Lt collet eigheeoteh Maem iclaeanaap ete Pea rwer gem Ow a ct eee be ee ae eee see pareticor drow Spay sraine Cen TTR Oe WR eee aoe pad one b spo - Fe Poe Te pO A Nig eet ae cme aD ak ey Se eegtecterenen are irn ee a en eel , ae Al », Ley icy Mite See oS Saspihe nas ; ee wnt apes - Ay SRY A gl or re * - j Wea Sally: A tele As) i ae th ress 4 Binh Digitized by the Internet Archive in 2009 with funding from University of Toronto http://www.archive.org/details/journalofexperim25broo THE JOURNAL OF EXPERIMENTAL ZO0OLOGy EDITED BY Wiuiiam E. Castie Jacques LoEB Harvard University The Rockefeller Institute EpWwIN (Gi. CoNKLIN EDMUND B. WILSON Princeton University Columbia University THomas H. MorcGan Cuarues B. DAVENPORT ; eh Columbia University Carnegie Institution GEORGE H. PARKER HERBERT S. JENNINGS Haeard Unseiunty Johns Hopkins University RAYMOND PEARL FRANK R. LILLIE Maine Agricultural University of Chicago Experiment Station and Ross G. HARRISON, Yale University Managing Editor HO aa VOLUME 25 | 1918 | 6 1 /17 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. CONTENTS No. 1. FEBRUARY Auice M. Bortna AND RAayMonD PzarRL. Sex studies. XI. Hermaphrodite birds. Nine text figures and nine plates............--.-+--+++-s++-+> Ropert Stantey McEwen. The reactions to light and to gravity in Droso- phila and its mutants. Three figures...........----------+seeeeceees CG. V. Morritu. Some experiments on regeneration after exarticulation in Diemyctylus viridescens. Ten figures (three plates). 282. .c ae tele - Epuvarp Usntenuutu. Is the influence of thymus feeding upon develop- ment, metamorphosis and growth due to a specific action of that gland? J.M. D. Otmstep. The regeneration of triangular pieces of Planaria macu- lata. A study in polarity. Fourteen figures........-.-.-.-+--+--+++-- Manton Coretanp. The olfactory reactions of the marine snails Alectrion obsoleta (Say) and Busycon canaliculatum ((Bamitis eee xc eehet aah Sener Setic Hecut. The physiology of Ascidia atra Lesueur. I. General physi- ology. Witteemeneures).... 2.52.2 Hee. 2 ee ee ee ye 2 a Sexic Hecut. The physiology of Ascidia atra Lesueur. II. Sensory physi- OLOGYs | AOMINENIEES 005.2 Fen oe Fee soe nina re neeintnin cin ole ee atin maine No» 2 APRIL Cuester A. Stewart. Changes in the relative weights of the various parts, systems and organs of young albino rats underfed for various periods. Orie pe IR A ts ga ee c eRe aye coke a encase cine»

“(4a ies) V84e3) Sile2) OS OO-2 OO es Mests7..-: 21 22 23 24 25 26 27 28 29 30 Indices... 66.8 65.6 81.8 68.7 61.2 64.9 58.7 59.3 60.6 61.6 Mests: x-c: 31 32 33 34 35 36 37 Indices... 58.1 58.0 59.3 61.2 58.0 53.1- 55.5 48.0 50.6 46.8 Testsper 41 42 43 44 45 46 47 48 49 50 Indices... 45.6 52.4 49.9 52.4 49.3 52.9 48.1 55.3 45.6 43.0 ing all the groups in each series. This was done by averaging the corresponding tests in each group of the same age and sex (table 1, graph 1). Fatigue curves for flies of different ages Average of seven cards of flies 4-6 days old ote ee toe es ea ee Average of seven cards of flies 4-6 days old ~...Average of five cards of flies 16-18 hours old + Average of five cards of flies 16-18 hours old Per cent of phototropism 7] > 10 St 20 aoe 30 3S, 40 45 ar) Graph 1 It appears that the younger flies are rather less phototropic, or at least less active, than those 4 or 5 days old. Also they tire more easily. Furthermore, although in the younger groups the females are consistently above the males, in the older groups the males, after the first few tests are quite the equals of the females. This result is substantially in accord with that obtained in my preliminary tests. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 59 As has been noted, the older females gain rather less rela- tively in speed than the older males. This, it appears, might very likely be due to the fact that the female is gradually weighed down by the growth of her eggs. Indeed, it is known that these do not reach their full size for 4 or 5 days, the time varying some- what with the abundance of food. This suggested the experi- ment of running some insects of both sexes through a series of tests to be made daily for a period of two weeks, during which time they would be given fresh food each day. Three groups of flies with ten males and ten females in each group were selected for this purpose. Five trials were given to every set of ten flies at each of the testing periods, and the results of these five trials averaged for the period in question. This average has been taken as the index for the set of ten flies for this test. It is from an average of these averages that table 2 is made up. The tests for the first day were given at the ages of 4, 12, 15, 18, 21, and 24 hours. Afterward there was one test a day consisting of the usual five trials. The results indicated in table 2 and the cor- responding graph show that the females, though starting ahead of the males, fell away much more rapidly than usual, thus tend- ing to confirm the conclusion that a good share of their falling off with age is due to the increased weight of their ovaries. In- cidentally, it is again evident that the strongest reaction does not come at 18 hours. The next step was to run a similar series of daily tests for flies whose food was not changed daily. In this case a small amount of banana food was put in the vial in which a group of flies was kept and allowed to remain there. It gradually dried up so that the flies could derive less nourishment from it each day. That they must have derived some is certain, for Drosophila can not live 24 hours without food. Besides the drying, however, there is also a chemical change in food in which larvae are not work- ing. This, as well as the drying, tends after a few days to make the food unfit for the insects. This fact accounts for the death of those flies to which a little fresh food was not given on the sixth day. From this, as well as other experiments, it appears that 6 days is about the average time that flies will live under 60 ROBERT STANLEY McEWEN such conditions, though I have not infrequently had them live longer. For some reason not recorded no B and C groups of males were started. Hence this experiment is not particularly satisfactory. However, so far as it goes it tends to uphold the views already stated. The males gain on the females as age in- creases, and both sexes show a general increase up to the time when feeding was necessary. The results are summarized in table 3 and the accompanying graph. There remain to be described a couple of experiments which throw more light on the fatiguing effect of frequent tests given TABLE 2 Males themperavuressa-seeeeeres oie 20° mile Oe D3 2e 24° 24.1° PB | 24° | 22.5° Hours Days Age in hours or days.... 4 12 15 18 21 24 2 3 4 Group TNb, eater 8-6 ocrchatie 82.5 | 60.5 | 57.5 | 62.0 | 69.5) 72.5 | 71.0 | 84.0 | 83.0 BE e:, one ee ee eae es (SROM OS LOE EON O4eOR ECD eou| OS.Obln@2 Onl Sleds leo [ Or o> aR bake ga So20) old) | G2.571Ga20 5 | 89.5 | 95.0 | 96.0 | 92.0 AVC AL Che ashes (OES Obes GOO eI a 5 OeSaleniticor| (6.6 |e (Qual Sy sll oaee Temperature..... 21.0° 21.6° | DE yin 24.6° | 22.8° 215° PPG 20.0° 23.0° 24.5° Days Age in days..... 5 6 7 8 9 10 11 12 131 14 Group: Atel MOO SON OES GOROM IES Hal llviotas I iG2) i ILS. Sill zaesl a ets lean lm SG) Be. eS leon CORON eile On sOS250 6565146920) || 56500 c44e0n oS om mos Ceee ee SOLO AGIe on Ole On|nS9e5) |) 82-00 (he 8520) | SSei ez Geon eet OmmesonO Average...... COLON SLe Os eiGe2alShe9! |c7325 7220) | G424s e56eSalnioeD alee 1 On this day one fly from the A group of males was lost. As the records indicate that very probably the fly lost was a slow one, a new calculation was made on the assumption that had this fly been present it would always have remained in the zero section. The figures resulting from this calculation are 72.1 for the thirteenth day and 73.8 for the fourteenth. This changes the aver- ages for these days to 70.5 and 69.6 respectively. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 61 TABLE 2—Continued. Females Temperature..... Tie 20° 21252 23-27 24.0° | 24.1° 23.0° 24.0° PPL S 21.0° Hours Days Age in hours or VSien caganeice 4 12 15 18 21 24 Z, 3 4 5 Groups Acs. 92.5 |.69.0 | .90.0)| 7625.) 84.0 | 84.0 | 75.5 |.76.5 | 70.5 |.72-0 1B EN Goes 95.0 | 79.5 | 77.0 | 77.5 | 90.0 | 87.0 | 83.0 | 83.0 | 76.5 | 65.5 Gr O4IS 7 Ser lanOsoL ll eiSeor lkOdeos aoe. Ol meneon || Slo) (o-o I OlLeo Average...... 94.0 | 73.8 | 79.1 | 76.8 | 90.5 | 87.6 | 81.6 | 80.3 | 74.1 | 69.6 ‘emperaturesc.s s2ose-1 5 21.6° Date 24.6° | 22.8° | Pa ase Dee 20.0° | 23.0° 24.0° Days fAlreninidayssuts ses cee 6 7 8 9 10 11 12 13 14 Groups Lin 6 {re 65.0 | 58.5 | 56.5 | 51.0 | 38.0 | 33.5 | 33.5 | 44.0 | 43.5 13. eso Se ee 655m os ON eAteOL le S8e5r |) So20) eobron eaten! oS sonon -O CL. 5 eee 63.5 | 43.5 | 43.5 | 32.5 | 26.51 37.0] 18.5 | 29.0 | 29.5 Averipewers.......| 64.6.1-52.0 | 47.0 | 40.6 | 33.1 | 35.6] 26.5 [3711 | 36.0 Showing change in phototropism of wild flies when fed daily with fresh food Per cent of phototropism Hours °9 827% # Days A ~ Temperature at each test dlso indicated, 8 7 b Age of flies Graph 2 62 ROBERT STANLEY McEWEN TABLE 3 Males Temperature....| 22.7° leo Deve 215° Dy hs 24.5° PBL 21:6° Zi e1e | PP BIG Hours Days Age in hours or AYS ence 4 12 15 18 21 24 2 3 4 5L Group A: ; 7 flies)... 91.4 | 80.6 | 69.9 | Sil aey Hi SAP eteytall |) Pane) || LekDacch || Acc) |) We atss Temperature ss -7seeeer 23.0° 22.6° 21.8° | 23.0° 23.0° 22.6° PAGS PA pA? 24.7° Days Aven daysieeeeeeeeeeen: 6 7 8 gz 10 11 12 13 14 Group A: (eieSece.: eee GUSH Wisi 93.3 | 94.9 | 92.4 | 82.4 | 88.3 | 80.1 Females Temperature.....| 23.5° 22.4° 221° Qo 22.0° 23.0° 23.0° Deelie 23.0° 21.6° Hours Days Age in hours or RYS vce 12 15 18 21 24 2 3 4 51 Giflies Aes al S028) BOFS | iGEG Nie | 642986823) | 85.8) | 9626 |e 93esntOres 9 flies B.....| 99.4) 96.6 | 85.5 | 92.1 | 77.4 | 93.8 | 97.7 | 98.8 | 94.9 | 97.7 10 flies C.....| 100.0} 96.5 | 95.5 | 98.0 5.3 | 98.0 | 87.5 | 96.0 | 94.0] died Average......| 93.4) 87.9 | 85.8 | 87.5 | 79.2 | 88.3 | 90.3 | 97.1 | 94.0) 97.5 Memperatures.-.eeeeeeeer 21.2° (for B and C females only). Days Age LINGSVS 2254 eee 6 7 8 9 10 ll 12 13 14 (Gabi A oes ee ocacec| Sl@ |) 7e.83 CoO yee eel oils) 1) 837/44 |] SHS | che! Olflres Breese eee sell OGL OR aired IAVCLAPE CHEE Ee eee RSSaS 1 Indicates that food was changed at this point to prevent flies from dying. The drop on the sixth day is probably accounted for by this fact. ? Indicates only six flies in this group from the ninth day on. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 63 1a0 een s dad 6” Jey.” asl =~ A y skipped “ g \ ye ws 2 \ ~ 2 \ Bor \ \ a belenging 1 \ 2, 80 o. A y ° from th F) 45; i 2 2 = 70 \ SSS Ses 9 Q a 6s ° Z «| ed ' S ' 5 1 ' A ' — ‘ ' 48 p ‘ v u AS \ ] Died 74 “a IS 18 12). a+ 2 3 4+ 5 ‘ 7 3 y io “ut f2 1s (O/4# Age in hours Age in days Graph 3 to flies of different ages, as well as on the problem of time of maximum activity. The first of these experiments was designed to show the effect of testing flies aged 18 to 24 hours three times with an interval of 2 hours between the tests. Seven groups of insects were thus tested, each group save three containing sixteen flies, eight male and eight female. Groups A and B only contained six flies of each sex, and Group E only five. The usual three trials constituted a test for each group, the sexes as always being tested separately. The average of the three trials is given as the index for the test in question.2. The following summary of the results was obtained by averaging the indices of the seven groups (table 4). Along with this series of tests there was run a parallel series similar in every respect except that the insects were 9 days old instead of 18 hours. Only six groups were used in this case, but there were eight flies in every group. A separate single * It is to be noted that in this experiment as well as in the one on fatigue, the testing tube was only divided into four sections instead of five. The sec- tions were then valued as follows: 100, 75, 25 and 0. With this variation ealcula- tions were made as described under method II. 64 ROBERT STANLEY McEWEN TABLE 4 MALE FEMALE TEMPERATU RE First testi sce.c..ahc eee ec eee 68.9 Sie 243° Second! test:2.c aa aon ee eee 59.9 85.3 DAA Third testes. eee ee 55 83.5 DAR ie Average: see teen emer ene Sek eee 61.4 85.5 Difference betweenmmalesvand temaleshe 4]. oe ase eee eee ee oe eo all Rotalentimberkotmalestusedss.4- eee eee eee 47 Motalenumibercottemaleshused..... ae. aan ee eee 47 SECON Fea a ve ADR ee Aee Cis, Ae fs ais ties SRS ood a teehee a tC 94 group was put through the series at 6 days, and the results were practically similar to those obtained from the 9 day flies. For the sake of uniformity, however, they are not included in the following summary (table 5). TABLE 5 MALE FEMALE TEMPERATU RE RiTsiti GEsthr cc taot eo ae one Sere ee 71.6 78.3 PIs 1 Second testacncckec cine ee eee ee eee 74.8 83.9 sy Al EHindsGesittemosccsct< eee tata eee 76.8 83.3 DAE ANGLER Coen ais iss aiede a OE cian 74.4 81.8 Differencesbetween malestand temales:+2a-....c.s0-21--s lose eee. 7.4 Notalenum|berrotemsleswused sweeeiaretes 4 415 lcs joel ane 48 TNotalinumberottemalessusedsasneeec.. sc. sss.e ecco 48 FOUAR ee ae Re aera: todd dhcelnw site deena 96 A glance at tables 4 and 5 is sufficient to indicate the general results. It again appears that the younger flies, both male and female, are fatigued by successive tests, whereas the older in- sects actually improve. Also it is clear that although the fe- males are more active than the males in both cases, they are relatively less so in the older groups. Not only this, but the older females are absolutely slightly less active than are the females of the younger group. In this particular instance, there- fore, 18 hours is actually the maximum age of activity. It may be added, however, that were the results from the 6 day flies REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 65 TABLE 6 | MALE | FEMALE X | ss TTS hei MM SON eRe Rune oe a cea IC Lo eiaick | Sos (Sey evap 0G [GLASS Pe Lies se Le SR party IS Ns Se oe eae anne | 90.0 aalO0Z0 ARIE GES tae nos ee Aone oe a a Ae aoe | 93.1 | 97.8 included in this. average this statement would no longer hold. The indices for these insects were as follows (table 6): The average male index is 84.9 and the female, 98.5, with a difference of 13.6. Thus while the males have again gained relatively, the females are also absolutely much faster than are younger females. Finally, a series of tests were run on flies 4 to 6 days old as fol- lows. An initial test was run at the same time of day at which it had been the custom to remove newly hatched flies from their bottles. They were then tested at the same relative intervals as the newly hatched insects had been in the experiments de- seribed above. In this case ten males and ten females were used in each group, and the number of trials constituting a test was raised to 5. The results were as follows (table 7): TABLE 7 Male indices 0 HRS. 4 HRS. 12 HRs. 15 HRs. 18 HRs. 21 HRS. | 24 HRs. 83.6 86.5 88.5 86.6 86.8 96.6 | 98.8 Female indices 96.8 98.3 97.6 97.1 G6USIG ir 2S6eI ay GUa0 Temperatures | BIKE | 99° Die | D2 | Pile | De 99° The main feature of this series is that there is no drop occur- ring in the middle of the series, such as was the case with the young flies tested at similar intervals. It is thus made more likely that the falling off in question was due, as suggested, to the more rapid fatigue of newly hatched insects. Incidentally, it will be noted that the males are not far behind the females, and that they gain on them during the series. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. 1 66 ROBERT STANLEY McEWEN In conclusion, it may be said that females are never twice as active as males. They are, however, somewhat more active, particularly when only 1 or 2 days old. As age advances the difference between the sexes decreases until in some cases at 8 or 10 days the males actually surpass the females. Moreover, in- stead of the maximum period of phototropic response occurring at 18 hours, it would seem rather that both males and females, if not too heavily fed, increase their response with age, reaching a maximum in the neighborhood of 4 or 5 days. After this point both sexes tend to become less active, the females more rapidly than the males. It may also be added that the young flies fatigue much more rapidly than do older insects. EFFECTS OF OPERATIONS ON THE REACTIONS TO LIGHT a. Removal of wings The operation of removal of the wings was suggested by Dr. T. H. Morgan as a laboratory experiment for one of his classes. Mr. 8. Safir was the first to try the experiment, and obtained the rather surprising result that flies so treated no longer showed any response to light. This effect was so unexpected that it was determined by the writer to investigate the matter as thor- oughly as possible. The first experiments performed in this connection were under- taken with a view to determining whether the insects would recover their normal response if kept a sufficient length of time after the operation. As these tests were made at the beginning of the work no apparatus was employed except the tube and light from the north window. One fly was tested at a time and its record calculated according to method I, for instance, the fly was placed alternately in the end of the tube toward the light and in the end away from the light, and the algebraic sum of the average of the two sets of records in inches crawled was taken. as the index of the fly in question. Five groups of animals were tested, in which the number of insects varied from one to five. When there was more than one fly in a group, the index of the group as a whole was computed by adding algebraically REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 67 the indices of the individuals constituting the group. It is the indices thus obtained that are set down in table 8. From these results it would appear that there is a slight recovery of photo- tropism. Nevertheless, in view of what occurs in normal flies TABLE 8! DAYS GROUP 1 2 3 4 5 tl. fl. tl. fi. tl. fi. tie fl. tl. fl. eee eS 24.0 W036 9.3 8.6 4.3 ETE Ue Si le 00.7 155.63 12.8 2.0 Bee IL aed Sa Cope eee 10.0 8.9 es 18.5 9.9 IES i ee eee 4.3 1.0 00.7 00.9 3.0 Ein eon Munna 00.4 00.3 2.0 Al Tf 0.4 WHotalSinge nee 00.7| 38.7 43.1 42.1 2.0 | 32.7 19.9 Atverage. 2.0.2.2. 00.7) 9.6 8.6 Ate 2 (ileal. 3.9 Differences..... 8.9 8.6 8.4 6.1 3.9 DAYS GROUP 6 if 8 9 10 tl. fl. tl: fl. tl. fl. tl. fl. tl. fi. Li. oe ed 7.0 15.0 18.4 10.0 T«.....0 see 10.3 3.6 16.0 Ae ey eo eee Bs revel cs Vato) 4.2 1A Sol 00.7 LV = 2. 2h eee 10.3 Ye 5a aaa 18.6 Vl, tele ie ee 4.4 1.7 4.3 3.6 otal See 24.5) 6.1 | 10.6) 8.8 | 41.0 | 14.2 | 38.5 42.3 | 00.7 A:vverage........ 8.1] 3.0 Oeoleeeon | LOe2 |) 14.2 Teeth 14.1 | 00.7 Differences.....}| 5.1 yy a! 4.0 Meal 13.4 ' No temperature was taken in these early experiments, but later tests made in the same room under similar conditions showed a variation of less than a degree during a period of two weeks. tl. indicates excess of inches crawled toward light, resulting from the alge- braic sum of indices in the group. fl. indicates excess of inches crawled away from the light, calculated in the same manner. Diff’s. indicates the algebraic sum of the fl and tl averages. 68 ROBERT STANLEY McEWEN TABLE 8—(Concluded) DAYS GROUP 11 12 13 tl fl tl. fl tl fl To 2 SR eee ae 113)-3 22.9 27.3 Ls. ER ee ee reser eer 25.0 22.6 TLE 2 See eee ee ero 4.3 LV xs. ee ee eee ee eee ae NW ee dari ead is 6 ha eteue cue ous oaco or ame AG Calls ee: Aas Maisie cies eecnerarersietck 2 38.3] 4.3 | 45.5 27.3 AVGRA GES 7 tame aa sere hans Aes cence slay: 19.1) 4.3 | 22.7 27.3 Differencestarne ewer e ear aes tae. 14.8 22.7 27.3 with advancing age, it seems likely that this recovery represents nothing more than the usual increase. It should be noted that the usual index for unmutilated flies under this system of record- ing is from 30 to 36. This will appear in the next experiment. In this experiment, also under method I, five groups of flies were used. In the first group the wings were not removed until the sixth day after hatching, while in the fifth group they were removed on the day of hatching. The results were the same in every case. The insects showed the usual positive reaction to light until the day the wings were removed. At this point the positive reaction disappeared, the insects being indifferent to light and remaining substantially so for the rest of the tests. This occurrence was perfectly regular and very striking. It will therefore suffice to give only a couple of illustrations. In Group I there was only one fly. On the day after hatching, a t. |. in- dex of 30.4 was recorded. Its record was about this each day until the sixth, when the wings were removed. For the subse- quent five tests its average was f. 1. 1.7. On the sixth test it went up to t. |. 4.7, but on the seventh it dropped to t. 1. 2 and on the eighth and last the record was f. 1. 1. In Group IV there were five flies. On the day after hatching they averaged t. 1. 33.4. The day following their wings were removed, one insect having been lost in the meantime. On this, the third day after REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 69 hatching, they now averaged f. 1. 3.1. Thus, it does not ap- pear that the age at which the wings are cut off has anus to do with the effect. It should be added that the slight f. 1. excesses recorded in the above experiments are probably not significant. From watching the actual behavior of the flies it did not appear to me that the operation did any more than to render them practi- cally indifferent to light. Indeed, I have never observed a clearly negative reaction in Drosophila. So far the apparent loss of phototropism might mean merely that the operation had made the insects inactive. However, since Drosophila is strongly negatively geotropic it was possible to use this reaction as a measure of general activity. For this purpose the system of testing several insects at a time, known as method II was used. ‘The flies were introduced into the usual testing tube and given one trial for the reaction to light in the regular manner, except that no agitation was employed. Following this the tube was fastened in a vertical position with the flies at the bottom, and at a distance of 41 ems. from a 100 watt tungsten lamp hung so that its tip just touched the table. Three such tests were given, alternating with three light tests, and the indices for the two sets calculated as usual for the above method. The elimination of agitation in these tests was made necessary in order to make comparable the records of the flies with and without wings. When agitated the former move to- ® Regarding the relative strength of the two stimuli, light and gravity, Cole decides in favor of the latter. He found that when flies were placed in a ver- tical cylinder illuminated from below the larger per cent went to the upper- most third. Carpenter, on the other hand, was able to attract the insects to the bottom of a similar cylinder without using as strong a light as did Cole. On account of the great variability of Drosophila, I suggest that this diserep- ancy may be due to the small number of flies used, Cole employing only twenty- one and Carpenter only six. My own results are not strictly comparable with those of either of these authors, because I used a type of apparatus which did not directly oppose the two stimuli, but such evidence as I have agrees with that of Carpenter. Thus, a reference to any of the tables where the light and gravity indices of normal insects are recorded will show that the light-index in any given case always ex- ceeds that for gravity. In any event, the matter of which stimulus is the stronger is not one of any great significance. 