PEAS OLA Veralavelatalatabctute htctatiee teeta bchet ‘ ay eyes { caratal ates Tatas sataorara eet Say erad hee. ah etter ibarat ox Cire heh a: te t titty Te ; Metin ‘ Ta mc abe tie x u Lee ONS 4 “4 ; Ct X > ates Daheatah ? Cy ; ee deci tate eke CUE ¥ v as 4 eer he Me erates Sette Gh. t haha higeg CeCe lg CU clog Gat. Cyietigh CPG E ata tateta ty Bes hale tr’ ee Huy d : a atae ras ae L Hips. t LyX ae Pt 474 tre gOS RS: 8: Lot ASCOT eE a MERE Le ky M4 ba heatat Parhtsiae tae taereaht secsenaatt eet é hte y .. 3 ‘ Lod f at Sates thy ne apardgecetes atelatate tr Ee tet seo hei Ry ba arcade PEC Nee it, ete 5 let MOLE tates oh : i Peete 1 eet t epee yah 17 4 + KEL IS oh rotetbres 5 aoe va rat ety fe? ots Roretiaes jtivetetySeirtatetnes ish Mai plese, z satay ‘ PEGE Gt ean ats if ate aoc kty eget CCCs fre behets ak tite ees ft rng As tee a q e th tee Ane ttt de cle ye ie f he d rors ~ " eratete is Lererers gteh 5 oe ean tat oan ieee cetety tain Book one Se Sy wee Puan al poche af, ; ‘ oh io: i % (x t Ps ‘ Lege aie! eter tearecaetn tats if ata Cette . rf has 1G Fety At if, x eh ati: ope ( i efivetnrces atthe ; ester pettetutytann Leh he EAE LEG Shea eatateacheb ii Parotichtatv lan sade ae 1 ie arate Ware : belle tent tees ee nbattatahaseicetahatny ayeae Mesos 5 ie ae Bi fe 5 Et 4 Ct LE Pantani: dibea roebrtacht Bits : 3 Pitiaess entra erne erate} ; ‘ et eh Tt : “ 44 : t i beet cet eS ofits 3 sabia Gots Se 0%, tee gt Eek fa terater] tel t Pest alene Slaeiprarapete mt Paine be a, eles Geber saree s2 eee ~ < a2 <= Satetsts eS ete trig oertrtee a rena a staat? ochelvty Lit ovep erty tees THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY Wiuuram E. Caste FRANK R. LILuIE Harvard University University of Chicago EDWIN G. ConKLIN JAcquEs Lors Princeton University Rockefeller Institute CHARLES B. DAVENPORT ‘Tuomas H. Morcan Carnegie Institution Columbia University HERBERT S. JENNINGS Grorce H. PARKER Johns Hopkins University Harvard University Epmunp B. WILSON, Columbia University and Ross G. HARRISON, Yale University Managing Editor VOLUME 19 1915 hy THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY . PHILADELPHIA, PA. CONTENTS NOSE JULY PAnViINED ew bRIDGESSAalinkacesin Drosophila nce neat selene) seers anseaiotor | 1 Jacques Lorn anp HarpotpH WastTENEYsS. The relative efficiency of various parts of the spectrum for the heliotropic reactions of animals and plants.................. 23 Lesuie B. Arty. The orientation of Amphioxus during locomotion.................... 37 Maynie R. Curtis AND RaymMonp Peary. Studies on the physiology of reproduction in the domestic fowl. X. Further data on somatic and genetic sterility.............. 45 Epwin Carteton MacDowe tu. Bristle inheritance in Drosophila. I. Extra bristles. SEM OTIDOS ee Ne eee esis Ge Rtas Seki. «= SIS Vege Re shee ROR auckeh Shc ats Meeoeney ich? ome 61 NO. 2 AUGUST C.M. Jackson. Changes in the relative weights of the various parts, systems and organs of young albino rats held at constant body-weight by underfeeding for various periods. ‘DURE VACA Des eee teeta UO tC eh RM Cac, CCR era Ci ompR ht 99 Kk. 8S. Lasuiey. Inheritance in the asexual reproduction of Hydra. Ten figures........ 157 Rosert H. Hurcuison. The effects of certain salts, and of adaption to high tempera- tures, on the heat resistance of Paramecium caudatum. Onefigure................ 211 Gary N. Cauxtns. Didinium nasutum. I. The life history. Twelve figures (one NO. 3 OCTOBER Morris M. Weuts. The reactions and resistance of fishes in their natural environment COMES ALS ewe AIG Gis LUQUE 5 a.cais Srnec horesa. m= » seks ERMERMIa, abn aerate Stays tae oan See onc sigs ona 243 T. H. Moraan. The predetermination of sex in phylloxerans and aphids. Five text EYRE STU a C6 HALA CO 6) 25 in ee ene oP =" ee PS SS eee Bs ae 285 CHARLES PackarD. The effects of the beta and gamma rays of radium on protoplasm. fwenty-averigures (taree plates)).2 2... .. .. Steppers ts Suis aos siareele some ¢ beneuee amb 325 TuHeEopPuHitus S. ParnrEer. The effect of carbon dioxide on the eggs of Ascaris. Fifteen HE XU LoMRES TAN Cb MLCes pla veSecvme is o>. sve CRM ee ee se Slee eves organ mene aS: 355 iv CONTENTS NO. 4 NOVEMBER ' : Ruts J. Srocxrne. Variation and inheritance in abnormalities occurring after conjuga- tion in Parameciim, caudatum. “Cwenty figures... .. 4§55:...<..-s26..<-.)25eeeee 387 Austin Rautru Mippieron. Heritable variations and the results of selections in the fis- sion rate of Stylonychia pustulata. Seventeen figures..................000cceeeeeee 451 CHARLES ZELANY AND C. T. Senay. Variation in head length of spermatozoa in seven Adgiional ‘Species of insects: Hight fieures 055.62. 2c cecosecee ce tee 505 CHARLES ZELANY AND W. E. Marroon. The effect of selection upon the ‘bar eye’ mutant of Drosophila. “Rive digures.... 2.16 2 ieck « Ds asses oe ele ene ee ee 515 Mary B.Srarx. The occurrence of lethal factors in inbred and wild stocks of Drosophila. Lwordiaeramsy qt oA Ss Ohh. EIT e NS boas 1 + ode SO eee 531 Jacgugs Lors anp Mary MircHett CHAMBERLAIN. An attempt at a physico-chemical explanation of certain groups of fluctuating variation............:......-.-++--+: 550 A LINKAGE VARIATION IN DROSOPHILA CALVIN B. BRIDGES From the Zoological Laboratory, Columbia University | In the breeding work upon Drosophila done in this laboratory, it has been the practice to allow a female to lay eggs only for a period of about ten days. This is the average length of time from the mating of a female to the emergence of her offspring. But this first brood does not by any means exhaust the eggs of a female; if she is transferred to a fresh culture bottle she will lay as many eggs in this as in the first, and will continue to lay for forty or fifty days. Ordinarily, then, we obtain a sample of from 200 to 400 flies from a female, although three times as many might be obtained. It seemed to me that the full output of each female would give a truer index than the one that we were using. Accordingly, in working out the linkage relations of several mutations, I raised from each of the F, females of a few experiments a second brood. In cases involving the second chromosome a remarkable relation came to light when the results of the second broods were compared with those from the first. There had been a change in the linkage so that both in the totals for each experiment and in a great majority of the individual cultures the percentage of crossing-over had fallen significantly. Or, in other language, the ‘coupling strength,’ or ‘gametic ratio’ had risen. This change, while very interesting theoretically, promises further to become an aid in the study of the mechanism of linkage. In the case of the first (sex) chromosome a large amount of data shows no change from first to second broods. In the case of the third chromosome present data show an increase in the percentage of crossing-over, but because of the small number of cases the rise may not be significant. 1 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 1 JULY, Le) CALVIN B. BRIDGES The most efficient experiment by which to determine the amount of crossing-over between gens is the back-cross. Here a multiple heterozygote is tested by mating to the corresponding multiple recessive. When this is done, in the next generation produced, there are two contrary classes representing the original or P;, combinations, and two other contrary classes of cross- overs. The percentage of crossing-over is given directly by division of the sum of the crossover classes by the sum of all the classes. It was in this way that the percentage of cross- overs was calculated for the following tables. TABLE 1 P, purple vestigial X wild. B.C. Fy wild type 2 X purple vestigial ica NON-CROSSOVERS | CROSSOVERS PER CENT | REF. ST UaInS alia aie F = aoe! LOLAL OF CROSS- | A ee |Wildtype| Purple — Vestigial » OVERS A 78) M202 16 lon |’ 402 ve) A’ 152 297 13 14 406 6.6 ee B 91 100; 4. a8 ie |. 922 14.0 | B’ 69 104 12 Si «| 193 10.3 =o C 165 | 150 17 19 351 10.3 Cc’ 191 216 18 17 442 7.9 | —2.4 D 140 149 20 15 324 102893) D’ 116 122 9 le OST 5D) nas E 191 14 | 20 Gee ai add 9.0 |. a! 196 299 11 22 458 7.2 Slee I 202 |* 226 20 22 470 8.9 F’ 9 228 25 20 470 9.6 ==Ou7 : 105 158 17 17 207 ii a a G’ 188 232 17 45 6.9 as H 123 140 26 30 319 17.6 H’ 129 179 | 11 20 339 | 9.1 —8.5 | Ists 1195 | 1339 | 154 151 2839 | 10.7 | 2nds 1238 | 1539 | 116 119 3010 | TAs = ong = ae . R40; Fi LINKAGE VARIATION IN DROSOPHILA a The mutant races used in these experiments will be fully described in a. series of papers by Morgan and Bridges. In this paper I am considering only the small fraction of data in which we have records of more than one brood from a single female. THE SECOND CHROMOSOME DATA The first case in which this relation for the second chromosome was clearly shown was that of purple and vestigial given in table 1. Of the eight females whose tests are given in the table, seven showed a decrease in the percentage of crossing- over and one (F) showed an increase, which, however, was smaller in amount than any of the decreases. Likewise, the totals for the second broods when compared with the totals for the first broods showed a decrease of nearly three units in the amount of crossing-over. That differential viability has little or nothing to do with this difference is evident from the regular totals. This.is shown even more clearly by the converse case (‘repulsion’), where the difference is entirely negligible. Tn all of these experiments contrary classes are affected similarly and to a like degree, which would not be the case if viability were causing the change. : Of the six females tested (table 2) all showed a decided drop, and the totals show a rather greater drop (5 units) than in the ‘coupling’ case. In obtaining data upon any linkage case it is best to have half the data in the form of ‘coupling’ and half as ‘repulsion’ experi- ments. It has been shown by comparative breeding tests that differential viability can be to a great extent eliminated by careful attention to the conditions of breeding—particularly by breeding in pairs in large culture bottles with the right amount of well prepared food. We may offset even this remaining disturbance by balancing the viabilty of a certain class against itself. For example, let us say that in the case of purple vestigial, the class vestigial is poorly viable. If, then, vestigial occurs in an experiment as a crossover class, that class will be too small and a false linkage value will be obtaihed. The remedy is to 4 CALVIN B. BRIDGES balance against these flies an equal number in which vestigial occurs as a non-erossover. In this case the error will be the opposite of the previous one, and by combining the two experi- ments the errors should balance and give a better approach to the true value. If each class is thus balanced the error should TABLE 2 P, purple X vestigial. B.C. F; wild type @ X purple vestigial ij oS NON-CROSSOVERS | CROSSOVERS | PER CENT | REF. = = |—— Tn > TOTAL OF CROSS- | A | Purple | Vestigial | Peal Wild type Oye I Dean) PZB. |. 26 21 8822" ao sam 1 | 200 165) | 12 14 391° | 64 | ==5x6 J | 198 176 | 23 23 A205 4) lO ay e242) 195 19 26. 482, | . 9:3. | ea ey, K 252 2279 340 a Naaas 551 1gsve" Ke 198 178 | 26,55 20 422 | 10.9 —2.2 M 205 | 158 Dill 32 422 14.0 M’ | 213901 246 | 14 23 496 ee —6.6 N | 66 54 6 11 137 12.4 NG | 66 64 4 7 141 1:8) || Gaeeo ak 180.8" 172. :\- 30 32 49380)" 14g 4 O’ Pl 225 13 18 473 Gro.) | Sat Ists 1067 965 146 157 2335 | 13.0 | 2nds 1136 1073 88 108 2405 8.1 | —4.9 otal...) 2202 2038 234 265 | 4740 | TABLE 3 Linkage of purple and vestigial with balanced viability (ists) CLASS pipe > | CROSSOVERS TOTAL pepe a? Wal tat ye ee ce ee ois ee: 1339 157 Bunpley seine bias ies): 1067 | 154 Westiowall args tis eee to! u 965 151 Purple vestigialoges.. ...-.- 1195 146 Ta talits 5.) aeae, ne anek e! » 4566 | 608 BIAS olden tS LINKAGE VARIATION IN DROSOPHILA 5 be very small. If an equal amount of data for ‘coupling’ and ‘repulsion’ be combined, each possible class will appear in the required manner both as a non-crossover and as a crossover. Table 3 combines in this manner the results for the first broods of tables 1 and 2. Of the 5174 flies of the first broods 608 or 11.8 per cent were crossovers. A similar balancing of the data of the second broods from tables 1 and 2 was made for comparison with the results of the first broods (table 4). TABLE 4 Comparison of firsts and seconds (purple and vestigial) with balanced viability } | [percent | Sp ag CROSSOVERS | TOTAL OF CROSS A BER CUNT ists 4,566 608) | Wmeeiae . MMeLiee | 2nds 4,984 431 5,415 | 8.0 | —3.8 | -32. Morale. 9,550 1039 10,589 | the fall of 3.8 units is 32 per cent of this original amount of crossing-over. This fall of a third in amount, in connection with the fall in thirteen out of fourteen cases, shows that a real variation in linkage has occurred. In this case the second broods gave a trifle over half of the 10,589 flies, showing that the egg-laying powers of the females had not diminished. The case upon which there are the fullest data (16,873 flies in back-cross experiments) is that of black and curved. The first experiment involves ‘coupling’ (table 5). Of the broods later than the firsts, four showed a decrease and two a smaller rise. In the totals the fall of 1.8 units is not large enough to be significant in itself. In culture 29 the female was carried through four broods and showed very little decrease in the number of offspring. Likewise the mother of 31 main- tained her output. It is interesting that the fourth brood of 29 showed more crossing-over than did the first. Perhaps the fall reaches a maximum in the second broods. This question 8) CALVIN B. BRIDGES P, black curved X wild. REF. 29 29’ 29'" g914 31 31’ Ba al Ists Others TOUR A 3 = 879 NON-CROSSOVERS Black P, black <* curved. REF. NON-CROSSOVERS Black | Curved | 144 | 150 £49 i193 134 11 159 2 157 | 148 125) 115 135 95 105 100 570 512 531 463 1101 975 Wild type curved 102 103 de 107 103 142 84 65 105 106 127 149 127 12 61 74 . 98 110 334 321 545 647 968 TABLE 5 B.C. F, wild type 2 X black curved io CROSSOVERS Black TOTAL Curved 34 40 27 24 36 42 29 30 27 44 44 32 34 55 18 33 38 31 95 139 192 192 287 331 TABLE 6 | PER CENT OF CROSS- | OVERS 26 22 24. 28 .« te bo on bo —_ lor) t “I CO ep bo A B.C. F, wild type 9 X black curved io CROSSOVERS Black "| Wild type | 55 45 31 26 53 4] 32 ota) 40 56 Biri 33 29 47 21 42 177 189 121 136 298 325 TOTAL PER CENT | OF CROSS- OVERS ~I oC LINKAGE VARIATION IN DROSOPHILA 7 is to be investigated. In the ‘repulsion’ experiment (table 6) each of the four females showed a marked decrease, and the totals show a decrease of nearly five units. The most convincing experiment is the following, in which three second chromosome loci (namely, black, purple, and curved) are run together in the same back-cross experiment. Such an experiment is much more satisfactory than three separ- ate experiments would be in studying linear arrangement. In spite of the labor of getting the triple recessive used, and the time of classifying the flies in a more compiex manner, there are many advantages. Since the same data furnish three linkage values, only a third as many individuals need ultimately be raised, or three times the amount of data may be obtained in the same time. The resulting values are more exactly and safely comparable, since they were produced under the same conditions. The order of gens is most strikingly shown by means of the smallness of the two contrary classes which are the result of double crossing-over. The amount of double crossing- over can be directly observed instead of being calculated from linkage values obtained in three separate experiments and perhaps not strictly comparable. A comparison of the observed amount of double crossing-over with the expected amount gives a measure of the amount of interference. Finally, if there is any disturbance of the linkage, it is important that the differ- ent values be derived from the same experiment, so that the disturbance can be shown to be localized or to affect the whole chromosome alike, as in the present case (table 7). — In the totals of the broods the change for each value is sig- nificantly a fall. The thirty-five cases of changed ratios are dis- tributed as shown in table 8. In this table the percentage of decrease are calculated from the brood totals of table 7. It is evident that the fall has affected each section of the chromosome by approximately the same amount. Of the twelve families of table 7, eight showed a fall and two a rise in both component values (black purple, and purple curved). It might therefore seem that there is some correlation, such that when black purple changes greatly purple curved 8 CALVIN B. BRIDGES would change greatly also. Calculation showed that the change for black purple in any particular female differs on the average from the change for purple curved in that same female by 3.5 units. But the change for black purple in any given female TABLE 7 P, black purple curved X wild. B.C. F, wild type @ X black purple curved 7S | y | NON-CROSS- SINGLE 7 OVERS CROSSOVERS DOUBLES PER CENT OF CROSSOVERS BETWEEN 7 | he | Bebe Cx |(BieriCy, SBsEr [Cv BI Pr | Cv sc o “ts y SS eee os Sea ia Ss 243 ye Sse S8i_ 6, 86 5/6 Fleiss 4 bee “ sae * Hae SS in Ge Oe O1Ms "ma leiae e a 5 m5 58 63 62 ee) i716 | ft | a Syal iD: 21.1 25.7 58’ | 80 95 3 18 914] =) =) 2g) 2.4 =7:9 15.1 —ehonizesmer es | 60 131 1 Wa 5 Al (34°| 4) | avaegpaleigelly | “= We2089 ae = ES a EOI 1 | 389| 4.9 19.1 22.9 62’ | 102 114 TES 1 e280!) eeeeesin 4a) 01800527) 34 ROE BRO Gon 147 men sI56 DD i PH ihe 9 | ee) Che 16.7 18.0 Gon 77 2) 04 Dimas 13 = = )\214/ 2.3 2.2) 13.1) (3:6) 15 4 ee | | | 68 89 “3. | Gt 24 -24)|) 3 ES Moagn720 22.3 26.6 68’ 80 so | 2 5 18. 9| 2 /: = /20tl 4:0 3:0: 13.9). =S8<4eele:o eee une 70 92 92 7) va 17 26| 1 peeaemoaeleraes 18.7 21.6 0’ 70 92 2 ea i pes — 8s =0S By SOOM. 7 72 TOMER I5Sy | TeunO 34 20). meee RS GOLDY 18.3 21.8 72! 69 mM | Bs 16 12} 1 “O) i189) 5.8 +00. 114 670 GOLO oes 74 103, 21, | 9 4 % 23) 1 1 | 287) 5.2 17.4 21.2 74! 79 97 9 5 180 129 fl || Spm CL SSG Ta 8.8 iO. =i 76 122 102 |12 9 23 20|/ 1 #2 | 296 8.1 17.2 23.3 76’ 75 97 4 3 gy ltl 1 | 207) 5.3 —2.8 13.8 —8.7 15.0 —8.3 | 78 140-148 Git 7 6) Uaouiy aD 358, 4.5 17s 20.1 797 OMICS 5 5 5 16.) Qeyeen ego 425 = a 5 6 eee bee S09 1133) elo) «| eu 8 26 24| 2 b=) | 3341 3.8 15.6 17.7 $07 |) 81 89 7.33 | = 1 |} 215] 5.1 41.8 16.3 4-0.7 20.2 -++3.5 sé | 76 nos] 1 14 18| 1 — | 289] 5.0 14.9 18.9 86’ 96 83 hen 10, S140 1 WW Bil Tl Spe 1.0 =s.9 88 103 ~—s«116 558 28) 19) vel 272, 4.0 16.5 18.4 gg’ | 109 111 iy & toe) il = 283 AO SYD 1A S50 OLD Bs.B | ists | 1476 1577 |96 74 339 330| 19 23 (3934) 5.4 1st 21.3 onds | 1028 1187 |55 41 200 166| 9 7 (2693) 4.2 —1.2 14.2 -—8.9 17.2 —4.1 Total...| 2504 2764 151 115 539 496! 28 30 |6627 LINKAGE VARIATION IN DROSOPHILA 9 TABLE 8 Cases from table ? of change in linkage 1B Rirkestt Sac a eh IW | BLACK PURPLE hipaa CURVED, BLACK CURVED | IDG CTE ASE tee ae te rt os eke a | 10 | 10 | | 2. lnCneasere i chica iene: | 4 | 2 | 2 | 8 5¢ | Percentage of decrease... .| 22a, 21.5 | 1953). | differs from the change for purple curved in any female by 3.8 units, which is practically the same amount. From this point, that there is a negligible correlative change for the values for black purple and purple curved, we obtain confirmation of the view that this change is a general one which affects different sections of the chromosome independently, but on the average to about the same extent. From the percentages of occurrence of the three mutants (table 9) it can be seen that differential viability has been al- most entirely eliminated. In the case of black, 98 flies hatched for every 100 expected—a very close approach to expectation. In the second broods the viability is somewhat poorer, but since the decreases are uniform no changes in linkage are to be expected from that source. TABLE 9 Viability coefficients (data of table 7) | BLACK PURPLE CURVED ink ee | 0.98 | 0.97 0.97 0.95 0.95 0.94 The results in table 7 are not balanced here by an equal amount of data for each of the converse experiments. To balance the data of an experiment involving three loci requires four sets of data, instead of two, as in the case of two loci. Fortunately, in the case of black and curved there are avail- able data for such a balancing as is given for purple and vestigial in table 3. Only flies in first broods can be used for this balanc- ing, as I shall make clear in the conclusion. In a paper by Bridges and Sturtevant (Biol. Bull. ’14), there appear records of 7419 such first brood flies in ‘coupling’ and 10 CALVIN B. BRIDGES ‘repulsion’ experiments. The data presented in this paper bring this total up to 11,353 flies and make the amounts for ‘coupling’ and ‘repulsion’ almost exactly equal, as shown in table 10. TABLE 10 Linkage of black and curved (1sts) with balanced viability Br le CROSSOVERS | TOTAL | pat = : rs Bae —_ =e \Wiirlle iting ofet" cetgc.s oroictea siceeeneae 2210 644 Biackoee. Sek Seco ton ek 2292 619 , CATRAW EC IRS hie tern cree Oe 2148 630. | Blackicurnvedsarcr ctr. see 2147 663 | Tiia Ae, See a 8797 2556 11,353 | 22:5 The last case on which I can now present data for the second chromosome is that of streak and morula. Streak is a domin- ant mutation which occupies a chromosome position far from black in the opposite direction (left) from the loci occupied by the other mutants so far treated. Morula occupies a locus likewise very far from black but in the opposite direction (to the right) so that a great section of the chromosome extending beyond the black curved section in both directions is tested by this experiment. Here also the single case so far tested showed a fall (table 11). TABLE 11 B.C. F, streak 9 X morula ox } j P, streak 2 X morula &. NON-CROSS- | OVERS CROSSOVERS | a | PER CENT OF REF. —|— —— | TOTAL | A ey . | Streak- Wild | | meisoenapics | Streak Morula | ocila eae 82 50 47 | 40 31 168 | 42.3 82" 50 31 26 DEI Ist 3871 =e In table 12, | have summarized by tables the cases for the second chromosome. Of the sixty tests, forty-nine showed a decrease from first to second broods. This decrease appeared uniformly in each experiment. A more satisfactory method of summarizing the data is that of table 13, wherein the data have been collected according to the linked pair of gens tested. The relative number of LINKAGE VARIATION IN DROSOPHILA AA TABLE 12 Cases of change of linkage in the second chromosome, from tables 1, 2, 5, 6, 7and 11 TABLES TOTAL 1 2 ) 6 7 bs! IDYXGRSDRES blo no. edb So vo ae Sart ee 7 6 4 4 27 1 49 IGAGREAEIAS OY 5 Ea 5 ORs 4 oe eee 1 0 2 0 8 ) 11 z ty | ARG 2 eee 8 6 6 Pao ml emis Ge TABLE 13 Change for each linked pair of gens PURPLE | BLACK BLACK | PURPLE STREAK o I VESTIGIAL CURVED PURPLE CURVED MORULA zor aDecreaseo” caer eh 13 18 a 10 1 49 NITCEEAS Chas eaites reir 1 4 4 2 0 1] Percentage of decrease O253 <7 222 AAS 10.1 cases of decrease and of increase for each pair is shown (table 13). The percentages of crossing-over calculated from the totals of the first broods and of the second broods for each pair of gens, have shown in every case a fall which is considerable in amount. Table 13 (last line) gives this decrease in amount calculated as a percentage fall from the value given by the first broods as a standard, as was done in tables4and 8. All of these methods show the same real change in the amount of crossing-over between second chromosome gens in the second brood as com- pared with the first brood from the same female. THE SEX CHROMOSOME DATA Although in the case of the first, or sex chromosome, the data are not as great in amount as for the second, the conclusion is quite certain that here there is no change in the amount of crossing-over with second broods (table 14). Between vermilion and fused (‘fused’ is at the extreme right end of the known plotted chromosome) there is a large amount of crossing-over, and the totals show that the value for the second broods is practically identical with that of the firsts. The five females tested showed a fall in three cases, balanced by 12 CALVIN B. BRIDGES a rise in the other two. More data on the same case is furnished by a triple experiment, which involves as the other locus, bar, a dominant mutant descr bed by Tice (Biol. Bull. 714). These data (table 15) add four cases of rise to the three given by table 14, and a corresponding four cases of fall to the two from the same source. The value for bar fused shows three eases of fall and five of rise. For the pair vermilion bar the cases stand four against four. The totals in every case show practically no change (table 16). The section from cherry to forked includes nearly all of the known sex-chromosome, and for this whole distance there is no change in the totals, and the females are balanced two against two (table 17). In this last case, that of cherry and sable, the slight change is a fall in the amount of crossing-over. TABLE 14 P, wild 29 X vermilion fused iS. Fi wild type 2 X Fi wildtype d' oh FEMALES | MALES ara | Noncerossovers Crossovers ey e eee A Ver OVERS sap |) Wildl). -\-<2 eal naViens Wild type | bles | en | Fused 52 G6y | 30 | 25-9) avewa) Reale 82 32.9 52! 176 || 64 | 59 | Faaeinaio || 166 25.9 | —7.0 53 G0") 22° | (20: SRROaal exe 57 26.3 53" eomerey) 27 |. 21 aaa te 69 30.5 | +4.2 | | 54 WSS) 38. | 3574 aie) "103 29.1 54’ 60))| 20 | 22) isaiese 59 28.8 | —0.3 57 Gi) 20.9| :22 ey alieal 60 30.0 ay WO} 54 | Ava eeea 19 |) 144 29.8 | —0.2 58 128 | 55 | 37 | 14 | 10 | 116 20.7 58/ 144 | 64 | 38 | 16 | 15) 188 23.3 | +2.6 ists (ey, 493" | 165 1185 eon | 54 ans 27.3 2nds 626 | 229 | 187 | 83 | 72) ||) S71 27.2 | —0.1 Potala eee 1059 | 394 | 326 | 143/126 | 989 | 27.2 LINKAGE VARIATION IN DROSOPHILA 13 The summary for the data on the sex chromosome is given in table 18, similar to table 13 for the second chromosome. The total of females tested shows 17 cases of decrease and exactly the same number of increase, so that we may safely conclude that there is no change here from first to second broods. The TABLE 15 P, bar 2@ X vermilion-fused 1. B.C. F; bar 9 X vermilion-fused io J E NON-CROSS- SINGLE peeve PER CENT OF CROSSOVERS BETWEEN OVERS CROSSOVERS | OVERS | 75 =. a Te | | —— = REF. |V_BrFu} V[BrFu V_BrjFu |_V_JBqFu | 5 kK ino] 3 he euita ee é : mo} mo) eer) = io} Eat bE zs |i, 2 & s8 |fe Z8/2/F 4 Se A Eeg A Sue (| | Syl fe SS eas: S| ie, Se eae SSChe 82 15M Ne LG5e| RGSS nS 7 1 = | 466) 25.9 3.4 28.9 RP 106 | GE 4 = OS OY SR eS Sat Sas 83 eS TGS) Bh eG 4 = 392] 23.0 2.6 25.5 Saul t00) o4) |) 28) 930) 4 4 =) |) |: 260)/oousweor7) Sule) 20) a e2pren—Onl 89 85 105] 28 | 24 2 = = (aXulios 2.9 22 89’ 7) Oil be Ve 2 Se hon ee] Op) Sy eS aT iS iy Oe +0.1 90 Ge BAR SO es — = | 234] 94.8 2 26.9 90! 33), 488 || 22 Led 1 — 1 |.413/32.7 +7.9 +1379) 36.3 9.4 Giese) eL07 |ad 3h 1 — — | 30] 93.5 0.7 24.2 Ol? |) OR GB Bik 1 — 2a es0lfoske) 0.3 see) 220d) 24188 F016 92 109 136) 41 24 4 2 — = | 316l90.6 1.9 92.5 92” || 100 105 | 29° 29) = 1 =. i e2doleooneiestety OnSee, 11) 022) Sue 02 cB | We GIO. 20 = 1 — — | 189) 91.4 0.6 22.0 OR? ||| GS MOAN eile Tees al 1 — one Moorg teh 009. f-048) 23.6) eIk6 94 SAle EvG6 | col neo peS 1 — — | 255] 95.9 3.5 29.4 94’ Gl 7) i), Baa 4 — — |185)99.7 —3.2 4.9 +1.4 27.6 —1.8 95 M709 || Be BB 3 — — |245)917 — 2.4 Se rt = OTR PR EVin SVT Flay a 2 — ailsvanentg: = 1aee siege = 97 81 96) 25 20 & 3 — — |230/19.6 — 3.5 = 6.0. 98 iy iD) 8) $3 “a 2 =. S Mpeeioasey ee = ORR eS ists | 1273 1383 | 4383 371 47 28 1 3537] 29.8 2.2 24.9 2nds | 635 677 | 208 188 20 8 5 jlzoilfepio 0.1 2:5 -+0.3 24.8 —0.1 seals Pad 4 ae id Ft ut Total...) 1908 2060] 641 559 67 46 1 . 6 5288) 22.9 2.3 24.9 | 8968=75% |1200=22.73% 113=21.4%| 7=0.13% | | 14 CALVIN B. BRIDGES TABLE 16 P, cherry 292 X forked oo. F, wild type 2 X Fi cherry vc. | FEMALES | MALES | | REF. ea Non-crossovers| Coron "Oo | Ge none: A = | . OVER gars hes Cherry Forked ee ea tars | 25 129 145 | 73 70 | 65 68 276 48 .2 25’ 167 148. | 74°) (820) \v6Griess: 310 49.7 | +1.5 36 | {96 88 «52, 7524) Saal 190 | 45.2 36’ ) “57 76 |. 41 32 | 24D 127 42.5 | —2.7 S4 i "76: 86 | 40) 34 38 26 13S ot) s4Gvoan 84’ 62 Tl ate © $89 2528, |e 116 45.7 | —0.6 85 ia SGN 48, 7 Avess | 215 | 43% 85’ 98-95 | 48 68 52 46 |: 209 |. 46:8 | -3al ists 415 405 | 208 234 | 179 198 | 819 | 46.0 | 2nds 384 390 | 187 216) | 1emeeig2 ) 762 | 45:8 | oie Total.......| 799 795 | 305 450 | 346 390 | 1581 45.9 | TABLE 17 P, cherry 22 X sable io. F, wild type 9 X F, cherry Jo. FEMALES MALES — PER CENT REF. ¥ Non-crossovers Crossovers TOTAL OF CROSS- A Cherry ys Cherry Wild | Silo OVERS ype |Cherry Sable | sable type ys) 131 101 63 52 38 = 48 201 42.7 55’ | 94 96 52 31 29>) 30)7 |) (142 | 41.6 —1.1 ; : a da ss 2 | Ae : ; = 225 197 | 115 83.) tn GGie ensn l « S43\; Viaes TABLE 18 Change for each linked pair of genes (1st chromosome) | | CHERRY | CHERRY VERMILION VERMILION | BAR TOTAI SABLE | WOOLY BAR FUSED | FUSED tx IDECrease:y. cee ston 1 2 4 7 3 17 Increases ne ener ee 0 WD) | 4 6 5 lg To tale se. cate: 1 4 | 8 hy pals 8 34 Percentage change..... =—2.6 —0.4 +0.4 | +0.8 | +13.6 LINKAGE VARIATION IN DROSOPHILA iL singe exceptional rise in the percentage in the case of bar fused is probably not significant, since the same females gave no rise but a fall in the case of other gens, and the distance involved is so small (2.3 units) that a very few flies make a great apparent difference. THE THIRD CHROMOSOME DATA For the third chromosome I am now able to report only the single case of pink and kidney (table 19). Here four of the five second broods showed rises, and only one a fall. The totals show a rise of 2.4 units, or of 17.1 per cent, on the basis of the first broods. This case in itself is too small for any con- clusion to be drawn. It is possible, however, that in the case of the third chromosome a rise from first to second broods may occur. TABLE 19 P, wild X pink kidney. B.C. F, wild type 9 X pink kidney oo NON-CROSSOVERS | CROSSOVERS | PER CENT. REF. \ eee a. ——————| fora | oF CROss- A | Mi Riines | Pink Kidney | ovens 17 fre agooL = 97 11 a ee 281.1) toes le [el VO8s) 84 27 adloaeniaenn 221 17.6 | +6.8 | | | 21 131 104 ishpas Tye ASS Qn 28.0 2117 30) fae 982 | 14.9 | 41.6 23 ie. OT 23 + 6 231 | lee 23’ 9 60 mil 164 | 15.2 | 42.6 25 85 92 2A ai DIS) | Wages 25! 107 90 14 29 33. | 15.4 | =3.4 172138 33° 20 863)! | 01436 7’ 121 105 30°) QUOT: | ants A Wee 3 8 Ists 608 522 QoS 1314 | » 14.0 2nds 528 456 i eal 1177 16.4 | +24 «LG 6 eae ce 1136 978 231 146 2491 15.3 16 CALVIN B. BRIDGES CONCLUSIONS Only data which have been obtained under like conditions can be used in dealing with any problem such as comparative link- age. If in the construction of a chromosome diagram the value used for A-B was that calculated from first brood data, and the value B—C was based on the total output of a female, the prediction of new values from the diagram would be in- accurate, unless the linkage remained unchanged throughout the life of the female. But if the diagram is constructed wholly from first brood values, predictions will be accurate. The practical point to be derived from this study is the breeding of only one brood from each female, especially in the second chro- mosome work. Any other condition as a standard could not be fulfilled with certainty for any large body of data. Linkage has been explained by Morgan on a chromosome basis, in accordance with the cytological evidence. It is assumed that gens occupy fixed positions, linearly arranged within the chromosome. In diploid groups each such linear series is represented by two homologous chromosomes, A and a, every locus in the one (A) corresponding to the same locus in its homologue (a). Before maturation homologous chromosomes become paired, side by side, and the members of each pair be- come twisted about each other. At some of the points of con- tact the two strands twist in two, as it were; moreover, the end of A fuses to the other end of a as they lie opposed. Any gens that were in strand A but on different sides of a chiasma point will emerge in different strands because of the crossing-over, and hence will be segregated to different gametes. It is obvious that the closer together in the strand any two given gens lie, the less is the chance that in any given maturation a chiasma will occur between them, the chiasmas being distributed ac- cording to chance. The basis of linkage is that two gens lie in the same chromosome so close together that in less than half the maturing germ cells a crossing-over takes place between them. There are two simple ways in which this scheme could be modified to give the change in linkage here described. We LINKAGE VARIATION IN DROSOPHILA 17 must suppose that normally the tightness of twist for any chro- mosome varies, within limits, in the different maturing cells. The length of the section between nodes—the internode—will hence vary correspondingly, but around a modal length normal for that chromosome. It is also probable that crossing-over does not take place at every node but only in a certain per- centage, specific for the chromosome. If the average length of the section between nodes, the inter- node, remains unchanged (that is, if twisting becomes neither looser nor tighter) such a decrease in the amount of crossing- over between given gens can be explained as failure to break and re-fuse, i.e., cross over, at as many of the nodal points as normally. If, on the other hand, the average length of the internode becomes greater (that is, if twisting becomes looser) such an effect as described would be produced while the percentage of breaks per node remains constant. A possible means of de- termining which of these views obtains here is offered by a study of the interference effects in the first and second broods. Interference stands in about the same relation to linkage as linkage does to free Mendelian assortment. In linkage there is a hindrance to the independent assortment of two pairs (A, @ and B, b) of allelomorphic gens. Such a case in sweet peas is that of round pollen versus long (pair A, a) and red flower versus purple (pair B, 6). In the phenomenon of interference there is a hindrance to the independent linkage of the members of two couples of linked genes, A—B and C-D. In any case of free Mendelian assortment one expects as great a percentage of A to be at the same time }, as of a to be at the same time B; thatrise is Abs: ab sab (as, e.g. mi Oesicod 2 orl : 1a de). In cases of linkage this relation is altered by a deficiency of the classes arising by crossing-over. Likewise in cases where two couples A—B and C—D in the same linkage group (chromosome) are crossed together, the number of individuals in the double crossover class may be considerably smaller than expectation according to the proportion above. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 1 18 CALVIN B. BRIDGES The development of the idea of interference is an illustration of the advantages of the chromosome hypothesis. The existence of this phenomenon was originally deduced by Muller and Sturtevant from a consideration of linkage as a chromosome process. Linkage had been explained as the result of two loci in the same chromosome being so elose together that a chiasma occurs between them infrequently. If the chiasma must occur at a node, then, since the internode has considerable length, there must be on either side of the chiasma, space which is free from crossing-over. A couple of gens C—D might le at such a dis- tance from A—B that when a chiasma occurred between A and B, the length of the internode would cause the next chiasma to occur most often to one side of C—D rather than between C and D. In any one maturing egg a chiasma between the members of one such couple would tend to prevent one between the members of the other couple; consequently, few gametes would be formed which would be the result of both occurring simultaneously, that is, double-crossing-over would be interfered with. While the phenomenon of interference is thus a corollary of the chromo- some hypothesis, it is almost unexplainable upon any other view of linkage. Since the accuracy of the calculation of the amount of inter- ference depends upon the accuracy of the smallest class—the double crossovers—in practice we derive the index of inter- ference from a comparison of the observed percentage of double crossovers with the percentage which would be expected if there were no interference. The percentage of double cross- overs expected without interference is the product of the total percentage of crossing-over between A and B by the like value for C and D. That is, if there is 5 per cent of crossing-over between A—B, and 10 per cent between C—D, then 5 per cent of 10 per cent would give the percentage of cases in which both occur. The case is exactly similar in treatment to the 9:3 :3:1 rat'o of two freely assorting pairs of gens, where 25 per cent of all cases are a and 25 per cent are b, and 25 per cent of 25 per cent gives the percentage occurrence of ab. In order to obtain LINKAGE VARIATION IN DROSOPHILA 19 a convenient index, the converse of interference namely coinci- dence, is calculated as the percentage which the observed double crossover class is of the expected value. If, as described, interference is a function of the chromosome twist, then from observing how the change in linkage here re- ported affects interference, we can deduce the method by which the linkage has been altered—whether by a decrease in the percentage of breaks per node or by a decrease in the number of nodes, that is, by a looser twist. If the twist remains normal and the decrease is due to a de- crease from the normal percentage of breaks per node, then each linkage value will be reduced proportionally. Whenever a crossover does occur it occurs in the same position in which it would normally have occurred, so that within any given section of chromosome as great a percentage of crossovers would be doubles as‘ in the normal condition. The effect is to make the new condition a replica of the old, except that every cross- over value is reduced in a common ratio. In this sort of change in the mechanism, the interference would remain unaltered, for in the ratio of the expected percentage of double crossovers to the observed percentage, both terms decrease proportionally, so that the value is unchanged. If, however, the average looseness of the twist is increased, a totally different result will be produced. Here the nodes become actually further apart, so that whenever crossovers occur they are no longer in the same average position as formerly but are wider spaced. It then requires a longer section of the chromo- some in order that double crossing-over be possible. This means that for any definite section the number of double cross- overs becomes less, that is, interference rises. On this second hypothesis, then, the closer spaced in relation to the length of the internode the members of the pairs A—B and C—D are, the higher is the interference. The new condition should show the same interference as would be shown under the old condition of a shorter internode by gens correspondingly more closely spaced. By considering the amount of apparent displacement (the decrease in the percentage of crossing-over) 20 CALVIN B. BRIDGES we might calculate how much interference should rise ‘if that change be due to an equivalent looser twist. Thus an experi- mental method is provided for the analysis of the mode of twist- ing and the distribution of chiasmas, not only under the changed condition, but also under the normal condition as compared to the changed condition. It is, however, no small task to secure data for such a study, and in any other material than Drosophila the problem would be wellnigh hopeless of solution. At present the unfavorable case of black purple curved furnishes only enough data to give a suggestion as to the mode followed. The data for the calculation of interference in the case of black purple curved are given in table 7 and for the first broods may be summarized as follows: Beer Cy, 1B) if ise 5 (hy Beer Cr iy Be. | Cw 3,053 170 669 : 42 Percentage 77.61 4.32 17.4 1.07 Here the total amount of crossing-over between black and purple is 4.32 + 1.07 per cent, and between purple and curved is 17.4 + 1.07 per cent. The expected percentage of double crossing-over is therefore 5.39 per cent of 18.07 per cent, which is 0.97 per cent. The observed percentage (1.07) of double crossovers was somewhat larger than this. Although the difference is so small that it may be due to chance fluctuation, yet it will be instructive to consider its meaning on the assump- tion that it is not due to chance. The actual increase of 0.1 per cent is a relative increase of 11 per cent over the expected 0.97 per cent (percentage of coincidence 111). But inter- ference which increases the percentage of doubles is a reversal of the ordinary type and the explanation of this ‘negative’ interference is as follows: . The preceding considerations have applied to the case in which the length of the average internode is such that when one node lies between A and B, the next node will most often lie beyond C—D. This relationship between the relative position of the gens and the length of the internode is such that a chiasma between one couple will tend to prevent one between the other LINKAGE VARIATION IN DROSOPHILA 21 couple, so that positive interference will result. Let us now consider two arbitrary points, M and N, so chosen that the dis- tance between them is equal to the length of the average inter- node. Whenever a node chances to occur near M, then the next node will most often occur near N. That is, the chances of a crossover at either point are very greatly increased by the occurrence of one at the other point. In this case, instead of getting less than the expected percentage of double crossovers, we would expect, on the internode hypothesis, to get more than the expected percentage. Interference for two such points may be termed ‘negative.’ The second broods of the black purple curved cross give the following data: B Pr Cv By [err Cy, BSPré rcv, oe B.. | oP Cv 2.215 96 366 16 Percentage 82.24 3.57 13.6 .594 From this we find that the interference is zero (percentage of coincidence 100). There has been an 11 per cent rise in inter- ference concomitant with the decrease in crossing-over. This would suggest that the decrease in linkage here studied has been due to an increase in the length of the average internode rather than to a decrease in the percentage of chiasmas per node. Unfortunately, the number of double crossovers obtained is not great enough to establish this change in interference as significant rather than due to chance fluctuation—1.e., the vari- ation is within the limits of probable error. THE RELATIVE EFFICIENCY OF VARIOUS PARTS OF THE SPECTRUM FOR THE HELIOTROPIC REACTIONS OF ANIMALS AND PLANTS JACQUES LOEB AND HARDOLPH WASTENEYS From the Rockefeller Institute for Medical Research, New York I. INTRODUCTION While the older authors had treated the motile reactions of animals to light as an indication of their love for or antipathy to this kind of energy, one of us in 1888 pointed out that we are dealing in these cases with phenomena of orientation comparable to the orientation of plants to light.!. In order to indicate the identity of the mechanism in both cases he proposed the same term for both, namely, heliotropism (or phototropism). Loeb stated in his first full pamphlet (1889)? that (if one source of light be given) the animals orient themselves so that their plane of symmetry falls into the direction of the rays of light, ‘‘ whereby the symmetrical points of the surface of the body are struck by the light at the same angle.’’ In 1897 the same writer expressed the idea that the action of light which caused the heliotropic reactions was chemical.’ Since it is reasonable to assume that symmetrical elements of the surface of the body are not only morphologically but also chemically alike, we must suppose that if the symmetrical elements of the surface of the animal are struck by the rays of light at the same angle, the velocity of the photochemical reactions in symmetrical elements of the surface (e.g., the eyes or skin) are the same, since the intensity of the illumination of a surface element varies with the cosine 1 Loeb, J. Sitzungsb. d. Wiirzburger physik.-med. Gesellsch., 1888. 2 Der Heliotropismus der Tiere und seine Ubereinstimmung mit dem Helio- tropismus der Pflanzen. Wiirzburg, 1889. 3 Loeb, J. Zur Theorie der physiologischen Licht- und Schwerkraftwirkungen, Pfliiger’s Archiv, Bd. 66, p. 439, 1897. 23 24 JACQUES LOEB AND HARDOLPH WASTENEYS of the angle of incidence. The influence of the light upon the tension and action of symmetrical muscles must in such a case be identical. The light, if it remains constant, will therefore not cause the animal to alter the direction of its motions. ‘If, how- ever, the light strikes symmetrical elements at a different angle the velocity of photochemical reaction is not the same in the two symmetrical elements and the symmetrical muscles on both sides of the animal will receive unequa impulses (by reflex) from this source. This will lead to an automatic turning of the animal until its plane of symmetry again falls into the direction of the rays of light. If these premises were true, it followed that the heliotropic reactions of animals should obey the law of Bunsen and Roscoe which says that (within certain limits) the photochemical effect of light is equal to the product of the intensity into the duration of illumination; and one of us predicted that this law would probably be found to hold for animal heliotropism.4 This prediction proved true for the heliotropic curvatures of Euden- drium, as the experiments of Loeb and Ewald’ showed. Ewald could also show that Talbot’s law holds for the orientation of the eye of Daphnia by light,* and Talbot’s law is an expression of the fact that the physiological effect of light is equal to the product of intensity and duration of illumination. The same aw holds for the heliotropic reactions of plants, as Blaauw and Fréschl had shown.7, It also holds for the human eye.’ In all these cases it should be remembered that the law of Bunsen and Roscoe is a threshold law, inasmuch as it holds only within certain limits. With the reduction of both groups of heliotropic reactions, those of animals as well as of plants, to the same law, namely, that of Bunsen and Roscoe, it is idle to consider further the idea that animals are led to the light because they are ‘‘fond”’ Loeb, J. The mechanistic conception of life. Chicago, 1912. * Zentralbl. f. Physiol., Bd. 27, p. 1165, 1914. 6 Science, N.S., vol. 38, p. 236, 1913. 7 Blaauw, Rec. des Travaux Botaniques Néerlandais, Bd. 5, p. 209, 1909; Fréschl, Sitzungsb. d. Akad. in Wien, 1908. 5 Charpentier, Arch. d’Ophthalmol., tom. 10, 1890. SPECTRUM, HELIOTROPIC REACTIONS 25 of it, or that they are turned away from it because they hate it; or that the reactions are the result of ‘trial and. error.”’ We may, therefore, conclude that the heliotropic reactions of animals and plants are due to photochemical reactions and that the turning of the animal to (or from) the source of light is brought about automatically if the velocity of photochemical reaction is no longer the same in symmetrical areas of the photosensitive surface. This automatic turning results when the mass of photochemical react:on products on symmetrical points of the surface of the anima (eyes or skin) exceeds a certain value; and the variations of this value determine the relative sensitive- ness of different heliotropic animals. Since this has been stated more fully in former publications of Loeb we may refer the reader to these publications. ° , If the basis of heliotropic reactions is a photochemical process, it follows that heliotropic animals must possess a photosensitive substance, and the question arises: Is this substance identical in all heliotropic organisms or do the photochemical substances differ in different heliotropic organisms? Especially does this question become of interest in respect to the question whether there is a specific difference between these substances in animals and plants. ; The method to decide this question consists in comparing the relative heliotropic efficiency of different wave lengths in different organisms. If we find that the optimal heliotropic effects occur for one form of organisms in one kind of wave lengths, for another in a widely different wave length of the same spectrum, we may conclude that the photochemical substances in the two cases are different, if deduction be made for the possible screen effect of secondary substances contained in the sensitive organ. The older experiments of the botanists were mostly made with colored screens, which yielded the result that behind red screens only weak or no heliotropic reactions of plants occur, while behind blue sereens they occur as well as in mixed daylight. Loeb was able to show in his earlier experiments that the same holds 2The mechanistic conception of hfe. Chicago; 1912; article on Tropisms in Winterstein’s Handbuch der vergleichenden Physiologie, Bd. 4, p. 451, 1912. iw) for) JACQUES LOEB AND HARDOLPH WASTENEYS TABLE 1 DURATION OF LOCATION OF ILLUMINATION, THRESHOLD IN THE IN SECONDS SPECTRUM, IN MICRA 6300 534 1200 510 120 499 15 491 5 487 4 478 3 fee, 4 466 6 448 good for the heliotropic reactions of animals. But these experi- ments are not adequate to decide the question of perfect identity of the photochemical substances in all cases. For this purpose experiments with spectral colors are required. The most re- liable experiments made on plants are apparently those of Blaauw on the spore bearers of Phycomyces and the seedlings of Avena. Blaauw proceeded in the following way: He exposed a row of seedlings of Avena to a carbon are spectrum for a certain time. The seedlings were than placed in the dark and after the proper time it was.ascertained which part of the spectrum had induced heliotropic curvatures. By varying the duration of time of exposure to the spectrum it was found that with a minimal time of exposure only certain blue rays, namely, those of a wave length of 478 wu, caused heliotropic bending, while with longer exposure longer waves also became efficient. In this way the minimum duration of exposure for various parts of the spec- trum was ascertained. Table 1 gives his result. The red and yellow parts of the spectrum were ineffective for the intensity and time limits used and the optimum of efficiency was in the blue, in the region between 466 and 478 uu. A shorter series of experiments was made on the fruit bearers of Phycomyces, with the following results: 44 to 47 per cent of the Phycomyces showed heliotropic curvatures after 192 seconds of illumination at 615 uy after 192 seconds of illumination at 550 py after 16 seconds of illumination at 495 pu after 32 seconds of illumination at 450 uu after 64 seconds of illumination at 420 pu SPECTRUM, HELIOTROPIC REACTIONS af The number of experiments was limited but they indicate an optimum between 495 and 450uy, in this respect agreeing with the results on Avena. We were anxious to know whether for the heliotropic reactions of sessile animals the optimum is situated in the same region of the spectrum. The number of sessile animals which are sensitive to light is rather limited and we had to make use of the hydroid Eudendrium, which also served in the experiments of Loeb and Ewald. II. THE HELIOTROPIC REACTIONS OF EUDENDRIUM The newly formed polyps are positively heliotropic to light and they react by bending towards the light. The bending occurs in the region near the stem; the method of procedure was as follows: Immediately after the colonies were brought into the labora- tory good stems with from 4 to 8 or more polyps were selected. The polyps were cut off and the stems put into glass troughs filled with sea-water, where they were held in position by being fixed in little holes of a layer of paraffin, on the bottom of the trough. The troughs had plain parallel walls. The stems were exposed during the first day to ordinary light—since light is necessary for the regeneration of polyps'9—and were then put into the dark-room. In the dark-room the polyps developed during the next day. These newly formed polyps are very sensitive to light and were used for the experiment. The trough was then exposed to a carbon are spectrum, the visible portion of which was about 20 cm. wide. The spectrum was in a dark- room and all precautions were taken to guard against any reflected or other light from reaching the polyps. The stems were in a row and each one was exposed to a different part of the spectrum. The position of each individual poylp was marked in a diagram at the beginning of the experiment and the polyps were exposed to the spectrum for times varying from five min- utes to five hours. Then the polyps were put into the dark again 10 Loeb, J. Einfluss des Lichtes auf die Organbildung bei Tieren. Pfliiger’s Archiv, Bd. 63, p. 273, 1896. 28 JACQUES LOEB AND HARDOLPH WASTENEYS and after the proper time (two hours or more) the number of the polyps which had bent to the light in various parts of the spectrum was ascertained. As the reader will notice, the bending of the polyps took place after they had been put back into the dark. This corresponds with the method followed by Blaauw in his experiments on plants and of Loeb and Ewald in their experi- ments on the applicability of the law of Bunsen and Roscoe to heliotropic reactions. We have also made experiments in which the stems were ex- posed long enough to the spectrum to enable them to form their polyps while still exposed to the light. The determination of the wave length to which each stem was exposed was rendered possible through the use of the absorption bands of a solution of didymium nitrate. With the aid of these bands, we could define accurately the position of each stem and polyp in the spectrum. We are indebted to Dr. E. Butterfield for the exact location of these bands in the spectrum. In noting the result the reader must keep the following facts in mind: When a short exposure to light influences the orienta- tion of the polyps of a stem, this will show itself in the fact that the majority of the polyps of that stem will bend straight to the light. If there is no effect of light, the polyps will grow in any direction but it may happen according to the laws of probability that one or the other may bend to the light. In order to be sure that the light influences the direction in which the polyps bend we must require that so great a percentage bend toward the light that chance may be excluded, before we draw the conclusion that we are dealing with a heliotropic effect. We considered it a positive result when 50 per cent or more of the polyps bent to the light. That they should all bend to the light cannot well be expected, especially in cases of short dur- ation of exposure. The new polyps are extremely delicate and they are not all healthy or strong. Moreover, certain polyps will be partly screened from the light by the stems. Blaauw, as well as Loeb and Ewald, had to use the bending of 50 per cent of the specimens as a criterion of a positive result. The fact that each stem has only a limited number of polyps creates SPECTRUM, HELIOTROPIC REACTIONS 29 an additional difficulty, in this way: that if by chance one polyp is directed to the light and there are only five polyps on the stem it may appear as if 20 per cent of the polyps had reacted positively, while in reality the stem was not influenced by light at all. To avoid misinterpretations of this kind from influencing the interpretation of results we always give the number of polyps in a stem in the following tables. In indicating the wave length it should always be remembered that the region given is usually the center of a small zone which contained the stem; since the latter is not straight and since the polyps are irregular in position it is not possible to indicate the position by one line in the spectrum. We will now enumerate some experiments. In the first verti- cal column is given the wave length, in Angstrém units, to which the stems were exposed (indicating in the next column the color of the region). In the third column we give the fraction of the number of polyps bending to the source of light over the total In the fourth column we give the percentage number of polyps. of the polyps bending to the light. with different material on different days, unless the contrary is stated (table 2). Each experiment was made From this experiment we may deduce, first, that for this dur- ation of exposure (five minutes) the rays to 5700 A.u. (i.e., the TABLE 2 Experiment 1: Exposure of Eudendrium polyps for five minutes to the spectrum WAVE LENGTH IN ANGSTROM UNITS COLOR OF THE SPECTRAL REGION FRACTION OF POLYPS BENT TO THE LIGHT About 6500 About 6000 About 5700 About 5300-5345 About 5100 * About 4900 About 4735 About 4690 4600 4400 orange-red yellow yellow yellowish-green green blue blue blue blue indigo 1/29 0/4 0/13 5/15 3/12 11/32 30/49 4/21 5/22 5/52 PERCENTAGE OF POLYPS BENT TO THE LIGHT 30 JACQUES LOEB AND HARDOLPH WASTENEYS orange and yellow rays) are absolutely ineffective; that the rays for 5300 to 4900 (i.e., yellowish-green, green, and greenish- blue) are only slightly effective; while the rays of the wave length of 4735 A.u., (i.e., blue) constitute the optimal portion; and that in the indigo the efficiency diminishes again. This result is al- most identical with the one obtained by Blaauw with the seed- lings of oats (table 3). TABLE 3 Experiment 2: Exposure, five minutes WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT About 5760 yellow 0/2 0 5711 green 1/2 (50)? 5100 bluish-green 4/6 66 4800 blue 9/22 4] 4735 blue 3/3 100 4700 blue yall 27 4676 blue 8/10 80 4500 indigo 3/3 100 4400 indigo 2/2 100 ultraviolet 0/11 0 This result was less striking than the previous one through the combination of two circumstances: (1) The material was more sensitive than that used in the previous experiment, and TABLE 4 Experiment 3: Duration of exposure, five minutes WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT 6200 orange 0/3 0 5600 yellow 0/5 0 5300 yellowish-green | 1/10 (10) 5000 bluish-green 1/10 (10) 4850 blue 0/6 0 A735 blue 8/13 62 4700 blue 1/7 14 4450 indigo 0/3 0 4432 indigo 0/2 0 4000 violet 0/6 0 3850 violet 1/2 ? 3600 ultraviolet 0/1 0 SPECTRUM, HELIOTROPIC REACTIONS ol (2) the number of polyps was so small that the error was greater and the results not so uniform. Positive results were obtained in the region between A.u. 4400 and 4735, that is, in the blue and indigo and also possibly in the bluish-green; yellow was ineffective, as before (table 4). The most effective region of the spectrum in this experiment is again the region about 4735 in the blue, the same which has proved the most effective in the two previous experiments. In all the experiments red, orange and yellow, and extreme indigo and violet, were ineffective. The next two experiments (4 and 5, tables 5 and 6) were made TABLE 5 Experiment 4: Exposure to light, four minutes WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT About 5600 yellowish-green 0/5 0 5400 yellowish-green 6/23 alee) 5000 bluish-green 4/22 18 4800 blue 4/13 31 4735 blue 18/30 60 4700 blue HAZ ae 4670 blue 9/21 43 TABLE 6 Experiment 5: Exposure to light, three minutes WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT 5760 yellow 3/11 27 5600 yellow 1/13 8 5200 green 2/16 13 4900 blue 2/4 50 4800 blue 4/10 40 4735 blue 3/3 100 4700 blue 7/20 35 4676 blue 0/1 0 4500 blue 0/1 0 4432 indigo 3/12 25 4100 violet 0/2 0 4000 violet 1/7 14 3800 violet 0/2 0 o2 JACQUES LOEB AND HARDOLPH WASTENEYS with shorter exposure of the polyps to the spectrum, namely, . four and three minutes. The exposure of three minutes is the minimum from which a result can be obtained, and the results of table 6 are difficult to account for. The time of exposure is so short that slight differences in the sensitiveness of various stems make themselves felt. Both experiments agree in their result with the previous ones, namely, that the region around 4735 A.u. (in the blue) is the most efficient. It was to be expected that in a longer exposure polyps would bend to the light in both indigo-blue and in green but not in yellow and red. Table 7 gives the result of an experiment with an exposure of fifteen minutes. TABLE 7 Experiment 6: Exposure to light, fifteen minutes WAVE LENGTH IN COLOR OF THE SPECTRAL FRACTION OF POLYPS PERCENTAGE OF POLYPS ANGSTROM UNITS | REGION BENT TO THE LIGHT BENT TO THE LIGHT 3700-4100 extreme violet 14/30 45 4100-4900 violet indigo and blue 72/95 76 4900-5400 green and bluish-green 14/37 38 5400-6700 orange-yellow toyellow- ish-green 0/? 0 Unfortunately, no record of the total number of polyps formed in the yellow and red was preserved; the number, however, was large. . The experiment confirms that the blue and indigo are the most efficient rays while the green are markedly less efficient. The yellow and red rays are inefficient. The longer exposure brings out the heliotropic effects in the extreme violet which do not show with shorter exposure. In the following experiments an attempt was made to ascer- tain the influence of longer exposure. The first question was whether by making the duration of exposure considerably longer we should be able to induce heliotropic curvatures in the yellowish-green, yellow and red. Second, we wished to find out whether solarization effects might be observed in the case of too long an exposure. SPECTRUM, HELIOTROPIC REACTIONS 33 The method was as follows: Different stems of Eudendrium with young polyps (prepared in the way described above) were put successively into the same limited part of the spectrum. Each stem was exposed a different length of time. The purpose was to find out how much time was required to cause the maxi- mum number of polyps to bend toward the light (table 8). TABLE 8 Experiment 7: Eudendrium exposed to light of 4700 A.u. (blue) DURATION OF FRACTION OF PERCENTAGE OF EXPOSURE IN POLYPS BENT TO POLYPS BENT TO MINUTES THE LIGHT THE LIGHT 5 4/25 16 10 8/11 73 20 8/29 28 40 1/3 33 80 18/23 78 160 24/24 100 The experiments show that an exposure of more than ten minutes (for the light intensity used in our experiments) did not essentially increase the percentage of polyps bending to the light. TABLE 9 Eudendrium exposed to light five and a half hours WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT 4800 blue 7/16 44 4950 blue 1/6 16 5300-5500 green 3/13 24 5720 yellow 0/21 0 6000-6550 orange and red 0/32 0 We made with the same material an experiment in which the polyps were exposed for five and a half hours to the spectrum from blue to red with the result shown in table 9. The experiment shows again, first, that even in five and a half hours the rays from yellow to red are without any heliotrepic effect. Second, that the efficiency of rays with wave length of THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 1 34 JACQUES LOEB AND HARDOLPH WASTENEYS 4800 and above is less than that of wave length 4700. It is possible but not certain that there are solarization effects. In the next series of experiments a region in the neighborhood of 5600 A.u. in the yellow towards the yellowish-green was selected for the stems (table 10). TABLE 10 Experiment 8: Eudendrium ex- posed various lengths of time to the same wave length (about §600 A.u.). DURATION OF EX- POLYPS BENT TO POSURE IN MINUTES THE LIGHT 10 20 40 80 160 Seo) &) ~e) In order to make sure that this negative result was not the fault of the material experiments with the same lot of material were carried out in the blue-violet part of the spectrum (table aay TABLE 11 Eudendrium exposed for one hundred and fifty minutes WAVE LENGTH IN ANGSTROM COLOR OF THE FRACTION OF POLYPS PERCENTAGE OF POLYPS UNITS SPECTRAL REGION BENT TO THE LIGHT BENT TO THE LIGHT 4220-4500 violet and indigo 21/24 87 4500-4660 indigo-blue 9/11 82 4660-4710 blue 7/10 70 710-4750 blue 15/15 100 4750-4850 blue 14/16 85 4850-5000 blue-green 7/8 87 This experiment again shows strikingly that the blue and vio- let part of the spectrum is the effective one. The region between 4700 and 4750 A.u. is again the most efficient. In a third experiment of this series the stems were exposed to 2 wave length of about 4900 A.u. (blue towards the green) for various periods of time with the result shown in table 12. SPECTRUM, HELIOTROPIC REACTIONS 30 TABLE 12 Experiment 9: Eudendrium exposed to light of 4820 A.u. (blue) PERCENTAGE OF POLYPS BENT TO THE LIGHT DURATION OF EX- |FRACTION OF POLYPS POSURE IN MINUTES| BENT TO THE LIGHT 0 2/9 22 10 5/11 45 20 8/12 75 40 7/1 64 80 11/18 73 160 5/18 28 With the same material an experiment with long exposure (five and a half hours) in the region of shorter wave lengths, 3750 to 4700 A.u. (namely, from the blue to extreme violet) was carried out. In all, 52 hydroids out of 66 (i.e., 79 per cent) were bent forward. There was not much difference in the vari- ous regions, probably due to the long exposure. III. CONCLUSION. AND SUMMARY OF RESULTS ® These experiments have shown that the most efficient region in the spectrum for the production of heliotropic curvatures in the hydroid Eudendrium is situated in the blue at-’ ~ 4735 A.u. This region coincides approximately with the one found by Blaauw for the seedlings of oats namely \ ~ 4780 Ave The regions in the red, orange and yellow are practically with- out effect in both Eudendrium and Avena. The heliotropism of the sessile animal Eudendrium and that of the sessile plant Avena are therefore identical even as regards the most efficient wave length. We expect to discuss the effects of different wave lengths upon the heliotropic reactions of motile animals and plants in another paper. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. NO. 259. THE ORIENTATION OF AMPHIOXUS DURING LOCOMOTION! LESLIE B. AREY Observers have differed regarding the question as to which end of Amphioxus is in advance during swimming. Rice (’80, p. 8) seems to have been the first to record ohservations on this subject: These movements were executed sometimes upon the back, sometimes upon the abdomen in the position of ordinary fishes, it seemed to make very little difference which side was uppermost, but I have never seen them move backwards or tail-end foremost. After circumnavigating the vessel once or twice gradually moving slower and slower, they would stop and sink down upon the sand at the bottom. In a more general statement Steiner (’86, p. 497) came to the same conclusion as Rice, ‘‘sie stellen sich so auf, dass ihre Breit- seite in die verticale Ebene fallt und rasch entfliehen sie (from the stimulus) mit grosser Geschwindigkeit, das Kopfende voran, indem der K6rper schlingelnde Bewegungen macht, an denen der Kopf nachweisbar theilnimmt.”’ Two years later (’88, p. 41) he expressed the same opinion in almost identical language. Parker (08, p. 441) took the opposite view: The locomotion of amphioxus is a rapid, curiously irregular wriggle, often accompanied with somersault-like movements which make it impossible to be sure at any moment whether the animal is swimming backward or forward. The results of momentary stimulation, however, show very conclusively that amphioxus can swim both backward and forward, and that the direction of swimming at the beginning of any course is dependent upon the part of the animal’s body that was stimu- lated. But how long amphioxus keeps to one form of movement I was unable to discover. The fact that it usually buries itself in the sand tail first leads me to believe that, though it can swim forward, as maintained by Rice and by Steiner, it usually swims backward. 1 Contributions from the Bermuda Biological Station for Research. No. 36. 37 38 LESLIE B. AREY Parker and Haswell (710) in their ‘‘Textbook of Zoology”’ refer to the more or less upright position which Amphioxus as- sumes after burrowing in sand and then add the following astonishing and somewhat ambiguous statement (vol. 2, p. 46): “‘Tt also swims in the vertical position, - - - - - - .”’ However this assertion may be interpreted, it certainly is contrary to fact as far as the West Indian Amphioxus is concerned, and from the description of other writers (see also the frontispiece in Willey’s 94 book), the same criticism undoubtedly applies to the closely allied European species as well. Anyone may easily convince himself that the foregoing quotation either presents a gross error or is highly misleading (according to the alternate possi- bilities of interpretation), if he will observe for a short time the locomotor responses, regardless of the antero-posterior orienta- tion, exhibited by Amphioxus under various natural or experi- mental conditions. While working recently at the Bermuda Biological Station opportunity was afforded me for making observations on the swimming habits of the West Indian lancelet, Branchiostoma caribbaeum Sundevall, a species very similar to the common European Amphioxus. This animal is found in abundance in the coarse coral and shell sand of Flatts Inlet, which connects the water of Harrington Sound with the outside ocean. Ordinary mechanical stimulation, as by a finely drawn glass rod, gave too active a response for observation and accordingly a milder stimulus was sought. This was found in a weak stream of sea-water forced from a rather large canula; if the jet was weak and was directed vertically through 10 cm. of air and 10 cm. of water, the entire body of the animal became subjected to a gentle stimulus formed by the wave front. When thus stimu- lated the locomotor response tends to be less energetic and the disadvantages of local or directional stimulation are obviated, while it has a further advantage over other mild stimuli such as jarring the containing-dish, inasmuch as individual animals may be singled out for experiment and watched from the begin- ning of their course. ORIENTATION OF AMPHIOXUS 39 About thirty individuals were placed in three glass Jars con- taining a layer of ‘amphioxus sand’ and were stimulated by the method just described. All responses in which the orientation was doubtful and those responses which included wild dashes or ex- cessive somersaulting were disregarded; of the responses recorded, some were observed during the whole course and others were judged chiefly by the orientation at the beginning and partic- ularly at the slowed-down finish of the swimming reaction. A tabulation of observations obtained in this manner is as follows: shotalinumber of Fesponses...........ct- oes eon Oe ciel eiteeiccse 50 IATIEELION CHOC AGC VATICE A: < <0 sactote ce = ORME oe ol oeiete detec erasers 41 ROstenlomendunyagwance cc. -.... .satance cnet aaa 9 In a considerable number of animals the first movement was backward, but the direction was quickly reversed bringing the anterior end in advance; this condition will be discussed later in connection with other experiments. When, however, Amphioxus are free to move in unlimited space, the somersaulting and the quick reversals of direction make accurate observations as to what is occurring during the response very difficult. To obviate this difficulty a large por- celain pan was partly filled with water, giving depths which varied from 0.5 em. at the edge to 1.25 em. at the center. The movements of Amphioxus under these conditions were as follows: When a stimulus (a bristle or finely drawn glass rod) was applied to the anterior end, the animal responded with a backward spring; if locomotion was now continued, this direction was not main- tained for more than a few centimeters, for a quick reversal occurred, exceedingly difficult to follow, but which seemed to include a doubling end for end succeeded by a partial rotation of the long axis of the animal about its middle point; as a result the Amphioxus swam away with its anterior end in advance and usually at an acute angle with the direction toward which it was heading before stimulation occurred. During subsequent swimming these reversals occasionally occurred and as a result the posterior end of the animal might be in advance 40 LESLIE B. AREY for a short time; in this event, however, another reversal soon restored the former orientation and the normal direction of continued movement was plainly with the anterior end in front. When the posterior end was stimulated the animal sprang forward and if it continued to swim, it proceeded head fore- most, although subsequent reversals usually occurred from time to time. One lancelet of this set was especially instructive; after stimu- lation and the usual energetic response, it would continue swim- ming at a rate of about 1 cm. per second for a considerable distance without reversal; in several instances it more than circumnavigated the dish, a distance of over a meter, yet the anterior end was always carried in advance. Several other animals showed the same behavior but in a less degree; in general, after several responses the reaction was as long but less vigorous than that of a rested animal, and hence was easier to observe. A circular trough 30 cm. in diameter, 0.5 em. wide at the bottom and 1.25 em. wide at the top was constructed by placing a porcelain pan, bottom up, inside a slightly larger pan; the purpose of this arrangement was to give the animal free swimming room but to limit its locomotion to one direction. The results were in agreement with those described previously; it was im- possible to make an Amphioxus swim backward for more than a few centimeters before somersaulting and forward locomotion occurred. A final method, which gave more conclusive evidence concern- ing the orientation of the animal during normal locomotion in unlimited space, consisted in treating one end of the lancelet’s body with an intra vitam stain, whereby through direct obser- vation one could be certain of the animal’s orientation even during the wild dashes that often occur. Objections may be raised to the artificial conditions of the experiments described above, which were devised for limiting the animal’s move- ments; thus it is entirely possible (although I do not believe it to be actually the case) that Amphioxus has a more complicated swimming behavior in the open than when locomoting in close quarters where some of its movements, including perhaps back- ORIENTATION OF AMPHIOXUS 41 ward swimming, are omitted. The whole locomotor response is undoubtedly much more rapid during unimpeded progress, hence it might be argued that although the animal in slower and more deliberate swimming carried its anterior end persistently in advance, yet when moving rapidly, if once reversed to back- ward swimming (and I have shown above that this reversal actually occurs from time to time), the physiological inertia of the animal as it travelled at its highest speed would tend to keep it so oriented until the swimming movements grew less energetic and the animal returned to the deliberate swimming that is characteristic at the end of the course. The data given at the beginning of the paper, in which nine out of fifty animals were recorded as oriented with the posterior end in advance while swimming freely in large aquarium jars, would tend to strengthen this suspicion; on the contrary, the actual results obtained from a study of stained animals did not substantiate such a line of reasoning. If an animal, with the exception of one end, is wrapped in a fold of wet absorbent cotton and laid on a glass plate, the ex- posed end can be immersed without difficulty in the stain; in this case a weak solution of neutral red made up in sea-water was used. The anterior end was the one stained in most cases, for the more open structure of the pharyngeal region offers a larger surface for the reception of stain. After the stain had caused coloration to a deep pink or light reddish shade the animals were allowed to recover over night. When stimulated after such treatment, the stained extremity could be followed with comparative ease, and observations made in this way corroborated my earlier conclusions. Amphioxus does not locomote backward for any considerable distance, even when the response is extremely vigorous; but, after a somersault brings it tail-end in advance, either another reversal follows directly or the animal changes its course and returns more or less in the direction from which it came. 4 I believe the observations recorded in the first experiment of nine lancelets, which were supposed to swim backward, can be explained as follows: When the swimming response is nearing 42 LESLIE B. AREY completion, the vigor of the muscular movements rapidly de- creases and ends in complete collapse (Rice ’80, p. 8; Parker ’08, p. 441). After cessation of movement the animal is carried on a short distance by its own momentum and then sinks slowly to the bottom. According to notes taken at the time, five of the nine animals thus observed were judged chiefly by the finish _ of their response; if a reversal occurred just previous to the ces- sation of swimming, it is reasonable to expect that the nearly exhausted animal would not reverse again but would continue tail first with the last feeble strokes which precede complete exhaustion. When Amphioxus has been kept in the laboratory for a short time the anterior half of many animals begins to turn pink and in a few days that end may become decidedly colored. This is presumably a manifestation of an approaching moribund con- dition, although the reactions of the lancelets appearto be practically normal. Observations of a number of Amphioxus in this condition led to conclusions similar to those gained by the study of artificially stained animals. Referring to the quotations above, it will be seen that the views of Rice and of Steiner, although agreeing with mine in the main, show some difterences. Rice’s statement that he never saw an Amphioxus move ‘tail-end foremost’ is not only con- trary to the results given in my tabulation, where nine out of fifty observed animals were so oriented during normal swimming, but is also not in accordance with Parker’s (’08, p. 431; pp. 437— 440) experiments, in which resting animals stimulated mechanic- ally or chemically on the anterior end or mid-body, responded with a backward spring. Steiner’s simple statement that Amphioxus locomotes ‘das Kopfende voran’ is certainly too general and omits entirely any mention of the characteristic backward movements just referred to in the criticism against Rice. My own observations agree perfectly with Parker’s, to the effect that Amphioxus burrows in the sand tail first, but his belief, obtained as an inference from this habit, that the animal usually swims backward is directly opposed to the con- clusion reached by me. ORIENTATION OF AMPHIOXUS 43 It is interesting, however, to see how near Parker was to the real solution of the matter, although it must be said that the whole question was not of major importance in his work; thus (08, p. 431) he says: When the anterior end of an amphioxus resting in a shallow dish of sea water was touched even lightly with the bristle, the animal usually sprang backward, though occasionally forward. The back- ward spring was often accompanied by a somersault-like movement, whereby the animal became turned end for end. When the stimulus was applied to the posterior part of the body, the result was almost invariably a forward leap. The somersaulting is only mentioned by him in connection with the backward spring, and this I have shown is characteristic- ally present at the time of reversal to normal swimming, while in a forward leap it is unnecessary and is usually omitted. The animals mentioned above, which swam slowly for a con- siderable distance in a pan of shallow water, afforded an oppor- tunity to observe the movements of the body during locomotion. The head and tail were bent simultaneously toward the same side; the posterior of all the flexures, which is by far the most prominent, occurs approximately at the level of the atriopore; the next prominent flexure is at about the region of the first gonadic pouches but is much less extensive than the former. When swimming slowly no other flexures are evident except a suggestion of one rather close behind the anterior flexure last described. The occurrence of the two largest flexures just anterior and posterior to the gonadic pouches suggests that these pouches materially increase the rigidity of the body throughout the region where they occur and thus actually determine the position of the major flexures. As might be expected, when a forward spring occurs the first flexure is initiated at the anterior end and muscular activity extends posteriorly like a wave; when an animal leaps backward the reverse is true. As regards orientation during locomotion, I thus conclude that while Amphioxus can swim backward for short distances, its normal orientation in continued:swimming is with the anterior end in advance. 44 LESLIE B. AREY LITERATURE CITED Rice, H. J. 1880 Observations upon the habits, structure, and development of Amphioxus lanceolatus. Amer. Nat., vol. 14, no. 1, pp. 1-19, pls. 1-2; no. 2, pp. 73-95. SreinER, J. 1886 Ueber das Centralnervensystem des Haifisches und des Amphioxus lanceolatus, und ueber die halbcirkelférmigen Caniile des Haifisches. Sitzungsber. kgl. Preuss. Akad. Wiss., Berlin, Jahrg. 1886, Halbbd. 1, pp. 495-499. 1888 Die Functionen des Centralnervensystems und ihre Phylogenese. Zweite Abtheilung: Die Fische. Braunschweig, 8 vo., x11 + 127 pp. Parker, G.H. 1908 The sensory reactions of Amphioxus. Proc. Amer. Acad. Arts and Sci., vol. 48, no. 16. pp. 415-455. ParkER, T. J.. AND HaswELt, W. A. 1910 A text-book of zoclogy. Vol. 2, Macmillan, London, 8 vo., xx + 728 pp., 537 figs. Witiey, A. 1894 Amphioxus and the ancestry of the vertebrates. Macmillan, New York, 8 vo., xiv + 316 pp., 135 figs. STUDIES ON THE PHYSIOLOGY OF REPRODUCTION IN THE DOMESTIC FOWL X. FURTHER DATA ON SOMATIC AND GENETIC STERILITY! MAYNIE R. CURTIS AND RAYMOND PEARL Some time ago one of the authors (Pearl ’12) called attention to the fact that any single record of non-production or low production could not be accepted as evidence for the absence of the genetic factor for high production, since the failure of a bird to lay might be due entirely to somatic (physiological) causes. Later a detailed description of two such cases was published (Pearl and Curtis 714). In both of these instances the ovarian eggs were formed, but did not enter the oviduct. In one ease the funnel mouth was too small to admit a full-sized mature yolk, and in the other there was an apparent lack of tone in the muscles of the oviduct and its ligaments. In both cases the yolks were ovulated into the body cavity and resorbed without causing any apparent disturbance in metabolism. The purpose of the present paper is to record some recent observations on other cases of the same general nature. MATERIAL On September 1, 1914, there were in the Maine Station’s flock of first-year birds, eight apparently healthy birds which had laid few or no eggs, and one which had laid well up to the beginning of the breeding season and then stopped laying. All of these birds were hatched between April 7 and May 21, 1913. Several were from high laying strains. In order to determine if possible the cause of the partial or complete sterility, these nine birds were killed and carefully examined. The observations 1 Papers from the Biological Laboratory of the Maine Agricultural Experi- ment Station, No. 73. 