STUDIES OF INHERITANCE IN GUINEA-PIGS AND RATS BY W. E. CASTLE AND SEWALL WRIGHT PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON WASHINGTON, 1916 CARNEGIE INSTITUTION OF WASHINGTON, PUBLICATION No. 241 PAPER No. 26 OF THE STATION FOR EXPERIMENTAL EVOLUTION AT COLD SPRING HARBOR, NEW YORK FROM THE LABORATORY OF GENETICS OF THE BUSSEY INSTITUTION Copies of this Book were first issued SEP 20 1916 f 0 *f U PRESS OF GIBSON BROTHERS, INC. WASHINGTON CONTENTS. PART I. — AN EXPEDITION TO THE HOME OF THE GUINEA-PIG AND SOME BREEDING EXPERIMENTS WITH MATERIAL THERE OBTAINED. BY W. E. CASTLE. PAGE. Introduction Some observations on guinea-pigs in Peru 5 Hybridization experiments with Cavia cutleri 8 Life history of C. cutleri 8 Crosses of C. cutleri males with guinea-pig females 13 Color inheritance among the F2 hybrids 13 (a) Cross 9 albino (race B) X cf C. cutleri 13 (6) Cross $ albino (race C) X c?1 C. cutleri 14 (c) Cross 9 brown-eyed cream (race C) X cf C. cutleri 16 (d) Results from (6) and (c) combined 16 (e) Intensity and dilution among the hybrids 17 (/) Significance of the results observed 18 Hybridization experiments with a race of feral guinea-pigs from lea, Peru 20 Origin and characteristics of the lea race 20 Crosses between the lea race and guinea-pigs of race C 23 The F2 generation 24 Summary on the lea race 29 Hybridization experiments with a domesticated guinea-pig from Arequipa 31 c? 1002 and his Ft offspring 31 F2 offspring of 2 2 One black with white foot. 505 504 ] 4 505 507 * J 5 One spotted with red and with white. 505 605 1 1 . . 533 529 f » 1 . . 533 540 t 1 One spotted with white. To -.al 3> $ 18 2 C. Both parents silver agouti (red-eyed). Silver Silver Father. Mother. agouti Father. Mother. agouti young. young. 565 527 5 569 587 and 588 8 565 528 8 602 60S 1 565 573 1 798 701 1 565 593 2 798 872 1 565 601 and 604 3 565 607 1 Total 31 The question at once arises whether the stock obtained by me from lea was really a feral stock, in origin like the animals described by Von Tschudi, or whether they were present-day domesticated animals concerning whose origin I was deceived. Since I did not myself see the animals captured or see similar animals running at large and did not even visit lea, I can make no positive statement as to their feral origin, but I believe the report made to me by the agents of W. R. Grace & Co., that they were caught feral in the neighborhood of lea, to be correct for the following reasons: (1) The animals were placed 22 INHERITANCE IN GUINEA-PIGS. on board our steamer during a stop made in the night at Pisco, the terminus of the short line of railway which leads down from lea to the coast. I found them the next morning in the " butcher-shop, " con- signed from lea to W. R. Grace & Co. in Callao. I conclude that they really did come from the neighborhood of lea. (2) I saw no domesti- cated guinea-pigs in Peru which were self-colored like these animals. All domesticated ones which I saw in Peru, except albinos, were spotted with white, or with yellow, or both. Self varieties are not fancied in Peru. Varieties of this sort are not uncommon among the guinea-pigs kept by European and American fanciers, but apparently they have been established only by careful and long-continued selection from the pied stock originally introduced from South America. (3) The spotting and the rough character which have cropped out as recessives among the descendants of the three lea animals are feebly expressed characters which appear to have been almost obliterated, but which still come to expression feebly under inbreeding. This is what we should expect to find in a feral race acted upon by natural selection, conspicuous variations like spotting tending to disappear. TABLE 13. — Parentage of pure lea animals whose mating s are recorded in table 12. Individual. Father. Mother. Individual. Father. Mother. Individual. Father. Mother. cf 5011 9527 505 507 9601 565 527 95021 9528 505 503 d"602 565 527 95031 9529 505 503 9604 505 503 9504 501 502 or 503 9530 505 503 9605 505 503 c?505 501 502 or 503 c?533 505 544 9607 565 528 9507 501 502 9540 505 509 9608 565 528 9509 501 503 c?565 505 507 9625 533 509 9510 501 503 0*569 505 504 9701 565 601 or 604 9573 505 502 c?798 569 587 or 588 9587 505 507 9827 798 701 9588 505 507 9593 505 504 'Original stock. The 3 original lea animals or their inbred descendants mated inter se have produced 114 young, of which 62 have been golden agoutis, 49 sil- ver agoutis, and 3 blacks; 4 of the 114 have shown a small amount of white spotting, 3 have shown yellow spotting, and 1 has shown a small amount of roughness of the coat. The various matings which have produced these young are classified in three groups in table 12, and the parentage of each animal which took part in a mating is shown in table 13, from which pedigrees may readily be drawn tracing back to the original trio. It will be observed from table 12 that silver agouti was derived from golden agouti as a recessive and has bred true without exception (31 silver agouti young being produced by silver agouti parents). GUINEA-PIGS FROM ICA. 23 CROSSES BETWEEN THE ICA RACE AND GUINEA-PIGS OF RACE C. An albino male guinea-pig (cf 54) of race C was mated with 5 golden agouti females of the lea stock. It was hoped from this cross to learn as promptly as possible the gametic composition of the lea race, since race C contained a larger number of recessive Mendelian factors than any other race in the laboratory. In this hope we were not disap- pointed. Race C has already been described. It contains two different recessive variations of the color factor, dilution and albinism, which are allelomorphic with each other and with ordinary color, thus forming a system of triple allelomorphs, C, Cd, and Ca, with dominance in the order named (see Wright, 1915). It lacks agouti, black, and extension factors. Visibly the animals of this race are either brown-eyed cream or albino. Male 54 was an albino, bearing the color allelomorph Ca, which is recessive to the color allelomorph C^ found in brown-eyed cream individuals of race C. The mating between d71 54 and the 5 golden agouti females of the lea race produced 13 young, 7 of which were TABLE 14. — Fi result of mating the albino <^54 of race C ivith golden agouti females of the lea race. Dark-eyed Red-eyed young. Mother. young. Golden agouti. Silver agouti. Sepia. 502 3 1 503 1 1 1 504 . . . . 2 507 1 1 . . 509 2 Total . . . 7 3 3 golden agouti (like the mothers), 3 silver agouti, and 3 a dull black or slate color, which will be called sepia. The silver agouti young were like those produced by lea animals bred inter se. The sepia young represented a new class not previously observed. In common with the silver agoutis they had no yellow in their fur. The ticking and spotting of silver agoutis was of white, as was also the spotting of the sepias, which had no ticking. It seemed probable, therefore, as proved to be the case, that the silver agouti and the sepia young differed from each other only in the presence or absence of the agouti factor. But these two classes of young taken together differ from golden agoutis in lacking yellow pigmentation with which the golden agouti fur is ticked. They also differ from golden agoutis in the intensity of the eye pigmentation, which is very great in golden agoutis and blacks, but ordinarily shows such reduction in silver agoutis and sepias that the eye by reflected light has a deep red glow. It will be convenient to distinguish them as red- 24 INHERITANCE IN GUINEA-PIGS. eyed, it being understood that the red eye is invariably associated with no-yellow in the coat. Four of the five lea mothers which were mated with of 54 had pro- duced silver agouti (red-eyed) young by lea mates. Each of them produced red-eyed young by c? 54 ; together they produced 5 dark-eyed young (golden agouti) and 6 red-eyed (silver agouti or sepia). The fifth lea mother (9509) had produced 11 golden agouti young when mated with lea males known to be heterozygous for silver agouti. (See table 12.) This is good evidence that she did not carry red-eye as a recessive character and was accordingly homozygous for dark-eye. By 6*54 she produced 2 golden agouti young. From these several facts it appears that dark-eyed lea animals capable of producing red-eyed young when mated inter se, produce equal numbers of dark-eyed and red-eyed young when mated to albinos, but produce no albinos. This indicates that albinism is recessive both to red-eye and dark-eye, an indication which the F2 result confirms. It will be shown further that the three conditions are mutually allelomor- phic, so that a zygote may contain any two of the three, but not more. Red-eye is in fact a fourth member of the albino series of allelomorphs, which includes the following conditions in order of dominance : (1) ordi- nary dark-eye and colored coat, such as is seen in Cavia cutleri and in golden agouti animals of the lea race; (2) dark-eye with dilute coat, seen in colored animals of race C; (3) red-eye and non-yellow coat; (4) albino. (See Wright, 1915.) For convenience these allelomorphs may be designated by C, Cd, Cr, and Ca. The cross of lea females with the albino cf 54 involves animals of the formulae CC or CCr mated with an animal of the formula CaCa. The 7 golden agouti young are expected to be of the formula CCa; the 6 red-eyed young of the formula CrCa. We may now compare the experimental with the expected results of breeding such animals in various ways. THE F2 GENERATION. One of the Fx silver agouti males (c?517) was known from his pedi- gree to be heterozygous in four characters, viz, red-eye vs. albinism, agouti vs. non-agouti, black vs. brown, and extension vs. restriction. His formula was accordingly CrCaAaBbEe, and we should expect him to form gametes of 16 different sorts, all equally numerous. This animal was mated with all three kinds of F: females, with the results shown in table 15. The golden agouti females produced 25 young, distributed among 10 classes very distinctly different in appearance. These golden agouti females were known from pedigree to be hetero- zygous for the same 4 factors as cf 517, but to contain a different allelo- morph for albinism. Both he and they carried albinism as a recessive character, but whereas he carried red-eye (Cr) as its dominant allelo- morph, they carried dark-eye (the ordinary condition of the color GUINEA-PIGS FROM ICA. 25 factor, viz, C). His gametes accordingly could transmit either Cr or Ca, but theirs would transmit either C or Ca. Accordingly their young should be in the ratio 2 dark-eyed to 1 red-eyed to 1 albino . The observed numbers were 13: 8: 4. Each of these three groups might theoretically contain 8 different kinds of individuals, but certain of these would be visibly indistinguishable. The classes visibly different which might themselves be expected to be composite are black-eyed reds and brown- TABLE 15.— Fi young from the cross 7 4 1 1 1 3 2 1 3 1 2 1 1 1 yellow agouti .... 3 sepia (dark-eyed) . 2 silver agouti . 2 sepia (red-eyed) . . 2 non-agouti (pink- eyed) 2 agouti (red-and- pink-eyed) Total 4 1 16 G 10 8 7 4 5 0 By pink-eyed daughters, c? 1002 has produced 7 pink-eyed young and 8 with eyes not pink — complete agreement with the expected equality. By daughters not pink-eyed, but which nevertheless are clearly hetero- zygous in pink-eye, he has produced 23 pink-eyed and 58 not-pink-eyed young; expected, 20 and 61 — an excess of pinks capable of explanation on the same ground as the excess of red-eyed young. By dilute-colored daughters cf!002 has produced 52 dilute-colored young, but no intense-colored ones, as expected, since dilution is recessive to intensity. By intense-colored daughters heterozygous for dilution he has produced 10 intense and 9 dilute young, equality being expected. MISCELLANEOUS MATINGS OF THE DESCENDANTS OF cM002. Matings of the descendants of cf 1002 beyond the F2 generation were made chiefly with a view to test further the genetic character of the new varieties. Their results are presented in tables 23 to 28 and serve to confirm the interpretations already offered. GUINEA-PIGS FROM AREQUIPA. 37 Matings of red-eyed animals inter se have in most cases produced only red-eyed or albino young, but two matings have also produced pink- and-red-eyed young, i. e., animals which are pink-eyed but develop no yellow in their fur, in which last respect they differ from ordinary pink-eyed and agree with ordinary red-eyed. (See tables 24 and 27.) TABLE 23. — Young produced by matings of red-eyed males, descended from tflOO'2, with dark-eyed females of race B. Nature of mating. Dark-eyed. Red-eyed. Albino. Golden agouti. Black. Silver agouti. Sepia. Silver agouti X black (homozygous) .... Silver agouti X black (heterozygous in albinism) .... 1 20 5 7 6 9 7 10 Sepia (red-eyed) X black (homozygous). . Total 21 IS 9 7 10 Most of the red-eyed animals, when bred inter se, produce albino as well as red-eyed young, showing themselves to be heterozygous for albinism and so of the formula CrCa. This is not surprising when we recall that all the F! red-eyed animals must by hypothesis be of this formula, and that two-thirds of the F2 red-eyed should be of the same sort. In a few matings of red-eyed with red-eyed, which failed to produce albino young (table 24) , it is probable that one or both parents TABLE 24.— Young produced by matings inter se of red-eyed descendants of& 1002. (See also table 27) . Nature of mating. Red-eyed young. Albino young. Silver agouti. Sepia. Both parents agouti 56 18 19 6 9 22 11 Only one parent agouti Neither parent agouti . . . Total 74 34 33 were homozygous for red-eye. Matings of red-eyed with albino animals (table 25), which failed to produce albinos in 6 or more young, afford clear criteria for red-eyed animals free from albinism and so of formula CrCr. Only one mating of a red-eyed animal with an albino has pro- duced pink-eyed young. (See table 27.) The red-eyed parent in this case (cf 48) was mated with 4 other albinos (all of race B) without producing pink-eyed young, but only red-eyed (13) and albinos (17). The female which produced pink-eyed young was his sister, derived like 38 INHERITANCE IN GUINEA-PIGS. himself from parents known to transmit pink-eye. This indicates that the character pink-eye in guinea-pigs (as in mice) may be transmitted by albinos. The fact should be emphasized that the pink-eyed young TABLE 25. — Young produced by red-eyed descendants of c?1002 mated with albinos. (See also table 27.) Nature of red-eyed parent. Red-eyed young. Albino young. Silver agouti. Sepia. Silver agouti, heterozygous for albinism 21 11 31 11 20 17 34 Sepia, heterozygous for albinism Silver agouti (homozvgous for red-eye) 18 Sepia (homozvgous for red-eye) Total . . . . 39 73 51 produced in this mating were also red-eyed, i. e., were non-yellow, for red-eyed animals may carry pink-eye as a recessive character, and con- versely pink-eyed may carry red-eye as a recessive character. How- ever, if these recessive characters crop out as recessive individuals from a mating of two like parents with each other, it can in either case occur only in the form of the double recessive, both pink- and red-eyed. TABLE 26. — Young produced by pink-eyed descendants of tfl002, mated inter se. Natuie of mating. Pink-eyed. Pink-and- red-eyed. Albino. Agouti. Non- agouti. Agouti. Non- agouti. Both parents agouti 15 4 2 4 1 1 7 One parent agouti, one non-agouti . . . Total 19 6 1 0 8 Pink-eyed animals (with yellow in their fur) have made their appear- ance as recessives produced by mating dark-eyed animals inter se. (See tables 20, 22, and 27.) In some cases red-eyed young have been produced by the same matings, or pink-and-red-eyed or albinos, for pink-eye seems to be quite independent of the color factor in its inherit- ance. Pink-eyed animals mated inter se have produced only pink-eyed, pink-and-red-eyed, and albino young. (See table 26.) Any of these three forms so derived will doubtless be found to transmit pink-eye in every gamete. Pink-and-red-eyed animals of whatever origin have been found to produce (when mated with each other) only pink-and-red-eyed young or albinos. But the record as regards albinos is doubtful. Two GUINEA-PIGS FROM AREQUIPA. 39 albino young have been recorded as produced by cf 88 mated with his daughters, 9204 and 9205; but this same male mated with albino females of race B produced 1 1 red-eyed young but no albinos, for which reason it seems very doubtful whether he transmits albinism. More probably the two young by pink-and-red-eyed mothers were not albinos, but very pale-colored non-yellow young, possibly lacking the extension factor, in which case their fur would be pure white, their eyes being uncolored because of the pink-eye factor. If so, they would be in appearance indistinguishable from albinos, though behaving very differently in crosses. TABLE 27. — Matings of descendants of &1002 which have produced pink-eyed young. Nature of mating as regards — Dark-eyed young. Red-eyed young. Color factor. Agouti factor. Golden agouti. Black. Silver agouti. Sepia. Dark-eye X dark-eye. . Dark-eye X red-eve . . Agouti X agouti 4 2 1 8 2 4 5 Do Do Agouti X non-agouti 3 Red-eye X red-eye Agouti X agouti Red-eye X albino Non-agouti X non-agouti . . Total 7 2 9 11 Nature of mating as regards — Pink-eyed young. Pink-and-red- eyed young. Albino young. Color factor. Agouti factor. Agouti. Non- agouti. Agouti. Non- agouti. Dark-eye X dark-eye . Dark-eye X red-eye Agouti X agouti 1 1 1 3 2 7 5 Do 2 1 Do Agouti X non-agouti .... Red-eye X red-eye .... Red-eye X albino Agouti X agouti Non-agouti X non-agouti . Total 3 2 4 2 12 Nevertheless, it is to be expected that pink-and-red-eyed animals can be produced which are heterozygous for albinism. Such animals necessarily would be heterozygous for red-eye also, which is an allelo- morph of albinism, and so would be of the formula CrCapp, for it is known (1) that pink-eyed animals may transmit albinism; (2) that red- eyed animals may transmit albinism; and (3) that pink-eye and red-eye are independent of each other in transmission. Consequently, there is every reason to suppose that albinism may exist as a recessive allelo- morph of red-eye in animals which are both pink-eyed and red-eyed. A pink-eyed animal mated with a dark-eyed one produced 3 dark- eyed young and 1 albino, which adds to the evidence that pink-eye is recessive to dark-eye and may be present in the same zygote as albin- ism. A pink-eyed animal (9307) mated with a pink-and-red-eyed 40 INHERITANCE IN GUINEA-PIGS. male (cTl40) produced 2 pink-and-red-eyed young. A pink-and-red- eyed male mated with 2 dark-eyed females of race B, which were hetero- zygous in albinism, produced 4 dark-eyed and 4 red-eyed young. Another pink-and-red-eyed male (140) mated with a dark-eyed female descended from cf 1002 produced 1 pink-eyed young, in which red-eye was undoubtedly recessive. Most of the foregoing matings are tabu- lated in table 28. TABLE 28. — Character of young produced in matings of pink-and-red-eyed descendants of ' '*< ^J rH SO ng 0*9 JM rH X t** ^ C^ O *H t-~. rH O *3* rH t*- O iOiO»OiO iOcOCO*O COCO COCO »OCO X X (N (N iO 10 CO CO '£ -o •^ WOi i^p 'jO ^t1 cO t^- n cO ^r ou rH t*» co x r*- t o x o -H rH rH -«s fr 2.9-Q'LQ w *e-a & r< 6 99-9 ' 99 : •"" ; • • : : : rt f 99-0 '99 b' "W 6 99-9 ' 99 : : : : : -* rt § e f ' 99-0 ' 99 : : : : : ^^ rH 3. 6 ' ^9-9 ' TO rH CO • CO a, i f TO-O'TO • rH i-H IO 6 S9-9 ' 29 • rH rH (M Tj< CO >O 03 f Z9~0 ' Z9 : : : : •* N : : M : *° -1 8 e 6 19-9 19 • t^- C>) CO »O rH O5 • 3 ^ 09-0 ' 09 • rH rH • O5 Ol ' IN ' »O ' ' ^ • rH •& & % 6 ' 69-9 ' 69 •CON • OS CO I-H rli -l> iMC ^ CO • rH "e s **-. T3 fr ' 69-0 ' 69 • Tj* ^* • • t>» rH rH • ^H • C*l T-H • r H • rH e •ea 03 6 ' 89-9 ' 89 • co o • co • •<* -co • r 0 !- 53 t2 ,—• , 00 r89-0'89 •COTfl •COrH'* --f rHrHr H ... 3 VJ t$ 43 6 Z9-9'Z9 • 1-H CO 'CO ' a * 99-0 ' 99 . rt 0. (M ; ; ; o> mrr Olfl) OO OC rMr^ Ut3 " '~ c3rf 03 ^ O4 £2 '-'-xxxx 22 XX XX Vy o- cr a- CT A ^ 'C ffl -5 -g "SPQ-g-g S£ £2 ess •** O •*"* u o3 3 flj ^ N r^OS""15^ <-!« — "CM C3 *" ' C oooo rb'bpbrb CM-O ^b1^, ooo i do t-C t-l o o o> a> o o o3 o3 : xx S g , j^J 'b 'b'b 50 INHERITANCE IN GUINEA-PIGS. TABLE 30.— A tabulation of skull width measurements of Cavia cutleri and of certain races of guinea-pigs and of their hybrids. •uop -13lA9p pXBpU'B^g 1C CO O 1C O O O 1C O 1C »CO OO CO 1C CO -^ COO5t>O -* O Ci •<*< COCO 0 000 t^ Tf rH OO O5 O rH rH COrHrHlM OJ^t1 rHrH Oi CO rH rH rH rH rH rH rH rH rH 00 CO CM rH rH rH rH _* Tfco^fco co co o: ••* r-ic ot^- xco COCOCMCM COCOt^CM COCO -^CO CMrH CO O CO CO CO O CN t- O^fllCCO rHCOt-lC COt^- OJOO COt-» cocococo cocococo coco coco coco lO 00 OJ OO CO CO CO CO TWJOX OCO^CO t^COCC^f COrH t^«CO OOt^ rH C>J IM CO CO CM C) rH Tf CM 1C ^ o co r*- rH rH rH Classes (in millimeters) and frequencies. Q-Z*-S-Zf • • • • • • • • • • • • • • • • • fZWZ* Q-l^-l* . . . . . . . . . . . ... ... flf-0'lf CM CM rH 6'(»HrO* rH CO CO • rH fof-o-of- rH • CO ^ rH CN rH 6'68-9'68 COM* ^^ rH rH CO £ ' 68-0 ' 68 — * *o i>- »— 1 CM CM CO *• 88 -0-88 i— IIO COlO rHt^- rHrH rH CM 6-Z8-9-Z8 -H OOlOC^ i-HrH lOO • T— ' rH '•CM *-Z8-0'Z8 rH C^C^rH dlC CO rHC"lt-H r-trH 6 '98-9 "98 CM CM rHC^rH (Ml> rH(M(N rH ^98-0-98 CM rH CM CO ^ rH O rH r- rH rH CO CO 6 '98-9 "98 rH CO i— t^H»O (M C^lrHrHrH rH ^•98-0-98 oo co CD 1-1 c co iC ^8-0-^8 T}H rH CM 6 88-9 ' 88 •^ CO 1C f ' 88-0 ' 88 CO • O3 6 £8-9 '28 rH rH rH CO r- Y se-o'se • rH rH • ... 6 '18-9 '18 CN rH C* riS-O-IS 6 -08-9 '08 CN rH C ^08-0-08 CN 6-6S-9-6S CN 0 1 OO OO QQ Q(J f-l t-i h r* • • , o o pq pq pq CQ pqpq pqpq & & ga> pqpq pqpq 22 22XXXX 11 11 xx xx ^ ^ xx xx ^3 j3 3 p ^jj ^3 o o o o -+3 u -*^ o c3 o3 oooo fbrbpbrb oo 'b'b ooo 'b'b'b SIZE. 51 1 o» •5, e 1 t- 8 -e c o CO H A n •< H •uop iO O *O O O O O O *O O ^O O ^O *O *O O O O OtOCOCD C^Jt^OOtN COCO OOO N-OOO ^HCDQO -1?tA8 3 pJBpUtJ^g C— 1 o a CD 3 *-Sfr-0-g* ^^1 ^^ „,„ Nb. ^^l ^^^ £ £ 6-SF-(H t7 UF U Uf tn a> m 6 '68-5-68 CO CM CO • rH rH • rH • rH 03 • co_n • «o o ob-0 ob 6' QO O ' Of », • op— ri • op O _'- V Z«o rH C1! _ . . r>_f, • f p ' v z,b-u z,b 6' OP O ' OP »o a y& T? * OP — n ' OP v yb^j yb Q'OP D ib i ib • CD 0 , an assemblage of unanalyzed factors which determine white spotting. INTENSITY OF GENERAL COLOR DEVELOPMENT (I B). In this group fall albinism and its variations. These factors affect all color, but not wholly irrespective of the kind of color. There are several peculiarities which are discussed more fully in a later section (page 70) . The most important is the fact that the level of intensity of the color factor at which yellow can develop at all is higher than the threshold for black or brown. This does not affect the differentiation of the fur into yellow and dark pigmentation areas by factors of group 2, but involves the result that with certain albino series factors, yellow areas appear white, while dark areas are quite strongly colored. Indeed, in albinism itself, dark pigmentation areas can often be distinguished from yellow areas by a slight sootiness in the former, absent in the latter. C. Determines the highest intensity of color of skin, fur, and eye which is to be found with a given array of other factors; dominant over Cd, Cr, and Ca, where distinguishable in its effects. In the following table, and in the similar tables under Cd, Cr, and Ca, are given the ranges of intensity in the yellow, black, and brown series to which these colors develop when the factor under consideration is present. In the case of black and brown, factor P is assumed to be present. When p is present, black and brown undergo a two-fold dilution. P is also considered present in the case of eye-color. Yellow series — redo to yellow2 in guinea-pigs; yellows to creamg in Cavia cutleri. Black series — blacko to black2. Brown series — browno to brown2- Eye color — black, brown. COLOR. 65 Cd- Determines an intensity of yellow distinctly lower than does C, an intensity of dark pig- mentation usually, but not always lower than does C, and an intensity of eye color rarely distinguishable from that determined by C. More or less dominant over Cr and Ca where distinguishable. (Wright, 1915.) Yellow series — yellow2 to creamy. Black series — blacko to sepiay. Brown series — browno (?) to browny. Eye color — black, brown. Cr. Determines the complete absence of yellow, an intensity of dark pigmentation indis- tinguishable from that determined by Cd and an intensity of eye color lower than that determined by C or Cd- More or less dominant over Ca where distinguish- able. (Castle, 1914a; Wright, 1915.) Yellow series — white. Black series — blacko to sepias- Brown series — browno (?) to brownj. Eye color — red, brown-red. Ca- Determines an absence of pigment, complete with yellow, not quite complete with dark pigments of the fur and skin, but complete in the eyes. (Castle and Allen, 1903; Castle, 1905; Sollas, 1909; Detlefsen, 1914; Wright, 1915.) Yellow series — white. Black series — white, dark smudges on nose, ears, and feet. Brown series — white, brown smudges on nose, ears, and feet. Eye color — pink. DARK VS. YELLOW COLOR (2). Factors of this group affect skin and fur color, but not eye color. In this group come the factors responsible for self yellows, tortoise-shells, and brindles, on the one hand, and self blacks or browns on the other, as contrasted with the ticked or agouti patterns of the wild rodents. Where more than one factor is present which determines a yellow pattern, combination effects are produced, such as in yellow-spotted agoutis among guinea-pigs. The following factors are known in guinea- pigs: E. A condition for more than a trace of dark pigmentation in the fur; determines dark pig- mentation wherever yellow is not determined by other factors; dominant over e, found in the wild species, all agoutis, blacks, browns, etc., but very rarely in self yellows. e. Determines the presence of one of the yellow colors in all colored areas of the fur, aside from a slight sootiness; responsible for the yellow in most self yellows, for the white in red-eyed whites, etc. (Castle, 1905, 1907, 1907a; Sollas, 1909; Detlefsen, 1914.) A. Determines the presence of a yellow color in the light-bellied agouti pattern wherever there is dark pigmentation in which the yellow group ticking may show; dominant over A' and a, found in Cavia cutleri and light-belh'ed agouti guinea-pigs, includ- ing the red-eyed silver agoutis, in which the agouti pattern is in white. A'. Determines the presence of yellow colors in a more restricted agouti pattern than does A, a pattern usually characterized by a ticked belly not sharply distinct from the sides in color; dominant over a, found in Cavia rufescens and in ticked-bellied agouti hybrids which have rufescens ancestry. (Detlefsen, 1914.) a. Determines the absence of yellow group ticking in hairs of dark pigmentation; found in blacks, browns, etc. (Castle, 1905, 1907, 1907a, 1913; Sollas, 1909; Detlefsen, 1914.) Sj/. An assemblage of unanalyzed factors which determine the presence of spots of a yellow color, conditional on factors of group (1) ; found in black and yellow tortoise-shells, black, yellow, and white tricolors, and in some red-eyed black and white bicolors; probably responsible for an occasional self yellow, though never in the writer's experience. 