70 ROBERT STANLEY McEWEN ward the light both by crawling and flying, whereas the latter can only crawl. When not agitated, however, the response even of winged insects is almost purely a crawling reaction. Three groups of flies containing 20 insects each were now se- lected at random from the stock bottles. They were placed in vials in the morning and tested according to the above plan in the afternoon. After this first test all the insects were etherized and the wings removed from the two groups which had made the best record. These groups will be designated as B and C. The following afternoon all three groups were tested again. Following are the records of Groups B and C before and after the wings were removed, and also the two records of the control Group A (table 9). TABLE 9 Before removal. Temperature 24° A B G Gravity Light Gravity Light Gravity Light (First test) Males Males Males Males Males Males 29.1 57.6 28.3 75.0 21.6 65.0 Females Females Females Females Females Females 316 81.6 45.0 90.8 32.5 85.0 After removal (Second test) Males Males Males Males Males Males 50.0 86.6 35.0 5.8 39.1 20.8 Females Females Females Females Females Females 41.6 90.8 30.0 Thea 25.0 SED As a further check, Group A, the control, was tested a third time 48 hours after the second test. The wings were then re- moved and eight hours later a fourth test was given (table 10). It is evident from this data that the removal of the wings affects the light reaction specifically, and does not merely reduce the activity of the insects. The next point investigated was the effect of the removal of parts of these appendages. The first experiment was done un- REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA iol TABLE 10 Group A. Temperature 23° GRAVITY | LIGHT Males Females Males Females Third test, before wing removal......... 46.6 30.0 88.3 84.1 Hourthr test atterenemovalesssreec rece ce 43.3 40.0 116 6.6 der method I, the age of the insects being unknown. The appa- ratus was exactly the same as that described in connection with the effect of removing the wings at different times after hatching. Five groups of flies were employed in the first series of tests. In the case of the first two groups only half of the wings were removed throughout the tests, while in the case of the last three groups one or two tests were given with half the wing removed, and then a second operation was performed in which three-fourths of the total wing was taken. Table 11 summarizes the results. It appears that though the insects are very erratic and vary much from time to time, those animals which had had the wing completely removed, with one single exception, made lower rec- ords than any made by flies with only half of the wings removed. These experiments are unsatisfactory, however, in failing to show what, if any, effect the removal of half the wing has. Later on, therefore, another set of experiments was devised to answer this question. In this case method II was used with the improved apparatus. Likewise, the alternating gravity trials were intro- duced as a control. In short, the general method was precisely similar to that employed in the proof that wing removal has a specific effect on the reaction to light (table 9). Four groups of flies were selected at random from the bottles and run through the tests. The flies were then etherized and treated in the following way. Group I, which contained nine males and ten females, had made the lowest record and was re- stored to the vials without operation. Group II, which con- tained ten males and eleven females, and had made next to the lowest record, had only the tip ends of the wings removed. Group III, which contained ten males and nine females and had the second best record, had one-half of the wings cut off, and V2 ROBERT STANLEY McEWEN TABLE 11 Rot sy ee cae ONE-HALF WING REMOVED eoeheata eka Meeks OF FLIES ; itsale fel. Galle felis 6 I 1 144.4 6 I 2 28.8 6 I 3 129.1 6 I 4 61.3 AV GRAS C: Acme the roe err: 90.9 6 IT 1 57.0 6 II 2 8.9 6 Ji 3 52.3 AWVCTODC 4 2 nue eoe temenras toe Ss 39.4 6 III 1 104.3 6 Itt 2 38.6 6 Til 3 8.8 6 INGE 4 24.2 AViCRAC CUR errr: hee eee rae 104.3 = lets! 31.4 6 IV 1 69.0 5 JY, 2 20.3 IV 3 B85 IA VIET ALE seh icy eee oe ee ae 69.0 26.3 6 V 1 48.1 6 Vv 2 46.4 6 Vv 3 9.1 6 V 4 10.2 4 Vv 5 tel IAWER ACOA. «2% ..5 eee eee 47.2 6.8 Averacestorone=lhaliimvyal eSpace reser td erici sci) ciel «Re ee eto ual eameZ OR A-verapertor three tOUnbnGnWwANGS. i826. 5. oss A. «ways celeateoe dekeee flee 127 Group IV, which had made the best record and contained ten males and six females had three-fourths of every wing removed. The next day all groups were retested with the following results (table 12): REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 73 TABLE 12 GROUP II. WINGS CUT GROUP I. WINGS NOT CUT ONE-FOURTH Light Gravity Light Gravity MiAlapi|fiartals | Medio" |:arills ||) Malos)| Boranla pileden lepers Before cutting: Le) LoL. 66.6 | 82.5 | 27.7 | 26.6 | 80.0 | 88.5 | 27.5 | 24.9 NemperacGueesss- seca ee | | SPE 23.5° Avgen canbe. 4. 0..0hy 911.6) /-9205)\-48 00 | 4803 | 76.6 | 72.5 | 40.0 | 36.6 Mempersturess-- eee 24° | 24° GROUP IV. WINGS CUT GROUP III. WINGS CUT HALF THREE-FOURTHS Before cutting.............| 80.0 | 95.3 | 41.6 | 42.5 | 90.8 | 95.8 | 49.8 | 38.3 Memperatures.-..2-5-552-- 23.52 | a ae After cutting......./......] 19.1 | 21.2 | 65.0 | 49.0 | 25.8 | 26.3 | 61.6 | 49.9 sllemiperaturess.a-a.e 2s. oe 24° 24° It is evident from these figures that with a single exception, such as the case of the groups in which three-fourths of the wing was removed, these results support the conclusion that the de- crease in phototropism is directly proportional to the amount of the wings removed. Moreover, in considering this result and particularly the one exception, it must be remembered that the amount of wing cut off in each group was directly proportional to the height of the index originally scored by that group in the initial tests. Thus, the group having three-fourths of the wing removed was originally the fastest of all, and this may account for its still retaining enough speed to win out over the group with only one-half of the wing removed. This seems particu- larly probable when this experiment is considered together with the previous one. Taking the two together, I believe we are justified in the conclusion that at least roughly speaking the pro- portion between phototropism and wing length holds good. 74 . ROBERT STANLEY McEWEN b. Gluing the wings Several attempts were made to glue the wings in such a way that though uninjured, the insect could not use them. These at- tempts were made fruitless, however, by the fact that a fly whose wings are stuck thoroughly enough so that they can not’be freed, will spend all its time in an effort to do so, and will scarcely re- spond to any other stimulus during the process. This experiment, therefore, had to be given up. There remained two other possibilities. First, the effect of — operations as such could be determined, by operating on other parts of the insect. Secondly, the existence in this laboratory of mutations of all degrees of winglessness made it possible to discover the effect of the absence, or partial absence, of wings in Drosophila upon which no operation had been performed. The effects of other operations will be considered first. c. Cutting off legs For this purpose eleven males and ten females, newly hatched, were selected and kept in vials until five days old, this being the usual procedure when records comparable with those made by other groups were desired.. Before placing in the vials each in- sect was operated on, and the tarsus and tibia of the middle pair of legs cut off. It was thought that removal of the middle pair in this manner would interfere least with the animal’s balance and ability to crawl. On the fifth day these flies were tested with a resulting index of 53.1 for the males and of 86.3 for the females. Under similar conditions it will be recalled that a normal index would be approximately 95 and 97, though I have cases where it was considerably lower. Thus, though there may be a slight effect from this operation, it is obviously not very great. Furthermore, it must be remembered that however quickly and accurately these flies might orient, they were neces- sarily handicapped in their speed of movement except when they took to wing. Since this experiment was performed under agi- tation this was frequently the case. As a matter of fact, orien- tation and movement toward the light was perfectly constant REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA t with these insects, a statement that does not hold at all for flies without wings. d. Removal of antennae The other type of operation which was attempted was the removal of the antennae. Inasmuch as these organs in Dro- sophila lie close against the head, considerable difficulty was ex- perienced in cutting them off. At length, however, a technique was developed which though tedious was successful in almost every case in removing the entire organ close up to the head. In this experiment four sets of tests were run which may be designated as A, B, C, and D. Since the method varied slightly in each of TABLE 13 Controls without antennaz removed LIGHT GRAVITY SINS Vertical Horizontal Male | Female Male |Female} Male |Female 1 SiO ole bl 27 Ole 2o20 SGOPA 2.2 --: eam cere etn: 2 [ 3 shemperature tren eee oe. ner te dees Dow Doe . 1 100 | 97.5 20.0} 48.0} 20.5) 35.0 Set B.. SARA Aaa eee ort a ( 1 41.5] 48.01 29.0 27.5 TemPerahUnegerae eet ees eee sc 23° 232 an ie te Set Co 24 nae eee ee rae ( ! a ice 33.0 sie 2 Temperature (Group 4) en aoe. 4 23° 23° | ‘ 1 36.6 Sees CUAL cn Mei ena Sra oa ae { 99.7 10.5 tpiaMye ceo2x). <2: iaeee ee ae 100 | 97.5 | 166.8| 135.5| 105.3] 82.5 Averavest 0 7 1) re 100 | 97.5 | 33.3) 44.5! 21.0] 27.5 Wierticalvexcessy, mailed. wud tid Some RD, © s:)6 5s autre nitetad ran elg cies Se 12.3 Wertichilsexcess: female... tas aeteeras erties Sac. Seether lenin 17.0 76 ROBERT STANLEY McEWEN TABLE 13—Concluded With antennae removed LIGHT GRAVITY eS Vertical Horizontal Male | F emale} | Male | Female} Male | Female 1 91.0) 66.5 0:5 1.0 6.0 Set Ac uikce Seu 2 86:0). 81.5). 11.0) 6.0)7 2725), 6:0 Ue 11.0 17.0 Temperstureisas ore oo cera eee 23° 23° ; detest a) i 1 | 92.5] 99.5] 21.5] 17.5] 17.0| 27.0 ee \ fi Temperature Be iad coe Ske Ao ae fe%c 24° DAS 1 77.5| 100.0 fa20) 1025) 12-0 Sei Osea | 2. 0).210s0 Be Se | 2 | 71.0| 90.1} 15.5| 16.5] 14.0 Temperature (Group b)...-.... 2B 230 . | } 2 1 | Also wings ; 16.0 13.5 Sethe) hic 3: Seon ee ee { 9 élipped 12.1 55 ‘Temperatureseoo 0. sees eee eee 23° Totals .csesesee os ee eee 418.0} 427.6} 142.1} 104.0) 140.0) 85.5 Averages). Beth aet est sean eee 83268525)! 1527) = Lied), oso Wierticalsexcessh an alle. a ree eee wn EPs ius nick. 2 has es meuorenens 0.2 Vertical,excess iemales.- sere one eee Plo. ot oe te eee 0.2 these tests, each will have to be described separately. The re- sults, however, will be found summarized in table 13. In Set A, two groups of twenty flies were removed from the bottle shortly after hatching. One group was operated on at once, while the other was kept as a control. Five days later the group from which the antennae had been removed was tested for light reaction. Three hours later both groups were given a gravity test. In this case this test consisted of two series of five trials each. In the first series the tube was held vertically, and the index calculated as usual, the result being designated as the vertical index. The second series acted as a control, the tube REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA as being placed horizontally and the result designated as the hori- zontal index. The difference between these indices may thus be taken as the index of the animal’s reaction to gravity. After these tests the flies with the antennae removed were given fresh food and kept for 3 days. The light and gravity tests were then repeated. On the following day, for instance, 9 days after hatching, the males of this group were given one more gravity test. At this point the alternating system for the ver- tical and horizontal positions of the tube was introduced, and used in all the subsequent tests of this experiment. In Set B three groups of flies were taken and from one group the antennae were removed. The other two were used as controls. Five days later one control group and the one from which the antennae had been cut were given the light and gravity tests as in Set A. The second control group was given only a gravity test. In Set C two groups of twenty insects were taken and from one group the antennae were removed as usual. At 5 days the animals which had been operated on were tested for ight and for gravity. The control was tested for gravity only. Three days later the former’ group was again subjected to both light and gravity tests. Set D consisted of two groups of male flies only, each having been used previously in tests on the effect of wing removal. One group, which we will designate as a, had been used as a control and had not had the wings removed. In the other group, b, the wings had been cut off. This latter group was now etherized and the antennae as well as the wings were taken off. The con- trol group was etherized at the same time but no operation per- formed... Four hours later a gravity test was given to each group, and 3 days later these tests were repeated. From the results of this experiment it appears that there has possibly been a very slight reduction in phototropic response. However, it is certainly in no way comparable with the reduction which occurs regularly as the result of wing removal. Further- more, a study of the light responses of normal insects contained in other tables, shows such variation that it is extremely doubtful 78 ROBERT STANLEY McEWEN if the slight falling-off of some of the antennaeless groups in table 13 is of any significance at all. From this result, therefore, as well as that obtained by the removal of legs we are led to con- clude that any operation as such is not sufficient to cause a loss of phototropism. Incidentally, however, a rather interesting re- sult does appear here as to the specific effect of the removal of the antennae and reaction to gravity. From the small amount of data on hand, it- appears that the loss of these organs greatly reduces a fly’s negative geotropism. It also seems to produce a slight reduction in general activity. There are not, however, sufficient data collected on these particular points to do more than suggest a line for further investigation. EXPERIMENTS ON MUTANT WING-CHARACTERS We are now in a position to attempt the second method of analyzing our problem by testing the various sorts of wing mu- tants which have arisen in this laboratory. These mutants vary all the way from vestigial, in which the wings are mere stubs to curled, in which the wings though of normal length are turned upward at the end and are not very effective in flying. There are many other variants between these two, one of which is desig- nated as strap. Strap has wings almost as long as normal, but they are narrow, often cleft at the end, held off from the body at a peculiar angle and are useless for flight. These three mu- tants therefore were selected as bearing the closest approximation to insects with one-fourth, one-half and three fourths of the wing removed. ‘The flies in question are represented in figure 2. A normal insect is also included for the sake of comparison. The tests on the above mutants have all been made ac- cording to the plan already outlined in one of the experiments for testing the effect of the removal of wings on light and gravity (p. 69). In brief, three tests were made for light, alternating with three tests for gravity, with the tube in the vertical posi- tion only. No agitation was used during any of these tests. Only one new point needs to be mentioned and that is in regard to an improvement in the apparatus. Under the former system of testing for geotropism the lamp whose rays struck the tube at REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 79 right angles was nevertheless near its foot. It was recognized that this method was unreliable when we attempted to compare the gravity reaction of flies which were phototropic with that of those which were not. Thus, two sets of flies, one photo- tropic and the other not, but possessing an equal amount of Fig. 2. A, Normal insect. B, Specimen of ‘curled’ stock. C, Specimen of ‘vestigial’ stock. D, Specimen of ‘strap’ stock. SO ROBERT STANLEY McEWEN negative geotropism would show the latter more negatively geotropic than the former. This would result from the fact that though equally impelled to move upward, the phototropic animals would be constantly handicapped by the attraction of the light from below. In order to remedy this difficulty, there- fore, three lamps of the same candle power were arranged in a vertical line, so that one came opposite the foot of the tube, one opposite the center, and one opposite the top. The end of each bulb was a distance of 41 em. from the tube. The latter, more- over, was now held in its vertical position by a wire support, so that there was no danger of its wabbling. With these improve- ments, the following experiments were undertaken. In the first place, it was decided to run a test on some wild flies in order to get some data which should be comparable with that obtained by the same system for the mutants. Also in mak- ing these tests it was decided incidentally to run a few checks on the effect of wing removal, in order to make sure that the former tests were not invalidated by the position of the single lamp. For this purpose three groups of insects, each contain- ing ten males and ten females, were selected, and kept in the usual manner until 5 days old. They were then tested as de- scribed above. After the test, the males in the groups which we shall designate as A and B were etherized and in the case of group B the wings were removed. Eight hours later both groups were retested. The control males in Group A were now also operated on, and tested for the third time 10 hours later. In the case of the females, Group A was operated on after the first test and Group B kept as a control. Twenty-four hours later both these female groups were tested again. Group B wasnow operated on and tested after eight hours. The results of these tests are summarized below (table 14). This experiment confirms the conclusions alr eee set forth re- garding the effect of wing removal. In other words, the re- arrangement of the lights has produced no significant effect. The only thing of note in the record is the extremely high gravity index registered by the males of Group B after being operated on. To determine whether this is of any significance or not will require REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 81 TABLE 14 BEFORE WING REMOVAL AFTER WING REMOVAL : Gravity ; Gravity TESTS y Light (vertical) Light (vertical) | Male | Female} Male | Female} Male | Female) Male | Female 1 96.6 | 98.3 | 68.3 | 77.5 Group! Aa. ease 2 96.6 66.6 41.6 50.0 3 5.8 45.0) Mempenrabunese smear ee s: BB 8" 23.5° 1 98.3} 99.1) 67.5) 64.1 Group Bae Bt 2 97.5 55.0] 52.5 91.6 3 35.8 | 10.1 80.0} 42.5 Temperature........ te eee 23.5° 23 Oe Soup Gee | i) Usa G4itiles5(0| 2508 | | Memiperature...2.. sh.) las. 24° 24° Totals...................| 375.6| 389.0| 237.4| 222.4] 94.1 | 51.7 | 216.6| 92.5 VENA ESHA eae) 95-9) 9722) 59F3) 5526) Sirs 2578 | 72.2 | 46.2 more data. It is true that a somewhat similar tendency is man- ifest among the males in table 9, but the poor light arrangement in the earlier experiments makes the records of doubtful validity on this particular point. Let us now consider the reactions of vestigial flies. Three groups of insects, half male and half female, were kept for the usual length of time after hatching, and then subjected to the test just described for wild flies. The only difference was that in this case the single lamp was used in the gravity trials. As will appear from the results, however, this feature was of no consequence in this instance because the insects were only very slightly phototropic. Also since the wings were already only stubs, nothing was cut off. Table 15 summarizes the results. Strap stock was next tested. Three groups, constituted asin vestigial and wild were kept as usual till 5 days old. The only irregularity in this connection was the use of only nine instead of ten females in Group A. As to the apparatus, the single THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. 1 82 ROBERT STANLEY McEWEN TABLE 15 Temperature 24° LIGHT GRAVITY (VERTICAL) Male | Female | Male | Female Group tA... eee ee ee 20.0 7.9 31.6 24.1 Grotp: Bosc. aah ere aeieees eee oe ae 13.3 10.0 24.1 20.8 Group Spec a ecae nce oe a eee 10.0 11.6 42.5 35.0 Totalsteer #ycrn-. bees Ser oe cere. fs 43.3 29.1 98.2 Ty a) ANGTAT Oh crties soca oN Ae a Re ee 14.4 Vel 32.7 26.6 lamp in the gravity trials was used in the first tests of the males in Groups A and B. After the first test the males of Group A were kept as a control, while those of Group B suffered the re- moval of the rather poorly developed wings which they possessed. Both sets were retested 8 hours later. Table 16 gives the results: The results from this experiment are enough to suggest strongly the slight increase in phototropic response which might be ex- pected to distinguish these flies from the vestigials. The most TABLE 16 | BEFORE WING REMOVAL AFTER WING REMOVAL ‘ ate ( | ss 22) | 2a pe a oe Male | Female} Male | Female) Male | Female} Male | Femaie Temperature 24° . 1 32.5] 17.6 | 45.0} 30.8 Ce eT ae { ibs BOND 71.2 é 1 42.5 45.0 Group Bee. f 9 63.8 68.4 Temperature 23° Group) Czes ee 1 10.0) 24.1 | 41.6} 48.3 ARoGal eee eee 147.0} 41.7 | 202.8] 79.1 | 63.8 68.4 Averages......... 