45 46 M. R. CURTIS AND RAYMOND PEARL on them confirm and extend our previous conclusions in regard to somatic and genetic sterility, and also in respect to the ability of a bird to absorb rapidly through the peritoneum yolks dis- charged into the body cavity. DATA Data on the nine cases of partial or complete sterility are given in table 1. From the figures in this table it is possible to com- pare the performance record of each individual, both with the anatomical condition of the sex organs at autopsy, and with her genetic expectation, judged by the performance record of her sisters. SOMATIC STERILITY Birds Nos. 141, 81, 364, and 383 belong to high producing families. Not one of them had a sister which laid less than thirty eggs before March 1 (Pearl 712). These birds themselves would, therefore, be expected to be good layers. Examination of the final columns of the table shows that Nos. 141, 81, and 364 could not lay for anatomical reasons. The oviduct of No. 141 had burst near the upper end of the isthmus. In the body cavity was a yellow liquid composed evidently of egg mixed with serum. In this liquid were many short tubular pieces of egg membrane of approximately the length of the portion of the isthmus anterior to the tear. The opening in the wall was of such size that it was impossible for an egg to pass it. The oviduct on both sides of the tear was in normal laying condition. There was a normal egg (a yolk en- closed in thick albumen) in the posterior end of the albumen secreting region. In the ovary was a series of five normal yolks, the largest apparently mature, and seven discharged follicles. At the posterior end of the body cavity was an empty collapsed egg membrane on which was a thin layer of shell. The perito- neum was slightly thickened, so that it was translucent rather than transparent. The viscera were normal. The bird was evidently in perfect health. : *10q}8} OUIBS OT} Aq BIOISIS-J[VY 1OJ 918 01GB} OY} UI SeINZY oYJ, ‘esnoy-SutAv] oy} Ul S194SI8 [[N] OU peY Spstq ose, | : “sasnoy 94} JO Zuluve[D pus YG 94} Jo Surzios oy} Zutinp poyoojsou JeyMeulos o1aM Ady} (Z) puB ‘s19JIVNb 9suvI}s OJUT P9AOU useq pBY SpITq OYJ, (T) :SasNvd OMY OF ONP IqQnop ou SBA 8339 10H Jo UOTyIOdo1d Ysty ey], “Joop 94} uo useqINo}; put Sjsou-del} UT 8930 AYXIS PIB] YOR Sty} Z oequieydeg 07 GZ YsNSny UOIg ‘osnoy-Zutpeesq oy} Ul ued ¥ OJUT PaAoUI put YsNSny ut yoo 4sey og} SutMp asnoy-Zur Av] od ULOI] P940e]08 SAY-AQIOJ JO YOOH ¥ JO F1vd ¥ 919M PoOssNosip o1OY SPIIG oUTU OY} YVY} Pres oq P[NOYs FI UOTJOOUUOD sIY} UT ‘“s}soU-deI} OY} APIs}no pre] 210M SFBa Bsay} yey} Udy} eTqvqoid sulees 4] *S9/91[[OJ 94} JO WOIdIOseI jo o7e4S BY} YIM 9][qI}eduUI0D polled v UTY}IM s}sou-deI} UI Pre] oso} pus JoNP 04} UI Zo oY} WOIJ 10} poyUNODOL aq Jou P[NoOd YOM s9[oI] [OF posreqostp ‘9aI4} 88 “ON Pilg pue ‘ouo pey GFE “ON PIIq ‘IAAQMOFT ‘SAH prey [BUILOU OJUT PoyeIOdIODUT 9UIODIq 0} PoTTey YOIYM pozE[NAO ATJUsVeI uVeq pBYy SH[OA eq} 9Tqeqoidult wees 41 apeur ‘AITAvO Apo ey} UI 8999 IO H[OA Jo souOSge OY} pure ‘sovJINs [vaUOT}IIed oy} Jo souvIvedde [eUIOU AJOINJOSGe 9g} ‘IayIIN ‘SINOY MO} TG UT Pre] Ueeq oAVY P[NOM SBF asoyy Vey} UTeIIE0 JSOUTTE ‘a10Jo104} ‘SI 4] “3H OY} JO SurAV] oy} 0} UOTyoN.YSqGo yucIvdde Aus 9104} SVM OSB JOYJIOU UT PUT JONPTAO oY} Ul BZ0 [VUIION & PRY SPIT OM} 9894} Jo TORN, ; &I 9% $1 é T 0 0 0 0 “S820 9 ueyy aioul pues 0g uBy} 889] JO UOTZ -onpoid J030TM (JIM 81948I8 JO “ON, LZ E64 Or G I € £ 9 OT “t'"*8399 o10UI 10 0g jo uo -onpoid J9}UTM UIA S10}818 JO “ON, Lip 4¥¢ 1 7 z g g | 9 Or “ gsnoy SUIAG] | | UI S10}SIS JO "ON ‘el pues “qoq T “BW | “user ul spio GG (99(T Je¥je 199 38 AlpJsour -001 Surjseu [[V Spiodal Buljysou |sprooe1 Zurjseny [AB puwidy ut oo «| Ane ‘9 Avy cl '0Z- <8) “1-61 IV ‘92 A®¥W |1Z1dysso4svq [sprooer Buysoeu [|idy paooer Bur |'6z ‘po “Ady ‘FT |pue ‘F ‘qaq ‘gT ‘OT ‘I “IB ‘82 €-1 ‘ydog |pue ‘zt ‘GT ‘gf fuo1gonpord jjo ¢ ‘g AvwW |-seN ‘OZeuNL’s |qaq ‘0g “AON |'99C ‘9Z ‘6T ‘ST uory ‘0 ‘FS GOT ssdsq\9g “ue ,:Sdaq\ Ivy ‘Z°qoaq :sssqis9juUIM SABZo oel‘cG] idy ,:sdsqaoq ‘yur *8380|'62 990 :835q|‘Z]T ‘OT “990 :888q/-onpoid uo sajyoN 0 L 0 99 82 L T 02 jAUY Sey Pamela Sp10001 suysou jo ‘ON 0 6 £ g 19 g tf ZL 8 “* prey 8399 Jo “ON T&P 646 6FE 8SP 9I€ £86 F9E 18 TF - —_—— ——_——. —— ya ee $$ —______— $$ UALIVUVHO HaAAHWON GUula Aqyriajs ajajduoo 40 youind fo sasvo auru uo nyop Buimoyyg . T ATaViL : 47 0 OF 0°GE 0'1F Gog 088 08 CaLy. steers * "SUIBIS ul JONPTAO JO FAIA, 69 creeees ss gOT9 OT 4ONPTIAO Jo yQsUeT Gg se sees Sa 1D1T] OF pesivyostp Jo "ON “* *y]O4 JnOyyIM AIVAO JO 4YSIOM 0g Ciahy, 8g an 0°S¢ ¢0F DOGO OOOO §<,W0) UL Y[OA JO FYBIOAY "UID T 9AOqB syjox jo “ON G89T 9806 0202 6808 TPSG GIFS 1682 0666 sees eee eps ar 4ysIem Apog 66 FOF LT 1G Lg SP PP GCE GL 19 89 G86 8°98 GhP 6L 0°99 &¢ £6 &P SOOO Coy -onpoid 10}UIM §,1o}sis UMNUITUT]Y 8° Lg £01 T 02 ESSE Coy -onpoid Jo}uIM §,1oqSIS UINUIIXE]{ Peete sree se eeg7gg -sIs jo uorjonp -O1d 1d}UIM UOT Te? 6G 8oF Olé UGHWON Guia II panuyuojg—l ATAV.L UULOVUV HO 48 [BUION [eulION punoj au0N punoj uoroni4s -qo [eorulojeue punoj eaon [BWIION yeuI0N peus orgs Apysiys [BULLION pouo orgy Apysys pouo poue Od APUTIS-—TOrGs ApysIg umnouojzded jo UOTJIpUOD punoj au0N BIOOSTA suowe sduin} [[eus AIOA Moy VY APIABI Apoq Ul 9a} ‘sduin] [[euis uy] punoj ou0N punoj ou0N sduim| [[Bwis uy APIABO Apoq jo puae 1011 -03s0d ul auviq -wWaul posdey]]oo Ajdure pazis [[Ny B Spmbry stq4 ul = =9ueIquIeUr 359 Jo saqny }yIOYS Auvyy “BIOOSTA SUIT }EG wmni9as YIM poxtu yo yon] AJLABO Apoq ut ‘azo YIJOA Jo uoTyIpUOD ou ‘!yoe8q ped (‘SUIS ¢Z yon plao pesny SUIAR] B50 -diio9 pey pig punoy 9u0N PUuUNO}j 9UON|HO poT[eM AIVAG!'4M) JOWINY vI}sAD punojy eu0N fo JoUIN} o14sXkHO\[oeuuny Jo sdry] U0} JONPIAGO/O} UOTJZONAISGO SuIARI-uoU pue Sul |surAvy] u9eM4 |snuIyyst suse, puvys [loys uOIse1 UTUING Suvsl0 XOS Jo -Av[-uou Al}oII}G|-eq Oo}EIpeutiequ]|-ua Bea !Surdvy] suse] Burkeqjgt B50 SsulAeT SUIAvT SurAey|-[e ur 330 !Surkey]\uoyrpuos o1ppaAQ TEP 66 6FE 8oF OT £86 PIE 18 Tl YMUaHWON Gulag UaLOvVuv HO penusjuoj—T ATAVL 49 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 1 50 M. R. CURTIS AND RAYMOND PEARL An examination of the bird’s egg record shows that she started her year with two 2-egg clutches, and a separate egg in October, a record not unlike that of many birds which are good layers. There were only three eggs recorded after October 26; these were — scattered; their dates of laying were December 18, February 4, and February 15. However, in November the bird began to go on the nest daily and the nesting records follow rhythms very similar to the rhythm of laying of a good egg-producer. The three irregular eggs recorded after October were in rhythms of nesting and may have been (and probab y were) errors due to an egg having been left in the nest the last time it was used, or to the attendant accidentally recording an egg instead of a nest- ing record. An accurate determination of the time, or of the cause, of the rupture of the oviduct is impossible. It seems clear, however, that at the time it occurred there must have been a completed egg in the lower end of the duct. This egg was evidently carried up the duct and through the rupture into the body cavity by an antiperistaltic movement of the duct, which accompanied or almost immediately followed the rupture. This egg was evidently resorbed through the abdominal peritoneum, leaving the collapsed membrane found at autopsy. The nesting rhythm of the bird; the egg and serum mixture containing the short tubes of egg membrane found in the body cavity; and the normal naked egg in the lower part of the albumen-secreting region; all make it seem certain that after the rupture of the duct the sex organs passed through their normal cycles and that the eggs were formed in a normal manner, so far as the rupture of the duct allowed. After passing through the rup- ture they were absorbed directly through the general peritoneal surface. Bird No. 81 had the two lips of the funnel tightly fused. In order to test this observation the duct was filled with water and the union of the edges of the funnel lips proved watertight except at one point at the lower angle of the mouth of the ovi- duct; through this the water slowly oozed; it was impossible for a yolk to enter the duct. Both ovary and oviduct were in laying condition. There were small lumps of absorbing SOMATIC AND GENETIC STERILITY 51 yolks among the viscera. In the ovary was a normal series of five growing yolks, the largest one mature; there were also three discharged follicles. It was apparent that the bird was ovulat- ing into the body cavity and resorbing the yolk directly. An examination of the egg record of this bird shows nesting records in a rhythm similar to a laying rhythm, and records of seven scattered eggs. These eggs occur in the nesting rhythms and, as in the case of the scattered eggs of No. 141, probably represent errors on the part of the attendant? — The peritoneum was slightly thickened, as in No. 141. The fusion of the funnel lips may have been secondary, and due to peritonitis caused by accidental ovulation into the body cavity. However, there were no other visceral adhesions and no present evidence of sufficient peritonitis to cause adhesions. While it is impossible to say how long the oviduct had been closed, at the time of autopsy, at least, there was no possibility | for a yolk to enter. Bird No. 364 also had the sex organs in laying condition at autopsy. In the ovary was a series of five yolks and five follicles. Ten centimeters from the mouth of the funnel attached to the inner wall of the duct was a cystic tumor the size of a large egg yolk. There was no evidence of yolk or egg material in the body cavity. The visceral peritoneum, however, was thickened as in the birds which had been absorbing yolks. Evidently the bird had completely absorbed the yolk (or eggs) derived from the ruptured follicles. She had records of three scattered eggs, but only one nesting record. These recorded eggs may have been mistakes, but, since the hen was not nesting regularly, 2 The difficulties of getting an attendant to look after trap-nesting operations on a large scale, who will consistently maintain the maximum level of possible accuracy (Pearl ’11) are extremely great. During the past year we have been particularly unfortunate in this respect. While making every effort to do his best, the person who operated the trap-nests was psychologically poorly adapted to such work, and consequently the records are marred by a number of such easily detected errors as those referred to above in the cases of birds Nos. 81 and 141, and probably by some others not so readily detected. While this is a very regrettable circumstance, it really is not so serious as it might appear at first sight to be, since, after all, the total number of errors in the records is absolutely and relatively small. 5s M. R. CURTIS AND RAYMOND PEARL the probability of getting credited with eggs she did not lay was much smaller than in the preceding cases. Since the maxi- mum diameter of the duct is much greater than a normal yolk or the tumor it is possible that a few small yolks passed the tumor and became the yolks of normal laid eggs. Bird No. 383 when killed was in normal laying condition, with an egg in the shell gland and a series of five normal yolks and four folicles in the ovary. There was no apparent ob- struction to egg-laying. The peritoneum was normal and transparent. As explained in the footnote to table 1, it is probable that three eggs derived from the follicles on the ovary had been laid on the floor of the pen. The other was still in the duct. This bird’s production record shows only three scat- tered eggs and seven nesting records. At the time of autopsy she was apparently in perfect health and capable of producing eggs. Why she had failed to lay was not apparent. The idea occurred to us that possibly she had been habitually laying on the floor. Investigation of the floor egg record for the pen in the laying house in which No. 383 had spent the year did not indicate that this was the probable explanation. The total number of eggs laid on the floor in this pen of 125 birds to March 1 was 92. This was not higher than the average number of unrecorded eggs for a pen of that size, and was not as high as many of the individual birds’ trap-nest records. Also, the attendants pick up all birds seen laying on the floor and a habitual floor-layer is bound to be discovered in the long run. We must, therefore, attribute the failure of No. 383 to lay to some obscure physiological condition. That there can be no reasonable doubt that No. 383 was genetically a high producer in constitution is shown by a careful examination of her complete record in this respect. Her sire was o No. 623; all his daughters, which were hatched before June 1, with four exceptions to be noted below, made records before March 1, of over 30 eggs each: SOMATIC AND GENETIC STERILITY 53 Sire o' No. 623 XK 9 different 2 2 (fInL, * fIyLe) | Winter production Over 30 Under 30 Zero Observed 28 4 0 Expected 32 0 0 plus No. 383, the abnormal bird under discussion The four birds giving records under 30 laid respectively 26, 26, 23, and 19 eggs before March 1. These are undoubtedly to be regarded as somatic fluctuations from zygotes carrying the two factors? for high winter production (over 30 eggs). The bird which laid 19 eggs was not hatched till May 20. Bird No. 316 belonged to a small family of which a part were high and a part low producers. By her own winter production of 55 eggs she had shown herself to be genetically a high pro- ducer. After the first of March she began to make many nest- ing and a few egg records; the last egg record was on April 21; the nesting records continued in normal laying rhythms. Autopsy examination showed that there was a cystic tumor attached to the inner wall of the funnel mouth which practi- cally closed the funnel; the oviduct was in laying condition; the ovary had a series of six large yolks and three follicles; there was some free yolk in the peritoneal cavity; the peritoneum was slightly thickened. It was evident that this bird was normally a high producer, which had gradually developed a cystic tumor. This had at first hindered and finally prevented the entrance of yolks into the duct. From that time on the bird had gone through the normal laying cycles, ovulating the yolks into the body cavity and absorbing them. Bird No. 458 belonged to a family which included both high and low producers. At autopsy the ovary was walled off from the rest of the viscera by a large peritoneal pocket which was attached to the dorsal body wall on all sides of the ovary. The pocket was bulging with the contained ovary and a considerable 3 TZ, and LZ» of our former notation. * 54 M. R. CURTIS AND RAYMOND PEARL amount of free yolk. A little yolk was oozing through a small hole at the midventral angle of this pocket. There were a few small lumps of yolk among the viscera. When the pocket was opened and the free yolk wiped out the ovary was found to be in normal laying condition. It contained a series of four growing yolks, the largest mature, and seven resorbing follicles. The oviduct was in laying condition. At the time of autopsy, then, this bird was ovulating into the ovarian pocket and absorbing the yolks. The rate of absorption was evidently not equal to the rate of yolk formation. In consequence the pocket was stretched to its capacity and had given way at one point, allowing a little yolk to escape into the body cavity. The egg record of this bird shows neither eggs nor nesting records until December 22. From then on until the end of the year the nesting records follow a rhythm similar to a very slow laying rhythm. There is one egg recorded on February 2, a normal clutch of three on March 13, 15 and 17 and a single egg again on May 26. While it is possible that records of all these eggs are mistakes due to causes discussed in reference to the egg record of No. 141, it is probable that at least the record of the normal clutch in March is authentic, as the errors considered would hardly have resulted in this sort of a record. It seems probable that the pocket was neither congenital nor formed complete at once but that during its growth it first hindered and finally prevented the entrance of yolks into the oviduct. In any case, the immediate cause of the partial sterility exhibited is somatic, in the sense that it is not directly con- nected with or related to the genetic constitution of the bird in respect to fecundity. Bird No. 431 represents a case of somatic steri ty of a different sort from those so far considered, in that the difficulty was not primarily with the genital organs. This bird had a crippled back which interfered with the normal use of her legs; she was in poor flesh; she had never been in a trap-nest either to lay or nest. At autopsy the sex organs were in strictly non-laying condition. ‘There was no visible obstruction between the ovary SOMATIC AND GENETIC STERILITY a9) and oviduct. It is, however, improbable that the physiological tone of the bird had ever been sufficiently good to allow the formation of yolk. Genetically the bird might have been either a good or a poor layer. GENETIC STERILITY Bird No. 349 belongs to a family in which are individuals with, and individuals lacking, both factors for high fecundity. At autopsy this bird was in laying condition, with an egg in the oviduct and a series of five growing yolks and four discharged follicles in the ovary. Two of the three eggs recorded were laid respectively on the day of death and the second preceding day. These two eggs and the one in the duct account for three of the four follicles. The other may have furnished the yolk for a floor egg, explained as in the case of No. 383. There was no yolk in the body cavity; the peritoneum was normal. This bird had made no nesting records. Whether this bird was the extreme of genetically poor layers or whether her sterility was due to somatic causes too subtle for detection by rough autopsy examination is impossible to state absolutely. The probability, however, is extremely great that this bird genetically carries .only LZ, or L,—that is, has only one dose of any of the factors on which production depends. This is evident from the follow- ing considerations: Sire o& No. 627 x 9 different 2 9 (fInLe ° fle) Winter production Over 30 Under 30 Zero * Observed 10 13 1 (+No. 349, the bird under discussion Expected 9-6 12.8 3.2 Now the dam of No. 349 was No. 303 J, whose genetic constitu- tion was fL,L.. flil2, with a winter record of 23 eggs, and a record for the year of 79 eggs. Absolutely the most likely result of mating such a bird with a Type 4 male, where, as in the present case, there is only one in the family, is a bird which will make a winter record under 30—that is, one which carries but one dose 56 M. R. CURTIS AND RAYMOND PEARL of either L; or Lo. The record of 1 egg for No. 349 on January 26 would, on a strictly literal interpretation, put her in the ‘under 30’ class. It seems clear, however, in view of the rest of her record, and of the fact, already repeatedly pointed out, that March 1 does not represent biologically the absolutely invari- able time of beginning of the spring cycle of production, that she is really a zero winter producer. From a mating like that of o& No. 627 x ¢? No. 303 J one zero producer in every eight offspring is expected. We may next consider the case of bird No. 249. This indi- vidual had no full sisters. The nature of the mating from which she was produced is shown by the following pedigree. Sire o' No. 628 8b/ 3.01 | 4.66 | 75) 93,F children in Fi, at 148 |3/9| 12 | 2.95 | 4.23 | 39| 38} 413/5]/5]| 10 | 3.41 | 4.61 | 48] 42 132 |6/8| 14 | 2.78 | 4.97 | 50| 85] 417/6/5 | 11 | 3.98 | 5.73 46) 38 | | 329 | 616. 12a | 1.76 | 2.72 |102/ 80 C children in F, | | | || 4121616] 12b | 3.20 | 4.54 | 29) 37 141 |3|4| 7 |3.27| 4.54 | 77] 90] is | | 155 |4/4/ 8 | 3.00 4.63 11 11 | | 146 |2|7| 9 | 2.22 | 3.90 | 66 52) | 144 15/6] 11. | 2.67 | 3.71 | 80| 98| 142 |6/8| 14a | 2.60 | 4.73 | 41) 69| | | 153 |5|9}| 14b | 2.97 | 4.76 | 37 26 | D PS ewe aN Nol GGE | ea Alar No 35298 | Bore 1\O2c77 | e\ BOF? | WN / \ | \ ip $< / Nol 774 / 27800 216%? No 284 # 122 a0 WAC 2 ° Six groups of cousins; in each group the parents are sibs. Solid lines The parents are indicated by squares (males) The fractions follow- The numbers A, grand- Fig. 3 are females, broken lines males. and triangles (females) above the curves of their children. ing the numbers of the mating indicate the grades of the parents. of males and females involved in the curves are given at one side. children of no. 56, in the 4th generation; B, grandchildren of 115, in the 6th generation; C, grandchildren of 116, in the 6th generation; D, grandchildren of 143, in the 7th generation; E, grandchildren of 246, in the 10th generation; F, grandchildren of 274, in the 11th generation. ~I oo 14. E. CARLETON MACDOWELL averages, 14 a little lower, and 8a much lower. In the second eroup of families in F, (fig. 3C and table 8C) female averages from parental sums 7, 8, 14a, and 14b are similar, whereas those from sums 9 and 11 are a whole bristle lower. The male averages for parental sums 7, 8 and 14b are similar while those from 11 and 14a are lower, and from sum 9 a whole bristle lower. In F; is a set of cousins consisting of four families (table 8D and figure 3D). In this group the low parental sum has the low male and female averages, whereas sums 11 and 14a and 14b are much alike. In F,, there is a group of cousins in four families (table SE and fig. 3E). The female average for parental sum 12b is considerably larger than that for the sum 10, yet the female average for the sum 12a is about equal to that for the sum 10. Parental sum 11 has the lowest male and female aver- ages. In F,, a group of four families (table 8 F and fig. 3F) shows averages for parental sum 11 that are above averages for 10 and 12b. Parental sum 12a has averages much below any of the other families. This study of averages must be followed by careful observation of the figures referred to above, which show the frequency distributions of each of the families that has been considered. These curves have been plotted so as to enclose in each polygon similar areas, to facilitate com- parisons. The actual numbers of flies are given both in the tables and the figures. Further data bearing on the question of the relation between the grades of parents and children is to be found in the series of curves representing inbred lines through single matings (table 11 and fig. 6). One case must be especially emphasized. The last generation, No. 216, in the series beginning with No. 18 (fig. 6 A) was produced by flies with no extra bristles. This generation which follows seven selections has a higher distribu- tion than the preceding ones which were from high grade parents. There are no normals from these normal parents, and no females with less than two extra bristles. This extreme case is in ac- cord with all that has been found in regard to the independence of the parental and filial grades. This is further borne out by these inbred lines in the comparison of parents and children in successive generations. BRISTLE INHERITANCE IN DROSOPHILA 79 So the general conclusion seems to stand, that after the first few generations there is no close relationship between the grades of the parents chosen and the grades of the children. This appears to mean that in regard to extra bristles the flies in later generations do not differ genetically from each other; that in the later generations the variability in the number of extra bristles is due mainly, if not entirely, to conditions outside the germplasm. A further test of this would be to cross with normals flies with many extra and flies with few extra bristles and compare the results in the second generations. If the difference between the many and the few extra bristles be due to genetic factors, this would be shown in the F, of the crosses. Data is given that shows the distributions of extracted extra bristled flies from crosses of high and low eatras with normals of a different race. To avoid any complication that might arise from not consider- ing sexual differences, and at the same time to find evidence in regard to sex linkage, four types of crosses with wild flies have been made: high and low males, high and low females. In table 9 the data are arranged to facilitate comparisons between males and females of like grades, but by comparing alternate lines the relations between high and low grades are clear. The ratios and the distributions are practically alike in all cases (fig. 4). In order to make the curve for high males, including over 800 individuals, more easily comparable with the others, these data were plotted on the basis of 600. In conclusion, comparisons of families in the second generation of inbreeding show a tendency for higher parents to produce higher children, but this is not found in the sixth or subsequent generations; in single inbred lines there is found singular inde- pendence of the parental and filial grades, seemingly normal parents being able to produce all extra children; by the analysis of crosses it is found in the ninth and tenth generations that, as should be expected from the foregoing statements, high and low bristle grades are genetically indistinguishable; that the variability found in the late generations is apparently not due to genetic factors. 76 E. CARLETON MACDOWELL TABLE 9 Showing that low and high grades of extra bristles give the same results when crossed with wild normals; also that reciprocal crosses show that no sex linkage is involved Fy | Fe | RATIO | | __|NORMAL Pi aa | alle Py ea if TO ] ‘ei 7 EH lel 9 | normal | De ete Sul aoa an nGi allie |S). 1) SO -ail bao mareaiAt Low o'o' X Wild 9 9..../3411 | 1756 (156/147| 98| 57|.37| 5 3| | 1 |8.4:1 ) Low 22 X Wild 0’. 437 3 | 2) 1841 |188|159/119| 48 33) 10) 4] | | 2.8:1 High #7 X Wild 9 9... .282 4 | 1 | 2870 |307)223/183| 89 38) 134) | | 3.3:1 High 99 X Wild ac... 203 1954 |192/166/109| 64 33) 65) | |3.3:1 N ies SSSAN Low O"xwito 9 Ss Low9 xwito of ____. BN HIGH Okwito @) HIGH g@ xwito ot a me Ta Shaya Fig. 4 Extracted extra bristles appearing in F, of reciprocal crosses of high and low grades of extra bristles with normal wild, showing that there is no facto- rial difference between the high and low grade eztras crossed. 2. RESULTS OF SELECTION a. Total generations On the basis of the conclusions just reached, we should expect to find selection effective in the first few generations and later, ineffective. Data will be presented to show the results of selection in the form of curves for total generations and for single lines. Owing to the difficulty of obtaining high grade females that were unquestionably virgin, the most rigid selection was not always possible, so that even in the later generations some matings were made of low parents. In order to avoid the possibility of hiding a real increase in the offspring from high grade parents by including the offspring from low grade parents, BRISTLE INHERITANCE IN DROSOPHILA 77 the children having either parent below certain grades were omitted. Along with these were excluded all flies whose grand- parents or more remote ancestors had been excluded. This reduced the numbers in the total generations very considerably, and this is especially so in the latest generations. In these resulting curves the parents of each generation are included in the curve of the preceding generation. In figure 5 the curves have been’ plotted in such a way as to include similar areas, al- though representing different numbers of individuals. For this reason the numbers of males and females included are put down beside the curves. The solid line represents females; the broken line, males. The small curves between the larger ones in each case represent the parents selected from the gener- ation above, to produce the generation below. Discussion of the curves for total generations. The original parents selected from wild stock were a male with one extra bristle (+1) and a female with two extra bristles (+2). Their children are represented in the first pair of curves. The large number of normal flies in this generation 44.90 per cent is due to the fact that the mother was not virgin. Disregarding these, the male and female modes are at +2. In the following generation the modes remain at +2 but the upper limit of the range goes up from +4 to +7. The flies chosen as parents from this generation, range from +2 to +7. In the third generation the modes are still at +2 but the male curve has moved a little higher and the female curve is markedly above the male. This is clearly seen by comparing the areas of the polygons above the vertical line drawn at +2. The flies chosen from this generation range from +4 to +7. In the fourth generation the female mode is raised to +3 and a clear advance is shown by the whole curve. Although the male mode is still at +2, the proportion of +1 males has decreased and the pro- portions of +3 and +4 have strongly increased. In the fifth generation the male curve has increased its range over the pre- ceding generation but as a whole has fallen back a very little. The female curve has a much increased range, and a slight sag above the mode, otherwise no change. In the sixth generation 78 E. CARLETON MACDOWEBLL the male mode still holds to +2 but the proportion of +3 flies has grown to nearly equal the mode. The proportions of +4 and +5 flies have also increased although +5 is the limit. The female mode is actually at +5 but the expected mode is easily seen to be at +4. So far the progress and increase has been steady and unquestionable. The following five generations do not show any such advance. In generations seven, eight, nine, and ten the male mode remains at +2 and the proportion of +8 flies is very slightly less. In the eleventh generation the +2 and +3 males are equal, however the proportion of +1 flies has greatly increased in this generation, so this does not mean any general advance. The female curves have modes at +3 in the 7th and 8th generations, with +3 nearly as high. In the last 3 generations the female mode is at 4 and the curves are in general of the same kind. To show what parents were used in the different generations more clearly than can be done by the curves, their distributions are shown in table 10. The averages of the parents are given in two columns at the left and the averages of their children are in two columns towards the right. This table gives a clear summary of the progress of selection. The averages of the children, males and females, show increase up to the sixth gener- ation, when minor fluctuations appear seemingly without any significant change. If the variability in the last five generations be amenable to selection, a gradual decrease in the standard deviation would be expected, even though some hypothetical physiological limit prevented the means and modes from advancing as before. Fig. 5 Eleven generations of selection. Original parents indicated by a square (father) and triangle (mother) above F; broken lines, males; solid lines females. The flies chosen as parents are arranged in small -curves above the curves of their children. In generations 8 and 9 the male parents were all grade 5 and are represented by a square the proper distance from the base line. A finely dotted line means males and females together. The curves are plotted in such a way as to include like areas, so the numbers of individuals included are given at the side of each curve. The upper limits are in some cases drawn on the base line to show that such extremes appeared, although in too small numbers to form a whole unit in plotting. Se \ i \ , | \ Ny , | \ : : | , | \ \ ; | | \ | \ , | 93407 ' | ‘ / /' Sea 4 \ Fi ae = \ eit 3 (Se = 12738 . | 2 | | \ / =, : / \ / \ / \ a | ‘ Ss 210500 (7 \\ Len MET aN a ‘ | \ , | ! fh S / | | | | | \ = ? ‘ F415 ot ‘ > S\N SOS ; / \ es 5) == = = ‘ x S06 | e / c fa ae | i \ , | \ | | \ | ] \ FS i B / \ 10900 / \ a Si \1 16599 yaa | | aa \ \ Fio 517 oa" ATTL? 4 ee _ rey \ s \ \ AS : 59200 SSN S26 #8. 80 ; E. CARLETON MACDOWELL But the standard deviations show a gradual increase while selection is effective according to the curves and the means, and when selection has seemingly become ineffective, the stand- ard deviations fluctuate, but do not show a gradual decrease. b. Single lines The most immediate facts are presented in the form of the his- tory of five inbred lines through single matings, as shown by curves in figure 6, A, B, C, D, E. The grades of the parents selected from each generation are indicated below the curve of their own fraternity and above the curves representing their offspring. Squares are the males; triangles are the females. Four of the lines start from the same pair of first generation flies. The first two generations which are common ancestors of all four lines, are represented in only one set of curves. The parents of families Nos. 87, 71 and 88 are all sibs in family No. 56. The parents of family No. 116 come from family No. 88. The line headed by family No. 17 starts with a different pair of first generation flies. As the history told by these curves is clear it does not seem necessary to discuss each set of curves in detail as has been done for the total generations. In Table 11, A, B, C, D, FE, the means for these curves are given. These curves give the clearest picture of the basic facts. Clear progress is shown in the first five or six generations. In each line are differences counter-balanced in the total generations, yet in no case do these differences modify the general conclusion drawn from the study of total generations. 3. SUMMARY In brief, selection appears to have made advances for six generations. When large enough numbers are observed to afford a counter-balancing of the fluctuations of individual families, the advance is steady. That this increase is really genetic is shown by the increase in the distribution of extras extracted from a cross of long selected flies, over the distribution of extras extracted from a cross of unselected flies. In later generations *s194}0 OY} SB SISBq OUTeS oY} UO 40]d 0} S|BNPIAIPUT Atay 00} 219M 910} PIT “OU T UT “PAINO Yo opIssUO|Y UOAIS O21" P2A]OAUT SOY Jo Sioqunu ayy {svere penbo yo suoBAjod usr0y 04 poyyoyd ore saamng) ‘sopSuvtsy ‘soyeuey ‘sorenbs ‘soyeur ‘AqruI0}V.y ITOyy JO SAAIND 94} MOA UMOYS ore UOTCI9MED ZUTMOT[OJ oy} Vanposd 0} pajoojes syuored ay], “SeUl] PIOs ‘seyeuray ‘saul] uayorq ‘soyeyy “SOY yy JO Ted yuoLoyTp B WOT] St | OUT] ! gf OUT] UT TZ Wor ore C{ OUT] UT S}uered ysiy O43 8881 | PPLE We op $8EZ1 {y aul] UT 9¢ ‘OU WOIZ Y}Oq OIE (78 PUB TZ ‘SOu) O pus g Sour] Jo syuored 4s1y OTT, ‘soul PaIquI OAL 9 “BU LPSO| aszeru ‘i sig! $4611 P61 | PPPG y 3292°N FISI°N 4 880S g 54 LP8r ae si AS Sie s3izy P PLRS $3251 secz°u xg PPE 1 #h2I°H ‘ 4 2°02! te ee a2e1, Vi $s£1 2 PPEE rein | al $502 68cS ssc PPLI 82S e BES°N \ £99 j #sIl = g 4 O¢1°N \; = ves ; #99271) $sc02z f g PPB / ease a eleieu 2 re Sas : - t = atic PCO! £9TON Kf §C21°H s8r6 i 4 ¢ ¢ Anne eparel a eel ae ees #901°N \J 4 zt v P 2868 we £9 Kao , d8ss P06 \ | #IGI°N \/ PPI a #SSIoN cd Peay B2ZE1°N y zg V PPED a | pies 316 PPSS THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No, 1 BRISTLE INHERITANCE IN DROSOPHILA 85 (after the 6th) as far as the eleventh, no further advance has been detected. This is what the comparisons of different fami- lies from the same grandparents, or successive generations of inbred lines, had led one to expect. In the second generation there is a tendency for parents of higher grades to produce chil- dren of higher grade, while this is not the case in the later gener- ations where high and low variates seem to have the same genetic potentiality, and in so far as this is true, the variability must be explained by extra germinal causes. THE ROLE OF THE ENVIRONMENT 1. INFLUENCE OF FOOD In seeking causes for variability in the number of extra bristles, studies have been made on certain factors of the environment, namely, food and temperature. It was early observed that if any normal flies came from a bottle of inbred selected flies they were apt not to appear among the first flies to hatch in that bottle. Also it became evident that more of the high grades did appear among the first flies hatched, than among the last. To make this clear the data was arranged to show the distribu- tions of the flies counted on successive days from individual bottles. In many cases it was clearly shown that successive days showed lower and lower distributions of extra bristles. In other cases there was no such decline. However, in all cases the highest flies appeared among the first flies drawn off. That the falling off in bristle numbers is not due to any differences in the fertilized eggs is shown by the fact that whenever there is found such a decline in bristle numbers at the end of a bottle, the first flies hatched from the next bottle into which the same parents had been placed, show bristle numbers as high as those found at first in the preceding bottle. This was found to hold good for all cases, even though the parents were moved into three or more successive bottles. When the changes from bottle to bottle were frequent there was less falling off at the end of a bottle. Since the facts are so clear it seems needless to publish a large number of examples. Table 12 gives 3 ex- THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 19, No. 1 86 E. CARLETON MACDOWELL amples from the 50 families so arranged. In mating No. 78, there was no falling off at the end of first bottle, but the second bottle shows very clear falling off. There is found, then, a certain relation between the time of hatching and the number of extra bristles, due to conditions external to the germplasm, namely, that the last flies from a bottle may be of lower grade TABLE 12 Daily counts of progeny from the same parents in successive bottles, showing that when the bristle numbers are lower at the end of a bottle, the high grades are regained at the begin- ning of the next bottle. | fame | | Gea (EO lao pax! 0 | 1/213) 4 | 54\cealms ane |2O worse totem | 2: | gl aaleamin see || 2 | M. iS lee °F. 1 { 10-35 wks. : be 35 wks. (1M. 1M. : : : : > 3-8 wks. 3 wks. S85...) Minneapolis fe iL at ff gos IIT Se sg ays ras 5 F. 3-10 wks. | 1M. IF 10 wks. 9 / 'S @oa c 4 eS : ne > 3-10 wks. = 10 wks. A 3M. M. S7 a 3-10 wks. + 3 wks. | feel yee ) ‘ = cS. Su. a 9 : BY 3-10 wks. | 41 F. pe 4 conte 1 1. 10 wks. 1M. 3 wks. 5 NM owe Sy PAS E = : 4 3-10 wks. 1M. 10 wks. c act 1M. | A. ie 3-16 wks. ‘ cS. G40. ae 1 2 5 wks LF. ' 3 wks ede . (1M 2M. . 3 wks. 98 (4M. | 2F j 3-6 wks. | (1 F. if pa (AF. leita oe 3-8 wks. | *1 M. 6 wks. DOE ide ih Veh 3-13 wks. Total 29 M. 15 M. 13 M. 36 F. 25 12oie As noted in this table, three of the litters are from the rat colony at the University of Missouri, and the remainder from a _loecal colony at the University of Minnesota. There are in- cluded twenty-nine males and thirty-six females, a total of sixty- five. In addition, two rats not listed in tables 1 and 2 were used as additional controls at three weeks in the study of the skeleton (table 7). In most cases, the experiment began when the rats were three weeks of age (time of weaning). These rats were held at con- 102 Cc. M. JACKSON Fig. 1 Photograph showing four albino rats at the age of ten weeks. The larger rats represent the normal, full-fed controls. The smaller rats are from the same litter, but have been held by underfeeding at constant body-weight since the age of three weeks. The rats are pure albinos (Mus norvegicus albinus) the dark spot on the head of one being an artificial mark of identification. stant body-weight for varying periods, from the age of three weeks up to the age of six weeks (8 rats) to eight weeks (3), to ten weeks (22), to thirteen weeks (1) and to sixteen weeks (1). A few were held at constant body-weight beginning at later periods,—from age of six weeks to age of thirty-two weeks (2), and from age of ten weeks to age of thirty-six weeks (3). Con- trols (25 in all) were also killed at the beginning and at the end of these time periods. At ten weeks of age, the normal controls are half-grown rats, sexually mature, of nearly adult proportions, WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 103 and five or six times as heavy as those held at constant body- weight from the age of three weeks (fig. 1). In addition, a large number of observations upon the normal rat previously published (by Donaldson, Hatai, Jackson, Lowrey and others) are available for comparison. Of the animals used in this experiment, the distribution of sexes is given in table 1. In table 2, the net body-weight (gross body-weight, including intestinal contents, is slightly higher), of the animals used is indicated. At each time period the average weight (and range) for each sex is given, for both controls and test animals held at constant body-weight. The cards containing the original in- dividual records for all animals used will be deposited in The Wistar Institute of Anatomy, where they may be consulted if desired. The rats were kept in ordinary wire cages (with wire-net bottoms, allowing the feces to drop through) and were individu- ally weighed daily before feeding. Those under experiment TABLE 2 Net body-weight of rats when killed; average weight and range indicated | CONTROLS (FULL-FED) | EXPERIMENTED (BODY-WEIGHT HELD CONSTANT) Controls No. and Weight and range: Weight and range: No. and | Test animals killed at sex grams grams sex constant for ~ ea (6M. | 24.4 (19.0-32.6) 3 weeks.. sept. = < see Salty land (21021 230).4) | ( 9 < 9° ¢ kos > ? if Ms ys J we 42.3 224 (2025-236) 3M. Bae eae 2eKSie | : 2 EES o-b WeeKS Bevel ior '| aga 29 OM 2ONGE280S)nl a oEe lh | ae 8 weeks | -ilkeseail 1M. | me eaes 7 o- PKS 21.2 (17.8-24.6) 2h) ORG ears 3M. 155 (141-167) 24.7 (20.7-81 .4) 8 M. 3-10 ee Se eebe sienti4scloo-118)) | 23.3 (iS 5-310), | 140. ie me il Je, 3-13 weeks 26.0 al Je 3-16 weeks 2M. 209 (203-215) Bad 1M. : 32 weeks. E 6-32 weeks SEES A ie | 166 | 50.4 iba | eI a 238 | 79.6, (732628525) || 2 ME) 3 5 weeks. EIS a8 a . 10-35 weeks eee ion. | 158. (153-162) 74.2 Lie’ Total 13 M. | 15 M. Debs 5) 25 F 104 Cc. M. JACKSON were fed an amount just sufficient to hold them constant at the initial body-weight. Of course slight fluctuations in the gross weight were unavoidable, but they rarely exceeded one gram above or below the initial weight. The food used was in all cases whole wheat (Graham) bread soaked in whole milk. Pre- vious experience has shown that, at least up to the age of one year, albino rats thrive and develop normally upon this simple diet. Water ad libitum was also supplied. Tt is a curious fact that under these circumstances the amount of food necessary for maintenance of body-weight apparently decreases as the experiment proceeds. Thus rats of about 25 grams gross body-weight when three weeks of age at the begin- ning of the experiment will at ordinary room temperature re- quire about 5 grams of milk-soaked bread daily for maintenance. Later, this will usually decrease to an average of about 3 grams toward the age of ten weeks.!. This is the opposite of what might be expected: (1) because at later periods the amount of avail- able food supply stored in the body has been greatly diminished; and (2) because the animals held at constant body-weight al- most invariably become much more active, requiring a greater expenditure of energy. Possibly the smaller amount of food required to maintain the animals at the later periods may be due to a greater absorption of water, thus maintaining a body- weight which would otherwise decline with the given amount of food. It is well known that during inanition in general the 1 Two examples may be cited. Six rats of litter No. 12 were held at constant body-weight (within a range of 1 gram) from the age of three weeks on June 21, 1914, average body-weight 23.6 grams, for seven weeks to the age of ten weeks on August 6, 1914, at which time the average gross body-weight was 23.8 grams. The average daily food-supply of whole wheat (Graham) bread soaked in whole milk for the seven consecutive weeks of the experiment was as follows: 5.1, 3.9, 3.7, 3.0, 3.3, 2.7, 2.7 grams. Similarly, six rats of litter No. 13, average weight 23.1 grams at three weeks of age on June 28, 1914, were held at constant weight for seven weeks until August 12, 1914, when at ten weeks of age their average gross weight was 22.6 grams. Their average daily food-supply for the seven consecutive weeks was as follows: 5.3, 5.0, 4.1, 3.9, 3.3, 3.2, 2.9 grams. Im all cases water (city supply, from the Mississippi river) was supplied ad libitum. The diminishing amount of food necessary for maintenance cannot be explained as due to increas- ing temperature, as this was fairly constant. Moreover, a similar condition has been found in other litters at all seasons of the year. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 105 percentage of water-content of the body increases, and it is quite probable that during chronic inanition resulting from main- tenance of a young, growing animal at constant body-weight the amount of living protoplasm in the body decreases. If the amount of metabolism is thereby decreased, a smaller food- supply would suffice for maintenance. As will be shown later, on account of the intensity of the growth-impulse, especially during the earlier periods of inanition, certain growth-changes occur which require the expenditure of energy. It is possible that this energy is supplied by the excess of food above that required for maintenance proper. Another, but less probable, explanation might be that during inanition the food-intake is In some way more economically utilized, a smaller quantity therefore being sufficient for maintenance. In the later stages of inanition, there is probably a decrease in the temperature of the body, which would therefore require less food. Rats held at constant body-weight from the age of three weeks to ten weeks, while becoming more active as the experiment proceeds, become at the same time less resistant to cold. They may die suddenly if the room temperature is lowered, or even without any apparent cause. Thus up to sixteen weeks, the longest successful period in those beginning at three weeks, it becomes increasingly difficult to maintain them alive at con- stant body-weight. When the experiment is begun later, the length of the time during which the body-weight can be held constant is considerably increased. Aron (’11) had a similar experience with dogs, finding it necessary after a time to feed sufficiently to increase the initial body-weight somewhat, in order to keep the animals alive. He explains this as due to the eradual exhaustion of available food-substanee stored in the various tissues of the body. At the end of the various age-periods of the experiment, and at the beginning and end for controls, the rats were killed by chloroform and dissected according to the technique described in previous papers (Jackson and Lowrey 712; Jackson 713, 715 ¢). The parts, systems and organs were carefully weighed, and 106 Cc. M. JACKSON portions preserved for microscopic examination (to be con- sidered in a later paper). As heretofore, in calculating the percentage weights, the ned body-weight (gross weight less intestinal contents) is taken. The percentage weights of the organs are thus slightly higher than if calculated upon the gross body-weight. The averages given in the various tables are the arithmetical means of the corresponding individual observations. In view of the comparatively small number of observations and the known variability, especially of some of the organs (ef. Jackson 13), the data are insufficient for treatment by statistical methods, and the values are therefore only fair approximations. They are, however, sufficiently accurate to show some of the more obvious and important changes in the young animal held at constant body-weight. It is hoped that they may be useful as prelimi- nary observations, which may lead to further and more extensive investigations of the various individual organs. In general, the amount of variation found is sufficient to necessitate great caution in drawing conclusions from a small number of observa- tions (sometimes upon a single animal), as frequently happens in experimental work. LENGTHS OF BODY AND TAIL The body-length is measured from the tip of the nose to the anus, and the tail-length from the anus to the tip of the tail. The measurements were taken immediately after death, the body and tail being extended by very slight tension. Measure- ments during life are not practicable, although they might be obtained by the use of anesthetics. In order therefore to discern the changes in the lengths of body and tail while the body-weight is held constant, it was necessary first to determine these measurements on the normal animal. For this purpose, 450 observations (267 males, 183 females) were available, varying from newborn to about 400 grams body-weight. Of these, 277 (130 males, 147 females) were from the Missouri rats described in a previous paper (Jack- son 713), and 25 (13 males, 12 females) from Minnesota. For WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 107 the remaining 148 (124 males, 24 females) observations upon rats from the colony at The Wistar Institute in Philadelphia, T am indebted to Professor Donaldson and Dr. Hatai. .,. ee eeeeeee ree ee nn. eee 2 42.4 Body-weight constant 26 weeks (age of 6 to 32weeks) 2 47.1 Normal at 10 weeks (Jackson and Lowrey)........ 10 131.0 ControlsiathH0kweekss 0 eee eeee ee eee eee 6 134.0 Bodyweight constant 25 weeks (age ot 10 to 35 weeks) 3 77.8 Controls at 32 and 35 weeks.... 6 189.5 percentage of net body-weight and — Amanwnwwwwnhd vw oO ABSOLUTE WEIGHT (AND RANGE): GRAMS .59 (3.86- 6.36) .81 (2.21- 3.21) 78 (2.27- 3.30) os _ => 2.63- 4.51) So _~ = 8.00- 8.08) 15 (5.70- 6.60) .70 (19 .40-82.10) .00 (12.30-17.20) .40 (28.10-55.20) RELATIVE WEIGHT (AND RANGE) PER CENT — wo C eoonrooorwnanans pp bo o cool (8.7 (1s. (9 (12 (16. (15. (12 (15 (17 —29.2) 4 -25.2) .82-14.1) (12.6 .3 -17.5) -15.5) 7 —25.9) 1 -19.0) .9 -13.1) .6 —22.3) .8 —22.0) (16.6 (18.0 —20.1) —23.2) 116 WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS Ey cerning the head and extremities, it follows that there cannot be much change in the trunk-weight, since the probable slight in- crease in the weight of the head is off-set by a slight decrease in the extremities. In the largest group, however, held at constant body-weight from three to ten weeks of age (fig. 3), the trunk would apparently decrease from an average of 54.1 per cent to 53.4 per cent. This apparent change is so slight as to be (prob- ably) insignificant. The results concerning the parts of the body therefore fail to indicate any decided change of proportional weights in young animals held for considerable periods at constant body-weight. There is apparently a very small increase in the head, counter- balanced by a corresponding decrease in the trunk and extremi- ties, but the change is so slight as to be of doubtful significance. During inanition in adult rats, there is apparently a relative increase in both head and extremities, counterbalanced by a relative decrease in the trunk (Jackson ’15 a, ” 15c). INTEGUMENT In the rats held at constant body-weight from the age of three weeks to six, eight, ten, thirteen and sixteen weeks, there is a very marked loss in the weight of the integument (including hair and nails; table 6; fig. 4). In the case of the largest (three to ten weeks) group, the decrease is from an average of 21.9 per cent to 14.5 per cent of the body-weight. In terms of abso- lute weight, the decrease is from 5.30 grams (5.59 grams, less correction on account of difference in body-weight, which aver- ages 25.1 grams at three weeks and 23.8 grams at ten weeks) to 3.41 grams, a decrease of about 36 per cent. The decrease would appear slightly greater if the difference in relative weight of the integument for different initial body-weights were taken into account. There is apparently even greater loss at the other ages. It would appear that this loss (which is perhaps chiefly a loss of fat) occurs rather early, as at six weeks (body-weight held constant three weeks) the loss is as great as at subsequent and longer periods. 118 Cc. M. JACKSON Integument integument 14,5 per cent. Integument 21.9 per cent. 20.0 per cent. Ligamentous skeleton Ligamentous skeleton Serene 21 2 per cere 10.7 per cent. 15.7 per cent Musculature Musculature Musculature 32.0 per cent 41.6 per cent. 31.2 per cent. Viscera Viscera 20.5 per cent. 20 a opericent Viscera 14.3 per cent. ‘Remainder’ ‘Remainder’ re F , emainder 13.4 per cent. 10.7 per cent. 10.1 per cent. Controls at 3 weeks. Constant 3 to 10 weeks. Controls at 10 weeks. Fig. 4 Diagram representing the average relative (percentage) weights of the various systems (integument, skeleton, musculature, viscera and ‘remainder’ ) in albino rats held at constant body-weight from the age of three to ten weeks, and in controls at three and at ten weeks of age. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 119 In the rats experimented upon at later and longer periods (ages of six to thirty-two weeks and ten to thirty-five weeks) there is also a marked loss in the weight of the skin, though apparently not so great as at the earlier periods. This less in the weight of the integument is in striking contrast with the results of inanition in adult rats (Jackson 715 a, ’15c). Here the loss is very nearly proportional to that of the whole body, so the integument nearly maintains its relative (per- centage) weight. From his experiments upon young dogs held at constant body- weight, Aron (’11, p. 29) states that: ‘‘The skin shows a slightly higher percentage of the body-weight in those animals kept at a constant weight than in the normal, control dogs. These figures indicate that, while the (body) weight was constant, the skin increased very slightly in weight.”’ The figures cited show the skin in animals held at nearly constant body-weight to form (in four cases) 12.2 to 14.6 per cent of the body-weight, whereas in three corresponding full-fed controls the skin formed 11.2 to 13.0 per cent. Aron, however, overlooks the fact that he is making his comparison with controls at the end of the experi- ment. In order to judge what changes have taken placed during the experiment, the comparison must be with normal control animals killed at the beginning of the experiment. Aron records but one case which can be used for this purpose. His Dog D (table 138, Experiment IV) killed at the age of 40 days, the begin- ning of the experiment, with body-weight of 1985 grams shows a skin-weight of 320 grams, or about 16.1 per cent of the body- weight. From Aron’s own data, therefore, I would reach the opposite conclusion, viz., that in young dogs held at constant body-weight, the skin suffers a marked loss in weight. This would agree with my results on rats. SKELETON The skeleton (table 7; fig. 4) was prepared in three ways. The bones, together with the cartilages, periosteum and liga- ments, constitute the ‘ligamentous skeleton’ (table 7a). The bones and cartilages, after removal of the periosteum and liga- 120 Cc. M. JACKSON ments by immersion for about one hour in 1 per cent aqueous ‘yold dust ’solution at 90°C., constitute the ‘cartilaginous skele- ton’ (table 7b). Finally the cartilagmous skeleton dried in an oven at 95°C. to constant weight constitutes the ‘dry skele- ton’ (table 7 ¢). An examination of the weight of the ligamentous skeleton (table 7a) reveals the striking fact that while the body-weight TABLE 7 The skeleton; average absolute weight, average percentage of net body-weight and range indicated Fs Sp 4 oD z 2 RELATIVE WEIGH DESCRIPTION OF RATS © g s 5 (AnD BANGRGaae CANS aca F O& »& es ae By ke PER CENT 5 2 Z, 9 a. Ligamentous skeleton Normal at 3 weeks (Jackson and Lowrey)......--- 13 24.8 4.12 16.6 (13.1 —21.10) @ontrolsatoaw COS meee ie clas cies eee inserter 12 24.5 3.90 (3.100- 5.46)} 15.7 (18.5 -17.00) Body-weight constant 3 weeks (ageof3to 6 weeks) 8 Zoe 4.04 (3.140- 5.50)| 18.0 (14.6 —23.90) Body-weight constant 5 weeks (ageof3to 8 weeks) 3 20.2 4.74 (4.290- 5.50)| 238.7 (22.3 -24.90) Body-weight constant 7 weeks (ageof3to 10weeks)| 22 23.8 4.98 (3.700- 7.14)| 21.2 (17.7 -24.60) Body-weight constant 10 weeks (age of 3 to 13 weeks) 1 Pa 5) 4.80 18.8 Body-weight constant 13 weeks (age of 3 to 16 weeks) 1 26.0 5.20 20.0 Normal at 6 weeks (Jackson and Lowrey).......-.- 14 64.4 9.02 14.0 (10.5 —20.10) WontrolstatiGlweeksiepec ccna: me cecil ence eclevees See 2 42.4 5.48 (4.750- 6.20) 13.0 (11.2 -14.70) Body-weight constant 26 weeks (age of 6to 32 weeks) 2 47.1 7.20 (6.800- 7.60)} 15.4 (15.1 -15.60) Normal at 10 weeks (Jackson and Lowrey)........ 10 131.0 15.30 11.7 (10.0 -12.90) Controlstatul Oaweeksre rect ccr tees si ee 6 134.0 14.00 (11.400-17.60) 10.7 (8.5 -14.10) Body-weight constant 25 weeks (age of 10 to 35 T.COLS) Sse ee tee rete cic seas Saige clelinere seer 3 77.8 10.30 (9.400-11.60)| 13.3 (12.7 -13.70) Controlsiatie2tand sbnweeKSe.. - ccs ota. ons we eee 6 189.5 18.60 (15. 600-24.30) 9.8 (8.8 -11.20) b. Cartilaginous skeleton (Fresh) Gontrolsyatiminvecksteerenin aot ete sas) eee 6 22.9 2.60 (2.120-3.00)| 11.4 (9.0 -12.90) Body-weight constant 3 weeks (age of 3 to 6 weeks) 3} 22.9 3.56 (2.930- 4.04)} 15.5 (13.3 -17.00) Body-weight constant 5 weeks (ageot3to 8 weeks) 1 24.6 4.50 18.3 Body-weight constant 7 weeks (age of 3 to 10 weeks) 8 22.4 3.16 (2.420- 4.00) } 14.6 (11.9 -17.10) Body-weight constant 10 weeks (age of 3 to 13 weeks) 1 251.9 4.00 Lieve Body-weight constant 13 weeks (age of 3 to 16 weeks) 1 26.0 3.96 15,2 Controls/atil Ohweeksmyereere ers aceon cee 1 115.0 6.40 5.6 c. Dry skeleton (cartilaginous) Controlsiatisiweekaiae eee eae eaeitel ears are 5 23.4 | 0.804 (0.710-0.964)| 3.43 (8.17- 4.03) Body-weight constant 3 weeks (ageof3to 6 weeks) 3 21.9 1.091 (1.033-1.172)| 4.98 (4.70- 5.13) Body-weight constant 5weeks (ageof3to 8 weeks) 1 24.6 1.351 4.59 Body-weight constant 7 weeks (age of 3 to 10 weeks) 9 22Ro 1.285 (1.076-1.485)| 5.84 (4.76- 7.00) Body-weight constant 10 weeks (age of 3to 13 weeks) 1 25.5 1.068 6.31 Body-weight constant 13 weeks (age of 3 to 16 weeks) 1 26.0 1.744 6.71 Controliatil0pweeks=neenec reer een errr sislacietieees 1 115.0 | 3.420 2.97 ee WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 121 is held constant the skeleton continues to increase in weight to a marked degree. In the rats beginning at three weeks, there is an increase in the relative weight of the ligamentous skeleton from 15.7 per cent of the body to 18.0 per cent at six weeks and to an apparent maximum of 23.7 per cent at eight weeks. This latter is probably an exceptional figure, as in the largest group, at ten weeks, the average is 21.2 per cent (fig. 4). This corre- sponds to an increase from an absolute weight of 3.90 to 4.98 grams, an increase of about 28 per cent (or slightly more, if correction be made for the difference in body-weight, average 24.5 grams at. three weeks and 23.8 grams at ten weeks). The two cases carried to thirteen and sixteen weeks, respectively, show a slightly smaller relative increase. The rats used at later and longer periods (ages of six to thirty-two weeks and ten to thirty-five weeks) also show a considerable increase in the skeleton, though relatively less than those beginning at the earlier period. The data for the cartilaginous skeleton (table 7 b) similarly show a marked increase in rats held at constant body-weight for various periods beginning at three weeks of age. The figures for the largest group (three to ten weeks) indicate an increase from 11.4 per cent to 14.6 per cent of the body. In terms of absolute weight, the increase is from an average of 2.60 grams (body-weight 22.9 grams) to 3.16 grams (body-weight 22.4 grams), an increase of about one-fourth. Subtracting the percentage weights of the cartilaginous skeleton from the corresponding ligamentous skeleton, there is (for the three to ten weeks group) an evident increase of the ligaments and periosteum from 4.3 per cent to 6.6 per cent of the net body-weight. This would indicate that the ligamentous component of the skeleton shares in the marked growth during constant body-weight. Professor Donaldson (in a personal communication) has kindly supplied a series of observations showing that the cartilaginous skeleton in the normal rat changes from a relative weight of about 10 per cent of the body at 20 grams to 7.5 per cent at 50 grams, 7 per cent at 100 grams and 6.7 per cent in rats above 122 Cc. M. JACKSON 200 grams. These weights, however, do not include the inter- vertebral discs. The data for the dried cartilaginous skeleton (table 7 c) indi- cate an even greater increase in the dry skeleton of the rats held at constant body-weight. Thus in rats beginning at three weeks the dry substance increases from 3.43 per cent of the body weight to 4.98 per cent at six weeks of age, 5.49 per cent at eight weeks, 5.84 per cent at ten weeks, 6.31 per cent at thirteen weeks and 6.71 per cent at sixteen weeks. Since the increase in the dry skeleton is relatively greater than that for the (moist) cartilaginous skeleton, it necessarily follows that the skeleton must be losing in percentage of water and gaining in percentage of dry substance. The percentage of dry substance has been calculated for each individual skeleton included in tables 7 a and 7b, and the averages for each group are as follows: controls at three weeks, 31.4 per cent; constant three to six weeks of age, 33.5 per cent; three to eight weeks, 30.0 per cent; three to ten weeks, 41.7-per cent; three to thirteen weeks, 40.2 per cent; three to sixteen weeks, 44.0 per cent; control at ten weeks, 53.4 per cent. Lowrey (713) finds the dry substance of the ligamentous skeleton in the normal albino rat to increase from an average of 33.3 per cent at 20 days of age to 39.2 per cent at six weeks, 45.9 per cent at ten weeks, 50.4 per cent at five months and 52.6 per cent at one year. From the foregoing it is evident that in rats held at constant body-weight beginning at three weeks, the growing cartilaginous skeleton steadily increases its percentage of dry substance. Thus it tends to change the proportions of water and dry sub- stance as during normal growth. The percentage of dry sub- stance does not increase so rapidly with age as during normal growth, however, but lags behind corresponding to the retard- ation in absolute growth. During inanition in adult rats, on the contrary, there is a relative decrease in the dry substance, and an increase in water-content (Jackson 715 ¢). It has already been noted in a previous section (‘‘Lengths of body and tail’’) that in the rats held at constant body-weight WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS $23 beginning at the age of three weeks there is an increase in the lengths of both body and tail. The latter increases more rapidly. however, so that it tends to assume the tail-ratio found in normal rats of corresponding age. This indicates that the skeleton not only continues to grow (though at a reduced rate) while the body- weight is held constant, but also tends to grow in a normal manner, so as to produce the normal ratio of tail-length and body-length. The preceding paragraphs have shown that the increased growth of skeleton affects the ligamentous as well as the cartilaginous and bony components, and that the chemical composition (per- centages of water and dry substance) also changes in a manner tending to assume the normal. The question naturally arises as to whether the skeletal growth during constant body-weight is merely a growth in mass, or is associated with the normal process of differentiation. During the present investigation a few observations have been made upon the development of the normal skeleton, indicating some of the more obvious changes during the age-periods of the rats under experiment, especially between the ages of three and ten weeks. While a detailed study of the developmental changes in the skele- ton is reserved for a separate paper, some preliminary con- clusions may be noted here. In skeletons of rats held at constant body-weight from the age of three to the age of ten weeks, the appearance and fusion of certain epiphyses may be noted as in the normal animal during ° this period, although in most cases the process appears to be retarded somewhat. The following examples may be cited. In the normal skeleton at three weeks of age, the epiphyses at the ends of the vertebral bodies have not appeared; the epiphysis at the lower end of the humerus is well developed, but not fused with the shaft; the maxilla and mandible present each two molars (on each side), with no visible trace of a third. In the normal skeleton at ten weeks, the epiphyses at the ends of the verte- bral bodies have appeared, and most of them have united with the corresponding bones; the epiphysis at the lower end of the humerus is firmly fused with the shaft; well developed third molar teeth have appeared, both in the maxilla and in the mandi- 124 Cc. M. JACKSON ble. In the skeleton of a rat held at constant body-weight from three to ten weeks of age, most of the epiphyses of the vertebral bodies have appeared, and they usually have united at one end of each bone; the lower epiphysis of the humerus is firmly united with the shaft, as normally at ten weeks; well- developed third molars have appeared, as normally at ten weeks. In a rat held at constant body-weight from age of three to six- teen weeks, the skeletal differentiation was more advanced, corresponding at least to the stage reached normally at ten weeks, and in some respects perhaps even beyond it. These observations will suffice to establish the fact that the skeletal growth during constant body-weight is accompanied by normal developmental changes, as well as changes in chemical composition (percentage of water). In other words, we find not only increase in mass but growth and differentiation appar- ently normal in character, though somewhat retarded in rate. These skeletal characters therefore tend to correlation with age, although influenced also by the general body-weight. The remarkable fact that the skeleton continues to grow while the body-weight is held constant was apparently first observed by Waters (08) who found that calves previously well nourished will continue to increase in height and in width of hip for a con- siderable time, even when increase of body-weight is prevented by under-feeding. He remarks (’08 b, p. 9): Apparently the animal organism is capable of drawing upon its reserve for the purposes of sustaining the growth process, for a con- siderable time and to a considerable extent. Our experiments indicate that after the reserve is drawn upon to a certain extent to support growth, the process ceases and there is no further increase in height or in length of bone. From this point on, the animal’s chief business seems to be to sustain life. This law applies to animals on a stationary live weight as well as to those being fed so that the live weight is stead- ily declining, and indeed to those whose ration, while above main- tenance, and causing a gain in live weight, is less than the normal growth rate of the individual. Such an animal will, while gaining in weight, get thinner, because it is drawing upon its reserve to supple- ment the ration in its effort to grow at a normal rate. Aron (711) experimented with dogs to determine the effect of a restricted amount of food upon young, growing animals. He WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 125 found that in spite of constant live weight, the animals con- tinued to increase in length and height for three to five months. Thereupon the emaciated animals became weaker and died un- less the amount of food was somewhat increased. From a comparative study of nine bones (the entire skeleton was not measured), Aron concludes that the skeleton during constant body-weight increases in mass and also changes in chemical composition (increase in water-content and protein (?); decrease in fat). The results of this very interesting investigation, while sufficient to establish the continued growth of the skeleton, would be more conclusive if the number of observations were larger, with an adequate number of controls at the beginning and at the end of the experiment. In a recent paper, Aron (’14) records a few observations indicating that malnutrition in children retards growth in length less than body-weight; so that the body may continue to increase in length while the body- weight is at a standstill, or even slightly decreasing. Thus the strong growth tendency of the skeleton during bare maintenance of the body-weight is manifest in the human species, as well as in the calves, dogs and rats. MUSCULATURE Although the musculature (table 8; fig. 4) in the normal rat at three weeks averages 26.9 per cent of the body, according to Jackson and Lowrey (’12), the controls in the present series gave a somewhat higher amount, the average being 31.2 per cent. As shown in table 8, the musculature in the controls of the present series also averaged slightly higher than the normal according to Jackson and Lowrey at six and ten weeks. In rats held at constant body-weight from the age of three weeks to six, ten and thirteen weeks, the musculature appears relatively very slightly higher than in the controls at three weeks. The apparent increase from three to ten weeks is from 7.40 grams (7.81 grams, less correction corresponding to the smaller body- weight at 10 weeks) to 7.62 grams, or an increase of 3.0 per cent in absolute weight. In the three to sixteen weeks experiment, 126 C. M. JACKSON TABLE 8 The musculature; average absolute weight, average percentage of net body-weight and range indicated ; = ” 8 Zz a z ABSOLUTE WEIGHT SPOS LEST TSE NOIRE Os DESCRIPTION OF RATS fs é d : (AND RANGE): GRAMS (AND pL ANGH). o> | 3% Z io) Normal at 3 weeks (Jackson and Lowrey)..... Pere |h eis} 24.8 6.67 26.9 (20.1-30.2) Controls'atseweeksy. peace tec ae eee ince eee 10 25.1 7.81 (6.90-9.82) 31.2 (29.5-35.3) Body-weight constant 3 weeks (ageof3to 6 weeks) 8 22.2 7.04 (5.51-8.05) 31.8 (25.1-34.7) Body-weight constant 5 weeks (ageof3to 8 weeks) 3 20.2 6.46 (5.68-7.70) 32.1 (31.3-33.2) Body-weight constant 7 weeks (ageof3to1l0weeks)| 22 23.8 7.62 (4.58-10.87) 32.0 (24.8-36.4) Body-weight constant 10 weeks (age of 3 to 13 weeks) 1 25.5 7.90 31.0 Body-weight constant 13 weeks (age of 3 to 16 weeks) 1 26.0 7.50 29.4 Normal at 6 weeks (Jackson and Lowrey)......... 14 64.4 | 21.10 32.7 (26.1-35.3) Control siatiosweekse saaenictics ace eiccieeee oes 2 42.4] 14.90 (14.85-15.0) 35.3 (35.0-35.6) Body-weight constant 26 weeks (age of 6 to32 weeks) 2 47.1 16.90 (15.30-18.5) 35.7 (34.8-36.6) Normal at 10 weeks (Jackson and Lowrey)........ 10 134.0 55.10 41.1 (37.4-49.1) Controlsiat: 10 weelkkssencprenne ron coe sei ace eee 6 135.0 | 55.80 (44.00-69.2) 41.6 (39.3-44.5) Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 77.8 31.20 (30.10-33.9) 40.1 (39.7-40.7) Controlsiatio2sandisbnweekss., oases e aeeieiae 6 189.5 | 81.20 (64.80-119.7) 42.6 (39.0-50.3) TABLE 9 Viscera and remainder; average percentage of net body-weight and range indicated e a : Z g2 RELATIVE WEIGHT OF a £ 5S |RELATIVE WEIGHT OF Roe DESCRIPTION OF RATS ie) = |vISCERA (AND RANGE) (nD RANGE) Of | we EERCENG PER CENT SE |) es a (=) Normal at 3 weeks (Jackson and Lowrey)......... 13 24.8 21.3 (20.1-24.8) 12.9 (4.0-19.9) Controlsiatiosweeks! ce mucceabincsk meahi aces sane nee 10 25.1] 20.5 (17.9-24.8) 10.5 (2.6-15.6) Body-weight constant 3 weeks (ageot3to 6 weeks) 7 PNA 25.0 (20.7-29.2) 13.6 (7.1-21.6) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 26.8 (25.0-28.5) PVs Taran Cat 87/5) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 | 22.2 (19.4-26.0) 10.0 (2.5-16.9) Normal at 6 weeks (Jackson and Lowrey)......... 14 64.5 | 20.4 (18.4-22.9) 12.0 (6.5-17.1) @ontrolsatiGuweeks: c-means oo: eee 1 42.4 19.6 19.1 Body-weight constant 26 weeks (age of 6to32weeks)| 2 47.1 | 19.5 (19.1-19.9) 16.4 (16.1-16.7) Normal at 10 weeks (Jackson and Lowrey)........ 10 130.5 16.0 (14.9-17.2) 12.5 . (1.2-18.4) ControlstatWOiweeksey smaeyeee riers. ie bes heist 6 135.0] 14.3 (13.0-15.7) 13.6 (9.7-16.4) Body-weight constant 25 weeks (ageof 10to35 weeks) 3 17.8 16.3 (16.1-16.8) 12.4 (9.7-13.9) Controlsiatio2;andysonweekssnee, awe ce cece 5 179.8 | 12.9 (11.9-13.5) 17.0 (13.7-19.5) WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS VAL the average is slightly lower. In the six to thirty-two weeks experiment, the musculature appears very slightly higher than in the controls at six weeks, while in the ten to thirty-five weeks experiment, the musculature appears slightly lower than in the controls at ten weeks. In the latter case, however, the controls are too heavy for comparison with those under experiment. In general, it seems clear from the foregoing that in young rats held at constant body-weight the musculature also remains nearly constant in weight, with perhaps a very slight tendency to increase in the mdjority of cases. In the course of normal growth during this period, the musculature shows a more rapid growth than any other system, increasing from about 27 per cent of the body at three weeks to 41 per cent at ten weeks of age (Jackson and Lowrey). During inanition in adult rats, the musculature loses approximately in proportion to the entire body, slightly less in acute inanition and slightly more in chronic inanition (Jackson 15 c) - Aron (’11) did not weigh the muscles in his experiments on dogs, but infers (p. 29) that: ‘‘Only the flesh, muscles and fat of the body remain as the tissues which must have lost during the course of the experiments.’’ From an analysis of samples taken from the leg muscles, he also concludes that ‘‘The muscles contained only one-half of the normal amount of solids,”’ the protein being greatly decreased and the water-content increased. Again, however, his comparison is with controls at the end rather than the beginning of the experiment, so that no conclusion can be drawn as to the changes taking place during the experiment in the animals held at constant body-weight. VISCERA AND ‘REMAINDER’ With the visceral group (table 9; fig. 4) have been included the brain, spinal cord and eyeballs, as well as the thoracic and abdominal viscera. According to Jackson and Lowrey ’12, this group decreases from about 21 per cent of the body at three weeks to about 16 per cent at ten weeks of age. This is in fairly close agreement with the controls in the present series, except at 128 C. M. JACKSON ten weeks. In this case the controls are considerably too heavy for direct comparison with the animals under experiment, which accounts for the discrepancy. In the animals held at constant body-weight from the age of three weeks to the ages of six, eight and ten weeks, the visceral group shows a distinct increase in weight. This is more marked at six and eight than at ten weeks, which perhaps indicates that the viscera may increase in the earlier part of the experiment, and lose weight later. The experiment from six to thirty-two weeks indicates no essential change in the weight of the viscera. From ten to thirty-five weeks there is a slight gain. On the whole, it may be concluded that during constant body- weight in young albino rats the visceral group as a whole under- goes but little change in weight, with a slight tendency to in- crease, especially in the earlier periods. As will be seen later, however, the individual viscera differ greatly in their reactions. Aron (711) concludes that in young dogs held at nearly constant body-weight the organs in general do not lose weight. On account of the small number of observations, however, and the lack of adequate controls, it is difficult to draw any satis- factory conclusion from his observations upon the viscera. The ‘remainder’ is obtained by deducting from the net body- weight the weight of the integument, skeleton, musculature and viscera. It therefore includes loss by evaporation and escape of fluids, as well as a few small unweighed organs and the masses of dissectable fat. The data in table 9 show a considerable variation, as might be expected. On the whole, however, it appears doubtful whether there is any material change in the weight of the ‘remainder’ in young rats held at constant body- ‘ weight for considerable periods. There is undoubtedly a loss in the fat, but this is probably counterbalanced by an increased water-content of the interstitial connective tissues. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 129 BRAIN The brain (table 10) in eleven controls at three weeks of age averaged 1.282 grams, or 5.31 per cent of the (net) body-weight. This corresponds fairly closely with Donaldson’s (’08) figure for the normal rat of corresponding weight. In the rats held at constant body-weight from the age of three to six and eight weeks, there appears (table 10) to be a relative increase in the brain, especially at eight weeks, where it forms 6.62 per cent of the body. In reality, however, this relative increase is only apparent, and due to the fact that these animals began the ex- periment at a lower body-weight, corresponding to which the brain is relatively heavier. The (net) body-weights of the two . rats at eight weeks were respectively 18.1 grams and 17.8 grams. According to Jackson (13, p. 22), the brain normally reaches its maximum relative weight of about 6.7 per cent of the body when the body-weight is about 15 grams. Thus the final brain-weight of the rats held at constant body-weight from the age of three to eight weeks of age is almost exactly that to be expected if the brain-weight has remained constant. TABLE 10 The brain; average absolute weight, average percentage of net body-weight and range indicated a | es t © a Z Bie a TOG RELATIVE WEIGHT DESCRIPTION OF RATS = 3 : ° Ricca Rane: anaes xp ene): as aS o7- re) a Z =) Normal at 3 weeks (Donaldson ’08, table 1)........ 52 25, 0F | e285 5.14 CWontrolsiatreiweeksess es. socks rhe ec nele dacdaeten ll 24.5 1.282 (1.187-1.364) 5.31 (4.14-6.20) Body-weight constant 3 weeks (ageof3to 6 weeks) 7 22.1 1.195 (1.035-1.297) 5.44 (4.50-6.33) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 1.183 (1.180-1.186) 6.62 (6.60-6.63) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 1.267 (1.136-1.379) 5.30 (4.38-6.20) Normal at 6 weeks (Donaldson ’08)................ 42 45.0*| 1.441 3.20 Controliat 6: weeks. saya horns eae einene es. 42.4 1.373 3.23 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 1.478 (1.459-1.497) 3.14 (2.96-3.32) Normal at 10 weeks (Donaldson ’08)............... 34 75.0*| 1.559 2.08 Contralsiatd0lbweekstit-yiecs nck sda eae sens anes 6 134.0 1.579 (1.512-1.636) 1.21 (0.96-1.44) Body-weight constant 25 weeks (age ot 10to35 weeks) 3 77.8 1.646 (1.603-1.723) 2.12 (2.02-2.18) Controlsyatiozandisonweeks. 5s). chicos d.s = ae. 6 189.5 1.777 (1.657-1.890) 0.97 (0.78-1.24) * Gross body-weight. + Body-weight ot controls at 10 weeks too high for comparison. 130 C. M. JACKSON The brain weight has also apparently remained nearly constant in the large group held at constant body-weight from three to ten weeks of age. The average absolute weight is slightly less, but almost in correspondence with the body-weight, so that the average relative weight, 5.30 per cent, is nearly identical with that of the controls at three weeks, the beginning of the experi- ment. Interms of absolute weight, there is a very slight apparent decrease from 1.274 grams (1.282 grams, less correction® for difference in body-weight, which averages 24.5 grams in the three weeks controls and 24.0 grams at ten weeks) to 1.267 grams, a decrease of about 0.5 per cent in absolute weight. In the series held at constant body-weight from the age of six to thirty-two weeks, there is an apparent slight decrease in the brain from about 3.23 per cent to 3.14 per cent of the body, and in the ten to thirty-five weeks series a slight increase (from 2.08 to 2.12 per cent). Considering the small number of observa- tions and the normal variation, however, these apparent differ- ences do not appear to be significant. I would, therefore, con- clude from the data above cited that there is probably no appreci- able change in the weight of the brain in young albino rats held at constant body-weight for considerable periods. Hatai (04) experimented with a series of young rats with initial body-weights corresponding roughly to those of mine at the ages of six to ten weeks. By giving an unfavorable diet (starch and beef-fat) their body-weight was reduced on the average about 30 per cent. The brain in these cases had appar- ently lost in absolute weight, the average loss being about 5 per cent. These results, however, are of course not directly com- parable with those in which the body-weight has remained . constant. In a later experiment, Hatai (’08) by underfeeding with un- favorable diet retarded the growth of a series of five rats, begin- 3 It should be noted here as in other cases that the correction for organ-weight is not in exact proportion to the difference in body-weight. Allowance must be made for the change in the relative weight of the organ corresponding to the change in body-weight. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS Jeol ning at the age of 30 days, so that at 170 days their average weight was only 91.5 grams, while full-fed controls averaged 146.5 grams. By comparison with ‘second controls,’ younger rats of body-weight similar to the final weight of the stunted series, he found that in the stunted rats the brain-weight was practi- cally identical with that of normal rats of the same body-weight. In other words, the growth in brain-weight had been retarded in the same proportion as the body-weight. On this principle, if the body-weight were retarded so as to permit no growth at all, that is held at constant weight, we should expect practically no increase in weight of the brain. This is in agreement with my results, as above stated. More recently Donaldson (’11) has experimented with a larger series (twenty-two litters) of rats held at nearly constant weight (34 grams) from the age of thirty to the age of fifty-one days. In the rats held at constant body-weight, the brain weight averaged 7.7 per cent less than in full-fed controls of the same litters. No direct controls were taken at the beginning of the experiment, but from the normal growth formula it is estimated that the initial brain-weight was slightly less than that found in the retarded rats at the end of the experiment. This would indicate an increase of 3.6 per cent in the brain-weight, while the body- weight was held constant. The large number of observations lends weight to this conclusion, -although it would be strength- ened if direct controls were available at the beginning of the experiment. It may be noted that if Donaldson’s normal (Wistar reference tables) rather than the direct controls be taken as the basis for estimating the initial brain-weight in my three to ten weeks series, the result would indicate a gain similar to that found .by Donaldson in his series. On the whole, therefore, we may safely conclude that there is but very slight if any increase in the brain-weight of young albino rats held at constant body-weight for considerable periods of time. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 2 32, Cc. M. JACKSON SPINAL CORD When the relative weights are compared with the controls at the beginning of the experiment (table 11), or with the theoreti- eal normal according to Donaldson (’08) there appears a very decided increase in the spinal cord at all the age-periods during the experiment. Thus while the body-weight has been held constant from the age of three to that of ten weeks, the spinal cord has apparently increased from an average of 0.179 to 0.243 grams, an increase of about 36 per cent (or slightly more if the initial weight be decreased to correct for the difference of body- weight at three weeks, 24.5 grams, and ten weeks, 24.0 grams). This corresponds to an increase from 0.74 per cent to 1.02 per cent of the net body-weight. The increases at the other age- periods are equally striking. Donaldson (’11) in the experiments previously mentioned also found an increase in the weight of the spinal cord in rats held at body-weight of about 34 grams from the age of thirty days to that of fifty-one days. He does not estimate this increase exactly but from the normal weight of the cord at the beginning of the experiment (cf. Donaldson ’08, table 1) the weight must have increased from about 0.223 to 0.2498 grams, an increase of about 10.7 per cent. While this is not so striking as my results (per- haps in part because my experiments covered a longer period of time) it agrees in indicating ‘during constant body-weight a much stronger growth tendency in the spinal cord than in the brain. This is in agreement with the well-known fact that in general the normal post-natal growth of the spinal cord is rela- tively much more rapid than that of the brain. This growth of the spinal cord is apparently correlated with the increase in trunk-length (Donaldson). EYEBALLS An increase even more striking than that of the spinal cord is apparent in the eyeballs (table 12). In rats held at constant body-weight from the age of three weeks, the eyeballs increase from a relative weight of about 0.50 per cent of the body-weight to 0.64 per cent at six weeks, 0.82 per cent at eight weeks, and WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS TABLE 11 133 The spinal cord; average absolute weight, average percentage of net body-weight and range indicated % gq ee Oa DESCRIPTION OF RATS ) [CFP a og a (o) a % Normal at 3 weeks (Donaldson '08, table 1)........| 47 WOUETO] SAGs WEEKS ss g 5 Z Z Q Normal at 3 weeks (Jackson '13)................... 24 21.2 Controlstatro weeks arcs cache Gee eet eia alee siete 10 24.€ Body-weight constant 3 weeks (age of 3to 6 weeks) 6 22k Body-weight constant 5 weeks (ageof3to 8 weeks) 2 18.0 Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 Normal at 6 weeks (Jackson ’13).... .............. 42 50.0 WontroliatiGrweeks scr se ns aorecioe cess tice Reinet, 1 42.4 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 Normal at 10 weeks (Jackson ’13).................. 75.0 WONTTOIS AEH WEEKSi eh tassels siesta de ele meet 6 134.0 Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 Mae Controlstatio2/ and SosweOkSis ss. cece cies saci cise - 5 184.4 ABSOLUTE WEIGHT (AND RANGE): GRAMS 0.105 (0.073-0.125) 0.120 (0.110-0.133) 0.142 (0.138-0.149) 0.146 (0.145-0. 146) 0.179 (0.130-0.196) 0.153 (0.125-0.175) 0.152 0.240 (0.230-0.250) 0.173 0.209 (0.195-0.228) 0.275 (0.249-0.323) 0.260 (0.235-0.296) * Body-weight of controls at 10 weeks too high for comparison. RELATIVE WEIGHT (AND RANGE): PER CENT 0.52 (0.31-0.73) 0.50 (0.34-0.69) 0.64 (0.60-0.72) 0.82 (0.81-0.82) 0.76 (0.57-1.00) 0.32 (0.18-0.40) 0.36 0.51 (0.50-0.52) 0.23 0.16 (0.13-0.21) 0.35 (0.34-0.38) 0.15 (0.12-0.17) 134 Cc. M. JACKSON 0.76 per cent at ten weeks. In terms of absolute weight, the eyeballs have apparently increased from an average of 0.120 grams at three weeks to about 0.179 grams (no correction made for the slight difference in body-weight) at ten weeks, an increase of nearly 50 per cent! In the normal, full-fed rat at ten weeks (average body-weight 112 grams) the eyeballs have reached a weight of only about 0.201 grams (Jackson 713). At this rate, the weight of the eyeballs at a normal body-weight of 75 grams (the body-weight indicated in table 12 as ‘normal at 10 weeks.’ to correspond to the body-weight of the animals held at constant body-weight from the age of ten to thirty-five weeks) would be only about 0.173 grams, or slightly less than that actually reached in the series held at constant body-weight of 24 grams from three weeks to ten weeks of age. The growth of the eyeballs in rats held constant from the age of six to thirty-two weeks, and from ten to thirty-five weeks, is equally striking. No data upon the growth of the eyeballs under these conditions have been found in the literature. I have shown elsewhere (Jackson ’15 a, 715 ¢), however, that the eyeballs lose but very little if any during inanition in the adult albino rat. In connection with the astonishing growth capacity of the eyeballs in young animals at constant body-weight, the possi- bility that the growth of the eyeballs is somewhat independent of that in the body as a whole may be considered, which I have already pointed out (Jackson 713, p. 24). When the large water- content of the eyeballs is considered (85.6 per cent in the rat at twenty days, according to Lowrey 713), it is, after all, not diffi- cult to comprehend the possibility of its continued growth, largely by water-absorption, when growth in the body as a whole is at a standstill. THYROID GLAND In young albino rats held at constant body-weight from the age of three weeks to six weeks, eight weeks and ten weeks, there is_ usually a well-marked loss of weight in. the thyroid gland (table 13). In the largest group, three to ten weeks, the thyroid has apparently decreased on the average from about 0.033 per cent WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 135 TABLE 13 The thyroid gland; average absolute weight, average percentage of net body-weight and range indicated ; aa 3 Hg a 3 RELATIVE WEIGHT DESCRIPTION OF RATS ¢ é Ss & fae mice Canund 4 (AND RANGE): BE ie AND RANGE): GRAMS sai GEIS + m o> Az Z rs) + Normal at 3 weeks (Jackson ’13) ats 26 18.7 | 0.0060 (0.0034-0.0104} 0.030 (0.018-0.041) Controlsiatisaweeksees Gasser. ceeeen ie. n. 11 24.6 | 0.0078 (0.0064-0.0094| 0.033 (0.022-0.042) Body-weight constant 3 weeks (age ot 3to 6 weeks) 3 23.1 | 0.0053 (0.0052-0.0056 | 0.023 (0.022-0.024) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 | 0.0058 (0.0047-0 9068 | 0.032 (0.026-0.038) Body-weight constant 7 weeks (age of 3 to 10 weeks) 18 *24.1 | 0.0059 (0.0034—-0.0087 | 0.025 (0.016—-0.036) Normal at 10 weeks (Jackson ’13)................-- 24 110.0 | 0.0165 (0.0100-0.030) | 0.015 (0.008-0.022) of the body to 0.025 per cent. Or, in terms of absolute weight, it has decreased from 0.0078 to 0.0059 grams, a decrease of 24 per cent. (A slight correction should be made on account of difference in body-weight.) No observations were made upon the thyroid gland in the experiments beginning at later ages. During acute inanition in adult rats, the thyroid gland appar- ently loses little or no weight; while in chronic inanition with an average loss in body-weight of about 36 per cent, the thyroid gland loses only about 22 per cent in weight (Jackson 715). There is some uncertainty as to the exact figures, however, on account of variability and difficulty in dissecting out the thyroid gland in an accurate manner. The same, of course, holds true for the present series. THYMUS The normal thymus (table 14) at three weeks forms 0.387 per ° cent of the net body-weight. This decreases, in rats held at constant body-weight, to 0.075 per cent at six weeks of age, to 0.030 (exceptional?) at eight weeks, and to 0.040 per cent at ten weeks. In terms of absolute weight, the thymus has decreased from 0.091 grams at three weeks to 0.017 grams (loss of 81 per cent) at six weeks, and to 0.0094 grams (loss of 90 per cent) at 10 weeks. No correction for the slight difference in body-weight has been made in these estimates. Normally at ten weeks of age the weight of the thymus should have increased to about 0.24 grams (0.30 grams in the controls). 