66 INHERITANCE IN GUINEA-PIGS. VARIATIONS OF DARK COLOR (3) Factors of this group are responsible for browns and pink-eyed sepias, as compared with blacks, in guinea-pigs; for browns and pink- eyed sepias in mice, and for the new pink-eyed and red-eyed dilute variations in rats. Where more than one factor of this group or of group IB determines dilution, combination effects are produced. Thus we have very pale sepias resulting from the combined effects of two independent dilution factors jB. Determines a color of the black -sepia series wherever dark pigmentation develops, including the eyes; has no influence where yellow pigmentation develops; domi- nant over b, present in the wild species and in blacks, sepias, albinos with black points, black-eyed yellows, etc. b. Determines a color of the brown series wherever dark pigmentation develops, including the eyes; has no influence where yellow pigmentation develops; present in browns, brown-eyed yellows, etc. (Castle, 1907a, 1908; Sollas, 1909; Detlefsen, 1914.) P. A condition for intense development of dark pigmentation in the fur and for eye colors more intense than pink; not necessary for intense development of yellow; domi- nant over p. p. Determines a low development of dark colors, i. e., below sepiag; has no influence where yellow develops; determines pink eye color. (Castle, 1914a.) TABLE OF FACTOR COMBINATIONS. In determining the color which corresponds to a given array of factors the groups of factors must be considered in the order given. Table 33 gives a list of the color varieties corresponding to the combinations of Mendelian factors. At the top and left of the table are indicated, by symbols, the factors present in each of the varieties named in the body of the table. The color of spots produced by Sw and Sy are given below. Only the varieties marked with an asterisk have not yet been synthesized. These include the pink-eyed yellows and creams and a kind of pink-eyed white which is expected to be indistinguishable from an albino in appearance, though breeding wholly differently. The pink- eyed brown series (bbpp) has not yet been produced and is not included. Some of the varieties have names given by fanciers which have been used in the literature. In this table, however, it seemed best to use a consistent scheme of naming, indicating at once the color and pattern. Agouti is used as the name for a pattern, the banding of hairs of pre- dominantly a dark color with a yellow color. The names preceding agouti give the two colors in each hair. The following table of syno- nyms may be useful : Black-red agouti = golden agouti. Sepia-yellow agouti = yellow agouti. Sepia-cream agouti = silver agouti. Brown-red agouti = cinnamon. Brown-cream agouti = light cinnamon. Sepia = blue. Brown = chocolate. Brown eye = brown eye (Castle), ruby eye (Sollas). COLOR. TABLE 33. 67 Factors Fur. Eve present. EA (agouti light-belly). EA' (agouti ticked-belly). Eaa. ee (A, A' or aa) . B P C Black-red agouti Black Red Black. CdCd Dark sepia-yellow agouti Dark sepia Yellow Do. CdCr. Dark sepia-cream agouti . Do Cream . . . Do CdCi Light sepia-cream agouti Light sepia . Do . Do CrCr Dark sepia-white agouti Dark sepia White (light Red Light sepia-white agouti Light sepia points) . Do Do CP White (dark points) White (dark ... Do . . Pink. Bno C Pale sepia-red agouti points) . Pale sepia Red Pink- CdCd Verv pale sepia-vellow agouti Very pale sepia *Yellow Do CdCr. Very pale sepia-cream agouti. Do *Cream Do CdCa. . Do Do . . .*Do Do. CrCr Very pale sepia-white agouti Do *White Do CrCa... .Do Do ...*Do Do White (light points) White (light . . *Do . . . Do bbP C . . . . Brown-red agouti points). Brown Red Brown CdCd . • Medium brown-yellow agouti . Medium brown . Yellow Do Medium brown-cream agouti . Do Cream Do CdCa. . Light brown-cream agouti .... Light brown. . Do Do CrCr . . . Medium brown-white agouti . . Medium brown White Brown-rod CrCa... Light brown-white agouti .... Light brown .... Do Do. White (It. br. points) White (It. br. . . .Do Pink points). Factors present. Sw. Sy- SwZy. Eye. C. . White spots (clear) Red spots Red and white tri- CdCd . . Do Yellow spots color. Yellow and white CdCr. . . Do Cream spots Cream and white CdCa. . Do . . Do . . tricolor. Do CrCr... Do \White spots, Sooty and clear CrCa... . . .Do / often sooty. white spots (Albino) (Albino) (Albino) HEREDITARY FACTORS AND THE PHYSIOLOGY OF PIGMENT. The definitions which have been given for the hereditary factors are based largely on the colors as seen without a microscope. It would be very desirable, however, to correlate color factors accurately with the variations in quality and quantity of the actual pigments and ultimately with the physiology and chemistry of pigment formation. Considerable progress has been made in recent years in the study of the chemistry of melanin pigments. The melanins are amorphous granular pigments found throughout the animal kingdom. A large number of researches have established the fact that substances which closely resemble the natural melanins can be produced by the action of 68 INHERITANCE IN GUINEA-PIGS. certain oxidizing enzymes on tyrosin and related aromatic compounds. Tyrosin is an important constituent of protein molecules and there is much reason to believe that tyrosin and related substances are the chromogens from which the natural melanins are formed. Tyrosinase, an enzyme, which can oxidize tyrosin to dark substances resembling melanins, has been found very widely among animals, including the skins of mammals, as will be discussed later. There have been many theories on the mode of origin of pigment in the cells. Early observations indicated that melanin was directly extruded from the nucleus. Recent studies by Hooker (1915) on in vitro cultures of mesenchyme and epithelium of the frog indicate that melanin granules form in the cytoplasm but at the point of known greatest efficiency of the nucleus as an oxidizing agent. Thus, prob- ably chromogen (tyrosin or derivatives) is in the cytoplasm, while oxidizing enzymes are given off by the nucleus. The color white in the fur of mammals is due to the absence of pigment. The theory of a white melanin seems effectively disproved (Gortner, 1910). A priori, the presence or absence of pigment might be conceived as due either to a deficiency of chromogen or of enzyme. In line with the first view, Gortner (1911) found that the pattern hi the elytra of potato beetles is due to a deficiency of chromogen. Fur- ther, Cuenot (1903, 1904), in the first attempt to correlate the facts of Mendelian inheritance with the physiology of pigment, suggested pro- visionally that albinos lack the power of producing chromogen, while the different colors which he demonstrated could be transmitted through albinos depend on specific enzymes. On the other hand, recent observations by Onslow (1915) demonstrate that absence of pigment in widely different cases in mammals depends on enzyme differ- ences. He found peroxidases in the skins of gray, black, blue, and brown rabbits, which produce a black pigment from tyrosin in the presence of hydrogen peroxide. In the skins of albino rabbits and mice and in the white part of the Dutch pattern in rabbits, all recessive whites, he was unable to demonstrate a peroxidase, although there was nothing present which prevented the oxidation of tyrosin to a black pigment when tyrosinase was added. In the white of the English rabbit, a dominant white, he did find an anti-tyrosinase. Finally, there is strong genetic evidence that albinism in guinea-pigs is not due to absence of chromogen. A diminution in quantity of chromogen should bring about the same diminution in quantity of all pigments, regardless of quality. But in red-eyed guinea-pigs (which we may consider as incomplete albinos, as they have an allelomorph of albinism Cr) no yellow develops, leaving white areas where factors of group 2 determine yellow differentiation, but there may be nearly as much black as in normal guinea-pigs. Indeed, in the albino guinea- pigs and Himalayan rabbits, there is no yellow, but some black. COLOR. 69 The physical or chemical differences between the pigments of the different fur colors are not wholly clear. According to Onslow (1915) the pigments of black, brown, and yellow rabbits can not be distin- guished, physically or chemically, when isolated. At first sight this seems hardly possible with such apparently different colors. A result thoroughly in line with this view, however, followed the matching of fur colors with Ridgway's charts, much to the writer's surprise at the time. Ridgway distinguishes 72 hues passing from red through orange, yellow, green, blue, and purple, back to red. The yellows, sepias, and browns of guinea-pigs and human brown and red hair all matched colors near hue 17, "orange yellow," in the classification. The differences depended merely on differing amounts of black and white. Bateson (1903), indeed, found that yellow pigment is dissolved from hair by potassium hydroxide very much more rapidly than brown pigment, which dissolved more rapidly than black. This, however, might be due merely to size or density of granules. This apparent qualitative difference in pigments has been attributed to several causes: (1) to variations in the chromogen acted on by a given enzyme, (2) to interruptions at different stages in the process of oxidation of a given chromogen, (3) to specific enzymes which in each case can only produce a certain result once the action on the chromo- gen is begun. Observations of Onslow indicate that for qualitative differences, as well as for the absence of pigment, enzyme and not chromogen differ- ences are responsible. He could find no peroxidase in self yellows, a recessive variation. There must of course have been some peroxidase at some time to produce pigment at all. Perhaps the apparent absence indicates a very low degree of stability in the yellow-producing enzyme. Again, grays differ from blacks by a dominant factor which causes yellow to appear in ticking over the back but white to appear on the belly. Onslow found a tyrosinase inhibitor in the belly and compared the case with that of the dominant white of the English rabbit. As grays differ from self blacks by only one Mendelian factor, it would seem likely that all of the changes in appearance— dorsal yellow tick- ing, ventral white — are to be ascribed to one physiological cause. If black is absent from the belly because of an enzyme inhibitor, it would seem likely that black is replaced by yellow in the dorsal ticking by the presence, for a certain period in the development of a hair, of the same enzyme inhibitor, which, however, is in this case merely an inhibitor of the black-producing reaction, not of the yellow. Reasons for which yellow can appear on the back of rabbits, but not on the belly, when black is inhibited will be discussed later. Thus, a recessive yellow and a dominant yellow-pattern factor are both due to enzyme, not chromo- gen, differences. 70 INHERITANCE IN GUINEA-PIGS. The second hypothesis — that yellow, brown, and black are due to interruptions of the normal process of oxidation at different stages is difficult to reconcile satisfactorily with the genetic facts in guinea-pigs. If brown and black pigments pass through a yellow stage, identical with the final stage of the pigment in yellow guinea-pigs, any factor which inhibits the development of yellow must a fortiori inhibit the development of brown and black. We have seen that with factor Cr there is complete absence of yellow pigment, but nearly full develop- ment of brown and black. We find nearly the converse of this in the effect of factor p. When factor p is present, the development of brown and black is very greatly reduced without the slightest dilution of yellow. This indicates that neither is yellow a stage in the develop- ment of black nor black a stage in the development of yellow. The most satisfactory hypothesis is the third — that there are distinct enzymes which produce yellow and dark pigment. There are a number of curious facts in connection with the albino series of factors in guinea-pigs which perhaps warrant further specula- tion. As has been mentioned, Onslow has shown that albinism is due to the absence of tyrosinase in the skin (and presumably the eye). It seems reasonable to suppose that the higher allelomorphs are quantita- tive variations in a factor which determines the power of producing tyrosinase. If this is so, we would expect to find that the different zygotic formulae could be arranged in a linear series with respect to their effects on pigments of all sorts. Following are the series with respect to black pigment of eye and fur, and yellow of the fur. (See plates 1 and 2.) Formula. Black eye. Black fur. Yellow fur. CC Black Black Red. CCd ... .. Do .... Do Do. CCr .Do. . . . Do Do. CCa. .Do .... Do Do. CriCd Do Dark sepia Yellow. CdCr . . .Do. . . . Do Cream. CdCa .Do .... Lis*ht sepia Do. Red Dark sepia .... White. Do Light sepia . Do. Pink. . . . White (sooty) . . . Do. The yellow series and the less accurately known eye-color series can be arranged in the same sequence. There is the striking difference, however, that the level of no pigment production is much higher in yellow than eye color. The black of the fur agrees with eye color in the level at which pigmentation becomes evident — between CaCa and COLOR. 71 CrCa — but the sequence can not be made to agree with either the eye- color or yellow series. CdCa is distinctly lighter than CrCr hi black fur but distinctly more intense hi eye color while, in yellow fur CdCa is above, CrCr below the threshold of any color. The effects could be explained by a complicated linkage hypothesis. We would need to suppose that there are separate series of allelomorphs acting on yellow, black of fur, and black of eye, respectively, and that Cr and Ca are complexes identical in the yellow-dilution factor, Cd and Cr identical in the black-fur-dilution factor and perhaps C and Cd hi the black-eye- dilution factor. But an hypothesis according to which it is a mere accident that the factors which dilute yellow, black of fur, and black of eye are perfectly linked in inheritance can hardly be taken seriously. Another escape would be to suppose that our four factors, Ca, Cr, Cd, and C, are, indeed, variations of the same thing but not linear quantita- tive variations. However, it seems most satisfactory to the writer to attempt to explain the results on the basis of four quantitative gradations of one factor, which determines the amount of the basic color-producing enzyme, if it is in any way possible. Let us see what assumptions must be made to do this. First, it will be convenient to assume with Little (1913) that the basic color-producing enzyme (I) acting by itself on chromogen, produces yellow pigment. The addition of a second substance (II) makes it a black-producing enzyme (I-II). We will further assume that I is relatively unstable and must be produced above a certain rate (that determined by CrCr) in order to reach and oxidize the chromogen in the cytoplasm. United with II it becomes more stable and even produces some effect at the rate of production determined by CaCa. The next assumption is that above the thresh- old for yellow, I-II and the excess of I compete for the chromogen. As a result of partial displacement by the paler color, the intensity of black decreases just above the yellow threshold. CdCa seems paler (and somewhat browner) than CrCr. In the eye, no factor ever brings out a yellow color. There is perhaps never an excess of I here and the intensity of black follows the normal sequence. Summarizing, the hypothesis to which consideration of the physio- logical and genetic seems to lead is as follows : (1) There is a basic color-producing enzyme (I) which acting alone on chromogen produces a diffuse or finely granular pigment which appears yellow. It is relatively unstable. Intensity of production and absence or inhibition in parts of fur and eye are determined by the various factors of group 1 — the albino series, " blue "-dilution factors, and recessive and dominant white-pattern factors. (2) There is a second substance (II) which may unite with I to produce a more stable enzyme, which reacts with chromogen to produce a coarsely granular pigment which appears sepia, brown, or black. When II is present, I is stabilized to such an extent that pigment is 72 INHERITANCE IN GUINEA-PIGS. produced at a lower rate of production of I than is the case of I alone. Above the level at which I alone produces yellow, the two kinds of enzymes, yellow- and black-producing, compete with each other for chromogen, producing a mixture of black and yellow, the relative importance depending on the rate at which II is produced. Because of the competition the intensity of black shows two maxima as production increases — one just below the yellow threshold and the other at maximal production of I. Intensity of production or inhibi- tion of II in patterns in the fur are determined by various factors (group 2) which produce self yellow, yellow spotting, agouti, etc. (3) There is a third group of substances which, added to the dark- pigment-producing enzyme (II), affect the intensity of dark color pro- duced but not the power of fixing chromogen in competition with the yellow-producing enzyme. They have no effect on the intensity of yellow. In this group are the brown factors of mice and guinea-pigs, and perhaps rabbits and dogs, the pink-eye factor of rats, mice, and guinea-pigs and the new red-eye factor of rats, i.e., the factorsof group 3. While based to a larger extent on the genetic facts in the albino series in guinea-pigs, the hypothesis explains many cases in other mammals in the sense that apparently complex variations are reduced to a single physiological cause. In rabbits, single Mendelian factors produce some rather complex variations. A single factor changes a self black to the gray color with a yellow-ticked back but a pure white belly. Another variation changes a self black to a sooty yellow with a black belly. These varia- tions combined in one animal give a white-bellied clear yellow. How can each of these apparently complex color changes be determined by a simple physiological change? Let us suppose that in all rabbits I is produced strongly on the back, but so feebly on the belly that it is below the yellow threshold, but not so feebly that black is greatly affected. Let us suppose that II is likewise more strongly produced on back than on belly. A factor which tends to produce an inhibitor of II is added. On the back II (the black-producing enzyme) is inhibited in only a portion of the development of the hair, leaving yellow ticking. On the belly all II is inhibited, leaving white. The result is a gray rabbit. The other factor causes a general slowing up in the production of II. On the back this enables the yellow-pro- ducing enzyme to predominate hi competition and sooty yellow results. On the belly — below the yellow threshold — what little black-producing enzyme does develop has no competition and only black can result. We get a black-bellied sooty yellow. The combination pattern can only be a white-bellied yellow. In many other mammals color phases are found which can be explained as due either to variations hi production of II or I. The red phase of the red fox has a white chest. The level of production of I is below the yellow threshold COLOR. 73 but above the black threshold on the chest. Increase in production of II produces the silver phase, nearly self sepia in color, including the chest. The colors of the varying hare seem to be due to variations in production of I determined by environmental causes. The white winter pelage gives way in blotches to a white-ticked sepia ; this gives way to yellow-ticked sepia as the intensity of production of the basic enzyme rises above the yellow threshold and in some varieties the full summer pelage is almost self red. Many other cases could be given in which two color phases of an individual animal or the color patterns of closely allied varieties seem to differ in many respects and yet can be explained on the basis outlined as due to a single physiological change. In the case of very complex color patterns, it is necessary to suppose that the power of producing the hypothetical enzymes I, II, or III may be distributed in quite complex patterns. But the hypothesis often gives a simple explanation for certain peculiarities in a pattern. In the tiger, the stripes on the back are quite intense yellow and black. The yellow stripes grow paler down the sides, becoming white on the lower sides and belly. The black stripes likewise grow lighter down the sides but at the point at which the yellow becomes white, the black stripes suddenly grow more intense, at least in some individuals, to become paler again on the belly. Again, on the legs, which are white on the inside, yellow on the outside, black stripes are visible on the white part, but either disappear completely or leave merely a streak of sooty red on the yellow part. All of this becomes intelligible if we assume that the basic enzyme (I) is produced at decreasing rates from back to belly and from outside to inside of leg, while the black-produc- ing supplement (II) is distributed in vertical stripes (horizontal on the legs). Two parallel stripes give a remarkable reproduction of the variation in black and yellow in the albino series in guinea-pigs. We have the same change from black and intense yellow to sepia and cream, then to darker sepia and white, and finally light sepia and white, illustrating the different thresholds for the appearance of black and yellow and the reduction in intensity of black above the yellow threshold due to the entrance of competition at this point. DISCUSSION OF EXPERIMENTS. MATERIAL. SYSTEMATIC POSITION. Guinea-pigs belong to the family Caviidae of the hystricomorph division of rodents. There are three living genera of Caviidse: Doli- chotis Desm., which contains the large Patagonian cavies; Hydrochcerus Brisson, to which belongs the capybara; and Cavia Pallas, containing the small cavies. Genus Cavia is divided into two subgenera, Cavia proper and Cerodon F. Cuv., distinguished most conspicuously by the greater complexity of the molars in the former. Seven living species are listed under Cavia proper by Trouessart (1904) : C. rufescens Lund, a small dark Brazilian cavy with subspecies in Guiana and Argentina. C. fulgida Wagler, a Brazilian cavy probably closely allied to rufescens (Thomas, 1901). C. aperea Erxl., a large pale-colored Brazilian cavy. C. azarcB Wagner, a cavy of Paraguay probably closely allied to aperea (Thomas. 1901). C. cutleri Bennett, a small pale-colored cavy of Peru. C. tschudii Fitzinger, a large, richly colored cavy described from lea, Peru. C. porcellus Linn., the tame guinea-pig, much larger than at least rufescens and cutleri. DESCRIPTION OF STOCKS. Four of these species are dealt with in the experiments to be described, viz, Cavia rufescens, C. cutleri, C. porcellus, and a type which is quite certainly that described as C. tschudii, although it is also quite certain that it is simply feral porcellus. Breeding experiments have been carried on with a fifth species, C. aperea, by Nehring (1894). The C. rufescens stock was derived from 3 individuals received from Mr. Adolph Hempel, of Campinas, Brazil, in 1903. The history of this stock is fully described by Detlefsen (1914). When received by the writer, most of the stock consisted of hybrids containing only from yV to -5^ rufescens blood. There were a few | and \ bloods and one % blood, 9 A68, which is still alive (August 1915) at the remarkable age of 8 years 1 month,1 a good illustration of the vigor of the first-generation hybrids. All of the pure rufescens stock has died out. The rufescens hybrids have been crossed with nearly all of the guinea-pig stocks to be described, and most of the color varieties may be found among them. The ticked-bellied type of agouti has been found only among them and in pure rufescens. C. rufescens was not completely fertile with the guinea-pigs (Detlefsen, 1914). Detlefsen found that while the female hybrids were fertile, all of the male hybrids obtained were sterile. In the \ rufescens, derived by crossing the females with guinea-pigs, the males were again all sterile. Not until the | bloods were obtained did 'Died October 1915, aged 8 years, 3 months. — TT. E. C. 74 MATERIAL. 75 a few fertile males appear. The percentage of fertile males gradually increased in later generations. The Cavia cutleri stock was derived from animals captured by Pro- fessor Castle in Peru in 1911. Like C. rufescens, these are much smaller than the guinea-pig. All show the agouti pattern. The color is described on page 59. Unlike C. rufescens, C. cutleri breeds freely in captivity and crosses readily with the guinea-pig. The male and female hybrids are fertile. The lea stock of guinea-pigs was derived from 3 guinea-pigs which were obtained by Castle near lea, Peru, in 1911. They were as large as or larger than average guinea-pigs, and of a rich golden agouti color, very different from C. cutleri. Two independent color variations appeared at once in the pure stock, viz, black (aa) and red-eye (CrCr). Such variations are very uncommon among wild species of animals; e. . Aglb. Aglb. Aglb. Aijtb. .Y«/i.-«(7. (I) aaA'a' X Aaa'a' = AaA'a' + Aaa'a' + aaA'a' + aaa'a'. (II) A'a X Aa = AA' + Aa + A'a + aa In both cases we expect 2 light-bellies to 1 ticked-belly to 1 non- agouti. Under (I) the light-bellied young which can transmit ticked- belly (AaA'a) must also have the power of transmitting non-agouti. Under (II) such light-bellies (AA') should not transmit non-agouti. Under (II) half of the light bellies should be of this type and the other half should transmit non-agouti but not ticked-belly (Aa). Thus, if a large number of young can be obtained from a light-belly from such a cross, which has had ticked-bellied young in crosses with non-agoutis, the presence or absence of non-agouti young is decisive between the VARIATIONS IN AGOUTI PATTERN. 