3084 2028 | 50.4) 39-5) 6858 68 .4 REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 83 striking point which distinguishes these data, however, is the failure of wing removal to affect light reaction in Group B. Indeed, the test after the operation shows an actual increase in the reaction to light. In view of the rather low light indices in all the flies of this variety, I am inclined to explain this as fol- lows: Even normal insects whose wings have been removed, vary a good deal from time to time in their degree of response to light. Also the general variability is such that a fly with three-fourths of its wings gone will not always show a lower re- _ sponse than one with only one-half gone. I therefore suggest that since these flies already have a low index on account of their imperfect wings, the removal of the remainder of the wing might not have sufficient additional effect to counterbalance some un- known change in the physiological condition of the animal. This statement is partially borne out by the fact that the males in Group A which were not operated on, also showed a markedly higher index for both light and gravity in the second test. Further experiments now under way will serve to show whether this is the true explanation or not. If it is not, we should have to accept the rather astonishing hypothesis that a fly with short wings and a low index to begin with, actually has its photo- tropism increased instead of diminished by the removal of such wings as it has. Finally, we have to consider the reaction of the flies desig- nated as curled. As usual, three groups of insects, constituted as in the previous tests of this series, were tested when 5 days old. As this particular group was really the first of the series the use of the single lamp in the gravity test was still customary. The change to the new system was, indeed, made during the work on this group, which accounts for the fact that only the males in Group A were subjected to the improved treatment. That this feature was really of no great significance, at least in the case of these flies can be told by comparison of Group A’s record with those of the others. The males of both Groups A and B had their wings removed after the first test and were re- tested eight hours later. Group C males were not operated on, but were retested after the same interval of time as a control. 84 ROBERT STANLEY McEWEN In both Groups A and B one fly was lost before the re-test. Table 17 shows the results. For the sake of comparison the average indices for light and gravity obtained from the above experiments on wild, vestigial, TABLE 17 Temperature 24° ‘ BEFORE WING REMOVAL AFTER WING REMOVAL TESTS Light Geer Light eae Male | Female} Male | Female} Male | Female| Male Female : 1 83.3} 34.1] 86.6} 40.0 GroupgAea cee: i > 55.5 83.3 ee f 1 83.3] 51.6} 80.8] 45.0 Group Beas \| 2 69.4 74.0 CG 1 75.8) 48.3) 81.6) 49.1 wroup U......... 2 83.3 85.0 Potala 325.7| 134.0] 334.0} 134.1] 124.9 157.3 ; Averages......... 81.4; 44.6) 83.5) 44.7; 62.4 78.6 strap and curled flies have been brought together in table 18. From this table it is evident that there is a steady drop in pho- totropic response which in a rough way is directly proportional to the lack of development of the wings. The curious effect of defective wings upon light reaction as in- dicated by the above experiments made it desirable to examine TABLE 18 Temperature 23°-24° BEFORE WING REMOVAL AFTER WING REMOVAL Gravity Light Gravity AVERAGES FOR I ight Male | Female| Male | Female| Male | Female} Male | Female NST Le eri aes ican anes 8 sitios 93.9 | 97.2 | 59.3 | 55.6 | 31.3 | 25.8 | 72.2 | 46.2 Cuplediss: 4 :4:98: eesoeee 81.4 | 44.6 | 83.5 | 44.7 | 62.4 78.6 SRD! 2s Geoiae d eae 36.7 | 20.8 | 50.7 | 39.5 | 63:8 68.4 6 Vestigial..................| 14.4 ath Ways CO |) Ao): REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 85 these organs carefully in order to see if they might possibly contain any sort of light receiving structures. A number of minute or- gans were found (fig. 3). Except for the seven larger ones which occur well out on the veins, the majority are arranged in groups near the base. They have in fact very much the same arrange- ment and appearance as have the so called olfactory organs de- scribed by McIndoo for the honey bee (14). In order to discover whether these organs have anything to do with the reaction to light three groups of twenty flies each were selected and kept for the usual 5 days. In one group the Fig. 3. A B, Line of cut made to isolate the larger sense organs. E F, (OLD) Line of cuts made as a control, on veins which contained no sense organs. vein upon which occured the largest number of organs was cut as shown by the line AB. In the other group the veins were cut along the lines CD and EF, while in the third group no operation was performed, though the flies were etherized as in the first two cases. Table 19 gives the result. These operations were performed on the assumption that if the structures on the veins were light receiving organs, the nerv- ous connection for such organs must pass along these veins. If such were the case, then the insects which had those veins cut on which some of the organs occurred, should have been most affected by the operation. As a matter of fact, however, flies in which the veins were cut which contained no sense organs were as much affected as the others. Furthermore, it should be noted 86 ROBERT STANLEY McEWEN TABLE 19 La VEIN CUT NO OPERATION A B. (O/ID)5 913} 1045 Males Females Males Females Males Females Before operation..... 100.0 99.1 93.3 95.8 93.3 80.8 After operation...... * 43.1 41.6 40.8 46.6 92.5 83.3 that in neither case was that part of the wing injured where the chief groups of organs occur. ‘These points make it pretty clear that the effect produced on light reaction due to injuring the fly’s wing is not the result of injury to these particular organs. The fact that the stimulus for the light reaction is received in large part at least through the eyes and not through the wings or other organs is attested by the following experiment. In one of the mutant stocks known as eyeless, the compound eyes are very poorly developed, and in the case of many of the females are entirely lacking. Two groups of twenty flies each were, therefore, selected, one group containing only the individuals with the best developed eyes, and the other only those males with poorly developed eyes, and those females with no eyes at all. At the age of 5 days a light and gravity test was given these insects with the following results (table 20). TABLE 20 EYES PRESENT EYES POORLY DEVELOPED OR LACKING Light Gravity Light Gravity Male Female Male Female Male Female Male Female 97.5 65.8 65.8 46.6 74.1 25.0 70.8 39.1 The data seems to indicate that at least a large share of the stimulus causing the light reaction is received through the com- pound eyes, since when these are undeveloped the response is greatly reduced, whereas that for gravity remains approximately the same. We may now summarize the work which has been done on the relation of the fly’s wing to its phototropic response as follows. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 87 If the wings of Drosophila are removed, the insect’s response to light is greatly reduced. Furthermore, if they are partially re- moved, the reduction in response is roughly proportional to the amount taken off. That such a reduction is really due to a loss of phototropism and not to a general decrease in activity is proven by the fact that the insect’s response to the stimulus of eravity is reduced, very slightly, if at all. It now remained to show that the effect was directly due to the loss of the wings and not to the operation in itself. This has been accomplished first by performing other than wing operations and noting their effect and, secondly, by using breeds of insects which are hatched with imperfect wings. The operations performed involved the removal of legs and antennae. However, except in so far as general speed of locomotion was affected by the former operation, it could not be concluded that such injuries specifically affected the response to light. One incidental suggestion arising from these operations, however, is to the effect that removal of the an- tennae may materially affect the reaction to gravity. There is no obvious explanation for this, since Colet has shown that the stimulus of gravity is probably received through the leg muscles. The second method, namely, the use of vestigial, strap and curled wing flies gave results which still further bear out the hypothesis that it is the condition of the insect’s wing as such that in some way directly affects the response to light. The possibility that sense organs on the wings were responsible for this peculiar re- sult was tested by injuring the wing so as to break the nerve con- nection with some of these sense organs. It was found, however, that these organs had nothing to do with the response to light. That the stimulus for this response is received chiefly through the compound eyes was proved by testing eyeless stock contain- ing individuals with and without these organs. Finally, the notion that the effect may be due to a variation in the weight of the wing is made very improbable by the fact that the wings of curled insects, though deformed, are apparently just as large as those of normals. ‘W. H. Cole, The reactions of Drosophila ampelophela Loew to gravity and air currents. Jour. Animal Behavior, Jan., Feb. 1917. 88 ROBERT STANLEY McEWEN INHERITANCE OF PHOTOTROPISM IN DROSOPHILA In the Biological Bulletin, 1911, Dr. Fernandus Payne gives the results of some phototropic tests made upon Drosophila which had been bred for 69 generations in the dark. In the course of the work he discovered, as I have done, that there was great variation among individuals and he therefore made an attempt to test the inheritance of the reaction. He was unable, however, to obtain any significant results and this, so far as I know, is the only effort to study the inheritance of this reaction in Drosophila that has ever been made. It was with great interest, therefore, that I discovered that a certain stock of flies in this laboratory showed very little re- sponse to light. This stock was a combination of three separate mutants which were being carried along together for the sake of convenience. It is known as eosin, tan, vermillion. Eosin and vermillion are eye colors while tan refers to a slightly tan tinge to the body and a clear tan in the antennae. The antennae of the wild flies are gray. At first thought, of course, it appeared likely that the peculiar reaction to light was due to the light eye color. Stock in which these eye colors occurred without the tan were therefore secured and tested. Neither of these eye colors, however, were any less phototropic than normal. It is unnecessary to give the figures for the tests here, since experi- ments performed later in connection with colored lights amply demonstrate the phototropism of these breeds. The fact still remained that the peculiar reaction of eosin and vermillion might be due to a combination in one fly of the factors producing these characters. As tan had arisen as a mutation in eosin vermil- lion stock, the only way to test this was to get the tan separated from the two eye colors. By a suitable series of crosses this was ultimately accomplished. It was then possible to test tan by itself. The test immediately showed that either the factor for tan itself or some other factor very closely linked with tan, was responsible for the peculiar light reaction. Whenever an insect was homozygous for tan it failed to react to light. It may be stated at this point that tan is a recessive sex linked character. That is, the daughters inherit the factor for the character from REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 89 their fathers, but do not show the character, while sons inherit the factor from their mothers and do show the character.® The independence of the light reaction from the color as such may be realized from the following facts. Tan, like most other characters, varies about a mode. Furthermore, this particular character is so delicate, that at the extreme of-variation toward the normal color it is quite impossible to distinguish the indi- vidual possessing it from wild stock. This being the case, it sometimes happened that an insect whose genetic constitution was really tan would be accidentally mixed with flies which were supposed to be normal and vice versa. For example, one case occurred of a male fly which was apparently tan. Its light test, however, proved to be that of a normal. To test it, therefore, it was bred to a tan female. If the father were really tan, as it appeared, its daughters would all be tan. If, however, it were really normal as its light reaction indicated, then all its daugh- ters would be normal. The latter turned out to be the case. The daughters were. normal both in appearance and in light re- action. Later a case arose where several normal appearing fe- males reacted as though tan. They were bred to a normal male. If they had been normal then all of the offspring should have been normal; if they were heterozygous then all the females would have been normal but half the males tan. What hap- pened, however, was that all the females were normal, since tan is recessive, but all the males were tan. This proved that the females were really all homozygous tans as they had indicated by their light reaction, though their appearance had belied the fact. Thus it developed that light reaction was a surer test for the character of tan than was the color itself. It remains merely to give a table showing the records of three -groups of twenty tan flies each. They were tested ac- cording to the most recent system of light and gravity tests (table 21). Table 21 offers conclusive evidence that the failure of tan flies to respond to light is not due to any general inactivity. This is 5 For a full discussion of Mendelism in Drosophila, see the Mechanism of Mendelian Heredity by Morgan, Sturtevant, Muller and Bridges. 90 ROBERT STANLEY McEWEN TABLE 21 LIGHT GRAVITY (VERTICAL) Male Female Male Female Temperature 24° Group Arts. Ie eae ee 16.6 7.5 53.3 41.6 Temperature 23° Group {Binns cece ie pace ee 1.9 5.8 78.3 44.1 Group iG cates si ee fae eee ro 18.3 20.8 81.6 70.8 otal sarees Sete eee re ae ae 42.4 34.1 213.2 156.5 AV. GTARESE hs) Sea those aaa 14.1 11.3 iW 52.1 further borne out by observation of the insects. They are fully as active as are those from normal stock. Finally, a number of normal, white and vermillion eyed and tan flies have been sectioned and examined, both with and with- out staining. So far, no histological abnormality has been dis- covered in the eyes of the tan insects, to account for their peculiar lack of response to light. EFFECT OF COLORED LIGHTS ON. NORMAL AND MUTANT EYE COLORS A very considerable amount of work has been done by various investigators upon the effects of different wave lengths on or- ganisms which respond either positively or negatively to light. Though it has generally been found that animals as well as plants respond more readily to the more refrangible rays of the spec- trum, such is by no means invariably the case. In Daphnia, for instance, Lubbock (Journ. of the Linnean Society, 1881) and others have found that the green and yellow rays are more effec- tive than any of the others, including the blue and violet... Also in the case of simpler organisms, it has been shown by Engel- mann (Mast’s account, Mast ‘‘ Light and the Behavior of the Lower Organisms’’) that Bacterium photometricium tends to form aggregations in the infra red. The general apparatus used in this work has already been de- scribed. The composition of the liquids used, together with the REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 91 wave lengths transmitted by each, were as follows (table 22). The spectrum tests were made often enough to make sure that no fading of the colors was taking place. TABLE 22 FORMULA WAVE LENGTHS \WERWSIE, co ansas sone .285.0 ce. 5380 A° (green)—4240 A° (violet Violet Ammonia. 47...- - 15.0) Ce: Green strong at 4950 A° ee eS Copper sulphate Violet strong at 4510 A° (HSA) pesscodode 7.5 grams | Blue, weak Waiter ttets sien oe 30020)Acec: 5660 A°-5050 A° ; ichitierunes sere 0.03 grams | Strongest at 5320 A° Green...... = ae Napthol yellow... 0.25 grams Napthol green.... 0.03 gram Red Waters. emcees 300.0 ce. 7200 A°-6325 A° ars Ponceau Red..... 3.0 grams | Strongest at 6570 A®° The above formulae were only selected after a long series of experiments, and are for the most part modifications of formulae contained in the ‘“‘ Methods of Studying Vision in Animals” by R. M. Yerkes and John B. Watson, Behavior Monographs, 1911. The red and the green are very satisfactory for colors obtained by ray filters, while the so-called violet is evidently not so good. It is, as a matter of fact, continuous from green to violet. The blue, however, is very weak, the green moderate and the violet band very strong and wide. ‘The results of the experiments show that it is probably not the green to which the effectiveness of this filter is due, and since the blue band is so slight, the probabilities are that violet is the effective stimulus. It is practically impossible to get a strictly violet filter. We find, however, that blue is obtainable, and it is intended to use such a filter in analyzing our results further at the earliest opportunity. Besides the wave lengths, the relative energy transmitted by the filters was also measured by means of a thermocouple, using the same source of light employed during the experiments. The results are indicated in graph 4. From this it appears that if the energy transmitted by the colorless flasks be represented by 100 92 ROBERT STANLEY McEWEN per cent, then the red flask transmits 103 per cent, the green 64 per cent and the violet 51 per cent. The fact that the red flask actually transmits more than the white is explained by the fact that the layer of clear water was slightly thicker than the red solution. It will be noted, however, that in the visible spectrum the red is somewhat less than the white, while the green and the violet are approximately equal. The first colored light experiments performed were under- taken in the Zoological Laboratory of Western Reserve Univer- sity, Cleveland, Ohio, during the summer of 1916. After some preliminary experimentation it was decided to make use of so) Curves showing relation between galvanometer deflections and scale readings wanes Bortle , Area~—635.4 cm \ evi. te Light, Area—Sl.7 cr7> 130 420| Areagsinyisible Spectra White Lignt—/3 cin Red Bottle — 9.8cm* Green Bottle—/.7 com> Ui Blue Bott/e—/-4¢m* 9o Blue Bottle, Area—3!.5c7n go 4o re Ss 60| > vy u = ~ U = 50 Ps 5 so a a 40 uv ~ Q 5 re) ~ 3 S 20 adings Graph 4 Seale re REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 93 method II.° Four sets of flies were employed, each set consist- ing of six males and six females, the sexes being tested separately as usual. Every test consisted of three trials which were aver-_ aged to obtain the index for that test. Each of the four sets of flies was tested once for each of the three colors, the successive tests coming at intervals of 2 hours. For each set, however, the arrangement of the colors in the series was varied. Thus for set A it was red, green, violet; for set B, red, violet, green; for set C, violet, green; red; and for set D, green, red, violet. The results of these tests when averaged together for the four groups were as follows: males—violet, 64, red, 29.8, green, 24.5; females—violet, 81.8, red, 64.6, green, 57.2. In every one of the four sets violet was first in each complete test for males and females. As be- tween green and red, red won in three out of the four sets for both males and females. Thus it would appear from this ex- periment that the colors are effective in the order—violet, red, green. A further and better test was later made according to the fol- lowing method. Two groups of insects each consisting of ten males and ten females, were picked out and designated as Group A and Group B. Each group was now tested six times at two hour intervals, with three trials to a test. In this case, however, the three trials constituting a test were not all of one color. Instead, there was a single trial for each of the three colors in every test. Furthermore, in each of the six tests the arrange- ment of the colors was altered according to a set plan. Group B was treated in exactly the same manner, except that the se- quence of the color arrangements for each test in the series was reversed. Thus for test one the Group A arrangement was V, G, R; for test 2 V, R, G; for test 3 ha GeV ofortestetuney Ve Gs for test 5 G, V, R; and for test 6G, R, V. For Group B, the series began G, R, V and ended with V, G, R. At the end of the tests the indices for all the trials of a given color were aver- aged together in Group A and Group B. Finally the averages thus obtained for Group A and for Group B were averaged. * The tube in this experiment was only divided into four sections instead of the usual five. 94 ROBERT STANLEY McEWEN Furthermore a record was kept in such a way that it was possible to see in how many trials a given color came out first, second, or third. It is evident that in this scheme, since every color was used in every test, the effect of previous tests would not change the relative value for any color in any given test. On the other hand the possible effect of the arrangement of colors in any given test is overcome by altering the arrangement every test. Finally, the method of recording gives not only the average of all the trials, but an analysis of individual trials. It, therefore, seems that a tolerably clear-cut result obtained in this way may reasonably be supposed of some significance. This method was now applied in testing a series of mutant eye colors as well as the normal stock. The eye colors were as fol- lows: white, an eye entirely lacking in pigment; tinged, almost white, but containing the lightest shade of red; eosin, a reddish yellow somewhat darker in the females; vermilion, a very good sample of this color; normal; and sepia. The last named color is virtually maroon on hatching, but grows darker with age until at five days it is practically black. The results from the tests on these stocks are summarized in tables 23 to 28. Graph 5 is based on the results for each eye color as indicated in groups A and B combined. Since there is no apparent sex differentiation as regards reaction to varied wave lengths, the graph has been constructed from the average male and female indices in every case. It is evident from these tables that in the case of all colors lighter than normal, the general tendency is for the order of ef- fectiveness to be violet, green, red. There are, however, three exceptions. First in group A of white eyed males, the red is ahead of the green both in the average index and in the number of tests in which this occurs. This case is more than oftset, however, by group B, so that in the average of the two the order of colors is as stated above. It should be mentioned, moreover, that white eyed insects are extremely erratic even for Dro- sophila. It is quite usual for them after making a few con- sistent trips up the tube to become very much excited and to simply buzz about convulsively. The second exception is that Male..... Female.. Male..... Female. . Male..... Female. . of group A tinged females. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA TABLE 23 White eyed flies Group A AVERAGE R Sex Violet Green Red 2 Male...... 72.9 49.9 52e5 1 | Female... 48.7 38.7 29.1 3 il : 0 Renae a. Tied once 5 b. Tied once Group B 0 IMG escce 61E3 282, 155-1 0 Female... 64.7 AY) al 43.2 6b 1 ; ; : 1 Mate a. Tied twice. Watanls a. Tied once 4b 7 b. Tied twice b. Tied once Groups A and B combined 2 Malena 67.1 39.0 33.8 1 Female... 56.7 42.9 36.1 9b 2 a. Tied once 1 Male /2 Tied twice Henle b. Tied once Odulses sr \b. Tied twice '~ \e. Tied once 95 d. Tied once slightly exceeds the violet. The table indicating the number times each color won, however, shows that from this standpoint violet is still well in the lead. The third exception is found in group B vermilion females where red very slightly exceeds green both in the average index and in the number of times which it was ahead. There is no special explanation for this ex- cept the fact that the eye color is approaching that of normal. In any case group A more than overbalances it. Here the average green index very of 96 ROBERT STANLEY McEWEN TABLE 24 Tinged eyed flies Group A | AVERAGE SEX Vv G R Sex Violet | Green | Red ‘| Ist] 6 0 0 Male...... 100.0 , 100.0 79.1 Male..... 2d | 0! 6 0 Female... 98.3 99.1 97.0 [| Say OM eorulee | Ist | 4 asta 0 (a. Tied twice Female.. ;|2d |0 | 2 | 1 Female < b. Tied twice adel e224 eabaeoe _c. Tied three Group B fst | 40n20. 100 | Males-e- 100.0 GSi7aaal\ (Per Oad Male..... ¢| 2d | 2a| 6b| 0 | Female... 100.0 97.9: | 97.5 i Se tans eon 86 | Ist| 6 | 0 a ae Re ras : ; Female..{/2d|0 | 2 |1 | Male{® 764 UV Female (#7186 Free 3d |0 | 4a] 5b Si Groups A and B combined ist}10 |0 |0 | Male....., 1000 | 99.3 | 74.6 Male.....{|2a | 2d|12b|0 | Female..| 99.1 | 98.5 | 97.2 [|3a}o | 0 [12 | Est 10) 4] 080 (a "Tied twic ‘a. Tied twice Female..4| 2d |0 | 4 | 2 Male ! = = j as © Female , b. Tied five 3d | 2a | 7b |10c Seago Le. Tied six Turning now to the results obtained from the tests on normal and sepia, we find that the early records for normal made during the summer of 1916 have been confirmed. The order of effec- tiveness is not violet, green, red, but violet, red, green. There is one exception to this in the group A males where the green slightly exceeds the red. Finally in the case of sepia, there is the same reversal of the relative effectiveness of red and green. This instance, however, is more clear-cut than is the case with normal, for with sepia the red exceeds the green in all respects with absolutely no exceptions. REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA Q7 TABLE 25 Eosin eyed flies Group A AVERAGE SEX VA G R Sex Violet Green Red Ist | 4 0 | Male..... g 93 .5 87 .9 70.0 Male..... 2do 2a) e510 Female... 89.5 85 .7 70.0 Sys) @) WO. We | SiG) 5, 0) 0 : : Female..<| 2d} la} 6b] 0 Male Ve Foie Mia Female ee aoa ae Bar Oncor 6 ue i Group B iste 0 .| OAMMale... |) £2100.0 948 | 71.8 Male..... 2d | 2a | 6b} 0 Female... 96 .2 92 .0 70.8 Bol) ©, PO) WG Ist | 5 1 0 : Be Female..{|2d}1 |5 | 0 Maley Pears Bralp OFF nOa | 6 Ue Groups A and B combined Isr} 8 if 0 Male..... : 96 .7 91 3 70.9 Male..... 2d | 4a |1lb | 0 Female... 92.8 88.8 70 .4 SanieOr | O° 112 Ist |10 1 0 Female... | 2d | 2a |11b | 0 Male 3d | 0 ONL fa. Tied three \b. Tied three fa. Tied once Henle \b. Tied once In order to discover a possible cause for the phenomenon just described, sections of the eye were made and examined micro- scopically. As expected, these sections showed that the pigment which imparts to the organ its color, is simply the pigment usually found surrounding the rhabdomes in the compound eye of arthropods. This pigment so far as its function is known, is supposed to be of a protective nature, placed so as to absorb all rays of light which do not fall directly parallel to the axis of the rhabdome in question. Of course, in cases such as we have un- der consideration the pigment is not black but colored, and will consequently reflect light of a certain wave length. At first, THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 1 98 ROBERT STANLEY McEWEN TABLE 26 Vermilion eyed flies Group A AVERAGE 3EX Vv G R Sex Violet Green Red sta (Gee Ole Omala Viale seser ; 90 .4 72 .0 59 .6 Male..... Jd OF 4 1 Female... 94.1 94.1 60.8 3d | 0 2a | 5b MSE @ NeOr |-o : Female..<| 2d |0 | 6 | 0 Male fa ae ee eG au . Tied once Group B TSS NO! WO: |) MIE 6 : 100.0 97 .5 90 .0 Male..... Dag ssanleoby|) Ld |i@kemalei a: 95 .0 72.0 76.6 3d.) 0) | 3¢ |"5e ist] 6 |0 | 0 (a. Tied three Female..<| 2d | 0 4 N83 b. Tied twice i Tied 3d 10 | 4a/] 38b| Male;c. Tied twice Female : pie ae | Ato enies \b. Tied once le. Tied twice Groups A and B combined (|ast| 9 | 0 [0 | Male...... 95 .2 84.7 74.8 Males.... 2d | 3a | 7b | 2d | Female... 94.5 78 .0 68 .7 3d |0 | 5e |10e (| Ist}12 | 0 | 0 a. Tied three Female.. {| 2d | 0 Salto b. Tied twice (a. Tied 3d | 0 4a | 9b| Male 4 ec. Tied three Female een ht b. Tied once d. Tied once e. Tied three therefore, when normal eye color was thought to be the only one with an increased effectiveness for red, it seemed possible that this might be explained by assuming the red of the light to be exactly the same shade as that of the eye. This would then mean that a larger percentage of the red light entering the eye would be reflected and therefore effective, than would be the case with any other color. When it was discovered, however, that a still darker shade of red still further increased the effect- REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 99 } TABLE 27 Normal eyed flies Group A AVERAGE SEX Vv R G Sex Violet Red Green 1atnie4s, 10) ~ 10: ae Male... , 100 .0 96 .2 97 .0 Male.....2| 2d | 2a} 2b | 1d.) Female. .. 100 .0 99 .5 95 .4 || 3d | 0 | 4e | 5c sti On LO (a. Tied twice (a. Tied three Female.. {| 2d | 8a | 4d | 0 | b. Tied once b. Tied twice 3d | 2b) 2e | 6c Male /c. Tied three Female { c. Tied twice ° d. Tied once | d. Tied three le. Tied three Le. Tied twice Group B Tet 2.24 lO |e Mater: ..<.- 93.7 93.3 86.6 Male:...- 2d | 3a | 4d | 0 Female... 97 .0 96.6 83.7 3d |1b|0 | 6e 1st | 2 LNW ( a. Tied twice Female.. {| 2d | 4a | 5b | 0 Male: bs Liedionces ae ale Tied three Sel C1 O We ) ce. Tied once 3 b. Tied three | d. Tied twice Groups A and B combined StiGe i 2.) (WON WeMiale:ce a: 96.8 94.7 91.8 Male..... 4| 2d | 5a | Ge | Ic | Female... 98.5 98.0 89.5 3d | 1b | 4f |11ld Stileomel ele fa. Tied four Tied six Female.. {| 2d | 7a | 9d | 0 |b. Tied once Tied t || 8d | 2b | 2e |12c |e. Tied once iar ae Male : Female 4c. Tied two \d. Tied four : ? 3 d. Tied six |e. Tied three le. Tied two \f. Tied three ; . ‘veness of that color this theory had to be given up. Thus, at the present time I have no explanation to offer for the increased effectiveness of red light which appears to accompany the dark- ening of pigment in the eye, other than the vague assumption that it may be due to some physiological difference which occurs in connection with this change of pigmentation. 100° ROBERT STANLEY McEWEN TABLE 28 Sepia eyed flies Group A AVERAGE SEX Vv R G Sex Violet Red Green stro 1 0 Male. =sea- 93.9 87.4 66.0 Male..... Zama 5 0 Female... 93.9 91.0 68.0 Srl Oy PO |e Ist 3 0 : Female.. {| 2d |2 |3 | 0 Female ia Eee Ong: ; aq | dastton |b b. Tied once Group B Ist | 6 0 | 0 Males 2a 9353 70.4 54.5 Male..... Pxol--{| 6 0 Female... 96.6 86.6 55.0 3d /08 ONG iE OO IC Female. . 2d | 0 6 | 0 \}3d | 0 |0 |6 Groups A and B combined Ist |11 1 0 Male...... 93.6 78.9 60.2 Male..... 2 el a aie 0 Female... 95.2 88.8 61.5 3d | 0 0 {12 Ist | 9 3 0 : Female... ;|2d |2 |9 |0 Female [a. Tied paps a Pa as \b. Tied once SUMMARY AND CONCLUSIONS 1. Females of Drosophila ampelophila react to light somewhat more readily than do the males. This difference is most marked in young insects and steadily decreases with age, until at 8 or 9 days it has almost vanished. The time of maximum activity for both sexes does not seem to come at 18 hours, but more probably at from 3 to 5 days. 2. The removal of the wings causes the fly to lose most of its phototropism. The effect is specifically on the tendency to re- act to light, as is shown by the fact that such an operation REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 101 affects little if at all the response to gravity. The effect is roughly proportional to the amount of the wing cut off. It is not a result of the operation as such, since other operations do not produce it, and because wingless flies and flies of other stocks with defective wings show the same deficiency of response. Certain organs (fig. 3) occur on the wings of Drosophila, but operations fail to show that they are connected with the response Showing relative tropism of flies of varied eye colors toward lights of different wave lengths White Eyed RAG tf recon aa Red Pee igi aa Tinged Eyed Flies Violet Cc. Red Eosin Eyed Flies Violet . aa Red Ree Bernese Vermilion Eyed Flies | ee, Gen Red [SSR Semone eae ay Normal Eyed Flies PZ Eco < Red De Sepia Eyed Flies .-' TE er Cn Red ee Graph 5 102 ROBERT STANLEY McEWEN to light. It appears fairly certain on the other hand that the chief light receiving organs are the compound eyes as shown by experiments with eyeless stock. 3. Operations on the antennae may produce a weakening of the response to gravity, though they have little effect on the reaction to light. 4. In a mutant stock of flies known as tan, there is clear-cut evidence for the sex linked inheritance of a character which may be described as indifference to light. It is apparently not due to any structural defect in the eye. 5. Colored lights which may be conveniently described as vio- let, green and red, are effective in the order named upon insects whose eye color is lighter than the red eye of the wild fly. In the case of wild flies, and flies whose eyes are of a still darker shade called sepia, red is more effective than green. GENERAL DISCUSSION In most of the earlier work on various organisms, both animals and plants, the conclusion generally reached was that the blue and violet rays possessed much more stimulating value than uid those which are less refrangible, particularly the red and orange. Thus Payer (’42) using both the solar spectrum and col- ored media found that seedlings turned toward blue and violet light but not toward red, yellow, orange or green. Sachs in 1864 obtained similar results, using coléred solutions and glass. Also in the case of animals Engelmann (’82) found that Euglena viridis collected in the blue of a solar and gas spectrum, while E. B. Wilson (91) working on Hydra viridis with colored glasses again found the maximum effect in blue. Finally, Loeb’s ear- lier work (90-93) led him to conclude that as between red and blue, the latter color was the more effective for fly larvae, plant lice, caterpillars of Porthesia chrysorrhoea, moths of Sphinx euphorbia, Geometra piniaria, various copepods, the meal worm Tenebrio molitor, the larvae of Polygordius, Limulus polyphemus, and the June bug Melolontha vulgaris. | Even some of the early investigators, however, found cases in REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 103 which the above condition did not hold. Thus Kraus (’76) using colored media, discovered that in Claviceps, a fungus, red light was nearly as effective as blue, while Engelmann (’82) showed by the use of a solar and gas spectrum that Bacterium photo- metricum actually collects most readily in the infra red. Further- more, the study of various other animals with more refined methods began to show that many forms were most affected by intermediate points in the spectrum. “Thus Yerkes (’99) has shown that for Simocephalus the point of maximum efficiency in a Welsbach gas prismatic spectrum is in the yellow. Bert (69) and Lubbock (’81) located this point for Daphnia in the green, while recently Loeb and Wasteneys (716) using the spec- trum from a carbon arc, have found the most effective point for Balanus larvae in the yellow and that for Chlamydomonas pisiformis in the yellow-green. Likewise, Hess, an opthalmolo- gist, (10) using the spectrum from a Nernst glower has fixed ereen or yellow-green as the maximum stimulating point for a variety of forms, including ichneumon flies, Culex pipiens, adults and larvae, Coccinella septenpunctata, Dasychira fascelina and cephalopods. In the last instance the reaction of the pupil of the eye is taken as a criterion of response. Lastly, 8. O. Mast has recently given an excellent summary of work previously done and the results of a recent series of experiments of his own on Arenicola larvae, blowfly larvae and a number of. unicellular forms. For the blow-fly larvae the maximum is in the green, while for Arenicola it is in the blue. From these results it is apparent that the variation in the point of maximum response for different animals and plants is very wide. To explain this divergence, the existence in differ- ent organisms of different chemical compounds varying respec- tively in the degree to which they are altered by light of differ- ent wave lengths has been suggested. That there are com- pounds of this kind we know, but their presence in phototropic organisms has not yet been proved: Aside from this view Hess believes that phototropic animals are all color-blind, and that, they go to the part of the spectrum which seems to them bright- est. He apparently gets this idea from the fact that he found 104 ROBERT STANLEY McEWEN so many organisms for which yellow-green is the most effective part of the spectrum, this being also the brightest part for color- blind men. This notion, has been criticized by Ewald (15) and Loeb (16). The peculiar fact about Drosophila is the reversal in the ef- fectiveness of red and green as the insects’ eye color grows darker. Thus for eye colors lighter than normal the order of effectiveness is violet, green, red, while probably for normal, and certainly for sepia, the order is violet, red, green. This case besides showing the peculiar reversal is remarkable as being the only instance so far discovered among the lower animals in which red is more effective than green, with the possible exceptions of Daphnia (Frisch and Kupelwieser, 713; Ewald, 714), and paramoecium bursaria (Engelmann, ’82). How to account for this phenomenon of reversal it is difficult to say. Were it not for the case of sepia it might be explained on the basis of the changed amount of red reflected by the normal colored rhabdomes as compared with that reflected by those of lighter shades. When two shades produce the same effect, however, it is difficult to see how this will suffice. It would thus seem as though we must fall back on the assump- tion that as the eye grows darker, the supposed sensitive chemi- cal substances on which the light has its effect change also. What this change could be, it is hard to imagine from what we now know of photo-chemical reactions. I am inclined to think, therefore that the explanation may yet be found in connection with some sort of differential absorption. It may be noted that my results with colored lights do not agree in one respect with those of Dr. A. O. Gross who also worked on Drosophila This writer found green more effective than red for flies with normal eyes, while my experiments re- versed this order. I suggest, however, that this descrepancy is due to the fact that Dr. Gross used lights which were equated in energy, whereas in the case of my filters, as is also true for the normal spectrum, the energy of the red is much greater than that of the green. This fact, nevertheless, does not invalidate or make less interesting the very evident increase in the effective- ness of red in the case of the darker eye colors, since whatever REACTIONS TO LIGHT AND GRAVITY IN DROSOPHILA 105 the relative difference in energy content, that difference remained constant for all the eye colors tested in my experiments. ' Lastly, it may be well to emphasize the peculiar relation which exists in Drosophila between general activity and photo- tropism This phenomenon has been clearly recognized by Carpenter and in general I agree with this author’s conclusions. The fact seems to be that this insect is not phototropic unless it 1s in a certain physiological state brought on by, or at least accompanied by, activity. When the fly reaches a certain de- gree of activity, induced by various means, it suddenly becomes phototropic. When it quiets down, however, it may still crawl about but ceases to be phototropic. Thus, when an insect has been exposed to constant illumination for some time, it no longer orients to light but wanders aimlessly up and down the tube. Eventually such an animal may even come to rest with its head away from the source of light. This phenomenon, Car- penter suggests, is probably due to slight fatigue. However this may be, it is certain that without a continuance of the me- chanical agitation or sudden increases in light intensity, the ani- mal’s general activity soon falls to the point where phototropic _Tesponse ceases. \ ‘ 106 ROBERT STANLEY McEWEN BIBLIOGRAPHY Carpenter, F. W. 1905 The reactions of the pomace fly (Drosophila ampelo- phila Loew) to light, gravity and mechanical stimulation. Amer. Nat., vol. 39, pp. 157-171. 1908 Some reactions of Drosophila, with special reference to con- vulsive reflexes. Jour. Comp. Neur. and Psych., vol. 18, pp. 483-491. Coir, W. H. 1917 The reactions of Drosophila ampelophila Loew to gravity centrifugation and air currents. Jour. Animal Behav., vol. 7, no. 1, pp. 71-80. ‘ Gross, A. O. 1913 The reactions of arthropods to monchromatie lights of equal intensities. Jour. Exp. Zodl., vol. 14, pp. 467-514. Hess, C. 1910 Neue Untersuchungen iiber den Lichtsinn bei wirbellosen Tieren. Arch. f. d. ges. Physiol., Bd. 136, pp. 282-367. Loxrs, J. 1906 The dynamics of living matter. New York, 233 pp. Lors, J. AND WAsTENEYs, H. 1915 The relative efficiency of various parts of the spectrum for the heliotropic reactions of animals and plants. Jour. Exp. Zoél., vol. 19, pp. 23-35. 1916 The relative efficiency of various parts of the spectrum for the heliotropic reactions of animals and plants. Jour. Exp. Zodl., vol. 20, pp. 217-236. Lutz, F. E. 1914 Biological notes concerning Drosophila ampelophila. Jour. New York Entomol. Soc., vol. 22, no. 2.” Mast, 8. O. 1911 Light and the behavior of organisms, New York, 410 pp. 1917 The relation between spectral color and stimulation in the lower organisms. Jour. Exp. Zoél., vol. 22, no. 3, pp. 471-528. McInpoo, N. E. 1914 The olfactory sense of the honey bee. Jour. Exp. Zool., vol. 16, no. 3, April, pp. 265-346. 1916 The sense organs on the mouth parts of the honey bee. Smith- sonian Miscell. Coll., vol. 65, no. 14, Jan. Payne, F. 1911 Drosophila ampelophila loew bred in the dark forsixty nine Payne, F. 1911 Drosophila ampelophila loew bred in the dark for sixty-nine generations. Biol. Bull., vol. 21, pp. 297-801. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 22 SOME EXPERIMENTS ON REGENERATION AFTER EXARTICULATION IN DIEMYCTYLUS VIRIDESCENS C. V.. MORRILL Department of Anatomy, Cornell University Medical College, New York City TEN FIGURES (THREE PLATES) The earlier writers (Phillipeaux and Fraisse)! on regeneration in urodeles seem to have held the opinion that the extremities of adult animals are completely replaced only when one or more bones are injured in the amputation, that is to say, not after total extirpation (exarticulation). Wound-irritation from an injured bone was considered necessary as a stimulus to the re- placement of a missing part. Also the bone supplied a ‘tissue- rest? to serve as a matrix. However in young salamanders and especially in their larvae, it was found that the regeneration of extremities takes place very readily since here the joints are only partially formed and wounding of the bones always occurs in amputations. This general conclusion, that in adults re- generation does not occur after complete extirpation, seems to have been shared by a number of the more recent investigators, some, Kochs (97) and Wendelstadt (01 and ’04) expressly confirming it, others, Towle (01), Morgan (’03), Reed (’03) and Glaeser (710), while not putting it to the test, seem to have taken care in their experiments to amputate through a bone. Kurz (’12) in the course of his experiments on transplantation of entire limbs in Triton, found that if the limb is completely extirpated (exarticulated) at the hip- or shoulder-joint, a new limb regenerates. Presumably no wounding of the bones of the hip- or shoulder-girdle took place although Kurz does not state 1 The works of Phillipeaux and Fraisse were not accessible to the writer. Their conclusions were obtained from Barfurth’s review in Merkel and Bonnet’s Ergebnisse, vol. 1, 1891. 