136 Cc. M. JACKSON A maximum absolute weight of about 0.29 grams is reached by the average normal thymus about the age of 85 days (Hatai 14). At one year, it has undergone a complete age-involution, and forms only 0.02 per cent of the body-weight (Jackson ’13). That the weight and structure of the thymus are markedly affected by various adverse conditions has long been known, and the process of invofution has recently been thoroughly investi- gated by Hammar and his pupils. Jonson (’09) experimented with young rabbits subjected to acute and chronic inanition. In the latter case, the diet was restricted so as to maintain the young rabbits at constant body-weight (similar to the present experiment with rats). Under these conditions, Jonson found the weight-curve of the thymus similar to that of the body-fat, although during acute inanition the fat decreases somewhat more rapidly. In young rabbits held at constant body-weight the thymus in four weeks is reduced to about one-thirtieth of its initial weight. The cortex suffers the greatest loss, being re- duced to one-twelfth of its initial weight within two weeks of chronic inanition at constant body-weight. It would therefore appear that the process of involution is much more rapid and complete in young rabbits than in young rats at the ages included in the present investigation. In both cases, however, the weight of the thymus in hunger involution decreases most rapidly in the earlier weeks of the experiment. HEART The heart (table 15) in the albino rats held at constant body- weight from the age of three weeks appears to have remained practically constant at about 0.70 per cent of the net body-weight up to the age of ten weeks. In absolute weight, the heart would apparently decrease from 0.167 grams (0.170 grams, less cor- rection for difference in body-weight) to 0.166 grams, a decrease of about.0.6 per cent, which is probably within the limits of experimental error. The very slight decrease at six weeks and increase at eight weeks are also probably not significant. -Simi- larly in the rats held at constant body-weight from six to thirty- WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 137 two weeks and from ten to thirty-five weeks the heart has ap- parently retained almost exactly its initial relative weight. The absolute weight therefore apparently remains unchanged (the differences in the table being due to different initial body-weights). During inanition in adult rats (Jackson 715 a, ’15 e) the heart likewise maintains its relative weight, losing in absolute weight nearly in proportion to the entire body (slightly more in chronic than in acute inanition). TABLE 14 The thymus; average absolute weight, average percentage weight and range indicated . me a RBRORRR TT CLT RELATIVE WEIGTH DESCRIPTION OF RATS ; e F 2 PERE acre (AND PANGS): ° e) vA a Normal at 3 weeks (Jackson ’13)................... 49 18.7 | 0.071 (0.034 -0.135) | 0.37 (0.23 -0.55) (Controle atioaveeks ny ee eeE nae eee eee ote Sec: il 24.5 | 0.091 (0.042 -0.123)| 0.37 (0.22 -0.51) Body-weight constant 3 weeks (age of3to 6weeks) 4 22.7 | 0.017 (0.011 -0.022)] 0.075 (0.046-0.0101 Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 | 0.0054 (0.0037—-0.0071)| 0.030 (0.021-0.040) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 | 0.0094 (0.0054-0.0170)} 0.040 (0.019-0.062) Normal at 6 weeks (Jackson '13)................... 42 50.0 | 0.108 (0.052 -0.284) 0.21 (0.14 -0.36) Normal] at 10 weeks (Jackson ’13).................. 42 107.2 | 0.24 (0.12 -0.44) 0.23 (0.13 -0.35) Controlsratil Onweelksie. cyanea ee A lok 6 134.0 | 0.30 (0.25 -0.34) 0-23 (0.16 -0.29) TABLE 15 The heart; average absolute weight; average percentage of net body-weight and range indicated x a DESCRIPTION OF RATS 6 2 5 Winee Nae Geaate of a oP | 6% V4 iso) Normal at 3 weeks (Jackson ’13)...............:.-- 49 18.7 | 0.135 (0.082-0.250) Controlsiatiomweeks = Ae cea cee tobe ose ll 24.5 0.170 (0.130-0.245) Body-weight constant 3 weeks (age of 3to 6 weeks) 7 22.1} 0.15) (0.123-0.187) Body-weight constant 5 weeks (age of 3to 8 weeks) yy 18.0 0.129 (0.128-0.130) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 | 0.166 (0.124-0.213) Normal at 6 weeks (Jackson ’13).................-- 42 50.0 | 0.277 (0.183-0.535) WontrolsratiOnweeks eae ties wis locos aetersisiers cs: 1 42.4| 0.237 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 0.271 (0.255-0.288) Normal at 10 weeks (Jackson '13)................-- 75.0 | 0.375 Control satel Olweelkss en. sence peices anacied teens: 6 134.0} 0.562 (0.489-0.€87) Body-weight constant 25 weeks (age of 10t035 weeks) 3 77.8 | 0.3889 (0.348-0.450) Controlsiatie2iand Sbiweeks.....-, <.0e oes<2e0600- 6 189.5 0.873 (0.688-1.29) * Body-weight of controls at 10 weeks too high for comparison. RELATIVE WEIGHT (AND RANGE): PER CENT 0.72 (0.56-0.93) 0.70 (0.55-0.84) 0.68 (0.63-0.85) 0.72 (0.72-0.72) 0.70 (0.55-0.83) 0.55 (0.44-0.68) 0 0.57 (0.57-0.58) 0.50 0.44 (0.40-0.48) 0.50 (0.47-0.52) 0.46 (0.37-0.56) 138 Cc. M. JACKSON LUNGS The lungs (table 16) in rats held at constant body-weight from the age of three weeks to the ages of six weeks and ten weeks appear to lose in weight. (The data at eight weeks are abnor- mally high, as noted in the table). Thus between the ages of three weeks and ten weeks, the lungs apparently decrease from 1.04 per cent to 0.91 per cent of the body-weight; or in absolute weight from 0.250 to 0.218 grams, a loss of about 15 per cent. (A slight correction has been made on account of the difference in body-weight). In the few observations upon rats for longer periods beginning at the later ages of six and ten weeks, there appears to be no material change in the weights of the lungs during the experimental periods. The lung infections frequently found in older and adult rats, rarely occur before the age of ten weeks, and thus do not affect the present series. As Hatai has already noted, rats during chronic inanition appear to be unusually free from lung infection. During inanition in adult rats, the lungs lose weight in about the same proportion as the whole body, thus nearly maintaining their relative (percentage) weight. The loss is slightly greater, however, during chronic inanition (Jackson 715 ¢). TABLE 16 The lungs; average absolute weight; average percentage of net body-weight and range indicated m a a3 RELATIVE WEIGHT DESCRIPTION OF RATS c é 2 S (GRD RANGE GER SS ae A : o> 5 Z Zz a Normal at 3 weeks (Jackson ’13).................. 49 18.7 | 0.216 (0.151-0.354) | 1.17 (0.86-1.46) Controlsiationwee ksi eh era eee ee eee 11 24.5 0.253 (0.201-0.290) 1.04 (0.8S-1.12) Body-weight constant 3 weeks (age of 310 6 weeks) 3 20.9 | 0.192 (0.190-0.197)} 0.92 (0.88-0.93) Body-weight constant 5 weeks (ageof3to 8 weeks) iD 18.0 | 0.280* (0.241-0.319) | 1.56* (1.36-1.77) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 | 0.218 (0.170-0.282)} 0.91 (0.78-1.32) Normal at 6 weeks (Jackson 713)................... 39 50.0 | 0.3383 (0.244-0.547)} 0.68 (0.58-0.94) ControllatiGtweeks: 25.5 5. seee oe coe aha hee 1 42.4] 0.264 0.62 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 0.319 (0.309-0.329) je 0.68 (0.65-0.70) Normal at 10 weeks (Jackson ’13).................. 75.0 | 0.49 0.65 Gontrolsiath Olweeksan < eeees se ec nee ee ee 6 134.0 0.74 (0.610-0.94) 0.56 (0.45-0.74) Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 77.8 0.49 (0.390-0.60) 0.63 (0.52-0.81) * Lungs in one case abnormally heavy on account of post mortem congestion. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 139 LIVER The liver (table 17) in the albino rats held at constant body- weight from the age of three weeks to six, eight and ten weeks shows an apparent increase. This increase appears greatest in the shortest series, three to six weeks, increasing from 4.95 per cent of the body-weight (in controls) to 5.89 per cent. Be- tween three and ten weeks the corresponding increase is from an absolute weight of 1.16 grams (1.20 grams, less correction for difference in body-weight; or 4.95 per cent of the body-weight) to 1.28 grams (5.25 per cent of the body-weight), or an increase of about 10.3 per cent in absolute weight. In the later and longer periods, however, there appears to be a decided decrease in the weight of the liver (six to thirty-two weeks and ten to thirty-five weeks series). This may be due to the fact that the latter experiments extended over a longer period of time, or it may be because they were begun at a later age. We may therefore conclude that in young albino rats held at constant body-weight there is, beginning at three weeks, an increase in the weight of the liver (apparently greater at six than at ten weeks of age) ; while in rats beginning the experiment later, at six and ten weeks of age (and extending over a longer period) TABLE 17 The liver; average absolute weight, average percentage of net body-weight and range indicated i = | ee So : mn 2 2 z RELATIVE WEIGHT DESCRIPTION OF RATS ss S . S aN ECERESS (AND RANGE): o& » PER CENT 5° 5 Z Zz Es Normal at 3 weeks (Jackson 713) ................... 49 18.7 0.87 (0.42-2.08) 4.50 (3.21-5.82) SSONETOIS A ORWEEKS a Gonna ees mete ace cleern ne 11 24.5 | 1.20 (1.06-1.38) 4.95 (4.03-6.00) Body-weight constant 3 weeks (age of 3to 6 weeks) 7 22.1 1.30 (0.89-1.63) 5.89 (4.32 -7.39) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 | 0.92 (0.78-1.06) 5.12 (4.36-5.88) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 1.28 (0.80-1.82) 5.25 (3.63-7.00) Normal at 6 weeks (Jackson ’°13)..................- 42 50.0 3.19 (2.18-4.76) 6.48 (5.39-8 46) CWontrollat Giweekss oe macd.ce oracisoestassoancess 1 42.4] 2.18 5.15 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 2.33 (2.21-2.45) 4.44 Normal at 10 weeks (Jackson ’13).................. 75.0 | 4.50 6.00 Won iro lst Ol weeks -o5. secs coe a eoac eke dercins @ 6 134.0 6.72 (4.91-8.66) 4.99 (4.28-5.72) Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 77.8 3.11 (2.30-3.51) 4.25 (3.91-4.74) (Controls/atise AUG ton WOCKS 6.5.2.) /s elo’ san af. era's 6 189.5 7.27 (5.84-9.86) 3.84 (3.47-4.14) 140 Cc. M. JACKSON there is a decided decrease in the weight of the liver. Thus the liver would appear under these conditions to have a tendency to follow the normal growth-impulse, which increases to a maximum at about the age of six weeks, and decreases thereafter (Jackson "NS D2 ols During inanition in adult rats, the liver loses in weight rela- tively more than the whole body, and toa greater extent in acute than in chronic inanition (Jackson 715 ¢). As the liver is nor- mally subject to great variation in weight, however, (Jackson ’13) great caution must be observed in drawing final conclusions. Hatai (713) found the weight of the normal liver distinctly higher than that in my series, and if his data instead of my controls were taken as a basis for the initial weights, the estimated losses in my experiments would be considerably greater, even the younger rats showing a loss instead of a gain in liver-weight. SPLEEN Taking the controls at three weeks as a basis for com- parison, it appears that in rats held at constant body-weight from the age of three weeks to six weeks the weight of the spleen (table 18) remains practically unchanged (the average at six TABLE 18 The spleen; average absolute weight, average percentage of net body-weight and range indicated & a a ng z Z : ABSOLUTE WEIGHT AOE AN TAA 72) NOME DESCRIPTION OF RATS ; : F : (om NCE aca GND BANG: oe | 6% z, i Normal at 3 weeks (Jackson ’18)................... 49 18.7 0.055 (0.019-0.145) 0.28 (0.15-0.42) Gontrolspatronweeks sain on ter ouic ene. ssa Ao enEnE 11 24.5 | 0.091 (0.051-0.130) 0.37 (0.27-0.48) Body-weight constant 3 weeks (age of 3to 6 weeks) 7 22h 0.087 (0.040-0.172) 0.38 (0.18-0.72) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 | 0.043 (0.036-0.049) 0.24 (0.20-0.27) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 | 0 053 (0.033-0.076) 0.22 (0.16-0.33) Normal at 6 weeks Gackson 713) .........-:....<- 42 50.0 | 0.1385 (0.086-0.204) 0.28 (0.19-0.47) ControliatiGsweeksi-.p- oade tise te ANH ORIES 5 5 6 1 42.4 0.107 0.25 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 0.139 0.29 (0.28-0.30) Normal at 10 weeks (Jackson 713).................. 75.0 0.225 0.30 ControlssathlOiweelks = .o5 ete isa ees, ever 6 134.0 0.350 (0.300-0.420) 0.27 (0.21-0.36) Body-weight constant 25 weeks (age of 10 to35 weeks) 3 77.8 0.230 (0.230-0.240) 0.30 (0.27-0.33) Controlstatrezjandisbiweeksmisensen 4.1152 sens eee 6 189.5 | 0.620 (0.480-0.750) 0.33 (0.29-0.39) WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 141 weeks being probably too high, on account of the inclusion of a litter with abnormally large spleens); while in the rats at eight and ten weeks there is a considerable decrease. In the case of the three to ten weeks series, the decrease would be from 0.091 gram (0.37 per cent of the net body-weight) to 0.053 gram (0.22 per cent of the body-weight), or a decrease of nearly 42 per cent in absolute weight. (No correction has been made for the difference in body-weight, which is 0.5 gram lower at ten weeks). If the normal relative weight of the spleen at three weeks (0.28 per cent, Jackson 713) be taken as a basis of esti- mate, however, the loss would appear considerably less. In the rats held at constant body-weight at later and longer periods (six to thirty-two weeks and ten to thirty-five weeks) there appears to be no appreciable change in the average weight of the spleen. It may be concluded therefore that in young rats held at constant body-weight beginning at the age of three weeks there is a marked tendency to a reduction in the weight of the spleen; while at later (and longer) periods the spleen ap- pears to undergo no material change in weight. It must be remembered, however, that the spleen is normally one of the most variable organs in the body (Jackson ’13), and final conclusions should be correspondingly guarded. In adult albino rats during chronic inanition the average loss in weight is nearly proportional to that of the entire body, while in acute inanition the loss appears very much greater (Jackson pliona.y Loe): STOMACH AND INTESTINES The stomach and intestines, including mesentery and pan- creas, are considered both with contents (table 19 a) and empty (table 19b). Considering first the empty canal, it appears in rats held at constant body-weight from the age of three weeks to increase from about 4.8 per cent of the body-weight to 8.0 per cent at six and eight weeks, decreasing to 6.0 per cent at ten weeks. In absolute weight the increase would be from about 1.13 grams (1.20 grams less correction on account of difference in body-weight) at three weeks to 1.45 grams at ten weeks, an * 142 Cc. M. JACKSON increase of about 28 per cent. Between six and thirty-two weeks there is apparently but little change; while from ten to thirty-five weeks there appears to be a decrease in the weight of the alimentary canal. The number of observations, however, is too small for final conclusion. The most remarkable increase occurs between three and six weeks of age, from an absolute weight of 1.20 to 1.78 grams, an apparent increase of about 48 per cent in absolute weight! (The increase would be even greater if allowance were made for differ- ence in body-weight). Thus, as in the case of the liver and TABLE 19 The stomach and intestines; average absolute weight, average percentage of net body- weight and range indicated f Eg mo S = ELATIVE WEIG DESCRIPTION OF RATS og : : (ae RA GEE enaE z LAND eee 38 ps 6 PER CENT oF aa 4 AQ a. Including contents Normal at 3 weeks (Jackson ’13)............-.+.+-- 49 18.7 1.78 (0.74-3.08) 9.3 (5.5-15.5) Controlsiationweekseseere eer ek eisteletetetelebareratellefeteerere uit 24.5 2.51 (1.96-3.23) 10.4 (7.8-15.4) Body-weight constant 3 weeks (ageof3to 6 weeks) 7 22.1 3.10 (1.67-5.07) 13.9 (7.4-21.3) Body-weight constant 5 weeks (age of3to 8 weeks) 2 18.0 3.64 (2.58-4.71) 20.2 ( 4.5-26.0) Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 3.29 (2.26-6.83) 13.5 (10 .3-20.2) Normal at 6 weeks (Jackson 713) ...............-.- 42 50.0 8.05 (4.48-13.40) 15.9 (10.6-23.4) Gontrolsiat Onmeeks saece eer tecieecieeeilee-lel-ei- = Sele 1 42.4 balls 12.1 Body-weight constant 26 weeks (age of 6to 32 weeks)| 2 47.1 5.98 (4.71-7.24) 12.5 (10.8-14.2) Normal at 10 weeks (Jackson 713) .................- 75.0 9.00 12.0 Controls at hOmweekstennemetr its lateteleciel state ousiate lieve 6 134.0 11.32 (8.82-14.49) 8.3 ( 7.3-11.2) Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 77.8 7.48 (5.$0-8.81) 9.6 ( 8.1-10.4) Controlsiatiscand: sp iweekse ema eme tcl) ser e)le ele el-ieiete 6 189.5 11.60 (7.51-14.04) 7.4 ( 5.4-12.2) b. Empty Normal at 3 weeks (Jackson 18)..................- 16 20.0 0.90 (0.37-1.61) 4.5 ( 2.9- 6.1) ControlsiatrouweckSs eerie eaeleoitel veers 10 25.1 1.20 (0.74-1.69) 4.8 ( 3.1- 6.7) Body-weight constant 3 weeks (ageof 3to 6 weeks) 7 22.1 1.78 (1.41-2.51) 8.0 ( 6.3-11.0) Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 1.44 (1.38-1.51) 8.0 ( 7.8- 8.3) Body-weight constant 7 weeks (age of 3to 10 weeks) | 19 24.0 1.45 (0.96-2.29) 6.0 ( 4.5- 8.6) Normal at 6 weeks (Jackson ’13)................--- 50.0 4.00 8.0 Control stationvee ks aye arei i ieieieicle sietolel> clei siariieae 1 42.4 3.00 eal Body-weight constant 26 weeks (ageof 6to 32 weeks)| 2 47.1 3.22 (2.85-3.59) 6.8 ( 6.5- 7.0) Normal at 10 weeks (Jackson 713).................. 75.0 5.30 7.0 Controlsixtel0sweeks*s. se eaper caer ieclic = cities eer 5 127.9 5.26 4.3 (3.9- 4.9) Body-weight constant 25 weeks(age of 10 to 25 weeks) 3 77.8 3.97 (3.48-4.35) 5.4 ( 4.7- 5.9) Controlsiatice andiso weekssene see scrmes- sie aeeeer 6 189.5 9.16 (7.57-10.64) 4.9 ( 4.1- 5.3) ee * Body-weight of controls at 10 weeks too great for comparison with those held constant 10-35 weeks. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 143 following the normal growth tendency, there appears to be in these retarded rats an early increase in the weight of the ali- mentary canal, reaching a maximum at about the age of six weeks, after which there is a decline in weight. It may be that the early increase in these organs is diminished later on account of the exhaustion of the stored food-supplies available else- where in the body at the beginning of the experiment. Individ- ual organs and tissues differ greatly from each other in their relative susceptibility to attack and absorption at different times during the course of inanition. The behavior of the stomach and intestines weighed with contents (table 19 a) is very similar to that of the empty canal. Contrary to what might be expected, during constant body- weight, with a restricted food-supply (water ad libitum), the canal does not decrease in contents. On the contrary there is an increase in contents (watery or mucous in character) which may even exceed in relative weight the normal contents in full- fed aminals. The maximum occurred in animals held at con- stant body-weight from the age of three to eight weeks, when the canal with contents formed 20.2 per cent of the body-weight! In experiments beginning at later ages and extending over longer periods (six to thirty-two weeks and ten to thirty-five weeks), the canal with contents appears to remain more nearly uniform in weight, with some tendency to decrease. During both acute and chronic inanition in adult albino rats, there is a very marked decrease in the weight of the stomach and intestines, both with and without contents (Jackson ’15 a, 715 ¢). SUPRARENAL GLANDS The suprarenal glands (table 20) from the age of about six weeks must be considered separately in the sexes, on account of a distinct sexual difference in their weight, as discovered independently by Hatai (13) and myself (Jackson 713). In the earlier ages, however, there is no apparent sexual difference, hence the sexes are combined. 144 Cc. M. JACKSON TABLE 20 The suprarenal glands; average absolute weight, average percentage of net body-weight and range indicated ABSOLUTE WEIGHT (AND RANGE): GRAMS 0.0072 (0 .0040-0.0140) 0.0088 (0 .0060-0 .0108) 0.0089 (0.0070-0 0110) 0.0109 0.0088 0.0101 (0.0072-0.0126) 0.0117 (0.0090-0.0147) 0.0128 (0.0070-0.0190) 0.0090 0.0105 0.0090 0.0220 (0.0150-0.0336) 0.0253 (0.0180-0 .0329) 0.0295 (0.0270-0.0326) 0.0318 (0.02Y0-0.0346) 0.0149 (0.0147-0.0150) 0.0145 0.0270 (0.0180-0.0310) 0.0320 (0.0280-0.0360) RELATIVE WEIGHT (AND RANGE): PER CENT 0.0400 (0.023-0.074) 0.0370 (0.025-0 .055) 0.0400 (0.034-0.050) 0.0610 0.0490 0.0420 (0.033-0.052) 0.0510 (0.040-0.061) 0.0260 (0.015-0.039) 0.0210 0.0240 0.0180 0.0185 (0.011-0.025) 0.0257 (0.017-0.033) - 0.0190 (0.017-0.023) 0.0280 (0.027-0.031) 0.0190 (0.017—-0.020) 0.020 0.0100 (0.009-0.013) 0.0200 (0.017-0.024) The kidneys; average absolute weight, average percentage of net body-weight and range 5 a is aS = 5 2 qm ac oa ae DESCRIPTION OF RATS =O Bo og] se pope. z Fs Normal at 3 weeks* (Jackson ’13)................. 49 18.7 Controls’atiS weeks*. eaeeeecstoncn evecare oon soa ces 11 24.5 Body-weight constant 3 weeks* (age of 3 to 6 weeks) tii 22/0 Body-weight constant 5 weeks (ageof3to 8 weeks) a ae Body-weight constant 7 weeks (age of 3 to 10 weeks) J an air Normal at 6 weeks* (Jackson ’713).................. 42 50.0 Controltati Giweeks aepeprice. veo aon Ade close eee baie lf 42.4 Body- weight constant 26 weeks (age ot 6 to 32 weeks)|< a a ZA . cee 20m} 117.0 Normal at 10 weeks (Jackson 713).................. 03f 99.0 3m| 154.7 ControlsatelOsweeksteacce css ae. | 64 Z a Normal at 3 weeks (Jackson 713)................... 49 18.7 Controls 'atisaweeks) -mccosas cele ce veils cick arciein rte eee 11 24.5 Body-weight constant 3 weeks (age of 3to 6 weeks) 7 22/1 Body-weight constant 5 weeks (age of 3to 8 weeks) 2 18.0 Body-weight constant 7 weeks (age of 3 to 10 weeks) 19 24.0 Normal at 6 weeks (Jackson '13)................... 42 50.0 *Controlaatibiweeks.ne, .csntess ences tose ise eee eee 1 42.4 Body-weight constant 26 weeks (age of 6 to 32 weeks) 2 47.1 Normal at 10 weeks (Jackson ’13)............... Bric 75.0 ControlsiatulOiweeksts..octtee memes acts sensitive 6 134.0 Body-weight constant 25 weeks (age of 10 to 35 weeks) 3 77.8 Controisiatrazand wouweeksse sp eceme cae: ae eee 6 189.5 ABSOLUTE WEIGHT (AND RANGE): GRAMS 0.271 (0.169-0.531) 0.393 (0.314—-0.487) 0.396 (0.328-0.487) 0.339 (0.331-0.347) 0.404 (0.325-0.515) 0.616 (0.500-0.943) 0.570 ‘ 0.613 (0.581-0.645) 0.750 1.200 (0.953-1.481) 0.753 (0.728-0.784) 1.539 (1.275-1.909) RELATIVE WEIGHT (AND RANGE): PER CENT .44 (1.19-1.87) .62 (1.33-2.16) 80 (1.46-2.05) .89 (1.86-1.92) .69 (1,441.97) .25 (1.00-1.55) 35 1.30 (1.28-1.33) 1.00 0.91 (0.81-0.99) 0.97 (0.92-1.02) 0.81 (0.71-0.86) a ee ero a oy * Controls too heavy for comparison. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 145 In the rats held at constant body-weight from the age of three weeks the suprarenal glands appear to increase slightly in weight at six and eight weeks; and especially at ten weeks, where the characteristic sexual difference has distinctly appeared. In absolute weight, the suprarenal glands have increased from 0.0090 gram (0.0088 gram, plus correction for difference in body- weight; or 0.087 per cent of the body-weight) at three weeks to 0.0101 gram (0.042 per cent of the body-weight) in the male at ten weeks, an increase of about 12 per cent in absolute weight. In the female, the corresponding increase is from 0.0084 gram (0.0088 gram, less correction for difference in body-weight) (or 0.037 per cent. of the body-weight) to 0.0117 gram (0.051 per cent of the body-weight), an increase of 39 per cent in absolute weight. Normally the suprarenals during this period are de- creasing in relative (percentage) weight. In the experiments at later and longer periods (six to thirty-two weeks and ten to thirty-five weeks), the changes are apparently not great, but a larger number of observations is necessary before definite con- clusions can be reached. In adult rats during inanition there is little or no loss in the absolute weight of the suprarenal glands, which therefore in- crease markedly in relative (percentage) weight (Jackson ’15 a, Bh ecG))). KIDNEYS The kidneys (table 21) in rats held at constant body-weight from the age of three weeks show a tendency to increase which is more marked at six and eight than at ten weeks. Between three and ten weeks of age the increase is from an average of 0.388 gram (0.393 gram, less correction on account of differ- ence in body-weight; or 1.62 per cent of the body-weight) to 0.404 gram (1.69 per cent of the body-weight), an increase of only about 4.1 per cent in absolute weight. At later and longer periods (six to thirty-two weeks and ten to thirty-five weeks) there is apparently but little change in the weight of the kidneys. The slight differences shown in the table are probably not significant. On the whole, it appears that in young rats held at 146 Cc. M. JACKSON constant body-weight there is a slight tendency to increase in the weight of the kidneys, in the earlier weeks, but little or no apparent difference later. In adult rats during acute and chronic inanition, the kidneys lose in weight relatively slightly less than the body as a whole, thereby gaining slightly in relative (percentage) weight (Jackson 715 ¢). TESTES AND EPIDIDYMI The weights of testes and epididymi unfortunately were not separated in some cases, which (together with their variability and the small number of observations) makes conclusions some- what difficult. For the testis (table 22 a), the clearest case is in rats held constant from three to ten weeks of age, in which TABLE 22 The testes and epididymi; average absolute weight, average percentage of net body-weight and range indicated R | Ee : ao Qs RELATIVE WEIGHT DESCRIPTION OF RATS e 6 2 5 (inp BANGe Gee (AND RANGE): o& pi PER CENT sé | 6% Zz Q a. Testes | Normaliat ouweeksi(blacaiels) an cose cee cece 20.0*] 0.090 0.45 Normal at 3 weeks (incl. epididymi) (Jackson 713)| 24 21.2 | 0.134f (0.07€-0.224) 0.63* (0.53-0.78) Controlsvat sameeks tes metre ecteela 41ers oie viele etees 6 25.0 | 0.144 (0.114-0.200) 0.57 (0.49-0.62) Body-weight constant 3 weeks (age of 3to 6 weeks) 2 22.1 | 0.133+ (0.129-0.137) 0.66* (0.65-0.67) Body-weight constant 5 weeks (age of 3to 8 weeks) 1 18.1 | 0.176T 0.98* Body-weight constant 7 weeks (age of 3 to 10 weeks) 7 25.2 | 0.193 (0.106-0.331) 0.74 (0.49-1.05) Normal atiGuwecksi(Elatal elo) sees re re ‘ 50.0*} 0.402 0.80 : Normal at 6 weeks (Jackson 713)..................- 20 49.0 | 0.592 (0.368-0.958) 1.19* (0.88-1.72) Body-weight constant 26 weeks (age of 6 to 32 weeks) 1 43.7 | 0.370T 0.84* Normal atelObweeks G@liataled3))erer c. sei = oc ates 70.0*| 0.774 ail Normal at 10 weeks (Jackson 713)..................] 20 117.0 | 1.747¢ (0.678-2.62) 1.51* (0.€3-2.41) ControlstatelObweeks ser acer ck iNe secre etree eral eh ena 3 154.7 | 1.850 (1.608-2.10) 1.217 (1.06-1.49) Body-weight constant 25 weeks (age of 10 to 35 weeks) 2 79.6 | 1.2944 (1.401-1.188) 1.62* (1.61-1.64) Gontrolstaties and so nweeks:eteesse sia. = deen 3 218.7 | 1.844t (1.766-1.932) 0.84* (0.74-0.90) * b. Epididymi Normal at 3 weeks (Jackson ’13)............. (estimated) 20.0 | 0.200 (?) 0.10 (?) Controlstatesiweeks cemeteries aes eres eeleree ae 6 25.0 | 0.0209 (0.0092-0.0246 | 0.084 (0.039-0.123) Body-weight constant 3 weeks (age of 3 to 6 weeks) 1 23.6 | 0.0144 0.061 Ms Body-weight constant 7 weeks (age of 3 to 10 weeks) 7 25.2 | 0.0209 (0.0084—-0.0360)} 0.080 (0.040-0.115) ControlsatlOGweekss.. 5 deere cree se ees cn ese 3 154.7 | 0.4700 (0.4480-0.4920)} 0.310 (0.290-0.330) * Gross body-weight. + Including epididymi. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 147 there is an increase from 0.144 gram (0.57 per cent of the body- weight) to 0.193 gram (0.74 per cent of the body), an apparent increase of 34 per cent in the absolute weight. (The average net body-weight was about the same in both eases, 25.0 grams in the controls, and 25.2 grams at ten weeks). For the epididymi (table 22 b), the conclusions are even more uncertain; but there appears to be a slight loss in the relative weight of the epididymi in rats held at constant body-weight from the age of three to six and ten weeks. For testis and epididymis combined (table 22 a) there appears to be an increase in rats held at constant body-weight, excepting the period from six to thirty-two weeks. While no final conclusions can be drawn, the evidence indi- cates that in young rats held at constant body-weight there is an increase in the weight of the testis, but not in the epididymis. During inanition in adult rats, the testes and epididymi ap- parently lose weight in about the same proportion as the entire body (Jackson 715 ¢). OVARIES In rats held at constant body-weight from three weeks to ten weeks of age, there would appear to be a decrease in the weight - of the ovaries (table 23) from 0.0066 gram (0.0068 gram, less correction for difference in body-weight; or 0.027 per cent of the TABLE 23 The ovaries; average absolute weight, average percentage of net body-weight and range indicated : a : 2 RELATIVE WEIGHT x O4 Bl ABSOLUTE WEIGHT 1B): DESCRIPTION OF RATS s : E : (AND RANGE): GRAMS (AND BANG); Ge Bz Z - Normal at’ 3 weeks (Jackson) "13).............--.--.- 24 16.2 | 0.0036 (0.0015-0.0067)} 0.0220 (0.011-0.0389) ControlsvatiSsweeksiaen tu. cbicmcimnee ent luis ae 5 24.5 | 0.0068 (0.0029-0.0104)| 0.0270 (0.012-0.034) Body-weight constant 3 weeks (age ot 3to 6 weeks) 3 22.4 | 0.0067 (0.0048-0.0084)|} 0.0300 (0.020-0.038) Body-weight constant 7 weeks (age ot 3 to 10 weeks) 12 23.3 | 0.0048 (0.0030-0.0064)) 0.0210 (0.012-0.028) Normal at 6 weeks (Jackson ’13)................... 20 54.9 | 0.0106 (0.0050-0.028) | 0.0216 (0.012-0.035) Normal at 10 weeks (Jackson ’13)........ Pyaoan Seee 23 98.8 | 0.0350 (0.0100-0.070) | 0.0340 (0.013-0.055) Controls ath Okweckseerer yy sane eee eRe on. 3 114.0 | 0.0237 (0.0297-0.038) | 0.0290 (0.027-0.033) THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 3 148 Cc. M. JACKSON body-weight) to 0.0048 gram (0.021 per cent of the body-weight). This would indicate a decrease of about 27 per cent in absolute weight. The ovary, however, is an organ which (like the thyroid gland) is quite variable and somewhat difficult to dissect out with accuracy, so a much larger number of observations would be required for a final conclusion. HYPOPHYSIS In the hypophysis (table 24) there is normally a sexual differ- ence in weight, observable in rats above 50 grams in body-weight (Hatai 13). As in the case of the suprarenal glands, the hypo- physis normally becomes relatively heavier in the female. In the rats held at constant body-weight from the age of three weeks to six weeks, there is what appears to be a sexual differ- ence, the one male with a hypophysis of 0.0016 gram (0.0068 per cent of the body-weight), while the two female hypophyses each weigh 0.0018 gram (0.0079 per cent of the body-weight). These are probably mere accidental variations, however, as in the three to ten weeks series the difference in the relative weight is insignificant (0.0084 per cent of the body-weight in the males, and 0.0086 per cent in the females). In any event, however, TABLE 24 The hypophysis; average absolute weight, average percentage of net body-weight and range indicated Dn oe OQ 2 =< RELATIVE WEIGHT om Sm ABSOLUTE WEIGHT DESCRIPTION OF RATS fe) zo cane (AND RANGE) & S fe (AND RANGE): GRAMS Sa OSNT ac ae o”7 ra) G Z ise) Normaliatiosweeks* (Hata). ...........ceneee 25.07|0.00175 0.0070 Controlsiatisoweekss eee tees ercick oases ccs eeeeene 9 25.5 |0.00178 (0.0012-0.0022)| 0.0070 (0.0052-0.0084) : fim. | 23.6 |0.0016 0.0068 Body-weight constant 3 weeks (age of 3to6 weeks) \ at. 22.9 10.0018 0.0079 (0.0076-0.0082) ney. ! peat 7m. | 25.2 |0.0024 (0.0016-0.0024) | 0.0084 (0.0057-0.0110) Body-weight constant 7 weeks (age of 3 to 10 weeks) a 23.6 |0.0020 (0.0010-0.0033) | 0.0086 (0.0043-0.0135) Ragas m. 120.0 |0.0054 0.0042 Normmaliat 10/weeksi(Hatarle)e... 2.2. coe eeeeeee a 130.0 10.0084 0.0065 Gontroleaeel0lweeks 3m. |154.7 |0.0062 (0.0059-0.0067) | 0.0040 (0.0038-0.0042) en atgeita a MRR Vor, ou ea 3f. 114.0 |0.0054 (0.0046-0.0067) | 0.0048 (0.0040-0.0062) . * No sexual difference apparent. + Gross body-weight. a WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 149 there appears to be a distinct tendency to an increase in the weight of the hypophysis in rats held at constant weight from three to ten weeks of age. In absolute weight, this corresponds in the male to an increase from 0.00178 gram to 0.0021 gram, an increase of about 18 per cent. In the female, the corre- sponding increase is from 0.00168 gram (0.00178 gram, less cor- rection for difference in body-weight) to 0.0020 gram, an increase of 19 per cent. As already shown, the suprarenal glands differ from the hy- pophysis in that they undergo a marked sexual differentation in weight during their growth while the body-weight is held con- stant. The impulse to sexual differentiation in these glands therefore appears stronger in the suprarenals than in the hy- pophysis. Final conclusions should be guarded, however, until a larger number of observations is available. In the adult rat during inanition the weight of the hypophysis decreases in nearly the same proportion as the body-weight (Jackson 715 ¢). DISCUSSION With reference to their growth tendency in young rats held at constant body-weight, the organs may be divided into three classes: (1) those in which the growth tendency is so strong that they continue to increase, even when the body-weight is held constant; (2) those which approximately hold their weight con- stant under these conditions; and (8) those which are unable to maintain themselves and lose in weight. According to this scheme, the organs in rats held at constant body-weight from the age of three to that of ten weeks are grouped in table 25. No grouping of this sort can be entirely satisfactory, beeause in some cases organs are intermediate in position, and especially because (as has been shown) in many ~ eases the weight of an organ will vary according to the age at which the experiment was begun and the length of the period. For example, the liver, which in the three-to-ten weeks experi- ment shows but a slight gain, shows a larger gain at earlier times, and a loss at later and longer periods. . 150 Cc. M. JACKSON On the whole, however, the grouping suffices to give a good view of the results. Group I shows the organs with marked growth tendency to be the skeleton, eyeballs, spinal cord, ali- mentary canal, testes, hypophysis and suprarenal glands. Group II, which approximately maintains constant weight, includes the musculature, brain, heart, kidneys and liver. Group III, those which fail to maintain their weight when the body-weight is held constant, includes the integument, lungs, thyroid gland, ovaries, spleen and thymus. In the same table 25 for convenience of comparison are also grouped the various organs of the adult rat according to their relative loss in weight during chronic inanition (Jackson ’15 ¢). TABLE 25 Comparison of growth tendency in young rats held at constant body-weight from age of three to ten weeks with tendency to maintenance in adult rats during chronic inani- tion. The figures indicate the apparent average percentage gain or loss in absolute weight during the period of experiment GROUP I “GROUP II GROUP II Marked tendency to growth in Weight nearly con- Marked loss in weight young held at constant body- | stant in young held | in young held at constant weight. at constant body- | body-weight. Strong tendency to maintenance | weight. Relative loss greater in adult chronic inanition Relative loss sim- | than that of entire body ilar to that of entire | during adult chronic in- body during adult | anition chronic inanition per cent of per cent per cent” change of change of change *eyeballs +50.0 | liver +10.3} lungs —15.0 *spinal cord +36.0 | *kidneys +4.1 | thyroid gland —24.0 testes +34.0 | *musculature +3.0 | ovaries —27.0 Vouns aticonetent | *skeleton +28.0} brain —0.5 | integument — 36.0 body eine ae {| alimentary canal +28.0 | *heart —0.6 spleen —42.0 4suprarenal glands ee se . sees 7Eee F. +39.0 é M. +18.0 hypophysis 7 +19.0 {| *skeleton +1.8 hypophysis —25.3 | liver — 43.0 Adults during *spinal cord —4.0 | *kidneys —26.8 | alimentary canal! —57.0 chronic inanition *eyeballs —5.8 spleen —29.0 (oss of body- } brain —6.6 | *heart —32.8 weight 36 per || *suprarenal glands —8.9 integument —38.5 cent) thyroid gland —21.8 lungs —40.0 thymus (2) testes —40.3 *musculature | —40.8 | ee * Indicates correspondence between the groupings in young and adult series. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 151 This enables us to answer the question as to whether the organs in young animals held at constant body-weight (and thus sub- jected to a chronic inanition) behave in a manner similar to adult rats during chronic inanition. In many cases, the group- ing shows an agreement. The skeleton, eyeballs, spinal cord and suprarenal glands have a strong growth tendency in young held at constant body-weight, and also a strong tendency to main- tain their original weight in adults: subjected to chronic inani- tion. The musculature, heart and kidneys approximately main- tain their weight in young held constant, and maintain their relative weight in adults during chronic inanition. In the major- ity of cases, however, the behavior of organs in the young differs materially from that in the adult. Thus the alimentary canal has a marked growth in the young at constant weight, yet it loses heavily during adult inanition. The converse is apparently true of the thyroid gland. To a greater or less degree, this inconsistency is seen in the case of most of the organs, as is evident from table 25. This inconsistency is perhaps to be explained in the following manner. In the adult during inanition the various organs lose weight relatively in inverse ratio to the ability of their cells to extract nutrition from the diminishing quantity available in the surrounding medium and to maintain equilibrium under these adverse conditions. In the young animal held at constant body- weight, the conditions differ in that the corresponding cells have the capacity not only to maintain themselves, but to grow. The growth capacity, as is well known, is different from and to some extent independent of the maintenance capacity. The adult has lost the capacity to grow, whereas the young animal has both the powers of growth and maintenance. Hence their organs behave differently during inanition. The alimentary canal, for example, apparently has a strong growth tendency, but-a weak power of maintenance. . It is further evident, however, that even the growth capacities of the various tissues and organs differ relatively from each other under different planes of nutrition. Thus under normal con- ditions of growth in the rat between three and ten weeks of age 52 Cc. M. JACKSON the muscular system shows the strongest growth capacity, and increases with greater relative rapidity than any other system (Jackson and Lowrey). The skeleton is of course growing stead- ily, but at a much slower rate, so that it is decreasing in relative (percentage) weight. Under the adverse nutritive conditions in young animals when the body-weight is held constant, how- ever, the musculature is barely able to maintain its weight, while the skeleton is able to absorb more than its share of the available nutrition, and to grow steadily (though at a retarded rate). There are similar differences among many of the indi- vidual organs, though in some eases (e.g., liver, alimentary canal) there is a certain degree of parallelism between the normal growth tendency and the behavior when the body-weight is held con- stant at corresponding periods. That also the power of maintenance may vary in organs according to the nutritional conditions is shown by the char- acteristic differences in the losses of organ-weight in chronic inanition as compared with acute inanition in adults (Jackson ahve Dae) E Finally, it should be remembered that the results of the present paper, as well as those concerning adult chronic inanition in a previous paper (Jackson 715 ¢), are based upon the use of a diet wholesome and balanced, but insufficient in quantity. It is probable that more or less different results would follow from other forms of inanition, such as ‘partial inanition’ from a chemically defective or highly unbalanced diet. For example, Hatai (15) finds a pronounced atrophy of the testis and other characteristic changes in albino rats whose growth had been retarded by a ‘lipoid-free’ ration. Bowin (’80), however, in dogs and rabbits fed dry food only (no water) found, with a loss of about 50 per cent in body-weight, the losses in organ- weight similar to those following total hunger. WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS 153 SUMMARY The principal results of the present paper may be summarized briefly as follows: Young albino rats may be held at constant body-weight for considerable periods by underfeeding. The amount of , food required for this purpose decreases as the experiment proceeds. As to the body-proportions, the relative weights of the head, trunk and extremities remain practically unchanged during the experiment. There is apparently a slight increase in the head, counterbalanced by a corresponding decrease in the trunk and extremities, but the change is so slight as to seem of doubtful significance. Of the systems—integument, skeleton, musculature, viscera and ‘remainder’—there is but little change in the weights of the musculature, visceral group (as a whole) and ‘remainder.’ There is, however, a marked decrease in the weight of the integument, counterbalanced by a marked increase in the skeleton. Thus on the low plane of nutrition in the young body maintained at con- stant weight, the growth capacity appears weakest in the skin and strongest in the skeletal system. This is in striking contrast with the normal growth process of corresponding ages, during which the musculature increases with relatively great rapidity and the skeleton lags behind relatively. The increase in the skeleton during constant body-weight appears to involve the ligaments as well as the cartilages and bones. The skeletal growth tends to proceed along the lines of normal development, as indicated by decrease in the water- content, and by formation and union of various epiphyses. Another evidence of the tendency to normal development of the skeleton is seen in the increased relative length of the tail as compared with the body-length. The teeth also continue to develop normally (formation and eruption of the third molars). The individual viscera may be classified in three groups: (1) There is during the maintenance of constant body-weight in young rats a well-marked increase in the weights of the eye- 154 Cc. M. JACKSON balls, spinal cord, alimentary canal (both empty and including contents), testes, hypophysis and suprarenal glands. The suprarenals undergo sexual differentiation in weight (as occurs normally), but the hypophysis apparently does not. (2) There is no marked change in the weights of the brain, heart, kidneys and epididymi. ‘The liver is variable, showing a definite increase in the earlier periods, but a decrease later. The lungs show a slight decrease in the early periods, but not in later. (3) There is a well-marked decrease in the weights of the thymus (‘hunger involution’), spleen, thyroid gland and ovaries. When the organs are similarly grouped according to degree of loss during chronic inanition in the adult (slight loss, loss pro- portional to body, and loss greater than body), many differences are found on comparison with the corresponding groups in the young during constant body-weight. This is explained as due to to the presence of both the growth tendency and the (more or less different) maintenance tendency in the young animals, whereas in the adult there is only the tendency to maintenance. Both the growth tendency and the maintenance tendency, however, show characteristic differences in the various organs according to nutritional conditions (normal nutrition, acute or chronic inanition). WEIGHTS OF ORGANS IN UNDERFED YOUNG RATS eae) LITERATURE CITED Aron, Hans 1911 Nutrition and growth. Philippine Jour. Science, vol. 6, pp. 1-o5l. 1914 Untersuchungen iiber die Beeinflussung des Wachstums durch die Ernihrung. Berliner klin. Woch., J. 51, 8. 972-977. Bowin 1880 Beitrage zur Frage tiber die Trockenernihrung. Dissert. St. Petersburg. Cited by Miihlmann, Russische Literatur iiber die Pathologie des Hungers. Centralbl. f. allg. Path., Bd. 10, 1899. Donatpson, H. H. 1908 A comparison of the albino rat with man in respect to the growth of the brain and of the spinal cord. Jour. Comp. Neur., vol. 18, no. 4. 1911 The effect of underfeeding on the percentage of water, on the ether-alcohol extract, and on the medullation in the central nervous - system of the albino rat. Jour. Comp. Neur., vol. 21, no. 2. Harms, W. 1909 Ueber den Einfluss des Hungers auf die Wirbelsiule der Tritonen. Verhandl. d. deutschen Zool. Gesellschaft, 19 Versamml., S. 807-312. Harat, 8. 1904 The effect of partial starvation on the brain of the white rat. Am. Jour. Physiol., vol. 12, no. 1. 1908 Preliminary note on the size and condition of the central nervous system in albino rats experimentally stunted. Jour. Comp. Neur., VOlwl Sonos ee 1913 On the weights of the abdominal and the thoracic viscera, the sex glands, ductless glands and the eyeballs of the albino rat (Mus norvegicus albinus) according to body-weight. Am. Jour. Anat., vol. UGS WaKoy5 Abe . 1914 On the weight of the thymus gland of the albino rat (Mus nor- vegicus albinus) according to age. Am. Jour. Anat., vol. 16, no. 2. 1915 The growth of the body and organs in albino rats fed with a lipoid-free ration. Anat. Rec., vol. 9, no. 1. Jackson, C. M. 1909 On the prenatal growth of the human body and the relative growth of the various organs and parts. Am. Jour. Anat., vol. 9. 1913 Postnatal growth and variability of the body and of the various organs in the albino rat. Am. Jour. Anat., vol. 15, no. 1. 1915 a Effect of acute and chronic inanition upon the relative weights of the various organs and systems of adult albino rats. (Abstract). Proc. Amer. Ass’n Anatomists. Anat. Rec., vol. 9, no. 1, p. 90. 1915 b Changes in young albino rats held at constant body weight by underfeeding for various periods. . (Abstract). Proc. Amer. Ass’n Anatomists. Anat. Rec., vol. 9, no. 1, p. 91. 1915 ¢ Effect of acute and chronic inanition upon the relative weights of the various organs and systems of adult albino rats. Am. Jour. Anat., vol. 18. Jackson, C. M., and Lowrey, L. G. 1912 On the relative growth of the com- ponent parts (head, trunk and extremities) and systems (skin, skele- ton, musculature and viscera) of the albino rat. Anat. Rec., vol. 6, no. 12: 156 Cc. M. JACKSON Jonson, Arvip 1909 Studien iiber die Thymusinvolution. Die akzidentelle Involution bei Hunger. Archiv f. mikr. Anat., Bd. 73, 8. 390 ff. Lowrey, L.G. 19138 The growth of the dry substance in the albino rat. Anat. Rec., vol. 7, no. 9. Morgutis, 8. 1911 Studies of inanition in its bearing upon growth. I. Archiv f. Entw. d. Org., Bd. 32, H. 2. Warers, H. J. 1908a Capacity of animals to grow under adverse conditions. Proc. Soe. for the Promotion of Agricultural Science. 1908 b The influence of nutrition upon the animal form. Presented at the Thirtieth Meeting of the Society for the Promotion of Agri- cultural Science. INHERITANCE IN THE ASEXUAL REPRODUCTION OF HYDRA K. 8. LASHLEY From the Zoélogical Laboratory of the Johns Hopkins University TEN FIGURES CONTENTS Peintroductioneeeseeaen ee gySiste a Sob Teele = ele eal rgeiapet ee sie > pre cing 157 j cberqayeranneXey ou si (CObk J gL) 0Y=) Cathe ee me reece ies ie ee ame 08 159 Griticismsrore danelasvexpeniim ents. 445-1. ae ae eee ee ie ee 159 PNTALY SISHOl LANE! SuCAGHes v.u..s\e 2s kok se eit gee meee ote ayer 163 ise ede qUestrons csg hisses. 3/5 0G,o dal gate a ene eNen ei = cee: eran 165 lle Wii Naonnep iia d hiohehiy oc 66 omic eee Ete cere cee os 6d a obaeenoneee dsc 166 Variation in the number of tentacles of H. viridis................... 167 Conditions producing like variation in parent and offspring......... 170 MUSH MIS teMceLOMGlvierse ma Ces .: occas c's. s% « Foci P RR Re ee ae eee 174 IE XELIMNETUG HE! TCT MOGS ec cates a tue 3.4 os + 400 oa ee eee ene ere ea eee 175 Detailed examination of two clones of H. viridis.................... 176 Differences in number of tentacles and size................... eetG, Ovherditerencessbetween the clones: ... Gace eee een oe nen 182 Nature of the difference between Clones A and D................. 183 xis tenceromothen Giviersetaces... >... .4sc sone mone eee een eee 186 SUUTCTVOMA ieee mae la chose Ce Rar eae eee eID tn. hm aia haa igre hold Bin Ree 190 iIWesinheritancerot variations within the clone: .s.sneeseee: eee eae 190 Riesemblanceromelosemelativuess 24> 04+ sce aoe Genes 191 He ChsvoriseleCtlONer eas iiss cc econ. 54, Sony Aa eee eee 195 IMEriGAMCEKOLASIIZOMee rsa. ais ce cok oe > olerale sh er BARTS Sie eae 199 Wa’ LDDTISYOIS eT Oat Eenes oa Fy ocr se Ree Ss OG CA le he noe tr ea 202 V1” BS TENBAWE POO ae OP ti Mi to a SERS ES nae Ratt Rear ch te rs 204 JUST SRE EW ATEN OI EEYG Lees oy Dt PRC ee Oy cide nes Aran tO tas Na ae 205 ANT OVSHING IS os ARs Ree A re, A Sg Oo oe os en oO i Me 208 I. INTRODUCTION The theoretical bearing of the work of Johannsen upon pure lines was not widely appreciated by students of heredity for the first few years after its publication; the general applicability of Mendel’s law had not been shown and, hence, the genotype conception was without the support since gained from the evi- 157 158 K. §S. LASHLEY dence: bearing upon the stability of the unit character. The first attempt to test the general application of Johannsen’s conclusions was made by Elise Hanel. In 1908 she published the results of two years of experimental study of inheritance in the asexual reproduction of the fresh-water polyp, Hydra grisea, and drew conclusions which were quite in accord with the results obtained with self-fertilized lines of beans. Her results were soon called in question, however, by the work of Hase (’09) and the eriticisms of Pearson (’10), who obtained such con- flicting results that a re-investigation of the subject became necessary. The results of such a renewed, independent study of variation and inheritance in Hydra are presented in the present paper.! In order to make clear the questions at issue and the bearing of the results presented here upon these and the general problems of genetics, it will be necessary to review briefly the work of Hanel and the criticisms raised against it, especially as the latter are complicated and not clearly valid. Before this is undertaken, however, some explanation of the termi- nology of the present paper may not be out of place in view of the past confusion in genetic literature. Johannsen (713), following Shull, has employed the word ‘clone’ to describe a family descended from a single individual by asexual reproduction. The name ‘biotype’ applies to any group of organisms which show the same hereditary constitution (‘genotype’). In the statistical terminology the ‘distribution’ is the natural arrangement of individuals showing variation (‘variates’) in classes. The ‘mean’ is the arithmetical average of the variates. The ‘standard deviation’ (c) is an expression of the average extent of variation from the mean shown by the variates. The ‘coefficient of variation’ expresses the average percentage of variation from the mean in terms of the mean as unity. The ‘coefficient of correlation’ (7) represents an arbitrary measurement of the average degree of resemblance be- tween correlated members of two series of variates. The ‘coefficient of regression’ (/) expresses the average extent to which the variations of one series follow the variations of the correlated series. The nomenclature of the genus Hydra has been subject to frequent changes. The names which have been used most frequently by ex- 1 This was undertaken at the suggestion of Prof. H. 8. Jennings, to whom I wish to express my indebtedness for his assistance and kindly criticism during the course of my work. I wish also to thank Prof. B. E. Livingston for the use of the facilities of the Laboratory of Plant Physiology of the Johns Hopkins University during the summer of 1912. INHERITANCE IN ASEXUAL REPRODUCTION 159 perimental zodlogists are those of Linnaeus. Brauer (’08) applied the rule of priority, thus requiring the nomenclature of Pallas. In 1912 Bedot criticised Brauer’s revision and showed that the species, H. viridis L. should be retained. Since the experimental literature has employed the Linnaean nomenclature this has been used here in order to avoid the constant repetition of synonyms. The names employed are given below at the left, with the revision of Bedét at the right. H. viridis L. : Hiovanidis i. H. grisea L. H. vulgaris Pall. H. fusea L. H. oligactis Pall. H. polypus A. Brauer. Experiments of Hanel Hanel began her work in 1906 and during the two years of her experiments bred clones from 26 wild Hydras, obtaining records of nearly, 7000 buds, the descendants, by asexual reproduction, of the original 26. Upon comparing these 26 ‘stem parents’ with their immediate progeny she found that those with a large number of tentacles produced buds having, on the average, a greater number of tentacles than the corresponding buds of par- ents with few tentacles. From the progeny of the ‘stem parents’ taken singly Hanel selected polyps with a high, an intermediate, and a low number of tentacles and continued this selection for from two to seven generations. Averaging the results obtained, she found that within the single clone the mean number of tentacles of the progeny of individuals selected for four genera- tions was slightly less in the group selected for a large number than in the group selected for a small number, whence she con- cluded that variations within the clone are not inherited. Critucesms of Hanel’s results In 1909 Hase, at the suggestion of Plate, studied the relation of the number of tentacles of Hydra to the age of the polyp and to the number of tentacles of the buds. He verified the obser- vations of Parke (’00) and Hanel that the number of tentacles increases with the age of the polyps, finding an average increase of 2.1 tentacles in a number of polyps kept under observation for 90 days. In recording her data Hanel had taken this fact 160 K. S. LASHLEY into consideration but, believing from observations upon mass cultures that the adult number of tentacles is proportional to the number which the polyps bear when they produce their first buds, she thought herself justified in computing the adult number of tentacles for young poylps which had produced a single bud. Hase concluded that the increase is not regular and that there is no true adult number of tentacles for any polyp. Hase further believed that the number of tentacles which the buds have at the time of separation from their parents is not proportional to the number of tentacles which the parents bear at that time. He gave data to show that while the number of tentacles of the parents is increasing with age, that of the suc- cessive offspring of these parents remains constant and does not follow the increase shown by the parents. He also cut a number of polyps into two or three pieces and found that they did not regenerate the same number of tentacles that they had borne originally. From this he concluded: Die verschiedene Tentakelzahl ist eine reine Somation und besitzt keinerlei Erblichkeitswert. Als eine reine Linie (im Sinne von Johann- sen) kann man daher die direkten ungeschlechtlichen Nachkommen einer Hydra nicht bezeichnen. Es ist daher auch nicht méglich char- akteristische Typen zu isolieren und sie erblich konstant zu erhalten. Aus gleichem Grund kann man auch keine reinen Linien (im vorigen Sinne) aus einer Hydrapopulation durch Selektion sortieren. This conclusion is far from justified, however, by the evidence which Hase presents. Records of two parent Hydras are given, which increased slowly in their number of tentacles from five to nine, together with records of all the progeny of each. Tf the two groups are averaged, as has been done in table 1, it is evident that the buds produced by these two parents after they had acquired a large number of tentacles have a higher average than those produced when the parents had few tentacles. Yet from these data, Hase concludes, ‘‘Wir sehen, zuerst, ein schein- bares Mitgehen der Tentakelzahl der Knospen mit der Mutter, aber bald tritt ein vélliger Riickschlag ein.”’ As I shall show later, this is not true for H. viridis, as it is probably not for H. fusca: the regression is only partial. INHERITANCE IN ASEXUAL REPRODUCTION 161 The experiment upon regeneration likewise fails to support the conclusion which Hase draws from it, since it shows only that the number of tentacles is not regenerated at its full value but fails to disprove that the number regenerated is not proportional to the original number. I have reviewed the experiments of Hase in some detail because they have been generally quoted as proving that the number of tentacles of Hydra is not an heredi- tary character, whereas, at best, they but serve to cast some doubt upon Hanel’s conclusions. TABLE 1 Increase in the average number of tentacles of buds produced while their parents bore the numbers increasing from 5 to 9 (modified from Hase) MEAN NUMBER OF TENTACLES OF BUDS PRODUCED WHILE THE PARENTS BORE THE NUMBER OPPOSITE NUMBER OF TENTACLES BORNE BY PARENTS or) oO or [op OH Or or (Se) ww 0 W > S ~I ie) rit) eo) o> or bo A much more serious criticism of Hanel’s work, and of ‘pure line’ work in general, was made by Pearson during the follow- ing year. He subjected Hanel’s data to a more thorough analysis than she herself had done and held that they by no means justi- fied the conclusions drawn. Concerning the evidence for the existence of strains, diverse with respect to tentacle number, he Says: Hanel begins with a very careful investigation of the growth and environmental changes in the character selected, the number of tentacles of Hydra grisea. There is a general agreement with Parke’s results that the number of tentacles changes with age, size, food and place of culture. Differences in these factors can produce very considerable differences in individuals and differences in the averages of differentially treated groups which can amount to as much as 0.5 to 0.8 of a tentacle. These are precisely the order of the average hereditary differences. Thus: 162 KS) Ashby: PARENT OFFSPRING No. of No. of No. 0, No. individuals tentacles individuals tentacles 8) 6 364 6.943 9 7 310 7.296 4 8 166 7.344 4 9 125 7.383 It will be at once recognized that the differences here are rather less than many of the environmental differences and that there is no security that these 26 foundation Hydras are really represented by differentiated hereditary numbers of tentacles. Yet this table, as it stands, embraces Hanel’s proof that the number of tentacles is an hereditary character in the ‘pure line.’ What evidence is there that any one of the numbers of tentacles attached to those 26 parents is really constitutional and not environmental? When we consider that Hanel’s experiments extended over a period of two years, that the different ‘stem parents’ were col- lected and bred at different seasons, and that no data are given by which it is possible to judge which of the races were kept under a uniform environment this criticism greatly reduces the value of her evidence for the existence of diverse genotypes. The second point of Pearson’s criticism deals with the inheritance of the number of tentacles within the clone. The data were subjected to statistical analysis by Miss Elderton and the cor- relation between relatives was determined for different degrees of relationship, within the population composed of the 26 clones. A comparison of the correlations showed that there is a greater resemblance between siblings and between parents and their offspring than there is between ‘uncle and nephew’ or between grandparents and grandchildren. This, as Pearson points out, can not be due merely to the inclusion of diverse races in the tables of correlation but indicates that there is an inheritance of variations within the clone. From this evidence Pearson concludes that Hanel’s records show, not what she believed them to show, but just the reverse; that regression is only partial, both in the population and in the clone; that variatons are inherited equally in both population and clone. INHERITANCE IN ASEXUAL REPRODUCTION 163 Analysis of Hanel’s data Since this method of treating the data does not deal with individual clones it seemed advisable to carry the analysis some- what farther. I have computed the mean number of tentacles of all individuals in each of Hanel’s clones and the correlation between parent and offspring for each.2. These are given in tables 2 and 3. The mean numbers of tentacles of the clones TABLE 2 The mean number of tentacles of Hanel’s clones arranged in the order of magnitude NO. OF TENTACLES OF CLONE NO. MEAN NO. OF TENTACLES o STEM PARENT AN Een ers anit eosin 6 .428+0.040 0.771 7 SESE ave Cnet ren OE MONS 6.438 +0 .020 0.475 7 Ate rcite chs clones 6.670 +0 .058 0.826 6 Sraaontek ase aes s 6.677 +0.049 0.816 7 De tb Cbe ama tee 6.8050 .029 0.734 6 A cise HS EI 6 .925+0 .054 0.692 6 Alpe yerah ry iresietens 6 .926+0.041 0.750 6 Wer eeieisral-e, stave e s.e%e 7.010+0.031 0.786 6 LP ae Ser ot hee are 7 .026+0 .024 0.780 6 US ertenerarivet cs Sie; ans. 1 7.170+0.031 0.721 7 2b... 7 .184+0.032 0.658 5 NOE tierce ces 7 .230+0 .034 0.798 8 tS ee ee 7.240 +0 .023 0.906 7 OP ep cists ieee fotaes 7 .264+0.041 0.902 9 AG eect ee ee ares 7.287 +0 .056 0.962 12 18S ond Sica 7.319 +0 .032 0.784 5 Sei topes Seen eae 7 .325+0 .029 0.786 7 DOr: 7 .361+0.050 0.850 oS) NSB ard Rieter oee cea oie 7.367 +0 .046 1.005 8 Gevertaet is oats G 7 .384+0 .025 0.914 9 35 7.393 +0 .023 0.763 8 BOR de Sve oka 2s 7.400 +0 .040 0.938 8 22... 7 .4380+0 .043 0.934 7 LCG 7 .456+0.049 IY 10 4. 7.500 +0 .039 0.862 7 Ae a CIRCE 7 .641+0.047 0.835 6 Drie rotete eee, cio. saa 7.712+0.044 0.942, a * Hanel records the variations in her clones in percentages; the tables contain many errors, the sum of the percentages included in one array ranging from 85 to 140 per cent. This leads to error in the means and correlations computed from the tables but these probably average out in the numerous constants determined. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 2 164 K. S. LASHLEY form an almost unbroken series from the smallest to the largest, and, lacking evidence upon the environmental conditions, give little indication of the existence of diverse races. Further analysis indicates that the clones tended to become more alike after they had been kept under cultivation for some time. The average number of tentacles of the offspring of each of the 26 ‘stem parents’ and the average of all their later descendants were computed and are given in table 4. The descendants of parents with extreme variations tend to revert to the mean. The coefficients of correlation between parent and offspring within the clones (table 3) are extremely variable, but the TABLE 3 Correlation between parent and progeny within each of Hanel’s clones CLONE NO. “CORRELATION CLONE NO. CORRELATION ee IRS Si catice ALE 0.170+0.054 14 0.057 =0 .050 D RRS ean: Rem e, P 0.336 +0 .067 15 0.323+0.040 PA Ys « 0.469 =0.040 16 0.0500 .066 SHON ee Reet ee ce any 0.010+0.029 17 —0.073+0.040 2 8 ee ors Ses Pay —0.009+0.040 18 — 0.060 +0 .037 Dee ee AIRE eles 0.2340 .037 19 0.072+0.041 Tiga as hc) Bs ie eee 0.142+0 .026 20 —0.087+0.041 Chere ac 35 Stee Re 0.0360 .028 21 0.708 +0.030 SER An A eer n eiee rd, 0.068 +0 .066 22 0.007 +0 .047 OS en ee Pe ee He ht 0 .006+0 .047 wy 0.178+0.059 LOR erie Merete es crs 0.009 =0 .042 24 0.000 +0 .067 ee en SI ERS Sap Net Mi 0.040 +0 .037 25 0.128+0.046 1 ce ER ee Re Oe a 0.031 +0 .030 26 0.1040 .047 TOA eee: erareeceye Beda 0.048 +0 .040 Mean of all...... 0.101 TABLE 4 Regression in the first and later generations of the descendants of Hanel’s 26 ‘stem parents’ ‘STEM PARENTS ”’; NO. OF TENTACLES 6 of 8 9 FIRST GENERATION; MEAN NO. OF TENTACLES ALL DESCENDANTS; MEAN NO. OF TENTACLES 7.086 7.102 7.347 7.347 INHERITANCE IN ASEXUAL REPRODUCTION 165 majority (75 per cent) are positive and the evidence favors the belief in an inheritance of variations within the clone. The great irregularity of the results, however, makes it impossible to draw from them any law of inheritance within the clone or even to be sure that there is any constant tendency toward the inheritance of variations. Unsettled questions Thus the evidence from previous work bearing upon heredity in Hydra must be looked upon as inclusive and the chief problems of heredity and variation as still unsettled. Hanel maintains the existence of diverse races of Hydra which are distinct in their hereditary constitution. This is denied by both Hase and Pearson but upon different grounds. Hanel holds that, within a population, variations are inherited and that selection is effective in isolating the diverse races, within which there is no: inheritance of variations. Hase denies that variations in the character studied are inherited at all and asserts that such variations are purely somatic. Pearson attempts to show that the variations are inherited both within the population and the clone. Analysis of Hanel’s data shows that there is usually a correlation between parent and progeny within the clone; that this may be very high in some cases, and in others, negative, varying around an average of 0.101. Experimental analysis of the cause of this correlation is lacking. The chief questions demanding further investigation seem to be the following: Will such a variable character as the number of tentacles of Hydra lend itself to genetic study? Are there other characters of Hydra which vary less within the individual? Given a character which is comparable in different individuals, do races, hereditarily diverse with respect to this character, exist? If there are such races, in what characters do they differ and are they numerous or few? Is there inheritance of individual differences within the population, and if so, what part do diverse races play in this inheritance? Is there actually a correlation between parent and progeny within the clone? 166 kK, S:- LASHLEY If so, how is it produced and what is its significance? Is there an inheritance of individual differences within the clone? What is the origin of the diversities between races? II. VARIATION IN HYDRA The only variable character in Hydra considered in detail by previous workers has been the number of tentacles. At the beginning of the present experiments I examined a number of polyps in the hope of finding other variable characters suitable for genetic study. Size, body-form, color, relative length of the tentacles, reaction to mechanical stimuli, distribution of the nettle cells, diameter of the mature nematocysts, and reaction to light were compared. Of these only size seemed at all suitable for statistical study. Variations in practically all the characters of Hydra appear, not only between different individuals, but in the same individual at different times. Hase has shown that the number of tentacles, alone, varies too much in the individual to form a character suitable for statistical treatment'and Plate (13) has stated the necessity for the use of the number of ten- tacles at some definite stage in the development of the polyps in future studies of heredity. The same individual variability appears in the size of the polpys so that the size at some definite stage of development must be used in comparative studies. The time when the bud separates from the parent marks the occurrence of certain definite physiological processes in the bud and, if the number of tentacles of the bud is taken at this time, it serves as an index of the state of development of the bud at the time when these processes are initiated. This number may be compared, in tests for heredity, either with the number borne by the parent at the corresponding age, or with the number which it bears when the given bud is produced. It is more difficult to find a stage where the size of different polyps is comparable. Growth is very rapid for the first few days after the buds are released and it is not always possible to take measurements of the new buds, as such measurements require a great deal of time. After four or five days for H. viridis, when the buds have matured INHERITANCE IN ASEXUAL REPRODUCTION 167 and begun to produce buds in their turn, growth is much slower, although the polyps continue to increase in size until the fifteenth day or later. For comparisons of parent and offspring, I have used measurements of the size at the age of seven days, or as near this as was practicable. The figures so obtained are greatly subject to chance variation but are problaby as dependable as any which can be obtained for so plastic an organism. Variation in the number of tentacles of Hydra viridis The most frequent number of tentacles (the modal number) found in wild populations of H. viridis is usually 6, sometimes 7, and the normal range seems to be from 4 to 9. Table 5 shows, in percentages, the distribution of variations in H. viridis found by previous investigators. Both local and seasonal differences probably: play a part in the production of the differences in these populations.. The data shown in table 6 have been obtained TABLE 5 Distribution, in percentages, of variations previously recorded for Hydra viridis NO. OF THNTACGLES,. ..4- s- <5 - 4 5 6 7 8 9 10 11 12 13 Rav clin( G99) sv. etowatars ete Sate 2 .5| 42).0|_40).0) 12,0) 220 Hathaway (@99)s2...-.- fi 25)) 5020)) 35:0)" 320 Pankey (ACO) dees. cow 32731484) 12-9) 458i IG 1 eae ate 4.8} 30.6] 44.5] 16.4; 2.9) 0.6 JUUE. 22.8] 46.5) 20.9} 7.4) 1.4) 0.9 INGE (CU escassa conan lu z 24205420 | Sa0 ee ? ? liasex COO) rata yihe actors De OuleAan Ole Ol 2820 pee OlRZoe0 a0 eon OLS aOrS TABLE 6 Distribution, in percentages, of variations in a population of Hydra viridis from Baltimore NO. OF TENTACLES MEAN NO. OF NO. OF TENTACLES 2 Y POLYPS 3 4 5 6 i 8 9 10 Oct., 1911. . 0.6) 0.6)/25.0)/51. 3.7| 0.6) 6.988+=0.044 | 0.8404 167 Oct., 1913. .| 0.1) 0.4] 5.5/30.9)50.9}11 0.9 6 .696+0.015 | 0.7332 | 1000 Apr., 1914.. 0.7} 0.7|11.4/41 .4/37 (ROBE teogo 020507 MeOrSs271 140 168 K, S.. LASHLEY from a population in a very small stagnant pond in the neighbor- hood of Baltimore. Very extensive seasonal changes are evident from the mean of the population at different times. Numbers of tentacles less than 5 or greater than 9 probably can not be considered normal for H. viridis from this neighborhood. Buds formed with less than four tentacles are usually small and do not extend their tentacles. They either produce more ten- tacles quickly, or die. The only individuals with ten tentacles which I have found were small-bodied, dark in color, and pro- duced buds very slowly. In H. grisea a large number of ten- tacles seems to present a rather unstable condition. Individuals with more than eight tentacles in my cultures have shown a tendency to divide lengthwise and thus reduce the number of tentacles to the mode. I have observed longitudinal division in only one specimen of H. viridis, which bore six tentacles at the beginning of division and gave rise to two polyps, each with six tentacles. As individual polyps grow older they form additional tentacles, a condition first observed by Marshall (’82) in H. viridis. Hase has demonstrated a like condition in H. fusca and Hanel in H. grisea. Hanel has found that the addition of tentacles occurs even in polyps which are starving, although hunger tends to inhibit the process, as is clear from table 7. Parke found that the number of tentacles might be reduced and Hanel’s table shows that this reduction is favored by starvation. My own TABLE 7 Changes in the number of tentacles of 129 specimens of Hydra grisea during one month’s cultivation; after Hanel WITH FOOD WITHOUT FOOD Number of tentacles added: -(.5-sheememetee. .-ebes 45 19 Number of polyps showing change................... 29 14 Number of tentacles absorbed..............,..-----. 2 8 Number of polyps showing change..................- 2 6 Number cf polyossmchanced- eee sate 31 47 otal mumberofpolvpss... 52 eee eee eee ert €2 67 Meankincrease perundividuall: .2 4. eeeerEemateocces 0.69 0.16 INHERITANCE IN ASEXUAL REPRODUCTION 169 data verify these results for H. viridis. Figure 1 shows the mean number of tentacles of about 30 polyps from two different clones at successive intervals of two days. Besides the specific tendency to change the number of ten- tacles with advancing age, there are environmental factors which may produce variation in the number of tentacles of Hydra. Hanel’s data, table 7, show that starvation partially inhibits 75 tentacles 6.5 6.0 2 2 10 20 30 Days Fig. 1 Increase with advancing age in the average number of tentacles of polyps from two clones, A and D. The averages are based upon about 30 polyps from each clone. Clone A shows an average daily increase of 3.6 per cent; clone D an increase of 2.0 per cent. the addition of tentacles and causes some to be absorbed. Parke has shown that other unfavorable conditions, such as stagnant water, have a like effect. Injury and regeneration of the oral end also leads to a reduction in the number of tentacles (Rand ’99). These factors produce variation in adult Hydras, but further, as will be shown, they have a like effect upon the buds produced by these adults. 170 Kk. Si LASHERY Conditions producing like variations in parents and offspring Age. Corresponding to the increase in the number of ten- tacles with the age of the parents, the buds produced by old polyps have, on the average, more tentacles than those pro- duced by young ones, that is, there is an increase in the number of 6.5 6.0; 5.5L * 5 10 15 20 25 30 Fig. 2. Increase in the average number of tentacles of successive buds pro- duced by the same parents. The ordinates represent the mean number of ten- tacles of the successive buds. The abscissae represent for Clone A the successive buds produced by 70 parents, for Clone D the buds produced on successive days after the 40 parents which furnished the basis of the curve began to produce buds. tentacles of successive buds with the increasing age of the parent. Figure 2 shows this condition clearly. Data from two clones are included in this figure; for clone A the average number of ten- tacles of the first, third, fifth, seventh, etc., buds are given in the graph; for clone D the averages of all buds produced during suc- INHERITANCE IN ASEXUAL REPRODUCTION VAL cessive intervals of two days furnish the basis of the graph. The general trend of both lines is upward, although more mark- edly so in A than in D. The same relation for the two clones was found for the increase in the number of tentacles of the parents (fig. 1). The increase in the parental number of ten- tacles with age is due to some internal factor. Its relation tothe increase in the number of tentacles of successive buds will be considered later. Besides this internal factor, there are various environmental agents which modify the number of tentacles of parent and bud in the same direction, and which must be taken into account in any study of inheritance. Effects of starvation. One hundred polyps from a mass culture of a single clone (clone A) were divided at random into two groups of 50 each. One group was fed every second day with small crustacea, the other every sixth day. Both were placed in clean dishes of fresh water every second day, and each polyp was kept in a separate dish. The result of this method was that one group received about three times as much food as the other. The first result of the partial starvation of the six-day group was a reduction in the rate of budding. The well fed group produced an average of 0.203 buds per day for each polyp; the starved group produced only 0.074 buds per day, which gives a ratio of almost three to one, corresponding to the amount of food given. There was also a difference in the average number of tentacles of the buds produced by the two groups, the offspring of well fed parents having a slightly higher average than the others. This is shown in table 8. The difference here is small and may easily have been due to chance. The experiment was TABLE 8 Effects of starvation of parents upon the mean number of tentacles of their buds; all buds recorded within 24 hours after their release PARENTS BUDS Mean no. of tentacles No. of polyps Mean no. of tentacles No. of polyps Dede etae nial: 7.18+0.06 50 6.32+0.04 153 Starved....... 7.18+0.07 50 6 .24+0.06 61 172 KG. DASHEE TABLE 9 Effects of starvation of parents upon the number of tentacles of their buds; second experiment PARENTS BUDS -Mean no. of | No. of | Tentacles | 14 Oral iNo: : fg ne-ot | Novel | ‘addea by | Meta nonof | No.of | iterate Clone A WAGs ssacancsncel Oss O(0s 2 0 6.59+0.03 WZ Stanvedeeeeeoeee 6 .80+0.05 25 De 6.37+0.09 44 | 0.22+0.09 Clone D WClocosco cnc scac|| DeQe=O (06 25 1 5.76+0.03 195 SIA EO gen oe al) Hoes O).0G 25 0) 5.20+0.06 44 | 0.56+0.07 repeated with polyps from two clones and results were obtained which made it certain that starvation of the parent leads to a reduction of the number of tentacles of its progeny. The results of this experiment are shown in table 9. The parents recorded in this experiment were old polyps taken from mass cultures. Their mean number of tentacles already exceeded the means of the clones from which they came, so that any considerable increase in their numbers of tentacles during the brief time of the experiment was not to be expected. Hanel’s data, however, prove that starvation tends to inhibit the addition of tentacles in parents and this experiment shows that the buds of starved parents have fewer tentacles than those of normal ones. This condition, occurring among a large num- ber of polyps subjected to unequal and fluctuating environmental conditions would undoubtedly lead to a slight degree of resem- blance between parent and progeny, even if there is no true inheritance of parental variations. Effect of injury. Twenty mature Hydras were cut in two trans- versely and kept until both halves had regenerated. Each half was kept in a separate dish and the buds which it produced after regeneration was complete were recorded. Altogether, 287 buds were obtained from the 40 freshly regenerated polyps. The mean number of tentacles of the clone from which the orig- inal 20 were derived was, at this time, 6.91+0.03. The mean INHERITANCE IN ASEXUAL REPRODUCTION 173 number of tentacles of the original 20, the oral ends, was 7.05 + 0.09. The mean number regenerated by the aboral ends was 6.42+0.09. The regenerated oral ends produced 146 buds, the aboral ends, 132 buds within 28 days after the operations. The mean number of tentacles of the buds from the oral ends was 6.88 + 0.03; that of the buds from the aboral ends was 6.59 + 0.04; the difference here is 0.29 +0.004 in favor of the offspring of the polyps which were required to carry out the lesser regenerative processes. Clearly, the same factors which cause a reduction in the number of tentacles in the parents (the call for an expendi- ture of energy in regeneration) cause also a reduction in the number of tentacles of their offspring. In order to test the length of time during which the after- effects of injury to the parents influence the number of tentacles of the buds, the mean number of tentacles of all first, all second, etc., buds produced after regeneration by the oral and aboral ends was computed. This is shown in figure 3. After the ninth bud the graphs are based upon too small numbers to be sig- nificant. It is clear from this figure that immediately after regeneration the buds produced have few tentacles and that the number of tentacles of successive buds increases rapidly until, after five or six have been formed, the normal condition is regained. These experiments show that certain environmental agents, acting upon Hydra, tend to produce variation in a given direc- tion in both parent and offspring, either, as in the first case, by acting simultaneously upon both parent and bud, or, as in the second case, by affecting the bud only indirectly through a reduction in the vitality of the parent. Although only two such agents have been studied experimentally it is probable that many others have a like effect. Parke has shown that the bacterial conditions leading to depression in his cultures brought about a reduction in the number of tentacles of the polyps. I have found that polyps which have just recovered from de- pression tend to produce one or more very minute buds with _ few tentacles. Polyps kept in water of high salt content become small and dark-colored and produce small dark-colored buds. ily! K. S. LASHLEY ‘ 6 | t 7.0 tentacles ‘ ' ‘ ' 0) 5 : 10 15 Successive buds Fig. 3 The mean number of tentacles of the successive buds produced by polyps after regeneration; (——) buds produced by regenerated oral ends; (——-—) buds produced by regenerated aboral ends; all polyps were taken from the same clone. Temperatures much above or below the optimum have an effect like that of high salt content. The results of the action of these agents resemble the cases of temporary parallel induction found in parthenogenetic reproduction (Agar ’13) but the mechanism of the action is probably much simpler since we are dealing here only with a congenital reduction in vigor. III. EXISTENCE OF DIVERSE RACES In addition to these variations resulting from environmental action, are there differences between the individuals of a popu- lation due to some internal, hereditary factors? This question has been tested by breeding a number of clones of H. viridis under similar cultural conditions. INHERITANCE IN ASEXUAL REPRODUCTION Lie Experimental methods Hydra is easily cultivated in the laboratory, provided that a con- stant supply of food is available. The polyps used in the experiments were collected from various sources, usually from ponds containing Nitella or Elodea. In the greater part of the work each polyp was kept separately in a small Stender dish with about 5 cc. of water. It was found necessary to wash these dishes in boiling water every second day in order to keep down bacterial growths and maintain the polyps in good health. A small species of Cyclops was used exclusively as food. Attempts to breed them in sufficient numbers in the laboratory failed and all food material was obtained by towing with a plankton net in an artificial pond. The supply so obtained was further con- centrated so that from one to two hundred Cyclops were given to each Hydra every second day. (A healthy specimen of H. viridis will eat ten or more Cyclops daily and the large number supplied allowed each polyp to capture as many as it could eat). At first, a growing stem of Elodea was kept in the culture with each Hydra but as this was found to be unnecessary and a possible source of contamination it was dis- continued after the first experiment. All the cultures were kept at approximately the same level upon a broad table-top where they were exposed to uniform conditions of light and temperature. The cultures of various clones were divided into small groups and the order in which these were fed and arranged upon the table was varied from day to day. While conditions may have varied in the different individual cultures, there seems to be no reason for believing that such differences did not average out in each of the clones, considered as a whole. Besides the individual cultures a number of mass cultures of different clones were kept under similar conditions. Numbers of polyps were placed in 6-inch battery jars filled with .water from the food pond. The water in the cultures was renewed weekly, and where. two clones were being bred for comparison the water from the culture jars con- taining the two clones was interchanged frequently. The number of tentacles of the polyps was counted under the Zeiss binocular, and in the case of the polyps in individual cultures the number of tentacles of the buds was recorded within 48 hours of the time when they were released from the parents. The numbers of ten- tacles of the parents were recorded also at the time when each bud was produced. Measurements of size were made in the following way: The polyps, contracted, were placed on a grooved slide under a com- pound microscope. When they expanded until the length was about four times the diameter, an outline of the body was made with a camera lucida. The area of this outline was measured with a planimeter and from this, treated as the area of the longitudinal section of an elipsoid, the volume was computed. This method is subject to about 25 per cent error, but is much more accurate than measurements of a single diameter. The constants given in the following pages were computed 176 K, Si DASHLEY by the simple statistical formulae devised by Pearson and others (Davenport ’04; Yule 712). In the computation of the -standard deviations for size Sheppard’s correction was not used. All calculations were made with the aid of the ‘Brunsviga’ computing machine, and later repeated de novo. Detailed examination of two clones On June 10, 1912, twenty-five specimens of Hydra viridis, collected from a limited area in a small undrained pond, were isolated in individual culture dishes. They began to multiply rapidly under the favorable conditions of the cultures. The new-formed buds were each placed in a separate dish, labeled, and the number of their tentacles recorded within 48 hours after they separated from the parents. After a few days it was found impossible to care for all the descendants of the stem mothers and it seemed best to discard the majority of the clones. Two of them, A and D, which seemed somewhat diverse were retained. It soon became necessary again to reduce the number of cultures, as not more than 300 animals could be cared for at one time. An equal number of each of the two clones was therefore selected for high, intermediate, and low numbers of tentacles, and the cultures were so arranged as to give not less than ten progeny from as many parents as possible. Since the selections were approximately equal in both directions from the means of the clones, the average number of tentacles of the two may be used for comparison, although the coefficients of variation may be affected and complications are introduced in the analysis of the data for the problem of inheritance within the cloné. The breeding was continued until the latter part of August, when the first experiment was brought to a close, and the polyps on hand were placed in mass cultures and left for two weeks. At the end of this time a single polyp was taken from a mass culture of each of the clones and used as the starting point of a subordi- nate clone (A’ and D’) bred in individual cultures as in the first experiment, except that no Elodea was included in the cultures. Differences in number of tentacles and size. The total number of pedigreed descendants of ‘stem mother’ A obtained before INHERITANCE IN ASEXUAL REPRODUCTION 177 September was 1353, of ‘stem mother’ D, 1395. The distri- bution of the variations in the number of tentacles of the two clones is given in table 10. The means and standard deviations computed from these figures are: Mean o ; ClonerAr reese: 6.468 =0.013 0.7371 Tentacles Clones ira cette 5.739+0.011 0.6086 Tentacles Difference (A—D) 0.7240 .017 Tentacles The difference between the average numbers of tentacles borne by polyps from the two clones is here many times its probable TABLE 10 The distribution of variations in the number of tentacles of the descendants of ‘stem mothers’ A and D a NUMBER MOR UMN TAGES cz... /ecsisvs-s cieversimicicleges,e siere-s1sissia'« 4 5 6 7 8 9 Number of polyps bearing each num- ber of tentacles CHGS) 2/4 ars a ec ore oe ncn en 9 93 | 590] 589] 69 3 Clones. ocehiss-: Sey resets oo 0's 30 | 394] 883) 86) 2 0 Percentage of all polyps bearing each number (CIIGING Al Mase hatotgo tue clei ae bration Aiea 0.6 || 6.9) 43.6] 43.59) 5:2) ‘Ob2 CHONG JD sy agers Se en ate iO ee PPA Meh. || (o8ioa) |), (O21 |), Weil error and proves that the two clones differed either in their hereditary constitution or in the environment to which they were subjected. The conditions of the experiment seem to rule out the latter possibility and further evidence against an environ- mental cause of the difference was given by further experiments with the clones. The second part of the experiment, beginning with the establishment of the two subordinate clones A’ and D’, from single individuals taken out of mass cultures gave 204 pedigreed descendants from stem mother A’ and 153 descendants of stem mother D’. The number of tentacles of these polyps is shown in table 11. The constants for the distributions given are: 178 K. °S. LASHLEY Mean number of tentacles o (COME RATS sree tee he cs an er eee ees 6.907 +0 .026 0.557 Clone Dies Ree pger cg sales ssce oo asa eee 5 .844+0.029 0.537 Differences (Ava——D))\.0..n eras weno 1.0630 .039 After two weeks in mass culture the difference between the two clones persisted and was, indeed, increased by 35 per cent. The significance of this increase will be discussed later. TABLE 11 Variations in clones A’ and D’ NUMBERORUTENIDAGUESS « o..