99 two hypotheses. Cross 10 gives the results of the tests of light-bellied young from such a cross as described (table 44). TABLE 44. Females. Males. Aglb. Agtb. Non-ag. 10a 7 aglb . . . Non-ag . . 17 20 lOc Non-ag . . 4 aglb. . . . 26 33 106 10 aglb... Non-ag . . 18 30 Wd Non-ag . . 5 aglb .... 45 50 In no case has the same animal had both ticked-bellied and non- agouti young. Some of those which have had ticked-bellied young have been quite thoroughly tested. Male M138 had 20 light-bellied and 19 ticked-bellied young. Male B121 had 13 light-bellied and 13 ticked-bellied young. Male M91 had 5 light-bellied and 11 ticked- bellied young. The chance that these can represent 2:1:1 ratios is negligible. Thus hypothesis (I) may be dismissed. Light-bellied agoutis demonstrated to carry ticked-belly have been crossed inter se (cross 11). They have given 25 light-bellies and 9 ticked-bellies, no non-agoutis. This agrees reasonably well with the expected 3 to 1 ratio. The remaining tables give the results of miscel- laneous crosses. All of them are in harmony with the hypothesis of triple allelomorphs. MINOR VARIATIONS. Thus there seems no doubt that the light-bellied agouti of Cavia porcellus, the ticked-bellied agouti of C. rufescens hybrids, and non- agouti form a series of triple allelomorphs. The question remains whether light-bellied Cavia rufescens hybrids possess a different allelo- morph from the ticked-bellied ones, or whether the difference lies simply in the residual heredity. There are no wholly satisfactory data bearing on this point. Nevertheless the fact that the darkening seems associated especially with certain stocks of guinea-pigs seems to favor the second view. The writer has crossed ticked-bellied agoutis re- peatedly with the intense blacks of BB or BW stock. Young have been obtained which were self black, except for a few ticked hairs in the chest and whiskers. One ignorant of their history would probably have classified several of them as blacks. Before they became adult, however, these black ticked-bellies acquired a uniform though very slight yellow ticking throughout the entire fur. On crossing such black ticked-bellies with a dull black stock (4-toe) there is a return to a more strongly developed agouti pattern. The young are uniformly ticked when born. Thus these variations in the agouti pattern seem related to the residual heredity of the stocks, possibly with the same residual heredity which determines the very intense development of pigment, 100 INHERITANCE IN GUINEA-PIGS. especially black, in the BB and BW races, the feebler development in the 4-toe race, and the much feebler development in the wild species. If this is correct, the resemblance of light-bellied rufescens to light- bellied agoutis, like that of the pale color of C. cutleri to dilute guinea- pigs, is secondary. In both cases the wild species possess a different allelomorph from the guinea-pig in the principal series of factors involved, but owing to different residual heredity, have a superficial resemblance. THE INHERITANCE OF THE AGOUTI OF CAVIA CUTLERI. The writer has had the opportunity of working with the agouti of Cavia cutleri. Repeated crosses have been made with blacks of the BW race of guinea-pigs to see whether a ticked-bellied agouti could be obtained. While some ventral ticking has been observed in some cases, the f-blood cutleri hybrids are still on the whole good light-bellied agoutis. The cutleri agouti is unquestionably more resistant to dark- ening influences than was rufescens agouti. No results have been obtained yet which serve to differentiate it from the guinea-pig agouti. This is additional evidence that C. cutleri was ancestral to porcellus. The experimental results are given in crosses 68 to 78. Only one cross has been made between a cutleri and ticked-bellied rufescens hybrids. Male K56, a black f cutleri (f 4-toe), was crossed with two ticked-bellied agoutis, which of course had some rufescens ancestry. There were 6 young (3 blacks and 3 ticked-bellies) of which one was quite light and one was black, except for a few ticked hairs on the chest and whiskers. There was thus no very conspicuous tendency toward light-belly introduced by the cutleri hybrid. It seems safe to assume that C. cutleri has a different member of the agouti series of allelomorphs from C. rufescens, but the same or nearly the same as C. porcellus. Wild species of the same genus seldom differ as much superficially in any one character as do many varieties of domesticated animals. Yet while very large variations in the latter have been shown in many cases to behave as simple Mendelian units in inheritance, the char- acters by which wild species differ usually seem to be highly complex in heredity. Few well-defined Mendelian factors are recorded in the literature of hybridization. It is, therefore, interesting to find that the darker agouti of Cavia rufescens differs from the lighter agouti of C. cutleri by a clear-cut Mendelian factor. INHERITANCE OF ROUGH FUR. In the wild species of cavy, and in the ordinary smooth guinea-pigs, the hair shows a definite direction of growth, which is always away from the snout on the body and toward the toes on the legs. This is at least the general tendency of the hair in most mammals, and it is ROUGH FUR. 101 obviously the most advantageous; the hair lying thus is not ruffled or caught by obstacles when the animal is moving. This direction is not directly imposed on the hair by outside agencies, as might be supposed, but is due to the direction of growth of the hair follicles (Wilder, 1909). Certain fancy varieties among guinea-pigs, as the long-haired rough "Peruvians " and the short-haired rough " Abyssinians," showa striking deviation from the normal hair direction. In these varieties the coat can be divided into a number of areas, within each of which all hair directions radiate from a definite center. The boundaries of these areas, where contrary hair-currents meet, are marked by crests. The centers with then- radiating hair-currents are called rosettes. Many mammals, including man, naturally show rosettes, crests, or other peculiarities of hair direction, but less conspicuously than the rough guinea-pigs. (See plate 7.) The positions in which rosettes may occur in guinea-pigs are quite definite. Following are the rosettes and irregularities given by Castle (1905), with the addition of L, irregular roughness on the chest. A. Forehead, unpaired. E. Sides, between shoulder and hip. I. Navel, unpaired. B. Eyes. F. Hips. J. Front toes. C. Ears. G. Above the groin. K. Hind toes. D. Shoulders. H. Mammse. L. Irregular roughness of chest. In the grading of the young guinea-pigs, large letters, as above, have been used for well-defined rosettes and small letters for feeble rosettes or slight deviations from normal hair direction in an area, indicated only by crests at a boundary. Thus a mid-dorsal crest or mane (e) without any side or hip rosettes is characteristic of a certain grade of partial roughs. The ear rosettes (C) are usually only revealed by a crest between the ears. The shoulder rosettes are seldom well developed. The side rosettes are sometimes doubled in the roughest animals (E,E2). The number of rosettes present varies from the full set described above, through a continuous series of intermediate grades, to one pair. The variations are not merely haphazard, but may easily be classified. In the first place, it is necessary to distinguish two series. A slight roughness found in certain stocks (the BW, lea, and Arequipa stocks) does not fit into the usual series of variations and will be discussed separately as series II. The roughness of the remaining stocks and also of the fanciers' "Peruvians" and " Abyssinians " we may call series I. In series I all the variations found may be arranged with con- siderable accuracy in a single linear series. Thus Castle (1905) used six grades, passing from rough A with the maximum number of rosettes to rough F, smooth except for the hind toes. These grades will be used in this discussion, with the exception that it has been found con- venient to combine grades C and D, leaving five grades, A to E. These letters used for grades must not be confused with those used to name the rosettes. 102 INHERITANCE IN GUINEA-PIGS. Reversal of hair direction on the hind toes is the most constant feature of the roughness of series I and has been found in all rough guinea-pigs of series I without exception. Following is the usual order of succession of the additional rosettes and irregularities found in pass- ing from a smooth to a full-rough: Hind toes K Front toes J Dorsal crest e Side rosettes, crest between ears E,C Forehead, hip, ventral rosettes. . . A,F,H,I,L Eye rosettes B Groin, shoulder, second side ro- settes G,D,Ea There is seldom more than a slight amount of asymmetry. As a rule the paired rosettes are present or absent as pairs. The most common exception is in the side rosettes in low-grade partial-roughs. Among these it is not uncommon to find a good rosette on one side and merely a slight change in hair direction on the other. In classifying, six of the most distinct sets of rosettes have been used as the principal criteria, viz, forehead, eyes, sides, hips, front toes, and hind toes. CLASSIFICATION. Rough A. — The forehead and five critical pairs of rosettes must be well developed. In addition, there is always some ventral roughness and a crest between the ears. Rough B. — This includes various conditions intermediate between rough A and rough C. Rough C. — There is only one pair (or half pair) of well-developed rosettes, usually the side rosettes. There is always, in addition, rough- ness on at least the hind toes and usually a crest between the ears. Rough D. — A mid-dorsal crest is present and roughness of at least the hind toes, but no well-marked rosettes. Rough E. — Roughness is confined exclusively to the toes, usually to the hind toes. The following list shows the variations which have been met with in each grade; the letters represent the rosettes: Rough A. ABcEFHIJK to ABCDEiE2FGHIJKL. Rough B. AcEJK, AcEHJK, abFGJK, ABcFHIJKl, ABcDEHIJKl. Rough C. EK, EJK, cEK, cEJK. Rough D. eK, cK, eJK, cJK. Rough E. K, JK. PREVIOUS WORK. Nehring (1894) made crosses between rough guinea-pigs and the wild species Cavia aperea. He described the young as smooth, but noted that a mane developed along the middle line of the back. Castle (1905) demonstrated that rough fur behaves as a Mendelian unit ROUGH FUR. 103 character, dominant over smooth, and is thus an example of the com- paratively rare class of dominant mutations. He found that the grade of roughness, while fairly constant in some stocks, could be reduced by crossing with smooth guinea-pigs of a particular stock (tricolor) , which he described as prepotent smooth. Detlefsen (1914) found that rough fur in hybrids between rough guinea-pigs and the wild species Cavia rufescens continued to be inherited in Mendelian fashion, but that domi- nance ceased to be complete. The writer began experiments in 1913 at Professor Castle's suggestion, to investigate further the heredity of variations in the rough character. MATERIAL. Several stocks have been used as material: (1) 4- toe stock. — A full-rough was crossed with members of the 4- toe stock, and by repeated back-crosses into the latter a stock has been produced which is practically pure 4-toe. No partial-roughs have ever appeared in these crosses. Pure 4-toe smooth animals have been very useful in the experiments, since it has been amply proved that when crossed with full-roughs they never reduce the grade of rough. (2) Tricolor stock. — Most of the partial-roughs experimented with are of very mongrel stock, with, however, more or less tricolor ancestry. In this section on rough fur the term tricolor stock will be used for convenience for these animals, without implying that all of them actually were tricolors. (3) Lima stock.— This stock as has been described was derived en- tirely from 8 guinea-pigs (2 nearly full-rough (rough B) and 6 smooth) brought from Peru in 1913. (4) Rufescens hybrids. — The writer has worked with a few rough animals descended from Detlefsen's hybrids and containing from £ to Cavia rufescens ancestry. (5) Cutleri hybrids. — Crossing with Cavia cutleri has been found by the writer to have a similar effect on the rough character to that described by Nehring (1894) for C. aperea and by Detlefsen (1914) for C. rufescens. The behavior of the roughness in these cutleri hybrids has been investigated. (6) Miscellaneous smooth guinea-pig stocks have been used in some of the experiments. All used resemble the 4-toe smooths in giving no partial-roughs when crossed with full-roughs. (7) BW, lea, and Arequipa stock occasionally have shown a slight roughness which is distinct from the usual type and is discussed later under series II. The BW stock has been used in a few cases as a source of smooth guinea-pigs. Aside from the slight roughness of series II all of those used have behaved like 4-toe smooths. Smooth 104 INHERITANCE IN GUINEA-PIGS. Sm 4 toe stock leas, on the other hand, behaved like the wild cavies in reducing the roughness. PROBLEMS. The following figures show the kinds of roughs which have appeared in the experiments with the four principal stocks and the nature of the variability in the rough character, the inheritance of which it is desired to analyze. All roughs are included, but only such smooths as had at least one rough parent. Inspection of figure 7 at once shows striking differences in the variability of the rough character in the different stocks. In the 4-toe stock there is a wide gap between the lowest rough and the smooths which come from the same parents. Only one individual in the 4-toe stock was graded rough B and this was close to rough A (lacked only groin and hip rosettes) . Most of the individuals were strong rough A. In the Lima stock most of the individuals which were rough at all were rough B or a weak rough A, but 5 were rough C or D. Among the tricolor and cutleri hybrids a continuous series can be formed passing from the best roughs to the smooths. Both show a distinctly bimodal distri- bution of the roughs, the modes being at rough A and rough C. Such a distribution would of course be purely artificial if rough B were a more limited category than rough A or rough C, but the FIG. 7. — Distribution of grades of roughness of the definitions show that TOUgll B in- fur in four stocks of guinea-pigs. eludes perhaps the widest range of possible variation. Further, the large number of rough B's in the Lima stock shows that this class may be practically as numerous as rough A under the right hereditary conditions. Thus there are strong intimations that in the tricolor and cutleri stocks, rough A and rough C differ by a unit hereditary difference. The problem of the inheritance of the variations in the rough char- acter thus seems to resolve itself into three phases: (1) The inheritance of roughness of any sort as opposed to smoothness; (2) the inheritance of a more or less full-rough type averaging about rough A as opposed to a partial-rough type averaging between rough C and D; (3) the inheritance of the variations within the full-rough and partial-rough types. 109 63 Tricolor Stock Lima stock Cutleri hybrids ROUGH FUR. 105 INHERITANCE OF ROUGH AS OPPOSED TO SMOOTH. Castle (1905) demonstrated that all roughs differ from smooths by a Mendelian unit-factor and that rough is dominant over smooth. The writer's experience fully confirms this conclusion. In nearly 3,000 young recorded, smooth by smooth has never given rise to rough (of series I) in spite of much rough ancestry, with one possible exception. This exception was of a kind which was expected and was being tested for when found. One of the smooth parents was undoubtedly like rough E genetically. The case will be discussed later. TABLE 45. Rough X rough. Rough X smooth. Formula and stock. Rough. Smooth. Formula and stock. Rough. Smooth. RR X Rr: 4003 (tri) 6 10 88 18 17 0 3 28 7 2 RR X rr: 4003 (tri) 11 29 104 56 60 54 0 32 125 55 63 48 Rr X Rr: 4-toe Rr X rr: 4-toe Tricolor. . . . Tricolor Lima Lima Cutleri hybrid Cutleri hybrid Total Miscellaneous . . Total 133 (130) 40 (43) 303 (313) 323 (313) Expectation Expectation Largely Rr X Rr: Miscellaneous Largely Rr X rr 46 12 16 12 On the other hand, rough by rough has often given smooth. In the cross of rough by smooths from stocks in which roughs have never occurred, all or half of the young are rough. All wild cavies, for ex- ample, are smooth and have only smooth descendants when crossed with smooth guinea-pigs. They have numerous rough young in Fx when crossed with rough guinea-pigs. Thus it is clear that rough is domi- nant. Table 45 is a summary of the rough crosses made by the writer. Grades of roughness are ignored. No special attempt has been made to obtain homozygous roughs. Male 4003, rough E, is the only one which has been adequately proved homozygous. Male R197, rough A (cross 46), is another which is probably homozygous. In the cross Rr X Rr above, only matings are included in which both animals are known to be heterozygous, either because of a smooth parent, or, having had 12 or more young, because of smooth young. Tabulated in this way, the expectation is not appreciably different from 3 rough to 1 smooth. Other litters of rough by rough are tabulated above. Probably most of these are of the type Rr X Rr, although in some cases there were no smooth young. In crosses Rr X rr, the only cases tab- 106 INHERITANCE IN GUINEA-PIGS. ulated are those in which the rough is known to be Rr because of a smooth parent, or because of a smooth young one in 6 or more. Expectation is here 1 to 1. The remaining cases of rough by smooth, in which expectation is still probably not far from 1 to 1, are also given. These results are in harmony with the view that rough differs from smooth by a dominant unit factor. INHERITANCE OF MAJOR VARIATIONS. Before giving any hypotheses, it will be well to present the experi- mental results with the immediate deductions which can be drawn from them (tables 46 to 55) . (1) In some stocks there is very little variation in the rough char- acter and there is a wide gap between the lowest rough and smooth. The 4-toe stock is an excellent example of such a stock. It is worthy of note that we get a similar result in the Lima stock if we exclude the litters of L6, L24, L56, and L99. Female L6, smooth, was one of the original 8 in the Lima stock. Female L24, smooth, was her daughter. Female L56, rough C, was the daughter of L24, and male L99, rough C, was the son of L56. Most of the tame guinea-pig stocks (BW, dilute selection) seemed to be like the 4-toe stock in the above respect when- ever rough was introduced into them in a cross. TABLE 46. Cross. Stock and grade. A B C D E Sm (45) Four-toe stock: A X A ... . . 10 3 (49) A X Sm 28 1 32 (58) Lima stock, except L6, L24, L56, and L99: B X B 7 5 3 (59) A X Sm . ... 21 5 17 (60) B X Sm 5 9 ?1 (63) (65, 66) Miscellaneous stock: A, B (Lima) X Sm (4-toe) . . A X Sm 4 31 4 1 4 36 Most of those graded rough A or rough B above must be heterozy- gous. As nothing higher than rough A appeared in the crosses A X A and B X B, it seems clear that the homozygotes in these stocks at least are no more rough than the heterozygotes, i. e., dominance is complete. (2) When a wild species of Cavia (smooth), or a smooth of certain tame stocks is crossed with a full-rough, the rough young are of low grade — rough C or D. ROUGH FUR. 107 As has been mentioned before, Nehring (1894) crossed a rough male guinea-pig with Cawa aperea of Argentina. He described the young as smooth at first, but developing a mane along the back later. This was evidently the dorsal crest of rough D. The reversal of hair direc- tion on the hind toes might easily have been overlooked. Detlefsen (1914) described crosses of C. rufescens of Brazil with full-rough guinea- pigs. His rufescens was undoubtedly a different species from the aperea used by Nehring, since the latter found complete fertility among the hybrids of both sexes, whereas Detlefsen found sterility among all the male hybrids. The rough young were rough D. The skin of one of them (A10) shows roughness on all the toes and a very slight dorsal crest. The writer has crossed rough A guinea-pigs with C. cutleri of Peru with similar results. Nine rough young have been obtained, of which 3 show good side rosettes and are rough C, while the others merely show a dorsal crest (and rough toes) and are rough D. A male of pure lea stock, which being from feral stock probably had consider- able wild ancestry, was tested by crosses with full roughs and also gave only partial-roughs — 5 all rough C. Castle (1905) obtained partial-roughs, C or D, on crossing full-roughs with smooths of tri- color stock. The writer has made further crosses of this kind with similar results. One smooth guinea-pig of the Lima stock, L24, had some partial-rough young when crossed with full-roughs of her stock. TABLE 47. Cross. Stock and grade. A B C D E Sm Smooth, of wild or feral stock: A X Sm (C. aperia) 1 Nehring (1894). (68) (67) A X Sm (C. rufescens) .... A X Sm (C. cutleri) A X Sm (lea) 3 5 4 6 7 15 4 Detlefsen (1914). (71, 74) Certain tame stocks and wild hybrids: A X Sm (5 cutleri) 10 3 10 1 23 C65) (59, 60) A X Sm (j, 3*2 rufescens) . A X Sm (L6, L24) 1 4 1 1 2 3 6 (51) A X Sm (Tricolor) . ... 6 3 3 19 The results given in table 47 show that in partial-roughs from a great variety of sources, the low grade of the roughness is not due to a vari- ation of the rough factor, i.e., to an allelomorph, nor is it due to an inde- pendent duplicate rough factor which produces somewhat similar effects to the factor of full-roughs. These partial-roughs have the identical rough factor present in full-roughs derived directly from the latter. (3) When a wild cavy is crossed with a partial-rough guinea-pig, rough young of the lowest grade (rough E) are produced. 108 INHERITANCE IN GUINEA-PIGS. No rough E young were produced in the crosses under (2) where one parent was full-rough. TABLE 48. Cross. Stock and grade. A B C D E Sm (69) C, D X Sm (C. cutleri) 1 1 9 12 (4) Partial-roughs crossed together may give all grades of roughness from full-rough (A) to the lowest partial-rough (E) . TABLE 49. Cross. Stock and grade. A B C D E Sm (76) cutleri hybrids: C X C 1 3 1 (52) Tricolor: C, D X C 18 6 19 7 12 17 (53) C, D X E 4 1 6 1 (55) EXE 4 3 (5) Full-roughs crossed together have never given partial-roughs. Crosses in the 4-toe and Lima stock have already been given. Below are crosses of rough A by rough A in the tricolor and cutleri stocks in which each rough A parent had one or both of its parents partial- rough (C, D) or in the case of the cutleri hybrids, an FI hybrid. TABLE 50. Cross. Stock and grade. A B C D E Sm (46) Tricolor A X A 17 2 7 (75) j cutleri A X A 3 (6) Full-roughs, one or both of whose parents were partial-roughs, have given no partial-rough young in crossing with smooths of 4-toe stock. TABLE 51. Cross. Stock and grade. A B C D E Sm (50) (73) A (Tri) X Sm (4-toe) .... A (j cut) X Sm (4--toe) 19 2 20 2 In (73) two of the \ cutleri rough A did not come from a partial-rough parent, but from smooth \ cutleri hybrids. ROUGH FUR. 109 (7) Partial-roughs crossed with smooths of 4-toe or a similar stock give partial-rough young, and also, in most cases, full-rough and smooth young. TABLE 52. Cross. Stock and grade. A B C D E Sm (54) C, D (Tri) X Sm (4-toe, etc.) 34 29 13 1 79 (64) (72) (61) C (j, |, g1} rufescens) X Sm (4-toe, etc.). . C, D (|, £ cutleri) X Sm (4-toe, etc.) .... L 56 C X Sm (Lima) 3 12 2 1 2 6 1 6 5 21 2 (56) E (Tri) X Sm (4-toe) 13 8 Crosses (3) to (7) show that full-rough can be recovered from partial- rough usually without trouble, but that partial-rough can not be re- covered from full-rough, either in crossing full-roughs together or with ordinary smooth guinea-pigs (4-toe, etc.). Cross (7) is more significant than appears at first sight. We know that 4-toe smooths transmit nothing which can reduce the grade of roughness. Thus, when we find partial-rough by 4-toe smooth giving partial-rough young, we see that the rough factor and the factor or factors responsible for the low grade of roughness can be transmitted in the same gamete. (8) Most partial-roughs crossed with full-roughs give a very similar result to the cross partial-rough by 4-toe smooth, except that fewer smooths are produced. TABLE 53. Cross. Stock and grade. A B C D E Sm (47) A X C (Tri) 10 1 5 1 10 (58) B (Lima) X C (Lima) 1 1 (70) A (4-toe, etc.) X C, D (£, j cutleri) 8 3 s 2 2 4 (48) A (Tri) X E (Tri) 2 1 (9) The lowest grade of roughs (E) very rarely have either a full- rough parent or full-rough young. They also very rarely have a smooth parent of such a stock as 4-toe. Of the icugh young, 507 have been recorded (excluding 8 from mixed bimaternal litters) ; 413 of these had a full-rough (A, B) or a 4-toe smooth parent; yet these include only 3 rough E young. On the other hand, 13 rough E are included in the 67 rough young from C or D X C ; 6 are included in the 11 from C or D X E, 9 are included in the 11 rough young from C X smooth Cavia cutleri and all of the 4 rough young from E X E were rough E. Table 54 shows the matings in which one or both of the parents were rough E. These are repeated from other crosses. 110 INHERITANCE IN GUINEA-PIGS. The offspring of male 4003 rough E are of special interest. He was undoubtedly homozygous rough; the chance that he was heterozygous is (|)u(f)6 = 0.0001. As he was the lowest grade of rough, he very emphatically disproves any necessary relation between homozygosis and high development of the rough character, or between heterozygosis and partial roughness. TABLE 54. Cross. Stock and grade. A B C D E Sm (56) E X Sm (4-toe) 13 8 (48) E X A 2 1 (53) E X C, D 4 1 fi 1 (55) EXE 4 3 (10) Occasionally a smooth from a cross which produces rough E will transmit the rough factor, breeding like a rough E. Rough E grades into smooth. From a cross which can produce rough E, 8 smooth animals were tested by crossing with 4-toe smooths to determine whether a smooth can ever be like rough E genetically. One such animal, female R201, was found. TABLE 55. Cross. Stock and grade. A B C D E Sm (57) 7 Sm (C X C) X Sm (4-toe) . . . 32 R 201 Sm (C X C) X Sm (4-toe) . . . 