107 108 C. V. MORRILL what precautions were taken to avoid this. The writer using the American salamander, Diemyctylus viridescens, obtained similar results some years previous to Kurz’s report but for various reasons they were not published.? Recently a new series of experiments were made to work out the histological details of the process and to determine how it differs, if.at all, from regeneration after Injury to remaining bones or cartilages. In addition, a number of more complicated operations were made to analyze further, if possible, the extent and power of regenera- tion after losses not usually met with in nature. MATERIAL AND METHODS A large supply of adult Diemyctylus was obtained through the kindly assistance of Prof. A. Treadwell, of Vassar College. Since many of these animals were in a weak, semi-starved condition when brought to the laboratory, they were kept for a month in glass aquaria before using and were fed on fresh liver. Under these conditions the animals became very vigorous, and with- stood the operations well. All operations were done under narcosis. At first ether was used, but this, owing to its irritating effect on the skin and to a certain percentage of mortality which followed its use, was soon discarded. Much better results were obtained by using a solution of chloretone, of 1: 2000, in which the animals were immersed. This acts very gently. After swimming around rapidly for a few minutes, the salamanders slowly come to rest and in about ten minutes are completely narcotized. The animals recover readily, though sometimes slowly after this treatment. There is no irritation of the skin and no mortality. After the amputations, to be described in detail beyond, the best results were obtained by closing the wounds with a stitch or two of fine silk thread. Although this is not absolutely necessary to the success of the experiment, healing then takes place more rapidly and there is less danger of fungoid growths. Immediately after operation, the animals were placed in a dark 2 The experiment was made at the suggestion of Prof. T. H. Morgan, in 1907. REGENERATION AFTER EXARTICULATION 109 chamber lined with moist filter paper for two days as recom- mended by Reed (’03). During this period the operated ex- tremity was moistened from time to time with a solution of permanganate of potash 1:1000. The animals were then returned to the aquaria. The above precautions almost entirely prevented the growth of fungus and consequent failure of the experiments. For microscopic study, the regenerating regions were removed and fixed for the most part in sublimate acetic or Gilson’s mer- curo-nitric fluid. Other fixatives, such as Zenker’s fluid, Bouin’s fluid and ten per cent formalin were occasionally used but on the whole the sublimate mixtures proved the most satisfactory. After hardening in alcohol for a few days, the objects were decalcified in a mixture of four per cent nitric acid in seventy per cent alcohol for three or four days. They were then imbedded in paraffin and sectioned. As a rule, good series were obtained, seven or eight micra thick, although the rather tough bone and cartilage from large specimens sometimes gave trouble. For staining Mayer’s haemalum followed by picro-acid fuchsin was most frequently employed. This gives a brilliant differentiation of tissues but is not always permanent. Other stains such as Mallory’s connective tissue stain, borax carmine and Lyons blue, haemalum and congo red were also used but none proved as satisfactory for most purposes as the haemalum and _picro- acid fuchsin combination. EXPERIMENTS Pari tf The fact that regeneration does occur after complete extirpa- tion (exarticulation) has been established by the observations of Kurz and the writer as stated above. In order to work out the detail of this process, two sets of operations were made, the hind limbs being used in both cases. In the first set the limb was amputated at the hip-joint, in the second at the knee-joint. . Great care was exercised in making these amputations. The skin and muscles were first carefully divided with a small sharp 110 Cc. V. MORRILL scalpel. Then the part to be removed was grasped with the forceps and slight traction employed to draw the joint surfaces apart. The capsular ligaments were then divided with the scalpel and the limb removed, care being taken not to touch the skeletal parts remaining (hip bones or femur according to the site of operation). A flap of skin and muscle was drawn over the wound and a couple of stitches taken. There was very little trouble from bleeding, but in cases where it was profuse, the specimen was discarded. 1. Amputation at the hip-joint. Nineteen animals were used for this operation divided into groups of eleven, six and two. All of the first group were killed and examined between thirty- nine and forty-six days after operation. Externally each showed a small bud at the site of amputation. Microscopically the bud was composed of a dense mass of indifferent cells with small round nuclei. No change in the hip-girdle was observed. The second group of six were kept for six months. At this time all had regenerated a new limb about three-fourths the normal size. Microscopic examination showed the normal number of skeletal elements in the limb, each represented by a bar of cartilage. There was a well developed narrow cavity in the femur and peripheral ossification had begun in all the cartilages except the tarsals which do not ossify in these animals.*? Joint cavities were well marked at this time. The third group of two animals _was lost. Owing to the small number of specimens and the lack of intermediate stages, the successive steps in this type regeneration could not be made out. The detailed account of this process will therefore be based upon the larger and more complete series of operations at the knee-joint (vid. infra). 2. Amputation at the knee-joint. About seventy-five opera- tions of this kind were made. Most of the specimens obtained were fixed at intervals, of from ten to fifty days and sectioned. The remainder were allowed to complete their regeneration to 3’ While it is true that peripheral ossification does not occur in the tarsalia, nevertheless an extensive marrow cavity is normally present and the irregular trabeculae of cartilage bordering it generally become calcified, if not actually bony. REGENERATION AFTER EXARTICULATION MOE determine whether the new part exactly resembles the old both in size and gross structure. In addition fifteen amputations were made through the distal end of the femur for comparison with the exarticulation experiments. Descriptive. Wendelstadt (’04) and Glaeser (10) have given very detailed accounts of regeneration in the limbs based chiefly on species of the European salamander Triton and on the Axolotl. In Diemyctylus the process is quite similar to that observed in Triton. A study of the specimens in which amputation was made through the distal end of the femur, ie., an operation corresponding to those of Wendelstadt and Glaeser, showed that the descriptions given by these writers for Triton, apply almost equally as well to Diemyctylus. It is true there is some slight discrepancy in their accounts but this can be discussed more conveniently when comparing the exarticulation experiments with those previously made. The earlier changes which take place in the stump may be passed over briefly here. They are concerned chiefly with the | over-growth of the integument, the breaking down of the soft parts, notably the muscle and the formation of a dense mass or bud of small cells with round, deeply staining nuclei over the distal end of the bone (femur). This mass lengthens out and forms a projection when seen externally but does not always lie in the axis of the limb. The origin of these cells could not be determined with exactness. Wendelstadt (704) encountered the same difficulty. Towle (’01) however, states that the accummulation of ‘nuclei’ in the bud is due to rapid (direct) division of nuclei in the old muscle fibers and the disintegration of these fibers. Undoubtedly the degeneration of muscle- fibers is largely responsible for the accumulation, but whether exclusively so or not, is difficult to decide. Connective tissue elements may also contribute something. Turning now-to the changes in the bone and cartilage with which this paper is chiefly concerned, it is here that an essential difference appears between regeneration after exarticulation and 4 Towle’s experiments are concerned almost entirely with the regeneration of muscle. 142 Cc. V. MORRILL regeneration after amputation through a bone. This difference has to do principally with the behavior of the distal epiphyseal cartilage of the femur. This epiphysis (fig. 1, Hp.f.) becomes slowly detached from the shaft by resorption of the bone and calcified cartilage (C.c.) proximal to it. The resorption seems to be brought about largely by the action of cells which arrange themselves along the surface of the bone or calcified cartilage and erode it. Many of them are giant cells. Their origin was not determined. Both Wendelstadt (04) and Glaeser (10) have described this resorption process by giant cells. Figure 1 shows the distal end of the shaft (£.f.) broken up into irregular fragments of bone and calcified cartilage. Coincident with the resorption process, a change takes place in the epiphysis itself. The cartilage matrix begins to break down. This occurs first on its distal and proximal surfaces (fig. 1, Hp.f.). There is no evidence, however, that the cartilage cells themselves undergo degeneration. Indeed, in many in- stances they have been seen dividing mitotically and further, as the lacunae are opened by the degeneration of the matrix, the cells pass out and mingle with the surrounding tissues. Those liberated on the distal surface could not be further traced but on the proximal surface, that is facing the marrow cavity (figs. 1 and 2, M.C.), they contribute to a mass of tissue (Az.C.) which is forming between the old epiphysis and the shaft. This mass which quickly takes on the appearance of young cartilage may be ealled axial cartilage adopting Glaeser’s (’10) term. This axial cartilage appears to have a two-fold origin: (a) from the liberated cells of the old epiphysis as stated above and (b) from the cells lining the marrow cavity and covering the trabeculae of bone and calcified cartilage at the distal end of the shaft; in other words from the endosteum. These latter cells appear to increase in number and, streaming out from the interior of the femur (fig. 1) become massed under the old epiphysis where they form the main contribution to the new axial cartilage. 5 It may be stated that there is normally a considerable amount of calcified cartilage at the plane of union of shaft and epiphysis. This is distinct from the uncalcified, hyaline cartilage of the epiphysis itself. REGENERATION AFTER EXARTICULATION i13 This substantiates Wendelstadt’s conclusion that ‘osteoblasts’ (in reality chondroblasts) arising from the lining of the marrow cavity contribute to the new formation (of cartilage). Glaeser, however, derives the axial cartilage from periosteal and connec- tive tissue fibrillae exclusively. He thinks that cartilage may oc- casionally arise from bone-marrow though this is doubtful. The origin of a part of this cartilage from cells of the old (distal epiphyseal) cartilage naturally could not occur in the experiments of Wendelstadt and Glaeser since the distal epiphysis was always removed in their operations. Glaeser nevertheless states that regeneration from old cartilage, either from the cells or the matrix does not occur. The writer is unable to determine from Glaeser’s account upon what experiment this conclusion is based. That giant-cells in certain regions break down the matrix and open up the capsules is true but in Diemyctylus, at least, the cartilage cells certainly do not degenerate. Evidences of further activity are abundant in this material including the frequent presence of mitosis. During the early stages in the formation of axial cartilage, the cells of the periosteum also begin to form cartilage. The site of this new formation is at first some distance from the distal end of the femur and entirely distinct from that of the axial cartilage. It may be called periosteal cartilage, or using Glaeser’s term, peripheral cartilage. An early stage of its growth is shown in figure 1, Per.C; a later one in figure 2, Per.C. The formation of peripheral (periosteal) cartilage is mentioned by most workers on regeneration and there is no need to discuss it at length. In all cases it first appears at some distance proxi- mal to the wound and gradually spreads distalward until it forms a continuous collar around the shaft reaching to the distal end (fig. 3, Per.C.). Cornil and Coudray (’03) described a similar formation in the healing of experimentally produced fractures in a mammal (rabbit). The first phenomena of repair, they state, are to be found at some distance from the fracture, that is the formation of peripheral (periosteal) cartilage. Whether or not the axial cartilage always appears first could not be determined. In some eases the peripheral cartilage was THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 1 114 Cc. V. MORRILL distinguishable earlier while in most the two appeared at about the same time. Glaeser seems to think the axial cartilage ap- pears before the peripheral, while Wendelstadt describes activity in the periosteum leading to the formation of peripheral cartilage before there is any outgrowth of cells from the marrow cavity to form axial cartilage. In the experiments on Diemyctylus the presence of an epiphysis may have had a modifying effect on the order of appearance. This point is probably not of any great importance. Soon after the appearance of axial and peripheral cartilage the tissue of the bud which has formed distal to the femur, begins to show signs of differentiation. This tissue as mentioned previously is at first composed of apparently indifferent cells. Cartilaginous masses now appear in it and, since they are formed in the same manner as in normal development, they may be spoken of collectively as embryonal cartilage following Glaeser (10). In figure 2, a mass of this cartilage (Hm.C.) can be seen lying distal to the old epiphysis (Ep.f.). It is at first entirely distinct from the axial cartilage.’ As the limb bud enlarges rapidly the growth of the embryonal cartilage keeps pace with it, the cartilage extending right into the growing tip (fig. 3, Hm.C.) In this way the skeleton of the new leg and foot is blocked out in cartilage at a young stage, as the early experimenters found. The formation of new skeletal parts is due chiefly to concentra- tions of cells in situ rather than to growth from the first formed mass or masses. There is, it is true, a continuous substratum or core of tissue running through the center of the bud from which the skeleton arises but it does not give rise to one or two con- tinuous bars of cartilage. There are always interruptions at points where joints are to be formed. During the growth of the embryonal cartilage, the axial cartilage continues to enlarge. This is due in part to growth at the expense of the old epiphyseal cartilage. The latter remains, however, for some time partly imbedded in the axial cartilage (fig. 3, Hp.f) where it can be distinguished by its different ° The space seen in figure 2, between the embryonal cartilage and the old epiphysis is probably an artifact. REGENERATION AFTER EXARTICULATION 115 staining reaction. The expansion of the axial cartilage finally brings it into contact with the embryonal cartilage and they become so closely united as to appear almost continuous (fig. 3). The peripheral cartilage (fig. 3, Per.C’.) meanwhile has spread to the distal end of the bone (B.f.) and there unites with the axial cartilage (Az.C). This results in a continuous cap of cartilage covering the distal end of the femur, and extending proximally for some distance. Figure 4, E p.f.n. from a much later stage than figure 3, shows this new cap. From it, the new epiphysis of the femur is formed and also the new bone of the intermediate zone between epiphysis and shaft which was lost in the early resorption process. A portion of the marrow cavity M.c.n. can be seen extending into it.? At Os.n., new bony tissue is spreading into the cartilage. This new ossification is shown under high magnification in figure 10, which is taken from the same region of a neighboring section. The new bone (Os.n.) appears as a mass of interlacing fibers extending through the matrix of the cartilage (C.n.). In this process, the cartilage cells are gradually enclosed by the newly formed bone and eon- verted directly into bone cells. Three such cells are shown at Os.c. Portions of the marrow cavity (M.C.) and bone of the shaft (B/f.) are included in the figure. Between the shaft and the marrow cavity some of the old calcified cartilage remains (compare fig. 3, C.c.). A direct transformation of cartilage into bone was noted by Cornil and Coudray (03) in the healing of fractures in the rabbit. They state that in most cases the car- tilage bordering the bone first ossifies along its edges while later the cartilage capsules themselves ossify, the cells being directly transformed into bone cells. It will be recalled that a portion of the distal end of the shaft is destroyed in the resorption process by which the old epiphysis is detached from the bone. This is replaced by ossification of a part of the new cartilage in the manner just described. The further history of this process was not followed owing to lack of material in the advanced stages. ’ Through an error, the leader from the letters M.c.n. does not point to this extension. 116 Cc. V. MORRILL Distally the new formed epiphysis comes into relation with the embryonal cartilage which forms the new fibula and tibia (figs. 4 and 9) and later a joint appears at this level. As pre- viously stated, before the new epiphysis is completed and while a portion of the old epiphysis is still present (fig. 3) a close attachment is developed between the axial cartilage (Az.C.) and the embryonal cartilage (Hm.C.). It is difficult to say whether or not the two cartilages become actually continuous. The tissue uniting them is composed of cells without definite bounda- ries and with small elongated nuclei. It does not appear to be cartilaginous. Small cavities soon appear in this connecting tissue, beginning usually at the circumference and spreading toward the interior. An early stage in the joint formation is shown in figure 9. The new joint cavity (J.c.n.) is Just appearing between the new femoral epiphysis (Hp.f.n.) and the new fibula (Fib.n.). In a later stage (fig. 4) the joint cavity has enlarged at the circumference and a definite capsule has been formed, the latter being continuous with the perichondrium of the new cartilages (epiphyses). A small joint cavity has also developed between the fibula and a tarsal cartilage (7.c.n.). At this stage the fibula itself has begun to show evidences of subdivision into shaft and two epiphysis. In the central portion, the car- tilage areolae are enlarging and becoming more spherical pre- paratory to the formation of a marrow cavity, while at the sur- face a thin layer of subperiosteal bone has been laid down. In the epiphyseal portions, the cartilage cells have a tendency to arrange themselves in concentric ares, a characteristic of epi- physes in general. No attempt has been made in this study to arrange the speci- mens in a series based on time after operation. The stages of regeneration .do not necessarily correspond to the intervals of time. For example figure 3 shows a stage obviously more ad- vanced than that shown in figure 2, vet it was taken from a specimen killed thirty-eight days after operation, while the latter came from one killed at forty days. This is very probably due partly to difference in time of healing of the wound and partly to temperature. It was noticed that among animals operated REGENERATION AFTER EXARTICULATION 7, on at the same time and as far as possible treated in the same manner, some always healed more readily than others. Re- garding temperature—the higher degrees were in general more favorable to rapid regeneration than the lower as might be ex- pected, but there was so much individual variation due to time of healing that no great reliance can be placed on this statement. Comparing, from the standpoint of time, regeneration after exarticulation with regeneration after wounding a bone, one finds that the former is appreciably slower. On the average, there is about ten days difference in corresponding stages. With regard to regeneration of soft parts (muscle, nerve, blood- vessels, etcetera), there is nothing to add to the accounts already published. Muscle regeneration has been very carefully worked out by Towle (’01) in Plethodon and Schminke (07) in Triton taeniatus and T. cristatus. These writers agree that the new muscle is formed chiefly by isolated cells (sarcoplasts) which arise from degeneration of the old muscle of the stump. The sarcoplasts are small masses of cytoplasm which contain at first several nuclei. They usually break up into small cells which are responsible according to Towle (’01) for the accumulation of nuclei (cells) in the growing bud distal to the stump of the old bone where they give rise to new muscle fibers. It was previously mentioned that the exact origin of this entire mass of small cells is somewhat uncertain (cf. Wendelstadt). Probably most of the cells originate from degenerating muscle and some perhaps from other soft tissues. In any event, there is no evi- dence that any are derived from bone or cartilage. Histologically there is at first no sign of differentiation in these cells, and it seems useless to assume that such exists. It is from a part of them, however, that new embryonal cartilage is developed in the midst of the bud. This seems to be an example of dediffer- entiation followed by redifferentiation in the sense of Child (’15). From the apparently indifferent mass both cartilage and muscle are formed, the cartilage showing the typical embryonic type of development. A somewhat similar process is to be seen in the behavior of the old epiphyseal cartilage of the femur. Here the matrix breaks 118 C. V. MORRILL down, liberating the cells which again become active and form new matrix. In this case, however, the cells do not dedifferenti- ate so far as to become indifferent; they remain cartilage forming cells. The formation of the peripheral cartilage and that part of the axial cartilage derived from endosteal cells is of course quite different in nature. Here the more or less undifferentiated cells of the periosteum and endosteum which have lain gormant, are suddenly stimulated to activity by the amputation. They form cartilage first and later a portion of the cartilage is trans- formed into bone as described on page 115. It was stated at the beginning of this account that some animals were allowed to complete their regeneration to determine whether the new skeleton was like the old. There seems to be no dif- ference whatever provided sufficient time is allowed for devel- opment. Externally, a slight deformity sometimes appears at first, since the new bud does not always le in the longitudinal axis of the limb. This is more common after operations at tbe knee-joint than at the hip-joint. In the course of time this irregularity disappears and the limb becomes normal in shape and position. Complete ossification, however, may take almost a year and sometimes even longer. . A glance over the literature of regeneration in amphibia shows that the power to regenerate a new normal skeleton does not extend to all animals of this class. Morgan (’03) found that in Amphiuma the new skeleton was abnormal and deficient al- though some specimens were kept under observation for nearly a year. Certain results which were obtained by the earlier experimenters Goette and Fraisse seem to indicate that some of the European urodeles (Triton marmoratus and Proteus) lack the power of complete regeneration but Kammerer (06) states that this is not the case if the animals are kept under favorable conditions and for a sufficient length of time. In the Anura the power is much more limited. New limbs will regenerate only if amputation is made in the tadpole stage. Barfurth (’94) was the first to find that the limbs of frog-larvae (Rana fusca) are capable of regeneration, but this power disap- pears in the progress of development. Ridewood (’98) obtained REGENERATION AFTER EXARTICULATION 119 regeneration of posterior limbs in the tadpole of the midwife- toad (Alytes obstetricans). The new skeleton was “normal or nearly so” in five cases. Byrnes (’04) using frog-tadpoles showed that the anterior limbs would regenerate while still under the operculum but the new limb is invariably smaller than normal and there is a tendency to reduction in the skeletal elements. Morgan (’08) (and Goldfarb) attempted to induce regeneration in the fore-leg of the adult frog by artificial means. Pieces of the leg, muscle and other tissues from the tail of the tadpole were grafted into the stump but with only small success. In some cases, however, incomplete regeneration of the leg with rudimentary toes was obtained, or a broad flat ‘foot’ with scant . toes. Histological details were not given. Glaeser (’10) more recently tested the power of regeneration in the hind limbs of adult frogs but found none except in two cases where a ring of peripheral (periosteal) cartilage developed around the stump of the femur. No artificial means were used to induce regeneration in this case. Rar TT To test further the power of regeneration in Diemyctylus a series of more complicated operations were made, involving losses not usually met with under natural conditions. These are to some extent a repetition of. those of Wendelstadt (’01) and Reed (’03) but with certain modifications. Expervment 1. Extirpation of the fibula and a portion of the femoral epiphysis but without injury to the tarsus. Number of animals, ten:—Of these two. escaped and one lost the foot. The remaining seven were killed at intervals of from forty-eight days to one year. There was no indication of regeneration of a new fibula, but the lost portion of the femoral epiphysis was restored. Experiment 2. Eextirpation of the fibula without injury to either femoral epiphysis or tarsus. Number of animals, ten:— These were killed at intervals of from thirty-three days to one year. For the most part no indication of regeneration was observed but in two specimens there was a narrow mass of eal- 120 Cc. V. MORRILL cified fibrous tissue or bone in the place occupied by the old fibula. This mass may have developed from fragments of the old fibular epiphysis which were accidentally left in the wound in two of the operations.