€ lero aise « «140 oe eee ioe hee enE 4 5 6 7 8 Number of polyps COM CA Bacio Sie athiccd aves Zee ee 3 33 148 20 Clone Reins seiko s.c clk ee eees es chen 2 29 114 i ft Similar results were obtained from populations of the two clones grown in mass culture. The results from five pairs of mass cultures will be considered. . Two battery jars were filled with pond water; to each afew Elodea stalks free from foreign Hydras were added; and each was inoculated with 50 polyps, the one from Clone A, the other from Clone D. Cyclops were added every week for three weeks. At the end of a month 100 polyps were taken from each culture and the numbers of their ten- tacles were recorded. The variations were distributed as shown in table 12, I. The constants obtained from this are: ‘ Mean number of tentacles o Clore Ane eens cs... 3 toad eee 6 .56+0.056 0.8284 ClO E AD rhe rs ‘avec ae ee 5 .86+0 .033 0.4903 DitierenceacAs—a))). 9.5. eee 0.700.060 Five weeks later samples were again taken from these cultures. Their variations are shown in table 12, II. The constants for these groups are: Mean number of tentacles o Glome tAs., 3. Gs oe oo ee eee 6.797 +0 .042 0.5038 CLOMID: . eR ee eh he te te) ee 5.457 =0.045 0.5539 Differences (CAg—D) 8. ee eee 1.340+0.061 INHERITANCE IN ASEXUAL REPRODUCTION 179 TABLE 12 Distribution of variations in samples taken from mass cultures NUMBER ORO TH NILA CLES «clotaapletriersin gece srels ais auerareialsjsianeierstsiayere 4 5 6 ia 8 I After three weeks with plentiful food Ome AR ten ee eA Rrra eet eras cotuee NOUNS il 9 33 47 10 ClOMCHD Rr ee eee ee re ee 20 74 6 Il After five weeks more under same con- ditions (Gov eVey 74 ee A eS A Ae a ee oO ED 1 13 48 2 Cloner tween ate ee see eee 2 34 34 III After one month with little food GlomesAbe ae Fas ee SH tees ashi s Boake 1 17 28 4 COMER ae Os AS Pies ee Oe Reva. 26 24 IV After three months in mass cultures lone vAve etre Cee ech tee ee phe oe Se 1 12 33 4 (Ciieov WO el Re PER Ue iat eee eae ae be 25 24 1 ty - Polyps from mass cultures to which no food had been added for one month showed the distribution of table 12, III. The constants for these groups are: Mean number of tentacles o Glome AU aay PR aa ey ae ANEE done ig MENS cies 515) 6.70+0.06 0.6403 Clonee EN erste ine ss 5s oo ROMA SOROS 0.4995 Diftenem cen CAb— MD eae eile fee ors cb gue es on0r 1.22+0.08 The samples recorded in table 12, IV, were taken from mass cultures three months old. The constants for these groups are: Mean numter of tentacles o (Toner Al eee teva a eer SE ee one «suche 6.800 .06 0.6000 CONCH) Reh eee ene een SS ys ==(0) 0 0.5380 Ditierentces CAR eID) heer Ns Fak igure cle cloves 1.28+0.08 * The Hydras in one pair of cultures were given very little food for three months and the water in the cultures was changed very rarely. As a result of these unfavorable conditions the polyps in both cultures were very much reduced in size and in number of tentacles. Clone A was most affected but the difference between the clones persisted even here, as the following constants for these populations show: THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 2 180 K. S. LASHLEY Mean number of tentacles o Clone Antes oo ee ee 5.519+0 .035 0.5825 Clone erence oes Sak Cee eee 5.346+0 .033 0.4970 Differences (4e— 7-1 days. cee a: Clone D 7.8 days The difference in size is even more pronounced than the differ- ence in the number of tentacles: individuals from Clone A were, on the average, more than twice as large as those from Clone D.* The difference in the appearance of the polyps is shown infigure 4. The numbers measured are relatively small but they extend 3 An attempt was made to determine whether the difference in size was due to a difference in the size of the individual cells, in accordance with the theory of the relation between cell and body size advanced by Popoff (’08). Because of the great mobility of Hydra, accurate measurements could be obtained only for the mature nematocysts. In this character the two clones were identical and if it furnishes an index to the size of other cells the difference between the clones was due to a difference in the number and not the size of the body cells. INHERITANCE IN ASEXUAL REPRODUCTION 181 ~y Imm Fig. 4 Diagram of the size relations of Clones A and D. From camera draw- ings of polyps of mean size. over more than half the period of cultivation so that environ- mental differences should have averaged out. The polyps measured were selected because of their relationships, so that the personal factor, an unconscious selection of extremes, can have played no part in producing the difference. All the differences recorded are between groups which had been kept under environmental conditions as nearly uniform as possible. The polyps in individual cultures were fed in irregular ‘order in order to insure a uniform distribution of food. In the mass cultures the parallel series of the two clones were given uniform treatment and an interchange of water in the parallel cultures was made frequently. The evidence seems to prove conclusively that some internal, hereditary factor caused the differences between the clones. This conclusion is still further verified by an analysis of the data from the individual cultures. In order to test the constancy _of the difference, the time during which the clones were culti- vated was divided arbitrarily into intervals of five days and the mean number of tentacles of all buds produced by each clone during each five-day period was computed. Table 14 and figure 182 K. S. LASHLEY TABLE 14 Constants for the number of tentacles of the buds produced by clones A and D during successive five-day intervals from June 25 to August 23 CLONE A CLONE D (A — D) NO. OF PERIOD aoe ten- 6 Meaulno: pt ten- a Toutacten I SRM IeUS Sa 6.5380 +0 .099 0.6056 | 5.500+0.168 | 0.5000 1.030+0.194 Zessooeensos|| Oa ls==(0) IIKs 0.8913 5.462+0.091 0.6919 0.653 +0.149 SoA oe ee 6 .223+0.046 0.7280 5.479 +0 .036 0.6204 0.7440 .053 CL Ade 6.142+0.055 0.8478 | 5.345+0.027 0.5679 0.796+0.061 Nossecoronoc|| O.28R=S0 055: 0:7754 | 5.441+0.036 | 0.6257 0.7920 .065 Geet ones reece 6.552 +0.031 0.6529 | 5.796+=0.029 0.5532 0.756+0 .042 Ue 3: es 6.795 +0 .030 0.6268 6.103+0.031 0.63886 0.692 +0 .043 Socdosbonoce| Wo oil==() 08S} 0.6334 | 5.958+0.0238 | 0.4542 0.723 +0.040 Us ere eet 6.439 +0 .0356 0.6499 5.757 £0 .029 0.5078 0.682 +0 .046 MWecsscescoedell Oo2bG==(0) (05S 0.7502 5.787 +0 .080 0.4980 0.4490 .042 LL emia ees 6.396+0.043 | 0.5958 | 5.787+0.036 | 0.5376 0.972+0 .056 UL Set Gone 6.3840 .090 0.68386 | 5.775+0.090 | 0.4875 | 0.609+0.128 5 show the result of this analysis. Throughout the entire experi- ment the difference between the clones persisted. By July 25 Clone D had increased in the average number of tentacles of its buds until it equaled the earlier average of Clone A (June 30) but this was accompanied by a corresponding increase in the average of Clone A. The figure shows that the difference re- mained fairly constant in spite of wide fluctuations in the absolute values of the means of the clones. Other differences between the clones. There is a marked differ- ence between the clones with respect to the age at which repro- ductive maturity is reached, the age at which the parents produce their first bud. This is shown in the following tabulation: Mean age of parenis when No. of Jirst bud was produced o parents ClonerAr re eee 4.806+0.103 days 2.422 248 CloneyOR cose eee 3.748+0.074 days 1.755 253 Difference (A — D) 1.058+0.127 days The rate of reproduction of the clones differed slightly, that of Clone A being somewhat greater than that of Clone D. The © difference, based upon 33 parents of each clone which produced more than 10 buds is 0.0542 buds per day. This difference is INHERITANCE IN ASEXUAL REPRODUCTION 183 N i?) | Sy Se 10 20 30 40 50 60 70 80 90 100 Fig. 5 The mean number of tentacles of the buds produced by Clones A and D during successive intervals of five days. The mean number of tentacles of Clones A’ and D’ are included at the right as these form a natural continuation of the earlier experiment. not greater than its probable error and is probably not signifi- cant. The equality in reproductive rate is shown by the fact that while no attempt was made to obtain equal numbers in the two clones during the experiment, the numbers bred were, after two months, almost equal (1353 and 1395). Other differences between the clones which can not be ex- pressed quantitatively were apparent in the living animals. Clone D was, as a rule, somewhat darker in color than Clone A and somewhat more resistant to unfavorable conditions. Polyps of Clone A showed a greater tendency to cling to the surface film and were more active than those of Clone D. Nature of the difference between Clones A and D. The fore- going data prove that the two clones studied are distinct through the presence of some internal, hereditary factor. The ‘Reak- tionsnormen’ of the two are different. But whether this differ- ence is genotypic in the sense of being the result of a fixed con- stitution of the germ-plasm or whether it is due to a spurious heredity like that by which the green color is transmitted is not clear. The agents, other than the hereditary constitution, which ~ might produce such clonal diversities are the direct action of the 184 K. S. LASHLEY environment, the persistent effects of former diversities of environ ment, variations in the commensal relations of the polyps and their zoochlorellae, parasitic infection of one clone, and differ- ences in clonal age from the fertilized egg. The first of these is eliminated by the methods of the experiment. It has been shown that the after effects of injury and ‘depression’ are only temporary. Experiments attempting to produce permanent modification by extremes of temperature, chemical agents, and artificially induced ‘depressions’ have been unsuccessful; diver- sities are readily produced but they persist for only a few days. Possible variations in the commensal relations of H. viridis are more difficult to control. Microscopic examination of the two clones failed to reveal any difference in the green bodies of the two clones. As a test of the influence of commensalism upon variations I partially removed the zoochlorellae from some individuals of Clone ZL (see below) by the method devised by Whitney (06-08). During immersion in glycerin the polyps became smaller but when restored to normal conditions they ate regularly and resumed their original size before beginning to bud. The buds, practically white in color, did not seem other- wise different from the normal members of the clone. I never succeeded in getting out all the zoochlorellae from the polyps so that not enough buds for statistical study could be obtained before the polyps resumed their normal color. Five parasites of Hydra have been reported: Amoeba hydrox- ena Entz, Balantidium hydrae Entz, a species of Ophryoglena (Entz ’12), Trichodina pediculus Ehr. (Clarke ’65), and Kerona polyporum Ehr. The true parasitic nature of only one of these, A. hydroxena, has been demonstrated. Trichodina did not occur in any of the pedigreed cultures recorded here, although it is common in wild populations. In October, 1913, an epidemic of A. hydroxena broke out in my cultures, destroying all of them. The symptoms of infection by this parasite are easily recogniz- able and there is no possibility that it occurred in the earlier cultures. Kerona occurs rarely in wild populations here but did not appear in my cultures; the other two parasites have INHERITANCE IN ASEXUAL REPRODUCTION 185 never been seen in this locality. The fungus mentioned by Rand (99) is evidently not a parasite of Hydra. Finally, there is no direct evidence bearing upon the question of the effects of the age of the clone upon its variations. It has been impossible to secure clones of known age since in this locality sexual reproduction in H. viridis seems to be almost completely suppressed. Hydras from wild populations have been examined at all seasons during the past three years and in this time not one with ovaries and only seven with spermaries have been found. The treatment by which Whitney (07) induced sexual repro- duction was tried but without success. These facts seem to justify the conclusion that sexual reproduction plays but little part in the life history of the green polyp. Sos Pa 100 150 200 250 300 350 400 Fig. 6 The number of polyps (ordinates) which died at the ages in days recorded on the abscissae, during Hase’s experiments. (——) H. fusea, (---—-) H. grisea. In the development of theories of ‘life cycles’ in micro-organisms the phenomena of ‘depression’ in Hydra have had a prominent. place. R. Hertwig, in particular, has sought to emphasize the relations of ‘depression’ to sexual reproduction, but the work of Frischolz (’09) seems to prove that no such relation exists. Practically all studies of depression have been carried out with polyps in mass cultures and there is no evidence that the epidemics of depression reported were not due to environmental rather than to internal factors. The ease with which depression may be induced by unfavorable environment and cured by putting the polyps in clean water makes it practically certain that depression is a pathological condition. The only other work bearing upon the question of age is that of Hase, who studied the age of individual polyps. He found a mean age of 55.2 days for H. fusca and 94.8 days for H. grisea, with the distribution 186 K. §. LASHLEY of ages at death shown in figure 6. The irregular form of the curves and the fact that the vast majority of the polyps died at a relatively early age scarcely speaks in favor of an ‘Altersschwiche.’ The curves are similar enough in form to suggest that the deaths: were due to two periods of unfavorable conditions (a and b) in his cultures. At no time have I observed an epidemic of depression in the individual cultures belonging to the same clone; less than 1 per cent of the polyps have shown depression and always the closest relatives of these remained normal, a condition which is not at all in harmony with theories of- clonal senescense. Further, during the four months that the clones were kept under obser- vation, including more than 20 asexual generations, there was no significant change in the character of either clone. The repro- ductive rate, the first character modified by a reduction of the vitality of the polyps, was as high at the end of the experiment as at the beginning. This furnishes some evidence against the belief that the diversities are due to the clonal age but it is not conclusive. The existence of other races A comparison of the standard deviations of the Clones A and D reveals the fact that clone A was always the more variable of the two. This difference suggested that Clone A was not pure and an examination of other evidence quickly confirmed this view. The variations of the clone tended to a dimodality which does not appear in any other race studied and the greater varia- bility and dimodal form of the variation curve did not persist in Clone A’ which was descended from a single polyp taken from Clone A. The records of the clone were reviewed in order to test whether the greater variability was characteristic of the race or was the result of the inclusion of diverse types as is sug- gested by the condition in Clone A’. One subordinate clone represented by the direct line of descent AlA2blala2ala (the 4 Agar (’14) has shown that diverse clones produced by parthenogenetic re- production may be produced by different eggs hatched at the same time. The differences here are clearly not due to the ages of the clones. Lang (’92) found that only one germinal layer of Hydra takes part in the formation of the buds, and this work, while unconfirmed, suggests that the differences between partheno- genesis and budding may not be so great as is generally supposed. INHERITANCE IN ASEXUAL REPRODUCTION 187 successive buds of alternate generations being labled with the series of letters and numerals) showed an average number of tentacles considerably lower than that of the remaining por- tion of the clone. The polyp A/A was recorded as small, with abnormal tentacles, rapidly becoming normal. The mean num- ber of tentacles of the fraternity from which it came was 6.809 + 0.062, that of its immediate progeny was 6.200 +0.023, giving a reduction of 0.609 +0.067 tentacles in one generation. Fig. 7 The distribution of variations in the number of tentacles of Clone A ) and of the subordinate clones A1A (———) and A — (A1A) (—--—--). ( The mean of all the descendants of A/A (437) was 6.121 = 0.020: that of the remainder of the clone was 6.624+0.013 giv- ing a difference of 0.503 +0.024. The distribution of variations in the two subordinate clones, separately and combined, is shown in figure 7. The dimodality of Clone A is clearly the result of the combining of these two groups (neither subordinate clone breaks up upon further analysis nor does any other method of dividing Clone A reduce its dimodality. The appearance of this diversity gives no evidence of a cumulative inheritance of variations. The change was quite sudden, resulting in a difference after one generation as great as the difference in any later gener- 188 K. S. LASHLEY, ation. The individual A7A must be looked upon either as a mutant or as, more probably, a polyp of another clone introduced by accident into the culture.6 It seems to be hereditarily differ- ent both from Clone D and from the remainder of Clone A. Diverse races which can be recognized at a glance are extremely rare in the populations near Baltimore. Several times fifty Libs Fig. 8 Diagram of the size relations of two clones kept under similar con- ditions, L and F. From camera sketches of individuals of estimated average size. or more clones have been started with polyps from different localities but only twice have distinct races appeared which were recognizable within a few generations. The first of these were Clones Aand D. In August, 1913, a small polyp was found which produced very small buds. At first it seemed unhealthy but the buds increased in size and began to multiply rapidly. Un- fortunately the culture could not be continued and the clone was 5 The opportunity for contamination was given by the use of Elodea stalks in these cultures, as in the early part of the experiment no certain method of free- ing them from foreign Hydras had been devised. INHERITANCE IN ASEXUAL REPRODUCTION 189 TABLE 15 A comparison of the mean number of tentacles of clones T, L, and R CLONE NOVO pee OEE meme enet eae NO. OF DESCENDANTS cD ee ee ane 6 6.553 +0 .058 56 TOR eee esr 6 6.282 +0 .049 46 [Ras cia he ROS OR 8 6 .650+0.081 40 ISCO e eR aoe 6.513 +0 .033 154 EM eee 6.184+0 .033 152 Difference (7’'— L) 0.828+0.046 lost before many individuals had been obtained. However, the difference between this and another clone, L, kept under the same conditions was so marked that it seems certain that the small race represented a distinct genotype. The size relations of the polyps of the two clones is shown in figure 8, taken from camera lucida sketches of estimated average individuals. More frequently the differences are less pronounced. Table 15 gives the constants obtained for three clones bred during July and August, 1918. Clones 7’ and Z were continued after R was discarded and the constants for their full numbers are given separately. They were grown in individual cultures under parallel conditions as in the first experiment and seem to be hereditarily distinct. These are the only races which I have cultivated under exactly parallel conditions and no evidence upon the number and variety of such races is available. Varieties of Hydra in the earlier literature. The extensive literature dealing with Hydra gives frequent suggestions of the existence of local varieties of the three generally recognized species, and one or two descriptions seems really to establish their existence. The statements concerning the size of H. viridis made by Baker, Trembley, Rosel, Pallas, and Kastner differ considerably but are not conclusive. Marshall (’82) de- scribed a variety of H. viridis from brackish water which remained distinct from the fresh water forms even after cultivation for 190 K. S. LASHLEY many generations in fresh water. The systematic literature and the studies of gonad production by Brauer, Downing Nuss- baum, Weltner, and Koelitz, suggest the existence of monoecious and dioecious varieties. The experiments of Tower show two types of H. viridis differing in their reaction to the ultra-violet rays (?) and in the time required for regeneration. Annandale has described a variety of H. grisea from India which is char- acterized by having a four-tentacled winter form producing gonads. Whitney has been able to establish varieties of H. viridis without green bodies. Summary To sum up the present section of the work: It has been found that within a wild population of H. viridis there are hereditarily diverse races which differ in their number of tentacles at sepa- ration from the parent, in their size at a given age, and less cer- tainly in other characters. The differences between such races are permanent so long as the races are kept under the same environment. The evidence favors the view that the differences are truly genotypic (with the reservation that they are possibly the result of differences in the age of the clones). IV. INHERITANCE OF VARIATIONS WITHIN THE CLONE Inheritance of variations within the clone may be tested in two ways; first, by an analysis of the resemblance between parent and progeny by statistical methods; second, by a con- tinued selection of variates, which should change the type of the clone if the variations are inherited. As has been shown in the discussion of Hanel’s data, selection of variates within the clone seems, on the average, to have no effect, yet there is usually a positive correlation of parent and progeny with respect to the selected character. Pearson has subjected Johannsen’s data for the selection experiment with beans to a similar analysis and ° I have made many attempts to repeat Tower’s experiments, using different types of arcs, with and without interposed glass, but have never obtained the sloughing of the ectoderm which he describes. INHERITANCE IN ASEXUAL REPRODUCTION 191 has shown that in these pure lines there is likewise a correlation between parent and progeny. His explanation of this seemingly anomalous condition and the criticism which has been urged against selection experiments in general where Vilmorin’s method is employed is based upon the small numbers selected. Pearson (10), discussing Hanel’s results, says: In selecting a few isolated individuals in each generation, where non-hereditary influences are so influential, we may break the influence of heredity at each step, and since such influences are equally effective with heredity, the chances are that we will do so once in every two selections. Only by taking large numbers of the high and large num- bers of the low would it have been possible to average out the effects of environmental changes. The only ways in which this criticism can be met are by very extensive selections involving large numbers of individuals, or by a study of variation and environment which shall make possible a differentiation between heritable and non-heritable variations, upon some other basis than that of ancestral correlation. Resemblance of close relatives within the clone Number of tentacles. In all breeding experiments the polyps have been so recorded that the number of tentacles of the buds can be compared with that borne by their parents, both at the time when the buds were produced and when the parents themselves were buds. Comparisons of both numbers must be made in order to answer two distinct questions: (1) do the offspring resemble their parents when the latter were in the same stage of development: (2) are the variations of the parents during their development transmitted? To test the first question the correlation between the initial number of tentacles of the parents and that of their buds was computed. Since there is reason to believe that Clone A is impure, it was divided into the two subordinate clones mentioned earlier and the constants were computed for these. If there is a cumulative inheritance of slight variations this division should affect only the probable errors of the correlation constants. The 192 K. S. LASHLEY coefficients of correlation obtained in this way are given in table 16. Only one of these, that of Clone D’ is more than twice its probable error, and this one is negative. The others are all too small to have any significance. The large negative corre- lation of Clone D’ is the result of the inclusion of a single parent (the stem parent of the clone) which produced a large number of progeny with few tentacles and was thus heavily weighted in the computation of the coefficient. The Clones A — (A1A) and D contain such large numbers both of parents and offspring that it is almost certain that the low correlation is not the result of chance, but truly expresses an almost total absence of similar variations in parent and progeny within the clone. In so far as the coefficient of correlation is a reliable test of heredity we may conclude that there is no inherit- ance of variations in the initial number of tentacles of Hydra viridis. For Clone D the fraternal and grandparental correlations were likewise computed; they are givenin table 17. Like the parental correlation, the grandparental is too small to have significance. TABLE 16 Correlation of the initial number of tentacles of parents and progeny within the clone r NO. OF PARENTS |NO. OF PROGENY Clonee RR mr tan obec ce eee 0.0038+0.018 251 1395 @lone Ay CAA iota sss 52 0.0011+0 .023 859 Clone sAViA Peano Sate 0.0342 +0 .032 439 ClOme De we crrwaseass. 5 4 aie 0 .2420+0 .051 18 153 @lonevAy sas seiGeeee seo carn 0.0310 +0 .047 28 204 (Gl losoVesnel Me Sarai ore oid oe ene 0.0750 +0 .054 51 154 TABLE 17 Ancestral correlations, clone D ip NO. OF PARENTS ae eataa a Baremtall’. ts: Seperate cyenvo fe 0.00380 .018 251 1395 Grand parentiallieeenneess seiner: —0.0495+0.018 68 1307 LOSER SY Oe IERIE cog o oubloteere enero: 0.07700 .006 10766 INHERITANCE IN ASEXUAL REPRODUCTION 193 The fraternal is only slightly greater and its value is doubtful. It will be considered again in the discussion of the cause of an- cestral correlation within pure lines and clones. The second question, do the parents transmit to their buds the characters which they have when the buds are formed, in- volves one of two concepts. Either the effects of environ- mental action must be transmitted or the organism must con- tinually change in its hereditary potentialities during its develop- ment in order that its transient characters shall be inherited. ‘Something of the latter conception seems to be implied in Pear- son’s doctrine of homotyposis. TABLE 18 Correlation between the original number of tentacles of buds and the number borne by the parents at the time when each bud was produced r NO. OF PARENTS |NO. OF PROGENY (CU aYS) IO ey eae ie re ea 0.096 +0 .016 251 1395 WlonerAre aay Geminis as eaten 0.240+0 .009 242 1353 Clone A. Grandparental.... 0 .229+0.013 1094 The correlation of the number of tentacles of the buds with that of the number borne by the parents when each bud was produced has been computed for Clones D and A; these are given in table 18. The high correlation of Clone A is obviously the result of the inclusion of the two diverse strains but a compari- son of the parental and grandparental correlations within this clone shows that there is a slight positive parental correlation which is not the result of the mixture of the two types. (The parental correlation would not be higher than the grandparental if the inclusion of diverse types were the only cause of correlation). This correlation is certainly very small, and that of Clone D is also too small to have significance when considered alone, but the fact that these correlations are greater than the ones obtained for the initial number of tentacles indicates that there is here some factor tending to produce a resemblance between the mature parent and its buds. This is also the impression 194 K. S. LASHLEY given by the average parental correlation of 0.101 found for Hanel’s clones. What is the significance of such a small coefficient of corre- lation in an organism so subject to environmental influence as is Hydra? Pearson holds that the presence of variations due to environment would tend to obscure the real correlation be- tween parents and offspring and hide any real inheritance which might exist. My data upon the relations of variation to environ- mental changes indicates that the latter may be more effective in producing a likeness between close relatives than in obscuring * such a likeness: Some conditions producing like variations in parent and offspring in Hydra have been considered already. Any great diversities in the conditions of the cultures would re- result in a correlation between parent and progeny, even though there were no inheritance of the variations studied. Such con- ditions have been found by Agar in daphnids and plant lice and probably account for the ancestral correlations found by Warren in these forms. The production of a correlation in Hydra by the action of diversities of environment may be illustrated by a consideration of the change in the mean of clones during long periods of culti- vation. Figure 5 shows that the mean number of tentacles of the buds produced by Clones A and D during the first few | weeks of cultivation was low; that it increased gradually during the first six weeks, and then decreased again. All the buds produced during each of these five-day intervals were correlated with one another. This gave the following results: Coefficient of correlation No. of pairs Clone. Ay pee eee Se 0 Ee 0 .0774+0.0015 95141 Clone 1)e.s 3A ie Soe tee 0.1313+0.0014 101872 Correlation of all buds produced in each five day period with all produced in the preceding period (a time interval correspond- ing to a full generation) gives for Clone A a correlation of 0.048 + 0.001 which is greater than any of the parental correlations found for the initial number of tentacles. INHERITANCE IN ASEXUAL REPRODUCTION 195 These correlations show a resemblance between the buds produced within a limited time as great as that found between the closest relatives, although these buds were no more closely related to each other than to those produced during other five day intervals. This same effect would be visible in the cor- relation between parent and progeny whenever several gener- ations are included in the same correlation table, as is necessary in the case of Hydra. Such effects of environmental action seem adequate to account for all the coefficients of correlation given by Hanel’s data and for the very slight positive correlations found for the initial number of tentacles in my own clones. The fact that the number of tentacles of successive buds increases with the increase in the number of tentacles of the parent accounts for the slightly larger correlation found between the number of tentacles of the parents and. offspring recorded when each bud was produced. Whether the increase in the number of tentacles of the buds is an inherit- ance of the variations taking place in the parent with growth, or is only a temporary effect of the increased vigor of the parent, must be tested by the success or failure of an attempt to modify the character of the race by the continued selection of variates. The effects of selection within the clone The second method of testing inheritance, that of continued selection of variates, offers a good many practical difficulties in Hydra owing to the sensitiveness of the polyps to environmental changes. Gonncee seu euaoe se 0 .058+0 .035 0.056 Grandparent and grandchildren........ 0 .018+0 .036 0.020 Clone A arent aU G OLS PEIN 5 Jot... ./ cua ayers 0.358 +0 .030 0.285 Grandparent and grandchildren........ 0 .030+0.036 0.048 Clone A 1A Parent and offspring...... Ce iene, 0 .009=0.106 Clone A — (A1A) Rarentigand OffSpLrING. ac.o00-5sa.4 -doe- 0 .255+0 .063 200 K. {S TASHDEY. correlation of Clone A shows that the high parental correlation is not wholly the result of the mixture of two races in Clone A. Computation of the parental correlation for the two subordinate clones confirms this, giving the following coefficients: Coefficient of correlation Clone Anam ie of Fo ce belt at ae ee hot ae hai s asters 0.009 =0.106 Clones Avs CAMA e. cote = fohaa,s OR RRP ECoG. o sins geen 0 .255+0.063 After this division the correlation is still significant and is four times as great as that found for Clone D. An examination of the data for possible environmental causes of correlation . S6L cu.mm. Fig. 10 Changes in the average size of the buds produced during successive ten-day intervals by Clones A and D. The ordinates represent the mean sizes in cubic millimeters of all measured buds produced during each ten-day period; the abscissae, the successive ten-day periods. gives the result shown in figure 10. As in the case of the number of tentacles, there was a considerable change in the mean size of the buds produced at different times during the experiment. The correlations between all buds produced in each of the five-day intervals were: Coefficient of correlation GLO ID x ce es Bs lok Oe rhs! sre RE oe ener: tape ae 0.0118 (Oj roneve ter ean ©4104), i RIPEN as .cre'c o ootln RaeoeGe me DON Hameo Bic 0.0952 There is here the same difference in the size of the correlations of the two clones which appeared in the parental correlation but INHERITANCE IN ASEXUAL REPRODUCTION 201 ' the correlation due to the changing clonal mean is too small to account for the parental correlation. No adequate data for a further analysis of the correlation in size is at hand, and it may be that variation in size within the clone is actually inherited but, although conclusive evidence against this is lacking, too much trust should not be placed in the parental correlations. The following points, while based only upon general impres- sions gained during the experiments and subject to correction by further experimental test, indicate some of the factors which may have been instrumental in producing a high parental cor- relation for size. Size is a character which is modified much more readily and quickly than the number of tentacles by changes in the environment. Hydras, placed in a 0.1 per cent solution of NaCl, in a very few days become dark in color, small, and produce small, dark-colored buds. Similar ehanges in size follow extreme changes in temperature. Upon the restoration of opti- mum conditions the normal size is resumed very quickly. Star- vation is effective in much the same way but to a lesser extent. When a Hydra has been injured slightly, or has passed through a slight depression it grows smaller and produces small buds but eventually both buds and parents resume the normal size of the clone. The size, indeed, seems to be a matter of the immediate state of nutrition of the polyps. Clone A was more readily modified by such agents than was Clone D. This is shown by the following data for the number of tentacles of polyps from mass cultures with and without food for three months: 38-month mass cultures With food Without food Difference G@Yomeg Ale ae eo 5 tei cate ee oie .. .6.80+0.06 5.52+=0.04 1.28+0.07 ClON CED Meee ean aia aes 5.02=£0).05 5.34+0.03 0.18+0.06 and by the changes represented in figure 10. Corresponding to this fact, the correlations between the relatives in Clone A are higher than in Clone D, which justifies the suspicion that the correlation between parent and offspring for size within the clone is really due only to the action of the environment. 202 K. S. LASHLEY V. DISCUSSION OF RESULTS The experiments reported show that populations of Hydra viridis consist of races which have different hereditary consti- tutions. The diversity between two such strains persisted for so long as they were kept under observation (143 days) and so long as the different strains were kept under similar and favor- able conditions they showed no tendency to approach each other in character. The distinguishing characters of the different races studied were not, however, fixed in the sense of remaining constant through fluctuations in the environment but under- went changes corresponding to changes in the environment. In general, the diverse clones responded to such changes in the same way and to the same degree, although in the face of very unfavorable conditions the larger strain was most affected and under conditions of almost complete starvation the strains became much more similar. Little evidence has been obtained as to the cause and funda- mental nature of the difference between the clones. The diverse characters noted, number of tentacles, size, color, and age at which asexual reproduction is begun, may all be modified by changing the environmental conditions and the changes thus produced in the first three characters mentioned are correlated in the same way that they are in the diverse races (large size, many tentacles, and light color occur together) so that it is pos- sible that the diversities in these three characters are due to a difference in some single set of physiological processes. Failure to obtain sexual reproduction in H. viridis has made it impossible to determine the relation of the diverse races to gametic processes and to the supposed life cycle of Hydra but there is certainly no relation between the diversity of the clones and the phenomena of ‘depression’ which have been thought to mark periods in the life cycle. The existence of diverse races of Hydra is in accord with the results of Jennings, Woltereck, Shull, Whitney and Agar gained from the study of clones of other invertebrates and, indeed, the volume of evidence from zoological and botanical literature INHERITANCE IN ASEXUAL REPRODUCTION 203 leaves no doubt that the existence of diverse races within the species is a general condition in all phyla. The problem of inheritance of variations within the clone presents much greater difficulties and there is much conflict between the results of different investigators. The work of Whitney has shown that in Hydatina diverse strains may arise in a clone descended from a single fertilized egg and Calkins and Gregory report similar results for Paramecium. In these cases: there is, however, no intimation that the inheritance of variations is a general characteristic of asexual reproduction or that the change is the result of the accumulation of slight variations. The four studies in which there is an appearance of inheritance of continuous variations within the pure line or clone are those summarized by Pearson in 1910; the studies of Johannsen on beans, Warren on Daphnia and Hyalopterus, and Hanel on Hydra. In all this work the evidence for inheritance can be drawn only from the ancestral correlations, while the evidence from the effects of selection seems to point the other way. ‘The question of the relative values of the coefficient of correlation and of selection experiments hence becomes of great importance. In Agar’s recent study of inheritance in parthenogenesis, where great precautions were taken to rule out the influence of environmental agents, there is no significant correlation between parent and progeny within clones of Cladocera and sufficient evidence of environmental causes of correlation is presented to account for Warren’s results with Daphnia. For Aphids he finds, with Warren, a slight ancestral correlation but the evi- dence for an environmental cause of this correlation, while per- haps not absolutely conclusive, is sufficient to make the ancestral correlation in these forms of very doubtful value as an index of inheritance. Ewing’s selection experiments with Aphis avenae offer further evidence against the inheritance of variations in these forms. The results recorded in the present paper are quite in accord with those of Agar on Cladocera and show that for Hydra also the ancestral correlation is untrustworthy as a measure of inheritance. 204 K. S. LASHLEY There remain only Johannsen’s data on Phaseolus giving evidence of an ancestral correlation within the pure line. The evidence adduced by Pearson indicates only a slight correlation here and the recent work of Harris (712) upon the transmission of the effects of unfavorable cultural conditions indicates that the real cause of this correlation is likewise environmental. VI. SUMMARY 1. Populations of Hydra consist of hereditarily distinct strains which differ in initial number of tentacles, size of body, color, age at which asexual reproduction is begun, and perhaps in other characters. 2. In the absence of selection these strains remain distinct. 3. Within populations there is a correlation between the - characteristics of parents and progeny and of other close relatives, which is largely due to the existence of these diverse strains. 4, Within the clone there is no significant correlation between the variations of close relatives in the initial number of tentacles. 5. Within the clone there is a slight correlation between the number of tentacles of the buds and the number of tentacles which their parents bear when each bud is produced. 6. Diversities of environment tend to produce like variations in parents and offspring and this likeness tends to disappear when the environmental cause is removed. The existence of such environmental agents is sufficient to account for the ances- tral correlations found, even though there is no inheritance of variations. 7. Continued selection of variates in tentacle number results in changes in the vigor of the selected groups. This results in an apparent diversity of the differentially selected groups but the diversity persists only during selection and disappears at once when selection is discontinued. Variations in the number of tentacles of Hydra viridis are not inherited. 8. There is a positive correlation between the variations in the size of parent and offspring within the clone. No statis- tical evidence of an external cause of this correlation is presented INHERITANCE IN ASEXUAL REPRODUCTION 205 but from general considerations it seems probable that. this, like the correlation of variations in the number of tentacles, is due wholly to the similar action of environmental agents upon parent and offspring. / LITERATURE CITED Acar, W. E. 1913 Transmission of environmental effects from parent to off- spring in Simocephalus vetulus. Phil. Trans. Roy. Soc. London, vol. 203 B, pp. 319-351. 1914 Experiments on inheritance in parthenogenesis. Ibid, vol. 205 B, pp. 421-489. ANNANDALE, N. 1906 The common Hydra of Bengal, its systematic position and life history. Mem. As. Soc. Bengal, vol. l, pp. 339-359. 1907 Notes of the fresh-water fauno of India. No. 10. Hydra orientalis during the rains. Journ. As. Soc. Bengal. vol. 3, pp. 27-28. Brepot, .M 1912 Sur la nomenclature des Hydres. Zool. Anz., Bd. 39, p. 602. Braver, A. 1908 Die Benennung und Unterscheidung der Hydra-Arten. Zool. Anz., Bd. 33, pp. 790-792. CaLxins, Gary N., and Gregory, Louise H. 1913 Variation in the progeny of a single ex-conjugant of Paramecium caudatum. Jour. Exp. Zodl., vol. 15, pp. 467-525. ‘ Crarkb, JAS. H. 1865 The anatomy and physiology of the vorticellidan para- site (Trichodena pediculus Ehr.) of Hydra. Mem. Boston Soc. Nat. Histor vol. py 11. Davenport, C. B. 1904 Statistical methods. New York. Entz, Giza 1912 Ueber eine neue Amobe auf Sitisswasser Polypen (H. oligactis Pall.) Arch. f. Protistenkunde., Bd. 27, pp. 19-45. Ewina, H. E. 1914a Pure line inheritance and parthenogenesis. Biol. Bull., vol. 26, pp. 25-35. 1914b Notes on regression in a pure line of plant lice. Ibid, vol. 27, pp. 164-168. Fries, 8. 1879 Mitteilungen aus dem Gebiet des Dunkelfauna. Zool. Anz., vol. 2, p. 154. Friscuyouz, KE. 1909 Zur Biologie von Hydra. Biol. Centralb., Bd. 29, p. 182. Hanet, Exist 1908 Vererbung bei ungeschlechtlicher Fortpflanzung von Hydra grisea. Jenaische Zeitschr., Bd. 43, pp. 321-372. Harris, J. A. 1912 A first study of the influence of starvation of the ascendants upon the characteristics of the descendants. Amer. Nat., vol. 46, pp. 313-340. 206 Kk. S, ASHLEY Hasr, A. 1909 Ueber die deutschen Siisswasser-polypen Hydra fusca L., Hydra grisea L., und Hydra viridis L. Arch. f. Rass. u. Ges. Biol., Bd. 6, pp. 721-753. Jennines, H. 8S. 1908 Heredity, variation, and evolution in Protozoa. II Heredity and variation in size and form in Paramecium, with studies of growth, environmental action, and selection. Proc. Amer. Phil. Soe., vol. 47, pp. 393-546. 1911 Computing correlation in cases where symmetrical tables are commonly used. Amer. Nat., vol. 45, pp. 123-128. JOHANNSEN, W. 1903 Ueber Erblichkeit in Populationen und in reinen Linien. Jena. 1911 The genotype conception of heredity. Amer. Nat., vol. 45, pp. 129-159. . 1913 Elemente der exakten Erblichkeitslehre. Jena. Lana, AtBpertT 1892 Ueber die Knospung bei Hydra und einigen Hydro-poly- pen. Zeit. f. wiss. Zool., Bd. 54, pp. 365-885. MarsHati, W. 1882 Ueber einige Lebenserscheinungen der Siisswasser-polypen und iiber eine neue Form von Hydra viridis. Zeit. f. wiss. Zool., Bd. 37, pp. 664-701. Parke, H.H. 1900 Variation and regulation of abnormalities in Hydra. Arch. f. Entw’mech., Bd. 10, pp. 692-710. PEARSON, Karut 1896 Mathematical contributions to the theory of evolu- tion. III. Regression, heredity, and panmixia. Phil. Trans. Roy. Soc. London, vol. 187, pp. 253-318. 1898 Mathematical contributions, etc. On the law of ancestral heredity. Proc. Roy. Soc. London, vol. 42, pp. 386-412. 1901 Mathematical contributions, etc. IX. On the principle of homotyposis. Phil. Trans. Roy. Soc. London, vol. 197, pp 4438-459. 1909 The theory of ancestral contributions in heredity. Proc. Roy. Soc. London, vol. 81 b. 1910 Darwinism, biometry, and some recent biology. I. Biometrica, vol. 7, pp. 368-385. Puate, L. 1913 Vererbungslehre. Leipzig. Poporr, M. 1908 Experimentelle Zellstudien I. Arch. f. Zellforsch., XX. Bd. 1, pp. 245-379. Ranp, H. W. 1899 Regeneration and regulation in Hydra viridis. Arch. f. Entw’mech., Bd. 8, pp. 1-384. Reese, A.M. 1909 Variation in the tentacles of Hydra viridis. Science, N.S., vol. 29, p. 433. SuuLt, A. F. 1911 Studies in the life cycle of Hydatina senta. II. Jour. Exp. Zoél., vol. 10, pp. 117-166. INHERITANCE IN ASEXUAL REPRODUCTION 207 Suutit, G.H. 1912 Genotypes, biotypes, pure lines, and clones. Science, N.S., vol. 35, pp. 27-29. Tower, W. L. 1899 Loss of the ectoderm of Hydra viridis in the light of the projection microscope. Amer. Nat., vol. 33, pp. 505-509. WARREN, i. 1899 An observation on inheritance in parthenogenesis. Proc. Roy. Soc., vol. 65, pp. 154-158. 1901 Variation and inheritance in the parthenogenetic generations of the Aphis, Hyalopterus trichodus (Walker). Biometrica, vol. 1, pp. 129-154. Wuitney, D. D. 1906 Artificial removal of the green bodies of Hydra viridis, Biol. Bull., vol. 138, pp. 291-299. 1907 The influence of external factors in causing the development of sexual organs in Hydra viridis. Arch. f. Entw’mech., Bd. 24, pp. 524-537. 1908 Further studies on the elimination of the green bodies from the ectoderm cells of Hydra viridis. Biol. Bull., vol. 15, pp. 241-246. 1912 Strains in Hydatina senta. Biol. Bull., vol. 22, pp. 205-218. WotrTeRECK, R. 1909 Weitere experimentelle Untersuchungen iiber Artver- inderung, usw., bei. Daphniden. Verh. d. deutsch. Zool. Gesell., pp 110-171. Yue, G. U. 1912 An introduction to the theory of statistics. London. 208 K. S. LASHLEY APPENDIX The correlation tables from which the principal constants given in the body of the paper were obtained are appended here. In all cases where the tables include an earlier and a later gener- ation the ascendants are arranged in the rows, the descendants in the columns. In those cases where there is no qualitative difference between the members of pairs, as when siblings are correlated, the pairs are entered only once, according to the method recommended by Jennings (’11). The tables from which the correlations for size were computed are not included because of their great bulk and the lack of certain evidence upon the environmental modification of size. TABLE 24 TABLE 25 Clone D: Correlation between the initial Clone D: Correlation between the initial numbers of tentacles of parents and numbers of tentacles of grandparents offspring and grandchildren 4 5 B® MP WB 4 5 Br ees 4 De least 30 4 2 9 27 38 5 Ye) WK; PBBY 394 IS) aM DK) BEER (7 375 6 76 198 542 49 23 883 6) 84) 211-5057 i 5 816 7 Gh ay dh 86 mi LL) 3. Sot 76 8 1 1 2 : 8 1 1 2 106 369 8383 59 28 | 1395 113 344 826 19 5 1307 r=0 .0088+0 .018 r=—0.0495+0 .018 TABLE 26 TABLE 27 Clone D: Correlation of the initial Clone D: Correlation between the numbers number of tentacles of theoffspring of tentacles of members of the same with the number of tentacles borne fraternity by the parents when each bud was produced. Ah SOMMER, 8 5 5 OU aS CC. enema, 4 9° 195 2 30 BAe vg, ace g ses | cea 5| 6 95 267. 21 5 | 304 2 cel eS I Mu 6| 8 119 623 a0) 23 \)) see 7 Ui 2, 79 8 1 1 9 6 1029 8326 1360 45/10766 14 255 955 1438 28 | 1395 r=0.077 =0 .006 r=0.096+0.016 INHERITANCE IN ASEXUAL REPRODUCTION 209 TABLE 28 TABLE 29 Clone A-(A1A): Correlation between Clone A1A: Correlation between the the initial number of tentacles of initial number of tentacles of parent parent and offspring. and offspring 5 6 7 8 9 4 5 6 7 8 : 2 : 4 Soh 4 5 1 Pee ban Wg 6 42 5 1 Ges Gs 50 6 | 14 140 109 31 2 296 6 A Gl 77 30) 4 278 @ | 23) 222 41338 “67 10 455 7 Cyl Ye le) Me 104 8 ches malsia adie se fm 62 8 Gye] 3 9 1 1 2) L0L) 276° Si 6 439 =—0.0342+0.032 42 412 275 117 138 859 r=0.0011+0.023 TABLE 30 TABLE 31 Clone A: Correlation of the initial Clone A: Correlation between the initial number of tentacles of the offspring number of tentacles of grandparents with the number of tentacles borne and grandchildren by the parents when each bud was produced 5 6 7 8 ) 10 5 6 7 8 9 t 3185 1 9 4 2 1 1 2 6 Ding. 32 lt 3 93 5 4 (24 "15 Jb, 3 61 6) 24 255 182-101 28 5 590 6 | 39 139 189 104 12 433 7| 16 125 221 159 63 5 589 7 | 36 53 155, 219 47 510 Sivest 928) 2h" 85. 9 69 8 8 i AS 7 AS 81 9 . Ayre 3 9 2 1 3 50 4382 468 293 100 10 | 1353 89 224 330 389 62 1094 r=0.240+0.009 r=0.229+0.019 TABLE 32 TABLE 33 Clone D’. Correlation between the ini- Clone A’. Correlation between theini- tial numbers of tentacles of parents tial numbers of tentacles of parents and offspring and offspring 4 5 6 7 6 7 8 9 : bot 2 i 3 2 ee a Gi) Meo eee Bs 33 6 Binh Ole “8 114 7\| aka 84)! de 148 g : Sates i 8 HOW 70% (8 20 8 1 1 Siomeon ol 153 28) LOE 252. 20 204 r=—0.242+0.051 r=0.031+0.047 210 K. S. LASHLEY TABLE 34 Clone A: Correlation between the numbers of tentacles of all buds produced in each arbitrary five-day period. Each bud is paired in the table with all the others produced in the same period 4 5 6 7 8 9 4 2 82 Bis) 270 A 729 5) 574 5572 4300 484 f 10934 6 17277 =35967 3859 102 57205 7 - 20742 4982 254 25978 8 264 30 294 9 1 1 2 656 23204 61279 9609 391 95141 r =0.0774+0.0015 TABLE 35 Clone A: Correlation of the numbers of tentacles of all buds produced in each five-day period with all buds produced in the preceding five-day period. Each bud is paired individually with all buds produced in the preceding period 4 5 6 7 8 9 4 9 87 374 372 42 1 885 5 66 846 4999 4552 461 5 10929 6 245 5084 32259 33039 3897 81 74605 @ 153 3973-32873 «38143 4650 = 94 79886 8 16 495 3866 4352 527 9 9265 9 4 55 93 11 163 489 10489 74426 80551 9588 190 175738 r=0.048+0.001 TABLE 36 Clone D: Correlation between the numbers of tentacles of all buds produced in each arbitrary five-day period; arranged as in table 34 4 5 6 7 8 + 91 1845 2936 261 12 5100 oS) 8993 33381 1993 72 44439 6 41613 9265 484 51362 7 797 168 965 8 6 6 91 10838 77930 12271 742 101872 r=0.1313+0.0014 THE EFFECTS OF CERTAIN SALTS, AND OF ADAPTA- TION TO HIGH TEMPERATURES, ON THE HEAT RESISTANCE OF PARAMECIUM CAUDATUM ROBERT H. HUTCHISON From the Zoological Laboratory of the University of Pennsylvania ONE FIGURE EFFECTS OF SALTS ON HEAT RESISTANCE The effects of certain salt solutions in increasing the resist- ance of various animals to heat are sometimes quite marked, and it seems probable that all animals of aquatic habit would be similarly affected under the proper conditions. Loeb and Was- teneys (Jour. Exp. Zool., vol. 12, p. 543), found that salts exerted a great influence on the ability of Fundulus to withstand sud- den changes of temperature. The maximum temperature into which these fish could, with impunity, be transferred suddenly, varied with the concentration of the sea-water or of a Ringer so- lution, ‘‘being about 25°C. for a concentration of M/128 or M/64; 27°C. for a concentration of M/32; 31°C. for a concen- tration of M/8; and almost 33°C. for a concentration of M /4.”’ Dextrose solutions were found to lack this protective effect against a sudden rise of temperature. In a previous paper (Jour. Exp. Zool., vol. 15, p. 143) the writer mentioned a few experiments in which the heat resistance of Paramecium caudatum was increased when the animals were transferred to solutions of NaNO;, NaCl, and KNOs, the whole death temperature curve being shifted two degrees higher on the scale. The results of experiments of this kind have never been satis- factorily interpreted, and there is still lacking a complete expla- nation of the protective action of salts against heat, and of the acclimatizing effect of exposure to moderately high temperatures. Loeb and Wasteneys from their experiments with Fundulus in 211 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 2 id ROBERT H. HUTCHISON sea-water, in Ringer solutions, and in CaCl, solutions, conclude that the protective action of these salts is not an osmotic effect, nor a case of antagonistic salt action, but ‘‘a specific effect of the salts of the sea water.’’ But we are left in doubt as to just what this specific effect is. Indeed the results of experiments described below seem to indicate that, in the case of Paramecium at least, the salts have no specific action, or if so, such specific action depends upon the nature of the medium in which the ani- mal has previously lived. No attempt is made to offer any complete explanation. It is desired merely to set forth the facts as we found them as a modest contribution to the knowledge of the heat-resisting properties of living cells, with the hope that from an ever-increasing mass of data some general law may eventually be worked out. The following experiments were carried out during the winter of 1912-13 while working at the University of Pennsylvania under the direction of Dr. M. H. Jacobs. Pure lines of Parame- cium caudatum were used throughout, and the method of test- ing their resistance to heat was the same as that previously de- scribed (Jour. Exp. Zool., vol..15, p. 133-134). In testing the effects of salt solutions a small quantity of the medium contain- ing the animals was centrifuged and two drops of the dense mass of animals were transferred with a pipette to 10 cc. of the solu- tion in question. Five drops of: this solution containing the animals were placed in each of the small glass dishes used. The dishes were covered and arranged in order on the floor of the blood-serum oven. The experiments were always conducted so that the rise from room temperature to 45°C. was accomplished in about one hour. Beginning at 37°C., or lower if necessary, the dishes were removed consecutively, one for each rise of one degree Centigrade. After at least one-half hour, for possible re- covery, the number of living and dead in each dish was counted and the percentage calculated. From the percentages the curves for the fatal temperature zones and the mean for each curve were worked out in the manner described in a previous paper. Two pure ‘lines of Paramecium were studied in some detail, the one growing in a medium of alkaline reaction, and HEAT RESISTANCE OF PARAMECIUM Pals the other in an acid medium. The results will be described separately and then compared. Paramecium caudatum in alkaline medium. The medium was prepared by boiling 20 grams of hay in 500 ce. of tap-water for five minutes. Two days later, i.e., about the time of maximum acidity, just enough N/20 NaOH was added to render the fluid neutral to litmus. The culture was then seeded with a single individual isolated from a pure line previously grown in the laboratory for some weeks. The animals developed rapidly and the culture remained densely populated for about four months. The medium soon became dark brown in color and decidedly alkaline to litmus, and retained its alkaline reaction throughout its history. The effect of NaCl solutions was tested when the culture was about ten weeks old. Control experiments were carried out in every case at the same time, using five drops of the unchanged culture medium in each dish. The results obtained with M/100 NaCl are summarized in table 1. Experiments with M/50 NaCl were carried out when the cul- ture was about three months old; the results are summarized in table 2. TABLE 1 Effects of M/100 NaCl on the heat resistance of P. caudatum from an alkaline medium TEMPERATURES IN M/100 nacl: SUM OF THREE Total subjected to each temperature... 383 pio22) 279) 374s F360 Numiberiofiideathsys, .c23< ao.) ose utes 0 0 19 | 149 | 365 Percentage dead at given temperature. . 0 0 a 40 | 100 CONTROLS: UNCHANGED MEDIUM, THREE | EXPERIMENTS : Totals subjected to each temperature. . 480 | 514 | 463 | 562 Mmber ot Gerth sin... 0es6 ©. os $4~- 0 6 | 248 | 562 Percentage dead at given temperature. . 0 1.1) 58.5) 100 MEANS OF THE ABOVE FATAL (me VI TOOP NaCl. 222 ene ty 42.03° TEMPERATURE ZONES ind controls e4ec...8 (eee 40 .9° 7, Wt ROBERT H. HUTCHISON TABLE 2 The effect of M/50 NaCl on the heat resistance of P. caudatum from an alkaline medium TEMPERATURES tN. Mi /50) Nacl|sUM ‘OF FOUR: |) =SU Rese ee EXPERIMENTS Bye 38° 39° 40° 41° 42° 43° Totals subjected to each tem- JOSIUNE, conasaaacacedcooanodes| OS | 200 | 370 | Se eo 458) | 480 Number ofideaths 2) 2045.61.) <. 0 0 0 0 0 | 200 | 430 Percentage dead at given tem- DCTALUTE ees Fae ote eae 0 0 OR a0 0 44 100 CONTROLS: UNCHANGED MEDIUM FOUR EXPERIMENTS Total subjected to each tem- DEENA coconnnbaueaacsnancco| Si | Ga -| aor ere | Gl Gee Numiberofudeathis.-. 2°... 09008 0 0 24 | 109 | 490 | 555 Percentage dead at given tem- DELUUUTC UA ae Renae ere eee 0 0 9.1) 238 84.3] 100 MEANS OF THE ABOVE FATAL hal MYO) INIEX CR RE BARES co So ancn se bal i° TEMPERATURE ZONES AC ONPRONG HS ae. 3's |. 5 ool se eee ae 40 .3° It will be noted from the tables that M/100 NaCl has raised the mean of the fatal temperature zone of this pure line 1.13°C., and that M/50 NaCl has raised it 1.8°C., the stronger concen- tration producing a slightly greater effect. Concentrations stronger than M/50 had a toxic effect on this race, deaths occur- ring at much lower temperatures than the controls. The effect of KCl was not studied except that it was found in one experi- ment that M/200 KCI had a toxic effect, all the animals being dead at 38°C. In weaker concentrations it would probably have shown some protective action. Solutions of CaCl, were tested on this same race, and weak concentrations had a marked protective action. When trans- ferred to M/300 CaCl, nearly all the Paramecia died within two hours at room temperature, while in M/3000 CaCl, this race remained alive at room temperature for at least two weeks. Table 3 shows the effect of CaCl, alone, and table 4, the effect of a HEAT RESISTANCE OF PARAMECIUM combination of NaCl and CaCh. transferred to a medium consisting of 10 ec. of M/100 NaCl plus one drop of M/100 CaCl, the results shown in table 4 were obtained. TABLE 3 The effect of M/3000 CaCl, on the heat resistance of P. caudatum from an alkaline medium IN M/3000 cacl.: SUM OF FOUR EX PERIMENTS Totals subjected to each temperature. . INumibenofedesdthssaseanmirmeee ness arias Percentage dead at given temperature. . CONTROLS: UNCHANGED MEDIUM, FOUR EXPERIMENTS Totals subjected to each temperature. . Number olf desith se 520.1544, ais facs'a! Percentage dead at given temperature. . 215 When the Paramecia were MEANS OF THE ABOVE FATAL TEMPERATURE ZONES TEMPERATURES 39° 40° 41> | ane 43° 44° 286 340 276 357 319 320 0) 0 0 194 310 320 0 0 0 54.2} 97.1} 100 444 409 316 554 0 58 314. 554 0 11.7} 99.3} 100 jt M/S000} CAC Ae see ee 1 O82 | dn controleee) sa ee 40 .3° TABLE 4 The effect of NaCl plus CaCl, on the heat resistance of P. caudatum from an alkaline medium TEMPERATURES IN Nacl + Cacl,: SUM OF THREE EXPERIMENTS 39° 40° 41° 42° 43° 44° Totals subjected to each temperature. .| 237 | 228 | 298 | 391 | 249 | 235 Nitmiberrotdesthses 8%. 2))..2252...) > 0 0 0 31 | 188 | 235 Percentage dead at given temperature. . 0 0 0 7.9) 75.5} 100 CONTROLS: UNCHANGED MEDIUM, THREE EXPERIMENTS Totals subjected to each temperature..| 312 | 294 | 316 | 369 INtmmiber oi denis. s/c dees ss cs 38 0 58 | 312 | 369 Percentage dead at given temperature. . 0 19 98.7} 100 MEANS OF THE ABOVE FATAL [ENO CCT wae. 42.56° TEMPERATURE ZONES an controle meter. 8 wee 40 .3° 216 ROBERT H. HUTCHISON The tables show that in this race the solution of M/3000 CaCl, has raised the mean of the fatal temperature zone 1.68°C. This is very nearly the same as the effect of M/50 NaCl, as shown in table 2. The most pronounced effect was produced by the addition of a little CaCl, to the solution of NaCl. In this case the mean was raised 2.26°C. The significance of these increases is all the more important when one notices that the means of the control experiments are almost exactly the same in all the tests, although they were made at very different periods in the history of the culture. The effect of KNO; on this race of Paramecium is summarized in table 5. The mean death temperature of this race was 1.9°C. higher in M/50 KNO; than in the corresponding controls. In three ex- periments with M/100 KNO; the mean was found to be 41.1°, a rise of 0.3°C. above the corresponding controls, which gave a mean of 40.8°C. TABLE 5 The effect of M/50 KNO; on the heat resistance of P. caudatum from an alkaline medium TEMPERATURES IN M/50 KNO3: SUM OF FIVE EXPERIMENTS Sn 38° 29° 40° 41° 42° 43° Totals subjected to each tem- WEKALUTC Nas ner eine reiisrs 5-3-8 Oe 295 | 781 | 634 | 623 | 735 | 694 | 691 Number of deaths........... : 0 0 0 Zi 190 | 210 | 691 Percentage dead Di given foe WELAGUTES wose eo Phoe eee pore ne 0 0 0 4.3| 25.8] 30.2! 100 . IN CONTROLS: SUM OF FIVE EXPERIMENTS Totals subjected to each tem- DETATULE Sy. Leer eos opie ye 224 | 380 | 584 | 677 2 725 Number of deaths. . Ree 0 ON 221 e500 S23.mn Eizo Percentage dead ne given ioe DErAvULes. =< ee sees ee sie cae 0 0 39.5} 73.8) 100 100 MEANS OF THE ABOVE FATAL Jf in M/50 KNO3...........-.----+-++--- 41.4° TEMPERATURE ZONES \ GC On brolss, 8... eee ae eee 395° HEAT RESISTANCE OF PARAMECIUM 217 Perhaps the most curious results of all were obtained when this race was subjected to the influence of M/600 Na,CO;, and to distilled water. Table 6 shows these results and those of the corresponding controls, all of which were carried out at the same time. The results in table 6 show that distilled water was just as effective in its protective action for this race as was M/600 Na»,COs, or M/50 KNOs, or as M/3000 CaCl,. The mean death temperature in distilled water was 1.78°C. higher than the corresponding controls. In cane sugar solutions there was no protective action mani- fest. In one experiment two drops of the medium densely popu- lated with Paramecia were transferred to 5 cc. of M/8 cane sugar. When subjected to heat some deaths occurred at 38° and above. This solution was much less effective than distilled water; indeed, it appeared to have a weakening effect. TABLE 6 The effect of M/600 NazCO3, and of distilled water, on the heat resistance of P. caudatum from an alkaline medium TEMPERATURES In M/600 Na2co3;: SUM OF TWO EXPERIMENTS 39° 40° 41° 42° 43° Totals subjected to each temperature......... MENS ee I) at) Nh ala 151 INtumberRoitdeathcten men meer rce sn eae 0 6 Salone lol Percentage dead at given temperature........ 0 4 23.3] 61.4) 100 IN DISTILLED WATER: SUM OF TWO EXPERIMENTS Totals subjected to each temperature.........| 147 | 187 | 165 | 148 | 188 iINumiberzotedeathseteaasasectiecteeicniaas eth eaose 0 0 4 97 | 188 Percentage dead at given temperature........ 0 0 2.4) 65.5} 100 CONTROLS: UNCHANGED MEDIUM, TWO EXPERIMENTS Totals subjected to each temperature......... 260 | 216 | 260 | 209 INiumbpersotydea thst: rays sai caus tases ote O | 115 | 241 | 209 Percentage dead at given temperature........ 0 53.2} 92.7) 100 J oo n PATAL TEMPERATURE in M/600 Nap,CO3 SOC O00 41 .62° Me oo in distilled water....... 41 .82° ZONES ; ° Tiny GOMUAKOUES cols dno snenese 40.04 218 ROBERT H. HUTCHISON The above results agree with those of Loeb and Wasteneys, in one respect at least, in that they show that the protective action of salt solutions is not an osmotic effect. But when we find distilled water just as effective as three of the salt solu- tions used and almost as effective as M/50 NaCl, and the combi- nation of NaCl and CaCl, it is apparent that in the case of Paramecium at least, the results are not due to the “specific action of the salts.”’ Some results of tests with another race of Paramecia in an acid medium detract still more from the specific salt action as a satisfactory explanation. An account of some of these results follow. Paramecium caudatum in an acid medium. The culture me- dium was prepared by boiling 20 grams of hay in 500 cc. of tap- water for one-half hour. The following day it was seeded with a single individual isolated from another pure culture which had been growing in the laboratory for some time. This culture medium retained its light straw color throughout. It was never as densely populated as the alkaline medium and died out sooner, i.e., in about three months. Most of the following experiments were performed when the culture was about two months old. It was still light colored and slightly acid to litmus. Tables 7 and 8 summarize the results of tests with M/4000 CaCl, with M/100 NaCl, and with distilled water. It is to be noted that the control experiments show that the normal heat resistance of this race was somewhat higher than that of the race from the alkaline medium, the average of all the con- trols of the acid medium being about one degree higher than the average of all the controls of the alkaline medium. This is just the opposite effect from that produced by. acids and alkalis on the coagulation temperature of proteids. Further, it will be noted that the same salts (NaCl and CaCl,) which gave the most pronounced protective action with the Paramecia from the alkaline medium, actually decreased the resistance of those from the acid medium. Considering the action of the salts alone it might be supposed that their effects were conditioned by the reaction of the medium in which the animals had pre- HEAT RESISTANCE OF PARAMECIUM 219 TABLE 7 The effect of M/4000 CaClz on the heat resistance of P. caudatum from an acid medium TEMPERATURES In M/4000 cacl.: suM oF Two EXPERIMENTS cA |) LU il 2 ia ie a Totals subjected to each temperature.........| 243 | 206 | 202 | 170 INTIM beroigd ea thsremwaeeere erie < oa. ee. ae OF 200") 202 | 170 Percentage dead at given temperature........ 0 97 | 100 | 100 CONTROLS: SUM OF TWO EXPERIMENTS Totals subjected to each temperature......... 238 | 380 | 1438} 300 | 295 Aci a Gre On GEA GIS 4 antes estrone ey ws ose Aci AO 5 Dow 224 a2 LON e295) Percentage dead at given temperature........ Weill” 7 15.4, 70 | 100 MEANS OF THE ABOVE FATAL TEMPERATURE {| in M/4000 CaClh........ 39.53° ZONES | a ecominole tee sel3,..<. 41 .55° TABLE 8 The effect of M/100 NaCl, and of distilled water, on the heat resistance of P. cauda- twm from an acid mediwm TEMPERATURES IN M/100 nacl: SUM OF THREE EXPERIMENTS 39° £08 | at 42° Totals subjected to each temperature............... 368 | 308 | 299 | 327 ASNT OCA S Mes ca yereen Gy che ties «ic cho SS Ss con Deere 0 49 | 242 | 327 Percentage dead at given temperature...........! fd 0 15:8) 8°) 10 IN DISTILLED WATER: SUM OF THREE EXPERIMENTS Totals subjected to each temperature............... 395 | 390 | 3382 | 429 INA eTRO Rid Cathishm es tatat ota cuca ahce see cies eran 0 72 | 326 | 429 Percentage dead at given temperature.............. 0 18.5) 95.1} 100 IN CONTROLS: UNCHANGED MEDIUM THREE EXPERIMENTS Totals subjected to each temperature............... 662 | 685 | 622 | 770 INtmiberxzotedeathst re bps ie cies enc oe soo oeeysus aeeeeee LS a eLSO Na ei Percentage dead at given temperature.............. 2-2)' 13055) 100 (in M/100 NaCl... . 40.5° MEANS OF THE ABOVE FATAL TEMPERATURE ZONES {in distilled water.. 40.3° (im controls.......... 41.1° 220 ROBERT H. HUTCHISON viously existed. But distilled water was found to have the same effect, and some influence other than the salts seems to be the important factor. Taken as a whole, the above experiments seem to point to the conclusion that certain properties of the medium are important factors in the heat resistance of P. caudatum, and that such properties will predetermine whether a given salt solution will have a favorable or an unfavorable effect. Whether this is true for other Protozoa remains to be determined, and a satis- factory explanation is yet to be proposed. EFFECTS OF ACCLIMATIZATION TO HIGH TEMPERATURES ON HEAT RESISTANCE A few experiments were carried out during the same season with a view to determining to what extent continued exposure to moderately high temperatures would influence the death temperature. Several attempts along this line failed, some on account of too sudden changes and some for other reasons. Some success was had with two cultures which were studied at fre- quent intervals during an exposure of over two months to tem- peratures ranging from 28° to 36°C. Pure line cultures of Paramecia were used in these experiments. On January 30, 1913, a culture medium was prepared by tak- ing 35 grams of hay in two liters of distilled water. This was heated for a period of ten minutes at a temperature of about 60°C. This was then divided equally between two culture jars so that each contained one liter of the fluid and about 173 grams of hay. After cooling, each of these cultures was seeded with a single individual taken from a pure line which had been growing in the laboratory for 86 days before this date. The two result- ing cultures, which will be referred to as ‘30-a’ and 30-b,’ may therefore be regarded as of the same pure line and growing in the same medium. On February 11,(12 days after the culture was started) 30-a was transferred from room temperature to a water bath at 28°C. The temperature of the water bath was regulated automatically and never varied more than 0.5° either way. The culture jar was so placed that the level of the cul- HEAT RESISTANCE OF PARAMECIUM DPA | ture fluid inside was about one inch below the level of the water of the bath outside the jar. The mercury bulb of the thermome- ter showing the temperature of the water bath was also placed about one inch below the surface of the water, so that it gave very closely the actual temperature to which the Paramecia were exposed. The culture jar itself was of course kept covered with a glass plate at all times. The constant temperature of 28° was maintained for a period of 27 days. On March 10 the tem- perature was raised to 30°C. and on the 12th to 32°C. After 10 days at 32° the temperature was again raised to 34°. After 16 days at 34° the temperature was raised on April 7 to 35°, and on April 15 to 36°, at which point it was maintained to the end of the experiment. On April 26 the culture was in very poor condition and very few animals were present. A little dry fresh hay was added to the medium but it did not have any favor- able effect and the strain had completely died out by April 29. ’ * Strain 30-b. was used. as a check on 30-a, and was kept on a table in a cool room, the temperature of which varied from 12° to 22°C. On April 29, when 30-a had completely died out, 30-b was still in good condition, although not as densely popu- lated as in earlier periods of its growth. During the course of the experiment the heat resistance of both strains was studied at frequent intervals. The mean of the fatal temperature zone for each experiment was worked out in the usual way and these means are plotted in figure 1. Another pure culture, 11-2, was started from an individual isolated on December 11, 1912. This culture was kept at room temperature until February 14, at which time it was transferred to the water bath at 28°C. From that time on until the cul- ture died out (April 25) it was; subjected to the same temperature conditions as was 30-a. The mean death temperatures of a series of tests with this strain are also plotted in figure 1. Ex- amination of the figure shows that the resistance of the control was by no means constant. There was considerable variation, the means varying from 40.5 to 42.3°C. . The resistance was more irregular and in general slightly higher during the later course of its history than in the earlier experiments. Mean death temperatures of three races of P. caudatum in degrees Centigrade TMT 1 MTU Th T TM TIN T Ti i | ara all : I i ! | [i ] aT | | \ Mt t 7 j | j i ATH | | i| i | {I | i Hi | ' i i H 1 i 1 i i | | hil ii UNE Wr 1 H Wh \ WH] i} | HTT i ! iit ul il | {i iT i | il ill NN | | | i] aH TN ATLA TTT TAT Whi HATE FE HY TTT TTT HATA ili iil | Viti 1 Hl i ] WT HTT ITT | AHH TAT ATL AAT TT TAT TT i} 1 ‘li wall i TTT i iit rit} T nih RTT TT ART TU it i 1 HIT! } Ili | HTT t | py TIT i T t Wt TTT Ti | AMAA ATTA ili il | | {II A a \| ATLL A li | HE AE W Au yj | T | ry HAGE T Tht i Wht WIE i | | | HN 1 iy i nH ii ily WH \ Wi | AA TE HH y2) ry iH} PTT AN Tint j ill | ATH Hy TT | HTT, Ht | el i | ny} | ill | | HTL | HITE RATT ITNT TATA i TUT H ili CVAD) ACLU AACUN ER CUTTLEEEU OOOH UAEY A | TTT ik HITT TTT NA | it aE WATT! | | All HTH Titl] itll . {IH iii! it i! - Th STUNT ETVTHIT Ltt 1 Hh i | \ | UTAH TITHE T nT | TTT WUT TT | TIA | Mh VHB - ; HH ill fit | 1 | | Il) | {Ht | ill 1 ) ND ni | | 1 | | | | | HTT HTT nth il! if H nT ut WT HTT H nH HVAAAATRAEHUITE WAIT Hii ; | HT | | HIE TTA TAT THAT TTA il | ; nl itll ATA ATVAA EH TTF mi | Il ii Hi AAT | l T Hit} HI NN Tit THT ill mM! iin TU UTARTANT 4 THT | (if ] | | H ml ii ij TTT if ‘HHT HI HTT i ity { | H i ti HU i HITTIN! HTT HUE HTT HUN iLL HH | i Wh il i ml TTT WT TT SHAT NET ST TTT | Ana iil} | iV | f j | ANANDA AEATTAT HTH | | itt ; rin} nT MATT i fi it | Wh i| I] | i} Wt HTH Wilt \ | NIT i TTT TTT TTT | UTE LEA ASANTE | | | WNT Wh ih UTAH TU ill HUTT UAT ATT TTT lll) | TAIT UUATUA HAAN ANNEAL TTT AAA OTOANUAAEA AOE AU HAAG AUTH ELT AUER OAGUNPEAGDAEHHAARRTAAGS coed TANUV AVES TOGEAGROUA TERT FRSR OGD AROOUEY AUTH AL HET OUNCE | | TT TTT TTL | Hilt Wy i NH ill | Wily! ty! | i {| Ht | | | i Ih THU UD LOTR | Aull ltt mI | ry rit TATE T i TID We H 1 HATTTAUIT MN HNN: MTT THTTTIN un ill AIT TTTTTT TTT TT i rH LTT i \ | WHITE! il} MT ' TTT i | if HUT WN | | i WE HA TAHT \| ii Wy | Ht | Hf i] | ui il} | HTT ATO i Hill TWH | i | | | T TTT! i | | i} 1 | | | | nt in| 1 Wi if | | if ! | \ it ATUUEANUTAUUAEELVED ALOUCREA TOTES AT AULTHLETACELOUUAAA AUUAUEQEAUULEEOUTGAAUOUAEOH EMEA AUTO EAE ET HA A NTN CN PAT UUNUT FAUT UNA TNANTAA AUTH ill AIT PTT TT TTT TT) Ni il HI | { | uu | 1 ; TET r Tie tt H ive nea sf TUDESLAUETUNAAE it n HOLSUEG HADES FRLDOCDESULOAED EADEA GEA DOAN ERGLS ARADO RANE EETO CACROQUNS CHARA DADA DORET LNBQOAATUGS QED LRT i IAA DANA ! { Fig. 1 Showing the mean death temperatures of two races of Paramecium growing at moderately high temperatures, and those of a control experiment kept in a cool room. 222 HEAT RESISTANCE OF PARAMECIUM 22 In the case of 30-a, which was growing in practically the same medium as the control, we sée a much greater fluctuation. The mean death temperatures for this race vary from 38.5 to 43.4°C. After one week at 28° the resistance decreases and is lower than at the beginning but after February 27 the resistance increases again and the general trend of the line is gradually upward. After the temperature was raised to 34°C. the resistance of this strain remained rather high, being in general from 1 to 2.5° above its initial resistance. However, the highest mean was only 1° above the highest mean for the control culture. Several experiments with strain “11-2 before it was put in the water bath gave an average mean fatal temperature of 40.5°C. As soon as this culture was subjected to the higher temperatures of the water bath we find that the resistance increases and re- mains at least 1° higher than the initial resistance and at times is more than 2° higher. The mean death temperatures of both 11-2 and 30-a at times approached 43° and sometimes exceeded it, while, as pointed out above, the death temperature of the control, 30-b, never exceeded 42.3°. However, this increased resistance of those strains growing at the higher temperatures was not maintained for any considerable period, and it is hard to see that any de- cided effect was produced. SUMMARY 1. The effects of certain salt solutions on the heat resistance of Paramecium caudatum were tested. Two pure lines of Paramecia were used; one growing in a medium of decided alkaline reaction, and the other in a slightly acid medium. Experiments with the race from the alkaline medium showed that M/100 and M/50 NaCl, M/3000 CaCl, and M/50 KNO; exerted a marked pro- tective action. The greatest increase was noted in a solution of M/100 NaCl to which a little CaCl, had been added. Distilled water also was found to increase the heat-resisting powers of this race. DOA ROBERT H. HUTCHISON 2. The Paramecia from the acid medium were adversely affected by M /4000 CaCl, by M/100 NaCl, and also by distilled water. , 3. The conclusion is drawn that certain properties of the medium in which the animals had been growing were the impor- tant factors in determining the ability of Paramecium to with- stand heat. No explanation is offered as to what these factors are, nor how they act. 4. The effects of continued exposure to moderately high tem- peratures on the death temperature of Paramecium were studied. Two cultures of Paramecia were kept in a water bath, the tem- perature of which ranged from 28° to 36°C. The mean death temperature of these two strains fluctuated considerably, and at times was about 1° above the highest mean death temperature of a control culture kept ina coolroom. But this increase above the control was not constant, and on the whole no very decided effect was produced. DIDINIUM NASUTUM I. THE LIFE HISTORY GARY N. CALKINS From the Department of Zoology, Columbia University TWELVE FIGURES—ONE PLATE The work of Woodruff and Erdmann (’14) on Paramecium aurelia showing the occurrence of periodic reorganization of the cell, which, like conjugation, has the effect of renewing vitality, raises the question as to the length of life of a ciliated protozoon and its progeny in which both asexual reorganization and conju- gation are prevented. Fermor ('13) has shown that a similar reorganization occurs in Stylonichia during the process of encyst- ment. Prandtl (’06) made the statement, unsupported by evi- dence, however, that in Didinium nasutum nuclear reduction occurs during encystment as well as during conjugation. These observations indicate that encystment in ciliates, when not for purposes of protection against adverse environmental conditions or for division (as in Tillina), is a process during which nuclear reorganization, or parthenogenesis, takes place. An encysting organism in which asexual endomixis takes place has advantages over Paramecium in the present problem because of the definite external advertisement of the internal processes taking place. Didinium nasutum was chosen for the experiments because of its large size, its easily controlled feeding habits and because of its readiness to encyst. The feeding habits have been worked out by Mast (’09) and the process of conjugation by Prandtl (’06). MATERIAL AND METHOD Two individuals—X and Y—of Didinium nasutum were iso- lated from fresh material brought into the laboratory from Van Cortlandt lake on October 28th, 1914. They were placed in ground-glass flat-bottomed culture dishes each containing 0.25 225 2260 GARY N. CALKINS ec. of clear pond water. Twelve individuals of Paramecium caudatum were picked out and placed with each of them, this number being purely arbitrary. After a few days’ trial it was found that a smaller number of Paramecium gave better results and a standard daily diet of 9 Paramecium caudatum was es- tablished and maintained throughout the experiments, which are still under way. Five lines of X and five of Y were estab- lished on the second day and one individual from each of the ten lines was picked out and placed with 9 Paramecium caudatum in 0.25 ec. fresh spring water daily. The usual history of Didin- ium in such an environment at the end of twenty-four hours is 8 Didinium and no Paramecium. At times we find only 2 or 4 Didinium and no Paramecium, showing that the appetite was good but the dividing power reduced. Again we find occasion- ally 2 or 4 Didinium and from 2 to 5 Paramecium, or some- times, only 1 Didinium and from 8 to 10 Paramecium, indicat- ing what I shall speak of as loss of appetite. In still other cases the single individual does not divide at all but, notwith- standing daily changes of water and food, dies, usually by the fourth day. Finally encysted individuals which have not di- vided are occasionally found at the end of twenty-four hours, together with from 9 to 13 Paramecium. The rate of division of Paramecium is of course very low owing to the scarcity of bacterial food. GENERAL DESCRIPTIONS Feeding habits of Didinium nasutum Next to the capture of Halteria grandinella by Actinobolus radians I know of nothing more spectacular or amazing in the whole realm of microscopy than the seizure and ingestion of Para- mecium by Didinium. Described by Balbiani (’73), by Thon (05), by Jennings (06) and by Mast (09) there is little in the process for me to dwell on. The actively rotating carnivore swims vigorously through the water, occasionally limiting its activity to side or bottom of the culture dish, making vicious jabs downwards or sideways until it hits something soft enough for its proboscis to penetrate. As earlier observers have pointed LIFE HISTORY OF DIDINIUM 227 out, there is no evidence whatsoever of choice of food nor any evidence of chemiotactic guidance of captor to prey. The en- tire process is apparently fortuitous, some one of hundreds of jabs is successful and a Paramecium, once hit, rarely gets away (fig. 1). The victim is partially or wholly paralyzed and is speedily swallowed, the walls of the Didinium being stretched around the prey like a rubber bag. If the Paramecium is seized at or near one end, this end goes in first (fig. 2) until it reaches the extremity of the captor, (fig. 3). It is then doubled on it- self until it lies like a U completely ingested (fig. 4). The Para- mecium protoplasm becomes highly vacuolated, broken up into small pieces and is quickly digested (fig.6). If the Paramecium is seized in the middle, this part goes in first and the two ends last, so that the victim is swallowed in the U form. Not only can a small Didinium thus capture and engulf a Paramecium six times its size, as shown by Mast, but it will swallow a dividing © Paramecium, and I have frequently watched one attack and swallow a pair of conjugating Paramecium. My imagination has pictured the surprise which such a Didinium might feel when, having completed its usual task, it found itself com- pelled to swallow another equally large meal. In such cases one of the free ends of the two victims is usually seized; this individual is ingested and the process is continued until the second individual is completely engulfed. It means a little more tension on the part of the elastic walls of the captor but, usually, he is equal to it. Such stuffed individuals are subject, however, to diffluence, especially if transferred shortly after feeding to fresh water, and I have watched more than one individual explode, victims of their gluttony. Structure of the proboscis and seizing organ The proboscis of Didinium is a conical projection in the center of the anterior end. It is supported by a dense layer of trichites which are anchored deep in the protoplasm. These are evidently strengthening organs and probably play a part in preventing rupture when a large food body is swallowed, in much the same THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 2 228 GARY N. CALKINS way that spiles in a ferry slip take up the strain. In the center of the conical proboscis is a column of protoplasm somewhat denser than the rest of the endoplasm. This structure, called by Thon the ‘mittlerer Strang’ and by Mast the ‘seizing organ,’ is the apparatus which fastens the prey and precedes it into the body of the captor. The open passage which it leaves by its migration through the protoplasm becomes the cytopharynx, the prey being drawn in largely, if not entirely, by its pull (figs. 1-6). | This seizing organ is such a remarkable structure that it well repays careful study. There is no doubt that it contains some toxic substance which partially or wholly paralyses Parame- cium, but no one, as yet, has shown where this substance lies. Balbiani (’73) held that trichocysts are discharged by Didinium and that these penetrate and paralyse the victim. Thon, how- ever, supported by Mast, denies the discharge of trichocysts and holds that the entire process of capture and retention is a func- tion of the seizing organ. Thon figures the seizing organ as stri- ated, and in this he is undoubtedly correct, but he evidently failed to note a zone of thickened granular striae near the apex of the seizing organ and clearly apparent when the organ is ex- tended (fig. 8). This zone is made up of the same sort of thing, apparently, as the granular zone at the apex of a tentacle of Ac- tinobolus. These granules were first described for Actinobolus by von Erlanger (’89) as trichocysts on observations confirmed by Moody (12) who speaks of them as ‘trichocyst material.’ They evidently are the poison granules which cause the instan- taneous paralysis of Halteria. In the absence of other physical evidence of poison in the seizing organ of Didinium we are justi- fied in regarding these granules, liberated on penetration of the cortex of Paramecium or other victim, as the cause of paralysis. The entire seizing organ may be regarded as a bundle of structures homologous with the distributed tentacles of Actinobolus. Like the seizing organ, each tentacle of Actinobolus is retracted into the endoplasm, dragging the attached victim to the surface of the body, where it is manipulated by the cilia until swallowed through the mouth. LIFE HISTORY OF DIDINIUM 229 Structures of the endoplasm Thon has given an excellent account of the finer structures of the endoplasm of Didinium. One or two points should be men- tioned here as they have to do with structures involved in different stages of the life history. The most important of these are the nuclei and their derivatives. The macronucleus is correctly described by Thon. In the resting stages it is characterized by deeply-staining spherical granules of chromatin embedded in a more feebly-staining matrix. These bodies in the nucleus behave during division like the chromatin bodies of Dileptus gigas, where they are dis- tributed throughout the cell. At periods of division of Didin- ium they form first a more or less complicated reticulum by elongating and fusing at one or more points. The strings of chromatin are ultimately divided, again as in Dileptus. At periods of encystment the distinct granules of the nucleus be- come much larger and are discharged from the nuclear mass until the cytoplasm becomes filled with densely-staining chro- matin bodies. After recovery from encystment the distributed chromatin masses are broken up into smaller metaplasmic gran- ules, which give a uniformly dense stain to the entire endoplasm. Finally, at conjugation, these metaplasmic bodies disappear from the endoplasm and are concentrated in a deeply-staining cortical armature in the ectoplasm. The micronuclei were entirely overlooked by Thon. They are extremely small and difficult to distinguish from the numerous spherical bodies distributed throughout the endoplasm. At periods of conjugation, however, they are plainly evident and their history may be followed with comparative ease. This was first done by Prandtl (’07), who also for the first time described the micronuclei in vegetative stages. The number, according to Prandtl, is variable, two or three being usually present and these are closely anchored to the macronucleus (fig. 7). In division one pole of the spindle is usually embedded in the sub- stance of the macronucleus. I have confirmed these observa- tions of Prandtl, finding as many as four micronuclei during the 230 GARY N. CALKINS resting stages of non-conjugating forms, four in organisms pre- paring to encyst, and as many as sixteen in the conjugating ani- mals. Owing to the difficulty in finding them even in the most carefully stained sections, no positive statement can be made as to the ‘normal’ number, but it appears to be four. They are very minute (4-6 uw) with relatively little chromatin, usually concentrated in a few central granules, and with definite nuclear membranes. The division spindles are narrow and _ sharply pointed with the chromatin in minute chromosomes (fig. 7). - In addition to the nuclei there are numerous curious and enig- matical structures in the endoplasm which I am unable to inter- pret. These are often in the form of spindles with curious rod- like bodies which suggest chromosomes (fig. 7). These are found during all stages of vegetative life. There is no evidence of their division and no reason to believe that they are nuclei, and the only suggestion I have to offer as to their function, is their possible connection with the formation of new seizing organs to replace those used up in food capture. ' THE LIFE CYCLE In other places I shall describe the details of eneystment and of conjugation, and will limit the present paper to the history of the race from October 28th to the present time (April 20). Thus far the organisms have gone through two completed cycles, each ending in encystment of all the living material and during which nuclear reorganization occurred. The initial cycle (October 31 to December 28): Individual X Five lines derived from individual X were each changed to fresh water and fed 9 Paramecium caudatum daily. The daily rates of division were averaged for 5-day periods and the results plotted to give the accompanying chart A. The daily records include the number of divisions which each individual had undergone during the preceding twenty-four hours, the number of living Paramecium in each culture dish, the number that LIFE HISTORY OF DIDINIUM 231 had eneysted, and the number that had died. In case of death or encystment of an individual in any line, its place was filled from among the descendants of some other line in the same X Series: 1st Cycle ~ 1 } o ia a | Gee ef a a raf PEELE EEE “EEEE EEE EEEEEHE +t am | [ | Balai anit + re LH at : | aan ie B _t | | cot SAH | go aE E Pe | EEE B ae om iD FE LECH EEHEE EE ; - 5 t iz tel | FEE : | ui Sess SeGRe re LE 3 t E E af aE ial nee | aE | EHH {- L E EH a NAPEEEELEEEE EEE EEE EEE EEE a a Chart A series. Superfluous individuals each day were supplied with food and kept as ‘stock’ in the moist chambers. The division rate averaged 1.95 divisions per day for the first 20 days of this cycle, 1.64 per day for the second 20 days, and 232 GARY N. CALKINS 1.18 for the last 20 days, falling finally to zero with encystment of all the living material in approximately the 131st generation (chart A). The percentage of encystment, computed from the total num- ber of individuals to encyst in the 5-day periods and the total number of individuals under observation, is shown at the bottom of chart A. There is but little ground for comment here, the fluctuations coinciding more or less with those of the division rate. The percentage was low at the outset but increased later until it finally rose to 100 per cent. The death rate, computed in the same way as the encystment rate, was comparatively low throughout the cycle, never rising above 8 per cent (chart A, top). The records of the numbers of Paramecium eaten, determined by the numbers found alive at the end of twenty-four hours, were not begun until one month after the cultures were started. These records furnish the basis for a study of the variations in what may be termed the ‘appetite’ of Didinium, shown in the dotted line of chart A. The data for this curve were obtained as follows: In 5-day periods the five lines of culture material are provided with 45 Paramecium daily, or 225 during the period. Add to these 25 per cent for approximate increase by division before being eaten, giving 280 Paramecium for the 5-day period for all five lines. The daily records give the numbers of Para- mecium alive at the end of twenty-four hours. These are aver- aged for 5-day periods and the average, divided by 280, gives the percentage of uneaten Paramecium, which subtracted from 100 per cent gives the approximate percentage of Paramecium eaten. Rough as this method is in illustrating the variations in appetite of Didinium the curve nevertheless follows that of the division rate with remarkable fidelity. It might be argued that the division rate should follow the eating rate and that the 5-day periods in the ‘appetite’ curve should be twenty-four hours in advance of the periods indicating the division rate. But it is equally true that the feeding rate depends on the vitality of Didinium, and as the records for feeding brought the initial dates of the periods forty-eight hours later than those for the LIFE HISTORY OF DIDINIUM Zao division rate, I have worked them out on this basis. The flue- tuations of the appetite curve follow closely those of the division rate and both are correlated with the fluctuations in the curve of encystment. When the latter reaches 100 per cent both division-rate and appetite-rate fall to zero. The over-lapping at this period (Dec. 25 to Dec. 30) is due to the fact that Para- mecium may be eaten prior to encystment but without division of Didinium. The second cycle (December 28 to March 5): Individual X . All the living material of Didinium in culture and stock dishes became encysted during the period beginning December 25. The race was recovered from encystment December 28, by pouring off the old water and adding fresh water and Paramecium to a Syracuse dish in which stock material of the X series had encysted the week before. The division rate immediately indicated a renewal of vitality, giving an average for the first 20 days of 2.01 divisions per day. It then fell to 1.76 for the second 20 days and to 1.25 for the third 20 days. In the last 5-day period (February 28 to March 4) it averaged only 0.52 divisions per day, after which all culture material and stock material en- cysted (chart B). The race passed through 148 generations during this cycle, or 279 generations since the culture experiments started. The rate of encystment began at 16 per cent for the first 5-day period but quickly fell and remained low during the first 40 days, after which it maintained an average of about 17 per cent until all individuals became encysted (chart B, bottom). An interesting feature of this second cycle was the increase in the average death rate over the first cycle. The average death rate for the entire first cycle was 1.8 per cent and for the entire second cycle it rose to 7 per cent. When we consider the relative infrequency of death of the individuals in culture this phenomenon becomes significant (chart B, top). As in the first cycle, the curve for appetite closely follows the curve of reproduction (chart B, dotted line). 34 GARY N. CALKINS X Series: 2nd Cycle « co Cer o Bi Coe | Co im Bal ae ESET + r | = 5 + Ht 4 isis | el + ai I 5 Co ao Cl LE Et ia a | t CI Pe ae aa | | 4 : are Hie 4 a a | j ia BEEEEEEEEEEEEEE EET H Eee ime came | i a a [a aq f Let ae [ ia i SS Se0 soo EEEEEI HS EEEEEEEEEEEHH f | tee | on - Seo aia teta pata Chart B The third cycle (March 9 to date, April 26): Individual X After encystment of all individuals in culture and stock dishes the race was recovered on March 8th as before, by adding fresh water and food to a stock dish set aside on the 27th of February. By the 3rd of March all of these had encysted. Five days later 100 cysts were picked out and placed in fresh water with Parame- cium caudatum. On the following day (March 9) there were LIFE HISTORY OF DIDINIUM 235 about 40 active Didinium in the culture dish. Five of these were isolated and furnished material for the third cycle. Twelve were killed and the remainder were fed and left as stock. All of the remaining cysts were killed for cytological study. The average division rate during the first twenty-four hours after recovery was 3.4, and for the first 25 days it remained high (2.19 per day). The encystment rate for the entire period was 8 per cent, while the average death rate for the period of 25 days rose to 14.4 per cent. This increase in the death rate as the culture series grows older, and already indicated in the second cycle, is interesting and significant. Cultural history of the Y series: Initial cycle The Y series, started with a second individual at the same time as the X series and treated in the same way, confirms the results obtained with the X series. The first cycle (Nov. 1 to Dec. 26) had the same general history as in the X series but with a slightly more regular descending curve of the division rate (chart C). The average for the first 20 days was 1.65 divisions per day; for the second 20 days, 1.19 divisions per day, and for the third 20 days it fell to 0.65 divisions per day. The curve for encystment is much more regular than that for the X series and, with one exception in the 7th 5-day period, shows a fairly steady increase (chart C, bottom). The death rate was very low (chart C, top) and the appetite curve is simi- lar to that for the X series (chart C, dotted line). The cycle came to an end with encystment of all culture in- dividuals a few days in advance of the X series, and in the 128th generation, on December 26. The second cycle (December 28 to February 6) The race was recovered from encystment December 28 from stock material which had encysted on the 22nd, and a second cycle was started with an initial division rate of 2.08 divisions per day (chart D). The vitality was not as great as before, the average division rate for the first 20 days being only 1.65 per 236 GARY N. CALKINS Y Series: Ist Cycle sea ee Sarees GEBOS FERRE SESS CORE Bees GSO oe De See Bees Peebs Sees | F | ZEEE BESS Bee - - OBS 00 BESSS SSSS8 SeeSeeeen4 Chart C day, and for the second 20 days only 1.05 per day, and the cycle came to an end with encystment of all living material at the end of 40 days. This cycle included only 84 generations, giving a total of 212 generations for the series in culture. The rate of encystment was fairly high throughout this cycle and went up rapidly during the last four 5-day periods (chart D, bottom), while the death rate showed a slight increase over that of the first cycle. Notwithstanding the relatively low division rate, the appetite was remarkably good, nearly 88 per cent of the Paramecium being eaten daily for the first 25 days (chart D, dotted line). LIFE HISTORY OF DIDINIUM Daevth All efforts to recover the series from encystment failed; not one individual could be pursuaded to come out of its cyst, and the race thus came to an end. Y Series: 2nd Cycle Chart D GENERAL Eneystment in ciliates has a three-fold purpose. © First, for protection against adverse conditions of the environment, which, as Fermor has well observed, are usually so delicately adjusted to the equilibrium of an organism that they baffle detection. Such encystment is characterized by no internal reorganization and the organism may be recovered in twenty-four hours or less 238 GARY N. CALKINS by substituting fresh water for the medium in which it had encysted. Second, for reproduction, a phenomenon observed in Tillina, Colpoda and a number of other ciliates, but by no means universal in the group. Third, for reorganization, which has to do with internal processes of the cell. In Didin- ium there is no encystment for purposes of reproduction, but it is frequent for purposes of protection and periodic for purposes of reorganization. In the latter case the approach of encystment can be predicted very often from the reduced activity in feeding and in dividing, from two to four days in advance. ‘This is shown not only by the averages for the entire race but also by individuals and their progeny watched from day to day. When in this condition, fresh water and food have no effect, nor will fresh water added daily bring such individuals . out of their cysts until a period of at least five days has elapsed. A very unexpected result was obtained in these experiments in connection with the phenomena of conjugation. During the first cycle no conjugating pairs were observed in any of the stock dishes although such material is prepared daily and always watched for at least five days. During the first week of the second cycle, epidemics of conjugation appeared in the stock dishes. This period of conjugation lasted about ten days, after which not a pair was seen. Conjugation epidemics appeared again in the third cycle and at a corresponding time. The first pairs were seen in the stock dishes on the third day after recovery from encystment (March 12) and pairings occurred in great numbers until March 20th, after which not one pair could be obtained from the material. During the height of the epidemic in the stock material two cases of conjugation occurred in the isolation cultures. One of these pairs (March 16) was the union of two individuals out of eight derived from one individual isolated the day before. The second case occurred on March 17 between two individuals among sixteen derived from an individual isolated the day before. April 26, 1915 LIFE HISTORY OF DIDINIUM 239 LITERATURE CITED BasBiaAntI, E.G. 1873 Observations sur le Didinium nasutum. Arch. d. Zool. Exper., tom. 2. von Ertancer, R. 1889 Zur Kenntnis einiger Infusorien. Z. w. Z., Bd. 49. Fermor, X. 1913 Die Bedeutung der Encystierung bei Stylonychia pustulata Ehr. Zool. Anz., Bd. 42, no. 8. JENNINGS, H. 1906 The behavior of the lower organisms. New York. Mast, 8. O. The reactions of Didinium nasutum (Stein) with special reference to the feeding habits and the function of trichocysts. Biol. Bull., vol. 16, no. 3. Moopy, J. E. 1912 Observations on the life-history of two rare ciliates, Spa- thidium spathula and Actinobolus radians. Jour. Morph., vol. 23. PranpTtL, H. 1906 Die Conjugation von Didinium nasutum. Arch. Prot., Bear Tuon, K. 1905 Uber die feineren Bau von Didinium nasutum. Arch. Prot., Bd. 5. Wooprurr, L. L., and Erpmann. R. 1914 A normal periodic reorganization process without cell fusion in Paramecium. Jour. Exp. Zool., vol. 17. PLATE 1 EXPLANATION OF FIGURES! 1 to 6 Stages in the process of swallowing Paramecium caudatum. Camera drawings from preparations mounted in toto. 7 Section of dividing form of Didinium. The macronucleus section shows the characteristic reticulum of the early stage of division. Two micronuclei in full mitosis lie in the margin of the macronucleus; several large endoplasmic bodies are shown in section, and two enigmatical bodies, which may be seizing organs in the process of development. 8. Section of Didinium preparatory to encystment. The trichites of the proboscis and the basal fibrils of the membranulae are clearly shown at this stage; the peripheral protoplasm is denser than at other periods; the macronu- cleus shows the enlargement of the contained chromatin bodies. The micro- nuclei at this stage leave the hollows in the margin of the macronucleus, swell, and prepare to divide; one of the four is shown at the left of the macronucleus. The seizing organ is extruded and shows the zone of dark granules near the tip. These probably represent the poison of the ‘trichocyst material’ instrumental in paralyzing the prey. 1 Drawn by Mabel L. Hedge from permanent preparations. 240 LIFE HISTORY OF DIDINIUM peELATE 1 : GARY N. CALKINS 241 we Tal mh THE REACTIONS AND RESISTANCE OF FISHES IN THEIR NATURAL ENVIRONMENT TO SALTS MORRIS M. WELLS THREE FIGURES Neale Cae tuo Meer crexcstis ieee neve lave id sls 6 sos REO ere ae sae 244 TUTE MIRIAYEy ee ea ates Sie ee ieee oe eR oh onc AAG eas Clo hears Sepa Ate 244 Me Niethodstand, apparatuses ie... 2). 4 <> ae eee el ore ots nyse 245 IVA PresentatvonzOl Gata scm cc foc os, 5 seek’ sss oe OE ee ae ar 248 AC MRICACTIOMV eX PCTIMMeNtSs ....5.< c/s «/.... 2s hee ae aes 250 2 MNEACCLOMVGO, MGT ALESn.c6:.6.8 sos oe Cee eee 250 7a SAMMOMUM MItTAteli.. acc... 00 eee eee ee eee 250 bi@Potassimmitrates: ) 2). js eee eer Sea or eal CEA DO CTUIMMIMTGTAGEY ts 0c. S12 ee ee eee 251 dia© alicnumenitratesss'./5 5 s.0 +c. aeRO oe ee oe ee 252 evavViagnesiumeanitrate.:...... .. dase ence eee el actors 252 Se INEACHLONEO SUlpWaesis.-.<...ccls fae ee ee eer cee 253 a. eATHmMoniim sulphate... .\..,.. 2c sae peeereee oe «alee 253 beweotasstum: sulphate...”..... - << aac eee en eis setae 253 Croodiumsulplate: 14. 0: ae eee eee ree eee a eee 253 dCalcwumy sulphate). A... 25.) che eae eee eee 254 ey Magnesium) sulphates. ;...: 4: n semen eae ne ee 254 4. Conclusions from reaction experiments..................... 255 B. Antagonising salts and the reaction of fishes..................... 255 C. Physiological states and the reactions of fishes.................. 260 [i Reaction ot .starved fishes to Ca@le tesserae. ssn. eee 262 Des Reachionvotstanved: tishes to) lowsomyceneeste aes oon eee 264. D Acclimatizationyand the reaction of fishes:s.....2.. 1.2 eees ce 265 Ho esistancesor tishestorsalts... << 2 5... Rae oe ee eee 267 il, IRENE DINOS tO) EHaohaavopavHUNON ENNIS Gsocccosncddoomoncaneabenaas 267 Zohesistance to;potassium salts....sseveecmeckc: ieee a 271 SLVCSISFAMCeNLOMSOGIUI! SAlLShis sce eer Ieee) Scie ee eee 272 4. Resistance to calcium and magnesium salts................. 273 Wes (Generales cuisstOml+:)ce. os \ais,.« « sich Wo RS oe eR a ee 274 Wilton Ge Wena lECOIMGIUSTONS lance ¢ ciecacl Sey ately wic kag ea eta oe ee eB 280 ES Tipe Tapo yee me acter tele 2 oles. is Ree ab ce Wo Hae on RE Ree SW EOE eed 282 243 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 3 ocTOBER, 1915 244 MORRIS M. WELLS I. INTRODUCTION In a previous paper (Wells ’15 a) the reactions and resistance of fresh water fishes to alkalinity, acidity and neutrality were discussed upon the basis of experimental evidence which seemed to indicate that the chemical reaction of the water (i.e., acid, neutral or alkaline) in which the fishes live, is a matter of con- siderable importance to fresh water fishes and probably to marine fishes also (Shelford and Powers 715). In the present paper a large number of experiments bearing upon the reactions and resistance of fresh water fishes to salts is presented. Practically no previous work has been published upon the reactions of fishes to salts and the main part of the data presented here has to do with this phase of the subject. Some interesting relations be- tween acidity and resistance to salts are also presented. This latter phase of the subject has been worked out in a prelimi- nary way only; the more definite relations are left for further investigation. The present investigation was begun at the suggestion of Prof. \V. E. Shelford and was carried on at the University of Chicago during the years 1912 and 1913. In the fall of 1914 operations were transferred to the University of Lllinois as the author accompanied Dr. Shelford in his transfer to that place. Il. THE WATER The differences in the water of the two institutions have been discussed in the first paper of the series (Wells, l. c.). The chief differences are the following: The water at Chicago comes from Lake Michigan; as it flows from the tap in the laboratory, it is slightly acid with carbon dioxide (2-3 cc. per liter), is super- saturated with O, (8-10 ce. per liter), contains 32 cc. per liter of half-bound CO, (bicarbonates) and a proportionate amount of other salts. The water at the University of Illinois comes from deep wells. As it flows from the tap it is strongly acid (18 ee. CO, per liter), contains practically no O, (0.12 cc. per liter) and the half-bound CQ, equals 101 cc. per liter; other salts are in proportion. Aeration brings the two waters to more nearly REACTIONS OF FISHES TO SALTS 245 the same condition and fishes can live in either after the proper amount of aeration. Too much aeration causes the Illinois water to become alkaline to phenolphthalein and fresh water fishes cannot live in such water. lil. METHODS AND APPARATUS The reaction experiments have been performed in the gradient tank used in the acid gradient experiments (Wells *15 a, fig. 1, pe223): The salts used have been, the chlorides, nitrates and sulphates, of ammonium, potassium, sodium, calcium and magnesium. In presenting the results of the reaction experiments the salts will be grouped with reference to the anion, as the similarities in behavior, in the different salt gradients, make this a rather natu- ral division. They will also be taken up in the order of increas- ing toxicity of this ion as worked out by Lillie (’10) and others. Thus the order of consideration will be, chlorides, nitrates, sul- phates. In considering the resistance experiments, on the other hand, the salts will be grouped according to the kation. In the gradient experiments the concentration of salt introduced at the salt end has been in nearly all cases 0.01N. In a few experiments the concentration was made 0.02N or even higher in an attempt to drive the fishes out of the salt end, to which they were giving a positive reaction. These experiments will be cited as they come up. The gradient in the salt experiments was obtained as follows: Tap water was set to flowing into one end of the tank at the rate of 500 cc. per minute, and into the other end at the rate of 400 ec. per minute. A 0:05 N solution of the salt was made up with tap water and run into the flow at the 400 cc. end at the rate of 100 ce. per minute. This made the volume of the flow at the two ends equal. The salt solution was mixed with the tap water, in a mixing bottle, outside the experimental tank. From the mixing bottle a single outlet led to the experimental tank. At first the gradients were tested before and after each experiment. Later, after a very careful study of the gradient 246 MORRIS M. WELLS had been made at Chicago, by determining the conductivity of the water at various points in the tank, tests were no longer made. Thus the actual concentrations existing throughout the tank have not been determined in each experiment but the study that was made indicated very clearly that under the given con- ditions this concentration is almost constant for a given salt. Thus there always exists a gradient of the dissolved salt, between the two ends of the tank. The presence of this gradient is shown by the reactions of the fishes as well as by the conductivi- ties and titrations. That the gradient is not perfect is to be expected; its peculiarities were brought out in the study which -OOSN .0035N | ,003N -OOIN sO01N -007N -OO6N | -OO6N ! Fig. 1 Longitudinal section through the gradient tank. The figures indicate the concentrations of the salt at the depths indicated by the arrows. These concentrations were ascertained by determining the conductivity of samples taken from the different parts of the tank; in determining the gradient 7 samples along any given level were taken; only five are shown in the figure. was made by means of the conductivity method. Figure 1 shows the gradient as it existed after the flows at the ends had been running for some time. It will be noted from figure 1 that at any given level there is a gradient of salt from end to end of the tank. The concentra- tion at the bottom of the tank was much higher than that near the surface of the water, and thus the fishes at times reacted to the vertical gradient, which was much sharper than the hori- zontal one. This reaction to the vertical gradient did not inter- fere greatly with the experiments, however, because the fishes tend to swim back and forth in the tank at whatever level they may be. Furthermore, most of the fishes worked with, remained near the bottom for a large proportion of the time. alone, 2) eee. Ne Ca (NO3)2 + trace NaNO.... .. Ca (NOs). + trace Mg(NOs)o. .. KIND OF WATER USED IN GRADIENT REACTION OF FISHES IN PER CENT OF TIME SPENT IN HALVES OF TANK. POSITIVE = IN SALT HALF; NEGATIVE = IN TAP OR DISTILLED WATER HALF Slightly acid tap Slightly acid tap Strongly acid tap Slightly acid tap Slightly acid tap Distilled water Distilled water Strongly acid tap 256 Per cent Per cent positive negative 30 70 79 21 96 4 23 77 76 14 19 81 87 13 53 47 REACTIONS OF FISHES TO SALTS D5 binations of salts in a gradient, and I present a number of experi- ments of this sort here. They show that fishes recognize and react to combinations of salts according to the prediction which might have been made from the results of previous work upon antagonism. It will be remembered that fishes are slightly positive to 0.01N NaNO; in moderately acid water. It was found, however, that they are negative to this salt in water that is but faintly acid and for this reason the following experiments were run in water which contained less than 8 ec. COs, per liter. The experiments were first run as regular salt experiments such as have been described before, and then the antagonizing salt was added to the salt flow. The concentration of the original salt was always 0.01N and that of the antagonizing salt 0.0002N, i.e., a bare trace was added. The results of these experiments are shown in table 2. A considerable number of experiments was run with each combination in the tap water, and then some check experi- ments were run in distilled water. The results in the distilled water gradients are very similar to those in the tap water. Discussion of the experiments with antagonizing salt combinations Table 2 shows clearly that the antagonistic action of the salts is detected and reacted to by the fishes. This is shown, for instance, in the sodium nitrate experiments; here the fishes were 70 per cent negative to this salt in slightly acid water but when a trace of calcium nitrate was added the negative response fell off to 21 per cent and the positive rose from 30 to 79 per cent. Then in strongly acid water the positive response increased to 96 per cent. The reactions of fishes in any gradient are due to their tendency to move about until they reach an environment that neither over- nor under-stimulates them. Thus they will not remain quietly in water that is strongly acid nor will they do so in water that is neutral. A slight degree of acidity (1-6 ec. CO» per liter, Wells 715 a) furnishes their optimum stimula- tion as far as H ion is concerned. The reversal in reaction of the fishes in gradients to which a trace of an antagonistic salt 258 MORRIS M. WELLS has been added, must then be due to the fact that this trace of salt lessens the stimulation in the salt end of the gradient. There are three principal factors affecting the degree of stimulation of the gradients referred to in table 2, namely, the original salt (e.g., NaNO;) the antagonising salt (e.g., Ca(NOs3)2, and the acid. Before the antagonising salt was added the fishes were negative to the original salt, even though this meant spending most of the time in a degree of acidity which was slightly above their optimum. With the addition of the antagonising salt, however, they reversed their reaction and became positive to the salt end. The antagonising salt must have diminished the original stimulation in the salt end or have increased the stimu- lation in the acid end, or both. The work upon the effect of acids and salts upon permeability suggests that both factors were concerned. Lillie (10) has shown that calcium salts de- crease the permeability of egg membranes while the salts of sodium increase the permeability. Osterhout (12, a and b) has shown that sodium salts increase the permeability of plant cells while the addition of a trace of calcium salt maintains nor- mal permeability even in the presence of an excess of the sodium salt. Osterhout has also shown that there exists a mutual antagonism between certain acids and salts as for instance be- tween HCl and NaCl, but the salts of calclum and magnesium work with rather than against the acid. In the above experiment, therefore, the addition of the cal- cium salt to the end of the gradient which contained a sodium salt in concentration strong enough to cause the fishes to give a negative reaction, resulted in the fishes becoming positive. This reversal in the reaction of the fishes must have been due to the decrease in the stimulating power of the salt end. It has already been shown that increasing the acidity of water will cause fishes to become positive to concentrations of sodium salts to which they are normally negative (table 2) and it was found that the higher the acidity the higher was the concentration of sodium salt selected by the fishes. . Table 2 also shows that the antagonistic action between cal- cium and sodium salts is detected and reacted to, when the con- REACTIONS OF FISHES TO SALTS 259 Bull-head No. 2 | | yee ml Say eat aes x | S | SiS) | I O | |©O o | £2) | & Tap |Z es Tap || 2 Tap | a+4 Tap | Z We pas roel la & | aera | 3 2 a eos iS 4 | Sey IS 4, 5 | ! ! ! | I | | | Fig. 2. Showing the reaction of bull-heads (Ameiurus melas) to Ca(Nos)s, alone, and in combination with a trace of NaNO;. The concentration of the calcum salt was 0.01 N throughout and that of the sodium salt 0.0002 N; experi- ments performed in distilled water; numbers at left indicate time in minutes. centrations of the two salts are reversed, 1.e., when the calcium nitrate is present in 0.01N concentration and the sodium as a trace (0.0002N). The data for table 2 were obtained from graph experiments and also from readings. The same fishes were used in the gradient with the different conditions, i.e., they were first graphed in the gradient with the sodium salt (e.g.) alone and then again after the calcium salt had been added to the flow. To illustrate more accurately this method of experi- ment figure 2 is inserted. The graphs shown in this figure are 3 of those made by 4 bull-heads. The experiments were run as follows: The gradient with only Ca(NO3;). flowing in at the salt end, was obtained by allowing the flow to continue for 30 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 3 © 260 MORRIS M. WELLS minutes. The fishes were then taken from the large aquarium and placed in pans of water, numbered 1,ete. The fish from pan No. 1 was placed in the gradient and its movements graphed for 15 minutes. It was removed and No. 2 was placed in the gradient and graphed. This was repeated for Nos. 3 and 4. A trace of NaNO; was now added to the inflow at the salt end; after 20 minutes fish No. 1 was again placed in the gradient and its movements graphed for 15 minutes. This was repeated for the three remaining fishes in the same order as before. The graphs show the marked difference in the reactions of the fishes before and after the trace of sodium nitrate was added. C. PHYSIOLOGICAL STATES AND THE REACTIONS OF FISHES In the discussion so far attention has been called to the fact that in most of the series of experiments, there was a small percentage of the fishes (usually 3-5 per cent) which gave re- actions more or less the reverse of those given by the majority. Such exceptions to the general behavior are common in experi- mental work of all sorts and probably indicate physiological differences upon the part of the organisms. That such physio- logical differences, i.e., physiological states, exist and that they influence very markedly the reactions of the animals has been proven beyond doubt (Child 718, and Allee 712). Allee and Tashiro (’14) have shown that the reactions of isopods are very closely correlated with the metabolic activity and Allee (’12) has shown that by changing the rate of metabolism he can alter and even reverse the reaction of isopods to current. A correla- tion between the rate of metabolism and the reactions of amphi- - pods has been shown by Phipps (’15). At Chicago during the winter of 1913-1914, a study not yet published was being made of the effect of starvation upon the resistance of fishes to KCN and low oxygen; it was thought that the starving fishes furnished good material for ascertaining dur- ing the same period something of the effects of starvation upon the reactions of fishes in gradients. Accordingly a series of 89 experiments was run with the starving fishes in gradients; 50 REACTIONS OF FISHES TO SALTS 261 of the experiments were in gradients of CaCl, since it seemed best to confine the experiments to a few salts at the most. It was decided that the starving fishes should not be handled to any great extent during the obtaining of the data for which the material was originally intended. A few experiments were run in gradients of Ca(NQO;). and MgCl, the results of which were much like those for CaCl.. Nine expriments with starving fishes in low oxygen gradients are included as they are significant. The experiments with starvation and resistance of fishes showed in brief the following points: The fishes as they began to starve became more resistant to KCN and low oxygen. This rise in resistance which is a decrease in susceptibility, continued for some weeks (varied with species). There was then a rather sudden decrease in resistance (increase in suseptibility) which was found to be a close fore-runner of death. In terms of metabolism, as starvation in certain fishes proceeds the rate of metabolic acitvity is at first decreased. After remaining below normal for some weeks (or even months) the forces which are inhibiting the rate of reaction, give way and the rate runs up rapidly to, and beyond, the normal rate. Whether the changes in the physiological condition of the fishes are wholly quantitative is not certain. It is very probable that a change in the rate of metabolism does not express all that takes place but there may be alterations in the kind of metabolism also; in other words starvation in fishes may produce qualitative as well as quantitative changes in metabolism. Starvation experiments were run with several species of fishes including the rock bass (Ambloplites rupestris), small mouth black bass (Micropterus dolomieu), pumpkin seed (Eupomotus gibbosus), mud minnow (Umbra limi), and the black bull-head (Ameiurus melas). The fishes seemed to be divided into two groups as far as their starvation reactions are concerned. The bull-heads made up one group and the other fishes a second. Most of the work was done with the bull-heads and the rock bass as representatives of the two groups. In the case of the rock bass some quantitative data can be presented. 262 MORRIS M. WELLS 1. Reactions of starved fishes to CaCl, Normal bull-heads are negative to 0.01N calcium chloride in a gradient. It was noticed, however, that when food was given these fishes they often became positive to the salt half of the tank. To check this reaction 23 experiments with normal, well fed and starved bull-heads were run. Table 3 shows the results obtained. It shows that normal fishes (bull-heads) are negative to 0.01N calcium chloride well-fed ones positive, and starved negative again. The well-fed bull-heads were in fact given all the food they would eat and thus were really over fed, as they ate until their abdomens were much puffed out. The data in TABLE 3 Showing the reactions of normal, over-fed, and starved bull-heads (Ameiurus melas) to .01N calcium chloride, in a gradient. Data shows per cent of time spent in the halves of the tank NORMAL REACTION OVER-FED REACTION STARVED REACTION FISH NUMBER CaCl Tap CaCl Tap CaCle Tap 1 29 71 66.5 33.9 32 68 2 44 56 78 22 63 37 3 34 66 57 43 29 TL 4 37 63 73 27 40 60 5 31 69 U 22 52 48 table 3 is taken from the graphs made with 5 fishes. The normal reaction of each fish was determined immediately upon bringing it into the laboratory from the streams. On the next day the fishes were fed all the beef they would eat and graphed again on the third day. They were then starved and graphed from day to day. The figures in column 4, table 2, are those obtained after from 5 to 10 days starving. Each day calcium chloride was run into the end of the tank opposite that of the day before. The method of experimenting with the rock bass in the resist- ance experiments was to bring them in from the creeks in which they live and to weigh them individually, and at once. The process of starvation was then kept track of by successive weigh- ings. Twenty-six experiments with these starving fishes were REACTIONS OF FISHES TO SALTS 263 run to determine the effect of the starvation upon the reaction to CaCl. It will be recalled that normal rock bass are negative to this salt in 0.01N concentration (except in the case of large fishes). The starving rock bass were therefore experimented upon in gradients of CaCl, at different stages of starvation, with results such as those shown in table 4. This table shows that starvation increases the percent of positiveness of these fishes. This is true for that period of starvation, during which the rate of metabolism is slowed up. The few experiments that were performed upon fishes in which the factors inhibiting starvation had broken down and the rate of metabolism had gone above normal, indicate that the fishes are again negative to Ca salts at this time. TABLE 4 Showing the reactions of normal and starved rock bass (Ambloplites rupestris) to .01N concentrations of CaClz in a gradient. Reactions are shown in per cent of tume spent in the two halves of the gradient tank REACTION IN PER CENT FISH NUMBER UA TELOS oe Sareea in ae oe ee bac g ta OVEMSICE ONS) MENT GRAMS EXPERIMENT CaCle Tap water 1913 1913 1 Nov. 20 | Nov. 28 9.9 8.9 30 70 2 20 23 23.1 22.5 43 57 3 20 23 56.2 54.5 38 62 4 20 23 70.6 68 .4 22 78 5 20 23 126.0 124.0 90 10 6 Oct. 16 23 él 18.6 100 0 7 16 Ze 66.0 61.7 30 70 8 16 23 90.9 77.0 10 90 1914 9 Dec. 6] April 9 97 .0 64.2 38) le ne G2 9 6 10 97 .0 64.0 73 27 9 6 10 Jends of gradient reversed 82 18 9 6 15 97 .0 67.3 34 66 9 6 15 jends of gradient reversed 30 70 9 6 16 ' 97.0 65.0 67 33 9 6 17 97 .0 62.3 76 24 10 6 5 83 64.2 46 54 10 6 10 83 61.7 61 39 10 6 10 jends of gradient reversed 58 42 10 6 16 83 61.0 60 40 264 MORRIS M. WELLS Note (table 4) that the normal fishes were negative to 0.01N CaCl,, that with the small fishes this reaction had become posi- tive by the end of a little over a month (fish No. 6) while the larger fishes were still negative. Fishes Nos. 9 and 10 show the reaction of fishes starved for almost four months. These fishes were kept in running water and probably obtained a little food but the successive weighings showed that the process of starva- tion was a continuous one. Note the reversal in reaction of fish No. 9. The first experiment with this fish shows it to be slightly negative. On the next day it had become positive, as was shown by two experiments, with the salt flow at one end of the gradient tank in one, and reversed in the other. The weigh- ings show that the fish had increased in weight since the day before and this increase must have been due to the securing of food in some way; the food had temporarily restored the normal reaction. However, by the next day the weight had again fallen off and the fish was once more positive to the salt, as is charac- teristic for starving fishes. 2. Reaction of starved fishes to low oxygen The results of the experiments with starved fishes (rock bass) in low oxygen gradients are seen in table 5, which shows that TABLE 5 Showing the reactions of normal and starved rock bass to low oxygen in a gradient. Reactions are expressed in per cent of time in the halves of the tank (work done at Chicago) DATE OF WT. AT REACTIONS IN PERCENT wisn wonmame |, 2AUEOF | emi | ONO | armen | OF MME IN Low O2 | Tap water Normal fishes 1913 1913 1 Nov. 20 | Nov. 22 ILS ¢/ IS 5 95 2, 20 22 9.9 9.5 25 75 3 20 DP, 2 eyall 2225 32 68 4 20 22 70.6 69.0 20 80 5 20 BP || 150) 124.0 91 9 Starved fishes 6 Oct. 16 22 PASI 18.6 44 56 @ 16 22 66.0 61.7 50 50 8 16 DD; 90.9 CA AD 34 66 REACTIONS OF FISHES TO SALTS 265 normal rock bass are negative to low oxygen (1 ce. per liter at the low end) as has been shown also by Shelford and Allee (713, p. 236). Large rock bass seem to be an exception to this general rule as they are not always negative to low oxygen, and in some cases seem to definitely prefer the low oxygen end of the gradient, spending a majority of the time there. The cause of this re- action has not been determined but it may have to do with the concentration of hydrogen ion which would probably be a little higher in the low oxygen end than in the high oxygen water, the difference being due to the difference in the effect of the two kinds of water upon the elimination of carbon dioxide by the organism. Fishes Nos. 6 to 8 (table 5) are the individuals occurring under the same numbers in table 4. There it was noted that their reac- tion to the CaCl, had become more positive than the normal re- action and in table 5 it will be noted that these fishes are less sen- sitive to the low oxygen also. Fish No. 5 is also the same in tables 4 and 5 and it will be noted that this fish was positive to both low oxygen and 0.01N CaCk. Experiments in low oxygen gra- dients were not performed with these fishes later in their period of starvation but the data given indicate that as they become somewhat starved they at the same time become less negative to low oxygen. This indicates that their metabolic rate is slower than normal. D. ACCLIMATIZATION AND THE REACTION OF FISHES During the course of the experiments, considerable evidence was accumulated concerning the effect of acclimatization upon the reactions of the fishes. A few experiments with fishes in CO, gradients indicated that these fishes after living for two to three weeks in water whose CO, concentration was 8 to 10 ce. per liter, were more sensitive to the CO, than normal fishes. To determine whether or not the presence of an excess of salt would result in similar reactions to the salt in a gradient, a series of acclimatization experiments with CaCl, was run. A medium-sized (45-gram) rock bass was graphed in a CaCl, gradient; its normal reaction was decidedly negative to 0.01N 266 MORRIS M. WELLS CaCl,. It was now placed in a 20-gallon-jar full of a 0.01N solution of this salt. Each succeeding day it was taken from the jar and its reaction in the gradient graphed, when it was re- turned to the jar. This was continued for 6 days; the con- centration of the solution in the jar was then riased to 0.05N. The fish was left in this solution 4 days longer, being graphed each day. It was then returned to the tap water and graphed again after 2 days. In making the graphs each day, the salt solution was run into the end of the gradient tank, opposite that of the day before. 0@,8 eo Oey ot Seacual forrake AS 4, P Male Pilar Eglin 5 Spundée g & First Seacuatl \ | 4 Se umalocyle Diagram 1 Illustrates the chromosomal cycle of Phylloxera caryaecaulis; to the left is shown the line that culminates in the sexual female; to the right is shown. the line that culminates in the males; this line splits into two subdivisions after the extrusion of the polar body in the male-producing egg. PREDETERMINATION OF SEX 293 When the polar body is thrown off all of the chromosomes divide. The migrant that develops from this egg comes to produce large eggs with the full complement of chromosomes in them. All of the chromosomes divide when the polar body is produced by this egg (plate 1, fig.n). The egg itself then develops into the sexual female that comes to produce a single egg. When this egg is ripe the chromosomes have united in pairs, of which there are three. In reality one of these pairs must be thought of as made up of the two large sex chromosomes = the two small chromo- somes attached to them. If we next turn to the line represented on the right of the dia- gram we see that the stem-mother contains the same group of chromosomes as in the other case (one of the small sex chromo- somes is marked). All of the chromosomes divide when the polar body of the egg of the stem-mother is given off. The egg gives rise to the migrant that comes to produce the small eggs. When the polar body is about to be produced in these small eggs (or before that event) a shifting of the sex chromosomes takes place, so that the two large X’s come together as do the: two small ones, the latter leaving their former loose attachment - to the large chromosomes in order to combine. When the polar body is given off the four autosomes divide, while the conjugated pairs of sex chromosomes separate; a member of the larger pair lags on the spindle and is precociously split lengthwise, while the small pair separate and the outer member does not lag on the spindle (plate 1, fig. 0). It will be seen that two kinds of eggs should result according to whether the marked chromosome passes out into the polar body, or remains in the egg. [Tf it passes out, the male that results will give rise, as shown in the figure, to female producing spermatozoa that contain the open x. If we suppose the sexual egg is fertilized by this kind of spermatozoan, the stem-mother that results will give rise to the female line (to the left). If we suppose that the other kind of sperm—the one with the marked x—fertilizes a sexual egg, a stem-mother will be produced that leads to the male line (to the right). The assumption that there are two kinds of males is not entirely hypothetical; for, in my former paper I pointed out 294 T. H. MORGAN that two kinds of males are found that differ from each other in the relation of the small x to the large X chromosome—in one kind of male the two remain in contact, in the other the two keep apart. And I pointed out that all of the spermatocyte cells of a given individual show one or all show the other relation. This is in accord with the assumption of two kinds of males that arise when the polar body is eliminated. In some of the preparations showing anaphase figures in this species, one of which is redrawn here in plate 1, figure o, the two lagging chromosomes appear unequal in size. In other figures, however, (figs. 5, 6, 21, of my 1912 paper) the two lag- ging chromosomes appear to be equal or subequal. For this reason I stated (12) that the two equal chromosomes represent one large sex chromosome divided into halves. On this view the conjugating pair of small x’s have been separated and moved to their respective poles while only the large X lags as it passes to the outer pole. It will be noticed that superficially the number of chromo- somes in P. fallax is double that in P. caryaecaulis. This sug- gests that the group in P. fallax represents the double number of chromosomes of caryaecaulis (tetraploid), or that caryaecaulis is P. fallax halved. But if the full number of chromosomes in P. caryaecaulis is looked upon as eight, this relation does not hold, unless one supposes that there are four small chromosomes present also in P. fallax (giving sixteen in all) that are attached permanently to other chromosomes. From this point of view the chromosomes of the one species would be double the number of the other and P. fallax would become XX and XY, while P. caryaecaulis would become X and Y. In other words the small chromosomes would no longer be reckoned as factors in the sex scheme. I have preferred, however, the interpretation given in the diagram because the two small chromosomes conjugate in the male egg when the large sex chromosomes conjugate; and, again, because one of them lags in the first spematocyte division along with the lagging X. But the latter relations may be only a con- sequence of the first, and be due to the absence of a mate at this time. Still, since it goes into the female-producing sperm, and PREDETERMINATION OF SEX 295 not indifferently to either pole, this would seem the more natural designation. Without wishing, therefore, to lay too much stress on the nomenclature, it seems to me that the one I have followed represents more naturally the facts as described in the text. SEX RATIOS IN PHYLLOXERA FALLAX Since the galls in this species do not open to release the sexual forms until a considerable number of them have hatched, and since all of the inhabitants (except for rare cases) are the de- scendents of a single female it is possible to get a fairly good sample of the output of each stem-mother. In order to antici- pate a possible objection, viz., that two or more stem-mothers might be included within the same gall, | opened a number of very young galls and found, with the rarest exceptions, that each gall is the result of the activity of a single female. We are safe in concluding, therefore, that the contents of each gall is the pro- duct of a single stem-mother. But whether her daughters, the apterous ‘migrants,’ are individually small-egg-producers or large- egg-producers can not be determined; because, in this species, the apterous ‘migrants’ deposit each egg as it matures. Since eggs that are not quite mature look like small eggs they can not be distinguished from them. Only by finding a gall in which a single apterous ‘migrant’ was present that had laid both kinds of eggs would it be possible to settle this question. I have al- ready reported that occasionally the migrants are winged, and that when this occurs each individual contains eggs that were all of one sort, namely, small eggs in this instance, but this does not settle the other question, however probable it may appear, that each individual of this generation produces only one sort of egg. In the following table the counts from 26 galls (July 11) are given. Those with the same letter come from the same leaf. The galls were old and ready to open. The old apterous ‘mi- grants’ were still present, but empty, as though they had about finished their productive life. There were present as many eggs unhatched as hatched; only the latter appear in the record, or 296 T. H. MORGAN rather the individuals that hatched from them. In some of the galls the dwarf forms recorded in my former paper were also found. Except in a few cases the females were in excess of the males. If galls are opened, when the first sexual forms have begun to hatch, males (one, two, or three, ete.) will be found. This means, TABLE 1. MALES SEXUAL FEMALES OP 28 36 OP 20 31 OP 6. 14 OP 15 2 OP. 30 15 A tale 4 11 TT 18 32 AMR 29 39 aa oy) 6 Atak 21 28 on 15 26 SS 24 33 SS 21 23 ss 2 7 Shy) 9 18 SS 2 2 SS 14 WZ SS 12 28 ss 18 12 4 20 10 21 WwW 7 51 WW 17 30 WW 30 32 WW 8 21 WW 36 66 405 628 no doubt, that males hatch first, or possibly that the first eggs produced are males. Since later more females than males are present, it follows that more female-producing than male-pro- ducing eggs are laid. The most striking departure from the ordinary ratios is that in the first count in WW where there were 7 males to 51 females. PREDETERMINATION OF SEX 297 Galls of the same size on the same leaf must have been sub- jected to as nearly uniform conditions as possible, but the counts from the same leaf are not strikingly at least more like each other than they are like other counts from other leaves. It is idle perhaps to speculate as to the factors that predeter- mine whether a given egg shall be a large or a small one especially as this may depend, as in the other species, not on the conditions that affect the migrant, but on the conditions that determined the nature of the migrant herself. The recent work of Whitney and of Shull shows that very slight differences in the environment turn the scale in sex predetermination. Differences like those described by them might accompany the differences in food con- ditions of the leaf during the day when starch is being made and during the night when starch is being converted into sugar. IH, perchance, such environmental changes affect the stem-mother, the differences as to output shown by the galls, might arise. THE SPERMATOGENESIS OF THE BEARBERRY APHID Stevens and von Baehr have described very completely the spematogenesis of several species of aphids. [| have also studied at different times a number of species but as they gave nothing new I have not published the results, except for one very brief account (Jour. Exp. Zodél. ’09. pp. 298-305). There is one species, however, which in clearness and simplicity exceeds all others that I have seen. I give here therefore an accountof some of the critical stages in its spermatogenesis and oogenesis. The data furnish, moreover, an occasion to make certain com- parisons between the phylloxerans and the aphids. The species in question, Phyllaphis coweni Cockerell, forms galls on the bearberry. My material was found north of Quissett, Mass., near Woods Hole, where I have collected material for four summers. I have found the galls on the bearberry during June, July and August. Each contains, as a rule, a single stem- mother—rarely two—and her progeny. As the gall gets older the progeny is seen to be made up of larger individuals with wing pads and in addition a series of immature forms, as well as afewmales. The large individuals with wing pads contain sexual 298 T. H. MORGAN eggs. As I have never found winged individuals within the galls the individuals must leave the galls when ready to expand their wings. The origin of the males puzzled me for some time until I discovered that the stem-mother produces them—at first spar- ingly, but later in larger numbers. The same stem-mother that contains the males also contains young sexual females. She also produces at times both males and individuals that contain par- thenogenetic eggs and embryos with six chromosomes. Throughout July and August young galls can always be found at the growing ends of some of the branches. These galls con- tain young stem-mothers, and later some of their progeny. These younger stem-mothers may possibly come from the old stem- mothers, or from belated eggs of the previous year, or from sex- - ual eggs of the same year in which they appear. A more de- tailed examination will be necessary to settle this point. The sexual eggs begin to pass through their synapsis stages while the young are still present in the stem-mother but even after the young are born and after some of the eggs have left the ovary, eggs at the outlet may show the chromatin contracted at one side. In the male the two reduction divisions may also take place while the young individual is still within the stem- mother, but other individuals do not contain these stages until after the young males have been born. Entire individuals were preserved in Carnoy solution, the ab- domen. cut into sections, and the sections stained in iron hemo- toxylin. The carly spermatogonial cell contains f've chromosomes (dia- gram 5). Ata later stage I found what seems to be a contraction feure, plate 2, figures 1 and 2, when the chromatin is shrunken and lies at one side, but as there is nothing specific about this condition it is with some hesitation that I identify it as the synizcsis stage. The early prophases, plate 2, figures 3, 4 and 5, are interest- ing. The three chromosomes are easily distinguished, even be- fore they shorten into rods. Miss Stevens has described a stage that she calls the synapsis stage; but from her figures it seems not improbable that she has PREDETERMINATION OF SEX 299 seen only the early prophases of the first division or possibly the stage between the first and second spermatocytes. The side view of the first spermatocyte division is shown in plate 2, figure 6. An equatorial plate is shown in plate 2, fig- ure 5, which gives the relative sizes of the three chromosomes. Stages in the later first division, some of them duplicates of each other, are shown in figures 7 to 11. The two autosomes di- vide; the X chromosome is drawn out, and is usually dumb-bell- shaped. It generally shows a well marked longitudinal split. The split is so distinct that the halves appear often like two parallel lagging chromosomes. Only in the latest stages of the division is it apparent that the whole X chromosome passes into the larger of the two cells. As in other aphids, the X chromo- somes are, during this division, often constricted near the middle, which in some species is sometimes carried so far that the two enlarged ends are connected by a mere strand. It is this ap- pearance that has led Miss Stevens in her earlier work to infer that the X chromosomes were really divided at this time. In the aphid of the bearberry the constriction appears at first at the middle of the chromosome. It seems later to pass more and more towards one end, until ultimately, as shown in figures 11 and 13, a small piece only is left at one end, which in most cases is later drawn into the thread; although once or twice I have found cases, as in figure 12, in which this terminal piece appears to have broken away from the strand connecting it with the rest. The difference in size of the two cells varies greatly in the earlier phases, as the figures show. But ultimately nearly all of the protoplasm passes into one cell, the one that contains the X chromosome. The small cell is left with two chromosomes and a small amount of cytoplasm. It never divides again, and later degenerates. Stevens was inclined to think that the small cell may sometimes show a division figure, which subsequently fades away, but I have never seen a case of this kind. The two autosomes in the functional cell begin to lose their condensed condition and spread out into loose masses, as shown in figures 16 and 17. The X chromosome is later in passing into this con- 300 T. H. MORGAN dition, but does so before condensation sets in again, prepara- tory to the next division. In one case, figure 15, the X chro- mosome failed to draw out of the smaller cell, and one end only lies within the nucleus of the larger cell. This end has begun to open out as the other chromosomes have already done. The condition of the other cells in the cyst makes this interpretation probable. During the resting stage, the X chromosome becomes con- densed, as shown in plate 2, figures 14, 15, 16, 17. The chromo- somes in the larger cells next begin to condense, preparatory to the second division. The X chromosome is distinctly larger than the other two; all three divide equally to form the sperma- tids (plate 2, figs. 18, 19, 20). The double nature of the X chromosome in the first spermato- cyte division is so apparent that it invites speculation concern- ing the nature of the division. Presumably the second divison occurs in the plane of the split, but it is impossible to follow this chromosome through its resting stage. Similar cases for the single X chromosome are known, and [| can but follow the usual interpretation, viz. that we are dealing with a precocious divi- sion. The tetrad formation that occurs in such forms as Asca- ris is interpreted as a double division, one of which is precocious. The rapidity with which the two reduction divisions take place, often without an intervening resting stage, indicates that each of the chromosomes, even though mated in pairs, has under- gone the preparatory stages of division. Hence two divisions are necessary to separate the four elements. But if this were the whole of the matter it is not apparent why, in a case like this one, the halves of the X chromosome do not separate from each other at the first division. It is perhaps little more than an evasion of the difficulty to suggest that the divisions in the X chromosome are not sufficiently advanced, when overtaken by the first division. Janssens has proposed a view of the necessity of the two re- duction divisions, based on his observations of the chiasma type. His studies of Batracoseps have shown that two of the four threads of the tetrad sometimes break and reunite so that two PREDETERMINATION OF SEX 301 new threads are made up of parts of each of the two original threads. Under such circumstances a single maturation divi- sion would often produce a dyad in which the two threads are genetically unlike. Janssens assumes that this is repugnant to the scheme of reduction, whose purpose is to give rise to a gamete with a single set of units (gens). This solution of the problem rests on the supposed necessity of pure gametes, and this in turn could be more directly accomplished, it would seem, if the conjugating chromosomes separated without interchanging parts, as they do, in fact, in the male of Drosophila. Granting, however, that ‘crossing over’ does occur as a necessity of the physical conditions prevailing at this time, Janssens’ hypothesis might appear to furnish a solution, if at the same time the ne- cessity for pure gametes could be shown essential to development. Obviously, however, the act of conjugation is a capital arrange- ment to give the zygote a heterozygous make-up, and why a quadruple set of gens instead of a double one would be a disad- vantage is not self-evident. A TETRAPLOID CYST IN THE BEARBERRY APHID In a male of the bearberry aphid one cyst was found in which all the cells have the double number of-chromosomes including the sex chromosomes. The cyst is in the first spermatocyte stage. Since the other cells in the same testis and in the testis of the other side are normal the tetraploid condition must have arisen from a spermatogonial cell whose chromosomes but not the cytoplasm divided. The other possibility, namely, that the four autosomes failed to conjugate, would not account for the presence of two X chromosomes, nor explain the doubling in all the cells of the cyst, because conjugation occurs long after the cells of a cyst have become separated. The former interpreta- tion is therefore to be preferred. The cells in this cyst, of which a few are shown in plate 2, figures 21-30, are completing, or else have completed, the first spermatocyte division. Taking the figures in order, we see in figure 21 four autosomes at each pole— only three show in the larger cell—and two X’s extending from 302 T. H. MORGAN one nascent nucleus to the other. Practically the same rela- tions are shown in figure 22, where however, the two X’s appear to be passing into the larger cell. In figure 23 the two X’s stand end to end as is also the case in figures 24 and 25. It is not pos- sible to determine into which cell they might have passed—pre- sumably, however, one to each cell. In figure 26 this result . seems to have been attained, since each cell contained four auto- somes, and one X. In figures 27, 28, 29 and 30, the division having been accomplished, only the larger cell is shown. In each case there are two X chromosomes and four autosomes in the cell. There are some questions here of theoretical interest. The autosomes appear to have acted as pairs, if one may judge by the equal distribution to the daughter cells, which seems to have taken place in many cases, although it can not be established for all cases. As there are four autosomes of each kind their copu- lation in pairs does not seem unexpected. The behavior of the X chromosomes is unique. In some cases they have passed to opposite poles and in this sense have acted as a pair, but in most cases they have passed into the larger cell. The first spermato- cyte division in phylloxerans and aphids is of such a kind, that the X chromosome passes into a particular cell, i.e., it does not pass indifferently to either pole. But as a matter of fact we do not know whether the larger cell into which it passes is prede- termined (by some polar relations in the cell) and the X chromo- some follows this preexisting condition, or whether that cell be- comes the larger one, which happens at the time of division to contain more of the X chromosome (owing, let us say, to its ac- cidental excentric position). If one may judge from the appear- ance of the early stages of the first spermatocyte division, the former alternative may seem more plausible. If this be the cor- . rect interpretation then the more usual case of the two X’s going to the same pole in the abnormal cyst would be due to their rela- tion to a particular pole of the dividing cell. The exceptional passage into the other cell would be due to one of them getting caught by the constriction, so that it was necessarily detained in the smaller cell. But the more nearly equal sizes of the PREDETERMINATION OF SEX 303 cells in the cases here figured, when an X passes to each pole, may seem rather to favor the other view, namely, that the size of the cell is determined by the chance direction taken by the X chromosome. If cells like these with duplex number of chromosomes should produce functional spermatozoa, and one such spermatozoon should fertilize a normal egg, the number of somatic nuclei would become nine instead of six and the resulting female would have three X chromosomes instead of two. In subsequent genera- tions the number of X’s might be further increased, if, in fact, viable forms could be produced in this way. It does not seem probable, however, that a condition like that described above for the phylloxerans, in which four X’s are present, could have arisen through irregularities of this kind, because the chance that such a rare phenomenon could supplant the normal type seems too small to make such a view possible. Still, the possibility of a tetraploid organism, arising through failure of a spermatogonial or oogonial cytoplasmic division must be conceded, especially in the light of the sudden appearance of tetraploidy in Primula and Oenothera. If it be assumed that some advantages, such as an increase in size, give the new type an advantage, then it might in time obtain an independent footing. A most striking and interesting relation is shown in this tetra- ploid cyst, namely, the chromosomes are only half as large as are those at the corresponding stage of the normal spermatocyte stage, as seen in other cysts of the same testis. This relation might be interpreted to mean either that the original mother-cell of the cyst, having divided (incompletely) one time more than the other mother-cells of the other cysts, never made good the size loss of the chromosomes. If the growth of the chromosomes be directly related to the size of the cell that contains them, owing to the amount of substance available in such a cell, the smaller size of the chromosomes in the tetraploid cyst might find a rea- sonable explanation. ‘Such a conclusion would indicate that the stage reached by a cell at a particular phase is determined by the cytoplasm, rather than by the size of the chromosomes. 304 - T. H. MORGAN THE OMISSION OF SYNAPSIS IN THE PARTHENOGENETIC EGGS OF PHYLLOXERANS AND APHIDS As is well known, the full or diploid number of chromosomes is present in most eggs that develop by means of parthenogenesis. Whether the presence of the full number of chromosomes has in itself anything to do with the phenomenon of parthenogenesis may well be disputed, because in some forms, as in the male bee, the eggs develop without being fertilized with half the number of chromosomes and in artificial parthenogenesis the half number of chromosomes occurs in some forms, at least. In the phyllox- erans and aphids there is a loss of one or two chromosomes from the male-producing egg that develops by parthenogenesis. There is a further question that is important from a descriptive cytological point of view, namely, whether the parthenogenetic eggs omit the synapsis stage and retain in consequence the full number of chromosomes, or whether they pass through such a stage and the chromosomes subsequently separate. In phyllox- erans and aphids the case is quite clear! and I wish to emphasize the ease and certainty with which the problem can be studied in them. In the bearberry aphid, the ovary that is producing sexual eggs (diagram 3) can with certainty be distinguished from the ovary that is going to produce parthenogenetic eggs (diagram 2). In the latter there is no contraction phase of the chromosomes. A prophase of an oogonial division of a parthenogenetic egg is shown in diagram 2,a. Six chromosomes are distinctly seen and the same number is found in the equatorial plate stage shown in diagram 2,6. At the beginning of the growth period, when the chromosomes begin to take the stain again, scattered threads or strands can be made out, as shown in c, which by further con- traction, d, give rise finally to the six rod-like chromosomes. In later stages, when the egg is about ready to leave the ovary and after that time while it is still acquiring yolk, the six chromo- somes can be distinetly seen and easily counted, as shown by most of the eggs in diagram 2, e and f. In the ovary of the sexual in- dividual the chromatin begins to condense into threads at the 1 Morgan, T. H. Proc. Soc. Exp. Biol. and Med., vol. 7, 1910. PREDETERMINATION OF SEX 305 beginning of the growth period, as shown in diagram 3, a, 6, c, and the threads later condense at one side of the nucleus, as shown in d. No details of the process of union of the chromo- somes, that must take place at this time, can with certainty be made out. The figures are as accurately drawn as possible, but Diagram 2 a, prophase of an oogonial division; b, metaphase of an oogonial division with six chromosomes; c, young parthenogenetic ovum just prior to the appearance of the chromosomal filaments shown in d, e, f, young ova for the most part in the lower end of the ovary or else just out of the ovary, showing six chromosomes whenever all of the chromosomes are included in the section. beyond the fact that the chromosomes are condensing, the in- terpretation of the details of the drawings is unsafe. When the chromosomes begin to emerge from the condensed stage, as seen in diagram 3, e, f, g, h, three clumps or rods, or sometimes open MORGAN We 1h 306 ‘pour1oy st Apoq repod oy} ot0joq WINAO Bunock ‘y fo8vys or;dvuds oy} Woy posiotia oAVY LO ‘SUISAIOUIA 918 SOMTOSOTUOIYS ddI} OY} UI S3Jo pUB PUS TOO} UG UL S[foo OATFTAYNA YM A -Moys ‘pryde Arroqivodq oy} JO S[ENPIATpUT [BNxXos IVAO B]OYAM JO UOT}OeS ‘p yoy ur Bao Bunod ‘4 ‘/ ‘a {pua oyyo 4v osvys stsdeuds sisdvuds oyut Surssed wao Sunok But 0 SOTUVAO OU} [O PUS TOTIO4SOd oY) UT SsTTI0 ‘9 ‘gq ‘D WIBIS BI I |! TST] ] I PREDETERMINATION OF SEX 307 rings, appear and in this condition they are carried onto the polar spindle. In the bearberry aphid, the stem-mother produces many male embryos as she gets older. Their presence in the mother serves as an index of the condition of her ovary at this time. An ex- amination of her ovarian eggs at this time failed to show in the chromosomes any process suggestive of synapsis, although no doubt the steps preparatory to the elimination of the two sex chromosome must be taking place at this time. As only two chromosomes are involved it is quite likely that even if they went through a contraction phase independently of the rest of the chromosomes (if such were possible) it would be difficult to recognize such a process. The negative evidence has no special value and is mentioned here only to show that the stages were examined for evidence of synapsis. In the phylloxerans the ovary of the stem-mother continues throughout her life to produce a series of eggs that develop by parthenogenesis. All stages in the development of the eggs can be found in almost any female. There is never in any of them the slightest evidence of a contraction phase, and, since the eggs show exactly the same conditions as do the parthenogenetic eggs of the bearberry aphid, it will not be necessary to repeat here what has already been said. The ovary of a young female that will later reproduce by parthenogenesis is represented in dia- gram 4. The migrant generation of the phylloxerans also produces eggs that will develop by parthenogenesis. In P. caryaecaulis all of the eggs develop at nearly the same time, so that the conditions for the study of parthenogenetic stages are not so favorable as in P. fallax, where the wingless ‘migrants’ produce one egg at a time over a considerable period. In neither species have I seen any evidence of contraction, nor have I seen any other evidence of conjugation of the sex chromosomes, with the exception already noticed where two chromosomes of double size were found in two eggs that would give rise to male embryos later. Here, however, the nucleus was ripe and ready to take part in the for- mation of the polar spindle. The observation only shows that > THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 3 308 T. H. MORGAN the two pairs of conjugated chromosomes have already united before the spindle is formed. Von Baehr describes for Aphis saliceti a synapsis stage in the spermatogenesis in which the chromosomes are contracted at one Diagram 4 Section of ovary of parthenogenetic female of second generation, with large nutritive cells at one end and small egg cells at the other end on the ovary. Diagram 5 Spermatogonial cells in prophase stages to left and one cell in metaphase to right; all cells show five chromosomes. side of the cell. As no further details that identify this stage were made out, one can not be certain that such figures repre- sent the true conjugation stage. I have studied the spermato- PREDETERMINATION OF SEX 309 genesis of several species of aphids in the hope of getting better figures of this stage, but have found none better than those given by von Baehr. In those species where the formation of the spermatozoa is a continuous process in the testis of the adult male it is possible to obtain in the same testis all stages, from the dividing spermatogonial cells to ripe spermatozoa. 1700. 23 Anaphase of the first cleavage; the injured chromatin is lagging behind. x 1700. 352 EFFECTS OF RAYS OF RADIUM ON PROTOPLASM PLATE 3 CHARLES PACKARD VE | 1s ee Ceo ew ae atts > TP 20 Ber (co: Grr By Ls Cages BOG ee CPU) E¥ Cy OR e. ae Qs) nec y & Bo 22 — @ cy — ® pire << “A Ws ate “ = is mat * same .@ 353 _ ys i Sat pene at “ie eee ; } Ai ie, oo. a ee ved nie ‘eo ee ‘i %e a ea a : wl hs yi Sl ye te Se, aoe : . ve Riis ‘ ha ‘ Wha 1 : Pal re AM , ' wi ) | ~ helps 2s . ‘ yom a J : : ¢ ie jek , a 7 vr Shed us f = « i ~ Ai ' Ce ’ 7 i ‘ . Ps . i . jer 1 . —_ - Al i ve Ps : < <— v) ry 2 ae 1 5 F irr, ve + ~—S oY 9 < i 7 -S =] ates ft =~ 2) any 1 eee i oe > | , A® : 4 * i “a z - . i » es x ) 3 D4 ‘ the i it The ; i a - ‘ ie : ae * a; ‘~ in y > . ‘ yee « A ee a Se bP . La ae ' ha apa « 4 SPS a oe Ye r teen aiut oo 7 we sD ‘ oie ~e , fe as ¥. - ’ fi 4 “ . +7 - 7 ; bit by ’ - « : ‘ 2 ” S 4 i] ‘ o 7 py sey, Me : * , 7 : ‘ he ; A ih oa | ‘* THE EFFECTS OF CARBON DIOXIDE ON THE EGGS OF ASCARIS! THEOPHILUS S. PAINTER (Instructor in the Sheffield Scientific School) From the Osborn Zoological Laboratory, Yale, University FIFTEEN TEXT FIGURES AND THREE PLATES The material upon which the present study is made was placed in carbon dioxide on July 14, and removed on October 9, of the same year (1913). Professor Boveri, in an attempt to keep the eggs of Ascaris for a long time without allowing them to undergo development, had a number of smears? from one female placed in a stoppered glass jar. The air of the jar was then re- placed by passing a current of carbon dioxide through it for an hour and a half and, after sealing it carefully to prevent the escape of the gas, it was placed in a basement room in the Zo6- logical Institute at Wiirzburg. At the time the eggs were placed in the gas, no ill effects were anticipated, consequently, no control smears were preserved to show the exact nuclear condition in which the eggs were at that time; and, after they were removed from the gas, they were placed directly on ice where they remained until used. When a few smears were allowed to undergo full development later, it was found that only part of the worms were normal. Professor Boveri called my attention to the fact and placed the material at my disposal with the suggestion that I determine the cause of the abnormal development which part of the worms showed. Although the material for this study was obtained in Wiirz- burg, the greater part of the work has been done since my return 1 A preliminary note on this subject has been published by the author (1914). 2 The smears were made on ordinary microscope slides according to Boveri’s well known method. 390 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 19, NO. 3 356 THEOPHILUS S. PAINTER to America. I take this occasion to express my thanks to Professor Boveri for suggesting the problem and for placing the necessary material at my disposal. In preserving the eggs, a mixture of 4 parts 95 per cent alcohol and 1 part glacial acetic acid was used. They were stained in toto and mounted in glycerine. DESCRIPTIVE Among the embryos which had been allowed to undergo full development, one finds perfectly normal specimens; specimens in which only one end of the body is developed; and lastly, totally disorganized embryos. The problem was _ primarily to determine the causes which produced the abnormalities observed, but as the work went forward a number of other questions of general interest came up which will be touched upon in the following paper.’ A drawing of a normal worm, which developed from an egg exposed for three months to carbon dioxide, is given in figure A. A blunt anterior end with the pharynx and the pointed posterior end may be seen, and, in addition, towards the posterior part of the body the large deeply-staining nuclei of the primordial germ cells. Typically there are only two of the latter but occasionally more occur; in one case six were present. The abnormal embryos are of two general types. One of these has the posterior end of the body fully developed while the anterior end is disorganized. The second type is characterized by the absence of organization in its blastomeres. The first type of embryos is shown in figure B. This occurs in roughly 33 per cent of the embryos (in 54 cases out of 165 examined for the point). The pointed posterior end is clearly seen together with the primordial germ cells, but there is much variation in the degree to which the posterior part of the body is developed. We 3 A brief description of the normal development of Ascaris is given on page 367. Any one not familiar with the cleavage, or with the nomenclature, in this worm, will find it helpful to read this over together with a glance at the schematic dia- grams given. The nomenclature of Boveri has been used throughout the present work. EFFECTS OF CARBON DIOXIDE ON EGGS 357 Text figures A to C may, in rare cases, find fully seven-eighths of the worm normally formed, or we may find nothing but a pointed stump; figure B gives a fairly typical case. | An example of the second type of embryo is given in figure C. Such individuals are found in about 40 per cent of the cases (66 cases in 165). The greatest variation is noted in the appearance 308 THEOPHILUS S. PAINTER of such embryos. Typically they consist of a mass of cells, among which one may distinguish the primordial germ cells, but no organization exists and quite frequently it is evident that no cleavage cavity was ever formed in the embryo, consequently, that gastrulation had not taken place. Eggs preserved as they were taken from the ice chest, where they had been since their removal from the CO:, showed a slight amount of development. All of them had divided at least once, and something less than half (227 out of 505 eggs counted) had reached the 3-cell stage. Occasionally, in such a preparation, a 4-cell stage is found. An examination of these 3-cell stages (fig. 1) shows that the S, blastomere has divided to form the A and B eells, while the P,; blastomere is in a ‘resting stage.’ The nuclear condition of the A and B cells is normal: as may be seen by the figure, waste chromatin occurs in one or both of the cells showing that the diminution process has taken place, but the two cells do not always lie pressed against each other as is normally the case (compare fig. 1 with text fig. F). In 50 eggs out of 78, examined at random for this point, the A and B blastomeres were separated, in extreme cases the two cells lying on opposite sides of the P; cell. Among the eggs in the 2-cell stage, 101 out of 278 cases showed the S, cell dividing, with the chromosomes in the equatorial plate phase. The remainder showed resting nuclei in both the S, and P,; blastomeres. A close examination of the equatorial plates in the dividing cells shows an abnormal condition of the chromatin (figs. 2a to 2d). Figures 2b, 2c, and 2d are drawn at a higher magnification. In practically every case (96 out of 101 eggs examined for the point) the chromosomes were found fused together. This fusion seems to affect the ends principally (figs. 2a, 2d) leaving the middle portion free, but in extreme cases even the middle parts are involved and all four chromosomes are clumped together into one mass (figs. 2b and 2c). One very constant feature of this fusion is the formation of what appear to be vacuoles in the fused ends. It is also to be noted that when only the ends of the chromosomes are involved the middle EFFECTS OF CARBON DIOXIDE ON EGGS 359 portions have the normal clumped or lumped appearance which precedes diminution (compare fig. 2d with fig. 9). In rare cases, both the §; and the P, cells were dividing at the same time (fig. 3). When this occurred, the chromatin of the S; blastomere showed the fused condition, while the chromosomes of the P; cell were normal. In the 2-cell stages with the nuclei in the resting phase, no departures from the normal could be distinguished. When the eggs are allowed to develop a short time before they are preserved, the P; blastomere begins to divide. The elon- gated chromosomes, so characteristic of the primordial germ cells, are always found and aside from the axis of division, the cell appears normal. Here and there a tendency for the four chro- mosomes to break up has been noted (text fig. J and L), but this is not to be regarded, I think, as an effect of the CO.. The point will be taken up in detail later under the heading ‘Anomalies.’ If the eggs are allowed to develop further until half are in the 4-cell stage, a variety of conditions are found in the 4-, 3- and 2- cell stages. Needless to say, these various conditions arise out of the 2- and 3-cell stages described above. | Among the 4-cell stages we find three different types of em- bryos. One of these (fig. 5) is perfectly normal, both in the position of the blastomeres and in their nuclear conditions. a @e i ~ 3 Pe a da we Poe S « Ls pl eal @e. es e — ts 9 ¢_° = Chart « ee atts Ns