1 1 The case is not quite as clear as could be desired, since R201 seemed to show a trace of irregularity on one hind toe when first graded. As an adult she is indistinguishable from a smooth. These experiments are sufficient, it is believed, to establish the mode of inheritance of the major variations of the rough character. First, it is clear that partial-roughs do not owe their roughness to an allelomorph of the rough factor or to an independent duplicate rough factor, but to the same factor found in full-roughs. Reasons were given under (2). Next, any hypothesis is untenable according to which partial-roughs are due to imperfect dominance and hence are necessarily heterozygous either with the ordinary smooth factor of guinea-pigs or with a more potent allelomorph of the latter present in wild cavies and special stocks of tame guinea-pigs. The latter hypothesis was suggested by Det- lefsen (1914), in the case of the partial-roughs among the rufescens hybrids. He represented the rough factor by Rf, the ordinary smooth factor by rf , and the smooth factor of Cavia rufescens by rf. He sup- posed that Rf is completely dominant over rf, but incompletely domi- nant over rf '. Thus Rfrf would be a full-rough, but Rfrf' a partia- ROUGH FUR. Ill rough. This hypothesis explains very satisfactorily all of the crosses given except those under (7), repeated under (9) and (10). It can not explain the case of male 4003 rough E, who was undoubtedly homo- zygous for the rough factor (RfRf) and yet was the lowest grade of partial-rough. Further, it can not explain the occurrence of partial- rough young coming from the cross partial-rough by 4- toe smooth. The latter are necessarily rfrf under the above hypothesis. Rfrf X rfrf = Rfrf + rfrf'. Rough C, D, E X smooth = rough A + smooth. Under this hypothesis, the rough factor Rf and the factor which reduces the grade of rough rf ' can not be present in the same gamete. But this cross actually gave 70 partial-rough young, 56 from tricolor partial-roughs, 12 from partial-rough cutleri hybrids, and at least 1, probably 2, where the partial-rough parent owed its low grade to Cavia rufescens. Female A606 rough C was | rufescens. Her parents were 2193, a full-rough guinea-pig, and A63, a smooth rufescens hybrid. The hypothetical factor rf' could only have come from the latter. The parents of A63 were A55, a pure Cavia rufescens, and 9586 of BW stock, a stock which has shown no tendency to reduce the grade of full-roughs on crossing with them. It thus seems clear that A606 owes her low grade of rough to her Cavia rufescens grandfather. She was crossed with a 4-toe smooth male, 166, and had two rough young, whose grades unfortunately were not recorded at birth. One, however, A1687, is still living (August 1915), and is a typical rough C. There is reason for believing the other to have been of the same grade. Thus in tricolors, Cavia cutleri and Cavia rufescens hybrids, the same gamete can transmit the rough factor and the factor or factors which limit the full development of the rough character. The formula Rfrf can not, therefore, be used for partial-roughs. There remains only one line of explanation. Partial-roughs must differ from full-roughs by possessing an independently inherited modi- fying factor (or factors). An incompletely dominant unit modifying factor will explain all of the results satisfactorily. Let us represent the wild condition (aperea, rufescens, cutleri) by rrSS. Let us suppose that the dominant mutation R is necessary for any roughness [of series I] and produces with rare exceptions at least reversal of hair direction on the hind toes (rough E). The second mutation, s, when heterozy- gous, may permit roughness to extend to grades D or C ; when homozy- gous it permits roughness to reach grades B or A. rrSS smooth Wild species, lea stock. rrSs smooth Most tricolor smooths. rrss smooth 4-toe, BW, BB, dilute stock, most Lima smooths. RrSS rough E (rarely smooth) R 140, etc. RRSS rough E (rarely smooth) 4003. RrSs rough C or D (rarely E or B) . . . Most tricolor partial-roughs. RRSs rough C or D. RrSS rough A (less frequently B) . . . . 4-toe, most Lima roughs. RRss rough A. 112 INtfERITANCE IN GUINEA-PIGS. This explanation fits very well the results which have been given qualitatively. As the numbers are rather small, and as it is necessary besides to assume some overlapping of class ranges, much emphasis can not be laid on the quantitative results. Nevertheless, the fit is in all cases reasonably close. Let us take up the qualitative results in order. (1) In some stocks there is very little variation in the rough char- acter and there is a wide gap between the lowest rough and smooth. Evidently if the 4-toe and similar stocks are pure for factor s, only full-roughs and smooths can appear (RRss, Rrss, rrss). Apparently 7 of the original 8 in the Lima stock were ss, while one, L6, was rrSs. (2) When a wild species of cavy (smooth) or a smooth of certain tame stocks is crossed with a full-rough, the rough young are of low grade, rough C or D. This result necessarily follows from a cross of the type rrSS (wild, lea, tricolor) by Rrss (rough A). The rough young RrSs should be rough C or D. (3) When a wild cavy is crossed with a partial-rough guinea-pig, rough young of the lowest grade, rough E, are produced. rrSS X RrSs = RrSs + RrSS + 2rr Sm C C E 2Sm (4) Partial-roughs crossed together may give all grades of roughness from full-rough (A) to the lowest partial-rough (E). Most of the partial-roughs handled should be RrSs. RrSs X RrSs = 3 Rss + 6 RSs + 3 RSS + 4 rr C C 3 A 6 C, D 3 E 4 Sm (5) Full-roughs crossed together have never given partial-roughs. A full-rough, whatever its parentage, must be RRss or Rrss. There is no way in which factor S, necessary for partial-roughs, can be trans- mitted by full-roughs. (6) Full-roughs, one or both of whose parents were partial-roughs, have given no partial-rough young on crossing with smooths of 4-toe stock. Smooths of 4-toe stock are all necessarily rrss and can not transmit factor S. (7) Partial-roughs crossed with smooths of 4-toe or a similar stock give partial-rough young and also, in most cases, full-rough and smooth young. RrSs X rrss = Rrss + RrSs + 2 rr C Sm (4-toe) A C 2 Sm (8) Most partial-roughs crossed with full-roughs^give a very similar result to the cross partial-rough by 4-toe smooth, except that fewer smooths are produced. RrSs X Rrss = 3 Rss + 3 RSs + 2 rr C A 3A 3C 2Sm (9) The lowest grade of roughs (E) very rarely have either a full- rough parent or full-rough young. They also very rarely have a ROUGH FUR. 113 smooth parent of such a stock as 4-toe. The parents and offspring of rough E (SS) must necessarily have at least one S (SS or Ss), while full-roughs and 4-toe smooths are ss. The three exceptional rough E's can only be interpreted as extreme minus fluctuations of type RrSs. (10) Occasionally a smooth from a cross which produces rough E will transmit the rough factor, breeding like a rough E. The discovery of a smooth (RrSS) which had a rough C young one (RrSs) when crossed with a 4-toe smooth (rrss), apparently violating the dominance of roughness, is the kind of exception that proves the rule, coming as it did where predicted. The descendants of male 4003 illustrate the theory very well. He was of constitution RRSS by theory. TABLE 56. A B C D E Sm P! RRSS X rrss 4003 4-toe Fi RrSs 11 F2 7 Rss + 13 RSs + 7 RSS + 9 rr . . . 8 5 8 3 5 7 7 A, B 13 C, D 7E 9 Sm Of those called rough B, 2 were close to rough A and 3 were close to rough C. POSSIBILITIES OF LINKAGE AMONG ROUGH AND COLOR FACTORS. In the mating just cited, factors R and S enter the cross from the same individual. The excess of full-roughs (Rss), probably 10 where 7 are expected, makes any linkage between R and S very unlikely. Another test is furnished by the cross of double heterozygotes, RrSs with 4-toe smooths, rrss, where it is definitely known whether R and S enter the cross together or apart. Cases which should show coupling if there is linkage are cross 54-1, 2, 4, 5, 6, 12, 15, 16, 17, and cross 72-5, 6, 7. Cases which should show repulsion are cross 61-1, cross 64-1, 3, and cross 72-1, 2, 8, 9, 10. TABLE 57. A, B C, D Sm Cross- Link- Rsrs. RSrs. rr — overs. ages. C, D Sm Coupling — RSrs X rsrs Repulsion — RsrS X rsrs 26 14 29 6 GO 16 26 6 29 14 32 43 The indication of linkage is too slight to be considered significant, especially in view of the excess of cross-overs in the F2 data. Thus there is probably no linkage between R and S. It is interesting to analyze the data with regard to possible linkages of these factors with 114 INHERITANCE IN GUINEA-PIGS. any of the 5 known sets of color factors. Four of the sets of color factors, those in which non-agouti (a), yellow (e), brown (b), and albinism (CJ are the lowest recessives, are known to be independent of each other (Part I). As regards the pink-eye factor (p), it is merely known that cross-overs occur between it and non-agouti, albinism, and yellow, and that it is not an allelomorph of brown (i. e., pink-eye X brown gives black-eyed intense young). Data on the possible repulsion of R and A are furnished by cross 72- 1, 2, 5, 6, 8, 9, 10 and cross 64-1. Coupling data are to be found in 64-3 and 66-1, 3. TABLE 58. Ag-Rf ARar Ag-Sm Arar B-Rf aRar B-Sm arar Cross- overs. Link- ages. Ag-Rf B-Sm Coupling — ARar X arar 4 7 11 3 18 7 Repulsion — AraR X arar 7 5 7 9 16 12 34 19 There is an excess of cross-overs. The most probable interpretation is that there is no linkage. Data on the coupling of A and S are to be found in crosses 70-1 to 9, 71-1 to 6, 72-1, 2, 3, 4, 5, 6, 8, 9, 10, 64-1, and 74-1. Repulsion data is to be found in 64-3 and 72-5. TABLE 59. Ag-C ASas Ag-A Asas B-C aSas B-A asas Cross- overs. Link- ages. Ag-C, D B-Sm (4-toe) Ag-Sm (£ cut) B-Rf A Coupling — ASas X asas 12 16 16 16 32 28 Repulsion — AsaS X asas 1 3 3 1 35 29 A and S are quite clearly independent of each other. Crosses 72-5 to 7 give a few data on the relation of R and S to C. We find in the case of R and C, 4 linkages to 4 cross-overs and in the case of S and C no linkages and 3 cross-overs, indicating probably independence in both cases. From cross 72-7 we find that cross-overs can occur between R and E, R and B, and S and B. From 61-1 we find that cross-overs can occur between P and R, and P and S. Summing up : R is quite certainly independent of S and A and cross- overs are known between it and all of the other known factors E, B, C, and P. S is quite certainly independent of A and cross-overs are known between it and B, C, and P, but not as yet E. It is hoped that more definite statements can soon be made on these points. ROUGH FUR. 115 SUMMARY OF ROUGH TABLES. Table 60, in which smooths are omitted, shows the closeness of fit of the hypothesis to the data as regards the inheritance of variations of the rough character. TABLE 60. Cross. 88 X S3 A B c D E 45 A 4-toe A 4-toe 10 46 A Tri A Tri .... 27 2 49 A 4-toe. Sm 4-toe 28 1 50 A Tri Sm 4-toe 19 58 B Lima B Lima 7 5 '59 A Lima Sm Lima 21 5 260 B Lima Sm Lima 5 9 63 A, B Lima Sm 4-toe, etc 4 4 365 A Misc Sm Misc 17 66 A Misc Sm Misc 14 1 73 A J cut Sm 4-toe 2 75 A J cut A 4 cut . . . 3 Total 157 27 Expectation ss 18 4 3 0 Cross. Ss X ss A B c D E 47 C Tri A Tri, 4-toe 10 1 5 1 451 Sm Tri A 6 3 3 64 C, D Tri . Sm 4-toe, etc 34 29 13 1 58 C Lima B Lima 1 1 559, 60 Sm Lima A, B Lima 4 1 3 61 C Lima Sm Lima 2 1 1 64 C ruf . hybrid Sm 4-toe, etc 3 2 665 Sm ruf. hybrid ... A 1 1 2 770 C, D £ cut A G. p 8 3 8 2 2 771 Sm $ cut A G. p 10 3 9 1 72 C, D 5, J cut Sm G. p . 12 6 6 74 Sm \ cut A \ cut 1 Total 90 11 69 ?,7 3 Expectation ss + Ss 10 0 1< )0 o Cross. Ss X Ss A B C D E 52 C Tri C Tri 18 6 19 7 12 76 C f cut C \ cut 1 3 1 Total 18 6 20 10 13 Expectation ss + 2 Ss + SS r r 3; J 17 1Omitting young of L24. 2Omitting young of L6, L24. 3Omitting young of A702, A605. 4Some of mothers may be ss or SS. 'Mothers L24, L6. "Mothers A702, A605. 'Including also one J cut. 116 INHERITANCE IN GUINEA-PIGS. TABLE 60 — Continued. Cross. ss X SS A B C D E 48 A Tri E Tri 2 1 56 Sin 4-toe E Tri 13 67 A 4-toe, tri Sm pure lea . . 5 68 A Srn pure cut 3 R Total 23 7 Expectation Ss C 3 0 0 Cross. Ss X SS A B C D E 53 C, D Tri . . E Tri 4 1 6 69 C, D Tri Sm pure cut 1 1 9 Total 5 2 15 Expectation Ss + SS c 1 1 1 11 Cross. SS X SS A B C D E 55 E Tri E Tri . . . 4 Expectation SS ( \ e I 4 The interpretation given is no doubt open to objections. In some cases the ratios seem rather aberrant. This is in part due to the small numbers, but also to the overlapping of class ranges. In most cases rough B must be considered as full-rough genetically (Rss) , but in some cases it is probably partial-rough (RSs) . Rough E usually seems to be RSS, but in some cases must be heterozygous (RSs). It has not been demonstrated that factor S of the wild species is identical with the similar factor of the tricolor stock. If not identical, however, the latter stock differs from the wild by two mutations which neutralize each other, while if identical we can consider that the original tricolor stock had simply persisted in the primitive condition, never having had the rough intensifying mutation, s, of the fancier's roughs. MINOR VARIATIONS. Probably part of the minor variations in roughness are due to chance irregularities in development which are not hereditary. This is indi- cated by the slight asymmetry not uncommonly present. This asym- metry seldom amounts to more than the absence of a member of one pair of rosettes. No Mendelian analysis has yet been attempted for minor variations, but certain hereditary differences between different stocks are quite ROUGH FUR. 117 clear. The Lima stock shows a distinctly lower level of development of roughness than is found in the 4-toe stock or even among the full- roughs of tricolor stock. A large part of the variation and overlapping in the remaining experiments in which various stocks have been mixed is made intelligible by assuming that the residual heredity is unfavor- able for roughness in the wild species and especially favorable in the 4-toe stock. If we let 2 + stand for favorable and S — for unfavorable residual heredity, the wild species and presumably the primitive guinea-pigs are rrSSZ — , while the good fancier's roughs, RRssS + differ by at least three independent sets of factors, all favorable for roughness. ROUGHNESS OF SERIES II. It has been mentioned that irregularities in hair direction have been found in certain stocks which can not be classified by the grades which have been defined. The BW race is a highly inbred race. No indi- viduals of the pure stock have ever been observed to have roughness on the face, back, or toes, but many of them show irregular partings and crests along the chest and belly. It will be remembered that in series I ventral roughness appears only in high-grade roughs — grades A or B. Thus the characteristic roughness of the BW stock is nearly the least characteristic feature of series I. The only distinction which has been made in these BW roughs is between strong-rough with two or more ridges and poor-rough with only one ridge or a mere trace of roughness. Table 61 shows the principal results. TABLE 61. Smooth. Poor rough. Strong rough. Smooth X smooth Poor X poor 11 14 6 1 G 1 Strong X strong 5 5 1G It is clear that this roughness is due neither to a simple dominant nor to a simple recessive. Aside from this, the results are exceedingly difficult to interpret, since poor X poor gives more smooth than does smooth by smooth. Probably the results will become more harmonious when more data are obtained. It seems safe to conclude at present that this roughness is wholly independent of ordinary roughness in its causation. Irregularity in hair direction on the back, not resembling anything in series I and not correlated with roughness of the hind toes, has been observed in a few individuals of Arequipa and lea stock. It does not seem to be like the BW roughness, but resembles the latter in the irregularity of its inheritance. 118 INHERITANCE IN GUINEA-PIGS. SUMMARY. The principal results which have been reached may be summarized as follows: 1. A classification of guinea-pig fur, skin, and eye colors is given with definitions of fur colors in terms of Ridgway's charts (1912). 2. Rodent color factors are conveniently classified as follows: a. Factors which affect the distribution and intensity of color largely irrespective of the kind of color. b. Factors which govern the differentiation between yellow and dark colors in colored areas of the fur. c. Factors which determine the kind of dark color in the areas with dark pigmenta- tion in fur and eyes, without influence on yellow areas. Definitions of all known guinea-pig color factors are given on this basis and a table of the color varieties arising from combinations of these factors is given. 3. Genetic and biochemical evidence on the physiology of pigment formation suggests the hypothesis that the three groups of factors determine respectively the distribution and rate of production by the nucleus of the following substances : a. A peroxidase which, acting alone, oxidizes chromogen in the cytoplasm to a yellow pigment but is so unstable that it must be produced at a relatively high rate to give any pigment at all. b. A supplementary substance which, united with the first, makes it a dark-pig- ment-producing enzyme and of such stability that color develops at a much lower level of production of peroxidase than when the supplement is absent. Above the level at which both produce effects, the dark and yellow-producing enzymes compete in the oxidation of chromogen. c. Additions to the second substance which cause variations in dark color but not in yellow or in the competition between dark color and yellow. 4. There is a continuous series of variations in intensity of pigmen- tation in the yellow, brown, and black series and in eye color. The ordinary dilute guinea-pigs are found to be imperfect albinos in the sense that dilution is due primarily to a member of the series of allelo- morphs— intensity, dark-eyed dilution, red-eyed dilution, and albinism, with dominance in the order of increasing intensity. 5. A further step in the analysis of the continuous series of variations of intensity is taken in the demonstration that dilution is imperfectly dominant over red-eye and albinism as regards the yellow series of colors, and that dilution and red-eye are imperfectly dominant over albinism, as regards the black series. Smaller effects are due to the residual heredity of different stocks and to age. 6. Evidence is presented which confirms the hypothesis of Detlefsen (1914) that the light-bellied agouti pattern of tame guinea-pigs, the ticked-bellied agouti of hybrids between the tame guinea-pig and Cavia rufescens, and non-agouti (as seen in self blacks or browns) form a series of triple allelomorphs in which light-belly is the highest dominant and non-agouti the lowest recessive. Evidence is presented which GENERAL CONCLUSION. 119 indicates that Cavia cutleri possesses the same agouti factor as tame agouti guinea-pigs. Light agouti of Cavia cutleri and dark agouti of Cavia rufescens are thus variations in a character in two wild species which differ in heredity by a clear-cut Mendelian factor.1 7. There is a continuous series of variations between smooth fur and very rough or resetted fur in guinea-pigs. The primary effects in this series are due to two independent pairs of allelomorphs. One factor, discovered by Castle (1905), is essential to any roughness of the common type, and is completely dominant over its allelomorph found in wild cavies and smooth guinea-pigs ; the other, an incomplete recessive to its allelomorph in the wild cavies and some tame guinea-pigs, is necessary for the higher grades of roughness. Second-order effects seem to be due to the residual heredity of different stocks, and probably to non-hereditary irregularities in development. There is a roughness of a different type from the usual which is inherited independently. GENERAL CONCLUSION. Most of the successful earlier attempts at Mendelian analysis of heredity naturally dealt with variations which were obviously dis- continuous. But in nature such variations are much less common than apparently continuous series of variations. It was thus a common reproach against the Mendelian analysis that it dealt only with excep- tional conditions. The work of Nilsson-Ehle, East, and others has shown how quantitative variation may be brought under a Mendelian explanation. MacDowell (1914) presents data on size inheritance from this standpoint and discusses the literature up to that time. Recently two very interesting papers have been published (Dexter, 1914, Hoge, 1915) which analyze the heredity of certain very variable char- acters in Drosophila by means of linkage relations. Several of the studies in this paper deal with inheritance in continu- ous series of variations. The only general statement which can be made about the results is that there is no general rule for such cases. Intermediates between varieties which mendelize regularly have been found to follow very definite modes of inheritance, which, however, are very different in different cases and could not possibly be predicted a priori. On the other hand, each mode of inheritance is exactly paralleled by cases among the most diverse groups of animals and plants. It may be interesting to summarize the modes of inheritance of inter- mediates which have been found. An intermediate condition is sometimes found to be due to an inter- mediate variation of the essential hereditary factor involved, i. e., to an allelomorph. Thus yellows are intermediate between red and albino 1 It should be pointed out, however, that the original stock of Cavia rufescens used in these experi- ments included individuals of the light-agouti character as well as those classed as dark agouti. It seems quite likely that dark agouti arose as a recessive mutation in C. rufescens. — W. E. C. 120 INHERITANCE IN GUINEA-PIGS. guinea-pigs in appearance, and we find an allelomorph intermediate in dominance between the intensity and albino factor to be responsible for their condition. Sepias are similarly intermediate between blacks and albinos and are due to the same allelomorph of intensity and albin- ism. The series, light agouti of Cavia cutleri, dark agouti of C. rufes- cens, and black, furnishes another example due to triple allelomorphs. In other cases, the intermediate type is an unfixable one, due to imperfect dominance. Thus cream is the heterozygote between yellow and albino. A "razor back" rough (rough C or D) is the heterozygote between a type smooth except for the hind toes (rough E) and a full- rough (rough A) . A series of deviations from the original type may depend on the presence of a certain factor necessary for any deviation whose effect is modified to different extents by independently inherited factors. Rough A contains the same rough factor (R) as does rough E, but differs in possessing an independent factor variation (s) favorable for rough- ness. Most of the variation which we have ascribed to residual heredity probably comes under this head. Deviations from type, which apparently form a natural series, may be due to wholly independent factors whose effects are merely super- ficially similar. A pink-eyed pale sepia superficially seems as good an intermediate between an intense black and an albino as does a black- eyed sepia, yet the former is due to a variation which is wholly inde- pendent of albinism; the latter is due to an allelomorph of albinism. White-spotted animals are sometimes called partial albinos and con- sidered as natural intermediates between the self-colored type and albinos, but genetically they are wholly distinct. Black, agouti, and self yellow form a series which is due to three allelomorphs in mice, but in guinea-pigs two wholly independent sets of factors are involved. Finally, we must recognize series of variations in which no Mendelian factors have yet been isolated. The series of white-spotted and yellow- spotted types and the series of polydactylous types are examples in guinea-pigs. Further, in all series of variations, to whatever extent analysis has been carried, there always remains some unanalyzed varia- tion. In many cases such variations are known to be hereditary and can be assigned to the residual heredity of particular stocks. Such unanalyzed variations, however, are probably in general complicated by variation which is not hereditary, due apparently to irregularities in development. If we can measure the importance of such non- hereditary variation by the extent of irregular asymmetry met with, it is very important in white and yellow spotting, in the variations in the development of extra toes on the hind feet, and is noticeable in varia- tions in roughness. In the continuous series of variations several of these phenomena have generally been found together. In the series from smooth to full- TABLES. 121 rough we find a primary unit difference, a modifying factor, imperfect dominance in the effects of the latter, effects of residual heredity, and probably some non-heritable variation. In the series from red through yellow and cream to white we find multiple allelomorphs, imperfect dominance, and small effects due to residual heredity. In the series black through sepia to white, we find independent factors, multiple allelomorphs which show imperfect dominance, and rather prominent effects due to residual heredity and to age. This last series is interesting as at least a close parallel in appearance to the series of variations in human hair — black, brown, tow-color, to white. Thus in each case a complex of the most varied causes underlies an apparently simple continuous series of variations. EXPERIMENTAL DATA. EXPLANATION OF TABLES 62 TO 137. Crosses 1 to 15 include all matings recorded by the writer which involve the inheritance of agouti and in which at least one of the parents had Cavia rufescens ancestry. A large part of the remaining crosses are non-agouti by non-agouti, producing only non-agouti young. All the young in which the agouti factor should produce a recognizable effect, if present, are classified under the heads Lb, Tb, and Non, which mean light-bellied agouti, ticked-bellied agouti, and non-agouti, respectively. Most of these are the typical (black-red) light-bellied or ticked-bellied agouti or black. Those which are not typical, e. g., brown-red agouti light-belly, red-eyed sepia, etc., are described further under the column "Remarks." Those young in which the agouti factor can produce no visible effect, even though present (albinos, reds, yellows, and creams), are described under the column "Unclassified." Thus the exact color of every one of the young from each mating can be found from the tables, with the exception that white and red spotting are not noted. The matings in each cross are numbered in the first column. The number, description, and descent of the mother and father are given in the second and third columns, respectively. As in the case of the young, black-red agouti light-belly or ticked-belly or black, depending on the heading of the column, are understood where no description is given. The descent is indicated in most cases by a reference to the mating from which the animal was derived. Thus 36-4 means mating 4 of cross 36. In other cases the stock is indicated as BB or BW. The symbol ArF2 means F2 from crosses of Arequipa cf 1002 with guinea- pigs. In some cases merely the amount of Cavia rufescens blood is given. Thus M49, in the first cross given, was an ordinary ticked- bellied agouti from mating 9 of cross la. Referring to this mating, we see that his parents were female 84, a black of BB stock, and male A1121, a ticked-bellied agouti with -^ Cavia rufescens blood. 