® The results of experiments 1 and 2 substantiate the conclusion of Wendelstadt and Reed that regeneration in a lateral direction (in the limb) does not occur. Experiment 3. Extirpation of both leg bones and a portion of the femoral epiphysis but without injury to the tarsus. Number of animals, five-—Considerable shortening occurred but in no case did the foot drop off. One was killed after sixty-three days and the remainder after one year. ‘Two of the latter showed some indication of regeneration. In these there were one or two car- tilaginous nodules, in one case fairly extensive, connected with the femoral epiphysis by fibrous tissue. The femoral epiphysis itself was restored in all. Experiment 4. Extirpation of both leg bones but without injury to either femoral epiphysis or tarsus. Number of animals, seven:— Shortening of course occurred. Two lost the foot subsequently and were discarded. Of the five remaining, one was killed at sixty days and the remainder after one year. All showed def- inite attempts at regeneration, in some cases quite well marked. Figures 5 and 6 from the same specimen cut in the dorsi-ventral plane show the extent of regeneration in the best marked case. The time elapsed was sixty days. Only a small portion of the femoral epiphysis (Hp.f.) and one tarsal cartilage (T.c.) appear in figure 5, since the section lies near the border of the limb. The new skeletal element of the leg (L.s.n.) consists of an epi- physeal portion which articulates with the femur and a long, narrow bar of cartilage which is partly overlaid by bone at either end. (The bone is darkly shaded in the figure.) At the proxi- mal end a distinct joint capsule with cavity (J.c.n.) has developed. Figure 6, from a section near the median plane, shows another portion of the new skeletal element (/.s.n.). This portion is 8 In the experiments described in part II, all bones removed were examined under a binocular microscope to determine whether any part of them had been accidently left in the wound. REGENERATION AFTER EXARTICULATION 121 a solid mass of cartilage united to the femoral epiphysis (/’p.f.) by a capsule with joint cavity (J.c.n.) as before. Distally it is in contact with a tarsal cartilage (7’.c.) but no joint capsule has developed. The two portions of the new element are continu- ous proximally when traced through the series. Unfortunately intermediate stages in the formation of the new skeletal element were not obtained. The epiphysis of the femur (fig. 6, Hp.f.) has the appearance of new cartilage similar to that of the new element (L.s.n.). This seems to indicate that the regeneration was centrifugal in direction and probably occurred in the same manner as described in the first section of this paper. _No changes were observed in the tarsal cartilages. Figure 7 is from another specimen killed after one year and cut in a plane passing through the borders of the limb. The new elements here consist of two large masses of cartilage (L.s.n.) united by fibrous tissue and connected with femoral epiphysis (Ep.f.) by a capsule containing a joint cavity (J.c.n.). Distally the new cartilages fall short of the tarsus. The tarsal cartilages themselves (7.c.) show signs of growth in a proximal direction (centripetal regeneration). They have become united proximally by a mass of cartilage which, however, has no connection with the new skeletal elements. The. arrangement produces what may be called a soft joint. Other specimens in this experiment showed the formation of irregular masses or nodules of cartilage but not so extensively as the two described above. There appears, then, to be a limit to the power of regeneration under the conditions of the experi- ment. This may be due to an inhibiting influence from the presence of the foot and to shortening of the limb which leaves very little room for the new growth. It may be well to state that as soon as the wound heals, the animal uses the limb con- stantly when creeping over the bottom of the aquarium. Wendel- stadt (’01) performed a similar experiment upon the anterior limb of the Axolotl but with entirely negative results though he kept the animals under observation for ten to fifteen months. The limbs shortened as in the case of Diemyctylus but the animals apparently made no attempt to use them. It is just possible 122 Cc. V. MORRILL that a certain amount of activity in the limb is necessary to start the regenerative process. Wendelstadt also tried the effect of leaving a small piece of one of the bones (ulna) in situ. For this operation he used one Axolotl and one Triton. In the latter the humerus was also wounded. The axolotl regenerated a new ulna which was shorter than normal while in the Triton a whole new forearm and a second hand were formed. This peculiar malformation in the Triton was never duplicated in any of the writer’s experiments on Diemyctylus, although in some cases the femur was purposely wounded. It is improbable that there is any difference in the power of regeneration of fore and hind-limbs in these animals. Experiment 5. Eextirpation of the fibula and removal of the foot entire without injury to the femoral epiphysis or tibia. Num- ber of animals, five:—In this lot two were killed at sixty-six and ninety-five days respectively and the remainder at the end of a year. ‘The first of these had regenerated a well-marked foot when killed and a new fibula. The latter consisted of a solid bar of cartilage with a layer of subperiosteal bone surrounding its proximal two-thirds. It was attached to the femoral epiphysis by a capsule, common to it and the tibia. Distally it was connected with the new tarsalia by ligaments, in places showing the beginning of a joint cavity. In the specimen killed at ninety-five days, the foot had regenerated but there was scarcely any indication of a new fibula. The remaining three specimens killed after one year regenerated a new complete foot including tarsalia and a new but incomplete fibula. A section through one of these is shown in figure 8. The old tibia (77b.) articulates with the femoral epiphysis (/p.f.) while the new fibula (F7b.n.) falls short proximally. Peripheral ossification has started in the new element but there is no marrow cavity as yet. The new tarsalia are seen at 7.c.n. The distal epiphysis of the tibia seems to be composed of new cartilage like that of the tarsals and fibula. This is to be expected since it was shown that the old epiphysis in a stump is always replaced by new cartilage (vid. Part I). In the present experiment one would expect first a new formation of cartilage from which a new tibial epi- REGENERATION AFTER EXARTICULATION 123 physis and the skeleton of the new foot is formed. This is followed by growth in a proximal direction to form the new fibula (centripetal regeneration). Apparently the energy of regenera- tion is not always sufficient to produce a complete fibula even in a year’s time although it may do so in two months as in the first specimen described (sixty-six days). A tendency to regenerate centripetally was also noted in experiment four, where both leg bones were removed and the foot allowed to remain. In this case however it was limited to the formation of a mass of cartilage (fig. 7, 7.c.) uniting the proxi- mal surfaces of the tarsalia. Wendelstadt (’01) obtained centri- petal regeneration in the axolotl by extirpating the upper ends of the radius and ulna. In three animals, the bones were com- pletely restored after fifteen to eighteen months. There was apparently no tendency to regenerate centrifugally from the femoral epiphysis. Morgan (’08) using Plethodon and Diemyctylus tried to dis- cover what kind of a structure would regenerate from the proxi- mal end of a limb. For this purpose the limb was cut off and grafted onto the stump in reverse position. Regeneration oc- curred but results were complicated, due to mixing of old and new material and to the turning of the graft in the skin-pocket. This method was discarded in favor of the following: The hind leg was cut off at the knee. Then the femur was cut off high up in the thigh and the distal portion reversed in position. A new limb regenerated. Its skeleton was composed of (1) proximal stump of femur, (2) connecting cartilage, (3) piece of reversed femur, (4) new tibia and fibula, and (5) foot. There was considerable evidence of absorption and in only a few cases did it seem probable that the material for the new limb came from the exposed proximal end of the grafted piece. In some cases the new cartilage from the proximal stump grew past the graft. On the whole it is not quite clear from Morgan’s account what part the graft played in regeneration, as histological de- tails are not given. Reed (’03) performed a series of experiments somewhat similar to those just described (Exp. 5) except that the distal end of the 124 Cc. V. MORRILL tibia was removed with the foot. Spelerpes ruber was used for these experiments. The results were, regeneration of the distal end of the tibia, a new foot and a new fibular element. The latter was usually incomplete but in one case it almost completed itself proximally, that is, reaching the femoral epiphysis. As in the present experiments there was no tendency to regener- ate from the femur. Wendelstadt, in his later paper (04) states that the experiments of Reed confirm his general conclusion that wounding of the skeletal elements is necessary for regenera- tion. The experiments described in the present paper show that this conclusion is too sweeping. It is true that if one bone only (fibula or radius) is removed (Wendelstadt, Reed and the writer) or if the proximal parts of two bones are removed (Wendelstadt) no regeneration occurs from the uninjured epiphysis of the femur (or humerus). In the first case, the pressure of the remaining bone against the joint surface of the femur (or humerus) and the tarsals (or carpals), that is the presence of a functional joint may inhibit regeneration from these points. In the second case the new growth centripetally from the remaining injured bones, which is always more rapid than from uninjured ones, may make up the deficiency in time to check any tendency to regenerate from the epiphysis of the humerus.’ Shortening of the limb which must occur in this ease would also be a factor. These, of course, are mainly suggestions. Further experiments are neces- sary before definite explanations can be made. SUMMARY AND CONCLUSIONS 1. In Diemyctylus regeneration takes place readily after com- plete extirpation (exarticulation) whether the operation is made at the hip- or knee-joint (Part I), or at the ankle-joint (Part I, Exp. 5). The time elapsed is somewhat longer than when a skeletal element is injured. 2. The new skeletal elements are similar to the old. There is no tendency to reduction. 3. The essential difference between regeneration after exarti- culation and regeneration after wounding a skeletal element lies in the behavior of the cartilaginous epiphysis which is present REGENERATION AFTER EXARTICULATION 125 in the stump in the former case. This cartilage becomes de- tached from the shaft, gradually breaks down and is, partly at least, reconverted into cartilage which assists in the formation of a new epiphysis. 4. The new cartilage which forms the basis for the skeletal elements appears independently in three localities: a) Around the shaft of the bone proximal to the epiphysis (peripheral cartilage). This cartilage is periosteal in origin. b) In the axis of the bone and in contact with the marrow subsequent to detachment of the epiphysis (axial cartilage). The origin in this case is twofold: (1) From the cells of the old epiphyseal cartilage and (2) from the lining of the marrow cavity (endosteum). c) In the tissue of the bud distal to the epiphysis (embryonal cartilage). Here dedifferentiation appears to have taken place forming a substratum of indifferent cells from which in turn new cartilage is formed as in early development of the limb. 5. If a single bone (fibula) is removed completely from the leg, it is not replaced either by proliferation from its fellow (lateral regeneration) or from the skeletal elements lying proximal and distal to it even when one of the latter is injured. 6. When both leg bones are completely removed they are replaced to some extent by new elements which, however, are always irregular and incomplete. The origin of the new parts was not definitely determined. 7. When one leg bone (fibula) and the foot are removed with- out injuring any of the remaining skeletal elements, a new complete foot is regenerated from the distal end of the remaining leg bone (tibia). This is followed by a slow and often incomplete regeneration of the lost leg bone (fibula) in a proximal direction (centripetal regeneration). 126 ' ©. V. MORRILL LITERATURE CITED Barrurtu, D. 1894 Sind die Extremititen der Frésche regenerationsfihig? Arch. f. Entw.-Mech., Bd. 1. Byrnes, E. 1904 Reconenanion of the anterior limbs in the tadpoles of frogs. Arch. f. Entw.-Mech., Bd. 18. 1904 On the skeleton of regenerated anterior limbs in the frog. Biol. Bull., vol. 7. Cuitp, C. M. 1915 Senescence and rejuvenescence. Univ. of Chicago Press. Corniu. V. et Coupray, P. 1903 (Note) De la formation du cal. Comptes rendus de l’Academie des Sciences, Tome 137, p. 220. GLAESER, Kart 1910 Untersuchungen iiber die Herkunft des Knorpels an regenerierenden Amphibienextremititen. Arch. f. mikr. Anat., Bd. 75. KAMMERER, P. 1906 Die angeblichen Ausnahmen von Regenerationsfaihigkeit bei den Amphibien. Zentralbl. f. Phys., Bd. 19. Kocus, W. 1897 Versuche iiber die Regeneration von Organen bei Amphibien. Arch. f. mikr. Anat., Bd. 49. Kurz, Oskar 1912 Die beinbildenden Potenzen entwickelter Tritonen. Arch. f. Entw.-Mech., Bd. 34. Moraan, T.H. 1903 Regeneration of the leg of Amphiumameans. Biol. Bull., viol 1908 Experiments in grafting. Amer. Nat., vol. 42, No. 493. Reep, M. 1903 Regeneration of a whole foot from the cut end of a leg contain- ing only the tibia. Arch. f. Entw.-Mech., Bd. 17. Ripewoop, W. G. 1898 On the skeleton of regenerated limbs of the midwife- toad (Alytes obstetricans). Proc. of Zoél. Soc. of London. ScumincKke, A. 1907 Die Regeneration der quergestreiften Muskelfasern bei den Wirbeltieren. Verhandl. d. phys.-med. Gesell. zu Wiirzburg. Bd. 39. Towts, E. 1901 On muscle regeneration in the limbs of Plethodon. Biol. Bull., vol. 2 Wenvetstapt, H. 1901 Uber Knochenregeneration. Arch. f. mikr. Anat., Bd. 57. 1904 Experimentelle Studie iiber Regenerationsvorginge am Knochen und Knorpel. Ibid., Bd. 63. PLATES” ~ PLATE 1 EXPLANATION OF FIGURES Figures 1 to 4 are from specimens in which complete amputation of the limb was made at the knee-joint. 1 Longitudinal section of a limb 30 days after operation. Ep.f., femoral epiphysis; Ax.C., axial cartilage; Per.C., peripheral cartilage; C.c.,calcified cartilage of the shaft; B.f., bone of the shaft (femur); M@.c., marrow cavity. The narrow space distal to the epiphysis is an artifact. Magnified about 30 diameters. 2 Longitudinal section of a limb 40 days after operation. Em.C., embryonal cartilage. Other abbreviations as in figure 1. The space between the femoral epiphysis, Ep.f., and the embryonal cartilage Em.C., is probably an artifact. Magnified about 30 diameters. 3 Longitudinal section of a limb 38 days after operation. Abbreviations as in figures 1 and 2. Magnified about 30 diameters. 4 Longitudinal section of a limb 48 days after operation. T.c.n., new tarsal cartilage; Fib.n., new fibula; J.c.n., new joint-cavity; Ep.f.n., new femoral epi- physis; Os.n., bone formation in new cartilage; W.c.n., extension of the marrow cavity into the new cap of cartilage (see foot-note on p. 115); B.f., bone of the femur. Magnified about 30 diameters. PLATE 1 REGENERATION AFTER EXARTICULATION Vv. MORRILL Cc. H. Murayama del. 129 1 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. PLATE 2 EXPLANATION OF FIGURES 5 and 6 Longitudinal sections of a limb from which both tibia and fibula were completely removed. ‘Time after operation, 60 days. T.c., tarsal cartilage showing normal marrow cavity and partly calcified lining; L.s.n., new skeletal element of the leg, in figure 5, partly ossified; Hp.f., epiphysis of the femur; J.c.n., new joint cavity; W.c., marrow cavity and B.f., bone of the femur. Magni- fied about 30 diameters. 7 Longitudinal section of another specimen after same operation as above. Time one year. ‘Tarsal cartilages blended proximally, 7’.c.; new skeletal elements L.s.n. Other abbreviations as in figures 5 and 6. Magnified about 30 diameters. 8 Longitudinal section of a limb from which the foot and the fibula were completely removed. Time after operation, one year. 7'.c.n., new tarsal ear- tilages; Fib.n., new fibula (incomplete); 7%b., tibia; Hp.f., femoral epiphysis. Magnified about 30 diameters. 130 REGENERATION AFTER EXARTICULATION PLATE 2 Cc. V. MORRILL H. Murayama del. 131 PLATE 3 EXPLANATION OF FIGURES Figures 9 and 10 are from specimens in which complete amputation of the limb was made at the knee-joint (exarticulation). 9 From a longitudinal section of a limb 41 days after operation. Ep.f.n., new femoral epiphysis; Fib.n., new fibula, proximal end; J.c.n., new joint-cavity forming. Magnified about 100 diameters. 10 From a longitudinal section of a limb 48 days after operation. C.n., new cartilage; Os.n., new bone spreading through the cartilage; Os.c., new car- tilage cells transforming into bone-cells; B.f., old bone of the shaft (femur) ; M.c., portions of the marrow-cavity. Magnified about 420 diameters. 132 REGENERATION AFTER EXARTICULATION PLATE 3 C. V- MORRILL H. Murayama del. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JANUARY 19 IS THE INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT, METAMORPHOSIS AND GROWTH DUE LO A. SPECIBIC- ACTION, OF THAT GLAND? EDUARD UHLENHUTH The Rockefeller Institute of Medical Research, New York City The experiments on thymus feeding thus far reported in the literature have given results sufficiently different to prevent the formation of a definite idea as to the réle of the organ in these experiments. This is even true if one has in mind only one of the various groups of animals which have been studied in such experiments. Concerning the Amphibia among which only the larvae of Anura have been studied carefully regarding their reaction to thymus feeding, it seems that most of the experiments showed a retarding influence of the thymus upon development and metamorphosis although some exceptions are reported. With respect to growth however, the results are so lacking in uniformity, that Gudernatsch as well as Romeis who studied the effect of thymus feeding in tadpoles doubted whether the effect produced by this organ was due to a specific action or only to quantitative conditions. Gudernatsch in his experiments on tadpoles noted accelerated growth leading to enormous size; but Romeis obtained completely normal growth in various series of thymus-fed tadpoles and pointed out that the thymus feed- ing never produces abnormally large animals. The following experiments which will be reported elsewhere in detail, seem to indicate that the accelerated growth of thy- mus fed Amphibian larvae is merely the effect of quantitative conditions and not the result of a specific quality of the organ such as a specific growth-stimulating agent. Furthermore, they yielded some very interesting results concerning development and metamorphosis although these are still difficult to explain. Finally they showed that in each thymus-fed larva severe tetany 135 136 EDUARD UHLENHUTH is produced. The last mentioned phenomenon will be discussed in another article; the effects of thymus upon development, metamorphosis and growth will be outlined briefly in the follow- ing pages. In the experiments to be reported, only larvae of Urodela were studied (Amblystoma punctatum and A. opacum). The ad- vantage of using Salamander larvae is that the quantity of food given to them can be controlled and measured with exactness, up to a certain degree, which cannot be done if tadpoles are used; and it should be emphasized that in order to avoid errors the possibility of measuring the quantity of food is very desirable and even should be demanded in experiments in which it is suspected that qualitative relations are involved. 