLe 14. Matings of wild hybrid females, heterozygous in agouti and with dark ticked belly, with guinea-pig males, also heterozygous in agoutt but with light belly. Parents. Offspring. Color. Dark. Light. Light. Dark. Non-agouti. Formula. 9 A’a o'Aa A’A or Aa A’a aa 108 1436 1 3 0 108 1917 4 i 0 131 2196 1 0 0 166 2196 4 3 0 172 1917 1. 2 0 198 2002 3 1 0 203 —98 1 0 0 219 —98 4 2 3 536 1917 0 1 1 Total......... 19 13 4 Most probable expec- | ——.~___ os fe tatION 6 sissieunee dees 9AA’+9Aa 9A'a 9aa Zygotic formula proven.| 3 4 4 108 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 15. TaBLe 17, Matings of 3; wild females, carrying Matings of 4 wild females in which wild and tame agouti, with guinea- all the offspring are black. pig males lacking agouti. Off- Parents. Offspring. Parents. spring. Light- | Non- | Light- | Ticked- 9 BB X o'BB or Bb BBorBb belly. | agouti. | belly. belly. 9 Bb or bb X o'BB 9 AA’ O'aa Aa A’a 127 1881 1 127 2034 2 399 40 3 2 135 1961 3 448 166 1 2 499 166 3 1 140 1881 6 166 85 2 Total........ 4 5 197 119 2 Most probable 222 1881 2 expectation. . 6 6 299 2034 5 252 85 2 264 85 3 TaBLeE 16. 296 117 1 Matings of 4 wild females, in 311 1961 2 which all the offspring are black. 311 2278 2 Parents. Offspring. 312 1961 2 OBB X @BB —| 312 938 1 @Bb X @BB | BBor Bb 471 96 5 475 96 1 87 170 2 90 214 3 574 G 2 91 214 3 576 99 4 96 12612 12 96 170 2 577 15 2 97 11030 1 599 2278 5 98 12612 2 101 12612 4 659 117 1 107 199 1 671 96 3 115 201 2 122 12612 3 723 117 1 124 11030 2 793 96 L 150 201 1 263 12612 2 of. 98 1 550 199 2 135 605 170 5 ra) 1961 2 606 201 3 135 642 199 2 645 199 : so} 1961 5 797 4 1 312 842 199 1 574 122 12612 6 96) 12612 5 ie 99 e 122 816 Total....... 67 Dotal:: 0 beens 76 TABLES. 109 Taste 18. TaBLe 19. Matings of 5 wild females, in which all the offspring are black. Matings of gy wild females, in which all the offspring are 0 black. neers Parents. ff Parents. ce 2 spring. spring. Parents. Offspring. 9BB X @BB,Bborbb| pp || 9 BB X o'BB,Bb or bb Q9Bb X #BB Bb || Bb X SBB BB 9BB X @BB Qbb X @BB or Qbb X GBB | Bb 9Bb X @BB_ | BBorBb Qbb X oBB 277 2132 4 554 104 2 277 72 5 556 104 3 383 12845 9 278 72 4 559 103 4 384 12845 5 304 223 3 560 64 4 385 2278 1 307 223 3 565 64 5 385 12845 2 317 163 1 580 103 5 403 2278 1 318 163 2 587 104 1 403 12835 3 329 223 8 589 104 2 488 12835 3 330 223 5 601 103 1 489 12835 2 333 223 10 613 224 4 529 12835 11 340 12815 8 624 224 1 547 15 3 341 12815 10 679 163 4 548 15 2 356 223 4 812 4 2 603 42 4 357 215 2 814 53 3 617 42 3 364 217 2 832 12815 1 618 42 4 392 12815 7 304 223 é 633 94 1 393 64 6 356 635 94 2 399 40 4 277 7s ‘ 662 12845 3 414 40 6 945 687 115 3 415 40 1 364 166 5 699 115 2 416 40 1 463 702 201 1 419 54 2 436 166 % 706 15 1 421 54 8 499 729 15 1 422 54 5 533 104 { 733 201 3 435 166 6 556f 740 115 3 436 166 4 eS 64 a 745 205 1 448 166 2 579 772 12835 3 460 53 5 613 801 4 2 461 217 3 oH 224 2 806 55 2 463 217 1 357 812 4 2 478 166 4 481 215 8 847 4 2 481 215 4 485 850 12835 2 484 12815 1 419 2430 2415 3 485 215 2 422 54 6 ar ieaar 4 492 217 2 454 749 503 144 1 | 533 es 12835 5 510 163 3 587 B-R Sp 2 772 519 144 5 || 589 ae 42 2 523 53 3 357 849 524 53 3 | 411 215 7 630 55 2 536 64 2 481 638 538 144 6 || 485 634 94 3 539 144 3 635 540 144 7 Total.......... 265 rd 42 3 544 40 2 Pee el 115 3 396 384 2278 3 385 Total...... 115 110 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 20. TaBLe 21. Matings of 3; wild females, in Summary of Tables 2, 6, and 16-20 (all offspring are which all the offspring are black pigmented). black. Table. Females. Males. BB or Bb Parents. Offspring. 2 Guinea-pig.....} Pure wild...... ar 9 BB X o'BB m . i OBb x @BB BB or Bb 6 Fowild’es ccccas Guinea-pig.... 83 16 4 wild......... Guinea-pig.... 67 515 45 3 17 $wild......... Guinea-pig.... 76 516 45 2 18 qe wild........ Guinea-pig.... 265 17 ‘i , ‘ bs = Z 19 gg WA iz xik asians Guinea-pig.... 115 630 55 1 20 ei wild........ Guinea-pig.... 37 638 55 2 759 170 6 DP Otalie ss ecase 2 eonscatusnau Weeks ba aioe 680 783 12845 2 783 170 2 515 45 4 517 758 | TABLE 22. a 12835 | 3 Matings of % wild females, heterozygous in Ee 45 | 1 k, with brown guinea-pig males. 629 638 55 3 Parents. Offspring. 806 Totals os .2. , 3 9Bb X Sbb | Bb + bb TaBLE 23. 89 9758 | 0 2 7 ‘ 90 1541 5 4 Matings of 4 wild females hetero- 91 1541 4 5 zygous in black, with brown 92 9246 0 1 guinea-pig males. 92 1541 0 2 Peet et 9g aie ae 95 1541 | 1 0 Parents. Offspring. 105 1541 6 3 oe ee et 106 1541 1 0 QBb X o'bb Bb + bb 106 0 1 0 SSS SSS 107 9758 2 1 127 1541 0 1 107 617 1 1 131 9758 0 1 107 25 2 1 131 2196 1 1 110 9578 1 3 145 2196 6 4 110 617 4 1 166 2196 6 0 13 1543 3 3 177 1923 3 0 115 1543 6 7 178 1923 10 3 119 1543 9 5 207 2083 1 1 147 1543 2, 0 208 2083 2 2 90 209 2083 0 2 95 1541 3 1 215 —30 0 1 105 215 —28 4 0 91 219 —9s | 4 5 ict tah | eT 240 2366 2: 2 113 242 —-2366 0 38 a to 367 2196 1 2 92 402 2366 | 2 1 A feat | C8 470 2036 1 2 Total.. 57 45 Total..... 43 31 Most probable | Expected..... 37 37 expectation. .; 51 51 TABLE 24, TABLES. Matings of 2s and 3, wild females with guinea-pig males, in which one parent is heterozygous in black and the other is brown. Parents. Offspring. QBb X cobb | Bbh+bb ps wild 275 2132 2 4 278 2132 1 3 275 ae 2132 3 0 gz wild 547 2132 | 0 1 Qbb X Bb ps wild 2030 2006 3 1 Totals 6:4cacenaiiave-s 9 9 Most probable expecta- MODS. 2 3.38. 4:5 enone ees 9 9 TaBLe 25. Summary of Tables 22-24, in which we expect an equality of black and brown offspring. 111 TABLE 28, Matings of 2; and 3 wild brown females with brown guinea-pig males. Matings of 34 Parents. Offspring. Qbb X bb bb gg wild 195 1436 1 195 2132 6 292 —67 2 298 —67 2 381 2366 5 382 2366 1 440 2366 1 456 2366 2 457 2366 4 536 1917 2 2030 2132 2 290 291 —67 2 439 440 2366 1 gs wild 546 2132 2 Total......... 33 TABLE 27, wild brown females with brown guinea- pig males. Parents. Offspring. Parents. Ofspeing. Table. Females. | Males. Bb + bb eb as ob: | bb 108 1436 4 22 Rwild......... Guinea-pig....| 57 45 108 1917 5 23 4 wild......... Guinea-pig....| 43 31 130 2036 6 ig wild ; : 130-1541 2 24 Ane wid }cuinea-pis. aos 9 9 170 —30 4 Syibat a gush u6 ee 5 TOTAL i cuciainngeneeeteanes ORM 10985 a Te 3 Most probable expectation............. 97 97 ion sce 197 —98 . TABLE 26. 