122 INHERITANCE IN GUINEA-PIGS. Crosses 16 to 44 include all matings recorded by the writer in which there was dilution or red-eye in either parents or offspring, except for a few cases among the Cavia cutleri hybrids and cases of intense by dilute with only intense young. Some other crosses are included for special reasons bearing on the inheritance in the albino series. There is some repetition from matings outside of 16 to 44, but most of those outside are intense by intense, with only intense and, in some cases, albino young. As in the agouti crosses, all matings are numbered in column 1 . The number, description, and descent of the mother and father are given in columns 2 and 3, respectively. All the offspring are classified under the heads Int, Dil, RE, or W, which stand for intense, dilute, red-eye, and white (albino) , respectively. A further description of all except the albinos is given under the column "Remarks." The attempt has been made to give the grade of dilution at birth for every dilute or red-eyed animal where known. Crosses 45 to 57 give the data on the inheritance of rough fur in the 4-toe and tricolor stocks. As before, the matings are numbered. The young are classified under the heads A, B, C, D, E, and Sm, which refer to the grades of roughness defined in the paper and to smooth. The parents and offspring were black (usually with red and white blotches) except for a few cases which are all noted. Such a symbol as red-B means a red of grade rough B. Crosses 58 to 62 give the results in the pure Lima stock and 63 the results in the cross of Lima with other stocks. Where no color is given black is always to be understood. Crosses 64 to 66 give the matings involving rough fur among Cavia rufescens hybrids which were recorded by the writer. Cross 67 gives crosses of pure lea with rough A stock. Crosses 68 to 78 give all the data in matings involving Cavia cutleri ancestry made by the writer. The following symbols are used: AgLborAg = Black-red agouti, light-belly. AgTb = Black-red agouti,ticked-belly. B = Black. BrAgLb = Brown-red agouti, light-belly. BrAgTb = Brown-red agouti, ticked belly. Br = Brown. R = Red (black-eye). R(Br) = Red (brown-eye). SYAgLb = Sepia-yellow agouti, light- belly. SYAgTb = Sepia-yellow agouti, ticked- belly. Sep, S = Sepia. BrYAgLb = Brown-yellow agouti, light- belly. BrYAgTb = Brown-yellow agouti, ticked- belly. LBr = Light brown. Y = Yellow (black-eye). Y(Br) = Yellow (brown-eye). Cr = Cream, used in compounds likeY. SAg(R) = Sepia- white agouti (red-eye). Sep(R) = Sepia (red-eye). W = White or albino. Red (p) = Red (pink-eye). = Sepia (pink-eye). Sep (p) In such expressions as S3Y3Ag the numerals stand for the grades defined in the text. In crosses 1-15, Lb and Tb are used at the heads of the columns to include any light-bellied or ticked-bellied agouti. Non means non-agouti. A, B, C, D, E, and Sm are used for grades of roughness and for smooth. TABLES. 123 TABLE 62. Cross /. — Matings of non-agouti (aa) with ticked-bellied agouti (A'a). Each of the latter known to be heterozygous because of a non-agouti parent. Expectation: A'a X aa = A'a -f aa (1 AgTb : 1 Non-Ag). la. Mother non-agouti, without rufescens ancestry. No. 9 Non-Ag. rfAgTb. Lb. Tb ^on Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 399 BW... 499 4-toe. M49 la Do _g 9 5 3 2 4 1 6 9 1 9 8 9 2 3 2 4 4 2 5 6 W 1 W 5 W 1 W 2 W 399 BW. . . 399 4-toe. . B5 ld-16. Do 299 BW.. . 65 BB . . . 299 BW. . . 399 BW. . . 84 BB . . . C22 Misc.. C35 Misc.. 399 4-toe. . 3W BW... 3 W Misc . . B27 Id B30 la B69 la B191 la A1121 & A1474 ^ . ...Do -14. -1 -1 2 W 4 W -5 2 8 3 2 1 1 1 4 B171 la B117SCrAgTb Id Do -4 6 2 3 1 2 1 W 5 W 1 W 2 W -11. 3 SCrAgTb, SCrAgTb, BJ 3 Sep. SCrAgTb 2 Sep .... •CrAgTb, D44Sep 16a-3. AA244 Sep 2-12 Do . .Do . . SCrAgTb BW 43 W BW. . . Sep (R) S.Am . Total D113BCrAgTb 3a-7 . AA433a 36-4 . SCrAgTb, S 3 AgTb, SAg Sep(R). up Tb(R), 2 1 W 62 62 Ib. Mother non-agouti, with rufescens ancestry. No. 9 Non-Ag. c^AgTb. Lb Tb Non Remarks. Unclassified. 1 2 3 4 5 6 7 8 9 10 11 A443 | A469 | 1 1 2 1 2 1 LE r . . A1390 3*5 A1227 W e1? • • • A1050 3*2 A781 s^ 3 1 1 3 3 1 4 1 SC rAgTb /A1413 gS, \A1291W & A1309 W gV !>A1449 *"*. W R, R(Br),Cr(Br) A1513 3^ Br Br AgTb . . A 1407 gg . A1449 J* A1413 0*5 ... Do M115W & /M114 lb-7... \M90Br M189 lc-3 \ Do . 4 1 SC Br rAgTb, Sep . . / Do CrAgTb, Sep Cr W M90 Br rU. A1330 rig---- Total A1331 fis 17 13 124 INHERITANCE IN GUINEA-PIGS. TABLE 62 — Continued. Ic. Male genetically, but not visibly ticked-bellied agouti. No. 9 Non-Ag. cfAgTb. Lb Tb Non Remarks. Un,cl,as- smed. 1 2 3 4 5 6 7 8 9 131 W G.p.. A412R(] 13a G. p Do Br) A, 2 Se 3 ... 20 G. p. . Do 3 3 1 1 3 1 1 SC . 3fc Se . SC SC 'rAgT SCrAj p "b 58 Sep Dil . . B42 W 17 30Cr(B) Dil Do la-3 . . /Tb 2 W 55 Cr(Br) Dil Do 1 2 1 1 JrAgl 'rAgT "b W M292 ^j... D18W M326 t'g Do . lc-5 . . ^b, 2 Sep. 4 W M353 /4 ... -Do Se Total 11 12 Id. Male non-agouti, without rufescens ancestry. No. 9 AgTh. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 ^606 AgLb i 166 4-to 2966 BB 3013 BB Do e 2 1 3 5 2 6 1 2 2 5 1 ^450 AgLb i ll 4 6 1 7 1 A340 ^ A341 r"« A357 j^g Do A1146 i'e . . . .Do A105H g'j . . . .Do A1171 £, . . Do . A1058, A1171 /2 2996 BB 1357 BAV Do 3 5 2 1 2 2 3 1 1 9 1 3 5 2 o Do . . 2 W A1117 ^3 SCrAgTb Do 2996 BB 3013 BB 1357 BW Do Do ... 4 1 1 1 2 11 A 1450 3*2 A1117, A1450 jW W A1582 JT 2996 BB Do A 1583 fa A 1677 -^j Do A 1678 6*5 Do B8 ld-7 Do 2 7 1 1 2 B23 ld-12 Do B2(j ld-14 Do Ml 13 16-7 C20 Mis 86 W BW c M442 BrCrAgTb lb-10 2 Sep Total ll 61 63 TABLES. 125 TABLE 62 — Continued. le, Male non-agouti (genetically) with rufescens ancestry. No. 9 AgTb. c?1 Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 A1146 Jg. A504 W TV 2 2 Sep B132 ld-3 M293 Y 42-14 2 1 SCrAgTb B95 ld-4 Do B24 ld-4 Do. . . . 1 B52 la-3 M201 W 42-13 2 2 B33 ld-18 Do W W W B110 la-1 Do Bill la-1 Do . 1 1 1 2 B128 la-1 Do B23 ld-12 Do Total 10 5 'AgLb, but agouti known to be derived from C. rufescens. SUMMARY OF CROSS 1. No. aa A'a Lb Tb Non la 9 9 Non-Ag (g.p.) .... cfcfAgTb 62 62 Ib 9 9 Non-Ag (hybrid) . . cfVAgTb 17 13 le 9 9 Non-Ag c^cfA'a (R or W) 11 12 Id cf c? Non-Ag (g.p.) .... 9 9 AgTb ... 1 61 63 le d1 d1 aa (Y or W) hybrid 9 9 AgTb . . 10 5 Total 1 161 155 TABLE 63. Cross 2. — Matings of ticked-bellied agouti (A'a) with ticked-bellied agouti (A'a), in which both are known to be heterozygous because of the parentage in each case. Expectation: A'a X A'a = A'A' + 2A'a + aa (3AgTb : 1 Non-Ag). No. 9 AgTb. cf AgTb. Lb Tb Non Remarks. Unclas- sified. 1 B15 ld-6.. B118 ld-6.. 1 2 2 B58 ld-15. Do 9 2 3 BrAgTb, 3 B59 ld-15. Do 6 4 SYAgTb.Sep 2 BrAgTb 4 B68 la-1 . . . ..Do 5 2 SCrAgTb 5 A529 BrAgTb & AA15 A 6 1 BrAgTb, SCr 6 A913 3*5 Do 5 AgTb, BrCr AgTb BrAgTb 7 A1273 SCrAgTb & AA16 gV . • 4 2 3 SCrAgTb . . . 3W 8 Do A1121 gV- • 1 9 A780 h . A781 g1? • 5 BrYAgTb . 3R.Y 10 A1306 lis--- A1307 1 2W 11 A1561 & A1050 3*5 . . . . 3 4W 19! A1566 & Do 5 1 BrAgTb, SCr }?,n A1566 AA15 A.. 2 1 AgTb, Sep SCrAgTb 13 A702 A . AA16 gV. . 2 2 Br . 14 A1450 3*5 AA433a 36-4 1 15 A1058 3*2 Do W 16 A1523 Jj . . A1449 g'j 2 1 BrAgTb W 17 AA176 41-4 AA177 SCrAgTb 41-4 5 1 SCrAgTb W 18 AA175 41-4.. Do W 19 M78 9-5 A1161 & 2 20 MHO 16-6 A1170 g's 1 9,1 D26 SCrAgTb lc-4.. D33 SCrAgTb lc-6 . W Total 6fi 19 126 INHERITANCE IN GUINEA-PIGS. TABLE 64. Cross S. — Matings of ticked-bellied agouti from cross 2 or 12 (A'A', 2A'a) with non-agouti (aa) made in order to test for the presence of homozygotes. Expectation: A'A' X aa = A'a (all AgTb) or A'a X aa = A'a + aa (1 AgTb : 1 Non-Ag). 3o. Heterozygous females. No. 9 AgTb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 »4 5 16 17 18 M19 2-16.. M203 2-19 . . AA211 2-6... AA240 12-8.. AA257 12-2.. AA285SCrAgTb 12-7. . AA242SYAgTb 12-8.. AA240, AA242 12-8. . - r r. '#- 7 females 393 4-toe .... 2 1 1 2 1 2 1 2 1 MllGSep 42-11 C20 Misc. . . . A1040 iV 2 2 1 1 2 A1040, 356 ^6.4-toe. 393 4-toe I5Sep(R) 21-1 BCrAgTb, 2 Sep . A1040 ^ 10 11 36. Possible homozygous females. No. 9 AgTb. d"1 Non-Ag. Lb Tb Non Remarks. f Unclas- sified. 1 2 3 4 5 AA209 2-11.. AA212 BrAgTb 2-6... AA213 2-12.. AA217 2-7 C21 Misc 4 2 8 8 3 C20 Misc C21 Misc Do AA298 2-13 . . 5 females 356 4-toe 25 JNot certain that both parents were heterozygous (A'a). TABLES. 127 TABLE 64. — Continued. 3c. Heterozygous males. No. 9 Non-Ag. cfAgTb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 M79 W 3V • • M77BrAgTh 2-16... 1 2 1 2 1 BrAgTb, Br.. . BrAgTb, Br.. . 3 W fM72 lc-2 . . . \M86 9-1 ... M42 LBr 42-12 . . M44 Cr 42-12 >.. .Do. / AA197 2-10 Do 3 1 2 5 1 SCrAgTb M99 42-13? Do M101 42-13? Do 1 3 [A 1407 ^5 JAA199 SCrAgTb 2-12 \A1413 & S6 zfcf.... S15 ,t~ AA223 BrYAgTb 2-9 Do 2 2 Br A1659 »i~ Do 2 2 B 9 Misc Do BrAgTb S2 Misc. . . A1665 Misc AA226 2-13 1 2 1 13 Do . 1 3 n 2 SCrAgTb . . 1 malG 12 JA litter with an unexpected AgLb. Paternity not wholly certain. 2Not certain that one of parents was heterozygous (A'a). SUMMARY OF CROSS 3. AgTb, from cross A'a X A'a, tested by cross with Non-Ag. Lb. Tb. Non 3a 799 A'a 10 11 36 59 9 ... A'A'(?) 25 3c 9cfc? . A'a 1 40 38 3d 1 d* . A'A' 12 128 INHERITANCE IN GUINEA-PIGS. TABLE 65. Cross 4- — Ticked-bellied agoutis, known to be homozygous because of test (cross 3), or parentage (cross 4), crossed together. Expectation: A'A' X A'A' = A'A' (all AgTb). No. 9 AgTb. 0*AgTb. Lb Tb Non Remarks. Unclas- sified. 1 AA213 2-12 AA253 SCrAgTb 2-7 9 2 SYAgTb, 2 AA217 2-7... Do . . . 11 SCrAgTb, BrAgTb, BrYAgTb, BrCrAgTb 3 SYAgTb, 3 AA613 4-1 ... ... Do 3 SCrAgTb 3 SCrAgTb . . . 4 AA671 4-1 ... Do W 5 AA577 4-2 AA573BrAgTb 4-1 1 BrAgTb 6 AA606 SYAgTb 4-2 Do 2 SCrAgTb Total 26 TABLE 66. Cross 5. — Homozygous ticked-bellied agouti (A'A') crossed with heterozygous ticked-bellied agouti (A'a) or with non-agouti (aa). Expectation: All AgTb (A'A' or A'a). No. 9 AgTb or Non-Ag. tfAgTb. Lb Tb Non Remarks. Unclas- sified. 1 M181 BrCrAgTb 15-15 AA253 SCrAgTb 2-7 1 SYAgTb ... . W ? Do AA573 BrAgTb 4-1 5 3 BrAgTb, 2 W 3 M442 BrCrAgTb 16-10. . AA573 BrAgTb 4-1 ... 1 2 BrCrAgTb BrCrAgTb.. . 4 M296 SAgTb 14-4 . . . Ml 16 Sep 42-11. 3 3 SYAgTb . . . fi D194Sep(R) 26-2.. AA670 SCrAgTb 4-1 3 2 SCrAgTb, 2 W SAgTb(R) Total 13 TABLES. 129 TABLE 67. Cross 6. — Matings of non-agouti hybrid (aa) with homozygous light-bellied agouti (AA). Expectation: AA X aa = Aa (all AgLb). 60. Female non-agouti. No. 9 Non-Ag. cfAgLb. Lb Tb Non Remarks. Unclassified. 1 2 3 A605 1. . 2597 G.p Do ... 2 2 5 A642 i A842 1 Do Total 9 66. Male non-agouti hybrid. No. ¥ AgLb. c? Non-Ag. Lb Tb Non Remarks. Unclassified. 1 2 3 4 5 15a G.p . . 3520 Cr(Br) G.p . . 3a G.p . . lla G.p.. 3392 G p . . Total A674 Sep J . . . Do 1040 jV. A504 W tV . . A1539 iV.. 6 3 7 3 1 2 SCrAg, BrCrAg . Y(Br),2Cr(Br),W SCrAgLb. . . 20 SUMMARY OF CROSS 6. No. aa (hybrid). AA (g.p.). Lb Tb Non 6a 9 9 Non-Ag. . . . cf a" Ag Lb 9 6b cf d" Non-Ag .... 9 9 Ag Lb 9,0 Total W 130 INHERITANCE IN GUINEA-PIGS. TABLE 68. Cross 7. — Matings of non-agouti (aa) with light-bellied agouti (Aa) of rufescens ancestry, known to transmit non-agouti, because of a non-agouti parent. Expectation: Aa X aa = Aa + aa (1 AgLb : Non-Ag). 7a. Female AgLb. No. 9 AgLb. cf Non- Ag. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 A601 j*g 103 4-toe... 224 4-toe . . . 1 1 2 3 2 1 W A614 jJg A953 3*2 A718 & 2 A1310 3^ 166 4-toe... A1311 3*2 Do 1 A1324 & A719 W 3^ . 3 W M102 66-1 A462 W ^s 2 2 2 Sep. Do M2 & 3 3 SCrAgLb /M357 SCrAgLb 106-7 \D95 SCrAgLb 106-8 D61 SCrAgLb 13-5 J20W BW .... BW 36 W BW 3 SCrAg, 2 Sep . 7 W W D63 SCrAgLb 13-5... Do . 2 2 2 1 2 SAg(R), Sep, Sep(R) 2 SCrAg, Sep, . . /D69 SCrAgLb 18-5 \M425 SCrAgLb 13-7 Total } . Do W 14 18 7b. Male AgLb. No. j 9 Non-Ag. c?AgLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 67 G.p . . M123 13-2. . D43Sep 16a-3. D94 SCrAgLb 106-8. M236Sep(R) ArF2. . M331 BrdAgLb 42-10. D45Sep 16a-3. D94 SCrAgLb 106-8. Total 2 1 1 1 SCrAg, Sep ... W W 1 4 1 SCrAg, Ag.... 3 SUMMARY OF CROSS 7. No. Aa (hybrid). aa Lb Tb Non 7a 9 9 AgLb.... cfcf Non-Ag. . 14 18 76 cfc? AgLb.... 9 9 Non-Ag . . 4 3 Total 18 21 TABLES. 131 TABLE 69. Cross 8. — Matings of ticked-bellied agouti (A 'a) with homozygous light-bellied agouti (AA). Expectation: AA X A'a = AA' + Aa (all AgLb). 80. Female AgLb No. 9 AgLb. c^AgTb. Lb Tb Non Remarks. Unclas- sified. \ 03 G p Bo ld-16 5 2 3392 G p Do ... 4 SCrAg 3 02 03 G p A1155 ^g 4 4 5a G p A1474 A 8 5 20a G.p. . . . .Do 4 SCrAg Total 25 86. Male AgLb. No. 9 AgTb. c?AgLb. Lb Tb Non Remarks. Unclas- sified. 1 B33 Id-lS 2597 G.p 3 2 B36 ld-18 Do 5 3 B37 ld-20 Do 2 4 B52 la-3 . .Do 3 5 /B40 W la-3 ) Do 6 6 \B37, B33 above A702 ^9 Do 2 7 A913 g's Do 2 8 ^AGOG SYAgTb 4-2 724 SAg(R) (lea) o 2 SYAgLb . . Total . . . . 25 Total cross 8 50 'AAGOG was A'A' TABLE 70. Cross 9. — Matings of light-bellied agouti (Aa) with ticked-bellied agouti (A'a), both known to be heterozygous with non-agouti. Expectation: Aa X A'a = AA' + Aa + A'a + aa (2 AgLb : 1 AgTb : 1 Non-Ag). No. 9 AgLb. cTAgTb. Lb Tb Non Remarks. Unclas- sified. 1 3015 G.p. A1474 rW 6 2 5 3R 2 M46 BrCrAg 13-3 ... . A1170 ^ 2 BrAgLb W 3 M57 13-1 Do 1 1 4 A1310 3*2.. A1449 3*2 1 1 Cr 5 Do A1513 3*2 1 6 A1311 3"^ A1449 g^ 1 7 Do A1513 ^2 5 1 2 8 A 1026 3*2 A 1050 3*2 2 1 Total 16 6 10 132 INHERITANCE IN GUINEA-PIGS. TABLE 71. Cross 10. — Matings of light-bellied agouti from such crosses as 8 and 9 (AA', Aa) with non- agouti, made in order to test whether light-bellied agouti can transmit both ticked- bellied agouti and non-agouti. Expectation: AA' X aa = Aa + A'a (1 AgLb : 1 AgTb). or Aa X aa = Aa + aa (1 AgLb : 1 Non-Ag). 10a. Females Aa. No. 9 AgLb. cf Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 B120 86-2 C20 G.p 2 3 2 B140 8a-2 M328B-Y 42-17... 3 1 3 Do 20 W BW . . 3 2 4 M195 9-7 . . MllGSep 42-11 2 2 Br . . 5 M217 8a-3 356 4-toe 1 2 6 M282 15-12 MllGSep 42-11 2 2 Sep. . 7 A1562 W fa AA83 B'Z . 2 8 A1691 86-7 Do 6 6 7 females 17 20 106. Females AA'. No. 9 AgLb. d"1 Non-Ag. Lb Tb Non Remarks. Unclas- sified. 1 A389 BrAgLb T18 ; . . . . A511W ^ 1 2 W.R 2 A499 BrAgLb &. . . . AA83 A . 1 2 BrAgLb.BrAgTb R 3 A 1688 86-6.. Do ?, 3 BrAgTb 4 A1690 86-7.. Do . . 1 5 3 R 5 B139 8o-2.. M328B-Y 42-17. 3 2 3 SYAgLb 6 Do 20 W BW 2 SCrAgTb 7 8 B 141 SCr AgLb 8o-2.. Do M328B-Y 42-17.... 20 W BW . 1 2 3 1 SCrAgLb, 3 SY AgTb 2 SCrAgLb, SCr 6 W 9 M25 9-1 . 356 4-toe 2 3 AgTb 10 M27a 9-1 ... 393 4-toe .... 2 5 11 M82 9-7 ... M116Sep 42-11 2 3 SCrAgLb, BrY 12 M92 8o-4 . . Do 1 1 AgLb, SYAgTb 10 females . 18 30 TABLES. 133 TABLE 71 — Continued. lOc. Males Aa. No. 9 Non-Ag. cTAgLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 75 BB 08 W 4-toe A581 &... Do 2 2 2 3 9 2 3 4 3 2 1 BrAgLb 4R.Y? 09 W 4-toe Do M7 T1* M133 8o-4... } Do SCrAgLb, Sep . . . . fM7 A. 1 A A58 TV Do 1 1 2 4 1 3 1 1 1 Ml 83 j*5 M261 Sep 7 Do 9 7 1 D240 BCrAgLbl4-5. . SCrAg.Sep 26 33 lOd. Males AA'. No. 9 Non-Ag. cfAgLb. Lb Tb Non Remarks. Unclas- sified. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 31 Mis M255 la- M253 la- M256 la- M254, M256 la- AA279 3a- C22, AA278 G.f B 9 9 (ab M168 3^ c M138 9-1 2 5 3 1 2 1 10 Do 4 4 1 1 1 4 5 3 1 1 10 Do 10 Do 10 Do 3 Do >., 3a-3 ive) Do Do 5 2 5 4 1 1 3 2 2 2 4 2 3 M91 8a-4... \ . Do. . SCrAgLb /M169 3>j /BrYAgLb, 2 BrCr I AgTb SYAgTb I IM171 J* / Do /' M177 lc-: M86 9-1 M79 W 3*2 > M210 15-14 Do 3 2 1 4 1 5 3 1 BrAgLb B31 lo- B53 la- B54 la- C29 G.p B 9 9 (ab M119 J* 1 B121 8b-2... Do 3 1 Do Do sve) Do B130 8a-l. . . Do AA174 14- 5 males 1 45 50 SUMMARY OF CROSS 10. No. AgLb (Aa or AA') tested by cross with Non- Ag (aa). Lb Tb Non 10a 7 females Aa 17 20 lOc 5 males Aa 26 33 Total 43 53 106 10 females AA' IS 30 lOd 5 males AA' 45 50 Total 63 80 134 INHERITANCE IN GUINEA-PIGS. TABLE 72. Cross 11. — Light-bellied agoutis (AA') crossed together. Both parents known to carry recessive ticked-belly by test (except in the case of AA533 with one young). Expectation: AA' X AA' = AA + 2AA' + A'A' (3 AgLb : 1 AgTb). No. 9 AgLb. cfAgLb. Lb Tb Non Remarks. Unclas- sified. 1 A1690 86-7 M91 8a-4 4 2 2 M25 9-1 Do 7 3 4 SYAgLb 3 M27a 9-1 Do 5 4 M25, M27a 9-1 Do 4 5 B139 8a-2 Do . . . 1 6 M25, M27, B139 . . Do . 1 3 BYAgTb 7 AA533 11-1 Do 1 8 AA588 11-2 Do 2 1 SYAgTb Total 25 9 TABLE 73. Cross 12. — Miscellaneous matings of ticked-bellied agouti with ticked-bellied agouti. No. 9 AgTb. cfAgTb. Lb Tb Noii Remarks. Unclas- sified. 1 2 AA203 BrCrAgTb 2-5. ... M19 2-16. . AA16 & Do . 3 3 1 SCrAgTb, LBr W w 3 AA497 SYAgTb 10d-ll AA284 12-7. 1 1 LBr R 4 AA598 3c-22. . Do . 2 BrAgTb W 5 A1523 3*5 AA199 SCrAgTb 2-12 3 SCrAgTb 6 M203 2-19... AA284 12-7... 3 2 Sep 7 AA202 2-7 AA15 ^ 4 2 SCrAgTb 8 AA206 2-9 . . AA177 SCrAgTb 16-3. . 7 4 SCrAgTb 9 Do AA507 3c-22 . . 1 10 AA342 12-8 Do 4 11 A1058 g^ M298 15-16.. 2 BrAgTb . . . 12 AA242 SCrAgTb 12-8 . . . Bl 17 SCrAgTb leZ-11.. 1 SCrAgTb . . TABLE 74. Cross 13. — Miscellaneous matings of light-bellied agouti guinea-pig with non-agouti hybrid. No. 9 AgLb. d* Non-Ag. Lb Tb Non Remarks. Unclas- sified. T3392 G.p.... , 3 1 SCr AgLb 1 \3444 G.p.... fOl G.p.... ^A1539 iV 5 1 3 R 2 >A426 R(Br) j^ 3 \02 G.p.... 3256 BrY AgLb G.p . A678 W iV • • • • 7 2 7 BrCrAgLb, 2 LBr 4 3220 BrAgLb G.p . Do 2 2 BrAgLb 5 271 SAg(R) G.p . M333 Y f$ 4 4 SCrAgLb 6 3a G.p . M34 Sep 16c-l 2 2 Sep 7 241 SAg(R) ArF*. M328 B-Y 42-17 3 3 SCrAgLb 8 29 9SAg(R) ArF2 M156R ils 4 3 2 SAg(R) 4 W 9 12, 16 SAg(R) 21-1 M34 Sep 16c-l 1 2 SCrAg Sep TABLES. 135 TABLE 75. Cross 14. — Miscellaneous matings of light-bellied with ticked-bellied agouti. No. 9 AgLb. cfAgTb. Lb Tb Non Remarks. Unclas- sified. i /20a, 5a G.p. . . . >4412 R(Br) A 3 1 3R 2 13015,3014 G.p 20a 5a 3014 G p A1474 Js • 3 1 3 AA173 14-1 AA199SCrAgTb 2-12. . 1 1 4 M82 9-7 A1161 3"z . ?, SCrAgTb 5 499 SA°-(R) ArFo AA508 3d-2 . . . 13 1 4 9 SCrAg, BY AgTb TABLE 76. Cross 15. — Miscellaneous crosses involving the inheritance of agouti in rufescens hybrids. No. 9 Misc. c? Misc. Lb Tb Non Remarks. Unclas- sified. 1 AA171 R 14-1 AA1161 AgTb 3^ 1 R, Cr 2 Ml 37 R 9-1 AA286 AgTb 3c-4 1 3 Do M29 Br 16-6 ? 4 M181BrOAgTb 15-15 C20 B G.p 6 5 A1472 BrAgLb TV 163 B 4-toe 9 6 f A1562 W iV ... JAA83 B g^ 4 1 8 \A1688AgLb S6-6. . /M27a AgLb 9-1 ... J393B 4-toe.. 3 6 2 9 \M19AgTb 2-16.. f Ml 75 AgLb 8a-3.. 1 A 1040 B rW 1 4 10 \AA242 SCrAgTb 12-8 . . [M195AgLb 9-7... |M116Sep 42-11 1 2 11 \M203AgTb 2-19.. ?B122AgLb 86-2.. JC20B G.p 9 9 2 12 \AA212BrAgTb 2-6... fM92AgLb 8a-4.. JA1513 3^ 1 ^ 1 BrAgTb 2R 13 \M106Y? 10c-2. /M85 R 16-5.. >AA1161 AgTb 3*2 1 1 BrAgLb . W 14 \M82AgLb 9-7.. f M56 AgLb 13-1 . . JA1170 AgTb 3^ 1 9 15 16 \M50AgTb 16-7.. AA28 W sS - . . . . A1523 AgTb 51* A1513 AgTb s^ . . . . M83 AgLb 9-7 2 1 3 1 BrAgLb, SCrAg Tb, BrCrAg Tb R, 2Cr 17 A1413 B ^ Do ? 18 M84 R(Br) 16-5 Al 161 AgTb s1*.... W 19 TA556 AgLb t^ \ A587 W jJg [l04 B 4-toe 9 3 [A533 Y ^'. '. '. '. f A495 AgLb ^6 1 3R 20 \A867B ^ /AA621SYALb 11-2 >163 B 4-toe.. BW36 W B W . . . 1 SCrAgTb 21 \AA621 SYALb 724 SAg(R) lea S 3 SYAgTb 22 23 39 9SAg(R) ArF2. . 198W ArF-2 M224 BrAgLb 9-2 ... M291 B A 21 1 1 10 SAg(R), SAg Tb(R) W 24 29 9 W ArF2 M2B ^ 2 1 3 W 25 D125 W la-13 133 SAg(R) 24-1 1 1 SAg(R) Sep(R) 26 D427 W la-14 133 SAg(R) 1 1 SAgTb(R), Sep 27 D86 W 76-3 I9G BWAg(R) >:>4-2 9 3 (R) 2 SAg(R) 3 Sep (R) 136 INHERITANCE IN GUINEA-PIGS. TABLE 77. Cross 16. — Matings of dilute with albino of intense stock. Expectation: CaCd X CaCa = CdCa (all Dil). CdCr X CaCa = CdCa + CrCa (1 Dil : 1 RE). CdCa X CaCa = CdCa + CaCa (1 Dil : 1 W). 16o. Females CdCd. No. 9 Dil. cfW (intense stock). Int Dil RE W Remarks. 1 AA621 S3Y2Ag 39-4 . BW36 W BW 1 S8Cr6AgTb 2 58 Seps Dil 75 W BW 2 2 Sep3 3 Do 20 W BW... 3 2 Sep3, Sep3-Cr5 4 Do B42 W la-3 . . . 3 3 S6Cr6AgTb Total q 166. Females CdCa. No. 9 Dil. cfW (intense stock). Int Dil RE W Remarks. 1 15 Sep6 Dil 75 W BW.... 4 3 2 Sep3, 2 Sep4 2 17 Cr6 Dil B42 W la-3 . . . 1 Sepg 3 30 Cr6 Dil . ... . . . .Do 2 4 55 Cr6(Br) Dil . ... .Do . . 1 1 S6Cr6AgTb 5 fM42LBr 42-12... [l5W BW.... 8 4 8 Sep6 6 \M44Cr6 42-12... M42 LBr 42-12 . . . Do 3 3Sep4 7 AA600 LBr-Crj 39-19 . . . Do 1 Sep4 8 M357 S4CrsAg 42-1 20 W B W . . 3 3 S3Y4Ag, SBCr6Ag 9 B141 S4Y4Ag 39-23... . . . . Do 3 fi Sep4 SoCr5AgTb, 2 S3 10 D43, D44 Sepa 16o-3. . . .Do 4 Y4Ag 11 D45 S3-Cr6 16a-3 .Do 1 1 Sep-i— CrB 12 D95 S3Y4AK 166-9 20 W B W 7, 12a D95, M357 Do 2 7, Sep3, S3Cr6Ag 13 D67 S3-CrB 16c-4 . . . Do 2 14 M384 Sep5 39-12. 80 W BW .... 3 1 Sep6, 2 Sep4 15 M442 BrCr6AgTb 39-12 .Do 2 Sep4, Sepj 16 D66 S>-O6 406-13 .Do 1 1 Sepo Total ... . 33 32 16c. Males CdCa. No. 9 W. cfDil. Int Dil RE WT Remarks. 1 132 W 4-toe.. . A674 Sep6 i 4 3 Sep6-Cr6, Sep 2 12a W 4-toe M34 Sep6-O6 16c-l. . 3 3 Sepg-Cre 3 5 W BW Do ? 1 2 Sep6 4 82 W BW B117 S4Cr6AgTb 39-14 9 5 S3-Cr6, S3Cr6Ag 5 fBWHW BW 1. . , .Do. 3 Tb 2 SsCr6AgTb fi \BW15W BW 120, 21, 23, 29 W 22- / 13 Cr8(Br) Dil 8 13 Sep4 2 S6Y4Ag, 3 S4- Y4, 2 S4> S» Total ?9 1Q TABLES. 137 SIMILAR MATINGS FROM CROSSES 19, 27, AND 33, AND SUMMARY OF CROSS 16. No. Dilute. W (intense stock). Int Dil RE W 599 Dil (CdCr) W BW 9 10 3 C? cf Dil (CdCr) Do 3 5 7 9 9 Dil (CdCa) . . .Do 5 7 1 cf Dil (CdCa) . . . . . .Do 4 2 16a 9 9 Dil (CdCd) W 9 166 9 9 Dil (CdCa) - . Do 33 32 16c d" d" Dil (CdCa) Do ?,?, 19 Total 85 15 60 TABLE 78. Cross 17. — Matings of intense from intense stock with albinos from dilute stock or from two dilute parents. Expectation: CC X CaCa = CCa (all Int). CCa X CaCa = CCa + CaCa (1 Int : 1 W). 17a. MaleCC. No. 9W. cFInt. Int Dil RE W Remarks 1 M117 W 42-11 3013 B BB. . 2 2 B 2 M327W 42-17 Do 2 2 B 3 D 37 W Dil Do 2 2B Total 6 176. Female CC. No. 9 Int. cfW. Int Dil RE W Remarks. 1 22, 23, 33 Ag Misc 11 W Dil 11 11 Ag 2 CIS, CSOAg Misc. Do 6 6B (C24 Ag Misc. \ 3 .Do 4 Ag, 3 B \C34 B Misc. f 4 22 Br Misc Do 4 4Br (S22 B ,to . 6 >M313 W 42-16 4 4 B 1A1665B -is 7 3223 B Misc Do 2 2B 8 B232 B ld-21 M201 W 42-13 . . 2 2B 9 B23AgTb ld-12 ... .Do 1 AgTb Total 34 17c. MaleCCa. No. 9W. 9 Int. Int Dil RE W Remarks. 1 D37 W Dil 06 B BW 2 2B 2 9 W Dil Do 1 2 B Total 3 2 INHERITANCE IN GUINEA-PIGS. TABLE 78 — Continued. 17d. Female CCa. No. 9 Int. cfW. Int Dil RE W Remarks. 1 D48 Ag 176-1 . . . 11 W Misc 2 2 D49 Ag 176-1 . . . . .Do .... 1 2 Br. 3 D224 Ag 176-1 Do 1 1 R 3a D224, D226 Ag 176-1 . Do 3 ?, Ag, 2B 4 B33 AgTb ld-18 M201 W 42-13 2 1 2 AgTb 5 B52 AgTb la-3 Do 2 2 AgTb 6 B110 AgTb la-1 Do 1 7 Bill A°Tb la-1 Do 3 1 AgTb, 2 B 8 B128 AgTb la-1 . . .... Do 1 AgTb Total . 13 10 SUMMARY OF CROSS 17. No. Intense (intense stock). White (dilute stock) . Int Dil RE W 17a cf cflnt CC.. W 6 176 9 9 Int CC... . Do 34 17c cfcflnt CCa. . Do 3 2 17d 9 9 Int CCa • • Do 13 10 Total . . . 56 12 TABLE 79. Cross 18. — Intense guinea-pigs, each of which had a dilute parent known to transmit albinism, mated with albinos or red-eyea to test whether the same intense animal can transmit both dilution and albinism. Expectation: CCd X CraCra = CCra + CdCra (1 Int : 1 Dil). CCa X CraCra = CCra + CaCra (1 Int : 1 RE or W). 18a. Male CCd by 1 ,est. No. 9 Red-eye. cflnt. Int Dil RE W Remarks. I 515 SAg (R) ArF» D 10 R (Br) 35-1 3 3 3 Ag.SCrAg, 2 S3Y4Ag 2 774 SAg (R) ArF;. Do 1 Ag 3 515 774 SAg (R) ArF2 Do 3 3 3 Ag,2 S"Y4Ag, S4Cr5 4 514 Sep (R) ArF2 Do ... 4 3 Ag 4 B,2 S2-Cr6, S4- Y4 5 281 SAg (R) ArF2 D 30 R (Br) 36-1 3 8 3 Ag,S3Y4Ag, S^^g 6 741 SAg (R1 ArF2 Do 3 3 S6Cr5Ag, 2 S,Y4Ag 2 S4Cr6Ag, S6Cr6 Ag 3 Ag,2 S3Y4Ag, S6Y4 7 241 SAg (R1 ArF2 AA508 AgTb 3d-2 . . . 2 1 Ag 2 B,S3Cr6Ag 8 413 SAg (R) ArF2 Do 3 2 3 Ag,BiY4AgTb, S3Cr6 9 759 SAg (R) ArF2 Do 1 7 Ag Ag.SsY^g, S2Y4Ag 10 M430SAg(R) 18r-4. . . Do 2 4 S4Cr6Ag, S6Cr6 Ag 2B Total, 3 males 95 30 TABLES. 139 TABLE 79 — Continued. 186. Female CCa by tes t. No. 9 Int. cf Red-eye or W. Int Oil RE W Remarks. 1 M292 B ^2 D 18 W 166-3 2 3 AgTb, B.SeCreAgTb, 2 M353 B g3¥ Do 1 1 2Sep6 AgTb Sep 5 3 D6a R (Br) 35-1 724 SAg (R) lea .... 2 1 (Cross 20) Total, 3 females. . . 5 5 18c. Male CCa by t est. No. 9 Red-eye or W. cTInt. Int Dil RE W Remarks. 1 2 271 SAg (R) ArF2.. 278 SAg (R) ArF2 . . M224 BrAg 406-12 . Do 5 2 3 5 1 5Ag,2S4Ag(R),S7Ag (R) 2 Ag,5 SAg(R) 3 716 SAg (R) ArF... Do 4 3 4Ag,SBAg(R),SAg(R) 4 233 SAg (R) ArF. . . M156R(B) i^e 5 2 3 S4AgTb (R) 2 Ag, 3 B,2 SAg (R) 5 773 SAg (R) ArF2.. Do 1 6 7 236 Sep (R) ArF2.. 264 Sep (R) ArF2.. AA433a AgTb 36-4 . . . Do 1 ?: 2 1 1 AgTb,S8AgTb (R) Sep8 (R) 2 AgTb,Sep4 (R) 8 9 10 485 Sep (R) ArF2.. 235 Sep (R) ArF2.. 693 W ArF2 D7 R (Br) 35-1 . . . D13 R (Br) 35-1... Do 4 2 a 3 1 1 ? 4B,Sep6(R),2Sep(R) 2 Ag,Sep4 (R) A" B 11 D42 W 166-1 . Do 3 12 AA578 W 3c-18 . Do 3 2 Br, Ag 13 3 9 9 W ArF2.. M291 B 3*2 6 q 5B, Ag 14 4 9 9 W Misc. . M339 B 40a-14 11 13 11 B 15 3 9 9 W ArF2 M2 B T^ 10 7 2 Ag, 8 B Total, 8 males . . . . 57 ?0 41 140 INHERITANCE IN GUINEA-PIGS. TABLE 80. Cross 19. — Matings of dilute from cross 18o or 43, with albino. Expectation: CdCr X CaCa = CdCa + CrCa (1 Dil: 1 RE) (1-6). CdCa X CaCa = CdCa + CaCa (1 Dil : 1 W) (7-13). No. 9 Dil (or W). c?W (or Dil). Int Dil RE W Remarks. 1 D72S3Y4Ag 18a-5.. BW36 W BW .... 1 2 S4-Y4, SsAg (R), Sep3 (R) 2 D63 S4Y4Ag 43-2 . . . Do 1 ;-? Sep4, 2 SsAg (R), Sep4 (R) 3 D121 W 18c-9.. D71 S6Cr5Ag 18a-5 9 9 S3Cr6Ag, S4-Cr6, SsAgtR) 4 157 W 22-3 . . . Do . . . . 2 Sep3 (R) SsAg(R) Sepi(R) 5 D148 W lc-8... D240S2Y4Ag 18a-9. . 2 S4Cr5Ag, Sep4 fi D239W 18c-6.. Do 3 1 2 S6Cr6Ag, S6-Cr6, S4Ag 7 D61 S6Y4Ag 43-2 . . . BW36 W BW .... 1 (R) 8 D241 S6Cr5Ag 18a-9. . B W50 W B W . . . 1 1 SeY^g 9 (D69S4Cr5Ag 18a-5. . JBW36 W BW 3 1 2 S4Cr5Ag, Sep4 10 \M425 SYAg 43-1 . . . S772W ArF2... D70 S4Y4Ag 18a-5 1 S4Cr6Ag 11 256 W ArF2 . . . . . . . Do . . 3 3 3 S4Cr6Ag 12 S781W ArF2... Do 2 1 SaY^g, S4-Cr6 13 D69S4Cr6Ag 18a-5. . BW36 W BW .... 3 3 S4Y4Ag 14 D206S4-Y4 18a-4.. BW50 W BW .... 1 SepB 2 females CdCr. 2 5 2 males CdCr. . . 7 5 3 females CdCa . 