4 381 PLATE 2 EXPLANATION OF FIGURES 6 A 3-cell stage showing the unequal distribution of the chromatin between the A and B cells. 7 Showing the tetraster in the 8; blastomere. 8 Showing the tetraster in the S; blastomere. 9 Showing normal 2-cell division after treatment with COs 10 Showing the A and B blastomeres dividing in abnormal planes. 11 Showing the result of the division when the A and B cell occupy abnormal positions. EFFECTS OF CARBON DIOXIDE ON EGGS THEOPHILUS S. PAINTER eet Ry Fn oe & Ps 6 BHESES » ‘eae Sats Se weer,” oF ee? € @e & @ ¢.e8 ew tars? % ave xt _e cere ‘ e as Se = ee res P es 6 S . oS, x o Se tf Pi. ee os -s 4 g pany ge @ *é& CF a ee” 4% ie? es * a Ce “fy we EMSt 8 e B A y Dy fay tal ee ee | ofl a % A Fee > ‘? * rg & we ° e 4 s . $ cw? a e a Paes) ee? @* ms 10 383 PLATE 2 EMSt i fh 4 Pp ; = ~~ > > 2 % = ~ » ~ Va * Oe & r a 4 ’ | e ia & g os $ a < ~~ = =f Ret Sag ‘= oe o™ ¥ “7 aa a = 7 w Ms <4 = cd = - Pode g = Ped oS - “4 uate - PLATE 3 EXPLANATION OF FIGURES 12 Showing the effect when the A and B cell divide in abnormal planes; the ectodermal cells number 8 in this egg. 13 Aslightly later stage of the same; note that the MSt cell does not lie in the median plane of the embryo. 14 An egg showing the effect of the unequal division of the chromatin between the A and B blastomeres. 15 An egg showing the effect of the tetraster formation in the §; cell. 16 Showing the P. and EMSt cell fusing; note particularly the protoplasmic ball projecting from the fusing cells. 17 Showing the later effect of the fusion of the P. and EMSt cells; note that diminution is taking place in this cell. 384 EFFECTS OF CARBON DIOXIDE ON EGGS PLATE 3 THEOPHILUS 8. PAINTER MSt e © : . $ e @-e are E ae hd ¢ aa s *% ( } e ge Te & .e s * af 4 se ¢ ot e 3 bi rs * ov e ea «* ea, ee yy fs a . ~ s PRS ns Ae 2 a ' @ Pp 28 6 Sa oor a MSt oes ee bis : fae g® - age 9 © 4 Ae ss ee ae Po wy e>% hs oe e8 pera J ghd Pik &% 9 ¥ Py. a > &S He > * 3 * = 2 . EMSt 14 re 15 >~s aus Fo i . chs = jit f ; *% ,* 2'@ @ é % - , & ¥*> pha a S20 | ais ‘ % Dak ge ek Seg rin oP nye a Ve Chee. 8 terns Pet ‘2 oe fi ee wg xt ‘i ee e & & Pe gen Z & & 4 ee oy 17 385 ; i ; . 4 ‘ = rhe 0 j 4 a . — 7 ot pom ir) ‘ i ‘ 0 d ? ae vey ' * i is) va) oS . ‘ - wy \ ‘ ch > > ‘ i } . Aes . ; . “ i ‘ ri . J . Y VARIATION AND INHERITANCE IN ABNORMALITIES OCCURRING AFTER CONJUGATION IN PARAMECIUM CAUDATUM RUTH J. STOCKING From the Zoédlogical Laboratory of the Johns Hopkins University! TWENTY FIGURES CONTENTS Palin RO GUCTIOMER corks aris Sicha sys SE 0 oaths. dc ee eee eae ee 387 OV Te tit Gl Sst netics Scrat sce Megat ac. waite Gpvaha< food tna eee ae Ee ree ee es 390 hiesE wpermenvall Cultures: sc sc 5.5: > ac eo «lente eee ee eee eels sees ake 391 Vey pes ot Abnormal iiacese:..... 4.655) odes ob BE Gate A RA Anna aE 397 Vee Nia tunevo titer Ab monrmallitlese . cit < <5 ciate aa eee ae arn ects 407 VI. The Abnormalities as Hereditary Characters: Variation, Inheritance, and Selection.................... etn cute 412 IDpsqjorsnruaaenmas) Tin SVG son cokedodenscs coSelesac seen dobeeeude 414 jARelation! to) Biparental Imheritamce:.2.......5..4.08455.....-. 440 Vitiea Summanzands Dis cissionnote Results: see Eee eee een I 444 AVATAR era TUE Se cr aaie Re 5.528 cit 01d ol Rane Na a ee Peay 449 Ey INTRODUCTION The fact that there are two kinds of teratological variation, one caused by gametic constitution and therefore heritable, the other occasioned by the action of environmental conditions, and not heritable, has been recognized for some years. There was for a time an endeavor on the part of many investigators to show that all abnormalities were due to conditions of nurture; but the discovery of the strict mendelian inheritance of a large number of malformations has overthrown that theory. In many of the metazoa both kinds of teratological variation, gametic and en- vironmental, have been studied, and their causes and behavior partly determined. In the protozoa the abnormalities caused by the action of external factors have received by far the great- 1 Part of the work here reported was completed while the author held the Alice Freeman Palmer Research Fellowship of Wellesley College. 387 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 NOVEMBER, 1915 388 RUTH J. STOCKING est amount of attention, heritable abnormalities having been but little studied. This paper presents an investigation of herit- able abnormalities in Paramecium. The principal points examined in this presentation are the following: The origin and nature of the abnormalities; their relation to conjugation; their inheritance and variation in uni- parental reproduction, with relation to the present day prob- lems of ‘pure line’ work; how precisely inheritance occurs; whether there are variations of kind and degree of abnormality; whether such variations are themselves inherited; whether by selection abnormal stocks multiplying asexually can be altered in their hereditary characteristics or differentiated into two or more hereditarily diverse stocks; their relation to biparental inherit- ance; their relation to survival. Previous work on the abnormalities found in protozoa has dealt mainly, as before remarked, with the results of environ- mental action. Jollos (13) subjected a race of Paramecium caudatum to changes of temperature and obtained by this method differences in size, as did Hertwig (08), Popoff (08, ’09), and Rautman (’09). Jollos found that these changes are transi- tory and soon disappear as the paramecia become adapted to their new conditions. In a short preliminary note, which does not present the evidence for his conclusions, he states that in one case he obtained a permanent change. He subjected a race of Paramecium to a high temperature and obtained a race which was permanently small at high, normal, and low tempera- tures. Popoff (’09) was able to produce by experimental means two abnormal races of Stentor, one giant and the other dwarf. He centrifuged a dividing Stentor, and so caused an unequal distribution of the nuclear material. The daughter cell which received the smaller part was one-fourth the size of the cell which received the larger part. These two multipled nor- mally for about a week and during that time retained their abnormal sizes. The cultures were then lost. He obtained another giant race of Stentor by suddenly cooling a dividing individual. The division did not take place and the animal reorganized into a single individual which grew and subsequently INHERITANCE IN ABNORMALITIES 389 divided, forming a giant race which was kept for about forty- five days, and during that time bred true. Lewin (’10) obtained abnormally-nucleate races of Paramecium by cutting an indi- vidual through the macronucleus. The production of form abnor- malities in Paramecium by cutting and other experimental means has been studied by Calkins (11), Peebles (12), Balbiani (’93), McClendon (’09), and Jennings (’08). . They all found that such experimentally produced abnormalities are mechanically handed on to one daughter cell at each division for a longer or shorter time, But such abnormal forms are gradually remodelled dur- ing successive generations, or die, and their normal sisters show no tendency to produce abnormal progeny. The teratological variations that arise spontaneously in a culture multiplying vegetatively have been studied extensively by Jennings ('08). With one exception he found that all such forms either die very soon or give rise to a race of normals. In one case he did find that the abnormal individual gave rise to a race of abnor- mals, the deformity being such as to prevent the daughter cells from separating after division, thus forming what have heen called ‘double monsters.’ In all other cases the abnormality was not a race character but an individual character and was not inherited. During the course of a year’s work at the Johns Hopkins University I followed the history of several abnormal forms which had arisen in the cultures of normals being carried on by other workers in the laboratory. In no case did these forms give rise to a race of abnormals; in one case a race of normals resulted; in all other cases death occurred either before any divisions had taken place or after one or two irregular divi- sions. In rare cases, Jennings (713) found, abnormalities may arise in the members of split pairs; that is, in the members of pairs separated before conjugation has been completed. But these also never gave rise to a race of abnormals, either dying very soon or becoming entirely normal. Teratological variations arising soon after conjugation have been described in a few cases. Simpson (’01) observed four daughter cells of a normal exconjugant Paramecium, three of which were normal while the fourth was posteriorly split. This 390 RUTH J. STOCKING animal divided eight times and its history was similar to that given above for the experimentally produced abnormalities in Paramecium. The abnormal character was handed on to one cell at each division, and the abnormal animal produced by the eighth division died. All of the sister cells and their progeny were normal. The only other observation known to me on the abnormalities arising after conjugation is made by Jennings in his 1913 paper quoted above. While he found only extremely rare and transitory abnormalities among his ‘split pairs’ and ‘free’ individuals, he found a large proportion of malformations among his exconjugants and their progeny. He describes their different types, characteristics, and constant and continued appearance throughout the course of his experiments; and adds, ‘“A precise study is greatly needed, as to the minute character- istics of these abnormalities, their heritability, their experi- mental cause, and their cytological basis.’”’? It was at his sug- gestion that a series of experiments was started for the purpose of studying these problems, and under his direction that these experiments have been carried through. I wish to express here my most sincere thanks for the constant help he has given during the entire course of the work. Il. METHODS In all of the cultures on which this work is based the method of handling, cultivating, and recording was identical with that described by Jennings (13). Conjugation was first induced; after the pairs had separated the two members were isolated and cultivated separately in ze per cent Horlick’s malted milk (Peebles, ’12). Each pair was numbered, and the two members of a pair designated a and b; the races which arose from these exconjugants were called after them in the same way, 2a, 2b, 3a, and so on. Beside the slide cultures, mass cultures were also kept for each race. These were kept in bottles and each bottle was labelled correspondingly, 2a, 2b, 3a, and so on. The method of recording described by Jennings was ‘supplemented in my work by drawings of the abnormal forms. In a few cases INHERITANCE IN ABNORMALITIES 391 this was impossible because of the rapid movements of the ani- mal. For about two weeks after conjugation camera lucida drawings were made of the abnormal forms. These were often less active than the normal forms and after being in the same drop of culture fluid for two days all of the individuals were more or less sluggish. So with a little patience camera lucida drawings could be obtained of almost all the abnormals. When the number to be examined and drawn became very large, how- ever, there was no time for that procedure and free hand draw- ings were made in the majority of cases, camera lucida only in such cases as seemed of more than ordinary interest. The slide cultures were examined every other day during the cool weather, every day when it became very warm, the individuals counted, the abnormals drawn, selection made of those by which the slide cultures should be carried on, and the others put into the bottle mass-culture corresponding to that race. The individuals selected for slide culture were then washed in the milk solution, and put into fresh fluid on clean slides. III. EXPERIMENTAL CULTURES The material presented in this paper is derived from three experimental cultures carried from the fall of 1912 to the spring of 1914. The first experiment, conducted from December 2, “1912, to June 11, 1913, was carried out with exconjugant mem- bers of a wild population which doubtless included members of many diverse stocks. The chief aim of this first experiment was to ascertain whether or not abnormalities were ever inherited, and to what extent. From November 25, 1913 to December 22, 1913, a second experiment was conducted on a group of animals all descended from one individual, constituting there- fore a ‘pure line’ or ‘clone.’ It was concerned mainly with deter- mining the effect of conjugation among the members of a single clone, particularly as to the production of abnormalities; this was brought out by a comparison of exconjugants with mem- bers of split pairs belonging to the same clone. In the first experiment some work on the effect of selection was attempted; but the third experiment, carried out on a wild population from 392 RUTH J. STOCKING January 14, 1914, to April 1, 1914, was concerned wholly with that problem. In it an attempt was made to ascertain the efficacy of selection as a means of breaking a single clone into two or more lines differing hereditarily in amount or kind of abnormality. In all three of the experiments, in two of which very large numbers of exconjugants and their descendants were studied, a large proportion of the resulting races were abnormal, just as was the case in the exconjugants studied by Jennings (’13). In Experiment 1, with a wild population, the proportion of abnor- mals was 36 per cent of the entire 262 exconjugants. In Experi- ment 2, with the members of a clone, the proportion of abnor- mals was much higher, being 81 per cent of the whole 200. In Experiment 3 the number of exconjugants studied was small; the abnormal races formed 43 per cent of the total 28. In the 54 members of the 27 split pairs of Experiment 2, twenty of which were cultivated for 19 days, and thirty-four for 7 days, no abnormals at all appeared. Whenever sets of individuals are isolated without conjugation and cultivated in the same way, no such proportion of abnormals are observed; indeed as a rule no abnormals whatever are obtained. We shall return later to the relation of the abnormalities to conjugation (page 412). The exconjugants from large cultures of paramecia may be divided, on the basis of their subsequent history, into a number of diverse classes, which are summarized for our three cultures in table 1. 1) A few pairs, in some cultures, never separate, but die while united. 2) A considerable number, in some cases, die within 24 hours after separation. 3) Others live after separation—often for a long time—but never divide. 4) A fourth group divide, but produce individuals that are in some way and to some degree abnormal. 5) Finally, a certain pro- portion propagate normally after conjugation, giving rise to typical progeny. The proportions of these different classes are shown in table 1. } The single pair that did not separate lived united for five days. The exconjugants that never divide after separation form a large proportion of the abnormal individuals, rising to 28 per cent of INHERITANCE IN ABNORMALITIES 393 TABLE 1 Diverse types of exconjugants, their number and proportion in the three experimental cultures 1) NEVER 2) DIED 3) NEVER 4) PROGENY 5) PROGENY SEPARATED FIRST DAY DIVIDED ABNORMAL NORMAL EXPERI- RENN fae TOTAL No.| Percent |No.| Per cent |No.| Fercent |No.| Per cent ee Per cent 1 0 25 10.0 74 28 .0 21 8.0 |142 54.0 262 2 2 1.0 3 Wh) 27 14.0) |183 66.0 319) fh 5) 200 3 0 0 0 2, 43 .0 16 57.0 28 Aonells|| 0.4 28 Sete LO 20.6 |166 33.9 |193 39.4 490 all in Experiment 1. These often live for some time, the length of life varying greatly. The length of life in days for the 101 exconjugants that never divided was as follows: Henechyonlitesim (davis sacra sales aeiecia dee ates elec ae era til * <8} 5 ahem Mee LS 15 ile total Number of exconjugants: EEO CLIN CMU Alera ete aoe ee fata'e sols. o 3 ONO AGH Oro mm On no) melon inset JEP GOSS NEON) CHS S ones acieieo.do cate eae 2) (Si Rare 2 el ORO Rn On BOM E24 TRON Bee nile pan bil aera Crochet A GO IP Be Il GS ily Aol All of the animals of Experiment 2 which never divided were dead by the thirteenth day after conjugation; in Experiment 1 nine such individuals lived for some time after this. Those that live for a number of days often change a great deal in size and shape; many of them become immensely large. Twenty- eight of these individuals were measured; for their lengths and diameters on successive days after conjugation see p. 474. The diversities of size and shape among these individual® is illustrated in figure 1, which shows nine of the individuals of Experiment 1 which never divided, all camera lucida drawings to the same seale. The usual size of Paramecium caudatum on the same scale is shown in figure 5. ’ The largest single individual Paramecium that I have ever seen was the first individual listed below for Experiment 1; it measured 520 by 150 microns, as against a usual length of 150 to 200 microns. The smallest individual that never divided was likewise in Experiment 1; it measured 128 by 10 microns. It is evident that there was much greater variability in size, shape, 394 RUTH J. STOCKING and length of life, in these individuals of Experiment 1, than in those of Experiment 2. It appears probable that this is con- Experiment 1 MEASUREMENTS ON “. Seventh day Ninth day Eleventh day Thirteenth day Fifteenth day 520 x 150 340 x 130 320 x 100 220 x 55 450 x 120 420 x 150 420 x 92 260 x 80 400 x 140 420 x 120 220 x 90 400 x 135 390 x 160 350 x 110 350 x 80 340 x 110 350 x 130 340 x 100 340 x 70 220 x 105 330 x 100 240 x 150 300 x 120 290 x 140 240 x 105 290 x 100 300 x 170 280 x, 140 210 x 115 200 x 20 ZOE 20 190 x 160 165ex 920 15; 0pm) 0) 129 x 10 200 x 35 250 x 60 Ave. 315 x 102 288 x 91 SIA le DOTS 220 x 90 Experiment 2 MEASUREMENTS ON Third day Fifth day Seventh day 240 x 75 240 x 55 DANO GO, 210 x 65 195 x 55 145 x 35 150 x 20 190 x 45 U7Ass 3% 3b) 145 x 35 Aves 209 O2 seer 145 x 35 150 x 20 INHERITANCE IN ABNORMALITIES 395 nected with the fact that the former culture consisted of wild individuals, probably of diverse stocks, while the latter consisted of members of a single clone. This matter will be taken up later. Ovid é ons Fig. 1 Nine of the individuals of Experiment 1 which never divided, show- ing the diversity of form and size among this class. In this figure and in figures 2, 3, and 4, the animals were all drawn by means of the camera lucida, and are all to the same scale. (xX 100) As a rule there appeared to be a tendency for these larg individuals to gradually decrease in size (as shown by the meas- @ urements above given). Figures 2, 3, and 4 show the changes in form and size in certain cases. The forms were frequently nearly normal; sometimes very abnormal, as the figures show. These individuals that never divided were usually black and granular, and often changed shape when transferred to fresh culture fluid, becoming swollen at the anterior end;-they later resumed their original form. A cytological study of these ani- mals might be of interest in connection with the ‘Kern-Plasm Relation’ theory, and with the mechanics of division. Such 396 RUTH J. STOCKING 4 Tig. 2. One of the individuals (15a) of Experiment 1 which never divided, showing the decrease of size and change of form it underwent from December 9 to December 15, inclusive. (x 100) Fig. 3 Two of the individuals of Experiment 1 which never divided, showing the changes*of form and size they underwent. The two figures at the left show 16a on December 9 and 11; the other two figures show 110a on December 9 and 13. (X 100) Fig. 4 Three of the individuals of Experiment 1 which never divided The first two figures show 42a on December 9 and 13; the second two figures show 89b on December 9 and 13; the last two, 19b on December 9 and 11. (x 100) INHERITANCE IN ABNORMALITIES 397 large individuals, never dividing, do not appear after conjuga- tion in all cultures; some later attempts to obtain them for study have not been successful. IV. TYPES OF ABNORMAL RACES Of chief interest for our purposes are the exconjugants which divided, but produced progeny some of which were deformed. The lines of descent to which these deformed individuals belong I shall call abnormal lines or races; it is they that provide the material for the study of heredity, variation, and selection here set forth. The abnormal lines differ greatly in the proportion of abnormal individuals produced and in their later history. Not all the individuals produced are abnormal; and not all the abnormal races remain abnormal indefinitely. Reserving many of the details for our section on the results of selection, we may divide these abnormal races into three classes: 1. In the first class are those races which ultimately became normal. There were 39 races of this class. The proportion of abnormal individuals produced ranges in the different races from one per cent to 41 per cent. In these races the abnormal indi- viduals gradually disappear, till finally only normals are pro- duced. Some of the abnormals of this class may be compared with the experimentally deformed animals; and with those which arise rarely in cultures that have not conjugated. They always form a very small proportion of their race, and either die out or produce normal daughter cells. Others however are not strictly comparable with the transient abnormalities described by other workers. They form a considerable and constant pro- portion of the individuals of their race; are continually produced for several generations; and are often descended from both nor- mal and abnormal sister cells. But through the action of natu- ral conditions and in some cases the selection of the normal indi- viduals, the abnormals become gradually eliminated and the race entirely normal. Data on these 39 races of the three experiments is given in table 2. The 39 abnormal races which became entirely normal EXPERI- LENGTH OF LIFE TABLE 2 NUMBER OF NUMBER OF MENT eee IN DAYS GENERATIONS NORMALS 1 96a 15 10 314 30a 17 10 32 14a 13 11 38 102a 15 a ie 63a 15 iG 13 2 84b 9 5 16 49b 9 5 14 53a 11 6 25 32a 11 8 35 63b 11 6 45 34a il 6 14 57a 11 5 14 29a 11 9 33 50b iil 5 12 30a 11 7 39 40a 9 4 10 15a 11 7 IG 73a 11 9 17 56a 11 9 57 47a 11 5 16 57b 11 9 31 92b 22 i 28 88a 13 6 13 35b 11 8 47 12b 11 3 3 26a 9 3 3 34b 11 4 9 54a 11 5 10 55b 9 4 10 95b 11 3 9 99b 9 2 Y 55a 15 3 4 96b 9 + 10 85b 9 3 it 38a 11 4 10 59b 13 7 9 16b 15 6 16 3 52b 17 9 250 538b 21 8 212 Range... 9 to 22 2 to LL 2 to 314 Average. 12 6 38 NUMBER OF ABNORMALS ow w — ra RrPOOrON Fr PPP WRrRRr oP WOR WHRENYNNTINN DOH NH HE — “Ib IL yefoy 15} 4 PER CENT ABNORMAL — 15 28 = 1 to 41 10 INHERITANCE IN ABNORMALITIES 399 Figures 5 and 6 give pedigrees of two of these transiently abnormal races, numbers 14a and 102a, which are fairly typical of this group. Race 14a (fig. 5) was kept for thirteen days. During that time it divided 17 times, giving an average of 1.13 divisions a day. It showed no abnormalities until December 11 (nine days after conjugation), when the three individuals which had arisen from the apparently normal animals selected on December 9 were, one large and very abnormal, one small and abnormal, and the third small and almost normal in appear- ance. The large one died after having divided once. The two small ones had divided to form four normal individuals on December 13, and on December 15 had given rise to eight per- fectly normal animals which were then discarded. Race 102a (fig..6) divided very slowly at first, averaging 0.5 divisions a day. The average division a day for the whole time the race was kept was 0.8, the animals dividing much more rapidly toward the last. This exconjugant showed no signs of being abnormal until December 11, when the three forms pres- ent were very abnormal, two being doubles and one a very much swollen individual. One of the double forms moved in a circle so swiftly that it could not be drawn. Both of the double forms died without dividing; the large swollen one gave rise to a perfectly normal race which was discarded on December Ls 2. In the second class of abnormal races are those which persistently produced abnormal individuals throughout their history. There were 9 such races in Experiment 1 and 88 in Experiment 2. They differed greatly in the proportions of abnormals produced, varying from races 100 per cent abnormal to those but 3 per cent abnormal. The pertinent data for these 97 persistently abnormal races is given in table 3. In many of these persistently abnormal races individuals appeared which were normal in form. But these if propagated eventually produced abnormals. Further details as to this will be given in our section on the effects of selection. Figures 7 and 8 give pedigrees of two of the short-lived races of this group, 39b and 40b, which show the typically persistent INHERITANCE IN ABNORMALITIES 401 ee ae ret Cee — | ae oy ABNORMAL ( ABNORMAL DEAD Sa" ( oe DEAD Fig. 6 Pedigree of 102a, one of the abnormal races which became entirely normal. On some dates the abnormal forms could not be dr awn; they are simply designated as ‘abnormal.’ 402 RUTH J. STOCKING TABLE 3 The 97 races which remained persistently abnormal EXPERI- RACE LENGTH OF LIFE NUMBER OF NUMBER OF NUMBER OF PER CENT MENT IN DAYS GENERATIONS NORMALS ABNORMALS ABNORMAL 1 2a, 15 3 3 1 25 16b 9 5 18 2 10 25b 9 2 0 2 100 39b 25 9 14 12 46 40b 17 8 10 8 44 47a 9 3 0 5 100 9la 15 2 0 2 100 106a 9 2 0 2 100 C 191 303 2027 2683 57 2 2b 27 11 55 19 26 6a 11 4 5 4 44 7b 15 6 + 9 69 8a 17 5 6 7 54 66b 11 6 4 13 76 8b 17 7 20 15 43 9a, 11 4 2 6 75 99b 11 2 2 11 30 lla 21 10 13 28 68 1lb 21 8 21 14 40 14b 15 6 17 8 32 16a 23 G 25 8 24 18b 11 6 18 5) 22 19a 17 6 21 16 43 19b 23 6 10 16 66 20a, 27 19 224 187 45 20b 19 5 11 10 48 21a 21 10 2) 17 40 21b 13 4 2 5 71 22a 13 4 6 5) 33 22b 11 4 6 6 50 23a 25 8 21 5 19 23b 25 8 33 25 43 2Aa, 9 2 2 1 30 24b 11 6 10 + 29 25a 23 a 13 10 43 25b 11 4 a 4 36 27a, 11 4 + 5) 56 27b 27 27 303 556 65 28a 15 5 19 9 02 28b 13 4 7 8 53 29b 19 8 26 18 41 INHERITANCE IN ABNORMALITIES TABLE 3—Continued . 403 EXPERI- ACE MENT a ——— 30b 3la 31b 32b 33a 33b 35a 39a “39b 43a 43b 44a 44b 46b 48a, 48b 54b 56b 58a 58b 59a 6la 62b 63a 65a 67a 67b 68a 68b 69b 70a 70b 71a (2a, 73b 74a 74b 75a 75b 76a 76b 79a 82a THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. LENGTH OF LIFE - IN DAYS — 19 13 ial 1 21 rb al 13 23 27 11 1 13 ra 27 23 13 9 17 ral 11 13 rn rb He 13 21 19 u 17 21 Dy) 13 15 rb 17 u in 1 a 13 27 9 NUMBER OF GENERATIONS NUMBER OF NORMALS — SEN FE MNOCAORaNAHaANY Ooo _ bo SEER TAWONANOARNONNMDRAONKRMODwWoORAY 36 w bo —_— [—) TS 189) fore) (3b) [) mee bh Nw ow Nowo re eS De Ct S Co) Gr Bm =7 60 i) le.) ee — SCOoOwoananwn _ ee) won NUMBER OF ABNORMALS re [——* top) nN bo No eK Wondna — w —y FP PRrAOON DAY BR BR O}OWO — 49 69 38 25 29 50 73 57 36 4] 32 20 50 jl 59 24 42 19, wo. 4 PER CENT ABNORMAL 404 | RUTH J. STOCKING TABLE 3—Continued EXPERI- RACE LENGTH OF LIFE NUMBER OF NUMBER ‘OF NUMBER OF PER CENT MENT IN DAYS GENERATIONS NORMALS ABNORMALS ABNORMAL 2 82b 9 4 4 8 67 84a 27 if 15 20 57 89a 11 5 10 66 37 89b ail 7 45 22 33 90a 19 4 11 4 27 90b 21 4 8 4 33 92a iif 3 7 5 42 94a 23 a 12 De 66 94b 13 3 7 5 42 96a 23 a 18 25 58 98a 15 5 8 2 20 99a 13 at 5 3 37 100b 19 3 6 5 45 Range... 9 to 191 2 to 303 0 to 2027 1 to 2683 3 to 100 Average. 18 9 41 47 53 appearance of the character in all their lines. Race 39b (fig. 7) divided seven times in eight days, forming only normal indi- viduals until December 11, when six of the eight individuals present were very much misshapen. After this date only abnor- mals were produced by this race, which died out on December 27. There was a very large proportion of double forms among the individuals of this race, eleven of the twenty-five abnormals produced being of this character. Race 40b (fig. 8) divided regularly to form only normals up to December 11, when three out of the five then present were abnormal. The two normals died out and the race was carried on through the abnormal individuals. These produced only abnormal progeny and the race died out on December 19. Race C, described on pages 414 to 420, is a typical long-lived race of this group. 3. As the third group of abnormal races we may distinguish those which did not become entirely normal, but from which normal races were obtained by continued selection. This group will be dealt with in our section on selection. INHERITANCE IN ABNORMALITIES 405 = — — a aes a. ABNOR MAL ABNOR a | ae ABNORMAL, ABNOR “s ABNORMAL 2 ae a ABNORMAL i i JEM : DEAD Q 7 yD 705 Fig. 7 Part of the pedigree of 39b, one of the short-lived races which re- mained abnormal throughout its life. The broken connecting lines designate the omission at that point of two entirely normal generations. DEAD DEAD Fig. 8 Pedigree of 40b, out its life. RUTH J. STOCKING ABNORMAL ABNORMAL aaa DEAD DEAD DEAD DEAL another short-lived race which was abnormal through- INHERITANCE IN ABNORMALITIES 407 V. NATURE OF THE ABNORMALITIES Descriptively considered, the abnormalities shown by the indi- viduals of the abnormal] races are of many diverse sorts. I have been able to distinguish 40 different types of abnormals. These are shown in figure 9, and their frequenties in the three experi- ments are given in table 4. Most of the races of Experiment 1 _ developed one or more representatives of a large number of these types; some of the races had representatives of all but two or three. The high frequencies of so many of these types in Experiment 1, as given in table 4 shows how great the diver- sity of form is among the individuals of that group. In experi- mentally produced abnormalities reported by other workers on other animals, this diversity of type is always shown; as Mall (08) says, there is never found any precise type of abnormality characteristic of certain conditions, but only a general character, a tendency to abnormality of many different types. But some races of my experiments were more or less characterized by a few types, which I have called their predominant types. The best examples of this predominance of type are race C of Experi- ment 1, and races 56a, 56b, and 59b of Experiment 3. . During the early part of their histories these races showed considerable diversity of abnormal type; but this diversity was rapidly re- duced and the abnormals became limited to a few types in all four of these races. During its whole history race C produced in its observed cultures 2683 abnormals; 1331 or 50 per cent of these were of type 33; 945 or 35 per cent were of type 27. The predominant type of the other three races was the same, number 27 of figure 9. Race 56a produced 1428 abnormals; 79 per cent of them were of that type; 59 per cent of the 780 abnormals of 56b were of type 27; and 50 per cent of the 2096 abnormals of race 59b were of the same type. In the abnormals of Experiment 2 there was considerably less diversity of type than in the abnormals of Experiment 1. Only 30 of the different types were observed in Experiment 2; most of these appeared only a small number of times, 90 per cent of the abnormals being of five types (11, 26, 27, 28, 30). Table 4 gives 408 RUTH J. STOCKING ay Hp 4 O ag & byyd 28 29 PUGCKAG Le 32 39 40 QO 9 DP Fig. 9 The 40 types of abnormals observed in the three experiments. In some cases two or more examples of one type are given in order to convey some idea of the variation within a type. The types which appeared in the three experiments and their number and proportion are given, in table 4. INHERITANCE IN ABNORMALITIES 409 the numbers of the thirty types that appeared among the abnor- mals of this experiment and their frequencies. Likewise in the abnormals of Experiment 3 there was less diversity of type than in those of Experiment 1. Table 4 gives the numbers of the 27 types observed and their frequencies; 91 per cent of all the abnormals of Experiment 3 were of five types (22) 265627, 305,39) The lessened diversity of type observed in the experiment with the members of a pure line (Experiment 2) as compared with the members of a wild culture (Experiment 1) is just what we might expect from the constitutions of the two groups. The wild culture probably contained many very diverse stocks; the conjugants probably were widely different in their gametic con- stitutions; so that there is the greatest possible opportunity for variation among their progeny. But all the members of the clone used in Experiment 2, by the theory of the pure line, have identical gametic constitutions. This may be and probably is highly heterozygous; but the diversity possible to the progeny of the different conjugants is limited by this heterozygosity. In Experiment 1 the diversity is limited only to the characters possible to 262 members of the species Paramecium caudatum, since every conjugant might possibly be different from every other; in Experiment 2 however this diversity is limited to the characters possessed by one individual Paramecium (the pro- genitor of the pure line used) and their possible recombinations. This fundamental difference in the constitutions of the two groups is very probably a cause of the production of greater diversity in form, and in the size, shape, and length of life of the indi- viduals that never divided, mentioned on page 393. The individuals used in the third experiment’ cannot be con- sidered in this discussion, since they were derived from a small culture kept for some time in the laboratory, and the history of only 28 exconjugants is known. Most of the abnormalities which have been described in the metazoa are explained .as arrests in development; as suppres- sions of some part of the normal process of growth and differ- entiation. Some of my abnormalities answer to this description; 410 RUTH J. STOCKING TABLE 4 Showing the frequencies of the different types of abnormals in the three experiments EXPERIMENT 1 EXPERIMENT 2 TYPE Number | Percent | Number 1 75 1.36 2 2 31 0.56 3 436 Teo2 32 4 26 0.47 5) 102 1.85 4 6 20 0.36 a 337 6.12 8 8 112 2.03 9 55 1.00 9 10 22 11 197 3.58 109 12 107 1.94 9 13 4] 0.74 3 14 123 2.23 11 15 23 0.41 1 16 4 0.07 17 We 202 3.67 25 18 3l 0.56 5 19 22 0.40 al 20 2} 0.40 3 21 43 0.78 22 43 0.78 2 23 25 0.45 5 24 30 0.54 25 17 0.31 26 278 5.038 384 Ze 1028 18.69 919 28 83 5! 102 29 9 0.16 30 211 3.83 | 1688 él 36 0.65 9 32 1 0.02 54 33 1430 26.00 87 34 56 OZ 15 ai 13 0.238 36 80 1.45 1 37 il 0.02 9 38 55 1.00 39 i 0.02 5) 40 93 1.69 16 otal eaeee o499 37.58 3567 Per cent i=) iw) bo (SSS) aye (8S) (SS) So ow (=) i=) on EXPERIMENT 3 Number | Per cent 6 0.10 3 0.05 90 1.61 10 0.18 88 1.58 6 0.10 1 0.01 4 0.07 2 0.03 134 2.41 32 0.57 42 0.75 6 0.10 15 0.27 5 0.09 10 0.18 By 5.78 2891 51.94 682 12.25 2 0.038 8 0.10 373 6.70 my 0.03 54 0.91 11 0.20 11 0.20 736 13) 2) 20 0.35 5566 38 .04 TOTAL Number | Per cent 83 0.56 34 0.23 558 3 8 36 0.24 194 132 20 0.13 351 2.39 113 0.77 68 0.46 24 0.16 440 3.01 116 0.79 76 0.52 176 1.20 30 0.20 36 0.24 227 155) 36 0.24 38 0.26 35 0.24 43 0.29 367 2-51 30 0.20 30 0.20 17 0.12 3553 24 .28 2629 17.97 187 W20 17 0.12 2272 15.52 45 0.31 57 | 0.39 1511 10.382 82 0.56 24 0.16 81 0.55 10 0.06 55 0.37 742 5.07 129 0.88 14632 100.00 INHERITANCE IN ABNORMALITIES 411 certainly the individuals that never divided, the double and the monster forms, and those abnormal races in which the division rate is very much lowered, can be explained in that way. But in a few of the abnormal races the division rate was normal, and there was very little tendency to die out; indeed they showed no abnormal characters at all except the development of a slightly bizarre form. Race C is a good example of such a race. It lived for 191 days, had an average division rate of 1.16 per day, and was consistently abnormal throughout its history. The mortality of the majority of the abnormal races was however considerably greater than that of the normal races. As the latter were all discontinued within a few days after conju- gation, nine days being the usual length of time they were kept, no exact comparison can be made between their mortality and that of the abnormal races. However, they may be compared in a few ways, that show there are differences between them in vitality and tendency to die out. In each of the three experi- ments a large proportion of the abnormal races died within two weeks after conjugation, only a few races being kept for any length of time. In Experiment 1 there were 21 abnormal races; 14 per cent (3) of these lived for over a hundred days (105, 131, 191); 29 per cent were lost or discontinued after, on the average, 9 days of life; 12 (57 per cent) died after an average length of life of 15 days, and an average number of generations, 6. This gives an average division rate of 0.4 a day. The normals, dis- continued 9 days after conjugation, had an average division rate of 1.12 a day. The facts in the other experiments are very similar to these; it is evident in all three that the abnormal races have, as a rule, a lower vitality and a greater degree of mor- tality. I explain this as largely due to the fact that the abnor- mal forms are hindered in their locomotion by their abnormal shape and so have not the normal facilities for meeting the exigencies of their existence. They often lie motionless on the bottom of the dish or slide for long periods; it was this lack of energetic movement which made possible camera lucida draw- ings of so large a proportion of them. It is probably for this reason that so few abnormals are found in the usual culture. 412 RUTH J. STOCKING VI. THE ABNORMALITIES AS HEREDITARY CHARACTERS; VARIATION, INHERITANCE, AND SELECTION As we have seen, the abnormalities which we are considering arise in consequence of conjugation. If half of a given stock are allowed to conjugate, the other half not, the former develops many of these abnormal races, while the latter develops none. This shows that the abnormalities cannot be considered due to infection, nor their reappearance in the stocks to the handing on of an infecting organism. | Since some lines are quite without abnormalities, while in others, under the same conditions, the abnormalities reappear for generations, it is clear that the tendency to abnormality is inherited. That is, the difference between a stock that thus produces abnormal individuals and one that does not, les in the constitutions of the stocks themselves, and is something that is transmitted during vegetative reproduction. Such hereditary diversities occur not only between normal and abnormal stocks, but also among the abnormal stocks them- selves. Precise types of abnormality are indeed not inherited exactly from parent to progeny; within a given line as we have seen there is great variation as to whether abnormality appears at all in a given individual, and as to the precise kind that occurs when the individual is abnormal. Nevertheless, as before set forth, certain types of abnormality are particularly common in some lines, other types in other lines. The diverse lines differ hereditarily in respect to something of which the diverse typical abnormalities are the outward results. It is only by keeping in mind the characteristic differences between lines that we shall be able to grasp their relation to the problems of heredity. The hereditary diversities thus far mentioned are between lines derived from diverse exconjugants. If anything like Men- delian inheritance occurs in infusoria, we could well expect such lines to show hereditary differences; this is of course as a rule true in any organism after the union of two parents to produce progeny. On the other hand, in vegetative reproduction, and in general in long continued uniparental reproduction of any sort, a remark- INHERITANCE IN ABNORMALITIES 413 able constancy in hereditary characteristics has been generally reported. All the progeny thus coming from a single parent have seemed uniform in their hereditary characteristics, though they may differ in their bodily appearance. And this is quite in agreement with the known cytological processes accompany- ing the two types of reproduction. In biparental reproduction there is a reduction and recombination of the nuclear elements, of precisely the same sort as the variation and recombinations of characters in the progeny in Mendelian inheritance. In uni- ‘parental reproduction, particularly of the vegetative kind,” such nuclear reductions and recombinations are not known; and the uniformity of the progeny is in agreement with this. These relations, with others not necessary to recount here, have given origin to the conception of the genotype as the hereditary constitution, in contradistinction to the bodily appear- ance. The genotype is commonly held not to change in vegeta- tive reproduction, or but rareky, and then by marked sudden steps, or mutations. In biparental reproduction the genotype does indeed change, but seemingly by mere shiftings and recom- binations, in numerically predictable ways; so that the relations here are quite in agreement with the condition sketched above for uniparental reproduction. A somewhat rigid, stereotyped scheme of heredity naturally results from the view of the facts as just set forth; in particular, evolution by gradual change, guided by natural selection, appears to be excluded. This becomes still more marked if we conclude with Bateson ('14) that all mutations consist in the dropping out of factors. However, certain investigators in genetics oppose this rigid view, holding that, over and beyond Mendelian recom- binations, hereditary variations of slight degree are frequently occurring, so that evolution may well be continuous and guided by selection. The recent papers of Castle give typical expres- sion to this point of view. 2 In their recent description of endomixis during the vegetative reproduction of P. aurelia, Woodruff and Erdmann observed neither reduction nor fusion of nuclear elements. 414 RUTH J. STOCKING If hereditary variations are frequently occurring, aside from Mendelian recombinations, it should be possible to find them in vegetative reproduction. Here we are freed from the mixing of types which makes these relations so difficult to interpret in biparental reproduction. The hereditary abnormalities with which the present study deals seem to offer a favorable oppor- tunity for the study of this matter. Within the same line of vegetative descent we find individuals that are in appearance normal, others that are outwardly abnormal. Can we by con- tinued selection of normal individuals on the one hand, of abnor- mal individuals on the other, break our single stock into two or more, differing in hereditary constitution? Experiments in selection To answer the question just proposed, selection was carried on for many generations in a considerable number of abnormal stocks. As before set forth, some of the races in which abnormalities occurred gradually changed character and became entirely nor- mal. In other races both normals and abnormals appeared for long periods, giving opportunity for long continued selection. We will first take up the large race C, of Experiment 1. Figures 10 and 11 give extracts from the pedigree of race C. This race, derived from exconjugant 101a, of Experiment 1, was kept for 191 days and produced during that time 4710 indi- viduals in the observed cultures; 2683 of these were abnormals. The early history of this race is shown in chart 1. CHART 1 Early history of Race C (n = normal; ab = abnormal) December 2 5 7 9 11 13 15) 17 : (Sn —1 4n —1 Discontinued In—l 4n—4 Sn—2 S8n—1 4n—1 ee lab—1 Dead lab—1 lab—1 Race C In this and in the succeeding charts the first figure under each date shows the number of animals present in that line on that date; m means that these were nor- mal; ab means that they were abnormal. The second figure, following the dash, denotes the number chosen to carry on the slide line, the others being put into ‘aps [Blo OY} pavaoy poArno adAy oY} JO Orv S[VUILOUGE ot} jo uorjytodoad aS.1e] B YVY} P9oTyOU Oq [[IM 4] “}USpTAS ST ‘UOTPIITOS IY} JO SsorpieSor ‘Soul, YOG UT SOTPTVULIOUG’ oY} JO gourivodde juoqsisiod oy], “poe}O9TOS 919M S]VUILOU YOIYA UT IUT] 10}STS B OPTS qySIA OY} UO {poxooTos OOM STRUILOUG’ YOIYM UT OUT] B Opts Je] oy} WO BurMoys “- aoVY JO oistpod oY} WOIJ JOVIPX, OT ‘BI RUTH J. STOCKING PEDDIE i i my) ‘One t 1 MY a different group of lines. ure 10, but from Fig. 11 Same as fig INHERITANCE IN ABNORMALITIES 417 bottle cultures, or discarded. In a number of the charts the second figure is omitted since the selection was the same on every date, one animal of a particu- lar kind being always chosen. When both normals and abnormals are present in a single line (Chart 2) the kind which are entirely discarded are placed above, the kind from which one is chosen is placed below, each with its number. D means that on that date that ine died out. Chart 1 therefore reads: On Decem- ber 2 the normal exconjugant was isolated; December 5, it had divided to form four normals, all of which were kept; December 7, these had divided into eight normals, two of which were kept; these two had divided on December 9 to eight normals, one of which was kept; December 11, this one had given rise to four normals, one of which was kept; December 13, this one had given rise to two abnormals, both of which were kept, and to eight normals, one of which was kept; December 15, the normal had divided to form four, one of which was kept, and the two abnormals, neither of which had divided, were both kept; Decem- ber 17, the normal line was discontinued entirely, one of the abnormal had died, and the other was kept and gave rise by repeated divisions to all the later individuals belonging to Race C. Race C therefore appeared to be entirely normal for eleven days after conjugation, at the end of that time producing two abnormals, one of which was double, the other small and de- formed. ‘This small one gave rise to all the later 4681 observed individuals of this race, comprising 303 generations. In 34 lines of this race a continuous selection of normals was made, in one case for 32 days and 24 generations. The data from these lines is given in table 5. Their character was not changed in any way by this selection; the average proportion of abnormality for these 34 lines is only 2 per cent less than that of the race as a whole (table 3). In one line only normals were produced for as long as six days, the line at the end of that time again producing abnormals. The history of this line with a few of the others is given in chart 2. In every case one normal was selected each TABLE 5 Data from the 34 normal-selected lines of Race C SUM NUMBER OF NUMBER OF Grown < an DAYS GENERATIONS NUMBER OF NUMBER OF PER CENT Las SELECTION OF NORMALS NORMALS ABNORMALS ABNORMAL bai CARRIED ON PRODUCED 2 5) 14 8 36 33° 48 3 29 32 24 233 294 56 Total 34 269 327 55 418 RUTH J. STOCKING day to carry on the line. It is evident that this selection’ has had no effect on the character of these lines. CHART 2 History of six normal-selected lines of Race C (n = normal; ab = abnormal; D = died out) April May 24-26 28 30 2 4 6 3 NO) ala! 16 ibs 40) BR 6ab lab 10n Sia 1D). lab 7ab 3n 6n D 7n fin PA IDE | fers Sab 2ab dab | | [ | | 3ab~— lab - Tad iim Sin | ab 10ab 6n D. lab 6ab 3n 2n D. be 2ab 2ab «8ab =e Babs 2ab 2ab 12n 4h Poy ibey Pan, | Pini In iba | Gin Zin In 483 lines of race C a continuous selection of abnormals was made, in one line for 40 days and 37 generations. Their char- acter was not changed by this sort of selection; their proportion of abnormality is the same as that for the race as a whole (table 3). The data from them is given in table 6, and the histories of a few in chart 3. In very case one abnormal was chosen on each day to carry on the line. CHART 3 (n = Normal; ab = abnormal; D = died out) History of five of the abnormal-selected lines of Race C May 10 12 14 16 18 20 Pa 24 26 28) a0 2n In 2n 2n 2ab 3ab dab 2ab 2ab 2ab D. 4ab 2ab 4ab 4ab lab lab 1D) 4n 4n 2n 2ab 3ab 4ab 3ab 4ab Jab 2a: 3n In 2n 2ab 3ab 3ab 2ab Sab 4ab 1D) 3n In 3n 6ab lab sab 2ab 4ab 2ab 3ab lab 2ab De Three times in the history of this same race all of the indi- viduals died with one exception; in every case this single sur- INHERITANCE IN ABNORMALITIES 419 viving individual was an abnormal animal which gave rise by repeated divisions to the further representatives of its race. On two of these occasions the two groups of lines which rose from the two first daughter cells of this single abnormal animal were treated differently; in one group the most normal indi- viduals were selected; in the other group, sister to the first, the most abnormal were selected. The data for these two pairs of groups is given in table 7. In this table, one abnormal-selected line in the first set lived longer than 56 days; in the second set one normal-selected line lived longer than 20 days; both however for purposes of comparison are counted in this table as having died out at the same time with the other lines of their sets. TABLE 6 Data from the 43 abnormal-selected lines of Race C NUMBER NUMBER OF NUMBER OF GROUP OF DAYS GENERATIONS NUMBER OF NUMBER OF SoA GER SELECTION OF NORMALS NORMALS ABNORMALS ABNORMAL ae CARRIED ON PRODUCED 1 3 40 30 58 83 a : 2 26 19 103 126 55 sail duaes 32 31 135 163 55 4 6 18 13 49 93 65 Total 43 345 465 57 TABLE 7 Data from two sets of sister groups of lines; in one group of each set the most abnormal were selected; in the other group, the most normal NUMBER OF NUMBER NUMBER NUMBER PER CENT 1 Abnormal- selected 15 56 48 387 489 56 Normal- selected 15 56 50 486 834 63 2 Abnormal- selected 20 20 25 131 157 54 Normal- selected 20 20 24. 211 320 60 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 420 RUTH J. STOCKING With these two exceptions, one abnormal-selected and the other normal-selected, the two groups ih both sets are very much alike. They show a great deal of similarity in length of life, number of generations, and proportion of abnormality, the diverse selec- tion having had no effect on the character of the lines. Indeed, in both sets, the lines of the normal-selected group show a larger proportion of abnormals. Thus in line C we have a race in which long continued selec- tion has no effect on the inheritance of abnormalities. And these results are typical of a large number of races—all those classi- fied on page 399 as class 2 (97 races all together). In all of these races where such a procedure was possible the selection of nor- mals was carried on as long as the race lived, in an attempt to establish a normal line; but with all these races this effort failed. The normals selected continued indefinitely to produce abnor- mal progeny in the original proportions. With these races there- fore the results of continued selection are the same as those obtained by the majority of investigators in uniparental repro- duction: selection does not alter the inherited constitution of the line nor produce two genotypes from one. In another set of abnormal lines however, different results were reached. In twenty-five races of our three experiments the single race was split up by selection into hereditarily diverse groups, one group composed of lines that continued to produce abnormals, the other composed entirely of normals. The main facts as to these twenty-five races are given in table 8. Many of these lines were kept but a short time, so there might be doubt as to the precise significance of the results from them. Certain lines however were kept for a very large number of gen- erations and give conclusive results. This was the case with the lines A and B of Experiment 1, and with the six lines of Experiment 3. Some details will therefore be given as to these races. The history of race A may be taken as a type. The excon- jugant from which this arose gave rise to normal individuals only for seven days, or until December 9, when eight abnormals and ten normals were found on the slide. After this the race did INHERITANCE IN ABNORMALITIES AD: TABLE 8 (first part) Data from the 25 races of the three experiments which were affected by selection ABNORMAL LINES EXPERI- f MENT HACK NGS a: = . ; Per cent bet ays Generations Normals Abnormals Abnoatall 1 48a 3 27 9 27 6 18 53b 1 11 3 2 2 50 93b 4 23 10 38 9 19 105b i 17 2 1 1 50 A 178 131 74 524 850 63 B 314 105 83 1328 1788 57 2 2a 1 1s: 4 2 7 78 4a 2 23 £5) 6 7 54 4b 4 27 14 52 22 30 7a 2 Dill 15 34 19 36 10a 1 13 4 8 3 27 10b 24 27 26 293 323 52 18a 20 20 28 634 748 54 40b| 1 17 6 6 2 25 4la il 11 4 6 9 60 62a 1 11 5 5 3 38 69a 9 27 25 172 213 55 83b 7 Dil 15 40 22 35 87a 7 27 15 55 23 29 3 54b 4 54 33 673 117 15 55b of 54 53 1785 467 21 56a 18 77 62 3650 1121 23 56b 20 UU 65 2818 567 17 59a 30 77 32 1051 369 26 59b 73 ae 50 3201 16438 34 TRotales. 733 16411 8341 ; Average 29 40 26 656 334 39 not return to complete normality, but always showed a large proportion of abnormal forms among its members. For some time all the normal individuals were discarded and all the abnor- mals kept. By December 23 this procedure had resulted in so many lines (178) that all but 45 of the most abnormal were discarded. These 45 very abnormal lines were grouped for con- venience into 13 categories. The genetic history of the lines 422 RUTH J. STOCKING TABLE 8 (second part) Data from the 25 races of the three experiments which were affected by selection NORMAL LINES . EXPERI- wiles, x* sia ie Days ney Normals | Abnormals| Per cent iL 48a 7 1 6 7 20 53b 7 if 5 7 Of 93b 9 4 6 5 50 105b 1 1 5 2 4 A 62 6 26 22, 134 B 5 to 49 3 20 21 126 2 2a, il 4 8 31 4a, 1 6 3 6 4b 2 1 6 4 77 ia 8 1 4 3 ih 10a 8 il 4 3 10 10b 4 1 4 3 8 18a 18 4 6 9 75 x 40b 1 8 3 8 4la 4 1 8 3 8 62a 6 il 6 4 16 69a 8 2 4 6 50 83b 1 2 4 6 87a 8 4 12 6 53 3 54b 9 to 42 15 35 20 3589 55b 9 to 11 2 12 13 272 56a 9 to 50 8 37 24 936 56b 9 to 44 11 24 26 1164 59a 9 to 37 2 47 35 425 3 0.5 59b 37 to 53 12 39 27 1334 Motel: 86 8346 3 AVC uaeine 24 3 14 ial 334 0.12 0.2 * X is the number of days that the selection of normals went on before the normal lines were isolated. which made up these groups is given in chart 4. Two of these groups of lines, A2 and A6 died very soon; the other lines lived for some time and the data from them is given in table 9. INHERITANCE IN ABNORMALITIES 423 TABLE 9 Data from eleven groups of Race A GROUP LENGTH OF NUMBER OF NUMBER OF NUMBER OF PER CENT LIFE IN DAYS GENERATIONS NORMALS ABNORMALS ABNORMAL AUIS hr rac. 31 16 22 35 61 IAS Hae Eee Ne 69 40 34 85 ql A Cae ee BY 22 13 44 77 JANG Sah ee oars Sem 131 64 61 182 75 JAY (SPR ae eee 33 16 18 54 75 INS ae ens kOe ee 47 16 14 55 80 JNO ee oe ae rae 69 30 35 154 81 /\O|O lone me tenied 25 11 31 19 48 2X] IN a lee. eae 27 11 13 26 67 PRONE cok carseat 51 13 26 54 67 AUIS Spee arte cas 115 74 391 132 25 Wangestic. 5. is. - 131 74 391 182 81 Motel asker. 658 850 56 Average......-. 58 28 60 77 56 CuHarT 4 Genetic history of the thirteen groups of Race A (vn = normal; ab = abnormal) December 3 5 a 9 veil 13 15 Lz Group an { n Al ab ab A2 | n A3 | | n Fe A4 ab n ab A5 (n a a ie A6 ab AZ te ab ab A8 i n n AQ9a x | n [* ab a ee A9b ab AQe | | n ab ce ab Al0a ab ab A10b fab ab ab n All ab ab ab n Al12 is hen l ah ms Ms ae ab Al3a n A13b 424 RUTH J. STOCKING After December 23 every one of the lines belonging to race A was carried on in two parts or sub-lines; in one sub-line the most normal individuals were selected; in the other sub-line the most abnormal individuals were selected. ‘This procedure kept going a number of normal-selected lines up to as many as 40; and an equal number of abnormal-selected lines. In almost all of the normal-selected lines the continuous selection of normals was not possible; some generations were entirely abnormal, no normals at all being produced. But 62 days after this normal selection had begun, one normal was isolated which gave rise to six lines in which the continuous selection of normals was possi- ble for some time before they died out. Their history is given in chart 5, and the data afforded by them in table 10. The largest of these six lines was kept for 26 days after the con- tinuous selection of normals was begun and gave rise to 22 gen- erations of normals. There were 42 individuals in the observed cultures of this one line; one of these was abnormal; all the rest were normal. TABLE 10 Data from the six lines of Race A in which the selection of normals was continuous SELECTION CONTINUOUS LINES ENTIRELY NORMAL Genera- Per cent Genera- Days as Normals | Abnormals AL onal Days tions Individuals 12 10 PA 0 0 ee, 10 Pat 10 10 12 0 0 10 10 12 6 6 8 0 0 6 6 8 22 18 32 1 3 10 a 11 14 13 20 0 0 14 13 20 26 Ds 41 il 2 20 1185) 25 Greatest 26 22 41 1 3 20 11855 25 Total 134 2 1 97 As all the animals of these six lines were descendants of one normal individual present on February 27 (chart 5), they can all be counted together as one group, which will give 134 normal * INHERITANCE IN ABNORMALITIES 425 individuals to 1 abnormal individuals in 21 generations. This is a very low proportion indeed as compared with that for the race as a whole (56 per cent). Moreover, every one of these six races was entirely normal for some time before it died, as CuHart 5 Genetic history of six lines of Race A in which the selection of normals was continuous (n= normal; D = died out) Line Feb. Mch. 27 1 3 5 OF ell SSS Sala On 21) eo ome, ie (suman Qn Any wD) \4n 4n 4n D. Auelsis Sie ye ee lab | 2a 2n 20) 4n° Soin 2needne an) Dp: 3n 4n 4n 2n 2n 8n DPD. (iin, 40 4n 4neesne4nisaneeone On Ine DE shown in the table. Figure 12 gives that part of the pedigree of race A which includes these six lines. In 24 of the abnormal selected lines abnormals were produced in every generation for some time and their selection was there- fore continuous. The history of six of these lines is given in chart 6. Table 11 gives the data from all of them. The largest line was kept for 36 days after the selection of abnormals became CHART 6 Genetic history of six of the lines of Race A in which the continuous selection of abnormals was made (n = normal; ab = abnormal; D = died out) December January Pall 23 25 27 2S 31 Z 4 6 8 LOM 24 Gaels 2n In Zabeecabee2abewlabs saby. tabi lab: lab dabs sD: In lab 2ab tab lab 1D). In lab lab lab lab 1b), 2n 2ab 3ab- lab 1D): : lab lab 4ab Jab 2ab 2ab 2ab 2ab lab 2ab dab 2ab 2ab lab D. In In lab lab lab lab lab tab tab tab lab’ Dz. INHERITANCE IN ABNORMALITIES 427 continuous, and gave rise to 16 generations. Two normals and 18 abnormals appeared in the observed cultures of this line. The proportion of abnormality (79 per cent) shown by all these lines-taken together is much higher than that shown by the race as a whole (56 per cent). Figures 13 and 14 give part of the pedigrees of three of these lines. TABLE 11 Data from the 24 lines of Race A in which the selection of abnormals was continuous SELECTION CONTINUOUS LINES ENTIRELY ABNORMAL Days Sei Normals. | Abnormals Ger cent: Days ener Individuals 16 9 4 13 76 10 6 10 18 6 3 6 67 12 3 3 8 2 1 2 67 6 2 2 6 1 0 1 100 6 1 i 8 1 1 1 50 6 il 1 6 4 2 4 67 4 2 3 6 2 0 2 100 8 1 2 18 6 8 8 50 10 1 1 36 16 2 18 90 28 12 14 20 3 3 3 50 12 1 1 16 5 5 6 55 8 2 2 22 11 4 11 73 8 3 3 24 8 6 8 o7 12 2 2 20 9 4 8 67 8 2 2 22 8 3 6 67 12 2 2 20 6 0 6 100 20 6 6 24 5 0 5 100 24 5 5 8 4 1 4 80 6 3 3 12 4 1 4 80 10 3 3 14 4 2 4 67 10 2 2 12 3 0 3 100 12 3 3 10 2 0 2 100 10 2 Dy 6 8 8 10 56 2 1 1 14 8 22 12 35 10 5 5 Greatest 36 16 22 18 100 28 12 14 Mopale ss. 80 147 65 79 430 RUTH J. STOCKING We have therefore in race A a group of individuals all de- scended by vegetative reproduction from one animal, exconju- gant 1b, which for 74 generations and 131 days constantly pro- duced abnormals in an average proportion of 56 per cent of all the individuals of the race. From this clone, having this herit- able character, abnormality, were obtained through the action of selection, two diverse groups of individuals. One group had a constant proportion of abnormality, on the average, of 70 per cent; the other with almost total normality, the abnormality having been reduced to an average of 1 per cent; all of the lines were entirely normal for some time before they died out. In this clone therefore we have an inheritance of a variation, abnor- mality; and also we have permanent changes in this heritable variation, brought about by the action of selection. TABLE 12 Data from fourteen groups of Race B LENGTH OF NUMBER OF NUMBER OF NUMBER OF PER CENT CE OUE LIFE IN DAYS GENERATIONS NORMALS ABNORMALS ABNORMALS Bee ene he hee 81 47 43 73 63 1B ae ea 81 67 107 80 43 BSR eee ee 81 73 178 144 45 [Ryd ek: aoe aN 105 83 292 338 54 Bb Sereno ee 75 61 60 94 61 BOW et eee 61 46 46 67 59 BVBoe. bees Ere (al 66 114 77 61 Basset cae 73 58 66 118 64 1Bio We iN a eRe on 89 64 (e 110 60 BLO eS oe 103 74 145 148 50 IB da oes: Meee 67 46 40 137 Ui IBS eat he pene 65 41 38 66 63 Ae cone: 69 53 87 145 62 Bilbsceis cacao ql 15 52 96 65 Greatest 105 83 292 338 ad T ovale arene. 1371 1788 56 The history of race B shows some slight variations from that of race A. Its early history is shown in chart 7. It was en- INHERITANCE IN ABNORMALITIES 431 tirely normal until December 7, five days after conjugation. On that date two of the existing 27 normal individuals were chosen to carry on the race; one of these continued to produce only normals until it was discontinued six days later. By that time it had given rise to 53 normal individuals. This was the largest group of normals produced by race B. Cuart 7 (n = normal; ab = abnormal) Early history of Race B and formation of the fifteen groups December 3 5 7 9 11 13 15 lel Group 4n—1 4n—4. 50n Discontinued {lab—1 dab—3 lab—l Dead lab Bl Serie yt ae (6én—6 4ab—4 {lab B2 in B3 | Zab—2 fin B4 yn) po) ba lab B5 Ropes, : 2ab—2 ‘is erates lab B6 2n B7 In B9 Dya| pa FL Es TSG lab Bll (lab B12 lab B13 Jah? Leer 2 omen Bie lab B15 The other normal kept on December 7 gave rise to all later individuals of race B, which were kept for 105 days, and 83 generations, and had an average proportion of abnormality of 56 per cent. The history of this race from this point on is exactly like that of race A. On December 23, 83 of the most abnormal lines were selected from the existing 314, and grouped (table 12). 432 RUTH J. STOCKING Each one of the 83 lines was divided into two sub-lines, in one of which the most normal were selected, in the: other the most abnormal. In six lines the continuous selection of normals was possible. Their history is given in chart 8, and table 13 gives the data from them. Two of these lines produced abnormals the last day they were kept, and a third was entirely normal for only four days before death; the selection has slightly de- creased the proportion of abnormals produced, but has not elimi- nated the abnormal character. But the three other lines became entirely normal some time before death, as the table shows. Selection here has had a decided effect; has entirely eliminated the abnormal character. CuHaRT 8 Genetic history of the six lines of Race B in which the selection of normals was continuous (n = normal; ab = abnormal; D = died out) Line January February D022, O24 9G 008. 30h pele se 5D, mae go lab 4ab 2ab B2 4n8n 2n 4m. 2n> 4n 8m 2n) An 4n D. B3 Sn) 4 Snes One onronmen. 4ne eon 1D). Line December January Ds Di OK) RL A» 16 8 WO) aes al BO) BR} Dal 3ab lab 2ab \ B4b2 Pan ik kay ah Om kin Shh A oy i, i, SN Mn Si. 1D). 2ab lab 3ab 3ab lab 2ab B4il Ky Dn Wy An Way Bn iy eine hay tin Aim Aoy4bn ID; sab Qab 2ab lab 2ab lab 2ab lab lab 3ab lab lab B7 lO Bin “Pin On” Ba ilm iby key Pan lin in Bin An Fin 1D). 2ab lab B15 Pay oane Pan an Tay, Tbs 1D). In nine lines the continuous selection of abnormals was possi- ble. The history of three of them is given in chart 9 and table 14 gives the data from them. The proportion of abnormality in most of these lines has been somewhat raised by this selection over that of the race as a whole; but not to such an extent as in INHERITANCE IN ABNORMALITIES 433 race A. Pedigrees of two of these lines are given in figures 15 and 16. TABLE 13 Data from the six lines of Race B in which the selection of normals was continuous SELECTION CONTINUOUS LINES ENTIRELY NORMAL Genera- Per cent Genera- Days pee Normals | Abnormals Saat Days Mons Individuals 20 20 33 7 17 10 11 18 18 22 43 0 00 18 22 43 28 PA 30 6 17 20 17 29 26 13 15 12 44 4 4 5 30 13 16 20 56 0 0 0 12 a 5 3 38 0 0 0 Greatest 30 22 43 20 56 20 22 43 WOM os cote 142 48 e225 95 CHART 9 Genetic history of three of the lines of Race B in which continuous selection of abnormals was made (n = normal; ab = abnormal) Dec. January 31 2 4 6 8 10 12 14 16 Spee 20) 222426 In In In In labeesabmmcabemoabl cab) e2abremiab) = labs abe Dead 2n 3n 3n 3n 2n In Zab 2ab 4ab 2ab Sab lab tIlab 2ab 2ab 4ab lab lab Dead In 3n 2n 2ab «63ab-) «COolab) «6(2ab SC 6ab SC 4ab 0«C4ab) SColab) «2ab-)Ss dab ~=lab Dead We have therefore in race B a group of individuals consti- tuting a clone, which after the fifth generation and the pro- duction of 32 individuals, gave rise to two diverse groups, with different hereditary characteristics. One group was entirely nor- mal; the other group showed a large and constant proportion of abnormals. Later from this abnormal group, by continued selection in opposite directions, lines were isolated which showed INHERITANCE IN ABNORMALITIES 435 TABLE 14 Data from the nine lines of Race B in which the selection of abnormals was continuous SELECTION CONTINUOUS LINES ENTIRELY ABNORMAL Days Sones Normals | Abnormals eae Days Geusra: Individuals 20 9 2 20 83 8 5 “eR 18 8 t 11 73 10 2 Pe 18 6 5 8 62 8 2 2 24 11 14 16 53 2 1 1 18 9 1 10 91 14 8 8 22 13 6 20 ae 12 8 11 22 10 13 15 54 2 2 3 24 13 11 15 58 2 1 1 14 5 2 6 75 12 5 5 hereditary differences in degree of abnormality. One set of Jines showed a:very low degree, three lines becoming entirely normal. The other set showed a very high degree of abnormality, one line having 90 per cent of abnormals. We have then in race B the inheritance of a variation within the clone; and a splitting up of the clone, both with and without selection, into hereditarily diverse groups. The six races of Experiment 3 underwent a most strict selee- tion of normals throughout their history; and in every case this eventually brought about a change in the inheritance of the abnormality. In each race all of the individuals arising from the exconjugant were kept for the first six days; thereafter the abnormals were discarded and only normals kept, as far as possible. Abnormals were kept only when no normals were present with which to carry on the race. In all these six races this procedure had much the same effect. Some of their lines were not changed at all; they continued to produce abnormals from the normal cells selected. But otherlines became entirely normal. In every case there was inheritance of the variations which had arisen within the clone. Figures 17, 18, 19, and 20 give pedigrees of four of these races, 56a and b, and 59a and b. All six races are very similar in their history. On January 17 none of the six had completed their first division; all had formed THE JOURNAL OF EXPERIMENTAL ZOOLOGY, vou. 19, No. 4 ‘ a gee . a DEAD DEAD JEAD ( a Be ( bee ee ) Ye i ae ety aa Pes ee Sa Fig. 17 The first part of the pedigree of 56a, showing the origin of four of the normal lines. The other four normal lines were derived much later from one of the abnormal lines shown here. In this figure and in the succeeding ones, in each generation only those animals are shown which were selected to carry on the lines. 436 INHERITANCE IN ABNORMALITIES 437 Fig. 18 First part of the pedigree of 56b, showing the origin of three of the normal lines. The other eight normal lines were derived much later from the abnormal lines shown here. 438 RUTH J. STOCKING —— JOR ee Fig. 19 The first part of the pedigree of 59a, showing the origin of one of the normal lines. The other normal line,was derived 32 days later from one of the abnormal lines shown here. INHERITANCE IN ABNORMALITIES 439 Fig. 20 The first part of the pedigree of 59b, showing the origin of one of the normal lines. The other eleven normal lines were derived much later from the abnormal lines shown in this figure. double monsters of the L variety. On January 19, exconjugant 56a (fig. 17) had divided to form four abnormals; two of these died without dividing further. One of the other two produced 440 RUTH J. STOCKING one abnormal which died before dividing, and one normal which gave rise to two entirely normal lines which were kept for ten days and eleven generations. They comprised 115 observed individuals, all normal. The other abnormal present on Janu- ary 19 divided to form one abnormal (which died before further division) and eleven normals. Nine of these were kept. Six of them gave rise to lines which remained entirely normal; three produced abnormal lines. Two of these abnormal lines remained abnormal throughout their history, never responding positively to the normal selection they underwent. From the other abnor- mal line five entirely normal lines were split off, kept for a number of generations varying from 6 to 24. Three of these normal lines were kept until April 1; their eight abnormal sister lines were also kept until April 1, 77 days after conjugation. Since all six lines had very similar histories they will not be given in detail. The main facts as to the production of the normal lines can be gathered from the pedigrees (pages 436 to 439) and the table (pages 421-422). Thus on the whole the work on selection shows that with relation to these abnormal characteristics heritable variations are in many races occurring during vegetative reproduction. By selection the effect of these variations may be accumulated, so that while one part of the race remains abnormal, or even increases the proportion of abnormal individuals, another part greatly decreases the proportion of abnormality, or becomes entirely normal. The general bearing of these results will be discussed after the rest of the facts have been brought out. Relation to biparental inheritance Does conjugation tend to produce similarity in respect to abnormality or normality between the progeny of the two indi- viduals that conjugate? That is, if the descendants of one of the two members of a given pair are abnormal, is there a tend- ency for the descendants of the other member to be abnormal also? Ifa given stock is abnormal, is the stock derived from the mate of its progenitor likely to be abnormal also? Jennings and Lashley (13, ’13a) have shown that there is such a tend- INHERITANCE IN ABNORMALITIES 44] ency to resemblance between the stocks derived from the two members of a pair, in respect to fission-rate, death-rate, and size. The question with relation to abnormalities may be attacked by the same methods employed by Jennings and Lashley. The problem presents itself as follows. From a certain number m of exconjugants (forming m/2 pairs) a certain number n of abnormal races are produced. In some cases both races derived from a pair are abnormal; in other cases only one; in others, neither. Is the number of cases where both races are abnormal greater than would be expected if the pairing had no relation to the distribution of the abnormalities? If so, this is evidence that the pairing tends to make the two races alike in this respect. Jennings and Lashley (’13) give a formula for determining the most probable number of pairs that will be found to be alike in respect to any such character if the pairing has nothing to do with the distribution. This formula is: y ee) 2 m+3 in which the nearest integer below k is the most probable num- ber of pairs in which the two members will be found alike in the character in question; n is the number of cases (lines) in which this character occurs, while m is the total number of cases (lines of descent from exconjugants in this case). (This formula holds absolutely only when n is even, but by obtaining.the result for the two even numbers above and below the actual number, if the latter is odd, the difficulty may be avoided.) In examining this matter for our case with respect to nor- mality and abnormality, 17 of the 131 pairs of the first experi- ment must be omitted from consideration, since in these the characteristics of one or both members is unknown. This leaves 114 pairs to be dealt with, giving 228 lines of descent. In the second experiment 3 of the 100 pairs must be omitted from consideration, leaving 97 pairs to be dealt with, giving 194 lines of descent. In the third experiment there are 14 pairs to be considered, giving 28 lines of descent. This makes 450 lines of descent all together. 4492 RUTH J. STOCKING Of these 225 pairs, in the first experiment both members in 48 gave normal lines, both members in 26 gave abnormal lines, while in 40 the two members gave lines diverse in this respect. The total number of normal individuals was therefore 136, of abnormal individuals 92. In the second experiment, both mem- bers in 7 gave normal lines, both members in 69 gave abnormal lines, while in 21 the two members gave lines diverse in this respect. The total number of normal individuals of experi- ment 2 was therefore 35; of abnormal individuals 159. In Experiment 3, both members in 8 pairs gave normal lines, and both members in 6 gave abnormal lines; pairing was perfect. The total number of normal lines of Experiment 3 was there- fore 16; of abnormal lines, 12. How many pairs with both lines abnormal should we expect to find in these cases if pairing does not affect the distribution of abnormalities? In the first experiment n is 92, while m is 298. Applying the formula we find that the expected number of pairs with both members abnormal is 18; the actual number is 28. In the second experiment n is 159 while m is 194. By the formula the expected number of pairs with both members abnor- mal is 66, while the actual number is 69. In the third experi- ment n is 12 while m is 28. The expected number of pairs is 2, the actual 6. In the same way the expected number of pairs with both members normal is in the first experiment 40, while the actual.number is 48; in the second experiment the expected number is 3 the actual 7; in the third, the expected number is 4 the actual 8. It is therefore clear that in all three experiments the number of like members of pairs is greater than would be the case if the pairing had no effect on the distribution of the abnormalities. Conjugation tends to cause the descendants of the two memlers of pairs to become alike in respect to abnormality and normality. Some of the exconjugants never divided after separation. The same question may be examined with respect to them. Do the two members of a pair tend to have the same fate in this respect? There were 72 exconjugants that never divided, out of the 228 of the first experiment, and 27 out of the 194 of the second INHERITANCE IN ABNORMALITIES 443 experiment. In the first experiment both members in 19 pairs were affected; in the second experiment, both members in 8 pairs. Applying the formula (first experiment, n = 72,m = 228; second experiment, n = 27, m = 194) we find that if pairing has no effect on the matter, the most probable number of cases for both members in the first experiment is 11; in the second, 2. Similarly there are in the first experiment 61 pairs in which both members divided, while the most probable number is 53. In the second experiment the actual number is 79, the most prob- able is 72. Again we find that conjugation increases the resem- blance of members of pairs in this respect. It will be of interest also to give the relation of pairing to the date on which the lines finally died out. This is known for both members of but 51 pairs of the first experiment, and 41 of the second. The numbers in the third experiment are too small to be considered. In some cases the descendants of both members of a pair completely died out on the same date; in other cases the descendants of one member lived longer than those of the other. The main facts are these. In the first experiment, on Decem- ber 5, twenty lines were found to have died out; these included both members of 8 pairs. The probable number of pairs if con- jugation does not affect the date of death is but 2. On Decem- ber 7, 53 individual lines ended, including both members of 20 pairs. The most probable number of pairs would be but 14. On later dates the numbers are too small to be significant. In the second experiment on December 6, 31 lines were found to have died out; these included both members in 7 pairs. The probable number of pairs if pairing does not affect the date of death is but 6. On all the other dates the numbers were sosmall that they are not worth considering. Conjugation thus tends to make descendants of the two members of a pair alike in their - length of life (in their ‘vitality’). All together, the evidence shows that conjugation induces resemblances in the two members of a pair in respect to all the characters examined: in the tendency to fail to reproduce after conjugation; in the abnormalities produced by their offspring; and in the length of life of the stocks produced. 444 RUTH J. STOCKING VII. SUMMARY AND DISCUSSION OF RESULTS Among the progeny of a large proportion (from 36 to 81 per cent in the different experiments) of exconjugants of Paramecium caudatum, abnormalities appear frequently and constantly. These abnormalities show themselves to be hereditary in the following respects: 1. Lines derived from different exconjugants differ in respect to them: some lines show no abnormalities; others show a small proportion of abnormal individuals; others large proportions up to cases in which abnormality is almost or quite universal. The tendency to abnormality is transmitted in fission; definite proportions of abnormality being characteristic of particular lines. In one case there was inheritance of a specific type of abnormality carried through 303 generations (Race C). 2. The diversities in abnormality occurring within a single line of descent (derived from a single exconjugant) are in some lines not hereditary, so far as can be determined by long con- tinued selection. In a very large proportion of the races in which the abnormals were regularly discarded and only nor- mals retained to carry on the race, the abnormal character per- sistently reappeared, the selected normals producing abnormal progeny. In all the abnormal races there is a wide variation in degree of abnormality of the individual, from those perfectly normal to the monsters so deformed that they would never be recognized as paramecia if their history were not known. Yet, as stated above, in most cases the progeny of all these variations were alike, the daughter cells of normal individuals being often just as abnormal, or even more so, than the daughter cells of monsters. This of course agrees with the conditions found in most of the studies on inheritance in ‘pure lines’ or clones: the diversities within the lines are not inherited. 3. But in other lines diversities within the line showed them- selves to be heritable, so that selection gave very different results from those usually obtained in pure line work. By selection, single lines, derived by fission from a single parent, were divided into two or more races differing hereditarily. This was success- INHERITANCE IN ABNORMALITIES 445 fully accomplished in twenty-five races; from each of these were isolated two sorts of lines, one quite normal, the other continu- ally producing abnormalities—the two cultivated side by side. Calkins and Gregory (713) have in some cases obtained four diverse races from the four primary daughter cells, or ‘quad- rants’ of an exconjugant—these being the four individuals that receive the four macro-nuclel produced before fission occurs. It is to be noted that our selection resulting in the isolation of lines differing hereditarily in abnormality has often been brought about much later in the series of generations, so that the differ- entiation has occurred within the compass of a single ‘quadrant,’ or indeed within a much narrower fraction of the descent. In several cases differentiation through selection did not begin till after several weeks had passed with production of a great num- ber of generations. Thus the results of selection in the present case cannot be interpreted as due to a primary difference in the four original macronuclei produced during conjugation. Selec- tion is effective when begun with progeny of a single individual that has appeared many generations after conjugation. 4. In a race of Paramecium which upon extended examina- tion shows no hereditary abnormalities, conjugation results in the appearance of many lines which are hereditarily abnormal, others which are normal throughout (Experiment 2). This is of course an example of what Jennings (713) has de- scribed as the ‘production of variation by conjugation.’ It appears to fully meet the desire of Dobell (14, p. 172) for “convincing evidence of a concrete instance in which from a known race—constant in a certain character—a new race—per- manently diverse in this character—has arisen as a result of con- jugation.”’ For in Experiment 2, from a clone with a constant character of normality (as shown by the progeny of the 54 split pairs of this clone) 101 races which were permanently diverse from the original one, being hereditarily abnormal, arose as a result of conjugation. 5. In the diverse lines descended from the different exconju- gants of a conjugating culture, the two lines descended from the two individuals that have conjugated together tend to be 446 RUTH J. STOCKING alike in respect to normality or abnormality. That is, if the progeny of the exconjugant a are abnormal, the progeny of its mate 6 are more frequently abnormal than would be the case if the distribution of abnormal races were not affected by con- jugation. This is an example of what Jennings and Lashley (13) have called biparental inheritance as a result of conjuga- tion. As these authors point out and as Dobell (14) has recently emphasized, this does not mean that the characters of the progeny of the two exconjugants are known to resemble the characters which the two parents had before they conjugated. It means merely that the characteristics of the progeny of a are not determined by the nature of a alone, but partly also by the fact that a has conjugated with 6. Just what the resulting similarity of the progeny of a and 6 should be called is of little importance, as compared with a clear grasp of the facts in the case, yet it is perhaps worth while to point out that similar relations often appear in what is called inheritance in higher organisms. Two heterozygotic parents frequently produce prog- eny which differ from both of them, yet what the progeny shall be is determined by the constitution of both of the parents. 6. We are dealing here with characters that are called abnormal. What is the bearing of this on any conclusions drawn from this study? Can we learn anything worth while from the study of abnormal characters? What we can learn from abnormalities, and what the relation of their behavior is to that of other characters can be deter- - mined only by investigation; the present paper ‘is offered as a contribution toward answering such questions. But certain gen- eral principles appear worthy of consideration. What is the real status of the conception of abnormality? Does it mean anything more than that the condition so characterized is not the one usually found? It certainly does not mean that the condition is one not subject to law of any sort. If we charac- terize the course of inheritance of these characters as abnormal, we can mean no more than that they do not follow the usual course. But the course they do follow is one actually occurring in animals, and therefore one not inconsistent with the nature INHERITANCE IN ABNORMALITIES 447 of organic matter or the constitution of organisms. It is per- fectly possible for organisms to exist in which hereditary varia- tions occur during non-sexual reproduction, and in which these variations can be accumulated through selection; for in the characters here studied this does occur. No one therefore can say a priori what characters must descend in this manner, what in some other manner. It is then an open question whether a similar scheme of descent may not be followed by other characters, either in this same organism, or in other organisms. Jollos (14) has raised the question whether all the hereditary variations shown by Jennings to follow upon conjugation may not be of the nature of abnor- malities. It is entirely possible that all show the same scheme of descent as the ‘abnormalities’ considered in this paper; but if so then the ‘abnormal’ cannot be characterized as the unusual. In the same way it is possible that the hereditary differences among the progeny of single exconjugants observed by Calkins and Gregory (’13) may be of the same nature as these abnor- malities. It is quite possible that lines might be ‘abnormal’ in ways evident only through study of the fission rate, or the like. But with such a state of the case the distinction between abnor- mal and normal becomes evanescent. The facts are, that in Paramecium as a result of conjugation, there appear lines or races that are hereditarily diverse; and that within some of these lines hereditary diversities likewise appear even during asexual multiplication. Whether we can maintain some special category such as ‘abnormality’ for all these cases, can be discussed; but this does not do away with the facts as to their existence and the scheme of their descent from genera- tion to generation. It is of interest to compare the genetics of these abnormalities in Paramecium with abnormalities in other organisms. Almost all the characters studied by Morgan and his associates in Dro- sophila, on which have been reached results of such immense importance for the entire theory of genetics, may be charac- terized as abnormalities. A large porportion of them show, as is well known, typical Mendelian inheritance. Others show a 448 RUTH J. STOCKING more irregular course, resembling in this respect those of Para- mecium. Such a character is the ‘beaded wing,’ recently studied by Dexter (14). Lutz (11) had previously investigated an abnormal wing character in Drosophila, with results that are still more similar to the conditions shown in Paramecium. Lutz found that certain abnormalities of venation were herit- able, but, as in Paramecium, the inheritance was not precise; the abnormalities of parent and offspring might be diverse, but the tendency to produce some sort of abnormality of venation was inherited. Furthermore the proportion of individuals abnor- mal, and the degree of abnormality, were modifiable by selection, apparently through fine gradations. Lutz is of course working with bisexual reproduction, which greatly complicates the inter- pretation of the matter. Yet selection continued to be effec- tive after ten generations of inbreeding, at which time it would be expected that a homozygotic strain would have been reached, if the character were dependent on typical Mendelian units. In many other cases abnormalities have been shown to follow aberrant types of heredity; to attempt a general review of the matter here would lead too far. INHERITANCE IN ABNORMALITIES 449 VIII. LITERATURE BauBiaAnt, E. G. 18938 Nouvelles recherches expérimentales sur la mérotomie des Infusoires ciliés. Ann. Microgr., Paris, 4. Bateson, W. 1914 The Address of the President of the British Association for the Advancement of Science. Science, 40. Cauxins, G. N. 1911 Effects produced by cutting Paramecium cells. Biol. Bull., 21. and Cutt, 8. W. 1907 The Conjugation of Paramecium aurelia (cau- datum). Arch. f. Protistenk., 10. and Gregory, L. H. 1918 Variations in the progeny of a single exconjugant of Paramecium caudatum. Jour. Exp. Zodl., 15. Dexter, J. S. 1914 The analysis of a case of continuous variation in Droso- phila by a study of its linkage relations. Amer. Nat., 48. DosBeEtu, C. 1914 A commentary on the genetics of the ciliate Protozoa. Jour. of Gen., 4. Hertwic, R. 1908 Ueber neue Probleme der Zellenlehre. Arch.f. Zellforsch., 1. JenninGs, H. 8S. 1908 Heredity, variation, and evolution in Protozoa. I. The fate of new structural characters in Paramecium, etc. Jour. Exp. Zool., 5. 1913 The effect of conjugation in Paramecium. Jour. Exp. Zodl., 14. and Lasuiny, K. 8. 1914 Biparental inheritance and the question of sexuality in Paramecium. Jour. Exp. Zodl., 14. 1913 a Biparental inheritance of size in Paramecium. Jour. Exp. Zool., 15. Jotuos, V. 1913 Experimentelle Untersuchungen an Infusorien. (Vorl. Mit- teil.) Biol. Centralb., 33. 1914 Variabilitit und Vererbung bei Mikroorganismen. Zeitsch. f. Ind. Abst., 12. Lewin, K. R. 1910 Nuclear relations of Paramecium caudatum during the sexual period. Proce. Cambridge Phil. Soc., 16. Lutz, F. E. 1911 Experiments with Drosophila ampelophila concerning evolu- tion. Carnegie Inst. Pub., 148. Matt, F. P. 1908 A study of the causes underlying the origin of human monsters. Jour. of Morph., 19. McCuEnpon, J. F. 1909 Protozoan studies. Journ. Exp. Zodl., 6. PEEBLES, F. 1912 Regeneration and regulation in Paramecium caudatum. BioleeBull 23: Poporr, M. 1908 Experimentelle Zellstudien. Arch. f. Zellforsch., 1. 1909 .Experimentelle Zellstudien II. Ueber die Zellgrosse, ihre Fixierung and Vererbung. Arch. f. Zellforsch., 3. Ravutman, H. 1909 Der Einfluss der Temperatur auf das Grossenverhiltnis des Protoplasmakorpers zum Kern. Arch. f. Zellforsch., 3. Stmmpson, J. Y. 1901 Observations on binary fission in the life history of Ciliata. Proc. Roy. Soc. Edinburgh, 23. Wooprurr, L. W. and ErpMann, R. 1914 A normal periodic reorganization process without cell fusion in Paramecium. Jour. Exp. Zodél., 17. ' Re Ae A Meera oben nt ta avers 5 oy bal ia ory ails vo ie ion ap ’ PNA Ra 1 Ane sical lea ay “ar file ae, ; os 7 y Ce eye mM, Figen! POLIS aaa a iia ‘4 ya is a ay’ Coe BS r} v ‘ ‘ a Lee AA rt hae fall Alii? € - HNch RNa P Krave) A \3 +) Pe a 0 ; ‘ ne V\ : L@ spare ‘ 7X ty if ? sal Wy) if ‘ae | . . ' : = ot aw “ s ¢ 4 f } + 2 4 ' i, + j HERITABLE VARIATIONS AND THE RESULTS OF SELECTION IN THE FISSION RATE OF STYLONYCHIA PUSTULATA AUSTIN RALPH MIDDLETON From the Zoélogical Laboratory of the Johns Hopkins University SEVENTEEN FIGURES CONTENTS eG OGG DIO Mee et ss eh «SS knee takai ase rieaceneiey oes oe EE eae See 451 BIVE CHNINI CULE ees eee ton Sess. cas ov) soe MA Maes toro acoeny a oded, coon le ey as ae 454 Experiments to test the effect of selection on the fission rate within a SIMP1EKClOMES. 45 05a. ao hee ee be Soka De eo eect se 456 1. First series: Long continued opposite selection, followed by balanced selection, mass culture, and reversed selection....................- 456 2. Second series: Repetition among progeny of a single individual from SETS ple ses), 5k vccnisvalke me Nearest alls eck TS AUS A ae ERE ERS ee oo 487 3. Third series: Repetition among progeny of single individual not related to those of series 1 and 2; also, effects of conjugation.............. 