203 —98 1 ‘5 a z é ‘ 205 —98 4 Matings of } wild females with a guinea-pig 239 1543 1 male, all heterozygous in black. 939 2366 . 237 2036 4 Parents. Offspring. 238 2366 3 248 2157 3 137 i @Bb X @Bb |BBor Bb +bb oe ator 3 197 a 1541 3 198 2002 2 1 rae ae 264 2002 5 0 203 96 6 1917 Total..... 7 1 a 5 Most probable expectation. 6 2 Total. 78 112 TABLE 29. Summary of Tables 27 and 28; matings of brown GENETIC STUDIES ON A CAVY female hybrids and brown guinea-pig males. SPECIES CROSS. TaBLe 30. Matings of guinea-pig females, heterozygous in extension, with a wild Cavia rufescens Parents. Offspring. male. Table. Females bb. Males bb. bb Parents. Offspring. 27 «| Awild....... Guinea-pig..... 78 Ee X (EE | EE or Ee 28 qg wild...... Guinea-pig ceesene 31 oan a oe 28 gz wild...... Guinea-pig..... “a eins Ee : Total siasdes ase ncsate winwsesunieiarsnatson Total..... 5 TaBLE 31. TABLE 33. Matings of 4 wild females, homozygous in extension, with guinea-pig males carry- ing restriction. Parents. Offspring. QEE X o'Eeoree | EE or Ee 68 4 5 69 4 2 72 617 4 75 9246 2 253 617 1 4 a a} 9246 3 | 4 wild 9246 12 Total. i... s..as. 29 TABLE 32. Matings of 4 wild females with guinea-pig males, in which one parent only carries restriction. Parents. Offspring. QEE X CEe or ee QEe x GEE EE or Ee 90 1541 9 91 214 3 105 1541 9 107 9758 3 107 617 2 107 —25 3 110 9758 4 110 617 5 115 1543 13 119 1543 14 124 617 2 150 1543 8 160 617 2 160 —25 4 90 } 1541 4 105 Dotals suey 85 Matings of 4 wild females with guinea-pig males, in which one parent only carries restriction. Parents. Offspring. QEE X cEe or ee QEeoreeX GEE EE or Ee 108 1917 5 127 1881 1 127 2034 2 140 1881 6 197 —98 7: 197 119 2 198 2002 3 203 —98 1 205 —98 4 207 2083 2 208 2083 4 209 2083 2 212 2083 1 215 —30 1 219 —98 9 222 1881 2 222 2034 5 232 2366 2 234 —30 3 238 2366 3 240 2366 4 242 2366 3 264 2002 5 311 1961 2, 311 2278 2 312 1961 2 367 —98 2 402 2366 3 471 96 5 475 96 , 576 99 4 599 2278 5 671 96 3 574 577 99 4 197 203 —98 6 208 209 2083 5 212 141 143 144 2083 7 198 Total 128 Matings of js wild females with guinea-pig males, in which one parent only TABLES. TABLE 34. carries restriction. Parents. Offspring. Parents. Offspring. Q EE X Ee or ee Q EE X GEe or ee QEeoreeX GEE EE or Ee QEeoreeX GEE EE or ee 195 1436 1 524 53 3 195 2132 6 536 64 2 275 2132 6 538 144 6 277 2132 4 539 144 3 277 72 5 540 144 7 278 2132 3 544 40 1 278 72 4 B54 104 2 304 223 3 556 104 3 307 293 3 559 103 4 329 223 8 560 64 4 330 223 5 565 64 5 341 12815 7 580 103 5 356 223 4 587 104 1 357 215 2 589 104 2 364 217 2 601 103 1 381 2366 5 613 224 4 382 2366 1 614 224 2 393 64 6 621 224 1 399 40 4 679 163 4 414 40 6 814 53 3 415 40 1 304 416 40 1 aaa 228 . 419 54 2 277 421 54 8 aa ie _ 422 54 5 364 435 166 4 are alt 8 436 166 3 436 440 2366 1 el 166 2 448 166 2 613 453 5A 2 eal aaa 2 456 2366 2 357 457 2366 4 is 215 8 460 53 5 485 461 217 3 419 463 217 1 ia} 54 6 478 176 4 454 481 215 4 357 485 215 2 492 217 2 481 215 z 503 144 1 485 519 144 5 523 53 3 Totalicssiacsuenss 250 113 114 GENETIC STUDIES ON A CAVY SPECIES CROSS. Tas Le 35. TaBLe 36. Matings of 3, wild females with guinea- Matings of gz wild females with guinea-pig pig males, in which one parent only males, in which one parent only (or (or neither) carries restriction. neither) carries restriction. Parents. Offspring. Parents. Offspring. QEE X GEE or Ee EE or Ee QEE or Ee X o' EE | EEKor Ee 383 12845 9 515 45 3 384 12845 5 516 45 2 385 2278 1 517 45 5 385 12845 2 629 55 3 403 2278 1 630 55 1 403 12835 3 638 55 2 488 12835 3 759 170 6 489 12835 2 783 12845 2 529 12835 11 783 170 2 546 2132 2 ae 45 4 547 2132 1 Bae 547 15 3 \ 4 548 15 2 935, A 603 42 4 629 617 42 3 638 55 3 618 42 4 806, 633 94 1 635 94 2 TOtahieswsaaan 34 662 12845 3 687 115 3 702 201 1 TABLE 37. 730 201 3 : 2 e 745 201 1 Summary of Tables 30-36, in which animals of extended 306 55 2 pigmentation are expected, since only one parent (or 812 4 2 neither) carries restriction. 847 4 2 Soe 2415 3 Parents. Offspring. ed 4 Table. ae 42 2 Females. Males. EE or Ee 849 al 55 2 ss, 30 Guinea-pig..... Wildesicuice een | 6 oe 94 3 31 % wild......... Guinea-pig.... 29 on 42 3 32 4 wild......... Guinea-pig.... 85 384 3 33 $ wild......... Guinea-pig.... 128 ss) 2278 34 fa wild........ Guinea-pig.... 250 Titales ess vee: 96 35 gz wild........ Guinea-pig.. . “i 96 | 36 ez wild........ Guinea-pig.... 34 Totalecrs raw aelsguinmewak ae ase 628 TABLES. 115 TABLE 38. Matings of 4 wild females, heterozygous in the extension factor, with guinea-pig males lacking it. TABLE 39. Matings of } wild females with guinea-pig males, in which one parent is heterozygous in the extension factor and the other parent lacks tt. Parents. Offspring. Parents. Offspring. QEe X cee Ee + ee QEe X cee Qee X o'Ee Ee-+ee 91 1541 € 2 113 1543 3 7 197 1541 0 1 Total....) 10 9 130 2086 | 2 4 Most probable ey ae 2 i expectation. 1 - a isi ona i i s 145 = 2196 2 8 166 2196 5 1 170 —38 0 2 TaBLe 40. 172 1917 2 1 Matings of ps wild females with guinea-pig a a : : males, in which one parent is heterozygous 367 2196 3 7 in the extension factor and the other 470 2036 2 1 parent lacks it. oa 141 | 2 1 203 Parents. Offspring. 172 far 1917 1 4 9 Ee X cee Total....) 22 32 Qee X He He +b. 6 Most probable expectation.| 27 27 298 =—b7 5 1 303 1923 0 1 333 163 6 4 510 163 1 2 Tapue 41. Total....| 8 8 Matings of 3, wild and Q wild females Most probable with guinea-pig males, in which one expectation.| 8 8 parent is heterozygous in the exten- sion factor and the other parent lacks i. TaBLe 42. : ‘ 4 Parents. Offspring. Summary of Tables 88-41, in which we expect an equality of animals of extended pigmentation and restricted pigmentation. Q Ee X cee : Qee X wEe Ee + ee Parents. Offspring. Table. Females. Males. Ee + ee gy wild 772 12835 1 2 801 4 1 1 38 3 wild....| Guinea-pig....| 10 9 ou 12835 1 1 39 2 wild....| Guinea-pig....) 22 32 a 115 2 1 40 js wild....] Guinea-pig....| 8 8 éy wild seat 12835 2 1 41 gy Wild.... Guinea-pig.... 5 ild....| Guinea-pig.... 2 Petiliaces ones 7 6 = a Most probable expec- Potalioncs cas naw ovestes aa 47 55 tation............ o . 6 Most probable expectation......... 