2 3 1 male CdCa 6 4 2 females (?).... fi TABLE 81. Cross 20. — Matings of pure lea male 724 CrCr. Expectation: CC X CrCr = CCr (all Int) (1). CCd X CrCr = CCr + CdCr (1 Int : 1 Dil) (2-3). CdCd X CrCr = CdCr (all Dil) (4-5). CdCa X CrCr = CdCr + CrCa (1 Dil : 1 RE) (6). No. 9 Misc. cf Red-eye lea. Int Dil RE W Remarks. 1 5 9 9 B Trior4-toe 724 SAg(R) lea q 9 Ag 2 D6a R(Br) 35-1 Do 2 1 2 Ag, S2Y4Ag 3 D209 R(Br) 36-3 Do 1 Ag 4 AA606 S2Y3AgTb 40a-8 Do 9 2 S2Y4Ag 5 AA621 S3Y2Ag 39-4 . ... Do 3 3 S2Y4Ag 6 SA61 Sep4 32-2 Do . 1 1 BiCrgAg, SeAg(R) TABLE 82. Cross 21. — Matings of pure lea male 575 CCr. Expectation: CCr X CaCa = CCa + CrCa (1 Int : 1 RE). No. 9 W. cflnt lea. Int Dil RE W Remarks. 1 5 9 9 W BW .... 575 Ag lea. . 9 4 4 Ag, 5 B, 2 S4Ag(R), 2 Sep4(R) TABLES. 141 TABLE 83. Cross 22. — Matings of intense Fi lea (cross 21) with albinos. Expectation: CCa X CaCa = CCa + CaCa (1 Int : 1 W). No. 9 W (or Int. Fi lea. cflnt FI lea (or W). Int Oil RE W Remarks. 1 2 9 9 W BW 13 Ag 21-1. . . 4 3 4B 2 141 W ArF2 Do 4 3 M79 W 3*2 14 B 21-1 . . . 2 4 2B 4 D17 W 166-3 Do 2 5 19 Ag 21-1 d" W BW.... 2 2 2B 6 111 Ag 21-1 Do 4 4 Ag, 3B 7 17 B 21-1 . . . Do 4 6 4B Total 16 25 TABLE 84. Cross 23. — Matings of red-eye FI lea (cross 21) with albinos of intense stock. Expectation: CrCa X CaCa = CrCa + CaCa (1 RE : 1 W). No. 9 White. cf Red-eye FI lea. Int Dil RE W Remarks. 1 8 W BW . . . 15 Sep4(R) 21-1 3 3 Sep3(R) TABLE 85. Cross 24. — F2 from red-eye F! lea (cross 21). Expectation: CrCa X CrCa = CrCr + 2 CrCa + CaCa (3 RE : 1 W). No. 9 Red-eye FI lea. c? Red-eye FI lea. Int Dil RE W Remarks. 1 12 S4Ag(R) 21-1 15 Sep4(R) 21-1 . . 5 5 (So, S3, S5) Ag (R), S4(R), S5 2 16 S4Ag(R) 21-1 .Do 1? 1 (R) (B0, 3 S2, S3, S6) Ag (R) ; (B0, S2, 2S4, S6, S6)(R) TABLE 86. Cross 25. — Matings of F2 lea (cross 24) with albinos. Expectation: CrCr X CaCa = CrCa (all RE) (1-7). CrCa X CaCa = CrCa + CaCa (1 RE : 1 W) (8-9). No. 9 White. c? Red-eye F2 lea. Int Dil RE W Remarks. 1 D86 W 43-3 I26B0WAg(R) 24-2.. 5 S4Ag(R),S5Ag(R),S3 2 M431 W 18c-4 Do 6 (R), 2 S6(R) 3S4Ag(R),2S6Ag(R), 3 D76 D78 W 18c-14 137 B0(R) 24-2 5 Sep4(R) 2 S5(R), 3 S6(R) 3a D76 W 18c-14 . . Do 4 03, 84* 05, SG 4 D77 W 18c-14 . . Do 5 3S6(R), 2S6(R) 5 D125W 16c-4... 133 S2Ag(R) 24-1 . . ?, S6Ag(R), S6(R) 6 BW48 W BW Do 1 S6(R) 7 D427 W 44-3 Do 2 S6AgTb(R), S6(R) 8 D73 W 42-6 134 Sep4(R) 24-1 3 •? 3 S4(R) 9 S755 W ArF2 . . .Do 1 142 INHERITANCE IN GUINEA-PIGS. TABLE 87. Cross 26. — Matings of red-eye Fj lea (cross 21) with dilute. Expectation: CdCd X CrCa = CdCr + CdCa (all Dil). CdCa X CrCa = CdCr X CdCa + CrCa + CaCa (2 Dil : 1 RE : 1 W). No. 9 Dil (or Red-eye FI lea. c? Red -eye FI lea (or Dil). Int Dil RE W Remarks. 1 AA242 SsYsAgTb 40o-6 . . 15 Sep4(R) 21-1 3 BiCrBAgTb, Bt- 2 (AA244 Sep4 39-15 . . 1 .Do. 8 4 ft Cr6, Sep6 f5 B,, S4, SS-OB, S8, \ S3(R),2S4(R), 3 \M261 Sep4 41-2. . . I2S4Ag(R) 21-1. / M34 Sep6-Cr6 16c-l . 1 I S,(R) S3 4 16 S4Ag(R) 21-1 . . .Do 2 S7Cr5Ag, S7-Cr6 TABLE 88. Cross 27. — Matings of dilute from cross 26 with albinos. Expectation: CdCr X CaCa = CdCa + CrCa (1 Dil : 1 RE) (1-6). CdCa X CaCa = CdCa + CaCa (1 Dil : 1 W) (8-12). No. 9W (or Dil). d"Dil (or W). Int Dil RE W Remarks. 1 3 9 9 W Misc. . D89 Sepi 26-3 4 fi 2 S4, S6-Y4, S6-Cr8, 2 D221 W s"s . . . . D196 BI 26-2 1 2 S6(R), S«(R), 2 S7(R), Sg(R) S3 3 G30 W St . . . . . . Do . . 1 3 S5, S4(R),S5(R),S7 4 D115Bi-Cr5 26-1 BW50 BW 3 (R) 3 S3(R) 5 D197B, 26-2... JBW46 BW 3 2 3 S4, 2 S5(R) 6 7 D198 BI 26-2. . . BW43 W BW .... D113BiCr6AgTb 26-1 2 S4, S5-Cr6AgTb 8 482 W ArF2... D123 Si-Cr5 26-4 3 1 9 D42 W 16b-l D55 Ss-Ors 26-2 4 3 10 BW56, 57 BW D114Sep5 26-1 4 ?, 2 S4, 2 S4-Y4 11 D122 S7Cr6Ag 26—4 BW50 BW 2 1? D195 Sep4 26-2... Do 1 1 s« TABLE 89. Cross 28. — Matings of pure Arequipa male 1007 with black guinea-pigs. Expectation: CC X CdCd = CCd (all Int). CCa X CdCd = CCd + CdCa (1 Int : 1 Dil). No. 9 Intense. d" Dilute (Arequipa). Int Dil RE W Remarks. 1 2 9 9 B 4-toe ^007 SYAg 4 3 Ag, B 2 4 9 9 B BW . . . Do 10 5 Ag, 5 B 3 1442 B BW .... Do 3 2 2 Ag, B, SCrAg, Sep3 SYAg CdCd from 1001 BRAg CCd and 1002 SCrAg CdCr pure Arequipa stock. TABLES. 143 TABLE 90. Cross 29. — Matings of intense FI (cross 28) with d"1007. Expectation: CCd X CdCd = CCd + CdCd (1 Int : 1 Dil). No. 9 Intense ArFi. cf Dilute ( Arequipa) . Int Dil RE W Remarks. 1 SA4 Ag 28-3 ^007 SYAg 3 2Ag, R 2 SA8 Ag 28-2 . .Do . 2 2 AR, R, Bi-Y3, Ys (SA4, SA8 . . 1 _ 3 > .... Do . 3 9 2 Ag, B, BiYaAg, Y3 \SA10 B 28-2 i ^007 SYAg CdCd from 1001 BRAg CCd and 1002 SCrAg CdCr pure Arequipa stock. TABLE 91. Cross 30— Mating of dilute FI (cross 28) with c? 1007. Expectation: CdCa X CdCd + CdCd + CdCa (all Dil). No. 9 Dilute ArFi. 9 Intense (Arequipa) . Int Dil RE W Remarks. 1 SA3 Sep3 28-3 . . J1007 SYAg 9 BiY3Ag, Y3 '1007 SYAg CdCd from 1001 BRAg CCd and 1002 SCrAg CdCr pure Arequipa stock. TABLE 92. Cross 31. — F2 from intense FI Arequipa (cross 28). Expectation: CCd X CCd = CC + 2 CCd + CdCd (3 Int : 1 Dil). No. 9 Intense ArFi. c? Intense ArFi. Int Dil RE W Remarks. (SAG B 28-2 } 1 >SA2 B 28-2 17 3 17 B, 2 B-Y3, Br-Yi \SA11 B 28-2 2 SA13B 28-3 SA7 Ag 28-2 . . 1 B fSASAg 28-2. \ T^ 3 >. . . Do . 3 Ag, 2 B \SA10 B 28—2 . j 4 SA8 Ag 28-2 Do 1 2 Ag, B2Y3Af , B^-Y3 5 SA4 Ag 28-3 Do . . 1 ?! Ag, 2 BiYaAg Total .... 23 7 TABLE 93. Cross 32.— F2 from dilute FI X intense Fi (cross 28). Expectation: CdCa X CCd = CCd + CCa + CdCd + CdCa (2 Int : 2 Dil). No. 9 Dil ArFi. cflnt ArFi. Int Dil RE W Remarks. 1 SA3 Sep3 28-3 SA7 Ag 28-2 1 1 B, B2Y3Ag 2 Do SA12Ag 28-3 1 1 B, S4 144 INHERITANCE IN GUINEA-PIGS. TABLE 94. Cross S3. — Mating of dilute FI Arequipa with albino. Expectation: CdCa X CaCa = CdCa + CaCa (1 Dil : 1 W). No. 9 Dil ArFi. cf White. Int Dil RE W Remarks. 1 SA3 Sep3 28-3 75 W BW 1 TABLE 95. Cross 34. — Matings of intense FI Arequipa with albinos. Expectation: CCd X CaCa = CCa + CdCa (1 Int: 1 Dil). No. 9 Int ArFj (or W). c?W(orIntArFi). Int Dil RE W Remarks. 1 SA4 Ag 28-3 M313 W 42-16 2 2 Y4 2 SA10, 11,13 B 28-2,3. . Do 4 8 3 B, R, 2 Ss, 4 Sjr-Crs, 2 Y< 3 149 W 22-2 .... SA26 Ag 28-1.. . 3 2 2 Ag, B, S3Y4Ag, S3Cr5Ag 4 161 W 24-1 .... Do 3 1 2 Ag, B, S2 5 349 W ArF2 .... Do 1 1 Ag, 83 6 149, 349 W Do 2 6 Ag, B,S3 Cr6Ag, 2 S6Y4Ag, 2 SsOsAg, Ss-Os Total 13 ?n TABLE 96. Cross 35. — Mating of cream of dilute selection stock with a red stock free from albinism or dilution. Expectation: CC X CdCa = CCd + CCa (all Int). No. 9 Intense. cJ1 Dilute. Int Dil RE W Remarks. 1 49 9R(Bi) Misc 00 Cr6(Br) Dil 12 11 R(Br), Y2(Br) TABLE 97. Cross 36. — FI (cross 35) mated with father. Expectation: CCd X CdCa = CCd + CCa + CdCd + CdCa (2 Int : 2 Dil) (3). CCa X CdCa = CCd + CCa + CdCa + CaCa (2 Int : 1 Dil : 1 W) (1-2). No. 9 Intense FI. cf Dilute. Int Dil RE W Remarks. 1 ?, Dl R(Br) 35-1 D2 R(Br) 35-1 OOCr6(Br)Dil... Do 4 6 1 2 1 4 R(Br), 2 Y4(Br), 4 Cr6(Br) O6(Br) 3 D6 R(Br) 35-1 Do 3 4 3 R(Br), Y2(Br), 3 Cr5(Br) 4 D136R(Br) 36-1 Do ? 1 2R(Br),Cr6(Br) . TABLES. 145 TABLE 98. Cross 37. — Matings of dilute with dilute in the dilute-selection stock. Expectation: CdCd X CdCd = CdCd (all Dil). CdCd X CdCa = CdCd + CdCa (all Dil) (1). CdCd X CdCa = CdCd + 2 CdCa + CaCa (3 Dil : 1 W) (2-11). No. 9 Dilute. d* Dilute. Int Dil RE W Remarks. 1 D301 Y4 Dil . . D292 Y4 Dil 3 Y3, 2 Cr5 2 Do D298 Cr6 Dil 2 Y3, Crs 3 D300 Cr6 Dil Do 3 1 Y3, 2 Cr5 4 D299 Cr6 Dil Do 5 Y4, 3 Cr6, Cr6 5 D289Cr6 Dil D290 Cr$ Dil 6 1 Y3, 4 Cr6, Cr» 6 D291,Cr6 Dil . Do 2 7 /D260Y3(Br) Dil.... >D261 Cr6(Br) Dil. 4 1 Y3(Br),3Cr6(Br) S \D262, D263 Cr6(Br) Dil. ... D263 O6(Br) Dil.... . . .Do ... 3 2 Y3(Br),2Cr6Br 9 D262 Cr6(Br) Dil .Do 3 2 Y3(Br), Cr6(Br), 10 fD265Cr6(Br) 37-7.. > Do 4 Cr6(Br) Y3(Br),3Cr6(Br) 11 \D266 Cr6(Br) 37-7.. D265 37-7 Do 9 1 Y3(Br),Cr6(Br) Total (excluding 1) . . 32 10 TABLE 99. Cross 38. — Matings of dilute with albino in the dilute-selection stock. Expectation: CdCd X CdCa = CdCa (all Dil). CdCa X CaCa = CdCa + CaCa (1 Dil : 1 W). No. 9 White (or Dilute) . ^Dilute (or W). Int Dil RE W Remarks. 1 38a. D293 W Dil D292 Y4 Dil 4 4Cr5 2 D294 W Dil . . .Do 2 2CrB 3 D264 Y3(Br) 37-7 1 1 W Dil . 2 2 Cr6(Br) 2CdCd 8 1 3Sb, D293 W Dil D290 Cr6 Dil 2 2Cr5 2 D302 W Dil D298 Cr5 Dil 1 Cr5 3 D303 W Dil .Do. 2 2 Cr6 4 D 75 W Dil D267 Cr6(Br) 37-7 . . 4 5 Do D274 Cr6(Br) Dil ?, 1 2 Crs(Br) fD276 W Dil 6 >D261 Cre(Br) Dil 3 2 3 Crg(Br) 1 D272 W Dil 5 CdCa 10 7 146 INHERITANCE IN GUINEA-PIGS. TABLE 100. Cross 39. — All matings of intense with intense which have given dilute young, except those given in cross 31. Expectation: CCd X CCd = CC + 2 CCd + CdCd (3 Int : 1 Dil) (1-7). CCd X CCa = CC + CCd + CCa + CdCa (3 Int : 1 Dil) (8-33). 3, 9, 11, 24, 26, 30, and 32 not wholly certain. No. 9 Intense. cf Intense. ** Int Dil RE W Remarks. 1 M353 B e3} . . . . SA26 Ag 28-1 3 1 Ag, B, R, 82- Y2 2 A780 AgTb ^4 A781 AgTb ^ 7 2 4 AgTb, 3 R, BrYAg 3 B58 AgTb Id- 15 Bl 18 AgTb ld-6 9 2 Tb, Y 5 AgTb, 3 BrAgTb, 4 M25 AgLb 9-1 M91 AgLb 8-4 6 4 B, SsY^gTb, S2 3 Ag, 3AgTb, 2 SYAg 5 AA588 Ag 39-4 . . . .Do 2 1 Lb, 2 SjYjAgLb 2 Ag, S3Y3AgTb 6 ( M25, M27a Ag 9-1 } Do 3 1 IAg, 2 AgTh, BiYiAg 7 \B139Ag 39-23.. M177B lc-2 Do 4 1 [ Tb Ag, 3 AgTb, SYAgTb g M168 B 3^ Do 4 1 2 Ag, 2 AgTb, SCr6Ag 9 M169, M171 B 3^ .Do 3 3 3 AgTb, 2 BrCrAgTb, 10 B68 AgTb la-1 BUS AgTb ld-6 6 1 BrYAg 4 AgTb. 2 B, S3Cr6 11 A443 B i A469 AgTb $ 1 AgTb LBr 12 M90 Br i^g M 189 AgTb 39-30.. 3 BrCr6AgTb, SB, Cr6 /M90 Br I'B 3 AgTb, 3 B, S5Cr6 13 Do 6 2 14 \M114B lb-7 . . . A1117B 3lj 1357 B BW . . 2 1 AgTb, S6 AgTb, B, S4Cr6AgTb 15 A 1566 AgTb g1? A 1050 AgTb 3*2 4 9 3 AgTb, BrAgTh, SCr 16 Do AAlSAgTb SV 2 1 AgTb, S4 AgTb, B, SCrAgTb 17 A529 BrAgTb g^ Do 5 2 3 AgTb, B, BrAgTh, 18 AA202 AgTb 40f>-8 Do 2 2 SCrAgTb, BrCrAg Tb 2 AgTb, 2 SCrAgTb 19 Ml 77 B lc-2 AA235 AgTb 40b-7 4 1 AgTb, 2 B, BrAgTb, 20 M102 AgLb 6b-l M2 B A 4 1 LBr-Cr6 2 Ag, 2 B, S6Cr6Ag 21 3392 AgLb Misc . . A1539B ^ 1 SCrAg 22 3392, 3444 Ag Misc . . Do 3 1 2 Ag, B, SCrAg 23 3392 Ag Misc . . B5 AgTb ld-16 3 1 3 Ag, S4Y4Ag 24 20a Ag Misc . . A1474 AgTb fe. 3 1 3 Ag, SCrAg 25 A1310 Ag A A1449 AgTb ^5 . 2 1 AgTb, B, Cr 26 M203 AgTb 2-19 AA284 AgTb 39-18 4 1 3 AgTb, B, S 27 M82 Ag 9-7 A1161 AgTb sV- 1 1 AgTb, SCrAgTb 28 AA171 R 3*2 Do 2 1 AgTb, R, Cr 29 M183B yV AA299 AgTb 40-6 3 1 3 AgTb, SCrAgTb 30 20 B Misc A412 R(Br) ^g 3 1 2 AgTb, B, SCrAgTb 31 M7 B «V Ml 33 Ag 8-4 4 2 2 Ag, 2 B, SCrAg. S6 32 A1420B gV A811 Br gV • • • 1 1 Br, LBr 33 A385 B 3^2 12845 B 4-toe . . 4 1 4B, S Total 109 147 Excess of dilutes expected because the presence of at least one dilute young is used as a criterion for admission to the table. TABLES. 147 TABLE 101. Cross 40. — All matings of intense with dilute which have given dilute or albino young, except those of crosses 28, 29, 32 and 36. Expectation: CCd X CdCd = CCd + CdCd (1 Int : 1 Dil), 1. CCa X CdCd = CCd + CdCa (Int : 1 Dil), 2-5 (5?). CCd X CdCa = CCd + CCa + CdCd+CdCa (1 Int : 1 Dil), 6-19 (9, 10, 12, 14, 15, 16, 17?). No 9 Int (or Dil) . c^Dil (or Int). Int Di RE W Remarks. 1 40o. B139Ag 39-23.. M328B2-Y4 42-17. 2 3 2 AgTb, 3 82 Y2 2 AA606 S2YaAgTh 40a-8. . AA573 BrAgTb 40a-7 . 1 1 Ag A°Tb, S6Cr5Ag 3 4 9 9 B Misc. . . AA241 SYAgTh 40a-6. 5 3 Tb AgTb, 4 B, S« 4 169 B 4-toe... A656 Br-Y \ 1 3 Cr&AgTb. STb 41-6 AA253 S5O5AgTb 406-8 1 1 SY^gTb 10 3590 Y4(Br) Dil A674 Sep6 | ft 1 S,Y->ACr5AgTb, S6Cr5 AgTb, S6AgTb (R) 150 INHERITANCE IN GUINEA-PIGS. TABLE 105. Cross 44- — All matings of dilute with albino, except those of crosses 16, 19, 27, 33, and 38. Expectation: CdCd X CaCa = CdCa (all Dil) (1). CdCa X CaCa = CdCa X CaCa (1 Dil : 1 W) (2-6). No. 9 Dil (or W). d"W (or Dil). Int Dil RE W Remarks. 1 3256 BrYAgLb Misc. A678 W & q 7 SCreAg, 2 Br-Cr6 2 A505 Sep ]*e . . . . A868 W ^ 1 S 3 3 9 $ W Misc. . B117 S4Cr6AgTb 39-14 . . 5 1 S4, S6Cr6AgTb, Br 4 S176S-Cr ArF2.. 11 W Dil 2 3 Cr6AgTh, 2 S« 2Sj-Y4 5 S263 SCr6Ag ArF2 . . Do . ... 1 S4Cr6Ag 6 O7 W Misc. A674 Sepg | ?, S7Cr7Ag, Cr7 TABLE 106. Cross 45. — Rough A (4-toe) X rough A (4-toe). Rrss X Rrss = 3 Rss + rrss (3 A : 1 Sm). No. 9 Rough A. cf Rough A. A B C D E Sm 1 3769 4-toe 3609 4-toe 1 2 96 4-toe ... . . . .Do ?, 3 /3769 4-toe > Do 5 1 4 \3770 4-toe 3769 4-toe 3987 4-toe 2 5 3770 4-toe Do 2 Total 10 3 TABLE 107. Cross 46. — Rough A (tricolor) X rough A (tricolor) ; one or both of parents of each, rough C or D. Rrss X Rrss = 3 Rss + rrss (3 A : 1 Sin) (1-8). or Rrss X RRss = Rss (all A) (9-12?). No. 9 Rough A. c? Rough A. A B C D E Sm 1 4018 Tri 3775 Tri 1 2 3941 Tri 3940 Tri 3 3 3943 Tri . .Do 1 4 /R65 54-17 1 Do 7 1 \3943 Tri / 5 R65 54-17 Do 2 2 6 R171 47-3 Do 2 2 7 R278 52-14 R248 52-10 1 1 8 R357 Red 52-8 Do 1 1 1 9 R65 54-17 R197 52-14 2 10 R171 47-3 Do 3 11 R194 54-1 Do 2 12 R196 52-14 Do 3 Total 1 to 8 17 0 7 Total 9 to 12 10 TABLES. 151 TABLE 108. Cross 47. — Rough A X rough C (tri); all mothers of tricolor stock except R175— 4-toe. Rrss X RrSs = 3 Rss + 3 RSs + 2rr (3 A : 3 C : 2 Sm). No. 9 Rough A. c? Rough C. A B C D E Sm 1 R21 46-2. . . . R52 56-1 1 0 2 R23 46-2 .... Do 1 1 3 R42 50-1 .... . . Do 1 1 1 4 R21 46-2 . . . R99 56-1 . . . 1 2 5 R23 46-2 . Do 1 1 6 R42 50-1 .... Do 3 1 2 4 7 R175 49-1 .... Do 3 1 Total 10 1 5 1 10 TABLE 109. Cross 48.— Rough A (Tri) X rough E (Tri). Rrss X RRSS = RSs (all C). No. 9 Rough A. d" Rough E. A B C D E Sm 1 R42 50-1 4003 Tri 9 1 TABLE 110. Cross 49. — Rough A (4-toe) X smooth (4-toe). Rrss X rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Smooth. cf Rough A. A B C D E Sm 1 499 4-toe 3922 4-toe. 18 13 2 599 4-toe 3609 4-toe 10 1 19 Total .... 28 1 32 TABLE 111. Cross 50. — Rough A, B (tri) X smooth (4-toe etc.). Rough A, B with one or both of parents partial rough. Expectation as in cross 49. No. 9 Smooth (or rough B). c? Rough A (or smooth). A B C D E Sm 1 799 Sm 4-toe 3775 A Tri 13 14 2 R121 Sm 50-1 Do 3 3 R62 Sm 50-1 R22 A 46-2 3 3 4 ^1636 52-13... 99 Sm 4-toe .... 3 Total . . 19 20 may he RR. 152 INHERITANCE IN GUINEA-PIGS. TABLE 112. Cross 51 . — Rough A X smooth (tri) ; smooth with one or both parents partial rough. Rrss X rrSS = RrSs + rrSs (1 C : 1 Sm). Rrss X rrSs = Rrss+ RrSs + 2 rr (1 A : 1 C : 2 Sm). Rrss X rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Smooth. c? Rough A. A B C D E Sm 1 R13 52-1 R22 46-2 . . . 1 3 4 2 R123 54-3 R76 4-toe 2 4 3 R124 54-3 Do 1 2 4 R133 47-2 Do 2 5 R124 R133 Do 2 8 6 R142 54-3 Do 1 1 Total 6 3 3 19 TABLE 113. Cross 52.— Rough1 C, D (tri) X rough C (tri). RrSs X RrSs = 3 Rss + 6 RSs + 3 RSS + 4 rr (3 A : 6 C : 3 E : 4 Sm). No. 9 Rough C, D. cfRough C. A B C C E Sm Remarks. 1 3013 Tri .... 3780 Tri 1 3 1 1 2 2 3246 Tri Do 1 2 a Rll 52-1 .. .Do 1 2 3 4 R54 52-1 . Do 1 1 z 324P> Tri 4019 Tri 2 1 p. 3939 Tri R58 52-5 4 1 1 Red-A 7 3809 Tri Do 2 1 Red-E, Red-Sm g 3246 Tri . . . Do 2 1 1 Red-A Q 3724 D Tri Do 1 1 10 3724 3246 Tri Do 1 4 1 11 R51 56-1 R56 56-1 2 1 2 2 1 3 12 R57 56-1 Do . 1 1 1Q R51 R57 56-1 Do 1 2 14 R98 56-1 Do 5 2 3 1 15 R103 56-1 Do 1 1 5 1 1 Red-C Total 18 6 19 7 12 17 1AII rough C except 3724. TABLE 114. Cross 53— Rough C, D (tri) X rough E (tri). RrSs X RRSS = RSs + RSS (1 C : 1 E). or RRSsX RrSS = RSs + RSS (1 C : 1 E). or RrSa X RrSS = 3 RSs + 3 RSS + 2 rr (3 C :3 E : 2 Sm). No. 9 Rough C, D. c? Rough E. A B C D E Sm 1 R6 D 54-15 4003 Tri 1 2 2 R88 C 52-1 R200 Red 52-7 2 3 R286 C 52-6 R280 52-14 .... 3 4 R222 C 52-12 Do 1 2 1 Total 4 1 6 1 TABLES. 153 TABLE 115. Cross 54. — Rough C, D (tri) X smooth (4-toe, etc.). RrSs X rrss = Rrss + RrSs + 2 rr (1 A : 1 C : 2 Sm). No. 9 Smooth (or rough C,D). c? Rough C,D (or smooth). A B C D E Sm 1 399 Sm 4-toe 3780 C Tri 1 1 8 2 R62 Sm 50-1 .... Do 2 ?, 3 599 Sm 4-toe R12 C 52-1 4 12 17 4 399 Sm 4-toe .... R26 D 54-15 8 3 5 14 5 499 Sm 4-toe R52 C 56-1 2 2 6 6 599 Sm 4-toe R102 C 56-1 4 5 3 9 7 B31 Sm ld-9 R105 C 48-1 2 8 M253 Sm la-10 R106 C 48-1 2 9 M255 Sm la-10 Do 1 10 M380 Sm g\ Do 1 11 131 M253 M255 Sm Do 1 3 12 M384 Sep-Sm lb-9 R99 C 56-1 ?, ?, 13 R62 Sm 50-1 . . R112 C 52-11 . . 3 14 65 Sm 4-toe Do . 1 1 1 15 3246 C Tri 2967 Sm BB 5 1 2 3 16 3809 C Tri Do 1 3 17 3724 D Tri Do 3 1 1 8 Total 34 ?9 13 1 79 TABLE 116. Cross 55.— Rough E (tri) X rough E (tri). RrSS X RrSS = 3 RSS + rrSS (3 E : 1 Sm). No. 9 Rough E. c? Rough E. A B C D E Sm 1 R221 52-12 R140 52-3 2 3 2 JR201 Sm 52-7 Do 2 Total 4 3 'See note, cross 57. TABLE 117. Cross 56. — Rough E (tri) X smooth (4-toe). RRSS X rrss = RrSs (all C) 1. RrSS X rrss = RrSs + rrSs (1 C : 1 Sm) 2-5. No. 9 Smooth (or rough E). c? Rough E (or smooth) . A B C D E Sm 1 399 Sm 4-toe 4003 E Tri 11 2 1 9 Sm 4-toe R140 E 52-3 1 3 3 B31 Sm la-1 . . Do 2 4 !R201 Sm 52-7 13 W-Sm 4-toe . . 1 1 5 R221 E 52-12 Do . . . ?, Total (1) 11 Total (2-5) 2 8 JSee note, cross 57. 154 INHERITANCE IN GUINEA-PIGS. TABLE 118. Cross 67. — Smooth (tri) X smooth (4-toe, etc.); both parents of tricolor smooths were partial roughs (cross 52). rr X rr = rr (all Sm). If, however, RSS, normally rough E, is ever Sm: RrSS X rrss = RrSs + rrSs (1 C : 1 Sm). No. 9 Smooth. cf Smooth. A B C D E Sm 1 699 R131 52-4 14 2 R139 52-3 99 4-toe 2 3 R13 52-1 . . . . Do 6 4 R164 52-13 Do 2 5 R199 Red 52-7 13 W 4-toe 3 6 R249 52-10 Do 3 7 R263 52-11 Do 2 8 JR201 52-7 . .Do 1 1 'R201 was called rough E? at birth with the note that there seemed to be a trace of roughness on one hind toe. No roughness was apparent when adult and she was called Sm, but nevertheless was tested by mating with a 4-toe smooth. The result shows that she was, genetically at least, like a rough E. TABLE 119. Cross 58. — Rough B, C (Lima) X rough B (Lima). No. 9 Rough B. cf Rough B. A B C D E Sm 1 L7 Lima L5 Lima 2 2 1 2 L97 60-6 L26 58-1 1 3 L140 60-7 Do 1 1 4 f L97 60-6 .... JL98 60-6 3 2 2 5 \L81 Red 59-3 .... L99 Rough C 61-1 Do 1 1 6 LSI L97 L99 (above) Do 1 1 2 4 Total 8 7 ?, 1 7 TABLE 120. Cross 59. — Rough A (Lima) X smooth (Lima). No. 9 Smooth (or rough A). cf Rough A (or smooth). A E C D E Sm Remarks. 1 L13 Sm 62-2 L9 A 58-1 2 1 2 L14 Sm 60-1 Do 7 1 2 3 L24 Sm 60-2 Do 2 1 3 5 Red-B 4 L25 Sm 58-1 Do ... 4 1 3 2 Red-A, 2 5 L37 Sm 62-3 . Do 1 Red-Sm 6 L57 Sm 59-3 Do 1 ?, 0, Red-A 7 L24, L57 Sm (above) Do 1 ? 1 Red-B 8 L22 A 60-2 LI Sm Lima 4 1 6 2 Sep(p)-Sm 9 L62 A 59-3 Do 1 2 10 L100Sep(p)-A 61-1 L82 Sep(p)-Sm 59-8 2 2 Sep(p)-A Total ?4 8 3 ?,3 TABLES. 155 TABLE 121. Cross 60. — Rough B (Lima) X smooth (Lima) . No. 9 Smooth. 724 SAg(R) lea 3 2 \R252 49-2 R236 50-3 Do 9, 1 3 R205 46-4 . . Do 1 4 R213 49-1 Do ?, Total 5 4 TABLE 129. Cross 68. — Rough A X smooth (pure C. cutleri). Rrss X rrSS = RrSs + rrSs (1 C : 1 Sm). Young all light-bellied agouti. No. 9 Rough A. cf Smooth. A B C D E Sm 1 3986 4-toe C128 Ag 1 ? 2 3988 4-toe . . . . . . . Do 9; 1 3 3986, 3988 4-toe . . . Do ?, 3 4 A1691 Ag 65-5 ... Do . 1 2 3 5 AA567 Ag 66-3 Do . 1 3 6 f AA568 66-3 . . . \ Do 3 \R80 54-15.. / Total 3 6 15 158 INHERITANCE IN GUINEA-PIGS. TABLE 130. Cross 69. — Rough C (tricolor) X smooth (pure C. cutleri). RrSs X rrSS = RrSs + RrSS + 2 rr (1 C : 1 E : 2 Sm). Young all light-bellied agouti. No. 9 Black rough C. D. cf Smooth. A B C D E Sm 1 3245 Tri C128Ag 2 2 R54 52-1 .Do 1 1 1 3 R152 54-4 Do . . 1 ?, 4 3245 R110 Do 2 1 5 R110 51-1... Do 2 1 6 R170 47-3 Do 1 ?, 7 Rll 52-1. . . . . Do 1 8 R101 D 52-5 . .Do 3 g R154 D 54-4 Do 2 Total 1 1 9 13 TABLE 131. Cross 70. — Rough A (guinea-pig) X rough C, D (|, £ cutleri). Rres X RrSs = Rss + RSs + 2rr (3 A : 3 C : 2 Sm). All \ cutleri Rough C, D, except K58, J blood. R116 and R137 may be RRss. No. 9RoughA(orC,D). cf" Rough D(orA). A B C D E Sm Remarks. 1 R116B-A 46-4 K54 Ag-D 68-1 2 1 Ag-A, B-A, Ag-C 2 R117B-A 46-4 Do 1 1 B-D, B-Sm 3 R137 B-A 47-1 Do 9 2 B-C 4 AA608 B-A 66-3 Do 3 1 Ag-A, 2 B-A, Ag-E K K12 Ag-D 68-3 R31 B-A 45-3 2 1 2 B-C, B-Sm 6 K14 Ag D 68 3 . . Do 9 2 Ag-B 7 Do 3609 B-A 4-toe . 1 1 1 B-A, Ag-B, Ag-E 8 Do R76 B-A 45-4 1 1 B-A, B-D 9 K12 Ag-D 68-3 Do 1 9 1 Ag-A, 2 Ag-C, B-Sm 10 K58 B-C 70-5 Do 1 1 B-C, B-Sm Total 8 3 8 2 2 4 TABLE 132. Cross 71. — Rough A (guinea-pig) X smooth (£, i cutleri). Rrss X rrSs = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). All \ cutleri except K79, 5 cutleri. No. 9 Smooth. 6" Rough A. A B C D E Sm Remarks. 1 K7 \g 77 1 R31 B 45-3 s 2 1 7 Ag-A, 2 B-A, 2 B-B, B-C, 6 2 K15 Ag 68-3 Do 3 ?, 9 Ag-Sm, B-Sm Ag-A, 2 B-A, Ag-C, B-C, 3 3 K55 Ag 68-1 Do 2 1 Ag-Sm, 6 B-Sm 2 Ag-C, B-Sm 4 K7 K55 (above) Do 9 3 2 Ag-A, 2 Ag-Sm, B-Sm 5 K68 Ag 77-1 Do 1 1 2 1 B-A, B-B, Ag-D, 2 B-C 6 K116 Ag 68-6 .... Do 1 B-Sm 7 K81 Ag 69-1 3609 B 4-toe 1 1 Ag-C, Ag-Sm 8 K79 B 78-1 3922 B 4-toe 1 1 B-A, B-C Total 10 3 9 1 22 TABLES. 159 TABLE 133. Cross 72. — Smooth (guinea-pig) X rough C, D (|, \ cutleri). rrss X RrSs = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). K71a, K92 may be RR. No. 9RoughC,D(orSm). cf Smooth (or rough C, D). A E C D E Sm Remarks. 1 K12Ag-D 68-3... OOCr(Br)Sm Dil 1 o, Ag-A, 2 B-D 2 K14 Ag-D 68-3 . . .Do .... 3 6 2 Ag-A, B-A, 3 3 /K12 Ag-D 68-3 .Do . 1 3 Ag-Sm, 3 B- Sm /B-C, 2 Ag-Sm, 4 \A71a Ag-C 70-1 . . . K92 Ag-C 70-9 . . . / Do 1 2 1 B-Sm Ag-A, 2 Ag-C 5 K157Ag-C 72-12 . 13 Cr(Br)Sm Dil... 1 1 Ag— A, Sep-Sm 6 K147 AgTb-D 72-13 . .Do 3 2 B-Sm,Sep-Sm 7 K142 B-D 72-1 00 Cr(Br)Sm Dil 1 1 1 Red-A, W-D, 8 D40Cr(Br)Sm 36-1... K60 Ag-C 71-2 .. 1 1 B-Sm B-A, Ag-Sm 9 R173 B-Sm 49-1 . . . Do 1 1 Ag— A, Ag— Sm 10 20 B-Sm 4-toe . . K54 Ag-D 68-1 . . 2 1 9 Ag-A, B-A, Ag- 11 AA533 Ag-Sm 66-2 . . . K56 B-C 70-5 1 C, 2 B-Sm Ag— Sm 12 AA586 Ag-Sm 106-10 . . . .Do 1 1 Ag-C, B-D 13 /M382 AgTb-Sm 16-9. . . ... Do . 1 1 2 1 /B-A, 3 AgTb- 14 \B239 AgTb-Sm la-5 . . . B239 AgTb-Sm la-5 .Do 1 \ C+D,B-Sm B-Sm Total ]?, fi fi ?1 TABLE 134. Cross 73. — Smooth (4-toe) X rough A, B (\ cutleri). rrss X Rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Rough A, B. c? Smooth. A B C D E Sm Remarks. 1 K95 B-B 71-1 99 B 4-toe 1 B-Sm /K101 B-A 70-8 2 >13 W 4-toe 2 I 2 B-A, B-Sm \K106 B-A 71-2 Total 2 9 TABLE 135. C?-oss 74- — Smooth (|, J cutleri) X rough A (i cutleri). rrSs X Rrss = Rrss + RrSs + 2rr (1 A : 1 C : 2 Sm). rrss X Rrss = Rrss + rrss (1 A : 1 Sm). No. 9 Smooth. c? Rough A. A B C D E Sm Remarks. 1 K68 Ag 77-1 K59 B 71-2 1 B-C 2 K42 B 78-2 Do 1 B-Sm Total 1 1 160 INHERITANCE IN GUINEA-PIGS. TABLE 136. Cross 75. — Rough A, B (J cutleri) X rough A (| cutleri). Rrss X Rrss = 3 Rss + rrss (3 A : 1 Sm). No. 9 Rough A, B. c? Rough A. A B C D E Sm Remarks. 1 K50 Ag-B 70-6 K59 B 71-2 1 B-A 2 K53 B 71-1 Do 9 2B-A Total 3 TABLE 137. Cross 76. — Rough C (f cutleri) X rough C (i cutleri). RSs X RSs = Rss + 2 RSs + RSS (1 A : 2 C : 1 E : ? Sm ?). No. 9 Rough C. cf Rough C. A B C D E Sm '- ; Remarks. 1 K114B 74-1 K93 Ag 70-9 . . 1 3 1 (Ag-C, Ag-D, 2 B-D, I Ag-E TABLE 138. Cross 77. — Black (BB) X agouti (pure cutleri). Parents and offspring all smooth. No. 9 Black. cT Agouti. AgLb Black W 1 2 B 9 9 BB C128 Ag pure C 6 TABLE 139. Cross 78— Black (BW) X agouti XJ, 1, i cutleri). Expectation : 1 Ag : 1 black : some white. Parents and offspring all smooth. No. 9 Black (or agouti). , ' c? Agouti (or black). AgLb Black W 1 6 B 9 9 BW C67 Ag (J) 12 12 6 2 2 B 9 9 BW K4 Ae (}) 78-1 6 9 3 3 B 9 9 BW K24 Ag (|) 78-1 9 5 4 K20 Ag (!) 78-1 39 B BW 3 5 K22 Ae (i) 78-1 Do 1 1 6 K29 Ag (}) 78-1 . . .Do 3 3 K103 Ae (i) 78-4 Do 2 8 K66 Ag (|) 78-2 . Do 4 9 K109 Ag (|) 78-6 Do 1 1 10 3 B 9 9 BW K104 Ag (|) 78-4 5 3 Total 35 41 10 PART III FURTHER STUDIES OF PIEBALD RATS AND SELECTION, WITH OBSERVATIONS ON GAMETIC COUPLING BY W. E. CASTLE THE PROGENY OF HOODED RATS TWICE CROSSED WITH WILD RATS. In 1914 Castle and Phillips published a report on breeding experi- ments with hooded rats, in which it was shown that the hooded color pattern — itself a Mendelian recessive character in crosses with the entirely colored (or "self") coat of wild rats — is subject to quantitative variation, and that different quantitative conditions of the hooded pattern are heritable. (Compare fig. 36, plate 7.) It was also shown that by repeated selection of the more extreme variations in the hooded pattern (either plus or minus) it is possible gradually to modify the racial mean, mode, and range as regards these fluctuations, without eliminat- ing further fluctuation or greatly reducing its amount. We concluded that the unit character, hooded color pattern, is a quantitatively vary- ing one, but were at that time unable to decide whether the observed variability was due simply and exclusively to variation in a single Mendelian unit factor or partly to independent and subsidiary modify- ing Mendelian factors. Since publication of the above I have been engaged in further experi- ments designed to show which of the alternative explanations is the correct one, and these are now sufficiently advanced to indicate definite conclusions. Previous experiments had shown that when a race of hooded rats, whose character has been modified by selection (either plus or minus), is crossed with wild rats, the extracted hooded animals obtained in F2 as recessives show regression toward the mean condition of the recessive race before selection began. This result suggested that the regression observed might be due to removal by the cross of modifying factors, which selection had accumulated in the hooded race. If this view was correct, it was thought that further crossing of the extracted hooded animals with the same wild race should result in further regression, and that if this further regression was not observed a different explanation must be sought for the regression already noted. The entire experiment has accordingly been repeated from the beginning, with the same result as regards regression in the first F2 generation, but with no regression of the same sort in a second F2 con- taining twice-extracted hooded animals. So far from observing further regression as a result of the second cross with wild rats, we have unmis- takable evidence that the movement of the mean, mode, and range of the hooded character has been in the reverse direction. So the hypothe- sis of modifying factors to account for the regression and for the pro- gressive changes observed under selection becomes untenable. In repeating the experiment of crossing hooded rats of our selected races with wild rats, great care has been taken to employ as parents individuals of the greatest racial purity and to inbreed the offspring 163 164 INHERITANCE IN RATS. brother with sister, thus precluding the possibility of introducing modifying factors from other sources. In making the second set of crosses, the extracted individual has, wherever possible, been crossed with its own wild grandparent. In the few cases in which this was impossible, wild animals of the same stock have been used. This stock consisted of a colony of wild rats which invaded the basement of the Bussey Institution apparently from a nearby stable. Owing to faulty construction of the building they were able to breed in spots inaccessible to us, and it took many months of continuous and persistent trapping to secure their extermination. During this period we trapped a hun- dred or more of them, all typical Norway rats, colored all over, without even the white spot occasionally seen on the chest of wild rats. Two generations of rats from this wild stock have been reared in the labora- tory, and all have this same self-colored condition. The hooded animals used in the experiments to be reported on in this connection consisted of 4 individuals of the plus-selected series, a male and 3 females, as follows : TABLE 140. Individual. Grade.1 Generation. 