1. DEVELOPMENT AND METAMORPHOSIS Gudernatsch found that the development of tadpoles was delayed if the animals were fed on thymus. Similar results were obtained by Romeis in his first experiments. This led to the inference that thymus contains a substance whose specific property is a retarding effect upon development. Only recently Gudernatsch again published a paper based on this hypothesis. However, in his work published in 1915 relative to the influence of the. glands.of internal secretion on Anura larvae, Romeis reports upon two series of experiments, which cannot be ex- plained from the above-mentioned standpoint, and which caused the author himself to doubt whether inhibition of development were indeed a specific function of the thymus. ‘The first series consisted of fairly old larvae of the species Rana esculenta, the individuals of which developed in a perfectly normal manner, in spite of being fed with thymus; but in this case it might have been supposed that the thymus feeding had been started too late. In the second series, however, in which larvae of Rana temporaria were employed, the thymus feeding was commenced at a very early stage, in spite of which fact the development of the larvae was not delayed. On the contrary, they underwent metamorphosis at an earlier stage than did the larvae fed with INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 137 muscle tissue, and before the latter had developed front limbs. This last series of experiments suggests the assumption that the influence exerted by the thymus on development must be de- pendent on factors not specific for the thymus. The experi- ments on Salamander larvae now to be described led to similar conclusions. Both Gudernatsch and Romeis took as index of the rate of development, the growth of the hind and front limbs, the absorption of the tail and the abandonment of the water. The latter phenomenon however, which we will refer to as meta- morphosis seems in Salamanders, to be dependent on a mechan- ism different in many respects from that which controls develop- ment, such as the growth of limbs, etc.; for in the first place even under conditions of normal feeding, different individuals show a different stage of development when they leave the water, and secondly the effect of thymus upon development and upon metamorphosis does not seem to be the same in the Salaman- der larvae examined. Therefore we shall distinguish between development and metamorphosis; growth and differentiation of the limbs, certain changes of the fin and gills not being in direct relation to the abandonment of the water, and the changes of the color pattern of the skin which finally lead to the definite coloration of the skin, will be referred to as development; while the abandonment of the water together with the sudden reduc- tion of the gills to mere stumps and the complete absorption of the fin will be called metamorphosis. In a group of eight series, O 1916, in which larvae of Ambly- stoma opacum were used, four series were fed with small frag- ments of earthworms and four series with equal sized pieces of thymus. As will be explained later on, these experiments were conducted with the intention of feeding the respective animals with equal quantities of worms and thymus. So far as develop- ment and metamorphosis is concerned, it would seem at least possible, that besides the quality of food, the amount of food may also have some influence upon these two phenomena; but at any rate only if we make the quantities of food alike in the experi- mental series and the controls, can we be sure that the differences 138 EDUARD UHLENHUTH obtained in both series are the expression of the quality of the food. ; The development of the legs and toes was carefully noted, and the development of these organs was seen to occupy a period of from 6 to 9 weeks. As the thymus feeding began as early as the second week, the development of the legs took place under the influence of thymus feeding during 4 to 7 weeks. According to Gudernatsch and Romeis this length of time sufficed in the case of tadpoles to produce the retarding effect upon development of the thymus; but in the case of the Salamander larvae absolutely no retardation could be noted as a consequence of the thymus feeding. Indeed, in those series which as a result of the simul- taneous effect of a lowered temperature necessitated a longer period of time in order to attain complete development, a change in the contrary direction could even be noted in the latest stages, the thymus animals attaining complete development of their limbs more quickly than the worm-fed animals. Thus it can hardly be argued in this connection that feeding had not lasted long enough to produce a result, in view of the fact that in the latest stages of development of the legs when the feeding had lasted a- longer period than in the first stage, the opposite result was obtained. This result then indicates that a distinct difference exists between the Anura and Urodela with respect to the effect of thymus-feeding upon development. The development of the legs in such animals to which instead of equal quantities of food as much food was given as each animal was able to take, was not studied in sufficient detail. In one group of experiments (P 1719) which consisted of A. punctatum larvae, development of the legs was recorded during 14 days after the beginning of the feeding; in this group the relation be- tween the thymus-fed animals and the controls was the same as in the above experiments on A. opacum. The differences between the Anura and Urodela become even more accentuated as development proceeds. But a careful distinction must be made between animals fed on equal quantities and those which obtain as much food as they will take. INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 139 In the above mentioned group (O. 1916) which consisted of larvae of Amblystoma opacum fed with equal quantities, the development of further advanced characteristics of the legs, of the shape of the head, of the gills and of the color of the skin proceeded much more rapidly in the thymus-fed animals than in those fed on worms. With respect to the development of the gills, the following should be remarked: In larvae of Amblystoma opacum fed with worms and kept at an average temperature of 22.6°C. as late as one or more days before metamorphosis the gills attain a stage not only of considerable size, but one in which they are char- acterized by considerable redness and above all by the fact that they are bent upwards in a crescent shape. The long well developed branches are widely extended and the points of the stem inclined forwards so as to bend over. In worm-fed animals kept at a high temperature (22.6°C.) this condition of the gills was only attained in the 28rd week, but in thymus-fed animals kept at the same temperature as early as the 11th week; in worm-fed animals kept at a low temperature (in average 14.8°C. ) only in the 29th week; in thymus-fed larvae kept at the same temperature as early as the 11th week (although in the latter case the gills were less developed than in the case of the high temperature thymus-fed animals). A similar relationship is observed with respect to the color of the skin. In the case of the warm worm-fed animals the melano- phore spots only began to develop in the 13th week, at which time they had already attained very considerable development in the case of the warm thymus-fed animals; in the warm worm- fed individuals the blue-grey pigment did not appear until the 24th week, but in that of the warm thymus-fed animals as early as the 12th week. In the cold worm-fed animals the fusion of the melanophore spots into a uniform black-brown coat only began in the 30th week, and occurred as early as the 13th week in the case of the cold thymus animals; but in the cold worm-fed animals no trace of a silver-grey pigment can be detected after 32 weeks, although this appeared in the cold thymus-fed in- dividuals as early as the 13th week. These differences in the 140 EDUARD UHLENHUTH rate of development are doubtless sufficiently great to indicate distinctly the differences existing between the Anura and Urodela. Under conditions of quantitatively equal feeding (which alone can be taken into consideration in a study of qualitative effects) the feeding of thymus to larvae of Amblystoma opacum causes accelerated development. Nevertheless, the above mentioned experiments become even more clear if the results obtained by them are compared with the result in experiments made by a different method; for the fac- tor to be emphasized is not the time elapsed since the hatching but the size of the animals. The thymus-fed individuals attain the stated conditions of development while much smaller in size than the worm-fed animals. The latter must attain much greater size than the thymus-fed animals, in order to acquire the same degree of development. In a group of A. punctatum (P. 1916) kept at a high tempera- ture, one series was fed with pieces of thymus and another with Tubifex. In both series the animals were allowed to eat accord- ing to their inclination as a result of which the worm-fed animals consumed a considerably larger quantity of food than did the thymus-fed animals, and consequently grew much more rapidly. The development of the skin pigmentation also proceeded more quickly than in the case of the thymus animals, although the latter attained these various stages while much smaller in size. At the time no exact drawings were made to show the relation- ship between the size and stage of development. This will be taken up in a recently initiated experimental series of A. puncta- tum (P. 1917) as yet incomplete, in which the same system of feeding is being maintained as in Group P. 1916. Meanwhile it can already be noted that the worm-fed animals do not de- velop the yellow network until they have attained the average size of 62.91 mm., the minimum length being 59 mm. The thymus animals, which have only attained an average length of 32.22 mm., with a maximum length of 36 mm., have not yet shown signs of this network. In group P. 1916 of A. punctatum in which the worm-fed animals behaved like the worm-fed animals in Group P 1917 regarding the relation between size and de- INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 141 velopment of network, the first thymus animal attained the network stage when only 41.3 mm. in length. We thus see that the time at which the various phases of development are attained varies according to the quantity of food, with the result that sometimes the thymus animals, at other times the worm-fed animals appear to lead. But the constant factor is the size at which the various stages are at- tained; that is, constant to the extent that the thymus animals always develop more quickly than do the worm animals, if the various stages are referred to the size of the animals. This relationship is directly opposed to that of the Anura larvae, for in these animals the thymus-fed individuals must usually attain a considerably greater size than the worm-fed animals in order to arrive at the same degree of development. Identical relationships as occur in the development are also found in the metamorphosis; but in this case one or more addi- tional factors seem to play a réle to complicate considerably the phenomena, as we shall see. Here again we must differentiate between the experiments in which the food was quantitatively equal and those in which each animal was allowed to eat to the point of satiety. But it should be emphasized that only the first method permits of a correct comparison. For when the worm animals feed at will they eat approximately 10 to 20 times the quantity of food that is con- sumed by the thymus animals when the latter begin to suffer from tetany; as the worm animals also grow much more rapidly as a result, it would not be surprising that they also metamorphose earlier, since we might expect that if a definite size of the animal is indispensable to metamorphosis, metamorphosis will be ac- celerated if we accelerate growth by some external conditions. We will now turn our attention again to the group O 1916 of A. opacum in which each series was given approximately the same quantity of food. In this group the warm thymus animals were the first to undergo metamorphosis; thus in the warm thy- mus series (22.6°C.) the first animal underwent metamorphosis in the 13th week, in the warm worm series only in the 24th week; in the cold thymus series (14.8°C.) the first animal left the water 142 EDUARD UHLENHUTH in the 24th week; whereas in the cold worm series no animal has yet undergone metamorphosis (in the 32nd week). Thus, thymus-fed animals are seen to metamorphose earlier than worm- fed animals; that is, provided they receive equal quantities of food. The relationship of time however becomes inverted if the worm and thymus animals, instead of receiving equal quantities of food are allowed to eat at will. In a group of A. punctatum (P 1916) consisting of two series, which had been kept at a high temperature and in which the last-mentioned mode of feeding was adopted, the first animal of the thymus series underwent metamorphosis after five months, whereas the first of the worm series did so after only 35 month As in development, so also in metamorphosis, the relationship of time is seen to be inconstant and depends on the amount of food given to the animals. But a constant factor exists in the relationship between size of the animal and metamorphosis. Whatever method of feeding may be adopted, the thymus-fed individuals are always much smaller when they undergo meta- morphosis than are the worm-fed ones. In the Opacum group (O 1916) consisting of equally fed animals, the warm thymus animals averaged only 47.8 mm. in length at the time that the first individual underwent metamorphosis, whereas in the worm series at the beginning of metamorphosis the average size was 53.5mm. The same relationship can be observed at a low tem- perature; the average size of the thymus animals being only 57.5 mm. at the beginning of metamorphosis, whereas the worm animals had not yet begun to metamorphose when their average length was 65.1 mm. The same conditions apply in the above- mentioned Punctatum series (P. 1916); the thymus animals begin to metamorphose when their average size is 41.9 mm., but the worm animals only at an average size of 50.0 mm. As in the case of development, so in metamorphosis the rela- tionships obtaining in A. opacum and punctatum are exactly the reverse of those found in the Anura larvae, for in the former the worm-fed animals must attain a much greater size than the thymus-fed individuals before they can undergo metamorphosis INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 143 whereas in the case of the Anura larvae the thymus animals must be larger than the worm animals before metamorphosis can occur. However, in addition to the facts mentioned above, still an- other phenomenon must be described which seems to aid greatly our understanding of the relation between development and metamorphosis. If we refer metamorphosis neither to the time which has passed since hatching nor to the size of the animals but to the stage of development of certain structures, metamorphosis does not appear to be accelerated in the thymus animals but rather retarded. For example, when comparing the warm worm animals with the warm thymus animals of the Opacum group (O 1916) we see that as early as the 11th week the warm thymus animals attained the same stage of development at which the warm worm animals commenced to metamorphose. At this stage, however, a remarkable phenomenon is noted; the warm thymus animals fail to metamorphose while some of their organs continue to develop; the structures of their skin, which are responsible for the development of the color of the skin attain, while the animal is still larval a phase of development reached by the worm animals only some time after metamorphosis has been accomplished. After the warm thymus animals have entered upon the stage characterized by the crescent-shaped gills and the fusion of the melanophore spots, they should, if compared with controls, undergo metamorphosis, but instead they develop the silver- erey pigment and undergo reduction of the size of the fin. Simul- taneously (a point to be specially emphasized) they stop growing and become reduced in length, a condition which also occurs in the case of worm animals before metamorphosis. They assume an aspect which on the whole resembles that of a worm-fed animal which had undergone metamorphosis about two weeks previously. As can already be seen, these relationships can be noted much more distinctly in the cold Opacum series; but as the animals of these series have not yet all undergone meta- morphosis and the worm animals have not yet begun to meta- morphose, we will not describe the phenomena already noted. 144 EDUARD UHLENHUTH Exactly the same phenomenon can be seen in agroup of Punctatum (Group P 1916) maintained at a high temperature, such as the development of definite characteristics of a metamorphosed animal during the larval stage. In this case the yellow network was separated into yellow spots during the larval stage—a phenomenon which does not occur in the case of the worm animals before they have left the water. From what has been stated above we can see that even in those animals which metamorphosed first and, in the series of Opacum larvae (O 1916) fed with equal quantities, metamosphosed 11 weeks earlier than the worm animals, the process of metamor- phosis was disturbed. This becomes much more apparent. if for the date at which the first animal underwent metamorphosis we substitute that of the last animal metamorphosed. In that case we obtain the following relationship: In the series of Opacum larvae (O 1916) after 32 weeks have passed, 12 per cent of the thymus-fed animals are yet in a larval stage, whereas the worm- fed individuals all had metamorphosed as early as the 29th week. Thus, the period of metamorphosis in the worm series extended only over 5 weeks, whereas in the case of the thymus animals it has already lasted 19 weeks. In the repeatedly mentioned group of A. punctatum larvae (P1916) kept at a high tempera- ture, the last thymus fed animal had not left the larval stage even after 8 months, whereas the last worm-fed animal had metamorphosed after only 53 months; thus in the worm-fed animals, metamorphosis covered a period of only 24 months, whereas in the thymus animals it lasted 5 months. In a group of A. punctatum (P 1916 C) kept at a low temperature, a worm- fed series comprising individuals which out of a number of 300 larvae had not yet undergone metamorphosis was added to a thymus-fed series which had been under observation for about 5 months. In other words, this worm-fed series consisted of larvae which were abnormally late in undergoing metamorphosis. The first of these worm-fed animals left the water 5% months after hatching, the last 7} months after hatching, -the period of metamorphosis extending in this series over 2 months. Of the thymus-fed animals the first metamorphosed after 43 months, INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 145 the last (leaving two animals out of consideration) after 64 months. In the case of these thymus-fed animals the period of metamorphosis lasted 2 months, for instance, not longer than in the case of the worm animals; but 2 of these thymus animals not yet mentioned, behaved very differently from all the other animals. They both remained at a low stage of development, so far as coloring was concerned, and their tails underwent but slight reduction in size. On the other hand, the gills were re- duced to short stumps. Although neither of these 2 animals was shedding its skin (which should take place before meta- morphosis) at the time of the reduction of the gills, they were taken out of the water and placed in a vessel, the bottom of which was covered with filter paper and just enough water to keep the vessel wet. But neither of the animals showed any further change, until finally 123 months after hatching one of them shed its skin and its gills became completely atrophied, while at the same time the skin became darker in color although the yellow network failed to develop. The other animal is still in the larval stage, 133 months after hatching. We must not fail however to mention that it still appears very doubtful whether this is a direct effect of thymus, fora similar phenomenon was also noted in the case of worm-fed animals, although not to so extreme a degree. Out of approximately 300 worm-fed animals, only 1 individual showed such a condition; after more than 8 months it was still in a larval condition and had not developed a trace of the yellow network. The fin of its tail was but slightly reduced; its gills were more reduced and the animal was still undergoing growth and taking food spontaneously. It was used for the purpose of an operation, in the course of which it died. However, as has been said, at this stage it showed no trace of approaching metamorphosis. From this it seems very doubtful that the delay of metamorphosis in the two last mentioned thymus animals was actually due to the action of thymus and we must exclude them from discussion until the same phenomenon is obtained in a greater number of cases. TdE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 1 146 EDUARD UHLENHUTH 2. GROWTH In one series of experiments (P 1917) for which purpose larvae of A. punctatum which had hatched on the same day and were the offspring of the same mother were employed, it was assumed that where there is an unlimited supply of food, the amount spontaneously taken up by each animal is a function of growth, and that growth is not a function of the food quantity. For that reason in these experiments which were carried out at an average temperature of about 22°C., the animals were allowed as much food as they felt inclined to take. The group consisted of three series. The animals of the first series were given small equal-sized fragments of thymus with a pair of forceps, until each animal was satisfied. They took the pieces easily and owing to the softness of the material had no difficulty in swallowing them. The second series received frag- ments of earth-worms. Owing to the hardness of this food, however, the animals found great difficulty in swallowing it, and it took several minutes, or even hours for each piece to be swallowed. As they were fed only once a day, these worm ani- mals remained hungry and consequently were soon backward in growth, as compared with the thymus-fed animals. The latter finding coincided with the observations made in the case of the Anura; ie., that the thymus stimulates