51 51 116 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 43. TaBLe 46. Matings of hi female hybrids with guinea- Matings of wild hybrid females and guinea-pig pig males, in which both parents were males, in which one parent only carries rozygous in the extension factor. albinism. s Parents. Offspring. Parents. Offspring. Seo eee ¢ Or cc 9 Cc or cc X o'CC CC or Ce QEe X Ee EE or Ee + ee 2 wild 90 1541 9 91 1541 9 4 wild 797 4 0 1 a. ee ! 4 wild 166 85 1 1 110 617 5 170 —30 3 1 173 1917 4 3 113 1543 10 178 1923 12 1 1 ate a 248 2157 3 0 147 1543 2 252 85 2 0 150 1543 8 264 85 3 0 160 —25 4 574 G 1 1 606 201 3 577 15 1 1 oe oe : 7s wild 340 12815 ¢ 1 ‘ 392 12815 6 1 # wild 166 85 2 dy wild 706 15 0 1 oo oe e 740 115 2 1 23 2 178 1923 13 Total.........+- 45 13 ee. Sean s Most probable expec- 44 14 tation as 15 282 2366 3 Suda a bee en cents 312 98 : 402 2366 3 471 96 5 TaBLe 44, 475 96 1 Matings a albino guinea-pig females with fia 871 96 3 dig wild 195 2132 6 Cavia rufescens males. 275 2139 6 277 2132 4 Parents. Offspring. 277 72 5 303 1923 1 317 223 3 9ee X SCC Ce 318 163 2 329 223 8 330 223 5 1125 1 11 340 12815 8 1625 1 4 356 223 4 3024 1 1 392 12815 7 9536 33 2 399 40 4 414 40 6 Total. c+. 18 419 54 2 422 54 5 481 215 4 503 144 1 TaBLe 45. 519 144 6 Matings of 4 wild females with guinea-pig 523 53 3 males carrying albinism as a recessive 524 53 3 character. 560 64 4 565 64 5 Parents. Offspring. 580 103 5 679 163 4 9CC X &Ce CC or Ce a 72 3 945 63 2193 4 #2 wild 635 94 2 68 2193 2 687 115 3 75 9246 2 740 115 3 69 oe us} 2193 4 635 : 75 Total icstanaeascad 252 3 wild 9246 3 Table 44............ 18 4 wild 9246 12 Table 45........... ; 27 Totals cc ecei3.s 27 Grand total....... 297 TABLES. 117 TasBLe 47. Tasie 48. Matings of } wild females, heterozygous Matings of 4 wild females with guinea-pi, in color, with albino guinea-pig males, in which one parent is haerooigous S. in color, and the other is an albino. Parents. Offspring. Parents. Offspring. Q9Ce X ee QCe X dee | Cetec Geo Mey | Sere 143 2002 1 1 107 —25 3 i 144 2002 1 3 115 1543 13 Z 166 2196 6 3 207 2083 2 3 Total....} 16 8 208 2083 4 5 Most probable 209 2083 2 1 expectation,.| 12 12 212 2083 1 4 232 1541 1 0 312 1961 2 1 208 TasBLe 49. 209 2083 5 6 Matings of 2g and x wild females ane with guinea-pig males, in which one Total....) 25 27 parent is heterozygous in color and Most probable the other is an albino. expectation.| 26 26 Parents. Offspring. TasBLe 50. Summary of Tables 47-49, in which we expect an equality 9Ce X eee a of colored and albino offspring. Qee X Ce or oe Parents. Offspring. Table. qty wild 357 215 2 1 Females. Males. Ce + ce 461 217 3 1 484 12815 1 0 47 4 wild....) Guinea-pig....| 16 8 505 215 0 4 48 % wild....) Guinea-pig....| 25 27 587 104 1 1 49 wild....| Guinea-pig.... 8 8 832 12815 1 1 49 gz wild....| Guinea-pig.... 2 0 sz wild 546 = 2132 2 0 Totalicssreciciereseaorn gine. 5143 Most probable expectation......... 47 47 Total.......... 10 8 Most probable expec- THEOE 6 ce ctw 9 9 TABLE 52. Matings of 4 wild females with guinea-pig males, in which both parents are heterozy- Taste 51. gous in color. Matings of } wild females with guinea- Parents. Offspring. pig males, in which both parents are terozygous in color. Ce X aCe | CCorCe+ce Parents. Offspring. 141 2002 1 0 145 2196 8 3 198 2002 3 4 9@Cce X Ce CC or Ce+ ce 296 117 1 I 361 2196 0 2 574 G 2 1 107 9758 2 1 576 99 4 1 9758 3 1 577 15 2 1 1170 5 1 574 ed 99 4 0 Total...... 10 3 Total....| 25 10 Most probable {9 4 Most probable ie 9 expectation.| 10 3 expectation.| |27 8 118 TBALE 53. Matings of 25 wild females with guinea- pig males, in which both parents are heterozygous in color. Parents. Offspring. Q9Ce X SCe | CCorCe+ce 278 2132 3 2 278 72 4 1 290 —67 0 1 292 —67 2 1 341 12815 t 3 416 40 1 1 435 166 4 2 436 166 3 1 463 217 1 1 478 166 4 1 544 40 2 3 554 104 2 1 559 103 4 1 601 103 1 0 602 103 0 3 290 Gar —67 2 2 Total....| 40 24 Most probable expectation.| 48 16 TABLE 56. Matings of wild hybrid females with guinea-pig males, t zygous in roughness. he latter homo- Parents. Offspring. Qrirf X SRfRf Rfrf 4 wild 127 2034 2 222 2034 5 QRfirf X SRIRE gz wild 2430 ©2415 3 Lotalycasiccuhessusve 10 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 54. Matings of 3 wild females with guinea-pig males, in which both parents are heterozy- gous in color. Parents. Offspring. Q@Cce X Ce | CC or Ce+ ce 702 201 1 2 733 201 3 0 745 201 1 1 Total.... 5 2 Most probable ‘a 2 expectation. 6 1 TaBLe 55. Summary of Tables 61-54, in which we expect 3 colored ani- mals to 1 albino. Parents. Offspring. Table. Females. Males. CC or Ce+ ce 51 3 wild....| Guinea-pig.... 10 3 52 4 wild....| Guinea-pig.... 25 10 53 wild....| Guinea-pig.... 40 24 54 wild....| Guinea-pig.... 5 2 TOtal s ccanwaase cea vata tea eaacte 80 39 s 89 30 Most probable expectation......... ie 29 TaBie 57. Matings of guinea-pig females, heterozygous in roughness, with smooth wild Cavia rufescens male. Parents. Offspring. QRirf X erirf | Rirf + rfrf 1125 1 3 4 1625 1 1 3 Total....) 4 7 Most probable 5 6 expectation. 6 5 TABLES. TaBLe 58. 119 Matings of wild hybrid females with guinea-pig males, in which one parent is heterozygous in roughness. Parents. Offspring. Parents. Offspring. Qrfirf X oRfirf | Rf + rf 9 Rfrf X rfrf Rf + rf 3 wild be 2193 2 2 | 2s wild 539 144 2 1 540 144 5 2 a anee | El sa aeiia 815 4 / 1 2 118 2193 2 2 516 45 2 0 7s wild 2030 2006 3 1 517 45 2 1 516 Bie 45 1 3 QRirf X @rirf otal. so. 405-.0sn'o 27 26 Table 57.............. 4 7 + wild 606 199 1 2 642 199 0 2 Grand total..... 31 33 645 199 0 2 Most probable expecta- qe wild 478 166 2 3 HOD 655 cewinee e e484 32 32 538 144 3 2 TaBLe 59. Average weights of males of the parent races and hybrids, calculated at regular intervals. 4 Days for which the averages are calculated. Bs Class. Be 10 | 20 | 40 | 60 | 80 | 100) 140 | 180 | 220 | 260 | 300 | 340 | 380 | 420} 460 4] Wild........ 95] 128] 206] 265] 313] 341] 378) 393] 400} 407| 415} 420) 425) 427) 427 6 | 3 wild....... 170} 241) 363] 468) 555} 619) 722) 795) 845) 888] 920) 941} 954) 959) 962 15 | 4 wild....... 112} 163} 249) 323) 396) 475) 562) 645] 706) 742} 782| 797] 823] 856) 869 62 | 4 wild....... 144] 200} 301) 393) 478] 555) 673) 758) 820) 870) 894| 908] 913) 915} 916 *28 | Guinea-pig . .| 130) 184] 283) 369) 437} 494| 603) 677) 724) 768] 799] 812) 816)....)..... 53 | Guinea-pig . .