95513.... +4i 10 c?G348 +4 10 9 6600 .... +4£ 12 96955 +4 12 'See figure 35, plate 7, for significance of the grades. Each of these animals was mated with a single wild mate, and their children were weaned directly into breeding cages containing a male and two or three females (brother and sisters). In the case of two matings, F! males of the same parentage were at the time lacking and males from a different cross were used. The results of such matings are tabulated by themselves and serve a useful purpose as controls. The F! animals all closely resembled their wild parents, but many of them had a white spot on the chest. They ranged from grade +5| to +6 (self). The F2 animals are classified in table 141, where it appears that 73 of them were hooded and 219 non-hooded (i. e., like FJ, an exact 1 : 3 ratio. More than half of this F2 generation consists of the grand- children of 9 5513, produced by breeding her children brother with sister, those children all having been sired by the same wild rat. Her grandchildren include 41 hooded and 107 non-hooded young. The hooded young range in grade from + lf to +4, their mean grade being +3.05, a considerable regression from the grade of the grandmother, which was 4.25. Hooded rats of the same grade and generation as the grandmother, when bred with each other, produced young of mean grade +3.84. HOODED CROSSED WITH WILD. 165 (See table 10, Castle and Phillips.) The mean of the extracted hooded grandchildren in this case (being 3.05) shows a regression of 0.79 from that expected for the uncrossed hooded race. From the extracted hooded grandchildren of 9 5513, produced as just described by a cross with a wild male, 7 individuals, 2 males and 5 females, were selected for a second cross with the wild race. They ranged in grade from +2 to +3|. (See table 142.) They produced several litters of young of the same character as the first FI young, all being similar to wild rats in appearance, except for the frequent occurrence of a white spot on the belly. These second FI young were at weaning time mated, brother with sister, in breeding-pens, precisely as had been done with the first F/s. They produced 394 second F2 young, of which 98 were hooded and 296 non-hooded, a perfect 1 : 3 ratio. The hooded young varied in grade from +2 to +4, as shown in table 142, the data there being given for each family separately as well as for all combined in the totals. One family was very like another as regards the character of the hooded young, except that the higher-grade grandparents had grandchildren of slightly higher grade. Thus the average of all the 98 hooded young was +3.47, but the average of those descended from the 3 grandparents of lowest grade was less than this, while the average of those descended from the 3 grandparents of highest grade was greater. This is just what had been observed throughout the entire selection experiments. (See Castle and Phillips.) If we weight each of the grandparents in table 142 in proportion to the number of its hooded grandchildren, then the mean grade of all the grandparents is +2.95. Since the mean grade of all the 41 first F2 hooded grandchildren, from which these 7 were chosen, was +3.05, it will be seen that these 7 are, so far as grade is concerned, fair repre- sentatives of the 41, being in fact of slightly lower mean grade. It is therefore all the more striking that their grandchildren, the second F2 hooded young (table 142) , are of higher grade. They regress in an oppo- site direction to that taken by the first F2 hooded young. Thus the original hooded ancestor ( 9 5513) was of grade 4.25. The grade of hooded young expected from such animals is 3.84. What she produced in F2, following a cross with the wild male, was young of mean grade 3.05. Seven of these of mean grade 2.95 produced a second F2 contain- ing hooded young of mean grade 3.47. This is a reversed regression of 0.52 on the grade of their actual hooded grandparents, or of 0.42 on the group from which their grandparents were chosen. Their mean lies about midway1 between that which would have been expected from the original hooded female (5513) had no crossing with wild rats occurred and that which was observed in the first F2. 'In The Scientific Monthly (Jan. 1916) I have stated that a second cross showed "a return to about what the selected race would have been had no crossing at all occurred." This is obviously inaccurate and should be corrected. It rests on a comparison with the combined average of both the older and the more recent experiments. 166 INHERITANCE IN RATS. Obviously these facts do not harmonize with the assumption that the regression observed in the first F2 was due to loss of modifying fac- tors accumulated during the ten preceding generations of selection; for no further loss occurs in the second F2. On the other hand, a partial recovery is made of what was lost in the first F2. This suggests the idea that that loss may have been due to physiological causes non- genetic in character, such as produce increased size in racial crosses; for among guinea-pigs (as among certain plants) it has been found that Fx has an increased size due to vigor produced by crossing and not due to heredity at all. This increased size persists partially in F2, but for the most part is not in evidence beyond F\. I would not suggest that the present case is parallel with this, but it seems quite possible that similar non-genetic agencies are concerned in the striking regression of the first F2 and the subsequent reversed regression in the second F2. Whatever its correct explanation may be, the fact of the reversed regression in a second F2 is very clear, as other cases than those already discussed will show. A hooded rat of grade +4 and generation 10, c?6348, had by a wild female several young of the character already described for the young of 9 5513. These, mated brother with sister, produced a first F2 (table 141) of 90 rats, 22 of which were hooded, 68 being non-hooded, again a good 1 : 3 ratio. The hooded young ranged from +2 to +4 in grade, their mean being 3.28. Of the 22 hooded individuals, 1 male and 7 females were mated with wild rats to obtain a second F1; and the second F! animals were then mated brother with sister to obtain the desired second F2. The character of this is shown family by family in table 143. It contained 497 individuals, of which 121 were hooded and 376 non-hooded, a ratio of 1 : 3.1. The weighted mean of the 8 selected grandparents is 2.93, which is 0.35 below the mean of the 22 first F2 hooded animals which they represent. The mean of the second F2 hooded young is 3.22, which indicates a reversed regression of 0.29 on the grade of the grandparents, but shows no significant difference from the mean of the grandparental group (3.28). All except one of the 8 families classified in table 143 show unmis- takably the reversed regression. This exceptional family consists of the grandchildren of 9 9747. They have a mean grade of 2.90, sub- stantially the same as that of the entire group of grandparents but con- siderably lower than that of their own hooded grandmother. Appa- rently she did not come up genetically to her phenotypic grade. This the other grandparents of the group did. For those of lowest grade (2, 2f ) produced lower-grade hooded grandchildren than did the grand- parents of highest grade (3^, 4), as was found to be the case also in table 142. We may next trace the inheritance of the hooded character through a third but smaller family produced by two successive crosses with wild HOODED CROSSED WITH WILD. 167 rats, the hooded character in this case being derived from 96955, grade +4, generation 12. The character of her first F2 descendants is shown in table 141. They consist of 5 hooded and 27 non-hooded individuals. The mean grade of the hooded young is 3.51, but the number of these young is too small to make this mean of much signifi- cance. One of the hooded young (cf 9660, +3f) was mated with a wild female to secure a second Fx generation and from this in due course was produced the second F2 generation (table 144) . It consisted of 2 1 hooded and 44 non-hooded young. The hooded young showed the usual range (2 to 4). Their mean grade was 3.50, substantially identical with that of the first F2 animals, but 0.25 below that of the actual hooded grand- parent. This family history is less satisfactory than the two alreadjr discussed because of the smaller numbers which it includes. It con- tains nothing contradictory to the interpretation already given, though reversed regression is not in this case in evidence. In two cases FI females could not be mated with brothers and so mates were taken from other families. Thus "mixed F! matings" were made between children of 5513 and 6600 and children of 5513 and 6955. (See table 141.) The former mating produced 3 hooded and 12 non-hooded "first" F2 young; the latter produced 2 hooded and 5 non-hooded "first" F2 young. The grade of the hooded young pro- duced by these mixed matings was not different from that of brother- sister matings, so far as the small numbers permit one to judge. One of these mixed matings wras carried into a second F2 generation. The first F2 hooded cf 9711, +3|, was mated with a wild female, and the young were bred, brother with sister, producing 16 hooded and 33 non- hooded young. (See table 144.) The mean grade of the 16 hooded young was 3.28, nearly the same as that of the first F2 hooded grand- parent. No additional regression through loss of modifiers (or other agency) is here in evidence. The result is the same as that observed in families wholly unmixed. The attention of my pure-line critics, who think that in our mass-selection experiments insufficient attention has been given to individual pedigrees, is particularly directed to the foregoing case. Having now discussed each family history separately, we may com- bine all the second F2 families in one table, in order to get a clearer impression of the results as a whole. (See table 145.) The second F2 generation thus combined includes 256 hooded and 749 non-hooded individuals, a ratio of 1 : 2.9, an unmistakable mono-hybrid Mendelian ratio. The mean grade of the hooded individuals is 3.34. The weighted mean grade of their hooded grandparents was 3.02, which indicates a reversed regression of 0.32 for the entire second F2 group of hooded animals. Classified according to the grade of the (first F2) grandparent, they show a correlation between grade of grandparent and grade of grand- 168 INHERITANCE IN RATS. child. The lower-grade grandparent has lower-grade hooded grand- children, and the higher-grade grandparent has higher-grade hooded grandchildren. This shows that the variation in grade is (in part at least) genotypic. As the experiment yields no evidence that the varia- tion in the hooded character is due to independent modifying factors, there remains no alternative to the conclusion that the single genetic Mendelian factor concerned fluctuates in genetic value. Fluctuation accordingly is not exclusively phenotypic, as DeVries and Johannsen have thought, but may be genetic also. Hence racial changes may be effected through selection by the isolation of genetic fluctuations, as well as by the isolation of mutations. Moreover, genetic fluctuation makes possible progressive change in a particular direction, repeated selection attaining results which it would be quite hopeless to seek by any other means. A SECOND REPORT ON MASS SELECTION OF THE HOODED PATTERN OF RATS. The experiments in selection for the modification of the hooded pat- tern of rats, when reported on by Castle and Phillips in 1914, had been carried through 13 generations. Since then the experiments with the same selected races have been carried through 3 or 4 additional genera- tions, the results of which will now be described. Additional records have also been obtained for certain of the generations reported on by Castle and Phillips, which may now be combined with those previously pub- lished. Thus, revised data, based on larger totals, may be given for generations 12 and 13 of the plus-selection series and for generation 13 of the minus-selection series. These do not materially change the results previously obtained, but add to their trustworthiness. The additional generations of selection show a continued progressive movement of the racial character in the direction of the selection and indicate the exist- ence of no natural limit to the progress which selection can make in changing the hooded character. For details concerning the earlier history of the experiments and the methods of grading the animals the reader is referred to the publica- tion of Castle and Phillips. The grading scale (exclusive of the newer and more extreme grades) is reproduced in figure 35, plate 7. Atten- tion may be called to the fact that the entire selection series, both plus and minus, consist of animals descended from an original stock of less than a dozen individuals. These descendants number more than 33,000. In their ancestry, since the beginning of the selection experiment, not a single cross out of the race has occurred. At the same time no effort has been made to avoid inbreeding. Brother and sister and cousin matings are frequent in our records. Under these circumstances it is inevitable that the selected races should have become much "inbred." MASS SELECTION. 169 Our critics with a leaning toward the "pure-line" idea have insisted that nothing but brother-sister matings should have been employed in our experiments. We have several times endeavored to carry forward certain high-grade families on this basis, but have been unable to secure large enough numbers of offspring to make this possible; but we have in several cases produced families of considerable size, descended exclu- sively from a single pair of ancestors — notably in the case of our pure "mutant" race and in a race descended from one hooded and one wild rat, which race was continued through 8 filial generations. (See p. 21, Castle and Phillips.) It would have been impossible, in these and other races, to make as rapid progress as we secured through selection in our two principal races, for when only brother-sister matings are permitted, it often happens that a mate of proper grade can not be secured for an individual among its own brothers and sisters, though such a mate may be found among its cousins or more remote relatives. It being our first object to test the effectiveness of selection, we have made selec- tion of any individual within the group (series or family with which we were dealing) regardless of relationship, making the selection as rigid as the maintenance of a stock of considerable size would permit. More than once we have crossed the danger-line in advancing the standard of selection to such an extent that only small numbers of parents came up to it ; more than once we have had to relax our standard temporarily in order to keep the race alive.. That the long-continued inbreeding of our selected races has affected their vigor and fecundity is unquestionable. It is shown by the fact that the plus and minus races, which had a common origin many generations ago and have ever since been bred in the same room and under identical conditions, if crossed with each other, produce offspring of much greater vigor and fecundity than either parent strain. In this our observations on the effects of inbreeding are entirely in harmony with those of Darwin, Bos, Weismann, and of breeders of farm animals quite generally. Miss King is credited with the view that inbreeding of rats may increase their size, vigor, and fecundity, but this is cer- tainly contrary to common experience with these and other animals. It is probably true that under inbreeding it is possible, in exceptional cases, to isolate a strain relatively immune to ill effects from inbreeding (like Darwin's "Hero" morning-glory) or so inherently vigorous that it succeeds in spite of inbreeding. But it is very doubtful whether inbreeding of itself affects vigor other than disadvantageously. It is a sufficient test to cross-breed an inbred strain, in order to ascertain whether the inbreeding has increased or impaired its vigor. 170 INHERITANCE IN RATS. PLUS AND MINUS-SELECTION SERIES. The plus-selection experiment, when described by Castle and Phil- lips, had been carried through 13 generations, but the last 2 genera- tions were incomplete. The number of offspring included in genera- tion 12 (table 146) has now been raised from 590 to 682 and the number of offspring included in generation 13 (table 147) has been raised from 194 to 529. The mean grade of the parents for generation 12 has advanced from 4.09 to 4.10; that of the offspring has fallen from 3.94 to 3.93. Neither of these changes is of significant size. The correla- tion is now found to be 0.168 instead of 0.161. In generation 13 (table 147) the changes are greater, as might be expected from the greater change in the number of observations. The mean of the parents is now 4.13 (formerly 4.22) ; that of the offspring is 3.94 (instead of 3.88). The correlation is 0.117, as compared with 0.132, the value previously obtained. Generation 14 (table 148) includes 1,359 offspring of mean grade 4.01. They are descended mostly from parents of grade +4 or higher, mean 4.14. Generation 15 (table 149) includes 3,690 individuals, more than have been produced in any other generation of the series. The mean grade of the parents was in this generation advanced about a quarter grade to 4.38; that of the offspring advanced a little, to 4.07. Generation 16 (table 150) was also large, including 1,690 offspring. The grade of the parents was again advanced a little to 4.45; that of the offspring followed a similar amount, to 4.13. In the three generations (14 to 16) which have been added since the last report, the grade of the selected parents has been advanced by 0.32, from 4.13 to 4.45; that of the offspring has advanced 0.19, from 3.94 to 4.13 (the mean grade of the parents three generations earlier). The upper limit of variation of the offspring has meanwhile advanced from 5.25 to 5.87, the highest grade being found in a rat black all over except for a few white hairs on the chest. This rat has produced a few offspring of almost as high grade, though the most of his young are of much lower grade. In the minus-selection series, generation 13, in our previous report, contained 571 offspring. This number has now been raised to 1,006 (table 151), the mean grade of both parents and offspring being prac- tically unchanged by the additional young recorded. The parents are of mean grade —2.49, the offspring of mean grade —2.40. In the next generation (14) the offspring number 717, their mean grade being —2.48, that of the selected parents being —2.64. (See table 152.) Generation 15 includes 1,438 young of mean grade —2.54. The mean grade of the parents is —2.65. (See table 153.) MASS SELECTION. 171 Generation 16 is the largest in the minus-selection series. It includes 1,980 young of mean grade -2.63. The grade of the parents is -2.79. (See table 154.) Generation 17 (table 155) includes 868 young of mean grade -2.70. The grade of their parents is —2.86. Four generations of selection have thus been added to the minus series as it stood at the last report. The mean grade of the parents has been advanced from —2.49 to —2.86; that of the offspring from —2.40 to —2.70, the former is an advance of 0.37, the latter of 0.30. In the plus series the corresponding changes for one less generation of selection (three), were 0.32 and 0.19, respectively. In both series a change in the mean of the offspring attends that in the parents, coin- ciding with it in character but not quite equaling it in amount. The lagging behind of the offspring, as compared with their selected parents, gives a good illustration of regression, the phenomenon made familiar by Galton's researches, but explained away by Johannsen as due to a sorting-out action of selection on mixed races. The extent to which in these experiments the offspring lag behind their parents or "regress on their parents" is indicated in each table in the column headed "regression." Tables 146 and 150 illustrate particularly well how the offspring regress toward the general average of the race for the time being. The offspring of parents substantially the same grade as the general average of the race show no regression; the offspring of parents below this average show regression upward (indicated in the tables by the minus sign); the offspring of parents above the racial average show regression downward, the amount of the regression increas- ing with the aberrant character of the parents. If one examines either selection series as a whole (compare Castle and Phillips), he will notice that the point (toward which regression occurs) changes with the progress of the selection. At the beginning of the plus-selection series regression was toward a grade of about + 1.75 (see table 1, Castle and Phillips); after about 15 generations of plus selection it has advanced to +4.00. (See tables 148 to 150.) At the beginning of the minus-selection series, regression occurred toward a grade of 0 to -1 (Castle and Phillips, tables 16 and 17); in generation 17 (table 155) regression is apparently toward grade —2.62. These grades toward which regression occurs represent points of racial equilibrium or stability at which the race would tend to remain in the absence of further selection, but these points of equilibrium are capable of being moved either up or down the scale of grades at the will of the breeder, provided he has patience and persistency and will select repeatedly. Regression indicates that there is not complete agreement between the somatic and the genetic character of the parents selected. But the steady movement (in the direction of the selection) of the point of 172 INHERITANCE IN RATS. equilibrium toward which regression occurs serves to show that geno- typic as well as phenotypic fluctuations occur in the material on which selection is brought to bear. DeVries and Johannsen have damned the word fluctuation by ascribing to it purely phenotypic sig- nificance. Is it not worth while to rescue the term from its present odious position, since it is clear that variation having a genetic basis may in every way resemble somatic fluctuation, except in its behavior under selection? Fluctuation may conceivably be either somatic or genetic or both. No one, in advance of actual experiment, can tell what its nature is in a particular case. In the case under discussion the fluctuation is obviously partly somatic and partly genetic. The somatic fluctuation occasions regression, the genetic fluctuation per- mits a change (under selection) of the point toward which regression occurs — that is, in the general average of the race. Tables 156 and 157 show (generation by generation) the progress made by selection in modifying the racial character. It will be observed that as the mean advances in the direction of the selection both the upper and the lower limits of variation move in the same direction. The amount of the variation as measured by the standard deviation is less in the last half of the experiment than in the first half. It is also steadier, owing in part doubtless to the fact that the numbers are larger, and in part to a more stable genetic character of the selected races. But the genetic variability is plainly still large enough to per- mit further racial modification and there is no indication that it will cease until the hooded character has been completely selected out of existence, producing at one extreme of the series all-black rats, and at the other end of the series black-eyed white rats. FURTHER OBSERVATIONS ON THE "MUTANT" SERIES. Castle and Phillips described, under the name "mutants," 2 rats of the plus-selection series of very high grade. They proved to be heterozygotes between the average condition of the plus-selected race at that time, about +3.75, and a new condition, not previously known in our hooded races, but resembling that seen in "Irish" rats, which are black all over except for a white spot on the belly and would be classed on our grading scale as about + 5|. In later generations we secured animals homozygous for the darker condition just described (that of Irish rats). The homozygous "mutant" race proved to be very stable in color-pattern, varying only from 5| to 5f , with a majority of ani- mals graded 5|. Attempts to alter the modal condition of the race by selection have thus far proved futile because of our inability to increase the race sufficiently to afford a basis for selection. Its inbred- ness and its feebleness are perhaps causally related. The suggestion was made that the change from our plus-selected race, which had occurred in the mutant stock, might be due to some supplementary modifying factor, not to a change in the hooded factor itself. If so, a cross with a race lacking the hooded factor or its "modi- fiers" might serve to demonstrate their distinctness by separating one from the other. A wild race seemed best suited for a test of this hypothesis, since it would be free from suspicion on the possible ground of harboring either the hooded pattern or its supposed modifier, which had converted the hooded pattern into the mutant. It was to be expected, if the hypothesis were correct, that the mutant character was hooded plus modifier; that then a cross with wild should produce in F2 hooded young (lacking the modifier) as well as mutants and selfs. But if the mutant race had arisen through a change in the hooded factor itself, then the cross should produce only mutants and selfs, without hooded young in F2. Crosses have now been made on a sufficient scale to show beyond question the correctness of the latter alternative, which is entirely in harmony also with the results described in the preceding parts of this paper. Six homozygous "mutant" females of grade +5| were mated with wild males of the same race described in Part I. They produced 46 young, all gray like wild rats and of grades as follows: Grade 5^ 5| 5| 6 No 1 15 7 23 Exactly half of the 46 Yl rats bore no white spot, ^. e., were of grade +6. Seven