488 Discussionsandeconclustonsseweer ee esc. see err 497 SS UHMNATAAER TEV Wp SSA, cs lch SAI ces, cove) = saab nS a, Se 501 GTS Grote MEST GUE Ey... doa meatus Su eerae oe aise F< es SE Oe a eee 502 INTRODUCTION Organisms multiplying without admixture of two parents that differ in hereditary constitution have been found remarkably constant in their inherited characteristics. Most recent work agrees that in such uniparental reproduction inherited varia- tions occur rarely or not at all, and that selection has practically no effect in altering racial characteristics. ‘These results have given origin to the concept of the Genotype (Johannsen), as a designation for the permanent heritable constitution of the race. In view of the importance of these relations for the problem of the method of evolution, much further study of this matter is required. The present paper deals with the inheritance of varia- ; 451 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 452 AUSTIN RALPH MIDDLETON tions and the effect of selection in the case of a most delicately poised and readily modifiable physiological character, the rate of fission of an infusorian multiplying without conjugation. Most studies of the fission rate in infusoria have consisted in observations on the rate in given species, for descriptive pur- poses; or in examinations of the direct effect of changed physical and chemical conditions; or in the study of the effect of conju- gation on the rate of multiplication. Review of these studies is not demanded here. A certain number of observations have been made which bear on the inheritance of the fission rate and on the question whether there are differences among the progeny of a single individual in respect to these. The great papers of Maupas (’88 and ’89) contain some of the most important of these. Maupas made extensive studies of the effects of temperature and of other con- ditions on the fission rate in many species of infusoria; among the.rest in Stylonychia pustulata, the subject of the present study. Maupas became convinced, as a result of his studies, that under given conditions the rate of fission is uniform in allthe progeny of a given individual; that inherited variations in the rate do not occur in such a ‘pure race’ or to use the modern term, within a single clone. On this matter the conclusions of Maupas are in complete harmony with those of the ‘pure line’ workers and upholders of the constancy of the genotype, in more recent times. He gives detailed observations on the matter for a number of species, and resumes his results as follows: In all my cultures I have always seen all the normal descendants of the same ancestor, grow and multiply with the most perfect uni- formity. I have become convinced of the integral transmission of the faculty of development from one generation to another, and the most complete physiological equivalence must exist among all the normal individuals, produced in the successive generations (’88, p. 203). In long and numerous experiments on fifteen to twenty species, I have never observed anything which permits belief in the existence of morphological and physiological differences, not merely between the two products of a given fission, but even among all those which have descended from such a fission by regular and continuous generations CSS. elrAG)). FISSION RATE OF STYLONYCHIA PUSTULATA 453 The paper of Jennings on the Effect of Conjugation in Para- mecium (713) likewise deals to a certain extent with this matter. In a wild population many strains differing in rate of fission (under the same conditions) were found to occur. Furthermore, it was demonstrated that even in a population derived by fission from a single individual (that is, in a ‘pure strain’), conjugation produced inherited differences in the fission rate, so that after conjugation there were present strains showing constant differ- ences in these respects. © On the other hand, if no conjugation has occurred among the progeny of a single individual, the fission rate was found to be nearly or quite uniform. Jennings sums up as follows: It is found (1) that differences in rate of fission among those that have not conjugated since they were derived from a single parent are not inherited (unless possibly certain differences of a minimal character are to be excepted; differences of an order of magnitude far below those with which we are dealing); (2) that conjugation among the members of such a pure race does result in differentiations that are inherited (18, p. 366). The paper of Calkins and Gregory (713) on the other hand sets forth that there are in many cases differences in the fission rate among the four sets of progeny resulting from the first two divisions of an individual that has just conjugated. Other papers bearing less directly on this matter will be taken up in the discussion of the results of the present work. The specific problem When a single infusorian divides, often one of its two progeny again divides before the other does. In successive generations this same thing may be repeated. Thus, as shown in figure 1, one may have among the progeny of a single individual at a given moment, animals that are the products of four and others that are the products of two fissions. Hence there are differ- ences of fission rate among the descendants of a single indi- vidual—differences that afford the opportunity of selection with a frequency that made it appear worth while to determine whether these differences are inherited and whether slow lines 454 AUSTIN RALPH MIDDLETON and fast lines can, by selection, be isolated among the descend- ants of a single parent. I have carried out this work for Sty- lonychia pustulata. The animals were kept isolated and trans- ferred daily to fresh culture medium; for the fast lines the indi- viduals that divide first are uniformly selected; for the set to be developed into lines having a slow rate of fission the indi- viduals that divide last are taken. No attempts have ever been made heretofore to test the effect on the fission rate of selection among the progeny of a single individual. It appeared possible, though hardly probable, that such a physiological character might give results differing from those obtained from studies of the mainly structural characters hitherto examined. The work was undertaken at the sugges- tion of Prof. H. S. Jennings, to whom my sincere thanks are due for assistance throughout the work. The fundamental questions for examination are then as fol- lows: Can we, with respect to the character examined, get from a single genotype by selection two genotypes that differ charac- teristically from each other under identical conditions; and that retain these differences from generation to generation? Is selec- tion of small variations, such as appear within the pure strain or clone an effective evolutionary procedure? TECHNIQUE In any investigation of a physiological character which is so delicately respsnsive to all environmental changes as is the fission rate of infusoria, the statement of Calkins (02, p. 141)—- ‘“A correct method is the sine qua non of successful experiments with Protozoa;’’—applies with peculiar force. In order that re- sults may be of any value conditions must be uniform throughout. Jennings (713) has pointed out that to secure this uniformity the bacterial content must not vary. In addition to uniformity the technique used in work on the fission rate must insure the experimenter against the introduction of a ‘wild’ individual into the cultures and against the contamination of the lines inter se. These results have been secured by the adoption of the follow- FISSION RATE OF STYLONYCHIA PUSTULATA 455 ing method, which is a modification of that described by Jennings Gils): As culture medium, 7¢ of 1 per cent Horlick’s malted milk ‘was employed, a fresh supply being made daily. This is the medium adopted by Miss Peebles (’12). One gram of the malted milk powder was dissolved in a 100 cc. graduate in a few ce. of boiling spring water; this was then diluted to 100 cc. with more of the boiling spring water. Six and one-quarter cc. of this 1 per cent solution were next diluted to 100 ec. with boiling spring water and this 7s per cent solution was filtered and cooled. The animals were cultivated on ground glass slides having each two circular depressions capable of holding four or five drops of liquid apiece. These were kept in moist chambers. Three drops of culture medium were used in each depression and no cover-glasses were employed. The ‘fast’ lines were kept in the left concavities and the corresponding slow lines in the right concavities of the same slides, so that conditions were uniform for the two sets. Uniformity of bacterial content in the culture medium was secured by washing the animals in fresh culture medium before transferring to a new slide. The animals were allowed to swim about for a time in the fresh medium, in order to wash them- selves largely free of bacteria; they were then transferred to the definitive slide, in new fluid. The pipette used in transferring the individuals was invariably sterilized in boiling water after each transfer, thus absolutely preventing the unintentional introduc- tion of any individual which might cling to the pipette; there was thus no possibility of admixture of the ‘fast’ and ‘slow’ sets. The slides were labeled in lead pencil; the number of fissions and selections at each examination were likewise recorded on them, to be later transferred to permanent records. The individual lines were designated in accordance with the plan set forth by Jennings (’138). Each concavity contained characteristically two parent indi- viduals, the products of a single fission. The slides were ex- amined daily or oftener. When one of the two individuals was 456 AUSTIN RALPH MIDDLETON found to be divided, while the other was not, a selection was made (fig. 1). If the line is under selection for the production of a rapid rate of fission, the two progeny of the individual that has divided are transferred to the new slide, while the undivided - individual is rejected. In selection for a slow rate of fission, on the other hand, the individual not yet divided is selected for Fig. 1 Diagram of successive fissions among the progeny of a single indi- vidual, illustrating the variations in fission rate which were the subject of selec- tion in the present work. The left side of the figure traces a series of ‘fast’ selections, the right side a series of ‘slow’ selections, showing that at a given moment we may have, among the progeny of a single parent individuals of the fourth and of the second generation. propagating the stock. This process is continued: a ‘selection’ is made and counted whenever one of the two parents is found to have thus divided before the other. EXPERIMENTS TO TEST THE EFFECT OF SELECTION ON THE FISSION RATE WITHIN A SINGLE CLONE 1. The first serves of expervments Experiment 1, part 1. Direct selection in opposite directions, November 3 to December 8, 1913. On November 38, sixty animals of the sixth generation from a single individual which had been isolated from a ‘wild’ labora- FISSION RATE OF STYLONYCHIA PUSTULATA ADT tory culture of Stylonychia pustulata were isolated on ground glass slides, one animal to each concavity, thus dividing the sixty animals into two groups of thirty each; one group to be selected for a fast fission rate (‘plus selection’), the other for a slow one (‘minus selection’). For twenty-one days these thirty ‘fast’ and thirty ‘slow’ lines were propagated with frequent selection as described above. On the twenty-second day the fast and the slow lines were duplicated (by division) and on the twenty-third day the first forty of the resulting sixty fast lines and the corresponding slow lines were all duplicated so that now we had one hundred fast lines and one hundred slow lines, plus and minus selection continuing with all these. These were thus propagated until the end of the third ten-day period. The actual number of fissions per line and the actual number of selec- tions that were made during the three ten-day periods are shown for the first thirty lines of each set in juxtaposition in table 1. When the differences per ten days of the sixty lines are aver- aged it shows that on the average the ‘fast’ lines have pro- duced 2.03 generations more than the average for the ‘slow’ lines during the first ten-day period, 3.57 generations more in the second ten-day period and 2.40 generations more in the third ten-day period. Further, the fast lines have each averaged 2.67 generations more per ten days, during the whole thirty days, than the slow lines. In other words, this table shows that on the average each of these thirty fast lines has produced 0.267 generation more per day during this thirty-day period than has each slow line. Figure 2 is the curve of the difference between the daily averages of the two sets. It is clear that the direct effect of the selections made would be to produce a difference in favor of the ‘fast’ lines as long as plus and minus selection continues, even though the differences were not hereditary and were due purely to accidental causes. Our next test is therefore to determine whether any hereditary result has been produced; to do this, selection must cease, and we must determine whether the differences in rate still continue. The most obvious method would be to stop selecting and pick 9F 19 OFS 08 02 0L°S6 “g°é &P ST 00 61 £0°S €L LT 92°61 G@NIT Wad GDVUaAv up moys Ubis snurwu ay? Waarh saduasaffiq) ay) yj0q fo sour fqury, ysuy ay? Lof ‘porsad hvp-uaz wad aur) sad suoyoajas fo saqunu pun (suornsauab) suoissy fo saqunN OFS WH Ws ral 3 | FOOT 09 | 8$ | 95 | 9G | 4g | 2g GT ees | OLS |S eG air PFST 6¢ | 99 | €9 | 89 | $9 | 29 OL SE 20 St on or GL WG Pe Pe Pe 609 FZ | PG | GS | &% | ES | IZ BONG? We ee i ee 189 GZ | 9% | $2 | Ga | $2 | & NO eA ey ey Z0T 6 We I ip | £9F CEST OT SD She or PNG ANAS Ge ae OLg ST | 4T | 8f | &@ | 6T | 8ST | ty ee 19 G47? to 16 he | GES TZ | 6E | ST | 8T | 6T | 06 Wee ae We |e WP te £6¢ 9T | & | 02 | 02 | Te | Te 6 Wie Neh te IVLOL | 0€ | 6Z | 8Z | 26 | 9G | SZ 69 LT 02 FG &% &% CL cit 09 &T 8s a! 9g él 19 ai 0Z FG ne 8g I 06 &% &@ &% G9 él GS GG + Oo nN KS IT gg II Lg &1 [> i) nN L 9¢ 6 £9 reece see e eres aepe JO JOAB} UL SssIdXy *“*SuOT}eIBUOy) [BIO], *“***SU01}09T9G [BIO], :MOTS ““SUOIPBIOUOL) [BIOT, “***sm01409[9G [BIOL [48B iy rte eee cence ene nerae JO JOAVl UT ssoox— “ees "**"STOT} BIOMOL) FrsreeeeeseaorqaTeg :MOTS "ss" “STOTPBIOUSL) H+ seeesssss01}09]9g 24887 :poned ABp-0T PANEL Peete eee ce sees gegy jo JOAB] UT ssooxq sees" **“STOT} BIOMOL) ++ ssssssssaorqaaTag :MOT[S "ses" **"“SUOT} BIOMOL) P+ sseeesesss1701909] 39 14SByy :polleg ABp-9— puosag Fite e sees etees gaggy JO IOAB} UL ssdoxy “55 "**"SuOT}B1VUOx) H+ reessssgrq09 Tag ‘MOIS (i) Gl 10) CLENCLOCT) "“suorqoaag 48Bq :powoeg ABp-OT ISI ZT oT ST iat I al IT Or ‘yas qspf ayy fo Loans ur ssaoxa ay2 ypu fssasboud UL spn worjdajas azisoddo ajrym T HIaVL UAAWON SANIT (jas mops ay} fo ssaaxa sjas mo)s pup yspf [ 0d ‘J quamisadry 458 FISSION RATE OF STYLONYCHIA PUSTULATA 459 at random the animals for daily transfer to the fresh slides. But this method opened the possibility of the ‘personal equation’ becoming so important a factor that I rejected it. In place of it I devised what I have called the method of ‘balanced selec- tion.’ Balanced selection is the process of compensating for the effect of every selection that one is compelled to make by making the reverse selection at the next opportunity; in other words, one makes the same number of plus and minus selections in any given line during each successive time interval adopted. Thus, if on a given day one makes a plus selection, at the next oppor- 1.2 1 10 20 30 Fig. 2 Curve of the daily differences between the average number of gener- ations per line produced by each of the two sets of lines while under selection in opposite directions (Exp. 1, part 1). The ordinates show the average differences (in fractions of a generation) in favor of the fast selected lines; the abscissae are the days. tunity he would make a reverse, or minus selection. In order to be able to tell at a glance whether to make a plus or minus selection the character of the selection to be made was also recorded on the slide. In addition to carrying out balanced selection, in a portion of the lines reversed selection was practiced; that is, the fast lines were now subjected to minus selection, the slow lines to plus selection. Of course, if the observed difference in rate is heredi- tary and has resulted from selection, then there is no logical reason why the same result should not be obtained a second time, so that the difference would disappear, or eventually be 460 AUSTIN RALPH MIDDLETON replaced by a reversed difference. The experiment with bal- anced and reversed selection is that described next. Experiment 1-A Balanced selection and reversed selection, December 4 to December 23, 1913. This experiment was undertaken to test by the methods set forth above whether the difference in fission rate shown at the end of thirty days of opposite selection by the two sets of lines in Experiment 1, part 1, were hereditary. In the first place fifty of the slow lines and fifty of the fast lines were subjected TABLE 2 Experiment 1-A: Actual number of generations per ten lines per ten-day period for the 50 fast and 50 slow lines of Experiment 1, Part 1, when subjected to bal- anced selection AVERAGE | , VopacE NUMBER OF LINES mere ote sae ne Fae Reta: ee 10-DAy SSNS PERIOD First 10-day Period: Has telimess.- eee see elo alee GO meee 1600 43a 15252 iL GS Slowalintes sane ee 155 158 167 148 BS) || 17433 1.528 Differences in favor of the PASTIMES: yates As —3 2 —6 12 Uf 0.24 0.024 Second 10-day Period: Wastelinesneere ey wer se. 90 95 88 99 100 9.44 0.944 Slowealimes vale ere ce ee 88 89 87 97 86 8.94 0.894 Differences in favor of the fastwlaMes seers eee Oe 2; 6 1 2 14 0.50 0.050 to balanced selection for twenty days. Table 2 gives the actual number of generations that these fifty fast and slow lines pro- duced. The sets are divided into groups of ten lines each. The average difference of fission rate per line per day for the whole thirty days of Experiment 1, part 1, while selection was in progress was 0.267 generation in favor of the fast lines. For the twenty days of balanced selection it was 0.037 generation. This would seem to indicate that the direct selection of Experi- ment 1, part 1, had produced a very slight heritable difference in fission rate between the two sets, especially since the differ- ence 0.05 generation of the second ten-day period of balanced FISSION RATE OF STYLONYCHIA PUSTULATA 461 selection is considerably larger than for the first, 0.024 genera- tion; and further it suggests the conclusion that if direct selection were practiced long enough a greater difference might be established. TABLE 3 Experiment 1-B: Actual number of generations per ten lines per ten-day period for the fifty fast lines and fifty slow lines subjected to reversed selection (‘fast’ lines selected now for slow fission, ‘slow’ lines for fast fission) AVERAGE AVERAGE NUMBER OF LINES ene ene oe eae eae cee ee 10-Day TMS PERIOD First 10-day Period: Slow lines, plus selected: Selectrlonsmseeenise. ae 29 42 47 55 48 Generations... sa. Sele Sulla lA sen GGa ial S Om ieline2s 1h Fast lines, minus se- lected: Selectiomsheess 5.40. oo - 36 42 35 39 35 Generations... ....4- 147 160 150 137 145 14.78 1.48 Difference in favor of lines now plus selected. 10 27 24 29 35 2.50 0.25 Second 10-day Period: Slow lines, plus selected: Selections neice tanee e 24 28 23 38 25 Generations. . 96 108 88 110 96 9.96 1.00 Fast lines, minus se- lected: SelechiOnsayen stances 22 21 17 34 33 . Generationse. 2. so. 95 93 92 93 95 9.36 0.94 Difference in favor of lines now plus selected. 1 15 | .—4 17 1 0.60 0.06 Reversed selection. Table 3 gives the actual number of fis- sions of the fifty fast and slow lines under reversed selection. Again the lines of each set are divided into groups of ten each. Table 3 shows that reversed selection has made the average fission rate of the fifty slow lines faster than the average fission rate of the fifty fast lines. Now, as pointed out above, if the observed difference in fission rate is hereditary and has resulted from selection this result might logically be expected to follow from reversed selection. The net result of both phases of this 462 AUSTIN RALPH MIDDLETON Experiment 1-A, while favorable to the production of an heredi- tary effect through selection, was to show that opposite selec- tion must continue for a much longer period before the nature of the result can be established beyond controversy. Experiment 1, part 2. Continued opposite selection, Decem- ber 4 to December 23, 1913. While Experiment 1-A was in progress certain fast and slow lines of Experiment 1, part 1, were kept under direct selection as a precaution against the possibility that selection in that experiment had not yet produced heritable differences in the fission rate. For this purpose fast lines 6, 30, 48, 45, 60, 73, 76, 78 and 100 and slow lines 3, 4, 9, 41, 53, 71 and 85 were chosen. During the first of these two ten-day periods the ani- mals were transferred to fresh slides every twenty-four hours and during the second ten-day period, every forty-eight hours. The actual number of generations produced by each of these lines, during the two ten-day periods is shown in table 4. TABLE 4 Experiment 1, Part 2: Actual number of generations and selections per line per ten-day period, with the excess of generations in favor of the ‘fast.’ Direct selec- tion of the fourth and fifth ten-day periods of Experiment 1. LINES NUMBER sae 4 2 aa re e 3 a 78 100 AVERAGES omitted in aver- Fourth 10-day Period: ey Fast: Selectionsteeesee ieee also 6| 4 AS A ee 2, 3 (CENA MOM. oscecns se alh UE |) WG.) WO WS Nez | XO) Nee i SS | aS 17.14 Slow: SAGO .ccccscacaesccall 2 5 3 Gullo 3) ||) (CeMGEHMOIMNS sss5c55 seca WAIN. | 1G} WG |} TS I is |] TS 14.71 Excess in favor of ‘fast’...] 2 5 5 2/-1 5 |-1 2.43 Fifth 10-day Period: Fast: Selections apa tae |e 3 1 3 3 3 Dy 3 2 Generations .555 eee |e ono OF PO | S10 8 9 9 9 9 .28 Slow: : SESCHOMN.ceosccccccevesl| 2 2 2 2 2 2 3 G@enerationsss. 9454.40. 7 AS ed Hi] “2 Hi & Hl Excess in favor of ‘fast’...| 2 6 2 2 6 3 iI 3) oh FISSION RATE OF STYLONYCHIA PUSTULATA 463 This experiment is simply a connecting link between part one and part three of Experiment 1. There is a slight increase in the difference between the average number of generations produced by the fast and slow lines during the first and second ten-day periods. The fourth and fifth ten-day periods of figure 25 1 10 30 50 70 90 110 130 Fig. 3 Polygon of the average number of generations produced, per line, by the fast and slow selected sets of lines of Experiment 1, while the two were undergoing selection in opposite directions; averaged for ten-day periods. The polygon shows this average for each of all the 13 consecutive ten-day periods of Experiment 1. The continuous line shows the average for the fast lines, the broken one the average for the slow lines. The ordinates give the average number of generations produced and the abscissae the number of days since the beginning of the experiment—the successive ten-day periods. 3 show the average number of generations per line per ten-day period for this part of the first experiment and its first three ten-day periods show it for the first part; while figure 4 is the curve of the daily differences between the average number of generations produced by each fast and each slow line during these two ten-day periods. 464 AUSTIN RALPH MIDDLETON Fig. 4 Curve of the daily differences between the average number of gen- erations per line produced by the fast selected.set and the slow selected set during the fourth and fifth ten-day periods of opposite selection in Experiment 1, (i.e., Exp. 1, part 2.) The ordinates give the daily differences, the abscissae give the days. Experiment 1, part 3. Continued opposite selection, Decem- ber 24, 1913, to January 22, 1914. It will be noticed from table 1 and figure 3 that the differ- ences between the average number of generations produced by the fast lines and the slow lines during the three consecutive ten-day periods of Experiment 1, part 1, are 2.03, 3.57 and 2.40. That is, each fast line produced, on the average, 0.267 generation more per day than each slow line during these thirty days. Table 4 and figure 3 show that the corresponding differ- ences for the fourth and fifth ten-day periods of Experiment 1, were 2.48 and 3.57 which gives a daily average difference of 0.300 generation per line. This difference, in favor of the average fast line is slightly (0.033 generation), greater than the corresponding difference of Experiment 1, part 1 (0.267 generation). Does this mean that on the average, selection is gradually increasing this difference? Experiment 1-A indicates that opposite selection had produced a heritable difference in the average fission rate of the two sets of lines; may this herit- able difference be increased by further opposite selection? FISSION RATE OF STYLONYCHIA PUSTULATA 465 To answer this question the first seven of the fast lines of Experiment 1, part 2, were each increased (by division) to four and from the eighth two lines were derived, giving thirty fast lines in all. Each of the slow lines of Experiment 1, part 2, was likewise increased to four, with the exception of the last one which was increased to six, thus giving thirty slow lines. These two sets of lines were then selected for fast and slow rates of fission. During the three ten-day periods of this part of Experiment 1, the selection and transfer of animals to fresh slides was made every forty-eight hours instead of at the close of every twenty-four hour interval as heretofore. At the end of the first ten-day period the eight fast lines which had produced the highest number of generations were selected and thirty lines were derived from them, care being taken that each group of ten lines, the ‘fast’ occupants of a single moist chamber, were represented in the new set. Also eight of the slow lines which had produced the smallest number of generations were selected and from them thirty lines for continued slow selection were derived. The same precaution was taken in reference to the distribution of the selected lines among the three moist chambers. These two sets of lines were then selected for fast and slow fission rates during the second ten-day period. Finally this same method of reduplication of the fastest and the slowest lines and the subsequent continued selection of the individual variations of fission rate was followed through the third ten-day period. This method of double selection was devised with the purpose of crowding as much selection into the experiment as possible, in the attempt to determine whether continued selection would gradually increase the previously established difference between the average rates of fission of the two sets of lines. It will be remembered that the difference of average fission rates between the two sets of lines at the end of Experiment 1, part 2, was 3.57 generations per line per ten-day period as shown by the fifth ten-day period of figure 3 and by table 4. The sixth seventh and eighth ten-day periods of figure 3 show those differ’ ences for the three ten-day periods of the present part of Experi. 466 AUSTIN -RALPH MIDDLETON ment 1. For the first period of this part (sixth period of the entire experiment) the difference was 1.53 generations per line per ten-day period, for the next it was 3.41 generations and for the last it was 7.51 generations. The small difference of 1.53 between fast and slow in period 6, following upon a period when the difference was 3.57, is due to a slowing of the fission rate in all lines during period 6, owing probably to low temperature, and perhaps partly also to one of the rhythms emphasized by Woodruff and his colleagues. The percentage of difference in proportion to the total average fission rate is in reality greater in period 6 than in preceding periods. Thus, for period 4 the average number of fissions for all sets was 15.925, and the differ-, ence between the fast and slow was 2.43—this difference being thus 15.25 per cent of the average rate. In period 6, the small difference 1.53 is actually 27.34 per cent of the average rate for all. Table 5 gives the actual number of generations pro- duced by each of the lines during the three consecutive ten-day periods, the number of selections that were made in each line and the average difference per line for each ten-day period. Table 5 and figure 3 demonstrate in this experiment a con- tinued increase of the difference between the average rates of fission of the two sets of lines. As has each of the previous experiments, so also does this one show that the fission rate within the clone may be changed by selection. During its three ten-day periods each fast line produced on the average 0.415 generations more per day than each slow one. On only one day during the thirty days of this experiment did the slow lines produce more generations than the fast lines and on that day only an average of 0.3 generation per line, as shown by figure 5. The two sets indeed hardly overlap at all in their rates, practically all of the fast set being faster than any of the slow set. This is shown by the curves of variation of the two sets in figure 6. Only one slow line produced as many generations as the slowest fast line. FISSION RATE OF STYLONYCHIA PUSTULATA 467 TABLE 5 Experiment 1, Part 3: Actual number of generations and of selections per ten-day period per line of the third part of Experiment 1, that is, the sixth, seventh and eighth ten-day periods of continuous opposite selection, with the excess in genera- tions produced in favor of the fast-selected set AVERAGE PER LINES NUMBER 1 ro 10 11 To 20 21 ro 30 TOTAL LINE PER TEN-DAYS Sixth 10-day Period: Fast: SACHIOING ce oo benouas 13 14 14 Generawonsesee sates as. 62 65 64 191 6.36 Slow: Selections jana ssc: eens 5 8 11 Genenaiionsse ces.) le. 44 48 53 145 4.83 Excess in generations in favor of the fast........ 18 17 11 46 11 583 Seventh 10-day Period: Fast: Selechiomsearees a... 28 22 21 Generations... ...... 25. 121 101 104 326 10.87 Slow: . SEISCHOMS, .oodeseubeune 23 25 23 Generations. .s..5-..4: 69 88 67 224 7.46 Excess in generations in favarotuhetast.......- 52 3 37 102 3.41 Eighth 10-day Period: aes Fast: . Sclecilonssesnen oa ee Bi) 42 4] Generations. .........- 168 222 236 626 20 .87 Slow: Selectionssys.........-% 31 31 36 Generations se. 4. 132 133 134 399 13.30 Excess in generations in Haviorotuheniast.. 4... .- 36 89 102 227 7.56 Fast: Movaletors0\days......- 351 388 404 1143 38 .10 Slow: MotaleroroONdayise acu. 4 - 247 271 254 72 25.73 Miiienemcesstees se. . ser 104 117 150 371 WABI Experiment 1-B. Balanced selection, January 23 to April 22, 1914. At the close of Experiment 1, part 3, the two sets of lines of that experiment had been under continuous opposite selection THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 468 AUSTIN RALPH MIDDLETON for eighty days, and the average difference per line per day had increased from 0.267 generation for the first thirty days to 0.415 generation for the last thirty days of that period. To test the permanence of this apparent effect of selection, to determine the answer of our experiments to the question ‘‘Can we get from a single genotype by selection two genotypes that differ charac- 30 Fig. 5 Curve of the daily difference between the average number of gener- ations per line produced by the fast set and the slow set during the sixth, seventh and eighth ten-day periods of opposite selection in Experiment 1 (Exp. 1, part 3). The ordinates give the daily excess in favor of the fast-selected lines, the abscissae give the days. teristically from each other under identical conditions; and that retain these differences from generation to generation?’’—the lines were now subjected to the test of balanced selection. For this purpose all the thirty fast and thirty slow lines which were in progress at the close of Experiment 1, part 3, were continued. In order to make the test thorough it was prolonged through nine consecutive ten-day periods, or ten-days longer than the FISSION RATE OF STYLONYCHIA PUSTULATA 469 lines had been subjected to opposite selection. During this whole experiment the animals were transferred to fresh slides daily as described at the beginning of this paper. The results of Experiment 1-B, are so important that they are set forth in some detail in table 6. This table gives the actual number of generations produced by each fast and each 20 20 25 30 35 40 45 Fig. 6 Curves of variation in actual number of fissions of the two sets of lines of Experiment 1, Part 3. The ordinates give the number of lines, the abscissae the number of generations produced during the thirty days. The con- tinuous line is the curve of the fast-selected set, the broken one is that of the slow-selected set. slow line during each ten-day period of the experiment, and the difference between them. This difference is given as a positive quantity when the fast line has produced more generations than the corresponding slow one and negative when the reverse is the case. The latter occurred only thirty-one times among the whole two hundred and seventy differences. In twelve of the lines it did not occur at all; in nine lines, once; in six lines, twice; EERE DICE CY iC IT | &1 OL | GE | Zt | FI |Stoereuer ‘Mog GOTT Gps PLVOLHMOT 926 NSE INET GD -@ MnOkaeckaeGee| TES letra! PU 6: \eGe Gs | Chalets Ge OR OTA eh MOTT a\nO TOIT, | pial PuOMeie ues eH :poried ABP-OT WIIG = o =) _— oO ioe) a oO ive) nN oOo bl oD = sH ba) oD ben! nN ~ ss i>] Ven) —) Oo hen! o oD of nN ba) Ten) ~ oO of nN | = oD “4 | o ~ ~ loo) oD co N [33 — s o 4 sal in} > a> So oOo on co oO Oo eo lor) o o o ol oa Sol el Seal lor) oo 5 oe —_ _ nN — 19ST 96°% 68 e lel tesle dile ho |e Te |e Ihe 16 ASIC Wie VR OI WES GG ee pte pe ts Pe |e SE ee ean 00°8 om Nile We eel de AG We We Pe We Se ER ae Ss eee |i cok i GI aan PIE IP | GL | GL |Suorzerouey) MO]S OGY OTe MGC eme eral Gian Gie ROTen Gla ane Ten Sety DEMO MG) s(On a SSp mi NGuaiiciln | Zea ecis|MOTe GIy Sis Soe eee site €L | 6 |) SE || Gh) Th |StoneioueS) asa ported ep-O1 WA eh IT 981 9¢ G té ¥ § ¢ T T F G € T Gala G Vala iS) 9 g T F 0 F G 0 ¢ I 48 G | le fee coe NS GOL 91G 6 9 ¢ L ¢ 8 9 ¢ F ¢ L 6 L ¢ Or | 9 M ¢ 17 8 8 HELE HF 4) L Tr | 8 OLR: OL | OF |SUOTPBIOUOS) “MOTS 90°6 GLE IT | 6 6 OT | OT | 6 L 6 9 8 8 ML; 8 L 9 8 9 Or | 6 6 GE | IL | OL | 6 | ASE OL | GE | 6 |Suor;ereUEy “48R,q :poledg ABp-0T YWNOY €0 IT LG°T LP € id € G F F G € F ts (SH 6 4 0 0 0 I 0 T G—| G 14 G 6—| G 0 0 I G Bie see €8°9 06T F F ¢ tS 9 ¢ 9 ¢ F 9 8 M F i 8 9 L ¢ 8 L 8 9 8 8 or | 9 L 8 6 g |Sudiye1euer ‘Mo[g 06°2 LEG M 8 8 8 Or | 6 8 8 8 6 9 9 9 9 8 9 L 9 8 8 9 8 GL | OL | 8 8 L 8 OL | OL |Suorzereuer) “4yseq :poreg Aep-OT pany co FI L191 0g ¢ § 0 € G § G F G G 0 T 4 T i T ets o—| 0 T ( G 0 G § 0 € TS 0 cape amen Se Oo 06°F LPT I ¢ ¢ € § G 9 ¥ ¢ ¢ ¢ ¢ € ¢ ¢ 9 ¢ ¢ g L ¢ ¢ ¢ 9 ¥ ¢ 9 ¢ 8 8 sie ae) MOTS 5 1¢°9 161 9 8 ¢ 9 ¢ ¢ 8 8 it L ¢ 9 Uf 9 6 L $ 8 § L 9 8 L 9 9 8 9 8 L § |Suorzereuer) ‘4SBiT ‘poled ABP-)] PUodDg 16° &1 FI'& $6 § ¢ § G F 17 v 0 I € € 8 0 alO) 9 6 G UWS 9 ¢ g I € i} () G | °° °° -e0medayIq It | IL | FE | IL | PE | ST | OL | @T | ST |Suorywremy ‘MoTS LT} 70 | at |6 |6 | ey | 2E | ot | ot | et | zr | Gr | OF | ST | St | ST | 12 | AT jSuonessuSh “seq :poreg ABp-OT ISIN 9°01 | 808 | 6 OF eT | GOP | CT mon oa for) co o > oO a _ ol co oo = ON © i> -— i=) bol a ~ ioe) nN - oD nl Gov -UGAV UaAAWON SANIT nN — Tvnou| og | 6z | 82 | Le-| 9% || $2 | #2. | 8s | ee | 12 | 02 | GE | SE | 20 | OT | Sk | Fr | St | Gk} Tr} Ory) 6 | 8 | 4 19 1a |r Te GHLIVHLINGO udd TIV NI 'TIVLOL AHL 49 SI GONGTUAAAIA 3 ‘osjp waarb aup saur ysvf ay] fo Lo.vf ur suorssy fo waqunu Ur sanuasaffip YJ, ‘[ wuawuisaday fo sjund asayz fo worjoajas joaL1p fo shvp fizylra ay} panojjof hjaypipamwua yoryn worjoajas paounjpg fo sporsad fipp-wa} anvnoasuoo auvu ay) 1of ‘¢ pup @ ‘] sjind ‘7 quawmisadagy fo our) wad poriad fivp-ua} sad suoynsouab fo vaqunu yonjay ig-] quaurwdxgy 9 WIAVL ‘auT] Jod uolze1oues [¢°Z SI ‘JuoUTIedxo STy} Sutnp porsied Aep-u9} J9d ‘4svyz oY} JO IOAV] UT DOUDIOTJIP OBVIOAV OUT, ‘porsed Avp-ue} JSIY OY} BULNP YT “ON UT] MOJS Oj OS¥O OY} ST SITY} ‘ported Avp-ueq oy} Sutmnp ouo ysvj Sutrpuodsesi09 oy} uvy} suoTyei9UEs oIOWI poonp -o1d SvYy SUT] MO]S dy} OATZVSOU ST UUINOO sdUaIOYIP oY} UT JoqUINU oY} UOT A, 68 &T Ih €1 €S 41 &8 £8" 91° 09° £o £0" “GG aon eo 8L9 G8 896 SFE 18 18é 896 8 L¥G 1&é 66 II a £& ites) 9¢ x ST If t' ™~ oO SZ aa 9€ oOo 68 oO SG co 1d sH (a9 o sH s GE XN oO oD ao ad ST arn bYon} arn ‘T (ON OUT] MO]S UBYZ SUOT}BIOUNS OIOW Z poonpoid ‘T ‘ON our] 4SsVj poliod Avp-us4 ysiy oy} Sulmnp snyy +900 MOJS Sutpuodsed109 oY} ULY} SUOT}eIOUEs e1oOW AUBUT OS poonpoid svYy UOT4SENb UT oUT] SBI OY} FLY} OYwoIpUT OATJISOd VOYA SUUUNOD ,SOUOIOYIP, OY} UI SIOGUINU OY} 9[qQvI SIYy UT GG Oo id HH Or a! ~ oO 06 Or II Or NOOO II oO 29 FI &I Or al 0 Or Or Or I lor) 1 9 i FI I IT IT FI Or &I rae IT oT II &T 6P as ST a &T PI Il 8 FI FI €T al iat ZT a a oT Or 1 Or 1 ZT i) IT él Or IT I “++ SARP-06 ITO AA oY} 10} sooudIE IC. “+++ -9guogogTq. SuOT}BIOUVY) ‘MOTG SUOTPBIOUOL) ‘YSBiT :poleg Aup-01 UUIN “ss *-9QuTaIO UIC SUOT}BIOUIY) ‘MOTS SUOTPVIOUOL) ‘4SBIT : poled Aep-O1 Wyse 9s gouera NT SUOI}BVIOUOY) ‘MOTS SUOT}EIOUIL) ‘4SBIT : polled ABp-0T YPUBAVG 471 472 AUSTIN RALPH MIDDLETON in two lines, three times; and in one line four times. When the differences between the number of generations produced by the corresponding lines during the whole ninety days were ascer- tained it was found that in not a single case had the slow line produced more generations than its fast one. And in only two lines was the excess of the fast over the slow probably too small to be significant (fast line ten produced only five more gener- ations than slow line ten, and the two lines number fourteen produced the same number of generations). Furthermore dur- ing each ten-day period the total number of generations produced by all the thirty fast lines was much larger than that produced by the slow lines, and the average number of generations per line per ten-day period was uniformly greater for the fast than for the slow lines. Also the per cent of the difference in proportion to the total number of generations produced by both sets was calcu- lated and found to be remarkably uniform; these percentages are shown in the extreme right hand column of table 6. Finally, the average number of generations per line per ten-day period for each set of lines is plotted as a polygon in figure’7-a, giving a graphic representation of that phase of table 6. Figure 8 gives the curve of the daily differences between the average number of generations produced by each fast and each slow line. On only three days during the whole ninety days of this balanced selection experiment was this difference too small to be significant (on the eighth and sixteenth of February, 1914, both sets of lines produced the same number of generations and on February 18, 1914, the difference was 0.03 in favor of the slow set). The average per day for each fast line, including the three instances just cited, was 0.251 generation greater than each slow line. From table 6 and from figures 7 and 8 it is there- fore evident that, when measured by the test of balanced selec- tion, the eighty days of opposite selection had produced a differ- ence of fission rate between the two sets of lines that is heritable. Furthermore, analysis of the daily records shows that this result is not due to the chance isolation of a ‘mutating’ line in either set of lines and the subsequent development of the thirty lines of FISSION RATE OF STYLONYCHIA PUSTULATA 473 the set from that one, for we have four lines of the fast set that run all the way through and three of the slow lines that do the same. ‘Table 7 shows the total number of generations produced by each line of both sets. For the fast set these totals vary from 30 60 90 B Fig. 7-a Polygons of the average number of generations per line per ten-day period produced by the fast and slow sets of lines of Experiment 1-B during its nine consecutive ten-day periods of balanced selection. The continuous line shows the averages for the fast set, the broken line the averages for the slow set. The ordinates show the differences of the averages, the abscissae the ten- day periods. Fig. 7-b Curve of the difference between the average number of generations per line per ten-day period produced by the fast and slow sets of lines of Experi- ment 1-B during its nine consecutive ten-day periods of balanced selection. The ordinates show the differences of the averages in favor of the fast set, the abscissae the consecutive ten-day periods. “Q-) OINSY OATS ‘sporszod Aup-Uo} 10] poBvs9AV ‘OAM STY} JO SayVUIpPIO OY, “SAUp AJOUTU OY} OAV OVSSTOSE OY} /4OS SBF OY} JO OAV} UT SOBVIOAT A[rep oy} JO soouarayIp oY} MOYS SoVVUIPAO oY, “WOIdeJos poouL[eq jo sXvp Ayouru syt Suldnp q-] JuouMIedxy JO soul] Jo szos MO[S puv ysvy oy} Aq poonpoad our] dod suo1zeVA9UED JO aquINU oSeIOAG oY} WOOMJo OdUEJEyIP ATIVp oy} JO OAIND 8 “ST 06 08 OZ 08 0g LP ond g 89 LIIIS FASE SZ\GS BITS LZ|8F LT|L0 S|9F 82/99 OZ|FS FA\ES 1S|F6 OT\EL LT\SE FLITS G\es 8 99°SI|€9°6|S6 LIES SIGh S|ZE LZ|S8 6ISh GiZ8°8|19 2/20 ST\SP 8)99°9)° °° 18909 Jo St Sse a -xo yey} Juod Jog “tH 66 | S& | 9€ ST Ty | 9% | 98 | 68 | Ge | c | ce | ST Go | 0G |OL |Fr |0 9G |9T |rr |G |9T 6h |1@ |ct st CHT (Xa PALE EN ae eS nS! JO LOA} UT SSeOXG FOL | get | cpr lzct | GFE | EFL | 9ST | ZEL | Tek | SET | OST | 2ET | THT | SEL |ZLT JOLT J9FT | 99T J99T JOLT |L6T JOGT | GLE JETS jOGZ [FOS |LOT | E6T |10G |L6T |------- 69 j0S j\&S Deel eee Seen OOe RO Paen |S e OSes 6g 19 | 89 | 6G |T8 |82 j€4 | 02 | ““sTeqOy, 18 G9 |96 |POT \€6 |T6 FAN Mri [IGe PORE Pe OR OOOO OSS ENE 06 UL SUOT}BIOUD jo sequinu ‘MOS 86 €8 68 |98 &6 $8 96 88 &Z 68 16 OL €8 6Z |16 |26 |&2 96 |16 |$6 |LOL |SOL | PET JZIT |9TT |TTT |900 | TET \60r |For °° °° SABp 06 UI SUOT}eIIU0S jo sgoquinu ‘sR ~ — [2] co lor) Of | 60 | 8¢ | 26) 9% | SG | Fo "ES | co | Te | 02 | GT ST ZT 9 | ST | FT | &T Gils | SE OE 6 SA ee Wz ts Wt |) UaANWON SANIT suns ay) [0 au saouasaffip ay) oY? quao1ad ay} pup ‘au2r) nojs burpuodsalsod 872 PUD AU2} spf yona fig paonpoud suoynwauab fo waqunu ayy fo wns ay) OS)P *aU27 Mo{s Burpuodsastoo szv pun aur ysvf yoda fig paonpoud suornsauab fo saquau ay, Waanjaq aouasaffip ay} yyIm “Guauisadxa ay) fo shop fyauru ay) lof uoysajas paouvjpg Burunp aur) mops pun yspf yova fq paonpoud suoynsauab fo Laqunu 70}0} 94], :{@-] wauwaday 4 ATavL FISSION RATE OF STYLONYCHIA PUSTULATA A475 seventy-three to one hundred and seventeen and for the slow set from forty-eight to ninety-six with the exception of one line which produced one hundred and four generations. The first ten lines of both sets each produced a much larger number of generations than the remainder of the lines of its own set. These lines were the occupants of a single moist chamber, so that there was probably some environmental difference distinguishing this moist chamber from the other two. That environmental differences have a marked effect on the fission rate of infusoria has of course been shown by many writers. Figure 9-a gives the curves of variation in number of genera- tions produced by each fast and each slow line during balanced selection. The curve for the slow lines shows an apparent . bimodality. To determine whether this might be attributable to the environmental difference suggested by table 7 these same variation curves were plotted for the first ten lines of each set alone. These curves are shown as figure 9-b; the curves are mutually exclusive. Table 7 shows that only one slow line among these ten produced more generations than the slowest one of these first ten fast lines. Figure 9-c gives the curves of variation for fast and slow lines 11 to 30, the occupants of the other two moist chambers. Only four of these twenty slow lines produced more generations during the ninety days of balanced selection than the slowest fast line. It must be borne in mind in this connection that these “‘two sets of thirty lines” are parts of the same clone. That is, we have got here, from a single genotype by selection two genotypes that differ characteristically from each other under identical condi- tions; and that retain these differences from generation to generation. Experiment 1, part 4.. Further opposite selection, January 23 to March 14, 1914. While Experiment 1-B was in progress the original lines were continued under direct selection for two reasons: 1) As a pre- caution against the possible failure of the difference of fission 476 AUSTIN RALPH MIDDLETON 65 Fig. 9-a. Curves of variation of the lines of Experiment 1-B, the ninety- days of balanced selection. The ordinates give the number of lines, the abscissae the number of generations produced. The continuous line is the curve of the fast set, the broken line that of the slow set. Fig. 9-b Curves of variation of the first ten fast and first ten slow lines of Experiment 1-B (balanced selection). The ordinates represent numbers of lines, the abscissae, generations. The continuous line is the curve of the fast set, the broken one the curve of the slow set. Fig. 9-c Curves of variation of the last twenty fast and last twenty slow lines of Experiment 1-B (balanced selection). The ordinates represent numbers of lines, the abscissae, generations. The continuous line is the curve of the fast set, the broken one the curve of the slow set. rate to survive long continued balanced selection (Experiment 1-B); 2) To discover what would be the effect if selection con- tinued further on these lines. During the first three ten-day periods there was no further reduplication of extreme lines and the slides were changed daily. At the end of the third ten-day period the fastest lines of the fast set and the slowest lines of the slow set were again chosen for reduplication and continued selection, as described for Experiment 1, part 3. This was the only time this was done during the present experiment. The slides were changed every forty-eight hours during the fourth and fifth ten-day periods. The ninth to the thirteenth ten-day FISSION RATE OF STYLONYCHIA PUSTULATA 477 periods of figure 3 (which gives the average fission rates of the two sets of lines of the entire first experiment for the one hun- dred and thirty days they were selected), give the average fission rates of the lines of this part of Experiment 1. Table 8 gives the actual number of generations per 30 lines as well as the average number of selections and generations per line that occurred and the differences in the average fission rates. It has been pointed out that, on the average, for the first three ten-day periods of Experiment 1 each fast line produced TABLE 8 Experiment 1, Part 4: Actual and average number of generations and of selections per 80 lines per ten-day period during the ninth to the thirteenth ten-day periods of continuous opposite selection of the lines of Experiment 1 AVERAGE TEN-DAY PERIODS NINTH | TENTH | ELEVENTH TWELFTH |THIRTEENTH TOTAL PER LINE Fast lines: Average num- ber of selec- tions per line.| 2.57 | 2.47 SLES 1.63 2.30 Total number of generations} 341 | 230 206 250 279 1306 - 43.53 Average num- ber of gener- ations per line|11.37 | 7.67 6.87 8.33 9.30 Slow lines: Average num- ber of selec- tions per line.) 1.57 | 2.33 1.36 1.33 1.97 Total number of generations} 251 | 139 121 149 189 849 28 .30 Average num- ber of gener- ations per line} 8.37 | 4.63 4.03 4.96 6.30 Actual excess of generations in favor of fast IME Sey rhs hes 90 91 85 101 90 457 15.23 Average excess per line in favor of fast P TINGS re as ae) 3.00 | 3.04 2.84 3.37 3.00 3.05 478 AUSTIN RALPH MIDDLETON 0.267 generation more per day than each slow line. During the fourth and fifth ten-day periods each fast line produced, on the average 0.300 generation more per day than each slow line. During the sixth, seventh and eighth ten-day periods this daily average difference per line was 0.415 generation. Table 8 shows that during the ninth, tenth, eleventh, twelfth and thirteenth ten-day periods it was 0.305 generation, which is considerably smaller than the 0.415 generation difference of the sixth, seventh and eighth ten-day periods of Experiment 1. But, as in a pre- vious case, this is due merely to the fact that the average fission rate for all lines has decreased; relative to this average fission rate the difference between fast and slow lines has not decreased, but on the contrary has increased. For part 1 of Experiment 1, the difference between the fast and the slow lines in number of generations produced was 6.9 per cent of the total number of generations produced by all; for part 2 it was 12.8 per cent; for part 3, 19.3 per cent, and for part 4 it was 21.2 per cent. Conse- quently, throughout the entire period of selection (thirteen ten- day periods), the proportional difference between fast and slow lines has steadily increased. Figure 10 further emphasizes the genuineness of the differ- ence of fission rate between these two sets of lines for Experi- ment 1, part 4. It shows the curves of variation of these two sets of lines of that experiment. Experiment 1-C. Mass culture and balanced selection, March 17 to May 4, 1914. In order to demonstrate as conclusively as possible whether the apparent average difference of fission rate between the two sets of lines was hereditary or not, it was decided to subject them to a period of mass culture and then to balanced selection. This was to determine whether the average difference of fission rate had survived the mass culture treatment. On March 4, 1914, all the animals remaining in the ‘fast’ concavities after the transfer of the chosen individuals had been made to fresh slides were placed, unwashed in a single circular glass dish 32 inches in diameter and 2 inches deep in 50 ce. of #4; per cent _ FISSION RATE OF STYLONYCHIA PUSTULATA 479 Horlick’s malted milk; and all those remaining in the ‘slow’ concavities were similarly treated. These two mass cultures were placed side by side on the laboratory table and allowed to propagate for twelve days. Every three days 25 ce. of boiled and cooled spring water was added to each to compensate for evaporation. The animals placed in these mass cultures were taken from Experiment 1-B, and hence had been subjected to opposite selection for eighty days and balanced selection for 18 28 35 45 55 Fig. 10 Curves of variation of the lines of Experiment 1, Part 4 (continuous selection). The ordinates give numbers of lines, the abscissae the generations produced during the fifty days of the experiment. The continuous line is the curve of variation of the fast set, the broken one the curve of the slow set. forty days; they were now allowed to remain in mass cultures for twelve days. On March 17, 1914, thirty individuals were taken from the mass culture of fast lines and isolated on slides and thirty indi- viduals were also isolated from the mass culture of slow lines. These sixty lines were then subjected to balanced selection for a period of fifty days, the transfer to fresh slides being made daily. 480 AUSTIN RALPH MIDDLETON In order to be sure of the uniformity of the bacterial con- tent of the slides, on March 18, an equal small quantity of the liquid from each of the mass cultures was added to the fresh medium. It was added two or three drops at a time and each few drops carefully studied in a watch glass under the binocular before it was added to the fresh culture medium. On March 20, each animal transferred to the fresh slides was washed in a watch glass in a mixture of equal quantities of the culture medium from the mass cultures instead of being washed in fresh culture medium. Now at the end of this fifty-day period of balanced selection the two sets had experienced eighty days of opposite selection immediately followed by one hundred and two days of no selection. If the difference remains after this test it is evident that selec- tion has in this case produced an heritable difference in fission rate within the clone. Table 9 shows that this difference of average fission rate does persist. Figure 11 shows the rate of division of the two sets of lines, averaged for ten-day periods, during this balanced selection test. Experiment 1-D. Reversed selection, April 13 to June 1, 1914. A second experiment in reversed selection was started on April 13, 1914, from lines derived from Experiment 1-B, the ninety- day balanced-selection experiment described above. Hence they had been subjected to eighty days of opposite selection and this was followed by eighty days of balanced selection before re- versed selection was started. This experiment was prolonged through five ten-day periods with daily transfer to fresh slides and during the whole time the average fission rate of the fast- selected ‘slow’ lines was higher than that of the slow-selected ‘fast’ lines. And on the whole there was a gradual increase of this average difference as reversed selection proceeded. Here again selection has altered the fission rate and the new rates were hereditary. Figure 12 and table 10 show graphically the results of this experiment. Attention is called in table 10 to the large average number of selections per line for the fifty days of this reversed selection. ee 1 e111 90°€T L8°T £°6 02° TT LT GOVuUaAV €—SSE C+18& G—S8z L+608 &—G8S €+0E€ 9-8 b—LOE 9+818 9—€98 IVLOL ‘soury JO Jaquinu oy} ‘og Aq (1¢g) JepuleUIel ST} SuIpIAtp Aq u0}403 SI GT] eDvIEAG 94} Pus peonpoid suorerieues Jo (ggg) JEquUINU [B04 94} WIOIJ pozxOwIGNS SI xIs a10j “TOG, SOUL] SVE OY} UL SUOTJOI[OS MO]S UY} PPCUI OIOM SUOTIOTOS 4SBy O1OUI xIS ‘potiod ABp-OT 4SIGy 94} SULINP 99ULISUT IOF JVY} SO}BOIPUT STOTJOOTEg sN{ding, poyeqv, uurnjoo oJ, for} o Oe GL |e | 1a! &1 for) o onl ao i=) ol oD el o MN) WE 1 Bl WE | ut or II for) — - Nr mm oD of al | al | cor) — ol a So hol ol bol a hon for) bol ho IT &T aon Sa Oye ee ST |\ST | of | ST |] 9 | CL | ST C5 Sc SW WO. te) B= it Ne BW AE | IE |] CE 7h NE eat AL | Ge | OLN |B [es USO Oe | eS | i Se Le |e PafE sKa A NAMEN NO | eae || iis ay Ue Ws i az fo A OS TG PO I A 1 NE IG TO Oe ae EG al | S| | | EH is SO I ee ese COPY) VE TAL NE ESE | call 41 eal &@ | GG | TG | 02 | 6T | ST | ZT 0 9T (GO| Gita | fe | ae | ll Tf I So Cian |Coem |fGe | eme cae SOUT 4SBj JO LOAF UI SSoOXT OT | Gr | HL | cE | TL | TE] OL | IL | 2 | FL | GE | €k | €t | 8 | FT | SuoMeseuer ‘mojg GT | PE | &f | PE | PE | SE | 6 | IL | 2E | ST | EE | OF | ST |} OT | OF | ‘suonereuey 4Yseq ‘ported Avp-OT YYW bE ae PO Pe A a NB | | BSB] @ | Ba]Poe ee ee cvseccstise yong 4SBJ JO OAV} UT SSOOxHT jG |G IT} 12} | % | et} 2 | TL} OF} IL | FT | St | FT | Saomereuen ‘mos OE TAR te Me oe eh Ve get is i Gal GG IT | &f | If | “Suonuroues “4se7 :porleg ABp-OT YRINO PINS Ne WE OO We Sees ae =|] @ || GaaPoer econ aeoonessayeny } SBF JO IOAVJ UTSSOOX TE} 1 )6 |9 | OL} 2 |% | 6 | 8 | TL} OF | OF | Ct | Zt | OT | “SuoeIeUer ‘mojg OF | ot | IE | IE | ck | Gl} 6 | IL | HE | HE | IE | SE | et | ct | 2 | suoyerouen ‘4seq D :poreg Avp-oT pay, VSD ap Wee Ws PD eae ae [las ae at € | F% ; alee EL aeRCOULT 4SB} JO IOA BG} UT SSO0XTT WE OE ee WN Te eee GO PO TS NG | or ls “"SUOTPEIOUOY) ‘MOTE OT | OL | TE | TE | Of | 6 IT | TE | GE | GE | GE | GI | OT | €T | OF |’ ** “Saorqye1uey ‘4suq :polleg ABp-QT puoveg PWS Oe Wie ie 8 |e || 6 te NE Es 1s Tile Ieee eect ee tere artes SO CLT 4SBy JO IOABJ UL SSOOXT TAL IPC EDIE Ge 9) OVE OVE he |) ae | | IT | 6 | &f | OL | 8 |°"""suoreiouey ‘mojg SI | &T | OF | €L | €t | &t | GE | Gt | St | FE | FT | ZT | ZT | Zt | at | suoneseuer 4ysuq ‘polled Avp-O]T 4Sany ST | vL | €f | ol | 11 | OL} 6 | 8 BND We Wie PB ee | YadWON SANIT adNjNI SSD 4aj}{D UOLJIAIaS pasuDjDg ‘aur sad porsad finp-uap tad suorjnuauab fo vaqunu jonjoy 6 HWIAVL {Q-] quawmisaday 482 AUSTIN RALPH MIDDLETON It is 13.74 selections per line for the fast-selected ‘slow’ lines and 12.79 selections for the slow-selected ‘fast’ lines. Now the aver- age difference of fission rate per line per day for the eighty days of balanced selection immediately preceding the present experi- ment was 0.25 generation per line per day and as a result of the fifty days of reversed selection this difference was wiped out and 15 10 50 10 B 30 Fig. ll-a Polygon of the average number of generations per line per ten- day period produced by the ‘fast’ and ‘slow’ sets of lines of Experiment 1-C (balanced selection for 50 days after 80 days of selection, 40 days of balanced selection and 12 days of mass culture). The continuous line shows the averages for the ‘fast’ set, the broken one the averages for the ‘slow’ set. The ordi- nates are the averages per ten-day periods, the abscissae, the ten-day periods. Fig. 11-b Curve of the difference in favor of the fast lines between the aver- age number of generations per line per ten-day period produced by the ‘fast’ and ‘slow’ sets of lines of Experiment 1-C during its five consecutive ten-day periods of balanced selection after mass culture. The ordinates are the differ- ences of the averages, the abscissae the consecutive ten-day periods. FISSION RATE OF STYLONYCHIA PUSTULATA 483 an average difference of 0.27 generation established in the reverse direction. , If this difference shows itself to be heritable, it will appear that it is probably due to the large number of selections prac- tised within the relatively short period of fifty days; it will indi- 16 10 30 50 Fig. 12 Polygon of the average number of generations per line per ten-day period produced by the ‘fast’ and ‘slow’ sets of lines of Experiment 1-D (reversed selection). The continuous line of the curve of the former fast lines now slow-selected. The broken line is the curve of the former slow lines now fast-selected. The ordinates are the averages per ten-day periods, the abscissae the ten-day periods. ® cate that the alteration of fission rate is proportional to the number of selections made, rather than the length of time over which the selections are distributed. This matter is tested in the next experiment. Experiment 1-E. Balanced selection after reversed selection, June 2 to June 21, 1914. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 4 484 AUSTIN RALPH MIDDLETON TABLE 10 Experiment 1-D: Actual number of generations and selections per 30 lines per ten- day period during the five ten-day periods of the second reversed selection of the lines of Experiment 1 AVERAGE TEN-DAY P IODS FIRST | SEC THIRD J A A ERIO OND FOURTH] FIFTH TOTAL PER LINE Fast Selected Slow Lines: Average number of selec- ‘ 5AM Total number of generations| 263 | 309 | 300] 292 | 473 | 1637 | 54:57 Average number of genera- CLOT SOMME ce soca c cds Meteret 8.77 | 10.380) 10.00) 9.73 | 15.77 10.91 Slow Selected Fast Lines: Average number of selec- CUOMStHE Rs coo sae ye aes eee PR |) DETA0)| 9, 118}] PAGO) |) 83 Total number of generations| 237 | 273 | 217 | 225] 276 |. 1228| 40.93 Average number of genera- EONS Meret ais Se eer ee: 7.90 9-10) 7.23) 7-250) 9).20 818 Actual excess in favor of the fast selected ‘slow’ lines....} 26 36 83 67 197 | 409) 13.64 Average excess in favor of the fast selected ‘slow’ lines....| 0.86 | 1.20) 2.77) 2.23 | 6.56 2.73 The lines from the experiment in reversed‘selection (Experi- ment 1-D) just described were next continued under the bal- anced selection method for two ten-day periods in order to de- termine whether the differences in the reverse direction obtained in that experiment were true heritable differences of fission rate. The animals were transferred to fresh slides daily. Figure 13 shows the polygon of the average fission rates of these two sets of lines and table 11 gives the actual number of generations produced by each set. From these it appears that the average daily difference of fission rate per line between the two sets of lines is 0.88 generation in favor of the fast selected slow set. It is therefore clear that reversed selection had really pro- duced an heritable difference in fission rate in favor of the former slow lines; the relative rates of fission of the two sets was now reversed. As before remarked, this result is to be expected; the earlier experiments show that direct selection pro- duces hereditary differences in one direction; reversed selection FISSION RATE OF STYLONYCHIA PUSTULATA 485 in the same way produces hereditary differences in the opposite direction. General results of the first group of experiments. The experi- ments thus far described have all dealt with parts of a single clone of Stylonychia pustulata; selection has been practised on 16 12 10 20 Fig. 13 Polygon of the average number of generations per line per ten-day period produced by the fast and slow sets of lines of Experiment 1-E (balanced selection after the reversed selection of Experiment 1-D). The continuous line is the curve of the reversed selected former fast lines now balanced selected. The broken line is the curve of the reversed selected former slow lines now bal- anced selected. The ordinates are the averages per ten-day periods; the abscissae the ten-day periods. the variations in fission rate within the clone. For one hun- dred and thirty days two halves of this clone were subjected to selection in opposite directions; this produced a marked and steadily increasing difference in the average fission rate of the two halves. Expressing the excess of generations produced by the fast-selected lines as a percentage of the total number of SGle OTT ies sees SOUT] MOTY Jo osuLY TiaiE TOL yay IP ObPP GOR SOUT] ISBT JO OsuLy 8c8T j29 |89 |TZ |99 |T9 |T9 |F9 |T9 [29 |T9 |#9 |€9 |89 |P9 89 [go |89 |€9 |G¢ jog j9¢ Ife |P9 |Po Jee j9G jag |eo [eG {TQ | coor 4SBy OY} JO IOARF ut ssooxGT SL9E JOCT GIT JTLT JOCT |GET |SITE \GLT [GIL |SITE JECT JOGI |EZT JZoT |EZT JATT |SZT |8VI |oST |G [Sct JOST FET \ZZT |ZeT |SZE |8st |9ZE |2zE [Sst Pet J eae MOTS 90S ST |Z8T TST JOST |O8T /62T JE8T OST |S8T |FST |F8T JOST |S8T |Z8T |SS8T EST JOST SST |OST |TST |ZS8T |SZT |98T |TST |TST [FST |E8T |G8T Jest jest | nee nome SC | aNameny (RS) La AG I 7G | CRG KS |) TS I ELEN aS ANOS | Oe EH PE Oe Ie Nae sie Cae |) oe | OG IE WZ 3 Mets he HaaNAN ANIT T quausadauy fo spoiiad fivp-o] g1 2y2 Burinp mops pun yspf ‘aur yonva fig paonpoud swownsauab fo waqunu ojoy, :4 pun g ‘e ‘7 spupg ‘J quamuodxy ol WTAaVL 08'S AWE |S 1S Pe eG We We Se SAC GW ee a ip WS ae te Re GS eG) PG Pee Pssiecgi jemnaey -3S 4sey ayy , H JO LOAB] UTSSOOX 90°2 NO PAS SSS AT || ASAT I ES PPS IP SS STSCI YES EE HO Cigale tesa IPE IPM I ET Oe I a I Ate Wachee ES 98 oI 0 PKS NG) VI AN Ne LEN OI TAL IGE CALE | OES NT TEE TS NE MT Ae YP Ce Pee I) Le TONE Ye TT Te NT Gat Sie 2 |] SI OIE |) BIE || Sale eee eee zones SB seouestet AwBp-0] puovg —8°T PNG NE NOS OG GSI RNG NG NA Te ae aie ap NG We PTE I sO az PO I@ yer BS UT pete] -0S 4SBy 9 JO LOAB] UTSSOOXH 99°01 0 ANS NAG ECO TVS) VTE MP VT MEY) DEE YUE TE COE EP LTE I Te Te) Ve Yh UNE HH DST OTe Nf NE] COLE OLE YO TE Le Te I] Tee |) [ye eee ees eo ones 01S OP GI 0 Cots ESI Mab SIE AG OE tsi SIE RIE fleet Ve Ae Ae | TD ES A eT ET he NT CON Ge HT Te Te ee CSE I Tn Ee a ye] Taye Pees oe ee eer eons i :powed Avp-O] Sar NOIL ov i RIGS |IvLOL! OF | 6% | 8Z | 242 | 9% | Go| FZ | EZ] Z| 1} 02 | GT | ST} AL | OT | St | FT | EL | ST} IL] OF | 6 8 Z 9 g P tf G I HaadNON ANTI MG AY | sotauras uoijoa)as pasunjppg ‘au wad poisad hvp-of “ad suoynuauab fo saqunu ony :q-[ quauisaday It WIaViL FISSION RATE OF STYLONYCHIA PUSTULATA 487 generations produced by both sets, the difference was 6.9 per cent in part 1; 12.8 per cent in part 2; 19.38 per cent in part 3, and 21.2 per cent in part 4. Table 12 summarizes the genera- tions produced by each line of each set throughout all parts of the experiment. It shows that for the fast lines the number of generations ranges from 178 to 187, while for the slow lines the range is but from 116 to 128. Thus there is no overlapping in the two sets; the slowest fast selected line has produced 50 more generations than the fastest slow selected line. To determine whether the difference in fission rate thus pro- duced is heritable, parts of the two sets were removed at inter- vals and subjected to culture without selection (‘balanced selec- tion’). In every case it was found that the difference was herit- able. Also, representatives of the two sets after eighty days of selection and forty days of no selection, were subjected to mass culture for twelve days. Further line culture for fifty days without selection showed that the inherited difference in fission rate still persisted. Thus the inherited difference produced by selection had lasted for one hundred and two days without selection. Experiments with reversed selection showed that the inherited difference could be reversed as readily as it is produced; the originally fast set was thus caused to become the slower one, and vice versa. Continuation of these sets without selection showed again that the difference so produced was heritable. Thus in this case the selection of small individual variations in fission rate has split the single clone (derived vegetatively from a single parent) into two hereditarily diverse divisions (diverse clones). - 2. The second serves of experiments The results. of the first series of experiments appeared so important and in some respects unexpected that it was. felt necessary to control them by repetition, beginning again with a single individual, and endeavoring anew to procure from it by selection two hereditarily diverse sets. This was first at- A88 ; AUSTIN RALPH MIDDLETON tempted with a single individual taken from one of the fast lines of the first set of experiments, giving experiment 2, described below. It was later carried out anew with the progeny of a single ‘wild’ individual (‘third series’). Experiment 2-A. Selection among the progeny of a single individual from Experiment 1. The individual selected for repe-- tition of the experiment was one of those belonging to a fast line of the previous experiment. The progeny of this individual did not live well, so that in some cases one or both sets died out before any definite result was obtained. In one case, however, selection of fast and slow sets was continued through nine con- secutive ten-day periods (April 5 to July 3, 1914). During every ten-day period except the first set the fast-selected set produced more generations than the slow-selected one. The results are less striking than in Experiment 1, however, in the fact that on twenty-two days out of the ninety, the slow lines produced more generations than the fast ones. The irregularity appears connected with the high mortality in both sets. How- ever, the average difference in fission rate per line per day in favor of the fast selected set was 0.317 for the first thirty days, 0.757 for the second thirty days; and 0.61 for the third thirty days. Experiment 2-B. Now the two sets resulting from the ninety days’ selection in Experiment 2-A were subjected to balanced selection for ten days. The difference in fission rate persisted to the extent of an average difference in favor of the fast lines of 0.28 generation per line per day, though on one day the slow lines were faster by 0.01 generation per line. Before the end of the next period all lines were dead, owing perhaps to the hot weather. Results. The evidence from these two experiments is, so far as it goes, in the same direction as from those of the first set. Selection produced from the progeny of a single individual two sets differing hereditarily in average fission rate. FISSION RATE OF STYLONYCHIA PUSTULATA 489 3. The third series of experiments Experiment 3-A. September 22 to October 21, 1914. To further test the results thus far reached, and to determine whether they are based on conditions generally occurring in the organism studied, a new wild individual was obtained from a new mass culture brought into the laboratory September 15, 1914. From this, two sets of thirty individuals each, all belong- ing to the seventh filial generation, were obtained, and sub- TABLE 13 Experiment 3-A: Actual number of generations and of selections per ten-day period per thirty lines of the 30 fast and 30 slow lines isolated among the progeny of the single ‘wild’ individual subjected to opposite selections for 30 days TEN-DAY PERIODS FIRST SECOND THIRD TOTAL Uae Fast Lines: Average number of selections.| 2.53 1.60 1.50 Total number of generations. 973 634 603 2210 73.66 Average number of genera- LHLUOHASSg 0 \c.c) tr cen ORCI Reon or 32.43 21.13 20.10 Slow Lines: Average number of selections.| 2.90 1 2633 0.96 : Total number of generations. 935 581 523 _ 2039 67 .96 Average number of genera- | TONS: 0.5.0 oe Mo aoe occ 31.16 19.36 17:48 Actual excess in favor of the PAS tMIIMESH Rei sss ss ckaeee 38 53 0) 171 5.70 Average excess in favor of the Hevsie IGS ee Seen ee 2 SZ 2.67 1.90 Percent the differdice is of the totalsioroothnes) 4.4.6.0 1.99% 4.36%) 7.10% jected, in the manner previously described, one to ‘fast,’ the other to ‘slow’ selection. In this series only one individual, in place of two, was selected from each line at each change, and the selections were made daily. There was no reduplication of the fastest and slowest lines. Fourteen of the fast and twenty-one of the slow persisted intact throughout experimenfs 3-A and 3-B. Opposite selection was practised for thirty days. ‘The records are given in table 13. In all three ten-day periods the fast- selected lines produce more generations than the slow-selected, 490 AUSTIN RALPH MIDDLETON though in four days of the first period the reverse is true. In all the other twenty-six days the average of the fast lines was above that of the slow. The table shows a gradual increase in the number of generations produced by the fast lines relative to those produced by the slow, the excess in favor of the fast 15 30 B Fig. 14-2 Polygon of the average number of generations per line per three day period produced by the fast and slow sets of lines of Experiment 3-A (direct selection in opposite directions of the progeny of the second wild individual). The continuous line is the curve of the fast set, the broken line, the curve of the slow set. The ordinates are the averages, the abscissae the three-day periods. Fig. 14-) Curve of the difference (in favor of the fast lines) between the average number of generations per line per three-day period produced by the fast and slow sets of lines of Experiment 3-A (direct selection). The ordinates are the differences between the averages; the abscissae, the consecutive three- day periods. Note the progressive increase of the difference under opposite selection. FISSION RATE OF STYLONYCHIA PUSTULATA 49] being for the three successive ten-day periods respectively 1.99 per cent, 4.36 per cent, and 7.10 per cent of the total number of generations produced in the given period. The gradual increase of the difference between the two sets indicates that this difference was heritable. This gradual in- crease is well shown in the curves of figure 14, giving at a, the average number of generations produced per three-day period by each set, at 6 the curve of the differences in favor of the fast set. Figure 15 gives the curves of variation of the total number of generations produced by the two sets, showing that they overlap very little. The evidence indicates strongly that the effect of selection is cumulative. Experiment 3-B. 'To test whether the difference produced by selection is actually heritable, balanced selection was now prac- ticed for twenty-one days, October 21 to November 10, 1914. On every day but one (the second) the lines that had been sub- jected to fast-selection averaged higher than others. Table 14 gives the actual numbers of generations per line for the two ten-day periods. Table 14 shows that during both of its periods the fast lines averaged more generations than the slow ones and further that the per cent that this difference is of the total number of gener- ations produced by both together was practically constant. Figures 14-c and 14-d show the average difference of fission rate between these two sets of lines, averaged for three-day periods. These two figures emphasize the marked uniformity of this average difference. Hence this experiment shows that the oppo- site selection previously practised had produced a nearly uni- form heritable difference of average fission rate between the two halves of this second clone. The results are thus the same as in our first set of experiments. Experiment 3-C. Effects of conjugation on the results of selection. It appeared of interest to determine whether these inherited results of selection would persist through conjugation. It is well known that conjugation is an ordeal having many effects on eg p | Lo€ 86 WY Net Se Ose te tee ta SN Oe I ete EN ath ie @) fe e yaxek aS) ® ‘q jo dousreyIC, 90°CL OIG | 42 | 08 | P24 | TL | 69.) 82 | 92 | FL | GL) 92 | 8 | G8) 92 | 2 | BZ | G2 | HL | 89 | 02 | GO| 29 | 99 | 8G | G9 | T9 | 8G | TL | 69 | G2 | 82 | TOG eee CeCe (CeCe Mee Cn ac ieee? SUOT} BION [eIOL OF FE cer SE | GE |) S€ | FE | FE | 2E | SE | Ee | SE | LE | Th | OF | OE | LE | OF | GE | 2E | Ge | ee | 2c | Te | ce | 82 | 9% | 62 | 2 | OE | GE | SE] BE | HOTS SuOI}eioues = [BIO], 99°28 OSTE | 6& | Ih | tH | 48) SE) Ih) 8E | Ty | Le | 6& | &h | &h | OF | 9E | SE] OF | Ze | €& | Ze |-98 | 9e | FE | OF | GE | cE | 98 | Se} Ze | OF | 6E | pole tine (puke: SUOI}BIBUIG =[BIOT, ToS | 261 6¢ I Wo We NO ae Wee = he i th SiG I te 2) Oe 0 5 | (20. es ce eee | | ON eee ROO COTORE Gy, £9°8T 6S¢ 1G | 1G | 2T | ST | GI | OG | Ge | GI | 02 | OZ | €Z | FZ | OZ | 12 | 12 | OZ | 12 | ST | ZT | SL | FT | SE | GT | EL | LE | OL | ST | OF | 06 | Fe |SuoBIEUEy ‘mos 09°02 819 GG | GG | &% | OG | GI | Fo | GS | &S | GI | Ge | GZ | SS | ES | OZ | OZ | GZ | GZ | ST | 0G | GT | ST | OT | SL | Z| ST | OG | LT | 02 | Te | Fe |SaoreIOUSH “YseqT :poleg ABp-9T puoseg 246° | O& T 6 () | G@ | 1 OO Wie iO VO ho | eae S15 | =} GS] I (He WY We ae ae % | 0 |" eouere pT Gg gt | #t | st | 2t | st | ot | or | or | st | ot | or | zt | ot | er | zt | er | et | et | at | at | st | of | St | gt suoresouen ‘moyg AT | OL | St | st | 2t | st | st | zt} or | st | gsr | er | sr | zn] at | gt | st | gr | 2t | zt | or | st | 21 | GF | gt |suoryerouen ‘4yse,q :ponod Aep-OT SII 92° ST &LP ZEW SE 19h | Or 90° ZT org 20 | 6 | 82 | 21 wou vel Sl ~ — INGO aov udd -UGAV TIVLOL | OF | 62 | 8G | 26 | 96 | SS | FG | &% | GE | IG | OG | GE | ST | ZI EAM RSI Me Ca TG OS HG I fe Wh 22 eG ax lta 1 6 if 4UdHWON SANIT o I sjas yj0qg fig paonpoud 7p}07 247 fo $i aouasaffip sry} abpywaosad yoym smoys Juad sad payLDiu UwWNjOd ay], *sauy, pazoajas-jspf ay? fo Loapf ur saouasaffrp ay. yum ‘y-y quowisadagy fo worsajas ajzisoddo fo shnp fipuzyy ay] Buinozjof hjaypipauwi, worydajos pooueleq fo shinp auo-fhjyuan) ay) sof y-y quawisadag fo sauy ayy fo au sad porsad finp-uay wad suoynsauab fo saqunu yonpoy :g-g juaumwaday tL ATAVL 492 FISSION RATH OF STYLONYCHIA PUSTULATA 493 vitality and reproductive power; often changing the fission rate. Conjugation was obtained among the selected individuals of this third set of experiments, and its effect tested. Watch glass cultures were made in 34 per cent malted milk from each of the thirty fast lines and the thirty slow lines. On 90 20 Cc . Fig. 14-c Polygon of the average number of generations per line per three- day period produced by the fast. and slow sets of lines of Experiment 3-B (bal- anced selections after the direct selection of Experiment 3-A). The continuous line is the curve of the fast set, the broken line is the curve of the slow set. The ordinates are the averages, the abscissae the three-day periods. Fig. 14-d Curve of the difference (in favor of the fast lines) between the aver- age number of generations per line per three-day period produced by the fast and slow sets of lines of Experiment 3-B (balanced selection after the direct selection of Experiment 3-A). The ordinates are the differences between the averages; the abscissae, the consecutive three-day periods. Note the practical uniformity of the difference under balanced selection. December 4 and succeeding days conjugating pairs were ob- tained, some from the fast lines, some from the slow ones. Though many of the ex-conjugants died, eventually representa- tives propagating normally were obtained from six of the fast lines and four of the slow ones of Experiment 3-B. It will be recalled that these animals, before conjugation, had been subjected to opposite selection for thirty days, then to 494 AUSTIN RALPH MIDDLETON twenty-one days of culture without selection and twenty-three days of mass culture. On December 7, sixty ex-conjugants were isolated from the fast lines (ten from each of six diverse lines), and sixty from the slow lines (twenty from each of two slow lines, ten from each of two others). They were cultivated in the way previously described. Balanced selection (i.e., in effect no selection) was practised for fifteen days. Table 15 gives the numbers of generations produced by the fast and slow lines during the five three-day periods of this experiment. No lines were lost by death, and practically no selection of any sort was necessary. The daily record sheets of this experiment show that during its first five days the slow lines averaged more generations per line than did the fast lines. Table 15 shows this for the first TABLE 15 Experiment 3-C: Actual number of generations of the 60 fast lines of ex-conju- gants and of the 60 slow lines of ex-conjugants during balanced selection for 15 days. The two members of each conjugating pair were individuals of the same line THREE-DAY PERIODS FIRST SECOND THIRD |FOURTH] FIFTH TOTAL See tee Fast lines: Total number of gen- CLrAvlonShes ee Lee 378 302 188 237 192 | 1297 21.63 Average number of FenerawoMseneyn. os 6.30 DA035 Solos ongoma|oe20 Slow lines: Total number of gen- CLAUONSHe ene rek 416 310 178 225 181 | 1310 21.83 Average number of generations......... 6.93 5.16 |2.96 13.75 |3.02 Actual excess in favor j of the fast liness =... —38 —8 10 i, 11 —13 | —0.20 Average excess in favor of the fast lines...... —0.63 —0.13 {0.17 |0.20 |0.18 —0.04 Per cent of the excess in favor of the fast lines in terms of the total for both..............| —4.78%| —1.30%|2.738%|2.60%|2 .94% FISSION RATE OF STYLONYCHIA PUSTULATA 495 65 70 75 80 15 16 Fig. 15 Curves of variation of the lines of Experiment 3-A (direct selection of progeny of second wild individual). The ordinates give numbers of lines; the abscissae the number of generations. The continuous line is the curve of the fast set; the broken line, the curve of the slow set. Fig. 16 Polygon of the average number of generations per line per three-day period produced by the fast and slow sets of lines of Experiment 3-C (balanced selection after conjugation). The continuous line is the curve of the fast set, the broken line, the curve of the slow set. The ordinates are the averages; the abscissae, the three-day periods. two three-day periods and also that there is a marked decrease in both the actual and percentage difference. During the last three of the three-day periods the difference was uniformly in favor of the sixty fast lines and was remarkably constant. Figure 16, the polygon of the average number of generations produced by each set of sixty lines per day during the consecu- tive three-day periods of this experiment, represents this same result graphically. Figure 17-a and 17-b give the curves of variation from the diverse lines, the former for the first six days, the latter for the last nine days. Why the ex-conjugants of the fast selected lines should for the first five days be slower than the ex-conjugants of the slow selected ones is not clear. But this condition existed 496 AUSTIN RALPH MIDDLETON only during the period of reorganization after conjugation, when the fission rate in all was extremely low. As soon as the ‘nor- mal’ fission rate was resumed, the balanced-selected set of origi- 36 Fig. 17-a Curves of variation of the lines of Experiment 3-C (balanced selec- tion after conjugation) for the first 6 days of the experiment. The continuous line is the curve of the fast set; the broken line, the curve of the slow set. The ordinates are the numbers of lines; the abscissae, the number of generations. Fig. 17-b Curves of variation of the lines of Experiment 3-C (balanced selec- tion after conjugation). The continuous line is the curve of the fast set; the broken line, the curve of the slow set. The ordinates are numbers of lines; the abscissae,*generations. nally fast ex-conjugants at once resumed its place in advance of the balanced-selected set of originally slow ex-conjugants and continued to divide more rapidly throughout the rest of the experiment. FISSION RATE OF STYLONYCHIA PUSTULATA 497 Thus it is clear that the heritable difference in fission rate brought about by selection during vegetative reproduction is not lost when the animals conjugate, but persists through that ordeal. Statement of the general results of the third series of experiments. This third series of experiments has entirely corroborated the results of the first and second series. In a second clone, entirely unrelated to that used for the first and second series of experi- ments, opposite selection for thirty days produced a heritable difference of average fission rate, a difference that gradually increased as selection progressed, indicating again that the effect of selection on this physiological character is cumulative. This average difference persisted through twenty-one days of bal- anced selection, twenty-nine days of mass culture followed by conjugation and then fifteen days of further balanced selection. DISCUSSION AND CONCLUSIONS All the experiments thus give concordant results; through selection of individual differences in fission rate it is possible to divide a clone into two divisions differing hereditarily in rate of multiplication. The effects of selection are cumulative; the hereditary differences between the two divisions become greater the longer selection continues. By reversing the direction of selection the hereditary differences between the sets are reversed. Is this effect of selection due to the slow accumulation of small variations, or to the chance isolation of mutants differing markedly from the type? The whole character of the results indicates strongly that the former is the case, and this indica- tion is borne out by careful study of the records. There is no sudden change at a definite point, indicating the appearance of a mutant. The steady cumulative effect of continued selec- tion can not be explained on the mutant theory without giving such a meaning to the word mutant as removes any distinction between it and ‘slight individual variation.’ It would require us to assume the repeated appearance of successively faster and faster mutants in each of the thirty fast selected lines, of suc- cessively slower and slower mutants in each of the thirty slow- selected lines; a conception which coincides with the view that selection operates cumulatively on slight individual variations. 498 AUSTIN RALPH MIDDLETON Woodruff and Erdmann (14) show that in Paramecium during vegetative reproduction there are periodical reorganizations of the nucleus, the vegetative macronucleus being replaced by a new portion of the reserve micronucleus. It has been suggested that this new macronucleus may give an altered hereditary con- stitution, so that at each reorganization inherited variations may appear. Are the effects of selection herein described based on such variations occurring thus at the time of reorganization? The relatively short time’ required for producing inherited differences among the progeny of a single individual make it improbable that the variations in Stylonychia are to be accounted for in this way. As we have seen, marked differences appear after ten days of selection, and these are gradually increased in the next ten days, and again in the next, and so on. The per- centage of difference in proportion to the total average fission rate for the 13 consecutive ten-day periods of Experiment 1 are 5.415 10.35, 5:58, 7.62, 23:81, 13.69, 19.20;521-93" 15.19) 24 ae 26.05, 25.385 and 19.23. These percentages for the three con- secutive ten-day periods of Experiment 3 are: 1.99, 4.36 and 7.10. Now in Paramecium the interval between reorganizations is about thirty days. If in Stylonychia the interval is of about this length, it would be quite impossible to account on this ground for the cumulative effects of selection occurring within periods much shorter; selection should show sudden effects imme- diately after the reorganization in a given stock, and should then be quite without effect during the intervening periods. Nothing of this sort appears in the records of the present experiments. The main interest of this particular matter lies in its bearing on the question whether variations are definite and limited in extent and possible number, as in rigid Mendelian recombina- tions of invariable factors; or whether variations may be of indefinitely many diverse extents and are not limited by a pre- cise numerically definable factorial structure of the germinal material. If the nuclear reorganization described by Woodruff and Erdmann takes place in a definite way, comparable to the known reductions and recombinations in the chromosomal appa- ratus at the formation of the germ cells, then this could not be FISSION RATE OF STYLONYCHIA PUSTULATA 499 a source of indefinite and unlimited variation. After a time the possible combinations of factors would be exhausted, and such constancy would result as Johannsen claims to have found in his ‘pure lines’ of beans. Evolution thus could not make ex- tended and continuous progress in this way. If on the other hand the nuclear reorganization occurs in no precisely definable way, but with indefinite and unlimited variations, then this apparatus shows precisely the characteristics held to be common to organisms by those who believe in continued evolutionary progress through the accumulation of such indefinite and un- limited variations. Some material: basis for such variations would have to be assumed; the nucleus might furnish this as well as any other portion of the organism. But, as we have seen, the present evidence does not favor the idea that the hereditary variations in Stylonychia are dependent at all on these nuclear reorganizations. Our main result, that during vegetative reproduction among the progeny of a single individual selection of small variations produces cumulative hereditary effects, is in marked contrast with the results of most investigators, who, following Johannsen (03, ’09, 711), have found that ‘pure lines’ or ‘clones’ are heredi- tarily constant under selection. Johannsen’s results were ob- tained with self-fertilized lines of beans. Similar ineffectiveness of selection has been found by Hanel (’08) and Lashley (15) as to the number of tentacles in Hydra multiply by budding, by Jennings (’08, ’09, ’10) for size in infusoria; by Barber (’07) (in the main), and by Winslow and Walker (’09), in bacteria; by East (10) in the vegetative reproduction of the potato, by Agar (713 and 714) in Cladocera and aphids multiplying partheno- genetically; and by various other investigators on diverse organ- isms. Some discordant results have been recorded, but most of these are ill-defined or uncertain; it is mainly in bacteria, with their immense difficulties for precise technique in pedigree work, that heritable variations or modifications have been described. The immense preponderance of evidence has been that in uni- parental reproduction heritable variations do not occur (save as rare mutations of marked character), and that selection of slight THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, NO. 4 a 500 AUSTIN RALPH MIDDLETON individual variations is without effect in altering the hereditary characteristics. How are we to account for the discrepancy between the present results, and those just mentioned? In Stylonychia we are dealing with an organism which is large enough to be easily handled and followed individually, so that no question can arise as to the purity of the pedigrees (as sometimes occurs with reference to Bacteria). In this organism the facts as to the cumulative effect of selection are clear. We are of course dealing with a delicate physiological charac- teristic, and this is perhaps more readily varied (even heredi- tarily) than the characters examined by most: other investigators. Further, it is perhaps true that hereditary changes are more easily brought about in the Protozoa than in the more complex organisms, for in Protozoa the ‘apparatus of heredity’ is in close chemical contact with all the somatoplasm. But a certain feature of the experimental procedure in the present case may have more importance than these conjectural considerations. It has been possible in my work to make a much greater number of actual selections (where plus and minus cases were both present to choose from), than in most of the work that has given negative results. And it has been found that few selections give very slight results, and that a great number are required to give any marked differences between the sets. Thus, in my main experiment, on the average 39.86 plus selections were made in the fast-selected lines; 34.36 minus selections in the slow-selected lines. The difference between the two sets was thus the equivalent of some 74 selections extend- ing through an average of 150 generations. This resulted in the production of a constant average difference per line of 0.42 of one fission per day. Contrast with this great number of selections the siz made by Johannsen in obtaining his negative results with beans, the three or four made by East with potatoes, the two made by Winslow and Walker with bacteria, and similar small numbers made by most’ other investigators along these lines; even indeed the selection through fifteen generations made by Agar, in FISSION RATE OF STYLONYCHIA PUSTULATA 501 Cladocera. It appears not at all inconceivable that in these organisms an equal number of selections, covering as great a number of generations as were made in Stylonychia, would have given similar heritable effects. What all the work shows (and here my own is not in positive disagreement) is that heritable variations of considerable extent do not occur so frequently as was at one time supposed, so that a few selections are not suffi- cient for establishing a definite positive effect. But negative results from a few selections are not sufficient for disproving the occurrence of heritable small variations which may be gradually accumulated. This indeed has been admitted by many of those that have obtained negative results; thus Johannsen remarked that ‘there is the possibility that a selection of fluctuating vari- ants, during very many generations, might divert the type of a line’ (03, p. 62); Jennings says ‘‘ what the pure line work shows (agreeing in this with other lines of evidence) is that the changes on which selection may act are few and far between instead of abundant. . .”’ @l0i%p;7144) and Hast states that “as a result of these experiments I would not go so far as to say that variations in power of resisting physiological or fungus diseases do not occur in asexual reproduction, but I do believe that the relative possibility that the commercial grower will obtain dis- ease-resisting varieties in this way is negligible’ (10, p. 134). As a result of this work upon Stylonychia it is possible to substitute for such indefinite remarks, precise data as to the occurrence of heritable variations and their accumulation through selection, when sufficiently long continued. And this can hardly fail to have influence on the conception of the hereditary con- stitution or genotype as a fixed thing, changing only discontinu- ously by marked steps or mutations, that do not intergrade. SUMMARY In Stylonychia pustulata, by the opposite selection through more than 150 generations of small individual variations occur- ring among the progeny of a single individual, it was possible to produce two sets differing hereditarily in rate of fission. During selection there was a gradual increase in the average heritable 502 AUSTIN RALPH MIDDLETON difference between the two sets, showing that the effect of selec- tion was cumulative. This result was, at various intervals, subjected to the most rigid tests possible, by balanced selection throughout long periods; by mass culture without selection, and by reversed selection. In every case the results were corroborated. The hereditary differences induced continued through periods of balanced selec- tion lasting longer than the periods of direct selection by which they were induced; they did not disappear save under the effects of reversed selection. These results were first reached with the progeny of a single individual multiplying asexually. They were then confirmed by beginning anew with a single individual from among this set and obtaining the same results among its progeny. A third set, derived from a wild individual quite unrelated to the first two series, gave the same results. In this third series conjugation occurred within each of the diverse sets produced through selec- tion, and it was found that the hereditary differences persisted through and after conjugation. Thus in Stylonychia, from a single clone of given genotype it is possible to obtain through long continued selection during reproduction by fission, two sets (clones?) of diverse genotype, differing characteristically from each other in rate of fission, under identical conditions; and retaining these differences from generation to generation. The selection of small variations, such as appear within the ‘pure strain’ or clone, is then an effec- tive evolutionary procedure. LITERATURE CITED Acar, W. E. 1913 The transmission of environmental effects from parent to offspring in Simocephalus vetulus. Phil. Trans., London, Series B, v. 203, pp. 319-850. 1914 Experiments on Inheritance in Parthenogenesis. Phil. Trans., London, Series B, v. 205, pp. 421-489. Barser, M. A. 1907 On heredity in certain micro-organisms. Kansas Uni- versity Science Bull., v. 4, no. 1, 48 pp. CaLkIns,G.N. 1902 Studies on the life-history of Protozoa. 1. The life-cycle of Paramecium caudatum. Arch. f. Entwicklungsmechanik der Org., v. 15, pp. 139-186. FISSION RATE OF STYLONYCHIA PUSTULATA 503 SALKINS, G. N., and Grecory, Louisk H. 1913 Variation in the progeny of a single exconjugant of Paramecium caudatum. Jour. Exp. Zodl., v.15, pp. 467-525. East, E. M. 1910 The transmission of variations in the potato in asexual reproduction. Conn. Exp. Station Rept., pp. 119-160. Hane, Evise 1908 Vererbung bei ungeschlechtlicher Fortpflanzung von Hydra grisea. Jenaische Zeitschrift, v. 42, pp. 321-372. Jennines, H. 8. 1908 Heredity, variation and evolution in Protozoa. II. Proceedings of the Amer. Phil. Soc., v. 47, pp. 393-546. 1909 Heredity and variation in the simplest organisms. Amer. Nat., v. 43, pp. 321-337. 1910 Experimental evidence on the effectiveness of selection. Amer. Nat., vol. 44, pp. 136-145. 1913 The effect of conjugation in Paramecium. Jour. Exp. Zodl., v. 14, pp. 279-391. JOHANNSEN, W. 1903 Uber Erblichkeit in Populationen und in reinen Linien, Jena, 68 pp. 1909 Elemente der exakten Erblichkeitslehre, Jena, 516 pp. 1911 The genotype conception of heredity. Amer. Nat., v. 45, pp. 129-159. Lasutey, K.S. 1915 Inheritance in the asexual reproduction of Hydra. Jour. Exp. Zo6l., vol. 19, pp. 157-210. Maupas, E. 1888 Recherches expérimentales sur la multiplication des infu- soires cili¢és. Arch. de Zool. Expér. et Gén., (2) T. 6, pp. 165-277. 1889 Le rajeunissement karyogamique chez les Ciliés, Arch. de Zool. Expér., et Gén., (2) T. 7, pp. 149-517. PresLes, F. 1912 Regeneration and regulation in Paramecium caudatum. Biol. Bull., v. 23, pp. 154-170. Winstow, C. E. A. and Watker, L. T. 1909 A case of non-inheritance of fluctuating variations in bacteria. Journ. Infect. Diseases, v. 6, pp. 90-97. Wooprurr, L. L. 1905 An experimental study on the life history of Hypo- trichous Infusoria. Journal Exp. Zoél., v. 2, pp. 585-632. 1912 A five-year pedigree race of Paramecium without conjugation. Proc. Soc. Exp. Biol. and Med., vol. 9, pp. 121-123, Wooprourr, L. L. and Erpmann, R. 1914 A normal periodic reorganization process without cell fusion in Paramecium. Jour. Exp. Zodl., v. 17, pp. 425-516. Lay OF ese § ig ame PR a Se ; ee SO aay i) on ee ae rj ‘ee 2 ah py Re a on glue a : “45 SUE eM (ae nee eh Page | a er bt par Cire Oe ne ; ta oe + nln : : : aon ‘ = as dolar, i of . - < im ee re 1 a4 , = ; fe y he, : ' 7) 4 - “4 ‘ Via see * - ee ' L ty wey eS 4 ; Pia. rf — ‘ Ay » ¢ pe i] i és . 4 : ‘ é ; = 3 1 ) . : nk res ' A Oe trae Pa is . Pie 4 VARIATION IN HEAD LENGTH OF SPERMATOZOA IN SEVEN ADDITIONAL SPECIES OF INSECTS” CHARLES ZELENY anv C. T. SENAY EIGHT FIGURES A study of variation in head length among spermatozoa from single testes in fifteen species of animals from widely separated groups was made by the senior author and E. C. Faust (15a and b). The frequency distribution of the size groups in a great majority of the cases showed bimodality of such a character as to make it highly probable that in these species the spermatozoa are dimorphic as regards size. Furthermore it was shown to be highly probable that this dimorphism is the result of chromo- somal differences. There is a close agreement between the ratio in the two groups as calculated on this hypothesis and the actual ratio determined by measurement. These chromosomal differences as has now been abundantly demonstrated are related to sex determination and the size groups must be similarly related, fertilization of the eggs by spermato- zoa of the upper group yielding females and by those of the lower group, males. A probability is thus opened for controlling sex as soon as living spermatozoa of the two sizes can be experi- mentally separated. In view of the importance of the question of existence of two size groups and in further preparation for the experimental tests, measurements were made for seven additional species and these are described in the present paper. - Together with those formerly described twenty-two species are now available for drawing a general conclusion. This num- ber includes all that have been measured, with two exceptions. These two are Helodrilus, a hermaphroditic form, and the 1Contribution from the Zoological Laboratory of the University of Illinois No. 49. 505 506 CHARLES ZELENY AND C. T. SENAY Plymouth Rock Fowl concerning which there is a controversy on the cytological side. They are not included because they have a special interest not connected with our main hypothesis and because it is desirable to make at least one more series of meas- urements in each species before publishing the data. Emphasis is laid on the fact that all sets of measurements are published, because bimodal distributions may appear occasionally as a matter of chance in a uniform population. The frequency of their occurrence is therefore all-important. The details of preparation of material and the method of measurement are in all respects similar to those described by Zeleny and Faust (15a) except that fixation was in all cases in osmic fumes and staining in haematoxylin. The authors are indebted to Mr. C. A. Hart for identification of the species. DATA 1. Corizus lateralus, a hemipteran. The material was obtained at the end of March and gave an abundance of active spermato- zoa. Two sets of measurements, each of 500 spermatozoa, were made. As shown in figures 1 and 2 the two determinations agree closely. Each shows a pronounced bimodal curve with modes at 27.1 » and 29.5 u, giving a ratio of 1.00 :1.09. The intermodal depression is deep and wide and the two elements of the curve are approximately equal as regards number of indi- viduals. There seems to be no doubt of the existence of two distinct size groups with equal numbers of spermatozoa. The chromosomal history was worked out by Montgomery (06). He describes two kinds of spermatids, one with six and the other with seven chromosomes. ‘Two size groups are there- fore to be expected but the drawings given by Montgomery are not large enough to enable one to make a calculation of the chromatin ratio. In Anasa tristis, another member of the Family Coreidae, the expected ratio is 1.00 : 1.11. 2. Leptocoris trivittatus, a hemipteran. The material was ob- tained in early March and gave an abundance of active sper- matozoa. Nine hundred and eighty-four measurements were made. The frequency distribution as shown in figure 3 is some- what irregular but there are two principal modes, one at 25.4 u = VARIATION IN HEAD LENGTH OF SPERMATOZOA 50 and the other at 27.8 uv, giving a ratio of 1.00:1.09. The majority of the individuals are grouped around the higher mode. If these modes can be considered as related to the sex chromosomes the Fig. 1 Corizus lateralus; frequency distribution of head-lengths of 500 sper- matozoa from a single testis. Welling Til, oo obencuee 23.0 23.3 23.7 24.0 24.4 24.7 25.0 25.4 5 Pde Brequencys a...) - 7 13 1 ily 13 14 16 19 sal 4 3 26.4 26.7 27.1 27.4 27.8 28.1 28.4 28.8 29 3 7 5 9 bo 5) (o2) — =) 2 42 30 L 11 13 19 26 38 59 8" 30 duesOnon 3080) 3i20 31ke). Biko) a2r3 aang oS. C0 eEIa ch 10 8 6 9 5 Fig. 2 Corizus lateralus; frequency distribution of head-lengths of another sample of 500 spermatozoa from the same testis as those given in figure 1. AVENUE Tics eer 23.0 23.3 °°23.7 24.0 24.4 24.7 25.0 25.4 25.7 26.1 BeeQuUency...)....".- 4 6 7 7 5 7 itil 11 11 13 26.4 26.7 27.1 27.4 (27.8) 28.1 2874) 28.8) 29.1 29.5 f 2 20 21 17 19 28 45 3L.2 ‘ol.6 3129 3273" 32-6 a 8 4 5 5 IN DING 2 aia BIN-AgYS CHARLES ZELENY 508 6.5 15 VARIATION IN HEAD LENGTH OF SPERMATOZOA 509 group containing the female-determining spermatozoa exceeds the other in numbers. On the basis of the cytological evidence we should expect the two groups to be equal but there is a great deal of evidence in both cytological and experimental data to indicate that one kind is often more numerous than the other. E. B. Wilson (06) has described two kinds of spermatids for this species, one containing six and the other seven chromosomes. Tn the absence of figures it is not possible to determine an ex- pected ratio. Since this species also is a Coreid, the ratio of 1.00 :1.11 determined from the chromosomes of Anasa is of interest. It agrees fairly closely with the measurements of head- lengths in this as well as in the last species. 3. Reduviolus ferus, a hemipteran. The material was collected during March and gave active spermatozoa. The results of the five hundred measurements are given in figure 4. The fre- Fig. 3 Leptocoris trivittatus; frequency distribution of head-lengths of 984 spermatozoa from a single testis. WADI SR, otis ld ro 20 Or2ies 2126. 22.0) 32253. 2257 = 20,023.38 23dn0 Brequencyasnecr. fa- 2c 1 2 3 6 9 13 18 23 25 PASO ZACAY ANT 1250) (25 2oe7ee2bole 2624 2027 4 i . 26 28 30 41 53 37 3l 38 40 Pap Mk | Dil ak DPA PAS PISS aE eshte) ANGIE) 38) = 20) cs) 41 50 60 51 45 41 39 38 40 Oo B05 S00) Bl sil @ sl) srs SA asl) dl 24 20 15 12 11 8 a 5 Fig. 4 Reduviolus ferus; frequency distribution of head-lengths of 500 sper- matozoa from a single testis. Wialtecrmee ee eke ccc... 2000 eae 24/40 247 2580) 25245 252% 26a 26.4: IDRC WISIMNGN 74s. 5 6 op aOR poe 4 5 6 5 12 19 25 30 30 Ket De oil Berke ake 36 49 56 45 33 2 29 8 “3051 30.5: 3029) SIE2) “SIGs SLO a2e35 3266 18 10 9 8 it 2 1 Fig. 5 Euschistus variolarius; frequency distribution of head-lengths of 500 spermatozoa from a single testis. Valuicuimmicrons.........-l3.0) W3e4 1367 14 14 Aloe 1525) S1lo28 ReQUeNeVe sae. fa. o- ols 6 16 20 21 24 46 60 39 38 Isl AG UBS A ale A) UR AI 3 2 7 6 6 2 08 84 38 1 510 CHARLES ZELENY AND C. T. SENAY quency distribution gives a major mode at 27.4 u and a minor one at 28.8 1. The minor mode may be accidental though such an irregularity is not common in frequency distributions in a population known to be homogeneous. There are no data available concerning the spermatogenesis of this species but Montgomery (’06) has described two kinds of spermatids for other members of the family. 4. Euschistus variolarius, a hemipteran. Material was ob- tained during April. All the spermatozoa were not fully devel- oped and selection was necessary to insure the exclusion of the unripe ones. Five hundred measurements were made. The re- sulting curve as shown in figure 5 is distinctly bimodal with modes at 15.1 » and 16.5 u and with an inequality favoring the larger spermatozoa. The ratio between the modes is 1.00 : 1.09. KE. B. Wilson (06) described an ‘‘ X”’ and a ‘‘Y” chromosome for this species. The expected ratio as approximated from his figures of the chromosomes is 1.00 : 1.04. This does not agree at all well with the ratio between the modes as given above. This result is contrary to that obtained for several species by Zeleny and Faust (15) which showed a striking similarity be- tween the two calculations. 5. Cosmopepla carnifex, a hemipteran. The material was ob- tained in May and the spermatozoa were all active. Five hun- dred measurements were made. The resulting curve as given in figure 6 shows well marked bimodality with approximate equality in the two groups. The modes are at 18.5 » and 19.9 u and give a ratio of 1.00 : 1.075. There can be no doubt in this case of the existence of two fairly equal size groups. Montgomery (06) describes two kinds of spermatids for this species, one with seven chromosomes plus an ‘‘ X”’ and the other with seven plus a “Y.”’ The figures are not suitable for the determination of the chromatin ratio. The ratio in another Pentatomid, Euschistus variolarius, is 1.00 : 1.04. 6. Passalus cornutus, a coleopteran. Material was obtained at the end of March and the spermatozoa were uniformly ripe and active. Five hundred spermatozoa were measured. Neglect- ing the minor projection at the right, the curve as given in figure VARIATION IN HEAD LENGTH OF SPERMATOZOA 511 7 is unimodal with the mode at 11.7 u. This indicates either a single group or two so close together as to give a unimodal result (see Zeleny and Faust ’15 a, p. 198). There are no published spermatogenesis data for this species. Descriptions for different species of beetles show in all cases, except one, two groups as regards chromatin content. Leaving out the possibilities of random sampling and of selective elimina- tion within the two groups the present species seems to be either without distinction among its spermatozoa or else has two groups differing but slightly from each other. 7. Berosus striatus, a coleopteran. The material was obtained in early April and the spermatozoa were all ripe and active. The frequency distribution of the five hundred measurements is given in figure 8. The curve is distinetly bimodal with approxi- mate equality in the two groups. The modes are at 16.1 « and 17.2 » with a ratio of 1.00 : 1.07. There are no cytological data for this species but as was stated in discussing the last form the descriptions for all but one species of beetles give two kinds of chromosome groups among the spermatids. DISCUSSION The present data as a whole substantiate the view that di- morphism in size of spermatozoa is of common occurrence in those groups of animals in which two chromosomal classes of spermatids are of common occurrence. Of the seven species described five are hemipterans and two coleopterans. In both of these orders a great majority of the species so far studied show a quantitative difference in chromosomal content among the spermatids though in each case there are some species which show no difference or only a slight one. Among the five hemipterans three, Corizus lateralus, EKuschis- tus variolarius and Cosmopepla carnifex, show two well marked size groups. The other two species are not so clear. One, Leptocoris trivittatus, is probably dimorphic though the curve is irregular. The other, Reduviolus ferus, is doubtful because the upper mode is very low and may be a chance projection in iy CHARLES ZELENY AND C. T. SENAY a unimodal distribution. It should not, however, be forgotten that the two elements of a population will, when combined, give a unimodal curve in case the modes of the elements are close together. Among the coleopterans one, Berosus striatus, has two well marked and equal size groups. The other, Passalus cornutus, has a single size group or two with modes very close together. Taking all cases so far described there are twenty-three spe- cles involved, including the one given by Wodsedalek (’13). This is a sufficient sample to justify us in stating that there can be no doubt of the validity of the hypothesis presented. The chromosomal dimorphism of spermatogenesis is represented in the active functional spermatozoa by size dimorphism. Control of sex then merely awaits our ability to separate the two sizes in the living condition and to use them in artificial insemination. . Fig. 6 Cosmopepla carnifex; frequency distribution of head-lengths of 500 spermatozoa from a single testis. Values microns......55...16.5° 1688 7-2. 17.5 7 9N asia ees 18 Opie Re quientyere 4.4) seen 6 14 16 19 26 42 83 53 35 I) I) BOs AS 40)8) ails} ilo 2220) 42 68 By 14 11 11 9 8 Fig. 7 Passalus cornutus; frequency distribution of head-lengths of 500 sper- matozoa from a single testis. Valuehmemnicronss.-o.25. 4. 72 Opto OR3O) 10455 ORGOe 1OLSO tl 00M isieales fod He quenCyaeerier, oka se 1 5 a 9 14 26 27 1L30— 1150-1170" Bieg5 12700, 125205 eo 46 65 89 65 55 22 29 WMA Wt US OO PAD 14 12 7 7 2 Fig. 8 Berosus striatus; frequency distribution of head-lengths of 500 sper- matozoa from a single testis. Vales aii eros: she trs 5 ek eee [sei WA Ara Ae Lop Le oe OMmlons Bre gwen cys. 2. sen ee tee > ee 4 9 9 10 14 32 41 16.1, “16.5 1638) Wie 17.5 ies ise 56 42 30 54 40 36 30 18.5 18.9 19:2) 19:6 197.9" 2023 22086 23 17 16 14 10 4 2 513 IN HEAD LENGTH OF SPERMATOZOA VARIATION 17 IG. 514 CHARLES ZELENY AND C. T. SENAY BIBLIOGRAPHY Faust, E. C. 1913 Size dimorphism in adult spermatozoa of Anasa tristis. Biol. Bull., vol. 25. Monteomery, T. H. 1906 Chromosomes in the spermatogenesis of the Hemip- tera Heteroptera. Trans. Amer. Phil. Soc., vol. 20. Witson, Epmunp B. 1906 The sexual differences of the chromosome groups in Hemiptera with some considerations of the determination and heredity of sex. Jour. Exp. Zodél., vol. 3. WopsEDALEK, J. 1913 Spermatogenesis of the pig with special reference to the accessory chromosomes. Biol. Bull., vol. 25. ZELENY, C. and Faust, E.C. 1914 Size differences in spermatozoa from single testes. Science, N.S8., vol. 39, no. 1003, p. 440. ZELENY, C. and Faust, E.C. 1915a Size dimorphism in the spermatozoa fram single testes. Jour. Exp. Zodél., vol. 18. ZELENY, C. and Faust, E. C. 1915 b Dimorphism in size of spermatozoa and its relation to the chromosomes. Proc. Nat. Acad. Se., vol. 1. THE EFFECT OF SELECTION UPON THE ‘BAR EYE’ MUTANT OF DROSOPHILA! CHARLES ZELENY anv E. W. MATTOON FIVE FIGURES The selection experiment described in the present paper was made with a view to testing the germinal uniformity as regards the distinguishing characteristic in a recent mutant, the ‘bar eye’ race of Drosophila. In this race the ommatidia are reduced in number and the facets are restricted to a vertical band or ‘bar’ as shown in figure 1. The characteristic appeared in a single male during 1913 (Tice, 714)? and the whole ‘bar eye’ stock is descended from this individual. The race has under- gone no apparent change during the two years of its existence. Our material was obtained in January 1914 through the kind- ness of Prof. T. H. Morgan. There is a pronounced sexual dimorphism in the number of facets, the males averaging 98.03 and the females 65.06. In every case in the present paper the female number is transformed to the male basis by multiplying it by ea = (5b. There is often a slight difference between the number of facets in the right and that in the left eye. For one hundred indi- viduals the average difference was 0.245 per cent in favor of the left eye. This difference is obviously not significant. The right eye only is given in the present records. The number of facets seems not to vary with the length of the period of development. In five broods counts were made of the facets of the earlier emerging individuals and compared 1 Contribution from the Zodlogical Laboratory of the University of Illinois No. 50. 2 Tice, S. A. 1914. A new sex-linked character in Drosophila. Biological Bulletin, vol. 26, pp. 221 to 230. 515 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19. No. 4 516 CHARLES ZELENY AND E. W. MATTOON with those of the later emerging ones. The averages for the two are approximately equal. A sample of five hundred individuals, 250 males and 250 females, taken from the general population of the mutant showed a range of variation in the number of facets from 45 to 182 with A /5} Fig. 1 A. Normal full eye of Drosophila. B. ‘Bar’ eye. The dark areas are the faceted regions. Facet number 16-30 31-45 46-60 61-75 76-90 91-105 106-120 121-135 136-150 151-165 166-180 181-195 196-210 Frequency 0 2 18 56 113, «164 —Ss 73 30 24 17 3 1 0 Fig. 2 Variation of facet number in the unselected population of the ‘bar’ eye mutant of Drosophila ampelophila. Five hundred individuals are repre- sented, two hundred and fifty males and two hundred and fifty females reduced to the male level. a mean of 98.03 + 0.73 and a standard deviation of + 24.30. The variation curve is shown in figure 2. The average for ten individuals of the normal wild race was 701.1 facets. Unselected stock was carried as a control through the period of the experiment. There was practically no change in the THE ‘BAR EYE’ MUTANT OF DROSOPHILA 517 number of facets, the average of 250 individuals at the beginning of the experiment being 98.04 and of the same number at the end of the experiment 98.03. There is thus no change in facet number in the absence of selection. . In making the first selection, food containing eggs and larvae was removed from the culture of the ‘bar eye’ stock and placed in glass vials. Every twelve hours the individuals which had emerged from the pupal cases were slightly etherized and an estimate was made of the number of facets under the low power of the microscope. Males and females with high or low numbers were selected out, high being mated with high and low with low. Each pair was placed in a small bottle with sufficient food to last until all the offspring of the first brood had reached the adult stage. When larvae began to appear in the bottle the parents were killed by etherization and an exact count was made of the facets. As the offspring emerged in the adult form esti- mates of facet number were again made and high selections were made in the ‘high’ lines and low selections in the ‘low’ lines. The final exact facet counts were in all cases made in killed individuals. The experiment has proceeded far enough so that data are complete for three successive selections in each of three ‘high’ lines, called A, B, and C, and in each of three ‘low’ lines, called D, E and F. Fifty individuals, 25 males and 25 females from a single pair, were measured for each generation in each of the lines, with the exception of the third generation in line B where only forty-six were available. The data for the individual lines are given in tables 1 to 6 and a summary of the six lines is given in table 7. Figures 3, 4 and 5 give in graphic form the course of the selections, each figure combining a ‘high’ with a ‘low’ selection line. The mean values, the extreme variates and the mid-parental values are here represented in diagrammatic form for each of the selections. In all cases there is a significant shifting of the mean as a result of selection. The general population mean of 98.0 is changed in plus line A to 108.7 by the first selection, to 127.5 by the second and to 135.5 by the third. In plus line B the 518 CHARLES ZELENY AND E. W. MATTOON TABLE 1 Plus selection Line A GENERATION 1 GENERATION 2 GENERAL POPULATION 38 28 g 3 3 = a arian ence Ra = fae = pan IPATCUUSS. acute. =. era tonne Seer 127) 183) 1704 169 Mid=parental@yalues...ccce ae sees sone eke 130 174 OPiS primase tac: cc bene Cann eee 67 69 84 89 79 ts 88 91 86 74 90 94 91 76 95 97 92 83 95 | 100 92 88 97 | 103 94 89 98 | 110 96 91 99 | 110 99 94 99 | 112 101 100 102 113 102 101 103 115 103 |} 106 | 118} 119 105 | 106| 124] 122 107 107 125 125 107.5). LOD) ize Ss 109 | 115 | 130] 134 412) LION Baie s7 116 | 118), 142) 139 118 128 147 142 118 133 148 148 122} 148| 176] 153 127: | 449 G7 156 134 | 149] 180) 163 140 | 151 | 189 | 187 179 | 169.) 210 | 210 Mean of offspring..........98.03 + 0.73 | 108.7 + 2.3 | 127.5 + 3.1 GENERATION 3 cE a | 2 s |e 177 195 92 95 93 97 98 103 98 109 103 115 107 119 108 124 109 127 10 [7/ 127 117 131 118 ileal 126 134 126 136 135 137 139 ilBy7/ 140 142 141 143 142 151 154 t53 159 157 167 160 180 163 187 171 192 | 174 204 186 135.5 + 2.8 corresponding figures for the three selections are 110.1, 128.6 and 141.9, and in plus line C, 116.9, 133.5 and 141.0. line D the general population mean of 98.0 is changed to 88.3 by the first selection, to 85.5 by the second and to 81.7 by the third. In minus line E the corresponding figures for the three In minus THE ‘BAR EYE’ MUTANT OF DROSOPHILA 519 TABLE 2 Plus selection Line B GENERATION Il GENERATION 2 | GENERATION 3 ae BE n 33 n 33 o 3 o 5 3 38g az 3 oo Get oo = os = [ees= GENERAL POPULATION 33 n 38 oe = sep IR SECIIUS HRMS Sis eee Pav aes 1389 | 121 Mid=parentalivalues:......... 2.4.6.4 05 130 Ofapringeme mens... ye eee 68 57 74 76 iD 80 79 85 80 85 82 89 89 91 90 91 95 95 98 95 98 97 100 | 100 102 | 106 104 108 105 118 108 124 112 125 115" | 4128 131 131 133 | 133 145 137 147 145 156 148 167 149 184 | 165 Mean of offspring..........98.03 = 0.73} 110.1 + 2.7 128.6 + 3.2} 141.9 + 2.9 =p selections are 93.9, 89.6 and 84.8 and in minus line F, 94.6, 89.8 and 84.7. As a result of the three minus and the three plus selections the mean value of the individuals in the plus lines (139.5) is greater than the highest extreme individual of the three minus lines 520 CHARLES ZELENY AND E. W. MATTOON TABLE 3 Plus selection Line C GENERATION 1 | GENERATION 2 GENERATION 3 On os} On GENERAL POPULATION oe go Oe = pa = ea = ao IPSTEMUS Me eH Ue eRe .nrs se rae etnias ce tamer 178 157 | 208 159 194 198 Mid=parentaliivalwest:.)- seme cite: 167.5 183.5 196 Ofisprimnggee ect: . OGeul 88 97 98 0.99 :1 39 77 92 Onsarad 89 119 140 OL920 ot 40 125 83 1.51:1 90 140 128 eeehsel 4] 159 155 1202551 91 126 124 Le OG ree 42 102 95 1.07 :1 92 133 137 0297s 1 43 74 49 1.51 :1 93 155 157 0.99 :1 44 53 36 WAT 94 157 148 £081 45 113 77 1.49 :1 95 118 115 021 46 69 64 1.09 :1 96 111 88 12600 47 200 93 21521 97 148 138 Oil Sa 48 88 89 0.98 :1 98 105 103 ee eal 49 56 60 0.93: 1 99 127 126 1 Ole 50 76 64 eee 100 124 120 te Oly ek 533 534 MARY B. STARK TABLE 4 13 36 NO. OF PAIR FEMALES MALES SEX RATIO |NO. OF PAIR] FEMALES MALES SEX RATIO 1 194 95 2.04: 1 1 116 128 0:99 a 2 123 50 2.46 :1 2 114 124 0.92% 3 123 124)" |,0:99 3a 3 89 63 Lalor 4 131 113 ILI) gl 4° 101 107 0.94 :1 5 164 61 2 Ove oS) 100 74 a RASH eo! 6 108 86 1.14:1 6 102 124 O.S2eeh i 143 53 2.70 : 1 7 119 96 1.24 :1 8 82 104 OTe 8 92 94 0.98 :1 9 193 173 ies ts a 70 49 1.43 :1 10 148 82 1.80 :1 10 100 80 1 Zoned 11 159 143 I atih gal 11 194 153 1 2teaee 12 87 93 0.94 :1 12 109 47 2.382 :1 13 53 59 0290 e! 13 129 105 te 25e-e 14 119 173 tro) sat 14 124 105 118 sal 15 165 152 OF el 15 127 116 1 10h 38 47 1 112 56 2.00 : 1 I 137 141 Of ew 2 159 139 aa 2 150 157 0.96 :1 3 125 78 Ge yaat 3 140 105 Laoeeal 4 67 73 0.93 :1 4 85 95 0.39 si 5 58 46 126031 5 115 107 1OSsa 6 131 90 1.45 :1 6 127 77 165 5ah 7 90 79 HBO a a 185 on 2.03 :J 8 119 129 OF92) 24 8 201 $1 2.45 31 9 71 78 OS etl 9 IIS) 90 bS2e 10 153 86 UR Trompe 10 149 150 0.99 11 183 153 1) ga 11 201 92 2.20 :1 12 164 117 1.40 :1 low 1 doubtful and 4 high to 7 low ratios where equal numbers of each kind are expected. Further tests were therefore made. Pairs: were mated) from/13, 1; 13, 2; 135.5; 13,7; 36, 1254 ieee 47, 8; 47, 11. The counts from these matings are shown in table 5. As expected, the test of 13, 1 shows the presence of a lethal since there were 6 high to 6 low ratios. The test of 13, 2 indicates a lethal with 8 high to 11 low ratios. The test of 13, 5 indicates a lethal with 6 high to 13 low ratios. LETHAL FACTORS IN DROSOPHILA as) TABLE 5 13, 1 NO. OF PAIR FEMALES MALES SEX RATIO 1 186 103 1 S0cen 2 118 122 0.97 :1 3 121 128 1.06:1 4 116 ° 101 1.15 :1 5 171 81 ata 6 138 46 3.00 :1 7 120 116 1.03 :1 8 107 52 2.06 :1 9 115 110 1.05 :1 10 104 128 Orsi 11 127 68 1.86 :1 12 174 100 74 si 3502 NO. OF PAIR FEMALES MALES SEX RATIO 1 159 61 2.61 :1 2 128 146 0.88 :1 3 105 53 2.00 :1 4 212 94 2.36:1 5 139 101 13721 6 84 74 ees a 135 122 110 8 240 ; 100 2.40 :1 9 225 160 1.40 :1 10 Wi 79 0.98 :1 11 91 69 Iago 12 72 87 0.83 :1 13 109 81 1.34:1 14 108 84 i eligacal 15 175 91 1.92 :1 16 222 116 1-91 d 17 88 40 Drove 18 134 104 120) 19 145 AT 3.08 :1 Two high to 9 low ratios appear in 13, 7, which probably in- dicates a lethal as the parent also had a high ratio (viz., 2.7 : 1). These four tests of 13 show beyond doubt that a lethal was present, since each of the four families tested gave some high ratios. 536 MARY B. STARK TABLE 5—Continued 13, 5 NO. OF PAIR FEMALES MALES SEX RATIO 1 « 116 139 0.87 :1 2 103 100 1.03 :1 3 222 109 2.03 :1 4 140 140 1.00 :1 5 159 58 2.75 31 6 150 168 0.89 :1 7 205 210 0.98 :1 8 107 44 2.43 :1 9 239 140 Tayi al 10 164 137 1.20:1 ll 125 ie 107031 12 148 123 1.20:1 13 172 156 1.10:1 14 107 r 50 2.14:1 15 113 115 0.98 :1 16 100 134 0.75 :1 17 211 150 1.41 :1 18 116 133 0.87 :1 19 155 62 2.50 :1 13.7 NO. OF PAIR FEMALES MALES SEX RATIO 1 267 229 1474 2 180 161 Laie 3 136 123 1.10:1 4 208 89 2132 al 5 106 106 1.00 :1 6 162 87 2.00 :1 7 139 131 1.06 :1 8 119 140 0.85:1 9 121 119 rere i 10 109 76 1.43 :1 ll 113 83 1.36 :1 Of the 60 females tested, half (30) should have given 2:1 ratios, while in fact, only 22 gave such ratios to 38 giving normal ratios. Provisionally, this deviation may be ascribed to chance. LETHAL FACTORS IN DROSOPHILA 5387 The following tests were made of the daughters of number 12 of 36, table 4: TABLE 5—Continued 36, 12 NO. OF PAIR FEMALES MALES SEX RATIO 1 121 72 Wage Sul 2 151 56 Peso) © I 3 130 109 W743 8 ik 4 159 140 igi 2 5) 97 115 OFS Tal 6 177 108 Weta) 2 Al 7 135 88 1b) § 1 8 144 55 27 Orel In this test there were 5 high to 3 low ratios showing that 36, 12 was a lethal bearing female. The following tests were made of the daughters of number 7 of 47, table 4. TABLE 5—Continued 47, 7 NO. OF PAIR FEMALES MALES SEX RATIO 1 259 121 Pad ES AL 2 235 122 Opel 3 188 78 2.4:1 4 133 152 OZR 5 170 90 ibe) gall 6 185 95 Ona 7 147 148 (ORL 8 164 150 ipl Here there are 5 high to 3 low ratios indicating a lethal. In 47,8 the 4 high to 4 low ratios indicate a lethal. In 47, 1 there are 3 high to 7 low ratios. Of the three sets of tests of 47 of table 5 there were 12 high to 14 low ratios which establishes the presence of a lethal in this line. The general outcome of these tests leaves no doubt that a lethal was present in the original females that were tested. Since only half of the daughters of a lethal female are hetero- zygous for lethal and since these females are indistinguishable 538 MARY B. STARK TABLE 5—Continued 47, 8 NO. OF PAIR FEMALES MALES SEX RATIO 1 146 80 1S 2 93 94 1 OV 3 201 162 ipa 4 124 58 aad 5 147 156 0.9:1 6 196 69 2S 7 150 119 ile) 2 8 158 57 238). 47, 11 NO. OF PAIR FEMALES MALES SEX RATIO 1 142 79 ye) 3 2 168 173 0.98 :1 3 157 171 0.92 :1 4 159 162 0.98 :1 5 167 124 13a 6 247 124 2.0021 7 142 127 11 2et 8 190 97 1.96 :1 9 122 146 0.90:1 10 1 163 156 1.04: from their sisters, it is by chance only that one would choose a female with a lethal factor when testing out a stock. On the other hand if a lethal female is mated to a male having a sex TABLE 6 113}, i, abl NO. OF PAIR FEMALES MALES SEX RATIO iL 183 156 wil gil 2 138 118 Welle? 2 il 3 155 70 22222 4 158 139 Is 8 5) 154 82 1.90:1 6 142 152 0.90 :1 7 163 152 1 OIEn oor WN Re OCOANOOA PWN rH COON Oorhrwhd SMP aONOOARWN HE _ LETHAL FACTORS IN DROSOPHILA 539 TABLE 6—Continued 13) 1s 12 114 70 Gated 156 136 1:12:1 199 108 1.83 :1 154 82 1.88 :1 195 102 1.92 :1 154 103 1.50:1 ee 123 60 2.00:1 » 130 52 2.5021 98 105 0.93 :1 82 3 97.33.21 110 95 1.15:1 140 135 1.04 :1 170 50 3.40 :1 84 62 1.35:1 99 70 1.41 :1 18, 2,'8 83 61 1.36:1 150 62 2.4221 105 96 1.08 :1 156 162 0.97 :1 84 39 2.16 :1 134 134 1.00 :1 120 73 1.64:1 109 64 1.54:1 134 111 1.21:1 138-19 164 129 Oypen| 207 214 0.92 :1 225 P11 1.07:1 932 172 1.35 :1 182 185 1.00 :1 341 120 2.84:1 283 152 1.86 :1 193 93 2.07 :1 366 167 2.20:1 251 246 1.02 :1 540 MARY B. STARK linked factor close to the lethal factor in question a stock may be obtained in which the lethal females may be selected with great probability. For example: If a red eyed female carrying a lethal is mated to a white eyed male half of her daughters will have the factor for red and the factor for lethal in one X chromo- some and the factor for white and the factor that is the normal allelomorph of lethal in the other chromosome. If such a daugh- TABLE 7 13, 2, 8, 2 (four 1:1 ratios omitted) FEMALES MALES , CROSSING OVER ENO 2 [OR oD ANTOR ey | he SE A ee | SB LWP NEW ULE Red White Red White NS ee i 62 73 11 47 19.0 2 87 78 12 48 20.0 3 99 ue. 23 58 28.4 4 i 67 11 55 16.7 13, 2, 19, 6 (five 1 : 1 ratios omitted) i 177 94 30 88 25.4 2 110 117 19 81 19.0 13, 2, 19, 7 (three 1 : 1 ratios omitted) 1 83 107 23 69 25.0 2 112 118 22 93 19,1 3 98 81 26 68 PAL oll 4 162 85 14 79 15e1 5 89 94 20 71 22.0 6 98 87 aM 72 Dilines a 82 83 | 18 65 Dileais ter is mated to a white eyed male half of the female offspring will be red eyed and half white eyed. The former getting their red bearing chromosome from their mother will be the lethal bearing females since the red bearing chromosome also carried the lethal factor. By selecting the red eyed females, therefore, in each succeeding generation and breeding them to white eyed males the lethal stock can be maintained. Virgin females from numbers 13, 1) 11s 13. 12. ols 13, 2, 8; 13, 2, 19; of table 5 were mated to white eyed males. The results are shown in table 6. LETHAL FACTORS IN DROSOPHILA 541 The tests give 19 high to 22 low ratios which is the expecta- tion for lethals, i.e., equality is expected and is approximately realized. Daughters from 2:1 cultures all of which were heterozygous for white and half of which should be heterozygous for lethal also, were again mated to white eyed males with the results in table 7. There were 13 high to 13 low ratios shown by these daughters indicating a lethal factor. On the basis of these data the locus of the lethal is at 23.7. It is interesting to note that the lethal factor occurred in flies that had been inbred a year and that none appeared in the stock having been inbred only two months. Miss Rawls (Biol. Bull. 13) found her lethal in a stock that had been inbred a year. The lethals described by Quackenbush (Se. 710) and Morgan (Se. 712; and Jour. Exp. Zodl. 714) appeared in stocks that had been inbred for some time. To test whether, in general, lethals are more frequent in inbred stocks I mated 100 pairs from wild stock caught at Falmouth, Mass., and 70 pairs from wild stock caught at Harris, Minn. The results are shown in table 8. Tables 1 and 8 show that the counts made of offspring from 270 pairs of fresh wild stocks have no unusual ratios. THE SECOND LETHAL FACTOR On February 10 of 1914, sixty pairs from stock collected in the summer of 1910 were mated. The results are shown in table 9. The next to the last pair of the above table seemed to show the presence of a lethal factor. Sixty virgin daughters from this pair were mated to brothers. Nearly one-half of these gave a ratio of twice as many females as males, as shown in table 10. Virgin females from numbers 7, 8, 12, 23 and 59 were mated to white eyed males. About one-third of the counts of the next generation show the presence of a lethal factor (table 11). Some of these lethal females were again mated to white eyed 542 MARY B. STARK TABLE 8 (a) Falmouth Stock FEMALES MALES FEMALES MALES FEMALES MALES FEMALES MALES 98 111 129 slit 161 132 175 182 144 132 207 216 175 176 171 174 74 78 181 154 195 187 191 190 68 86 182 193 101 118 199 223 129 120 196 205 102 85 100 88 120 117 179 187 154 162 99 106 119 138 191 175 195 190 185 180 178 . 174 105 97 191 198 100 84 176. - 175 184 181 199 202 174 194 96 103 173 138 150 150 87 128 130 109 186 150 166 183 91 90 156 141 174 169 110 123 98 86 103 8s 170 197 175 177 100 108 125 152 187 229 176 180 100 93 105 95 220 182 192 171 146 136 84 92 186 157 189 72 108 110 140 188 95 110 175 193 100 92 128 151 192 200 159 173 85 102 127 139 174 171 171 159 100 101 157 193 113 111 177 173 172 160 120 145 167 119 166 184 183 189 109 106 123 120 117 81 192 189 102 100 191 210 159 163 176 180 148 125 134 129 110 107 164 170 167 155 106 102 191 199 iy 175 males. It was expected that only a few of the red eyed males would appear in the second generation if red and lethal were closely linked. Results (table 11) show, however, that a great number of red males appear. Such a result indicates that a new lethal had appeared which was located some distance from the factor for white, thus giving a chance for a greater amount of crossing over. In table 11 it will be noted that the percentage of crossovers is 46. This would place the lethal somewhere beyond the factor for sable and not far from the factor for bar eyes upon the sex chromosome. LETHAL FACTORS IN DROSOPHILA 543 TABLE 8 (b) Harris Stock FEMALES MALES FEMALES MALES FEMALES MALES 69 166 137 142 83 95 126 126 117 96 176. 167 158 139 104 104 91 86 124 121 97 105 175 130 | 159 216. 115 125 140 139 163 132 105 98 94 76 180 170 115 79 138 120 199 196 84 87 116 115 154 155 93 81 183 185 153 . 153 115 89 163 135 157 167 93 75 107 78 165 135 104 105 138 120 104 Male 93 82 164 161 183 179 110 136 218 229 136 142 112 110 215 210 170 156 139 129 211 196 173 161 144 132 240 244 151 133 90 87 195 236 98 98 125 93 162 165 129 99 88 82 102 98 122 116 121 119 - 236 230 106 95 137 117 144 154 70 90 134 135 166 167 96 79 122 96 172 165 To determine the approximate location of the lethal upon the chromosome virgin females of the fifth generation (table 11) were mated, some to sable males, and some to bar eyed males. The counts of the second generation of crosses with sable and with bar eyed males are shown in table 12. The 1 : 1 ratios are omitted. The data in this table show that the distance of this lethal from bar is 8.3. If the locus of the factor for sable is 43 and that for bar is 57, the data of table 12 show that the locus of the new lethal is 43 + 23.5 or 66.5, since the distance of lethal from the factor for bar is only 8.3. The locus of the lethal must be ‘to the right’ of both factors since 43 + 23.5 or 66.5 is approximately equal ° 544 MARY B. STARK TABLE 9 FEMALES MALES FEMALES MALES FEMALES MALES 146 151 164 161 164 169 157 168 82 125 148 131 83 78 183 136 266 239 88 99 197 193 201 216 95 104 114 93 167 160 108 103 114 111 133 129 76 78 96 106 112 140 140 138 88 86 106 86 190 213 93 120 116 97 105 85 114 95 169 173 100 fe 155 230 199 185 176 152 106 111 181 205 131 155 98 103 145 154 102 84 142 119 106 96 160 133 91 64 110 96 171 153 138 110 181 197 250 188 143 135 195 178 252 216 152 135 142 162 150 163 180 149 266 108 139 105 148 134 86 5) to 57 + 8.3 or 65.3 or at 65.6. The locus indicated by both experiments when the data are weighted proportionately and a correction is made for double crossing over is 66.2. In the spring of 1914, an interesting lethal turned up in my 1910 stock. Half of the males hatched out as normal males but of the other half, though able to pass through the different stages of metamorphosis, many of them were unable to escape from the pupa case. Those that chanced to do so fell over on one side when trying to walk. I examined all the appendages carefully but noticed no abnormalities. Nevertheless, the legs did not seem strong enough to support the body, nor did they seem to move codrdinately and for that reason would so often become entangled with one another that the fly could not get them separated and would die from exhaustion in a day or so. Not any of these males lived longer than two days. Whether the first lethal of the 1910 and the lethals of the 1911 flies allow any development of the lethal bearing male at all, is still under investigation. LETHAL FACTORS IN DROSOPHILA NO. OF PAIR FEMALES 1 211 2 197 3 161 4 123 5 201 6 64 a 277 8 253 9 63 10 112 11 96 12 244 13 155 14 69 15 72 16 119 17 168 18 val 19 38 20 90 21 213 22 70 23 208 24 117 25 64 26 85 27 69 28 88 29 231 30 217 MALES 99 87 91 83 100 76 129 125 44 67 48 116 92 67 73 69 102 50 12 36 141 60 90 74 61 60 35 69 123 119 SEX RATIO |NO. OF PAIR| FEMALES ee ORE EE OWE RB eB RB BP PE NON RP REP NWS ON KF Wb bo TABLE 10 OIL ale souls .58: .05: Al: 96: 28: .88: ES2E fee i i i tt rt ol 32 33 34 35 36 ov 38 39 40 55 112 64 58 164 73 27 103 56 191 193 148 MALES 49 54 103 dl 78 43 128 50 27 78 = FON NaAN® WwW bw oOnw bw oF FP CO ~~ or @ We) 545 SEX RATIO oNoW SOR M ora) Soe 19: DEFOR OrFPNONHHHRP HB HNN RNH YP NYP RP HEH OF Fb S (=>) 19: — ~J TN Since no lethal has been found in the hundreds of pairs fresh wild stock examined and since lethals have occurred many of the inbred stocks, it may appear that inbreeding (with its constraints of confinement and homogeneous feeding) tends to cause the mutation to appear or else the removal of the flies from the competition that takes place under wild conditions makes possible the preservation and continuance of any lethal factors that may appear. Pe Ce eC ee a ee ee ne ee ee ee ee ed of in A discussion of these alternatives would involve a fuller knowl- edge than we have at present of the causes of mutations and the TABLE 11 1910 Lethal o's No. of Pair 40 an No. of Pair 40 n~ 7 (277 4 (178 1 (208 ve 7 (169 6 (229 10 ( 52 12 (258 13 (207 gi@5se 125) ( 7 (146 | F, | Qs as | 12 (244 (266 108) tie) es } 4 (189 L 8 (170 = pena 23 (208 go) <= | 8 (228 [4 (228 x W gg) XW_, } 10 (210 [59 (193 (13 (212 * The 1:1 ratios are omitted. 546 101) ass 105) TABLE 11 = = 3S RQs Ros Ws W o's Ss R@Qs_ Ro’'s WQs Wo's = = 1 (148 56 135. 77) 1 (148 4D) 166 123) 3 (128 29 10g 61) wee we ICS iGeme 132 me gs) 4 (220° 92 124 135) XW_, 6 (132 42 92 68) fe 7 GsoeeeGt). 1218101) 7 (160 63 150 80) a 8 (158 61 152 79) cere en 5am 637) 4 ( 88 26 78 51) he 4 tae i eaWieae |p Oars 96 64) 6 (05) 45, 92mm 43) 8 ( 65 20 76 39) \ 9 (11S 61 93 61) 7 (120) 944 122. 66) 6 (175 64 201 95) 9 (109 66 118 46) 3 (151 52 135 82) iG@soeeioo” 135 > 68) Bist 59: | (1 85) 3 (188 76 145 97) 62)" 47 11m aez3) A(GiNeNG?: 16280) =" isoms 253 «1d 7160 57 165 84) 8 (124 57 112 58) 9 (128 61 137 68) 10 (101 42 80 56) 1 (138 68 146 77) Ciseemol. 118 58) 2u@i66==980->* 170 91) F@uGn 257 212° . 146) 1ONGAT 57% 108 83) Selle 47 96 62) 11 (158 «467 ~—Ss«139 79) 6(108 56 103 63) 12-2247 83 «9-194 86) pars 9.(35u 53 124 74) {5702005 460, 124 69) 11@24- 52 18 56) 2139). GO 72) 2 (121 45 106 72) 2 (109 34 139 81) 4 38 80 48) 8 46 90 57) 5) 67 112 63) 5) 49 78 70) 9 42 77 58) AN (Giicsemmie4y 115 63) GZ 95, 124 62) (3 9) (ia 49 131 57) eR 1OCCUAE2Ae 96 © 56) Sey iE (130 56 109 59) 1 1 13 (121 o4 123 69) mgs sg 111’ 71) = . & ) ow wo ie Bl 38 Seven 50) | 134. 76) SS 8(117 42 112 — 63) ae nL 9(124 65 113 65) Total no. of o's = 3053 No. of cross overs = 1405 The ratio of the cross overs to all the o's = 1405 + 3053 = .46 547 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 548 MARY B. STARK frequency of their appearance. For the present, therefore, the fact of the occurrence of the lethals in the inbred stocks is the one important result of this examination. THE PRESENCE OF TWO LETHALS Number 4 of 13, 2, of table 6 yielded a ratio of 27.33 :1. It seemed probable in this case that two different lethal factors were present, one in each chromosome. This might seem to prevent all the sons from developing since each son must get one or the other maternal sex chromosome; but the survival of the three males would be possible through the crossing over TABLE 12 (a) Heterozygous Bar Female * Bar Mate FEMALES MALES NO. OF PAIR CROSSOVER Bar Bar Noumal PERCENTAGE 1 284 108 5 4.4 2 214 93 13 12.3 3 120 47 3 6.0 4 187 86 12 12.2 5 173 72 4 5.3 6 146 64 8 Hibs il 7 187 78 8 9.4 8 247 122 5 4.2 9 138 9] 4 4.2 10 166 02 4 Gail 11 167 74 iil 13.0 12 183 75 12 13.0 13 123 75 10 12.0 14 230 129 10 CA 15 121 56 3 5.0 16 193 60 ac 10.0 17 241 107 10 8.0 18 159 84 2 2.3 19 143 77 9 10.0 20 139 50 4 8.0 1590 144 Total number of o's, 1734; Crossovers 144. Crossover percentage = 144 + 1734 or 8.3. LETHAL FACTORS IN DROSOPHILA 549 TABLE 12 (b) Red eyed 2 heterozygous for sable and lethal X sable 7 NO. OF PAIR SABLE NORMAL SABLE NORMAL Stain 1 45 40 35 6 15.0 2 111 101 80 20 20.0 3 167 oo 84 30 26.3 4 110 74 69 2A 24.6 5 162 104 95 37 28.0 6 ig 59 45 9 17.0 a 149 113 92 22 20.0 8 92 104 58 25 29.0 9 91 105 65 25 Py Ry 10 85 78 55 16 22.5 ll 134 115 78 29 Met 12 111 113 75 21 29-1 13 138 138 86 33 27.0 14 35 38 35 5 12.5 15 Dil 52 29 8 21.6 16 121 114 63 21 O5RS 17 54 ail 48 14 22.6 18 86 91 80 26 94.5 19 73 83 79 19 19.4 1254 387 The total number of males = 1641; Crossovers = 387. The crossover percentage = 387 + 1641 or 23.5. Therefore, the distance of lethal from sable is 23.5. of one of the lethal factors from one chromosome in the other, thus freeing one chromosome from its lethal factor. TABLE 13 13%2. 1 4 _ FEMALES MALES CROSSING OVER NO a ORE Ath ree | eee a 0 eee | BE RWEEND WHITE SEX RATIO Red White Red White TD etal 1 172 174 32 116 21.6 2.34 2, Pali 228 30 136 18.1 Pel 3 147 174 29 126 18.7 2.07 4 215 216 28 174 13.9 Dales 5 229 194 44 124 26.2 De 5 6 228 209 30 144 We 2 Pg ii vi 184 201 35 139 20.1 iol 8 214 212 25 147 14.5 2.47 9 193 155 38 138 21.6 1.97 550 MARY B. STARK Nine of the daughters were mated individually to white eyed males and gave the results in table 13. All nine daughters gave a 2:1 ratio which is expected if their mother had two lethals. Ten other daughters (from the very high ratio mother) were mated to their red eyed brothers! and gave the results in table 14. Six red females of table 13 were tested and gave the following results: ee ty FEMALES MALES CROSSING OVER NO.OF PATRAS. a et BETWEEN WHITE SEX RATIO Red White Red White AUST NES i 59 51 4 47 7.8 Of 3, 2 A 2 88 84 8 59 12.0 2268 3 87 66 1 72 15.4 Le Seca 135 24s il 90 80 ial 60 155 255 Die ae 2 73 79 10 73 12.0 its Bal 3 75 66 il 63 55,1 LOR a Three of the sisters by brother No. 1 failed to produce any progeny when transferred to separate bottles. Each of the other sisters, however, showed the presence of a lethal factor. Thus all 19 daughters of the female gave a 2:1 ratio. There can be no question but that the fligh sex ratio of the mother was due to two lethals. Virgin red eyed daughters of some of these females were then mated to white eyed brothers with results as shown in table 15. Since the mothers of the females used in table 15 were all heterozygous for white and for one or the other lethal, they would, when mated to red eyed males, produce two kinds of daughters; one-half heterozygous for a lethal and the other half 1 Tf the sons that came through were due to crossing over, then the X chromo- some that went into each is free from lethals and consequently they must be normal males. The normality of these three males was also tested by mating them to wild females. The sex ratio was normal. The daughters of the three males were tested individually and gave normal ratios. LETHAL FACTORS IN DROSOPHILA 551 TABLE 14 Ten sisters by brother No. 1 CROSSING OVER NO.OF PAIR | REDFEMALES | RED MALES | WHITE MALES | SEX RATIO | BETWEEN WHITE AND LETHAL 1 105 8 41 2.14:1 Moyer 3 224 23 108 1 (380531 6 5 64 6 18 2.1071 25.0 6 90 7 36 2.09 :1 16.3 a 35 2 7 3.88 31 22/2 8 38 2 6 4.75 :1 25 .0 9 36 3 12 2.40 : 1 20.0 Miss epeee 428 61 136 2 eee 30.9 Three sisters by brother No. 2 1 128 18 38 PA) SAL 32.1 2 176 12 78 20%: 1 13.3 3 197 ili 78 2ealeea 12.4 Massy. neo: 336 25 109 2 toes 18.6 heterozygous for white. The females heterozygous for white when mated to white eyed males would produce equal numbers of red eyed and white eyed males and females (table 15) except where the lethal may have crossed over to the factor for white (as was the case in numbers 5 and 1, starred in table 15).2. The females heterozygous for lethal when mated to the white eyed males produce the 2 : 1 ratio, also indicated in table 15. Matings were, also, made between the three males and daugh- ters of the sisters of the males with results as follows: (13, 2, 1, 4; Female by W male) F; Female by No. 2 male \ FEMALES MALES CROSSING OVER NO. OF PAIR SEX RATIO BETWEEN WHITE Red Red White UND) EDIE Mass 236 6 95 Qi ononl 5.9 1 172 16 85 aler(8 A 15.8 2 116 9 435% 22a UZ 333 3 125 a 40 yet = i 14.9 4 141 5 61 2). Veal 7.6 2 The other class resulting from such crossing over should contain neither the white nor the lethal, and one such female is recorded in table 15. BY MARY B. STARK TABLE 15 (Sister by No. 1 male), F2 FEMALES MALES CROSSING OVER NO. OR SPAT Rg |e ee a ATO: BETWEEN WHITE Red White Red White Sa ees en 1 72 80 69 78 1S O2EsI 2 63 48 49 49 Paseo: 3 18 8 2.2092 1 4 50 32 iPoone Ne 43 51 27 6 2.85 :1 18.2 (Sister by No. 1 male) 3 Fe ils 145 72 2.00551 2 141 15 2 00731 3 53 67 47 75 10021 4 30 42 32 41 100751 5 43 34 50 44 0.82 :1 6 161 34 67 44 WEG 2 Il (Sister by No. 2 male), F. 1 137 72 1.90:1 2 103 107 0.90 :1 3 72 86 81 74 1 (00% 4 164 86 2 .00)51 § 101 47 2 Anat (Sister by No. 3 male), 3 Fo i 101 85 64 14 2.40:1 19.0 2 29 24 24 20 1.00 :1 3 168 65 26021 (Sister by No. 2 male) Mass F2 1 97 69 77 70 12001 2 78 24 3} (010) 3 il 3 153 72 2.00 :1 The four daughters show the presence of a lethal. Summary: The nineteen tested females of 13, 2, 1, 4 (table 6) gave a 2:1 ratio. Other tests showed that the three males behaved like normal males. The explanation of her high ratio (2 :1) is that two lethal factors were present. LETHAL FACTORS IN DROSOPHILA 553 Table 16 gives some of the counts of the descendants of 13, 2, 1, 4 with white eyed males. TABLE 16 FEMALES seule CROSSING OVER eS BETWEEN WHITE ae SG as White AND LETHAL 162 85 14 79 15.0 228 209 30 144 7 @ 100 84 13 68 16.0 91 91 12 68 15.0 127 123 15 85 15.0 122 125 16 108 12.8 127 110 18 117 13.3 115 115 De 119 16.2 95 86 12 57 72 142 105 15 88 Lise 137 138 20 118 14.4 43 43 8 39 17.0 87 66 12 72 14.3 90 70 11 60. 15.5 108 ° 102 17 79 le SG, 72 65 15 53 22.0 76 eT 13 48 ; lee 120 140 30 120 20.0 119 100 99 90 24.3 130 98 18 69 20.7 149 133 22, 95 18.8 172 174 32 116 21.6 193 165 ae 38 138 ane 184 201 3D 139 Ail Numbers 5 and 1 starred of table 15 show a decided decrease in the number of white males. It looked as if the lethal factor had crossed over into the chromosome carrying the factor for white. To examine this, virgin females from number 1 were mated to white eyed miniature males as shown in table 17. A lethal factor connected with the factor for white is evidently present since all the white eyed females gave a 2 :1 ratio. The red eyed sisters of the white eyed females in table 17a should bear no lethal if, as the last table indicates, crossing over, 554 MARY B. STARK TABLE 17 (a) White eyed granddaughters of sister and brother by a white eyed miniature male NO. OF PAIR WHITE FEMALES WHITE MALES SEX RATIO 1 79 44 1eSiull 2 171 71 24 3 176 88 DAO real 4 182 90 ZeOi sal 5 187 88 Dealt 6 176 63 7) 8) I 7 139 48 Pas) 3. iL 8 182 72 Yh) Sek 9 170 104 L(G) 6 i 10 164 68 24 al had taken place, except when crossing over occurred again in the mother. This was tested (table 17b). TABLE 17 (b) Red eyed granddaughters of sister and brother by white eyed miniature male FEMALES MALES ~ NO. OF PAIR : SEX RATIO Red White Red White 1 48 59 20 43 1-541 9 76 78 74 58 116 cal 3 82 59 50 50 Te Aleeat 4 69 68 66 62 107 eat 5 76 52 67 51 108)31 6 65 63 80 61 0.90 :1 i. 88 53 29 59 1.60 :1 8 33 34 5 27 209 =a 9 42 53 6 35 2 32.2m 10 79 65 1 44 32200 ll 64 99 7 il 2 09a Four of the red eyed females gave a 2 : 1 ratio indicating that crossing over did occur in the mother. To determine whether this lethal is the new one or the original one whose locus was shown to be at 23.7 the white eyed females were mated to eosin? to find the locus of the factor in the chromo- 3 Eosin is an allelomorph of white. A female that has the eosin factor in one X and the white factor in the other X can be distinguished from a pure female with eosin in both X’s. LETHAL FACTORS IN DROSOPHILA 555 somes. If it is a crossover it should be found to have the same locus as the factor with the red. The white eyed females mated to eosin males gave white- eosin’ long winged females (with the factor for lethal): white- eosin miniature females; white eyed miniature males; and a few white eyed long winged males as crossovers. The white-eosin long winged females were again mated to eosin miniature males. This mating gave white-eosin long winged females with the lethal factor, eosin miniature females without the lethal factor, eosin long winged females and, white-eosin miniature females, eosin miniature males, white eyed miniature males, eosin long winged males and white long winged males. The single crossovers are the white miniature and the eosin long winged males, while the white long winged males are double crossovers. In table 18 some of the counts are given. The crossover value of white with the lethal involved is 15.6 and that of the lethal with miniature is 19.9. Therefore the locus of this lethal is at 16.7, that is, 1.1 plus 15.6. The locus of the original lethal was shown to be at 23.7, so that this lethal with a locus at 16.7 must be the new lethal whose advent led to the production of the high sex ratio. The mother of the high sex ratio carried in one of her X chromosomes the original lethal at 23.7 and in the other X chromosome, the one derived from the father, the new lethal at 16.7. If this female contained two lethals some of her descendants, should have one and some the other; and these two kinds should be expected to give slightly different linkage values with white. If, then, the values obtained from all of her descendants be plotted they should give a bimodal curve. Diagram 1 was made from such data except that the diagram does not include the data in table 18. In diagram 1 there is a strong mode at 15 which corresponds to the new lethal, but the mode which corresponds to the original lethal gives only a weak mode at 22; indeed the curve is not obviously bimodal, small number of determinations not being sufficient to distinguish clearly between the two lethals. The TABLE 18 FEMALES MALES LINKAGE LINKAGE NO. P Fosi Whi-e OF WHITE | OF LETHAL oF PAIR | White- osin | ‘eosin | Kosin | Eosin | White eosin | minia-| <>. minia- | minia- | Hsin | White ee Geena long ture ne long ain mane long long , W M 1 28 32 17 10 Dill 4 10 0 9.8 24.4 2 34 22 24 11 32 7 U Py Meters: 18.8 3 35 24 17 18 29 9 6 0| 20.5 14.0 4 39 35 14 9 29 4 9 0 9.5 21.4 5 81 62 69 36 87 15 23 il 120 19.0 6 66 82 42 25 74 17 20 3 |) le 20.1 7 31 11 25 7 25 4 5 0 11.8 14.7 8 34 30 26 17 41 2 9 0 3.8 17.3 G 39 42 22 10 42 5 8 0 9.0 14.4 10 48 48 51 27 59 9 18 0| 10.46 20.9 11 28 40 24 15 25 5 10 0 12.5 25.0 12 47 44 19 16 48 6 15 0 8.7 Pall afl 13 17 23 14 17 26 6 10 0 14.3 24.3 14 44 24 21 il 35 6 10 V0 Wile 19.6 15 88 62 62 31 63 12 23 0 1222 23.4 16 93 78 37 30 | 100 10 30 1 8.0 22.0 17 113 120 63 80 101 25 37 1 15.6 23.1 18 104 105 70 78 83 21 33 3 ee 23.5 19 58 42 39 34 49 15 17 3 21.4 23.8 20 87 73 51 46 78 Le 18 il 15.8 16.6 PA 60 60 41 38 44 9 13 1 14.9 20.9 22 60 44 38 22 47 16 18 2 21.6 24.0 23 48 45 45 22 30 9 13 1 19.0 26.5 24 48 35 38 23 58 11 We 1 13.8 20.7 25 95 76 54 26 76 26 19 3 23.4 Ile. 7 26 58 49 39 29 56 13 14 1 16.6 18.0 ith 56 35 24 37 64 Zl 11 3 24.2 14.1 28 63 60 49 37 69 20 16 2 20.5 17.0 29 54 aad 48 33 47 4 11 2 §.4 20.3 30 52 39 38 24 55 13 19 1 15.9 22).6 3l 49 38 37 25 53 5 13 0) 7.0 18.3 32 53 54 29 16 49 2 9 0 3.3 15.0 - 33 65 45 47 30 71 14 15 1 14.8 15.8 34 107 54 67 38 100 PAI 26 3 16.0 19.4 35 101 54 67 29 78 21 28 0 16.4 22.0 36 69 48 52 30 63 16 23 2 17.3 24.0 37 58 24 66 25 50 8 9 3 S227 17.0 38 5Y/ 38 47 25 BY 17 17 0 18.7 18.7 39 63 45 41 3s) 54 18 13 1 22.1 16.2 40 28 28 19 16 39 8 12 0 13.5 20.3 41 60 36 27 26 28 6 4 2 20.0 15.0 42 64 34 51 32 71 15 23 i 14.5 21.9 43 26 32 45 25 51 8 13 2 13.5 20.2 At D2 50 46 34 57 24 10 1 2720 12.0 2198 | 2174 | 1780 | 1201 | 2421 524 685 48 6633 8656 LETHAL FACTORS IN DROSOPHILA 558 MARY B. STARK determinations of table 18 are all of the crossover value white new lethal, and are expected to give a unimodal curve with the mode at 15. SUMMARY Lethal factors were found in inbred stock only. Fresh wild stock gave no unusual sex ratios. The first lethal (1) was found in stock caught in 1911 and has its locus at 23.7. One female of this stock gave an ex- traordinarily high sex ratio; viz., 83 females to 3 males. That this extraordinary sex ratio is due to the presence of two lethal factors, one in each X chromosome, is shown by the fact that all (not half) the daughters gave a 2 to 1 sex ratio. The new lethal (l,,) that appeared in the female with extra- ordinary sex ratio crossed over, in one case examined, to the X chromosome that carried the factor for white. Its locus is at 16.7. Two other lethals were found in the 1910 stock. The first (1,.) has its locus at 65.2. The other (lq) differs from all other lethals in that-the lethal bearing males emerge from the pupa case, and die almost immediately on becoming adult flies. AN ATTEMPT AT A PHYSICO-CHEMICAL EXPLANA- TION OF CERTAIN GROUPS OF FLUCTUATING VARIATION JACQUES LOEB anp MARY MITCHELL CHAMBERLAIN I There is a general tendency to visualize the factors which determine the hereditary characters as specific chemical com- pounds. If we wish to carry this view (with which we sympa- thize) beyond the limit of a vague statement, we must either try to establish the nature of these compounds by the methods of the organic chemist, or we must use the methods of general or physical chemistry and try ‘to find numerical relations by which we can identify the quantities of the reacting masses or the ratio in which they combine. Attempts in this direction have been made by the suggestion of Loeb! that phenomena of growth belong in the group of auto-catalytic processes, and by T.B. Robertson’s? and Ostwald’s investigations supporting and enlarg- ing this idea; by A. R. Moore’s* attempt to show that in hybrids the velocity of development of the dominant character is slower than in the pure dominant breed; and by Loeb and Ewald’s! proof that all the embryos of Fundulus have practically the same rate of heart beat at the same temperature. Since our new experiments are a sequence of this last mentioned paper, we may briefly discuss its contents. tJ. Loeb. Ueber den chemischen Character des Befruchtungsvorgangs Roux’s Vortrige und Ausitze, Leipzig, 1908. Biochem. Ztschr., 2, 34, 1906. 2T. B. Robertson. Roux’s Archiv, 25, 581, 1908; 26, 108, 1908; 37, 497, 1913. Am. Jour. Physiol., 37, 1, 1915; Robertson and Wasteneys, Roux’s Archiv, 37, 485, 1913; Wo. Ostwald. Ueber die zeitiichen Higenschaften der Entwicklungs- vorginge, Leipzig, 1908. 3A.R. Moore. Roux’s Archiv, 34, 168, 1912. 4J. Loeb and W. F. Ewald. Biochem. Ztschr., 58, 177, 1913. 559 560 JACQUES LOEB AND MARY M. CHAMBERLAIN C. G. Rogers® has shown that the heart beat of the embryo of Fundulus has a temperature coefficient of the order of the magnitude of a chemical reaction, i.e., that it practically doubles for an increase of temperature of 10°C. Loeb and Ewald found that the rate of heart beat is practically the same in each indi- vidual embryo (of a certain age) for a given temperature, vary- ing only in very narrow limits; so that the rate of the heart beat of any of these embryos could be utilized as a thermometer. The authors explained this fact on the basis of general chemistry as follows: given a sufficient quantity of substrate the velocity of the reaction is in proportion to the mass of enzyme. If we suppose that the rate of the heart beat is determined by the velocity of an enzyme reaction—which supposition agrees with the temperature coefficient—we must conclude that all hearts of Fundulus embryos must have the same mass of enzyme, since they all beat at the same rate when the temperature is the same. If we consider the rate of heart beat of the Fundulus embryo a hereditary character—which is legitimate—we are forced to the conclusion that each embryo of Fundulus inherits practically the same mass of those enzymes which are responsible for the heart beat. The hereditary factor in this case must consist of material which determines the formation of a given mass of these enzymes, since the factors in the chromosomes are too small to carry the whole mass of the enzymes existing in the embryo or adult. II While the rate of heart beat is approximately the same in each egg (at the right age) and for the same temperature, we notice slight variations, the usual fluctuating variation. It occurred . to us that this fluctuating variation might offer a chance for further testing the enzyme conception of the factors of certain hereditary characters. We selected, instead of the rate of heart beat, the velocity of cell division. Loeb* had shown in a former paper that the time from insemination to the first cell division 5>C. G. Rogers. Am. Jour. Physiol., 28, 81, 1911. 6 J. Loeb. Pfliiger’s Archiv, 124, 411, 1908. GROUPS OF FLUCTUATING VARIATION 561 in the egg of the sea urchin Strongylocentrotus purpuratus can be so sharply measured and is so nearly constant that it can be used for the establishment of a temperature coefficient and this was later confirmed by Loeb and Wasteneys’ for the egg of Arbacia. Since the influence of temperature is again of the high order characteristic of chemical reactions, we may make the assumption that each egg carries a definite mass of one or more enzymes or catalysers which determine the rate of cell division. If we fertilize a mass of eggs of the same female of Arbacia and keep them at the same temperature, we find that they do not all begin to segment at the same time, and that there is an interval between the cell division of the first and last egg of the group. If we assume that the velocity of the cell division is determined by the mass of enzymes and the temperature, the fact that at t° some eggs divide after 100, others after 101, 102, until, e.g., 113 minutes, we must con- clude that this difference in time is the expression of a corre- sponding difference in the mass of enzymes in different eggs, those dividing in 100 minutes having a greater mass of enzymes than those dividing in 102, 103, ete., and 113 minutes; and that the mass of enzymes varies in inverse proportion to the time required for cell division at a given temperature. On this basis we should have to assume that the latitude of variation in the rate of cell division of a group of eggs is the expression of a corresponding variation in the mass of enzyme in the individual eggs. This idea can be put to a test with the aid of the tem- perature coefficient. If we call m the minimum mass of the enzyme responsible for the first cell division in the slowest eggs, then we shall find a certain greater percentage of eggs with the enzyme mass m + a, a still larger percentage with the mass m -+ ds, and a small number with the mass m + an where m +. Gn is the greatest mass of enzyme occurring in an egg. If the eggs with the mass m + an divide at the temperature t° after 100 minutes, they will divide in about Qi. X 100 minutes at the temperature (t — 10)°, where Qi: is the temperature coefficient for 10°C. at this point; the eggs with the smallest mass of enzyme 7J. Loeb and H. Wasteneys. Biochem. Ztschr., 36, 345, 1911. 562 JACQUES LOEB AND MARY M. CHAMBERLAIN m, which at t° divide after 113’ will divide at (t — 10)° after Qio X 118 minutes, since the temperature coefficient must be the same for both types of eggs. If we call the difference in the time of segmentation between the slowest and fastest egg the latitude of variation, this latitude of variation should vary in direct proportion to the temperature coefficient for cell divi- sion if our theory is correct. IIl We will first give the temperature coefficient of cell division for the egg of Arbacia for different temperatures; i.e:, the results of measurement of the time required from the moment of insemi- nation to the moment when the first egg in the field was seen to divide. The eggs had been kept in a water bath with constant temperature, and a little before the cell division was expected to oceur (which time we knew from the former observations. of Loeb and Wasteneys) the eggs were put into a watch glass of the temperature of the eggs and the exact time ascertained when the first egg of the lot underwent cell division. Table 1 gives these times according to Loeb and Wasteneys, and according to our own observations. The reader will notice how closely both values agree.’ Our values are the average of a number of ” determinations, which show only a negligible variation. Beyond 31° no segmentation occurs. We tried no experi- ments on the latitude of variation beyond 25° or below 9°, since outside of these limits the segmentation is no longer entirely normal. From the results of table 1 we compute the temperature co- efficients for the time from insemination to the first appearance of cell division (table 2). In order to determine the latitude of variation of the time of segmentation—i.e., the interval between the time at which the first egg of a set begins to segment and the time when the last egg segments for a certain temperature, we proceeded as 8 The eggs were always used in the first hours after they had been removed from the animal. The time required for the first cell division was remarkably constant in different experiments. It is worth mentioning that such constancy is only possible when the temperature is kept constant. GROUPS OF FLUCTUATING VARIATION 563 TABLE 1 Time in minutes from insemination to the cell division of the firstegg in Arbacia TEMPERATURE LOEB AND WASTENEYS CRATE degrees 7.0 498 .0 8.0 410.0 411.0 (0) 308.0 297.5 10.0 217.0 208.5 11.0 175.0 175.0 12.0 147.0 148.0 13.0 129.0 14.0 116.0 15.0 100.0 100.0 16.0 85.5 iieo 4025 18.0 68.0 68.0 19.0 65.0 20.0 56.0 56.0 21.0 53.3 22.0 47 .0 46.0 23.0 45.5 24.0 42.0 25.0 40.0 39.5 26.0 33.9 ; 27.9 34.0 30.0 33.0 31.0 37.0 follows: The eggs were inseminated in sea water, and kept in a water bath at the desired temperature. The eggs remained in this water bath until about the time when the first segmen- tation was expected to occur. In the meantime, a second water bath was prepared on the stage of the microscope whose tem- perature was slightly below that of the desired temperature. This water bath contained the watch glass in which the segmen- tation of the eggs was to be observed. The watch glass had therefore the temperature at which the eggs were observed. The temperature of this water bath was also kept constant. When the temperature at which the latitude of variation was observed was very low and that of the air of the room was high a slight error crept in, in as much as the temperature of the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 19, No. 4 564 JACQUES LOEB AND MARY M. CHAMBERLAIN TABLE 2 TEMPERATURE COEFFICIENT FOR 8/18 sa Wy, eRe 6.0 9/19 etal 4.5 / 65 ee 208.5 1 SSS 57 0/20 56 Bot 175 9 pe 21 =e BS 146 12/22 SS 8 /2 1G 3 129 13/23 eS / 45.5 116 14/24 see 100 15 eae 5/25 AG 5 water in the watch glass rose slightly during observation. This error made itself felt in that in the case of low temperatures the actual temperature was occasionally a trifle higher than intended. We shall come back to this point later on. When the eggs had been put into the watch glass, a field with no less than 80 and often as many as 150 eggs was selected, and every minute the number of eggs which underwent cell divi- sion was counted until the last egg had divided. Very often a small percentage of the eggs had remained unfertilized and these of course did not divide.’ In table 3 we give a few examples of the actual measurements of the latitude of variation in the time required from the segmentation of the first to that of the last egg in a field. As far as the irregularities in the first two minutes are con- cerned, they must probably be attributed to the fact that the entrance of the spermatozoa into the eggs occurred somewhat irregularly, the moment of insemination differing in various eggs within one or two minutes. Table 4 gives the latitude of varia- ® When this number was great the material could not be used since in such cases the spermatozoa no longer entered the eggs simultaneously. GROUPS OF FLUCTUATING VARIATION 565 TABLE 3 Latitude of variation in segmentation time TEMPERATURE NUMBER OF 25° | as? | 22° 2° | 125 | 122 EGGS SBG- MENTED SSS EE EEE” AFTER Number of eggs in field 117 127 116 Y 126) & 116 100 minutes il 3 1 1 2 4 3 2 12 6 8 24 15 5 3 34 15 21 49 26 8 + 68 34 33 85 40 10 5 107 44 85 95 51 12 6 10 eggs not 62 103 111 60 16 a fertilized 79 110 l?/ 67 19 8 90 116 119 UU 20 9 95 7 eggs not 80 24 10 100 fertilized 80 28 11 109 88 32 12 18 eggs not 88 36 13 fertilized 88 38 14 90 49 15 92 60 16 95 75 lg 84 18 100 85 19 101 20 105 85 21 105 95 22 106 96 23 108 24 8 eggs not 25 | fertilized 98 ‘ 2 eggs not fertilized tion, i.e., the difference in time between the segmentation of the last and that of the first egg in a field for different temperatures for all observations made. The averages appear in the last line. This series illustrates the source of error to which we have already alluded, namely, that at low temperatures the times 566 JACQUES LOEB AND MARY M. CHAMBERLAIN _ TABLE 4 Differences in minutes between segmentation of first and last egg in a field at g° 10° it 120 13° NE SS CEN I eC? || SON SPOR Iho Ol) Ose 50 39) 25) | 22 ame | 17 | PS 24 TOURS: 1 9 | 7s 49 AQM26 |, 2041) see! 19° | 20 TSO tale? |) Oaes AT 27 | 22 @aim| 16 \s12) a3 tie AG SHON eee eS 64 moe 18) | 12 12 9) 8} 7 60 — 20 14 12 ans 19 46 18 14 14 SZ 8 20 12 8 14 8 14 7 8 8 Mean 52.6 39.5] 26 |22.5] 19.2 |17.5} 13) 12/12.5) 9.6] 8 | 7.8] 8 | 8 | 5 Min. were liable to be too short when the outside temperature was very high. Thus the value 13 minutes for the temperature of 13° is unquestionably too low, and probably the values 46 and 47 for 9°C. are also too low. At the higher temperatures the values differ much less, since the temperatures approximate much more the room temperature. We are now in a position to compare the expected with the observed result. The expected result is the series of tempera- ture coefficients for the time from insemination to the time when the first egg of the set begins to divide; the observed result is the series of temperature coefficients for the latitude of variation, i.e., the time which elapses between the segmentation of the first and last egg ina set. These two sets of coefficients should be identical and table 5 shows the degree of agreement. A comparison shows that the temperature coefficients for the latitude of variation are practically identical with the tempera- ture coefficients for cell division, and that where a noticeable difference exists it is always in the same direction, namely, the coetiicients for the latitude of variation are a trifle too small. We can account for this on the basis of the deficiency in the method we have already discussed, namely that when the tem- perature of observation was low and that of the room high, the GROUPS OF FLUCTUATING VARIATION 567 TABLE 5 Temperature coefficients for latitude of variation TEMPERATURES EXPECTED FOUND . 9/19 LO 52.6 DEG oe 10/20 3.8 39.5 a= 3.9 1 PA Cee 26 See 8 9 99 5 12/22 Sail 28 Le 7.8 13/23 Dag 19.2 = 2.4 8 ‘ i 14/24 2.8 17.5 ae 8 15/25 2.5 * Desas temperature in the watch glass may have risen slightly during the observations. Since in the determination of the temperature coefficient the value for the low temperature forms the numer- ator, it is obvious that the observed temperature coefficients are liable to be a little smaller than they would be without this error. We expect to test this idea next season. THEORETICAL REMARKS It was found in a previous investigation that the time which elapses from the moment of insemination to the moment of the beginning of cell division in the egg of Arbacia, is a constant for a given temperature. On the basis of the enzyme theory this was to be explained on the assumption that the mass of ferments contained in the egg of the sea urchin responsible for this process is approximately constant in each individual egg. This would mean that the hereditary factor. determining the rate of cell division consists in determiners for definite quantities of ferments. This idea was put to a test by applying it to the fluctuating variability of this process. While for a given tem- perature the eggs of Arbacia will always begin to segment at the same time, not all the eggs segment simultaneously. Assuming 568 JACQUES LOEB AND MARY M. CHAMBERLAIN that those eggs which segment first have a greater mass of fer- ment than the others, fluctuating variability would in this case be due to differences in the mass of ferment in the different eggs of the same female. If this idea were correct, eggs with the maximum and with the minimum amount of ferment should differ in the rate of segmentation by an amount of time which would vary in direct proportion to the temperature ‘coefficient for the process of segmentation. This theory was tested and it was found that the observed values agree very closely with the expected values; the slight variations found being in the direc- tion of the possible source of error of the method of the experi- ments. These experiments support therefore the idea that the hereditary factor responsible for the rate of segmentation is a determiner for a given mass of certain ferments, and that fluctu- ating variability depends in this case upon slight but definite variations in the mass of those ferments in different eggs. SUMMARY OF RESULTS 1. It is shown that the temperature coefficient for the lati- tude of variation of the segmentation of the egg of Arbacia (i.e., the time between the segmentation of the first and last ege of a group fertilized at the same time) is practically identi- cal with the temperature coefficient for segmentation. 2. It is shown that the fact is intelligible on the assumption that the fluctuating variation in this case is due to a variation in the mass of enzyme contained in the different eggs and sup- posed to be responsible for the rate of segmentation. AUTHOR AND SUBJECT INDEX 4 BNORMALITIES occurring after conjuga- tion in Paramecium caudatum. Variation and inheritance of 387 Adaptation to high temperatures, on the heat resistance of Paramecium caudatum. The effects of certain salts, and of 211 Albino rats held at constant body-weight by un- derfeeding for various periods. Changes in the relative weights of the various parts, sys- tems and organs of young 99 Amphioxus during locomotion. The orientation of 37 Aphids. The predetermination of sex in phyl- loxerans and 285 Arey, Lestin B. The orientation of Amphioxus during locomotion 37 Ascaris. The effect of carbon dioxide on the eggs of 355 Asoxal reproduction of Hydra. Inheritance in the 157 “Bak EYE’ mutant of Drosophila. The ef- fect of selection upon the 515 Body-weight by underfeeding for various periods. Changes in the relative weights of the various parts, systems, and organs of young albino rats held at constant 99 BripeGes, Carvin B. A linkage variation 1n Dro- i 1 _ sophila } Bristle inheritance in Drosophila. I. Extra bris- tles. 61 N. Didinium nasutum. [. 225 (CALKINS, GARY ‘ The TV ife history Carbon dioxide on the eggs of Ascaris. The ef- fect of 355 Caudatum. The effects of certain salts, and of adaptation to high temperatures, on the heat resistance of Paramecium 211 Caudatum. Variation and inheritance of abnor- malities occurring after conjugation in Para- mecium 387 CHAMBER A'N, Mary MitcHett,Lors, JAcQusEs, and. An attempt at a physico-chemical ex- planation of certain groups of fluctuating variation 559 Chemical explanation of certain groups of fluc- tuating variation. An attempt ata ph ysico. 55 Curtis, Maynig R. and PEARt, RAYMOND. Studies on the physiology of reproduction in the domestic fowl. X. Further data on somatic and genetic sterility. 45 The life history 2 DeNiuM nasutum. I. The Scat Dioxide on the eggs of Ascaris. ee 5 of carbon Domestic fowl. X. Further data on somatic and genetic sterility. Studies on the physi 4 ology of reproduction in the Drosophila. A linkage variation in ; : Drosophila. I. Extra bristles. Bristle inheri - ance in ; 5 Drosophila. The effect of selection upon the ‘bar eye’ mutant of 515 Drosophila. The occurrence of lethal factors in inbred and wild stocks of 531 5 BScs of Ascaris. The effect of carbon diox- ide on the 355 Eye’ mutant of Drosophila. The effect of selec- tion upon the ‘bar 515 FACTORS in inbred and wild stocks of Droso- phila. The occurrence of lethal 531 Fishes in their natural environment to salts. The reactions and resistance of 243 Fission rate of Stylonychia pustulata. Heritable variations and the results of selection in the 451 Fluctuating variation. An attempt at a physico- chemical explanation of certain groups of 559 Fowl. X. Further data on somatic and genetie sterility. Studies on the physiology of re- production in the domestic 45 |G pee) length of spermatozoa in seven addition- al species of insects. Variation in 505 Heat resistance of Paramecium caudatum. The effects of certain salts, and of adaptation to high temperatures, on the 211 Heliotropic reactions of animals and plants. The relative efficiency of various parts of the spectrum for the 23 Hurcuison, Ropert H. The effects of certain salts, and of adaptation to high temperatures, on the heat resist.nce of Paramecium cau- datum 211 ore Inheritance in the asexual reproduction oO ] NBRED and wild stocks of Drosophila. The occurrence of lethal factors in 531 Inheritance in Drosophila. I. Extra bristles. Bristle 61 Inheritance in the asexual reproduction of Hydra. 157 Inheritance of abnormalities occurring after con- jugation in Paramecium caudatum. Varia- tion and 387 Insects. Variation in head length of spermatozoa in seven additional species ‘of 505 JACKSON, C. M. Changes in the relative e weights of the various parts, systems and or- gans of young albino rats held at constant body-weight by underfeeding for various periods. 99 JT ASHLEY. K. S. Inheritance in the asexual reproduction of Hydra. 157 Length of spermatoz-a in seven additional species of insects. Variation in head 505 Lethal factors in inbred and wild stocks of Droso- phila. The occurrence of 531 Life history. Didinium nasutum. I. The 225 Linkage variation in Drosophila. A 1 Locomotion. The orientation of Amphioxus during 37 Lors, JACQUES, and CHAMBERLAIN, Mary Mrr- cHELL. An attempt at a physico-chemical explanation of certain groups of fluctuating variation. 559 Lorn, Jacques, and WasTENEYS, Harporpn. The relative effieiency of various parts of the spectrum for the heliotropic reactions of ani- mals and plants. 23 69 ; 570 MAcDowELL, Epwin Cariton. Bristle in- heritance in Drosophila. I. Extra peiscles 1 Martoon, W. E., ZELENY, CHARLES and. The effect of selection upon the ‘bar eye’ mutant of Drosophila 515 Mipputeton, Austin RawpH. Heritable varia- tions and the results of selection in the fission rate of Stylonychia pustulata. 451 Moraan, T. H. The predetermination of sex in phylloxerans and aphids. 285 Mutant of Drosophila. The effect of selection upon the ‘bar eye. 515 NASUTUM. J. The life history. Didinium & 225 @CCURRENGE of lethal factors in inbred and wild stocks of Drosophila. The 531 Orientation of Amphioxus during locomotion. The 37 packsRD: Cuarues. The effects of the beta and gamma rays of radium on pro- toplasm. 2 PaIntER, THEOPHILUS S. The effect of carb.n dioxide on the eggs of Ascaris. 355 Paramecium caudatum. The effects of certain salts, and of adaptation to high temperatures, on the heat resistance of 211 Paramecium caudatum. Variation and inheri- tance of abnormalities occurring after conju- gation in 387 Peart, Raymonp, Curtos, MAyYNIE R. and. Studies on the physiology of reproduction in the domestic fowl. Further data on somatic and genetic sterility. 45 Phylloxerans and aphids. The predetermina- tion of sex in 285 Physico-chemical explanation of certain groups of fluctuating variation. An attempt at a 559 Physiology of reproduction in the domestic fowl. X. Further data on somatic and genetic sterility. Studies on the 45 Predetermination of sex in phylloxerans and aphids. The 285 Protoplasm. The effects of the beta and gamma rays of radium on 323 Pustulata. Heritable variations and the results of selection in the fission rate of Stylonychia 451 RAvIUM on protoplasm. The effects of the beta and gamma rays of 323 Rats held at constant body-weight by underfeed- ing for various periods. Changes in the rela- tive weights of the various parts, systems and organs of young albino 9 Rays of radium on protoplasm. The effects of the beta and gamma. 323 Reactions and resistance of fishes in their natural environment to salts. The 243 Reactions of animals and plants. The relative efficiency of various parts of the spectrum for the heliotropie Reproduction in the domestic fowl. X. Fur- ther data on somatic and genetic sterility. Studies on the physiology of Reproduction of Hydra. Inheritance in the asexual 157 Resistance of fishes in their natural environment to salts. The reactions and 243 Resistance of Paramecium caudatum. The ef- fects of certain salts, and of adaptation to high temperatures, on the heat 211 ZELENY, INDEX ‘ GALT, and_ of adaptation to high tempera- tures, on the heat resistance of Paramecium caudatum. The effects of certain 2 Salts. The reactions and resistance of fishes in their natural environment to 243 Selection in the fission rate of Stylonychia pustu- lata. Heritable variations and the results of 451 Selection upon the ‘bar eye’ mutant of Droso- phila. The effect of 515 Senay, C. T., ZeuENy, CHARwES and. Varia- ation in head length of spermatozoa in seven additional species of insects 505 Sex in phylloxerans and aphids. The predeter- mination of 285 Spectrum for the heliotropic reactions of animals and plants. The relative efficiency of vari- ous parts of the 23 Spermatozoa in seven additional species of in- sects. Variation in head length of 505 Stark, Mary B. The occurrence of lethal fac- tors in inbred and wild stocks of Drosophila. 531 Sterility. Studies on the physiology of repro- duction in the domestic fowl. X. Further data on somatic and genetic 45 Stockine, Rutu J. Variation and inheritance of abnormalities occurring after conjugation in Paramecium caudatum 387 Stylonychia pustulata. Heritable variations and the results of selection in the fission rate of 451 (PEMPERATURES, on the heat resistance of Paramecium caudatum. The effects of cer- tain salts, and of adaptation to high 211 UJ NDERFEEDING for various periods: Changes in the relative weights of the various parts, systems and organs of young albino rats held at constant body-weight by 99 VARIATION. An attempt at a _physico- chemical explanation of certain groups of fluctuating 559 Variation and inheritance of abnormalities oc- curring after conjugation in Paramecium caudatum 387 Variation in Drosophila. A linkage 1 Variations and the results of selection in the fis- sion rate of Stylonychia pustulata. Hertable Variation in head length of spermatozoa in seven additional species of insects 505 WASTENEYS, HarpoipH, Lorn, JAcQuEs and. The relative efficiency of various parts of the spectrum for the heliotropic reactions of animals and plants 3 Weights of the various parts, systems and organs ‘of young albino rats held at constant body- weight by underfeeding for various periods. Changes in the relative 9 WELLS, Morris M. The reactions and resistance of fishes in their natural environment to salts. 243 Wild stocks of Drosophila. The occurrence of lethal factors in inbred and 531 V7 ELENY, CHARLES and Marroon, W.E. The effect of selection upon the ‘bar eye’ mu- tant of Drosophila 515 Cartes, and Senay, C. T. Varia- tion in head length of spermatozoa in seven additional species of insects 505 te Kt el ee, as ees Pareteratepecn tates ean caeat eategtne, , pt ite era te tat phstee pare sett ps sean ro ht Taide sy Ey Cas y pat eee eh? " Shay pap Wt ee ice ee ey dy Fy" Bre heh sey.) : 1D 6 or Se oe) “eee tex ee Kes beg on ete bf TRO I89 wise » a4 supe eek ot bree te <* oe «