| 165] 198) 310} 421) 508) 573) 650) 707| 754) 794) 821) 839) 853) 862)..... *This was a small inbred strain. TaBLe 60. Average weights of females of the parent races and hybrids, calculated at regular intervals. % 4 Days for which the averages are calculated. Bs Class. 2s 10 | 20 | 40 | 60 | 80 | 100] 140} 180} 220| 260 | 300} 340] 380} 420) 460 5 | Wild........ 83] 110) 157] 198] 230) 257} 297| 325) 348) 369} 383] 394! 412) 417) 422 9 | 4wild....... 178] 235} 331] 412) 480) 530) 601| 656) 703) 736) 766) 788) 823} 828) 832 22 | 4 wild....... 110] 162} 244) 310) 376) 419} 498] 566) 621) 663) 700} 742) 785) 801) 812 76 | $ wild....... 139] 197} 296} 390) 465) 527; 623] 678] 710) 734) 753) 764) 771) 777| 777 *17 | Guinea-pig . .| 132] 186] 283] 365) 432) 500| 570, 622; 654) 680; 690) 702) 713) 729] .... 56 | Guinea-pig . .| 135] 192] 299) 382) 460, 527) 623) 690 739} 770| 787| 801] 801} 809]..... *This was a small inbred strain. 120 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 61. Coefficients of variability for the weights of the males in the parent races and hybrids, at six successive ages. Ages in days. Number of Class Average of individuals. 7 coefficients. 100 180 260 300 340 380 Ais Gena Wild............ 7.29 | 6.17} 4.99 | 4.49) 4.59 | 4.80 5.39 Gaicwesd 4 wild........... 13.56 | 12.16 | 10.74 | 9.94 8.99 8.19 10.60 Besa seats 4 wild........... 12.29 | 13.38 | 10.24 | 9.16 8.00 8.08 10.19 G2 ys eras 4 wild........... 12.23 | 11.47 | 10.63 | 10.09 | 10.20 | 10.67 10.88 oo: Guinea-pig......| 10.24 7.84 8.16 8.16 9.60 9.90 8.98 SS ei ae Guinea-pig...... 8.22 | 10.41 | 10.19 8.17 8.03 6.34 8.56 *Small inbred strain. TABLE 62. Coefficients of variability for the weights of the females in the parent races and hybrids, at six successive ages. Age in days. Number of Average of individuals. ies coefficients. 100 180 260 300 340 380 Bis gasws Wild neicscencesees 17.38 | 13.80 | 13.16 | 12.98 | 12.23 | 11.89 15.24 ee # Wild sesc-cecuons 9.65 8.56 8.16 7.97 7.28 | 4.96 7.76 DD is 6.8 aczysius Awild........... 12.91 | 10.77 | 10.08 | 11.51 | 11.44 | 10.07 11.13 TO. sas oes 4 wild........... 11.09 9.51 | 11.08 | 10.77 | 10.61 | 10.64 10.62 Ly Guinea-pig...... 10.00 7.05 6.27 6.19 6.20 5.56 6.88 ee Guinea-pig...... 12.47 9.72 9.32 | 10.08 9.29 9.16 10.01 *Small inbred strain. TABLE 63. Averages in millimeters of sixteen different skeletal dimensions of the males in the parent races and in hybrids. Number of individuals in the different classes. Numbers designating measurements. 3 5 16 60 78 wild. 4 wild 2 wild. 4 wild. guinea-pigs. Nos Lecasss 60.13 | 70.58+0.38 | 66.74+0.24 | 68.98+0.17 | 68.48+0.13 Qicaae 54.30 | 62.46+0.54 | 59.43+0.26 | 60.87+0.15 | 60.98+0.13 Bioko 20.03 | 22.90+0.21 | 22.20+0.15 | 22.74+0.07 | 22.33+0.06 4....., 34.13 | 39.48+0.35 | 37.65+0.18 | 38.34+0.10 | 38.20+0.08 Biscia avis 27.43 | 32.48+0.21 | 29.883+£0.15 | 30.80+0.09 | 30.66+0.08 Ohi veseissect 32.13 | 36.22+0.24 | 33.90+0.19 | 34.68+0.10 | 34.41+0.08 eT arewsctechnt 37.13 | 42.64+0.46 | 41.01+0.18 | 41.38+0.11 | 41.48+0.08 Pena yas 24.83 | 27.90+0.26 | 25.56+0.11 | 25.46+0.09 | 25.20+0.05 Qu. exad 34.56 | 40.40+0.31 | 38.10+0.24 | 37.72+0.12 | 37.53+0.11 Tsu ceva 30.46 | 35.30+0.34 | 33.40+0.21 | 34.52+0.11 | 34.31+0.08 SL ae cme 25.26 | 31.06+0.36 | 29.22+0.14 | 29.90+0.08 | 29.67+0.07 A2yeeNan 45.23 | 51.80+0.59 | 47.93£0.21 | 48.47+0.15 | 47.99+0.11 ABR cued 24.60 | 28.16+0.33 | 27.42+0.11 | 27.86+0.08 | 27.92+0.07 14...... 34.53 | 40.00+0.16 | 37.14+0.18 | 38.10+0.13 | 37.95+0.08 Toy aad 40.53 | 44.90+0.54 | 41.28+0.20 | 42.41+0.14 | 42.06+0.10 Whesscd3 43.50 | 50.55+0.77 | 47.38+0.26 | 48.37+0.15 | 48.06+0.11 121 TABLES. TaBLe 64, Averages in millimeters of sixteen different skeletal dimensions of the females in the eared races eae in hybrids. Number of individuals in the different classes. Numbers designating measurements. 1 8 20 65 63 wild. 4 wild. 2 wild. 4 wild. guinea-pigs. NOs. deed acs 59.20 | 67.10+0.36 | 65.40+0.28 | 65.52+0.19 | 65.61+0.13 Disease 62.20 | 59.41+0.33 | 57.45+0.25 | 57.41+0.16 | 58.49+0.11 Doreen 20.70 | 22.156+0.24 | 22,150.10 | 21.64+0.06 | 21.75+0.05 Bowes 34.20 | 38.06+0.29 | 36.96+0.17 | 36.84+0.11 | 37.28+0.08 Bay iee 26.30 | 30.62+0.24 | 29.01+0.18 | 29.01+0.10 | 28.96+0.06 Cy cans 30.90 | 34.21+0.19 | 32.84+0.17 | 32.37+0.10 | 32.58+0.08 aaa ton 36.20 | 40.89+0.18 | 89.09+0.14 | 39.15+0.12 | 40.11+0.08 Bu ceags 23.80 | 27.14+0.21 | 25.23+0.14 | 24.78+0.07 | 24.81+0.06 Bowens 33.10 | 88.220.41 | 36.74£0.25 | 35.33+0.13 | 35.47+0.10 Ty 2eyaes 28.70 | 832.12+0.22 | 31.88+0.15 | 31.88+0.08 | 32.70+0.06 dda 2eh ad 23.30 | 29.387+0.15 | 27.81+0.21 | 27.81+0.09 | 28.65+0.05 Wo eae 41.60 | 560.45+0.30 | 44.46+0.25 | 46.96+0.11 | 47.67+0.11 1B. ce e4 23.50 | 27.46+0.21 | 26.61+0.14 | 27.22+0.07 | 27.72+0.06 14...... 31.30 | 38.44+0.36 | 36.54+0.25 | 36.88+0.11 | 37.51+0.08 IB axena 33.60 | 42.90+0.43 | 40.57+0.27 | 41.09+0.13 | 41.31+0.08 Ay ions 37.50 | 48.91+0.29 | 46.13+0.381 | 46.82+0.15 | 46.89+0.10 TABLE 65. Standard deviations in millimeters of sixteen anor skeletal dimensions of male guinea-pigs and hy! Number of individuals in the different classes. Numbers designating measurements. 5 16 60 78 4 wild. } wild. } wild. guinea-pigs. Now Tes ievee 1.27+0.27 | 1.40+0.17 | 1.92+0.12 | 1.74+0.09 D secaste- calor 1.80+0.38 | 1.54+0.18 | 1.76+0.11 | 1.65+0.09 Be jegede 0.68+0.15 | 0.90+0.11 | 0.75+0.05 | 0.72+0.04 Beeaeene 1.17+0.25 | 1.08+0.13 | 1.12+0.07 | 1.08+0.06 Bexvvaga 0.70+0.15 | 0.87+0.10 | 1.00+0.06 | 1.00+0.05 On etawks 0.78+0.17 | 1.14#0.14 | 1.09+0.07 | 1.08+0.06 cere 1.53+0.33 | 1.08+0.13 | 1.27+0.08 | 1.04+0.06 Bua niaae 0.78+0.17 | 0.66+0.08 | 0.98+0.06 | 0.66+0.04 D gx sie 0 1.03+0.22 | 1.40+0.17 | 1.833+0.08 | 1.40+0.08 MO peck ye sd 1.18+£0,24 | 1.26+0.15 | 1.28+0.08 | 1.08+0.06 Ricca vex 1.18+0.25 | 0.838+0.10 | 0.96+0.06 | 0.89+0.05 TQ aeiay es 1.75+0.42 | 1.22+0.15 | 1.76+0.11 | 1.37+0.08 1S. ca bs 1.09+0.23 | 0.66+0.08 | 0.89+0.06 | 0.84+0.05 TA iawn a0 0.46+0.21 | 1.00+0.13 | 1.50+0.09 | 1.04+0.06 TB divans 1.59+0.38 | 1.02+0.14 | 1.57+0.10 | 1.24%0.07 TGs enn 2.29+0.54 | 1.82+0.18 | 1.61+0.10 | 1.37+0.08 122 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaB_e 66. Standard deviations in millimeters of sixteen different skeletal dimensions of female guinea-pigs and hybrids. Number of individuals in the different classes. Numbers designating measurements. 