more bore only a few white hairs (grade 5|) . The remain- der were very similar to the mutant parent in grade. Several matings were made of the FI rats, brother with sister, which produced 212 F2 young. About a quarter of these were black (non- 173 174 INHERITANCE IN RATS. agouti), the rest being gray (agouti). Both sorts included about equal numbers of individuals with and without white spots. No difference was observed in this respect between the progeny of spotted and of unspotted parents. Table 158 shows the F2 young grouped family by family according to grade. Three of the four families are descended from a single mutant grandparent; the fourth family is descended from two different mutant grandparents which were bred simultane- ously to the same wild male in the same cage. The 10 F2 young of this family may have been produced either by full brother and sister, or by half-brother and half-sister; it is uncertain which. All other F2 young were produced by brother-sister matings. It will be observed that the F2 young (table 158) which are white- spotted are in no case hooded. Their range of variation does not fall beyond that of the uncrossed mutant race. It is certain, therefore, that the "mutant" condition is not hooded plus an independent Mendel- ian modifier. It is a changed form of white-spotting, alternative to the form of spotting found in the race from which it was derived (the plus- selection series, generation 10). It is, without much doubt, also alter- native to the self condition of wild rats, though fluctuation in grade ob- scures the segregation, which may, very likely, be imperfect. This serves to confirm the general conclusion that throughout the entire series of experiments with the hooded pattern of rats we are dealing with quan- titative variations in one and the same genetic factor. GAMETIC COUPLING IN YELLOW RATS. Two yellow-coated varieties of the Norway rat (Mus norvegicus) made their appearance as sports or mutations in England a few years since (Castle, 1914) and are now recognized as distinct varieties by fanciers. Both are similar in appearance except for the eye color. In one variety the eye is pink, showing under gross inspection only the color of the blood in the retina. In the other variety the eye is a reddish-black, owing to the combined effect of the red-colored blood and the black-pigmented retina. Since the retinal pigment is much less in this variety than in rats with gray or black coats, the eye is redder. It will be convenient to distinguish the dark-eyed yellow variety as red-eyed, reserving the name black-eyed for gray or black rats. In the coats of both the pink-eyed and the red-eyed varieties of yellow rats black pigment is very feebly developed. It is in fact of a pale cream color. But the true yellow pigment seen on the tips of the hairs of gray rats is retained in full intensity in the yellow varieties. For this reason agouti varieties of yellow rats are much brighter-colored than non-agouti varieties. A non-agouti yellow variety has fur cream- colored throughout its length ; the corresponding agouti variety has fur of this same cream color at its base, where the fur of gray rats is black- pigmented, but the hair-tips are of a bright yellow color of exactly the same shade as the hair-tips of gray rats. Hence it is clear that in these yellow varieties of rats a genetic factor for black pigmentation has been affected without any apparent change in the genetic apparatus for producing ordinary yellow pigment. This is quite different from the genesis of yellow coat in most ro- dents— for example, in guinea-pigs and rabbits — in which black pig- ment is not apparently changed in character but merely in distribution, being "restricted" chiefly to the eye. In the yellow varieties of rats black pigment seems to be affected in the same way as in the pink-eyed variety of guinea-pigs and mice, viz, to be greatly weakened without affecting in the least the development of yellow pigment. The genetic behavior as well as the appearance of the pink-eyed yellow variation in rats is in every way parallel with the behavior of the variations known by the same name in mice and guinea-pigs. But red-eyed yellow in rats is a genetically distinct variation, as we shall presently see. In no other mammal does there occur a parallel variation, so far as I know. Both red-eyed yellow and pink-eyed yellow were found to be recessive Mendelian variations in crosses with black-eyed rats. From a cross between black-eyed and red-eyed an F2 generation of ^609 rats was raised, of which 452 were black-eyed and 157 red-eyed; expected, 457 : 152. From a cross between black-eyed and pink-eyed rats, cer- tain F! females were back-crossed with a pure pink-eyed male. They produced 46 black-eyed and 39 pink-eyed ; expected, 42 of each. 175 176 INHERITANCE IN RATS. The pink-eyed yellow and red-eyed yellow of rats are complementary loss variations ; for when the two varieties are crossed with each other they produce F: offspring which are either gray or black pigmented, according as their yellow parents did or did not transmit the agouti factor. These F! reversionary grays or blacks are paler in pigmenta- tion than ordinary gray or black rats, indicating that neither character in a heterozygous form is the full complement of the other. But it is evident that in homozygous form each is the full complement of the other, since in F2 and later generations grays and blacks of full intensity are obtained. The FI black-eyed animals (blacks or grays) obtained by crossing pink-eyed yellows with red-eyed yellows, if mated with each other, pro- duce an F2 generation containing (1) black-eyed young (black or gray), (2) red-eyed yellow young, and (3) pink-eyed yellow young. We have obtained thus far 324 such F2 young, of which 162 were of class (1), 90 of class (2), and 72 of class (3). If, as suggested, red-eyed yellow and pink-eyed yellow are due to mutually independent Mendelian factors, then F2 should contain four classes instead of the apparent three; wherefore it seemed probable that one of the three classes was really composite and that the three should be as 9:3:4. On this basis the F2 expectation would be 182 : 61 : 81 instead of the observed 162 : 72 : 90. Hence there appear to be fewer black-eyed young than are expected. Further, when we came to test the other F2 classes to discover which of them was com- posite, we found very few individuals which would fall in the hypotheti- cal fourth class transmitting both pink-eyed and red-eyed yellow in the same gamete. Instead of 1 in 16 as expected, we have been able to discover a much smaller number of double recessives. Both a defi- ciency in double recessives and a deficiency in double dominants (the black-eyed class), which have been observed among the F2 rats, would be expected if pink-eyed yellow and red-eyed yellow are due to " linked genes," i. e., to factors located near each other in the germ-plasm. For in the cross under consideration each form of yellow enters the F! zygote in a different gamete. Hence, in the gametes arising from such zygotes we should expect the two forms of yellow to show mutual repulsion. If they did so, then the gametes formed by F! zygotes, of the four possible combinations, RP, Rp, rP, and rp, would not be equally numerous, but Rp and rP should be more numerous than RP and rp. That this is true is indicated by the facts presently to be stated. To test the gametic composition of the F2 yellows, those which were red-eyed were mated with pink-eyed yellows of pure race, and those which were pink eyed were mated with red-eyed yellows of pure race. For it was known that, since red-eyed yellow is a recessive variation, every red-eyed F2 yellow must be homozygous for red-eye, but conceivably it might be either heterozygous for pink-eye or might lack it altogether. A cross with the pure pink-eyed yellow race would GAMETIC COUPLING IN YELLOW RATS. 177 decide between these possibilities. Further, it was clear that every pink-eyed F2 yellow must be homozygous for that character, which is also recessive, but might be either homozygous or heterozygous for red-eye without affecting its appearance, or might even lack the gene for red-eye altogether. A cross with pure red-eyed animals would suffice to show in each case which possibility was realized. In accord- ance with this reasoning the proposed tests have been made in the case of 45 red-eyed and 40 pink-eyed F2 yellows. Of the 45 red-eyed yellows tested, 32 have given exclusively black- eyed young (blacks or grays), no test being considered adequate which did not produce 4 or more young; but 13 of the tested animals gave a mixture of black-eyed and of red-eyed young in approximately equal numbers. The former group, numbering 32, evidently lacked the gene for pink-eye, since they always produced atavists in crosses with pink- eyed yellows; the latter group, numbering 13, were evidently hetero- zygous for pink-eye, since only part of their young were atavistic. Of the 40 pink-eyed F2 yellows which were tested, 27 produced only black-eyed young; these evidently lacked the gene for red-eye. Ten others produced both black-eyed and red-eyed young, being evidently heterozygous for red-eye. Three have produced only red-eyed young, which shows them to be homozygous for red-eye as well as for pink-eye. Hence they are the double recessives, expected to be one-sixteenth of all F2 rats if no linkage occurs, but less numerous if linkage occurs. We are now in a position to estimate the strength of the linkage shown. If we designate by r the recessive gene for red-eye, and by p the recessive gene for pink-eye, then in the current Mendelian terminol- ogy the following F2 classes are to be expected in the frequencies shown, if no linkage occurs : Black-eyed. Red-eyed. Pink-eyed. 1 RRPP 1 rrPP 1 RRpp 2 RrPP . . . 2 rrPp . . 2 Rrpp 2 RRPp . . 1 rrpp 4RrPp 9 3 4 For the present we may pass by the black-eyed classes, since none of these were individually tested. The individual tests already described have shown the existence of the expected two classes of red-eyed arid three classes of pink-eyed young, but in proportions very different from those given in the table. Among the red-eyed, instead of the expected 1 rrPP : 2 rrPp, we observe 32 : 13. Among the pink-eyed, where we expect 1 RRpp : 2 Rrpp: 1 rrpp, we observe 27 : 10 : 3. These are very different frequencies from those expected, and they strongly suggest linkage. How strong is the linkage? We may estimate it 178 INHERITANCE IN RATS. from the actual proportions of the four possible kinds of gametes which the FI parents produced. With no linkage these gametes should be of four sorts, all equally numerous, viz, RP + Rp + rP + rp. Linkage would tend to increase the proportion of the two middle classes (Rp and rP, the original combinations) at the expense of the extremes (RP and rp, the double dominant and double recessive classes) . The latter may be called " cross-over" classes, the former "non-cross-over." In producing the 85 F2 yellow rats which were tested, twice that number of gametes were concerned, viz, 170. From the demonstrated genetic constitution of the tested animals, we can estimate how many cross- over and how many non-cross-over gametes entered into each. Zygotes. Cross-over gametes. Non-cross-over gametes. 32 rrPP . 64 13 rrPp 13 13 27 RRpp 54 10 Rrpp 10 10 3 rrpp 6 85 29 141 The estimated proportion of cross-over to non-cross-over gametes is seen to be 29 : 141 or 1 : 4.8. In the terminology of Bateson and Pun- nett this would be a reduplication series lying between 1:4:4:1 and 1:5:5:1; in the terminology of Morgan, 17 per cent of the gametes formed by FI individuals are cross-over gametes. We can test this linkage theory in another way. If linkage exists it should modify the proportions of the apparent classes in F2 as well as of the real classes, which we have just been considering. The apparent classes are three, viz, black-eyed, red-eyed, and pink-eyed, with observed frequencies of 162 : 90 : 72. If no linkage exists the expected frequencies are 182 : 61 : 81, which deviate considerably from the expected frequencies. But if linkage exists, it will lessen the discrep- ancies. Linkage of 17 per cent strength will change the expectations to 164 : 79 : 81 . This alteration shows agreement almost perfect in the case of the black-eyed class, a much reduced discrepancy in the case of the red-eyed class, and no change in the pink-eyed class — on the whole a much improved agreement between expected and observed frequencies. Sturtevant has called attention to the fact that double recessives could occur among our F2 animals only as a result of cross-overs occur- ring simultaneously in the gametes of both parents, a fact which Wright and I considered too obvious to demand comment in our preliminary paper, but recognized in our calculation by counting two cross-over gametes for every double recessive zygote. Sturtevant has questioned the adequacy of our tests in the case of these doubly recessive indi- viduals because apparently he had formed the idea, from studies made on insects, that crossing-over could occur only in the gametogenesis of GAMETIC COUPLING IN YELLOW RATS. 179 one sex. I may say, therefore, that the classification of two animals as double recessives made in our preliminary paper was based on tests which had produced 14 and 9 yellow young respectively. The only possible alternative classification would have involved an expected 1 : 1 ratio of black-eyed to red-eyed young. The chances are overwhelm- ingly great against the observed results being departures due to ran- dom sampling from this expectation. The additional case of a double recessive reported in this paper is so classified on tests which, to the present time, have produced all together 28 yellow young. The num- ber of young produced in each of the other tests is indicated below. Tests taken to indicate that the parent was of the formula rrPP produced only dark-eyed young (gray or black-coated), as follows: No. of young.... 4 5 6 7 8 9 10 11 12 13 16 17 19 Cases 244522 1 3 1 2 3 1 2 = 32 Tests showing the parents to be of the formula rrPp gave the follow- ing numbers in 13 tests: Dark-eyed : Pink-eyed ... 2 : 6 4:2 5:4 1:5 3:3 5:3 4:6 1:5 5:5 3:4 3:2 1:5 6:3 Pink-eyed animals were classified as of formula RRpp on the basis of the following tests, which yielded only dark-eyed young: No. of young.... 4 6 7 8 9 10 11 12 13 14 15 16 18 Cases 213222161311 2 = 27 Pink-eyed animals were shown to be of formula Rrpp by the follow- ing tests: Dark-eyed: Red-eyed... 10: 7 8:4 1:4 7:8 6:7 5:6 3:1 4:6 4:2 8:2 Both red-eyed and pink-eyed yellow rats, when crossed with albinos, produce an F: generation consisting exclusively of black-eyed (black or gray) young. F2 from the red-eyed cross consisted of black-eyed, red-eyed, and albino young, and F2 from the pink-eyed cross consisted of black-eyed, pink-eyed, and albino young. If no linkage occurs the expectation in each case is 9 : 3 : 4, and we at first supposed that this was the ratio approximated. But a summary of all litters thus far obtained indicates a probable linkage between albinism and the two yellow variations. Thus, red-eyed non-agouti yellows mated with albinos from our plus- selected hooded race produced 17 black Fl young. These have given us 58 F2 young, of which 30 are black-eyed, 18 red-eyed, and 10 albinos. A 9 : 3 : 4 ratio would call for 32.5 : 11 : 14.5. It is evident, therefore, that we have too many red-eyed young and too few black-eyed and albinos. Linkage (in this case, repulsion) between red-eye and albin- ism would tend to increase the number of red-eyed and to decrease the number of black-eyed without changing materially the expecta- tion for albinos; hence, linkage seems probable. Linkage involving 1 cross-over to 3 non-cross-over gametes, or 25 per cent cross-over 180 INHERITANCE IN RATS. gametes, would give an expectation of 29.9 black-eyed : 13.6 red-eyed : 14.5 albinos, which agrees much better with the observed numbers (30 : 18 : 10) than does the 9:3:4 distribution. But if red-eye is linked with albinism as well as with pink-eye, then albinism and pink- eye should be linked with each other. Apparently such is the case, for three F2 litters from the cross pink-eye X albino include 12 black, 12 pink-eyed, and 3 albino young. A 9:3:4 ratio (expected if no linkage occurs) would call for 15 black, 5 pink-eyed, and 7 albinos. Linkage of 5 : 1 would call for 14 : 6 : 7, and perfect linkage would call for 14 : 7 : 7. It is evident that the observed numbers of blacks and albinos are too small on any of these hypotheses, but the existence of linkage would tend to diminish the number of blacks and albinos in proportion to the number of pink-eyed, which is the nature of the deviation observed. To determine definitely whether linkage really occurs between the yellow variations and albinism, and if so, what is its strength, further experiments are needed, which are now in progress. It will also be desirable to determine whether the linkage strength is the same in both sexes. SUMMARY. Two yellow variations in rats which have recently arisen as muta- tions show mutual repulsion in heredity. When crossed with each other they produce an F! generation composed exclusively of rever- sionary dark-eyed individuals. The F2 young are of three apparent classes, dark-eyed, red-eyed, and pink-eyed. Their numerical propor- tions deviate somewhat from the typical 9:3:4 ratio. Further, the proportions of the several expected classes of red-eyed and pink-eyed young do not agree with those usually observed in an F2 Mendelian population. But in both cases the deviations are largely accounted for by the supposition that the genes of the respective yellow variations are "linked" (in this case showing repulsion) and that the proportion of "cross-over " gametes is about 17 per cent, or in other words, that non- cross-over gametes are about 4.8 times as numerous as cross-over gametes. NOTE. — In the foregoing discussion it has been assumed that the ratio of cross-over to non-cross-over gametes is the same among gametes which take part in producing yellows as among those which take part in producing black- eyed individuals. Theoretically it should be slightly different, as the following table will show: Ratio cross-over Per cent Per cent among Per cent among to non-cross-over cross-over gametes pro- gametes produc- gametes. gametes. ducing yellows. ing black-eyed. 1: 1 50 42.9 55.6 1:2 33.3 29.4 36.8 1:3 25 22.6 27.3 1:4 20 18.4 21.6 1:4-4 IS. 5 17.1 19.8 1:5 16.7 15.5 17.8 1:6 14.3 13.4 15.2 TABLES. 181 TABLES. Table 141 shows the classification of extracted hooded first Fz young obtained from crossing hooded rats of the plus-selected series with wild rats. TABLE 141. Hooded grandparents. Grade of hooded grandchildren. Total hooded. Total non- hooded. Means of hooded. 1* If 2 2} 2h 23 3 31 31 33 I 4 95513, -Hi, gen. 10. ... 1 3 1 2 1 1 7 2 8 4 6 3 1 5 4 1 1 7 6 3 1 1 41 22 5 3 2 107 68 27 12 5 3.05 3.28 3.51 3.17 3.37 cf 6348, +4, gen. 10 . . 9 6955, +4, gen. 12 95513, +41, and 96600, +4i gen. 12 ? 95513, +4}, and 96955, +4, gen. 12 1 1 Totals 1 4 2 2 9 14 11 12 16 L> 73 219 3.17 Table 142 shows the classification of extracted hooded second Fz young obtained from crossing first F2 hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of 9 5513, +4J, generation 10, on the side of both parents. TABLE 142. Hooded grand- parents. Grade of hooded grandchildren. Total hooded. Total non- hooded. Means of hooded. 2 21 2* 21 3 31 31 3! 4 99619, +2 1 1 2 1 2 5 1 1 2 2 4 11 6 2 1 3 4 7 8 7 30 1 4 1 5 2 5 13 10 30 22 16 8 28 24 22 104 68 42 3.37 3.40 3.06 3.62 3.47 3.55 3.70 c?9686, +21... 99620, +21... 99729, +2 1 1 1 2 1 1 1 c?9727, +3... . 99728, +3... 1 1 2 3 1 99621, +31 .. 1 Totals . . . 1 2 3 4 6 13 2S 11 98 296 3.47 Table 143 shows the classification of extracted hooded second ¥2 young obtained from crossing first Fa hooded rats (table 141) with wild rats. The hooded grandparents were themselves grandchildren of c?6348, +4, generation 10, on the side of both parents. TABLE 143, Hooded grand- parents. Grade of hooded grand- children. Total hooded. Total non- hooded. Means of hooded. If 2 21 01 63 2- 3 3| 31 Q3 •J4 4 c?9639, +2 1 2 1 1 3 6 1 4 15 4 6 1 39 6 1 27 16 21 9 2 110 16 10 76 47 74 40 3 3.24 3.17 3.50 2.90 3.28 3.48 3.36 3.87 99704, +2f . . . . 99765, +3. . . 1 4 5 4 2 4 2 8 3 1 1 2 2 1 7 9 9747, +3J 99703, +3£ . . . 1 7 1 1 1 1 2 1 1 7 2 5 2 1 1 1 1 9 9705, +3^ . . . 99748, +3£ . . 9 9796, +4 Totals. . . o 10 2 2 8 23 8 35 24 121 376 3.22 182 INHERITANCE IN RATS. Table 144 shows the classification of extracted hooded second F2 young obtained from crossing first F2 hooded rats with wild rats. The hooded grandparent, o" 9660, +3f, was a grandson of 96955, +4, generation 12, on the side of both parents. The hooded grand- parent, c?9711, +3^, was a grandson, on the side of one parent, of 9 5513, +4j, generation 10, and on the side of the other parent, of 96955, +4, generation 12. (See table 141.) TABLE 144. Hooded grand- parents. Grade of hooded grandchildren. Total hooded. Total non- hooded. Means of hooded. 2 21 2* oa •^4 3 01 •J* '•$ - OJ 3! 4 d"9660, +3f... c?9711 +3j 1 1 1 2 1 2 2 4 5 4 9 2 2 1 21 16 44 33 3.50 3.28 Totals 1 ^^~ 1 3 3 6 9 11 3 37 77 3.40 Table 145 is a combination of tables 142 to 144, in which the second F2 young are classi- fied according to the grade of their first F2 hooded grandparent. TABLE 145. Grade of hooded Grade of hooded grand- children. Total Total Means „!• grand- parents. U 2 2J 2| 2| 3 31 3J 3f 4 hooded. non- hooded. Ol hooded. 2 1 2 1 3 6 5 16 6 1 41 118 3.25 2| 2 1 2 1 3 4 12 8 1 34 90 3.29 3 1 1 2 4 7 18 15 5 53 182 3.48 31 1 7 2 4 9 6 10 13 7 59 151 3.22 3* 1 1 4 9 3 11 13 4 46 161 3.39 3i 1 1 1 2 5 9 2 21 44 3.50 4 1 1 2 3 3.87 4 6 3.02 2 12 15 32 27 72 65 21 256 749 3.34 Table 146 shows the classification of generation 12, plus-selection series. This is an enlargement of table 12 of Castle and Phillips. TABLE 146. Grade of parents. Grade of offspring. Totals. Means. Regres- sion. 21 2| 2f 3 31 31 Q3 •J4 4 41 4| 4i 5 51 3f 3| 4 4* 41 4| 4* 4| 4! 4| 5 4 3 12 25 11 3 20 23 62 106 25 6 1 3 7 21 66 91 35 14 4 4 4 6 12 30 16 10 1 3 35 57 164 267 95 45 7 11 3.83 3.94 3.91 3.87 3.95 4.17 4.14 3.91 -.08 -.06 .09 .25 .30 .20 .36 .71 o 5 6 5 7 2 1 3 2 3 1 1 1 1 2 1 1 1 2 3 2 1 1 1 4.75 .25 4.10 2 3 5 58 246 242 S2 25 12 4 3 682 3.93 .17 TABLES. 183 Table 147 shows the classification of generation 13, plus-selection series. This is an enlarge- ment of table 13 of Castle and Phillips. TABLE 147. G"ade of parents. Grade of offspring. Totals. Means. Regres- sion. 2i 3 3| 3| 31 4 41 4| 41 5 £i °4 3J 3! si 3| 4 4J 41 4| 4* 4! 41 41 1 1 1 3 3.50 0 1 4 33 11 7 2 1 3 9 60 32 23 8 15 1 3 11 7 59 33 33 7 17 2 1 9 3 4 25 13 13 6 10 1 1 1 1 17 5 1 1 5 1 3 2 2 3 1 1 1 1 1 1 21 31 205 96 80 29 50 4 5 5 4.08 3.90 3.90 3.92 3.93 4.03 4.05 4.00 3.95 4.05 -.33 -.03 .10 .20 .32 .34 .45 .62 .80 .82 1 1 1 1 2 7 1 1 1 1 4.13 1 3 11 61 155 172 76 32 13 4 1 529 3.94 .19 Table 148 shows the classification of generation 14, plus-selection series. TABLE 148. Grade of parents. Grade of offspring. Totals. Means. Regres- sion. 2* 21 3 31 3* Q3 "4 4 41 4| 41 5 51 5| 3* 3| 31 3J 4 44 41 4| 4* 4| 41 4J 5 5* 51 2 2 11 28 4( 19 6 2 3 6 9 32 52 84 74 25 24 12 11 5 1 3 4 45 63 122 72 48 36 31 13 16 7 1 12 24 115 184 306 241 130 100 101 58 45 33 3.83 4.02 3.90 3.97 3.92 3.99 4.04 4.04 4.16 4.23 4.17 4.33 -.33 -.39 -.15 - .10 .08 .13 .21 .33 .34 .39 .58 .54 7 18 28 50 56 42 29 37 15 14 14 1 5 8 6 15 8 6 12 11 8 6 1 1 2 1 1 2 3 2 3 1 2 3 5 2 3 1 1 1 1 2 2 1 1 1 1 1 4 1 1 1 6 4 4.25 4.75 .87 .50 1 1 4.14 1 3 4 113 335 461 315 89 24 9 4 1 1,359 4.01 .13 184 INHERITANCE IN RATS. Table 149 shows the classification of generation 15, plus-selection series. TABLE 149. Grade of parents. Grade of offspring. Totals. Means. Regres- sion. 24 .,:< *i 3 31 34 31 4 4| 4 1 •±2 I:,1 5 5} .V, 3f 31 4 4| 4| 4f 44 4f 4! 41 5 5| 51 1 2 1 2 1 1 1 3 5 16 29 22 18 9 3 3 3 10 58 184 183 207 99 37 25 13 1 7 30 156 670 721 969 506 237 168 146 43 18 19 3.57 3.80 3.91 4.00 4.02 4.06 4.10 4.11 4.22 4.37 4.27 4.37 4.36 .18 .07 .09 .12 .23 .31 .40 .51 .53 .50 .73 .75 .89 11 46 255 296 357 159 87 38 25 9 4 2 2 28 165 165 290 175 88 58 42 21 4 10 7 29 44 71 48 14 24 34 10 7 5 3 8 21 6 5 5 15 2 3 1 1 1 1 1 1 3 7 2 9 12 1 G 4 2 1 1 1 1 4.38 2 9 108 820 1,289 1,048 293 69 36 11 4 3,690 4.07 .31 Table 150 shows the classification of generation 16, plus-selection seiies. TABLE 150. Grade of parents. Grade of offspring. Totals. Means. Regres- sion. 31 34 3i 4 4J 44 4f 5 51 54 5f 51 4| 41 4| 44 4| 4| 41 5 5| 1 4 4 9 7 1 26 34 149 25 12 5 64 64 316 82 50 12 23 8 1 37 40 271 69 61 16 26 25 8 8 5 58 18 26 19 8 15 6 139 149 816 206 166 66 69 61 18 4.04 4.02 4.08 4.10 4.24 4.47 4.21 4.44 4.39 .08 .23 .29 .40 .38 .38 .66 .56 .73 1 9 5 11 10 3 7 1 1 2 4 G 2 9 1 1 1 1 9 1 1 2 1 1 1 1 4.45 25 252 620 553 163 46 16 9 4 1 1,690 4.13 .32 Table 151 shows the classification of generation 13, minus-selection series. This is an enlargement of table 28 of Castle and Phillips. TABLE 151. Grade of offspring (minus). Grade of Regres- parents. sion. If 2 21 01 •"2 2-1 3 3i 34 -21 7 43 25 28 12 1 116 2.25 0 -2| 8 74 80 87 56 19 5 329 2.39 0 -24 8 65 65 92 46 11 2 1 290 2.38 .12 -2f 3 23 33 58 44 4 4 1 170 2.50 .12 -2f 1 5 4 16 12 7 1 46 2.56 .19 -21 / 8 10 7 2 34 2.42 .45 -3 4 1 6 5 3 2 21 2.59 .41 -2.49 27 221 216 297 182 47 14 2 1,006 2.40 .09 TABLES. 185 Table 152 shows the classification of generation 14, minus-selection series. TABLE 152. Grade of parents. Grade of offspring (minus). Totals. Means. Regres- sion. 1 11 14 H 2 2| 2i 2f 3 31 3^ -2* -2| -24 -2f -21 -21 -3 -3* -31 0 3 — ,7< FI ^nd, race C x Cavia cutleri, adult. Fig. 21, Ft hybrid, race B (Plate 5, rig. 34) x Cavia cutleri, adult. PLATE 4 25 26 27 28 29 30 F2 hybrids, race C x Cavia cutleri. Fig. 25, agouti; 26, black; 27, chocolate; 28, cinnamon; 29, yellow; 30, albino. PLATE 5 31 ,. V 32 Some new guinea-pig color varieties. Fig. 31, red-eyed cinnamon; 32, red-and -pink-eyed black spotted with white; 33, pink-eyed golden agouti spotted with red and with white; 34, albino with sooty fur and black pigmented extremities, similar to race B. PLATE 6 Femurs of Cavia cutleri (male and female), of race B, and of their Fj and F, hybrids, showing complete range of variation in each. Natural size. 36 * 4 f 35 37 38 39 Fig. 35, a scale of grades for piebald rats. Fig. 36, a pair of piebald rats and their nine young. Fig. 37, a smooth guinea-pig. Fig. 38, a well-rosetted rough guinea-pig, grade A. Fig. 39, a poorly resetted rough guinea-pig, grade C. ' 10