8 20 65 63 4 wild. 2 wild. 4 wild. guinea-pigs. MWe, Tec ccewn 1.49+0.25 | 1.82+0.20 | 2.24+0.13 | 1.50+0.09 Dek taatann 1.38+0.23 | 1.68+0.18 | 1.88+0.11 | 1.29+0.08 Ps is wees 1.01+0.17 | 0.68+0.07 | 0.75+0.04 | 0.53+0.03 Ae ented 1.21+0.20 | 1.14+0.12 | 1.33+0.08 | 0.96+0.06 Be seas 0.99+0.17 | 1.17+0.13 | 1.13+0.07 | 0.72+0.04 a diene x 0.79+0.13 | 1.10+0.12 | 1.18+0.07 | 0.94+0.06 WU reysctiauehaie 0.76+0.13 | 0.95+0.10 | 1.36+0.08 | 0.93+0.06 Be caaderton 0.87+0.15 | 0.91+0.10 | 0.87+0.05 | 0.70+0.04 Ds. seveitstbaine 1.71+0.29 | 1.67+0.18 | 1.50+0.09 | 1.14+0.07 TOs sarextene’s, 0.94+0.16 | 0.99+0.11 | 0.95+0.06 | 0.66+0.04 a Vay appa cease 0.64+0.11 | 1.37+0.15 | 1.01+0.06 | 0.64+0.04 1D) sedeicsses 1.25+0.21 | 1.59+0.17 | 1.29+0.08 | 1.22+0.08 1S. cece 0.87+0.15 | 0.92+0.10 ! 0.84+0.05 | 0.67+0.04 PAE. sciaeseee. 1.50+0.25 | 1.60+0.18 | 1.833+0.08 | 0.83+0.05 DBs soe dieed 1.82+0.31 | 1.72+0.19 | 1.46+0.09 | 0.91+0.06 as LG chaste: 1.22+0.21 | 1.98+0.22 | 1.78+0.11 | 1.09+0.07 TaBLe 67. Coefficients of variability of sixteen different skeletal dimensions of the males in the guinea-pigs and hybrids. Number of individuals in the different classes. Numbers designating measurements. 5 16 60 78 4 wild. 3 wild. % wild. guinea-pigs. No: -Lavecces 1.81+0.39 | 2.09+0.25 | 2.78+0.17 | 2.54+0.14 Dis! Sve ne tae 2.88+0.61 | 2.59+0.31 | 2.89+0.18 | 2.71+0.15 Sis eee 2.99+0.64 | 4.11+0.49 | 3.29+0.20 | 3.23+0.17 Beha dee 2.96+0.63 | 2.86+0.34 | 2.92+0.18 | 2.84+0.15 D vce aleehe 2.15+0.46 | 2.93+0.35 | 3.23+0.20 | 3.25+0.17 Gr estas 2.16+0.46 | 3.36+0.40 | 3.13+0.19 | 3.13+0.17 ica acd iyo 3.59+0.77 | 2.63+0.31 | 3.06+0.19 | 2.50+0.14 Bice as 2.79+0.60 | 2.58+0.31 | 3.87+0.24 | 2.60+0.14 ee 2.54+0.54 | 3.67+0.44 | 3.52+0.22 | 3.73+0.20 10's 5 sxyeac 2.79+0.60 | 3.76+0.45 | 3.70+0.23 | 3.16+0.17 LL es eaters 3.79+0.81 | 2.84+0.34 | 3.21£0.20 | 2.98+0.15 Le keeRee 3.38+0.81 | 2.54+0.30 | 3.64+0.22 | 2.86+0.16 ema oo 3.88+0.83 | 2.41+0.29 | 3.18+0.20 | 3.00+0.17 TA) neuitare 1.16+0.28 | 2.68+0.34 | 3.94+0.24 | 2.74+0.15 LS eioveien 3.55+0.85 | 2.46+0.34 | 3.70+0.23 | 2.94+0.16 16 iho 4,.52+£1.08 | 2.80+0.36 | 3.33+0.21 | 2.86+0.16 123 TABLES. Pee TaBLe 68. Coefficients of variability of sixteen different skeletal dimensions of the females in the guinea-pigs and hybrids. Naik bees Number of individuals in the different classes. designating ‘Gai measurements. 8 20" = 65 ee } wild. 4 wild $ wild. guinea-pigs. Nos. Laci vauss 2.22+0.37 | 2.79+0.30 | 3.41=0.19 | 2.28+0.14 Dols ihe e 2.32+0.39 | 2.92+0.31 | 3.28+0.20 | 2.21+0.13 Des eo4s 4.54+0.77 | 3.07+0.33 | 3.4840.19 | 2.46+0.15 Bet 8 eau i 3.19+0.54 | 3.09+0.33 | 3.61+0.21 | 2.57+0.15 Dieta eau 3.23+0.54 | 4.05+0.44 | 3.91+0.23 | 2.47+0.15 Os cs ead 2.30+0.39 | 3.835+0.37 | 3.49+0.21 | 2.88+0.17 sien 1.87+0.32 | 2.43+0.26 | 3.49+0.21 | 2.32+0.14 Bins. antans 3.22+0.54 | 3.63+0.39 | 3.52+0.21 | 2.82+0.17 Wace 4,.47+0.75 | 4.54+0.48 | 4.24+0.25 | 3.2340.19 LO sands 2.92+0.49 | 3.11+0.35 | 2.98+0.18 | 2.00+0.12 Thceccgaid 2.17+0.37 | 4.93+0.54 | 3.60+0.21 | 2.240.138 ee aa 2.49+0.42 | 3.57+0.39 | 2.76+0.17 | 2.55+0.16 Lian aeen ey 8.17+0.53 | 3.460.387 | 3.08+0.18 | 2.41+0.15 14e essaas 3.91+0.66 | 4.37+0.48 | 3.61+0.22 | 2.21+0.14 - }- LBepyedan 4,.24+0.71 | 4.24+0.47 | 3.56+0.22 | 2.19+0.14 VGrge sec 2.49+0.42 | 4.29+0.47 | 3.81+0.23 | 2.32+0.15 TABLE 69. Averages of the coefficients of variability of sixteen measurements in the hybrids and guinea-pig.* Class. Males. Females. dwild......... 2.93+0.17 3.050.138 dwild......... 2.89+0.09 3.62+0.10 $ wild......... 3.34+0.05 3.45+0.05 Guinea-pig..... 2.94+0.04 2.45+0.04 *The probable errors were calculated from the formula: y | 2 2 2 . Be A ete +e, +... . according to which the probable error of the average of a series of statis- tical determinations is equal to the reciprocal of the number of determi- nations into the square root of the sum of the squared errors of the = - individual determinations. 2 Re Pas TaBLe 70. Occurrence of an interparietal bone. With interparietal. Total Percentage Class. SS | skulls with inter- Males. |'Females. | examined.| parietal. C. rufescens..... 0 0 7 0 C. porcellus...... 5 4 141 6.4 ewwildaen palesntets 1 1 13 15.4 F Wildes ccedee 6 9 46 82.6 3 wild........... 4 19 125 18.4 124 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLeE 71. Ratios of the average of measurement 9 to the average of measurement 11; and the averages of the ratios of measurement 9 to measurement 11 in the individual skulls. Ratios of averages. Averages of ratios. Class. Males. Females. Males. Females. Wildicctisoe 1.37 1,42 1.37 1.42 £wild......... 1.30 1.30 1.29 1.30 3 wild siiedudes 1.30 1.32 1.31 1,31 $ wild......... 1.26 1.26 1.27 1,27 Guinea-pig.... 1.26 1.24 1.26 1.24 TaBie 72, Totals of various classes of males tested by the different methods. Tested by} Micro- | Tested Total |Totalmi-| Total Generation of male hybrids. | breeding | scopic | by both | breeding | croscopic | number only. | test only.|methods.| tests. tests. tested. By, oF wildisc.s.2¢esc4200005 5 0 1 6 1 6 Fi) 239i acne essen 14 7 1 15 8 22 Fay F wild vs sees vesuecaa ss 25 28 21 46 49 74 Fy, pg wild................. 6 55 44 50 99 105 Fs, gb wild... 22.0... cc eee 0 123 27 27 150 150 Frei. dp Wildsays cnsni4.aeede 0 45 4 4 49 49 Fy, aheewild so deeswencae’ 0 14 1 1 15 15 Fa, abe Wild. ..........000. 0 1 0 0 1 1 Offspring of hybrid males and females)... cess causaw wens 0 36 3 3 39 39 Offspring of hybrid males and guinea-pigs.............. 0 22 0 0 22 22 Totals; s¢cccussacnes 50 331 102 152 433 483 TaBLE 73. Results of a simple breeding test alone. Z Tested by Generation of male | jreeding | Sterile. | Fertile. hybrids. only. F,, #wild....... 5 5 0 F,, 3 wild....... 14 14 0 F;, } wild....... 25 25 0 Fy, 2g wild...... 6 5 1 Totals..... 50 49 1 125 TABLES. TaBLy 74, Results of a simple microscopic test alone. Totals Percent-| Number | Percent- having | Number| Number' oe elena age with] with |age with Generation of only a | with no | with any pet be til ¥| any many | many male hybrids. micro- | sperma- | sperma- a ow ‘ oe | motile | motile | motile scopic | tozoa. | tozoa. ee See sperma- | sperma- | sperma- test. : “ | tozoa. | tozoa. | tozoa. Fy, 3 wild....... 7 5 2 28.6 0 00.0 0 00.0 F;, 4wild....... 28 18 10 35.7 0 00.0 0 00.0 Fy, zg wild...... 55 20 35 63.6 16 29.1 11 20.0 F;, gy wild...... 123 16 107 87.0 77 62.6 67 54.5 Fe, gy wild...... 45 2 43 95.6 33 73.3 31 68.9 Fy, ra wild..... 14 2 12 85.7| 12 85.7{/ 10 71.4 Fs. gt, wild..... 1 0 1 | 100.0 1 | 100.00 1 100.0 Offspring of hy- brid males and females........ 36 3 33 91.7 22 61.1 18 50.0 Offspring of hy- brid males and guinea-pigs.... 22 0 22 100.0 22 100.0 21 95.5 TaBie 75. Results of all microscopic tests. : is r A 3 > 2 7a. | Sa) eo, | ae] pe lag | ey lee 1D 3 “mo Bg ms So =“ gg ne og Ae/ 28) FS | ag] Fe) aalFo | as [Fes 2| FS o 8 = 8 og : Hloma] = g 2 As Generation of | 32| | 8 | #8 a | RF | 3 (/So8) Fs |S. male hybrids. | ,@| S86 | 23 | 82 | $8 | 8S |288) so |oag a2| ga] 82 a| 8, | #2 /888| ge |sas $°) 8?) so] B@| be) BS ses) BS [gaa a a Aa a me) 42g io 2g {a Fi, 3 wild........ 1 1 100.0 0 | 00.0 0 | 00.0 0 | 00.0 F,, } wild........ 8 6 75.0 2 25.0 0 00.0 0 | 00.0 F3, $ wild........ 49 27 55.1 22 44.9 8 16.3 7 14.3 Fy, py wild....... 99 30 30.3 69 69.7 46 46.5 33 33.3 Fs, ¢s wild....... 150 18 12.0 | 132 88.0] 102 68.0 91 60.7 Fe, @y wild....... 49 2 4.1 47 95.9 36 73.5 34 | 69.4 Fy, zt wild...... 15 2 13.3 13 86.7 13 86.7 11 | 73.3 Fy, stg wild...... 1 0 00.0 1 | 100.0 1 | 100.0 1 | 100.0 Offspring of hy- brid males and females......... 39 3 7.7 36 92.3 23 59.0 19 48.7 Offspring of hy- brid males and guinea-pigs..... 22 0 00.0 o2 100.0 23 100.0 21 95.5 Totals... .| 433 89 20.6 | 344 79.5 | 251 58.0 | 217 50.1 126 GENETIC STUDIES ON A CAVY SPECIES CROSS. TaBLe 76. Results in the combined microscopic and breeding tests. a é 4 al « + | moa | oh | ab Felt SUE Le (eles el alesi a! | sBlasi elas S[a8|3)3) 88] 21] 8 : or les/ £USs/ss) fies) £) Flee) &) & Generation of male ay E 3 A/ERS | Ss| 4] E A) 5 | 5 & E a} 2 hybrids. Balee| 2] 5 E So) s}e8) a] a)ea).4a] 8 —=o)/25| 2/26] g A 2 a o|2| 28 Bo] 2/2 Seleeleleels |] ss) 3) eles/E/ 2 o 6] 5 218 a | & ej 2 |S /Fo|2) 8 a a nla n al42g|/n|/me/29|/a |e Bij 4 wildese as sahecns 1 1} 1 0] oO} o 0 0| O| o 2 a eases 1 i, 1 0 0 0 0 0 0 0 Fs, Fwildeea paces candso% 21 9| 9] 12 4) 4 1] Oj 1 7; 1] 6 Fa, py wild...........4. 44 10 | 10 34 4) 4 8; 8 0 22 6/ 16 Fs, os wild............. 27 2] 2] 25 0; 0 1} 1] O]/ 24} 6] 18 Fe, dy wild............. 4 0] Oo 4 1/1 0} O| 0 3/ 1] 2 Fy the Wild vases sevases 1 0| 0 1 0} of o| of o 1; Of 1 Offspring of hybrid males and females.......... 2 0} 0 3 2| 2 Oo}; 0; 0 1 0 i THRs so cence se 102 23 | 23 79 11} 11 10 9 1 58 | 14 | 44 TaBLE 77. Theoretical percentage of ultimate recessive individuals eapected in back-crosses, on the basis of various numbers of factors involved. Number of factors. Generation. 1 2 3 4 5 6 7 8 9 F,, 3 wild...... 00.00 | 00.00 | 00.00 | 00.00 | 00.00 | 00.00 | 00.00 | 00.00 | 00.00 Fi, 4 wild...... 50.00 | 25.00 | 12.50 | 6.25 | 3.13 | 1.56 | 00.78 | 00.39 | 00.20 F;, $4 wild...... 75.00 | 56.25 | 42.19 | 31.64 | 23.73 | 17.80 | 13.35 | 10.01 | 7.51 Fy, 7g wild..... 87.50 | 76.56 | 66.99 | 58.62 | 51.29 | 44.87 | 39.27 | 34.36 | 30.07 F;, gz wild..... 93.75 | 87.89 | 82.40 | 77.25 | 72.42 | 67.89 | 63.65 | 59.67 | 55.94 Fo, gz wild..... 96.88 | 93.85 | 90.92 | 88.07 | 85.32 | 82.66 | 80.07 | 77.57 | 75.15 Fy, zg wild... .| 98.44 | 96.90 | 95.39 | 93.90 | 92.43 | 90.98 89.56 88.16 | 86.79 Fs, fg wild....) 99.22 | 98.44 | 97.67 | 96.91 | 96.15 | 95.40 | 94.66 | 93.92 | 93.19 Fo, gtz wild....| 99.61 | 99.22 | 98.83 | 98.45 | 98.06 | 97.68 | 97.30 | 96.92 | 96.54 with guinea-pigs, and female hybrids with male hybrids. Offspring of female hybrids and guinea-pigs. Offspring of female hybrids and male hybrids. TABLES. 127 TaBie 78. Different combinations of matings which have been made. Females. Males. 3 wild. | } wild. | $ wild. | py wild. | dy wild. | g& wild. | 74, wild enw Bewild......cccceeceeeecheeeeeee x Saal ll Preece | eens errr x Se Wild lieaan dy al yeeaieadd lh gemini x x | Whee wel bee nis a x dy wild......... Se Nii deel na eed ih x x Ke eee ee x Swill ca esas feet Bee ea ea tale x x x x STIG richie nas eel cs sean Saas hl oeahcavecse a sll maubacaviouanlolta ddssuaua-a SvG)| lm aea- peace dase coca DR, leeuenaases Guinea-pig....... x x x x x x x x TaBLe 79. Percentages of hybrid male offspring with many motile sperm in matings of female hybrids Generation of females. Percentage Percentage Number. with many Number. with many motile sperm. motile sperm. Wig aera 1 00.0 1 00.0 Pigs aatanns 8 00.0 2 00.0 Pgs cas mun 49 14.3 7 14.3 Bias a sens 99 33.3 17 58.8 Biles sata sae 150 60.7 11 63.6 Poss soaaee 49 69.4 aN ecietc Posgasaves 15 73.3 1 100.0 TasLeE 80. Average number of young per litter in the wild and tame parents and in the hybrids. G ti Total Number of | Average eneramon: individuals. litters. | per litter. Wid cscshaid Bd aed ah eect 46 34 1.35 Tae): siisCo eae ets aoe es 484 207 2.34 Tame (Minot’s results)........ 366 143 2.56 Tame females by wild males... . 37 16 2.31 By, 201d. ok ssa ena caneeewnas 83 52 1.60 Fo, } wild... .....0..ceeeeeee 217 114 1.90 Fac Wildeas, cacaeiwwnaouad 312 152 2.05 Fg. fg. Wild ines He saeaewes ses 344 172 1.98 Fey. Se WHI. sen ces screesmeiy ees 122 60 2.03 Fan Ge Wild. can cc eens tes 36 19 1.89 128 GENETIC STUDIES ON A CAVY SPECIES CROSS. TABLE 81. Ratios of sexes in the hybrids. Number of Generation. Males. | Females.| Total. males to 100 females. F,, 4 wild....... 14 23 37 60.87 Fy, } wild....... 31 52 83 59.62 F;, 4 wild....... 101 116 217 87.07 Fy, dy wild...... 159 153 312 103.92 Fs, ge wild...... 173 171 344 101.17 Fo, gy wild...... 58 64 122 90.63 Fy, zig wild..... 16 21 37 76.19 Total..... 552 600 1152 92.00 BIBLIOGRAPHY. AcKERMANN, A. 1897. Tierbastarde, Zusammenstellung der bisherigen Beobachtungen. Abh. und Ber. des Ver. fiir Naturkunde in Kassell, II Wirbeltiere, Bd. 43, pp. 1-79. ALzEzals, H. 1903 Etude anatomique sur le cobaye. Paris, F. Alcan. 170 pp., 58 fig. Bateson, W. 1913. Mendel’s principles of heredity. Univ. Press, Cambridge, Eng. xiv + 413 pp., 38 fig., 6 pl. Baovr, E. 1911. Hinfiihrung in die experimentelle Vererbungslehre. Berlin, Gebriider Borntraeger. pp. 1v + 293, 80 fig., 9 pl. Bovp, M. M. 1908. A short account of an experiment in crossing the American bison with domestic cattle. Amer. Breeders’ Ass’n. Annual, vol. 4, pp. 324-331, 4 fig. 1914, ronan bison and cattle. Am. Journ. Heredity, vol. 5, No. 5, pp. 189-197, g. BRAINERD, E. 1907. The behaviour of the seedlings of certain violet hybrids. Science, n. s., vol. 25, pp. 940-944, Castie, W. E. 1905. Heredity of coat characters in guinea-pigs and rabbits. Carnegie Inst Wash., Pub. No. 23, 78 pp., 6 pl. ; 1906. The origin of a polydactylous race of guinea-pigs. Carnegie Inst. Wash. Pub. No. 49, 29 pp. 1907. On a case of reversion induced by cross-breeding and its fixation. Science, n. 8., vol. 25, pp. 151-153. 1907a. Color varieties of the rabbit and of other rodents: Their origin and inheri- tance. Science, n.s., vol. 26, pp. 287-291. 1908. A new color variety of the guinea-pig. Science, n. s., vol. 28, pp. 250-252. and WaLtTEeR, MULLENIX, and Coss. 1909. Studies of sale cees in rabbits. Carnegie Inst. Wash. Pub. No. 114, 68 pp., 4 pl. 1912. On the origin of an albino race of deermouse. Science, n. s., vol. 35, pp. Cusnot, L. 19038. L’hérédité de la pigmentation chez les souris. (2me Note.) Arch. de Zool. expér. et gén. (Sér 4), tome 1, Notes et Revue, pp. 33-41. 1904. L’hérédité de la pigmentation chez les souris. (8me Note.) Arch. de Zool. expér. et gén. (Sér. 4), tome 2, Notes et Revue, pp. 45-56. 1911. Les determinante de la couleur chez Jes souris, étude comparative. (7e Note.) Arch. de Zool. expér. et gén., tome 8, Notes et Revue, pp. 40-61. Darwin, C. 1876. The variation of animals and plants under domestication. New York, D. Appleton & Co. Vol. I, xir + 473 pp., 43 fig.; vol. II, x + 495 pp. DETLEFSEN, J. A. 1912. The fertility of hybrids in a mammalian species cross. Am. Breeders’ Mag,, vol. III, No. 4, pp. 261-265. 129 130 GENETIC STUDIES ON A CAVY SPECIES CROSS. Drinxwater, H. . 1908. An account of a brachydactylous family. Proc. Roy. Soc. Edinburgh, vol. 28, pp. 35-57, 11 fig. Doruam, F. M. ae 1908. A preliminary account of the inheritance of coat-colour in mice. Journ. of Genetics, vol. 1, pp. 159-178. East, E. M. 3 1910. A Mendelian interpretation of variation that is apparently continuous. Am. Nat., vol. 44, pp. 65-82. and R. A. Emerson. 1913. The inheritance of quantitative characters in maize. Neb. Agr. Exp. Sta. Research Bull. No. 2, 120 pp., 21 fig. and H. K. Hayss. 1911. Inheritance in maize. Conn. Agr. Exp. Sta. Bull. 167, 142 pp., 25 pl. 1912. Heterozygosis in evolution and in plant breeding. Bureau of Plant Ind. U.S. Dept. of Agr., Bull. 248, 58 pp., 8 pl. Emerson, R. A. 1910. The inheritance of sizes and shapesin plants. Am. Nat., vol. 44, pp. 739-746. FaraBer, W. C. 1905. Inheritance of digital malformation in man. Papers of Peabody Museum, Harvard Univ., vol. 8. No. 8, pp. 69-77, pl. 23-27. Focke, W. O. 1881. Die Pflanzenmischlinge. Berlin, Gebriider Borntraeger. Iv + 569 pp. GARTNER, C. F. 1849. Versuche und Beobachtungen tiber die Bastarderzeugung im Pflanzenreiche. Stuttgart. Geppgs, A. C. 1910. Abnormal bone growth in the absence of functioning testicles. Proc. Roy. Soe. Edinburgh, vol. 31, pp. 100-150. GreseE1, C. G. 1855. Die Saiugethiere. Leipsig, A. Abel. x11 + 1108 pp. Go.pscHminT, R. 1912. Erblichkeitsstudien an Schmetterlingen I. 1. Untersuchungen iiber die Vererbung der sekunddren Geschlectscharaktere und des Geschlechts. Ztschr.{f. Abstammungs- und Vererbungslehre, Bd. 7, pp. 1-62, figs. A-X. Guysr, M. F. 1909. On the sex of hybrid birds. Biol. Bull., vol. 16, No. 4, pp. 193-198. Hayes, H. K. 1912. Correlation and inheritance in Nicotiana tabacum. Conn. Agr. Exp. Sta. Bull. 171, 45 pp., 5 pl. HeErtwic, O. 1910. Die Radium Strahlung in ihrer Wirkung auf die Entwicklung thierischer Eier. (1 and 2) Sitsungb. Kén. Preus. Akad. der Wiss., Jahrg. 1910, pp. 221-233, 751-771. Horst, C. C. 1905. Experimental studies on heredity in rabbits. Linn. Soc. Journ., Zool., vol. 29, pp. 283-324. Kine, H. D. 1911. The sex in hybrid rats. Biol. Bull., vol. 21, No. 2, pp. 104-112. Iwanorr, E. J. 1911. Die Frucbtbarkeit der Hybriden des Bos taurus und des Bison americanus Biol. Centralbl., Bd. 31, pp. 21-24. Lunn, P. W. 1841. Blik paa Brasiliens Dyreverden for sidste Jordomvoeltning. Kon. Danske Vid. Selsk., Afhand. 8, pp. 282-283, 1 pl. BIBLIOGRAPHY. 131 Marsaatt, F. H. A. 1910. The physiology of reproduction. London, Longmans, Green & Co. xvit + 706 pp., 154 fig. Minor, C. 8. : : 1891. Senescence snd rejuvenation. Journ. of Physiol., No. 2, pp. 97-153, 3 pl. Moraay, T. H. 1911. The influence of heredity and of environment in determining the coat colors in mice. Ann. N. Y. Acad. of Sci., vol. 21, pp. 87-117, 2 fig., 3 pl. von Naruusivs, S. 1912. Der Haustiergarten der Universitat Halle. Hannover, M. und H. Schaper. 77 pp., E2 fig. NEsRING, A. 1889. Uber die Herkunft des Haus-Meerschweinchens. Sitzungsb. der Naturf. Gess. zu Berlin, No. 1, pp. 1-4, 4 fig. 1893. Uber Kreuzungen von Cavia aperea und Cavia cobaya. Sitzungsb. der Naturf. Gess. zu Berlin, No. 10, pp. 247-252. 1894. Kreuzungen von zahmen und wilden Meerschweinchen. Zool. Gart., vol. 35, pp. 1-6, 39-43, 74-78. NILLSON-EHLE, H. 1909. Kreuzungsuntersuchungen an Hafer und Weizen. Lunds Universitets Arsskrift, N. F. Afd. 2, vol. 5, No. 2, 122 pp. 1911. Kreuzungsunte:suchungen an Hafer und Weizen. Lunds Universitets Arsskrift, N. F. Afd. 2, vol. 7, No. 6, 82 pp. OssBorn, H. 1908. The habits of insects as a factor in classification. Ann. of the Ent. Soc. of Am., vol. 1, No. 1, pp. 70-84, 1 pl. PuHIuirs, J. C. 1912. Size inheritance in ducks. Journ. Exp. Zool., vol. 12, pp. 369-380. PrzIBRaM, H. 1910. Experimental-Zoologie: 3. Phylogenese Inklusive Hereditat. Leipsig und Wien, F. Deuticke. 315 pp., 24 pl. Pott, H. 1910. Uber Végelmischlinge. Ber. tiber den V Intern. Ornithologen-Kongress, Berlin. pp. 399-468, 5 pl., 3 fig. 1911. Mischlingstudien VI: Eierstock und Ei bei fruchtbaren und unfruchtbaren Mischlingen. Arch. fiir mikros. Anat., Bd. 78, Abt. II, pp. 63-127, pl. v-vu, 1 fig. Ronta, A. 1903. Uber Saugetierbastarde. Der Zool. Garten, Bd. xiv. SHouz, G. H. 1910. Hybridization methods in corn breeding. Am. Breeder’s Mag,, vol. 1, No. 2, pp. 98-107, 1 fig. Souras, I. B. J. 1909. Inheritance of color and of supernumerary mamme in guinea-pigs, with a note on the occurrence of a dwarf form. Reports to the Evol. Com. of the Roy. Soc., London. Report 5, pp. 51-79, 1 fig., 1 pl. Sranpruss, M. 1895. Handbuch der Paldarktischen Gross-Schmetterlinge fir Férscher und Sammler. Jena, G. Fischer. x11 + 392 pp., 8 fig., 8 pl. Tamas, T. 1911. Das Verhalten fluktuirend variierender Merkmale bei der Bastardierung. Extract du Recueil des Travaux botanique Néerlandais, vol. 8, Livr. 3 pp. 201-288, 3 pl. 132 GENETIC STUDIES ON A CAVY SPECIES CROSS. Tuomas, O. 1901, On mammals obtained by Mr. Alphonse Robert on the Rio Jordfo, S. A., Minas Geraes. Ann. and Mag. Nat. Hist., London, vol. 8, No. 49, pp. 526-539. bE Vaiss, H. 1906. Species and varieties, their origin by mutation. Chicago, Open Court Publ. Co. xvii + 847 pp. Waipowzvon Waal, H. 1907. Fruchtbare Maultiere. Jahrb. f. wissensch. und prakt.T iersucht, vol. 2, pp. LI-11Vv, 1 pl. WarteErHouseE, G. R. 1848. A nator History of the mammalia. London, H. Bailliére. Vol. II, 500 Pp., <1 pl. Puate 1. Fig. 1. 2. 3. Puats 2. Fig. 4. 7. 8. 9. DESCRIPTION OF PLATES. Pure wild Cavia rufescens 3 33. 4 wild hybrid (C. rufescens 7 XC. poreellus 2)9. 3 wild hybrid (3 wild hybrid 9° XC. rufescens 7) @. Mid-dorsal hairs of 4 wild hybrid. The agouti was received from the wild and is about the same shade asin the tame. In some cases the pure wild agouti was a trifle darker. . Mid-ventral hairs of 4 wild hybrid. Its relation to the pure wild and pure tame is as in fig. 4. . Mid-dorsal hairs of o’804 (#; wild hybrid), a typically modified, darkened agouti. The agouti was received from a pure wild strain. The ticking is very slight. Mid-ventral hairs of o804, Compare fig. 6. o'804, 34; wild hybrid. Compare figs. 6 and 7. Ventral view. Thesame animal. - Pate 3. Photographs of male skulls in parent species and hybrids. Fig. 10. 11. 1Z. 1, the original wild male ancestor of all wild and hybrid animals in these experiments. o'86, male guinea-pig. 10, 4 wild bybrid. 13. 9151, 3 wild hybrid. 14. 206, $ wild hybrid. Notr.—These and all other reproductions of skulls and bones are natural size, and as near the calculated averages as possible, unless otherwise stated. Prate 4, Photographs of female skulls in parent species and hybrids. Fig. 15. 16. 17. 18. 19, 93, the original wild female ancestor of all wild and hybrid animals in these experiments. 9 12656, female guinea-pig. 9 63, 3 wild hybrid. 9 87, 3 wild hybrid. 9 264, 3 wild hybrid. Prats 5. Photographs of lower jaw-bone in parent species and bybrids. Photographs of Fig. 20. 21. 22. 23. 24, 25. 26. 27, 28. 29. 30. 31. 32. 33. a wild male, a 3 wild male, and a 3 wild female skull. ol, the original wild male ancestor of all wild and bybrid animals in these experiments. o'24, pure wild son of 71. o'78, 4 wild hybrid. @111, 4 wild hybrid. 6169, } wild hybrid. o’617, male guinea-pig. 98, the original wild female ancestor of all wild and bybrid animals in these experiments. 9 22, 3 wild hybrid. 996, 4 wild hybrid. 9171, 4 wild hybrid. 930, female guinea-pig. o'24, pure wild son of 71. o'28, largest 4 wild male. 9119, 3 wild female. 133 134 GENETIC STUDIES ON A CAVY SPECIES CROSS. Puate 6. Photograpbs of humeri and femora in the parent species and hybrids. Fig. 34. #1, the original wild male ancestor of all wild and hybrid animals in these experiments. o'24, pure wild son of #1. 23, 4 wild hybrid. 151, 4 wild hybrid. 192, 4 wild hybrid. 086, guinea-pig male. 35. 93, the original wild female ancestor of all wild and hybrid animals in these experiments. 9 22, 4 wild hybrid. 990, 3 wild bybrid. 9 207, 4 wild hybrid. 9 87, guinea-pig, female. 36. 1, the original wild male ancestor of all wild and hybrid animals in these experiments. 924, pure wild son of #1. 23, 4 wild hybrid. 151. 4 wild hybrid. 128, 3 wild hybrid. 2304, guinea-pig male. 37. 93, the original wild female ancestor of all wild and hybrid animals in these experiments. 975, 3 wild bybrid. 9 87, 4 wild hybrid. 9 108, 4 wild hybrid. 9 12600, guinea-pig female. Piate 7. Photographs of scapule and tibie in the parent species and hybrids. Fig. 38. 24, pure wild son of #1. #10, 4 wild hybrid. 151, + wild hybrid. 126, 3 wild hybrid. 02034, guinea-pig male. 39. 93, the original] wild female. 9 69, 4 wild hybrid. 9 208, 4 wild hybrid. 9 87, guinea-pig female. 40. #1, the original wild male. 24, pure wild son of #1. 23, 4 wild hybrid. 0120, 3 wild hybrid. o'246, 4 wild hybrid. 012267, guinea-pig male. 41. 93, the original wild female. 975, % wild hybrid. 91138, 4 wild hybrid. 9 208, ¢ wild hybrid. 912601, guinea-pig female. Prare 8. Camera-lucida drawings of the nasal-frontal sutures in the parent species and 3 wild hybrids. Fig. 42. Tame guinea-pig, C. porcellus. 43. Wild guinea-pig, C. rufescens. 44, 4 wild hybrids. Piate 9. Camera-lucida drawings of the nasal-frontal sutures in the } and } wild. Fig. 45. 4 wild. 46. 4 wild. Pate 10, Fig. 47. Camera-lucida drawings of the nasal-frontal sutures in the ;', wild hybrids. DETLEFSEN PLATE 1. 1. Pure wild Cavia rufescens & 33. One-half wild hybrid (C. rufescens dx C. porcellus Gg), Three-quarter wild hybrid (% wild hybrid ?x C. rufescensd) ¢. The figures are about one-half life size. one-half wild hy bric irs O1 4, Mid-dorsal ha rid hyl 1/64 wild rs of one-half wild 5, Mid-ventralhai 9 1 { SOF of 6. Mid-dorsal hairs ybrid). h i GD PHOTOGRAPHS OF FEMALE SKULLS IN PARENT SPECIES AND HyBRIDS 15. Q 3, original wild female ancestor of all 17. 2 63, one-half wild hybrid , and hvbrid animals these i OF . wild and hybrid animals in thes¢ 18. 2 87, one-quarter wild h 2xperiments. ; apa 2 E He} 19. 2 264, one-eighth wild hvb 16. 2 12656, female guinea-pig DETLEFSEN PHOTOGRAPHS OF LOWER JAW-BONE IN PARENT SPECIES AND Hvsrips. PHOTOGRAPHS 0 NILD MALE, AND A ONE-QUARTER WILD FEMALE § The detailed descriptions of the above figures will be found on page 133 HYBRIDS DPECIES AND THE PARENT PHOTOGRAPHS OF HUMERI AND FEMORA IN figures will be found on pag criptions of above les The detail *~HOTOGRAPHS PLATE 8 DETLEFSEN ‘spriqAy py HeeuO ‘TP ‘suoosoyn “> “SId-vouInd plIM “EP “snyjeoiod °C ‘sid-eeuind curvy, ‘Zp ‘spriqdy pia & pue soioeds quored oy} Ul Soungns [vjUOIJ-[esvU oY} JO SSULMBIP EpPlony-B1ouIED) YAR) vee aTIM vay pram aqTIM= ue rele ey ne a Waianae ‘ayy Nay eae INVA [Yoel e 1 \ CTeeodle \ 2 (Tal) egszil 5 wl isall 3 ( Jose 1 90821! S atozil 9 ar) erry pes a (Kale (Toy oor [Tee ran (ate T] Tro (Pool OY (Pacey (Tele) eS (Toul 1) (yatet) coal “siz Pree is6t} 9 TIE