growth; but it failed to prove a specific influence of thymus, for the reason that the animals which were fed in a normal manner were found to be starving. In a third series the animals were fed with small worms (Enchytraeus), which were at first given in small pieces; these worms were thrown into the containers in such large quan- tities that the animals never lacked food. Besides this, each animal was fed on pieces of earthworms which the fast-growing animals soon took readily and in large quantities. The individ- uals of this series grew faster from the very outset than did the thymus animals. As the latter did not develop tetany until the 5th week and were in a completely normal condition until the end of the 4th week, we may look upon the result attained up to that time as the pure effect of nutrition. The salamander INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 147 larvae failed to show that the thymus had exerted any growth- accelerating influence. On the other hand, the quantity of food given plays an important part in this connection, for the animals react in a highly sensitive manner to relatively slight differences in food quantities. From these experiments it would seem that in the experiments on tadpoles conducted by Gudernatsch and Romeis the factor revealed is not a specifically growth-promoting influence, but that the accelerated growth of the thymus animals should be attributed to the fact that the jaws of the tadpoles, although adequate to supply the body with a quantity of the soft thymus material corresponding to the needs of the organism, were nevertheless not the most appropriate instrument for pre- paring from the hard beef muscle sufficient nutriment for the purpose of keeping up normal growth. The very fluctuating results which Romeis obtained in his experiments indicate pro- nounced sensitiveness on the part of the tadpoles to small quanti- tative differences of food which often completely escape control, rather than the presence in the thymus of a specific growth- promoting influence. It should also be remarked that in the above-mentioned experimental group it was also noted that the animals fed with worms must consume a much greater quantity of earth-worms than the thymus-fed animals consume of thy- mus in order to grow equally quickly; the supply of earth-worm fragments which the second series consumed was only slightly smaller than that of the first series fed with fragments of thymus. This fact speaks in favor of relatively high nutritive value in the thymus. It should also be taken into consideration that in the fragments of earth-worms a not inconsiderable part of the volume consumed consists of indigestible substance (chitin, earth) which are later eliminated in the feces. In the preceding order of experimentation it is seen that at the moment that the tetany period begins in the thymus-fed animals we are confronted by an obstacle which prevents any quantitative judgment from being formed; for from this time on the thymus animals are seen to be abnormally placed and the amount of food taken in by them becomes abnormally low. This is all the more disturbing for the reason that it is uncertain whether 148 EDUARD UHLENHUTH under these conditions the quantity of food spontaneously taken is really a function of growth. On the contrary it appears very probable that the reduced amount of food taken must be at- tributed to disturbances caused in the swallowing apparatus by the convulsions. In such a case the animals would be in a condition of starvation and in contradiction to the idea of the experiment the rate of growth would be the function of the food quantities introduced into the organism. I thought to be able to overcome this obstacle in another group of experiments, in which I proceeded from the fact that when food is present in sufficient quantities equal amounts of food produce an equal rate of growth. In a group consisting of 4 series (O 1916) for which larvae of A. opacum were used, the food was given in small fragments at the point of the forceps in all the series; an attempt was made to make all the pieces of approximately the same size on the same day of feeding. The number of pieces given to each individual animal was noted, and on each feeding day approximately (for the week) the same number of pieces was given, so that all the animals of these 4 series received approximately the same number of pieces, the series comprising one thymus and one worm group at an average temperature of 22.6°C., and one thymus and one worm group at about 14.8°C. An effort was hereby made to distribute a quantitatively equal amount of food among the 4 series; but it must be remarked that this can only be roughly attempted and cannot be exactly carried out. As it can never be known beforehand how much food the animals may need on a given day in order to be satisfied, it would also be quite impossible to weigh the food. But even if this were possible, the distribution of equal quantities according to weight would not lead to the distribution of equal nutritive quantities as a given volume of thymus contains a larger quantity of sub- stances available for metabolism than does the same quantity of fragments of earth-worms, as has been shown in the first experimental group. Although this method is not exact, it has at least furnished an approximate idea as to how important it is to control the quantity of food in such experiments. INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 149 Of course the quantity of food to be given each day was always standardized from the series which desired the smallest amount to eat. At the beginning these were the worm series, and of these the cold worm series showed less avidity for food than did the warm worm series. As a result the thymus series at first received less than they would have liked to eat. The reasons for this comparatively small appetite in the worm animals have been specified above when discussing the first experimental group. From the time that the series of warm thymus animals began to undergo metamorphosis, the animals of this series showed the least desire to eat; after that it was the worm animals in general, and the warm worm series in particular which received less than they could have consumed. It may be emphasized at this point that when this method of distributing quantitatively equal amounts of food is followed, tetany exerts a very slight influence on growth. Sometimes the rate of growth is reduced at such points where the tetany curve reaches its apex, but in other cases, on the contrary it increases or reaches even a maximum when the tetany curve does. The condition which exerts an influence on growth in compari- son with which all other influences are reduced to insignificance, is metamorphosis, as will be apparent from the following description. During the first few weeks the warm thymus animals are seen to lead in size; next in order come the cold thymus animals, then the warm worm series, and finally the cold worm series. Nevertheless no special importance must be attributed to this relationship, for as has already been stated, given an equal volume of food, the thymus animals probably obtain more nourishment frem their pieces of thymus than do the worm animals from an equal quantity of worm fragments. T he relationship of size which has just been mentioned lasts until the 10th week, and the acute tetany which has meanwhile set in among the warm thymus animals and reached its climax has failed to influence this relation at all. In the 11th week a pronounced change sets in; at this stage the warm thymus animals are all ready for metamorphosis, the first individuals being 14 days removed from 150 EDUARD UHLENHUTH this step. During this week the curve of the body size of the cold thymus animals, which up to that time occupied the second position, can be seen to cross that of the warm thymus animals. From the time that the first animal of the warm thymus series entered upon metamorphosis, the warm thymus animals com- pletely stopped growing. Their curve, which of course does not include the metamorphosed animals, is soon after crossed by those of the two series of worm animals, and the warm thymus animals remain smallest in size for the rest of the experiment. The cold thymus series, the first individuals of which under- went metamorphosis in the 24th week, also increase in size only a little from the time of metamorphosis on; but as the first animals of the warm worm series which is most proximate to the cold thymus curve similarly undergo metamorphosis in the 24th week, and also because the curve of the cold thymus animals is higher above that of the warm worm animals than the curve of the warm thymus animals is above the cold thymus animals, the curve of the latter remains the first at the beginning; it is not crossed by that of the warm worm animals until the latter have all metamorphosed. Finally, in the 29th week, together with the curve of the warm worm-fed animals, it is crossed by the curve of the cold worm series, which now occupies the first position. As early as the end of the 29th week the largest animals of the cold worm series have attained a size greater than that of each non-metamorphosed (and of course of each metamorphosed) individual of the three remaining series. As for the present (after the 32nd week) the animals give no sign of ee metamorphosis and continue to grow. The above-reported circumstance appears to. us to a the most instructive with reference to the statement that thymus- fed anuran larvae attain a size by many denoted as abnormally large but stated by Romeis never to exceed normal limits, al- though sometimes exceeding the size of the muscle-fed animals. If we begin by comparing each of the two thymus series with the corresponding worm series, we see that the same relation exists between them as between muscle and thymus-fed tadpoles, inas- much as the animals which metamorphose later attain greater INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 151 dimensions than do those which first underwent metamorphosis. It can be seen that this is not connected with a specific influence of thymus feeding, from the fact that exactly the same relation exists between the warm and the cold worm series, the warm worm-fed series which first underwent metamorphosis meta- morphosing while smaller in size than the cold worm series, and the latter continuing to grow after the former has metamorphosed. While yet in the larval stage the cold worm-fed animals attain a size which when the largest animals of both series are used for comparison, already exceeds that of the largest warm worm-fed larva by 12.5mm.' From another point of view the salamander larvae of those species so far examined show the very opposite characteristics from those possessed by the anuran larvae; for it is not the worm-fed salamander larvae which first undergo metamorphosis but the thymus-fed individuals. Thus, the point to be primarily emphasized is not the greater size ulti- mately attained by the worm-fed salamanders and thymus-fed tadpoles, for we have seen that this does not depend upon the specific qualities of the thymus, but that it is a general phe- nomenon peculiar to amphibia and one dependent upon the time at which the animals undergo metamorphosis. The point of importance in both cases—the larvae of Anura as well as of A. opacum and A. punctatum is the circumstance that thymus- feeding produces metamorphosis in the Anura only when con- siderable size has been attained, whereas in the Urodela, on the other hand, this occurs while the animal is but small in size. To summarize, we may make the following statement: The differences in the rate of growth to be noted before metamorphosis are not the result of a specific growth-promoting influence of the thymus; they are based on the circumstance that animals which are better fed grow more quickly. In the experimental group of A. punctatum (P 1917) discussed in the preceding section, these ‘Although the cold worm larvae are at the time of writing larger than the largest metamorphosed warm worm animals, we do not here intend to take up the question of this relation; moreover, a comparison of the experiments hitherto conducted in connection with the Anura shows this not to be possible, as the respective authors never observed their experimental animals beyond the period of metamorphosis. 152 EDUARD UHLENHUTH individuals are obviously the worm-fed animals of the third series, which are allowed to have as much food as they wish; in the experimental group with A. opacum (O 1916) it is the thymus animals which take in a greater quantity of nutritive material through eating thymus. The experiments furthermore show that qualitative influences exerted on the rate of growth would have to be very considerable in order that they can be experi- mentally tested in the case of amphibia, for in these animals the slightest quantitative differences, such as can hardly be controlled, would bring about very misleading differences in growth. With respect to the ultimate size attained by the animals, Salamander larvae resemble tadpoles in the fact that under certain conditions the later they metamorphose the greater is their final size; this is not only true for thymus-fed animals in comparison to worm-fed animals, but also for worm-fed animals kept in high temperature in comparison to worm-fed animals kept in low temperature. The action of thymus on development and metamorphosis may be summarized in the following way: In animals fed on thymus the development presumably of the organism as a whole but certainly of the legs, gills, shape of the head and color of the skin, is greatly accelerated during the larval period. The thymus-fed animals, therefore, reach the stage at which worm-fed animals are ready for metamorphosis, much more quickly than worm-fed animals. As development at least to some degree may be dependent on growth, on the rate of growth and on size, it is impossible to examine the specific influence upon development of any substance without keeping alike the conditions of growth in both the experimental and control series; such was attempted by admitting an equal amount of food to both series. ; When the thymus animals have reached the stage at which worm-fed animals go into metamorphosis, the development of most organs seems to stop, while certain characteristics of the skin continue to develop; the skin of such animals then behaves very similarly to the sex-organs of neotenic larvae, since the skin at least with regard to the structures determining pigmentation, INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 153 develops characteristics of a metamorphosed animal, while the animal as a whole still is in a larval stage. At the time when metamorphosis should occur disturbances in the course of de- velopment begin to appear evidently due to the suppression of the development of some factor, without which further develop- ment is impossible. In most of the animals of a thymus-fed series this factor still develops much earlier than in the controls; but even in these individuals metamorphosis becomes a grave danger to the animals’ life. In high temperature some animals die during metamorphosis and those which survive metamorpho- sis die a relatively short time after metamorphosis. In some in- dividuals the development of the factor necessary for meta- morphosis is still more disturbed and becomes delayed in com- parison with the controls; at high temperature all individuals in which this is the case die on the day when the gills and the rest of the fin undergo the sudden reduction in size, characteristic of the entrance into metamorphosis. In low temperature they may survive metamorphosis. In low temperature a very small percentage of the thymus-fed animals may remain at a low stage of development and not metamorphose for more than a year; but whether this is due to the action of the thymus diet is not yet certain, as a similar phenomenon was observed in one worm- fed animal of the stock. It seems that we cannot understand the results reported in thymus feeding experiments if we assume that they are the pure expression of the influence of the thymus substance. The rather great fluctuations reported in individuals of the same species as well as the surprising differences between larvae of Anura and Urodela when fed on thymus, indicate that quite a number of factors are involved in metamorphosis, some of which were not controlled in the experiments. It is of course clear, that differ- entiation of the organism is one of these factors; that a certain degree of differentiation is indispensable for metamorphosis, or at least to facilitate it, was shown by Gudernatsch in some recent experiments on the influence of thyroid. That some of the individuals among a thymus-fed series of Salamander larvae metamorphose earlier than the controls may be due in some degree 154 EDUARD UHLENHUTH to the fact that in the thymus-fed Salamander larvae develop- ment and differentiation and consequently metamorphosis also depend on the general conditions of growth; the experiments on Salamander larvae reported suggest that rate of growth and size play an important réle in metamorphosis. The difference noted between Anura and Urodela when fed on thymus can be ex- plained only by assuming a fundamental difference between the organization of these two groups of animals. It will be pointed out in another article that such a difference, namely the absence in the Salamander larvae and the presence in the anuran larvae of the parathyroids, seems to explain why thymus- feeding should develop tetany in Salamander larvae and should not in anuran larvae. It suggests itself that metamorphosis in part must depend on a factor similarly being present in one eroup but absent in the other group. The development of that factor may be induced primarily by processes occurring in a - certain stage of differentiation, but also may be influenced and inhibited or disturbed by thymus diet. The action upon this factor of the thymus may be widely different from that upon developmental processes preceding its development; this is indicated by the fact that development while accelerated during the larval period is on the contrary retarded from the time at which metamorphosis should occur. It is this phenomenon which emphasizes the fact that metamorphosis to some degree must occupy a particular place among the processes of develop- ment. In this connection, finally, frequent reports may be remembered according to which thymus causes disturbances of the blood circulation; in metamorphosis of the Amphibians the blood circulation undergoes a fundamental change in the course of which the gills are absorbed, and in Salamanders, the absorp- tion of the gills according to Maurer, is a prerequisite for the formation of the parathyroids. It may be worth while to keep these facts in mind during further studies of the influence exerted upon metamorphosis by the thymus. Though the effect of thymus feeding on development and meta- morphosis is very evident, it appears to the writer that similar effects may be produced by other and purely quantitative exter- INFLUENCE OF THYMUS FEEDING UPON DEVELOPMENT 155 nal conditions, such as temperature and quiéntity of food and in general by all factors which modify growth, rate of growth, size and velocity of development. No doubt such factors are of great importance in determining at what time, at what size and developmental stage of the animal, metamorphosis will occur. Since the relations between these different factors are very complicated and the number of experiments relative to them is rather small, discussion of these conditions must be postponed. Finally it should be mentioned that the thymus gland appar- ently contains all substances which are necessary to build up the substance of an Amphibian organism to maintain the animal growing and to sustain life permanently. This is demonstrated by a number of specimens of A. punctatum kept at low tem- perature which have been fed on thymus since about the 14th day of their life and are now about 14 months old; they are in-’ creasing in size. AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JANUARY 19 THE REGENERATION OF TRIANGULAR PIECES OF PLANARIA MACULATA. A STUDY IN POLARITY! J. M. D. OLMSTED FOURTEEN FIGURES Morgan (’98), in his studies on the regeneration of Planaria maculata, describes two types of operation by which he was able to obtain regenerated pieces in which “the long axis of the new head” was ‘‘at right angles to the long axis of the original worm.”’ When he cut narrow strips from the side of a planarian, he found that the piece, through contraction, assumed the shape of a crescent, the cut edge forming the concave margin. In certain cases all the new tissue which formed in the concavity of the crescent was used in the production of a head. A similarly shaped worm was formed in several cases when he cut from the side of a planarian a triangle the apex of which lay within the body. Both these methods of cutting, however, produced other pieces, which upon regeneration nearly or quite retained their original polarity. Morgan remarks (p. 373) ‘The experiments do not show clearly, why, at one time pieces cut from the side give rise to new worms having the long axis in the direction of the original long axis, and at other times at right angles to the original long axis.’ Child (715, p. 165) states that in triangular pieces cut from the side of Planaria dorotocephala the regenerated “head often devel- ops nearly or quite in the direction of the transverse axis.”’ The possibility of producing regenerated planarians whose polar- ityshas apparently been so changed that their chief axis is at right angles to the chief axis of the worm from which they were taken having been demonstrated, at Dr. H. W. Rand’s suggestion a more detailed study of the regeneration of such pieces was under- taken, the results of which are given in this paper. ‘Contributions from the Zoological Laboratory of the Museum of Compara- tive Zoology at Harvard College, No. 302. 157 158 J. M. D. OLMSTED The species of planarian used in these experiments was Plan- aria maculata Leidy, and the specimens were taken from Fresh Pond near Cambridge, Mass. Worms of various sizes, from 12 to 5 millimeters in length, were used. Some specimens, after being brought into the laboratory, were fed on liver until at the time of operation they were of the maximum size. Others, medium and small worms, were kept without food for several weeks. Neither the condition of satiety nor of starvation noticeably influenced regeneration. At one time the mortality of one lot would be greater, at another time, that of the other. In the fed worms, however, it was found best to allow one week to elapse after the last feeding before the operation was performed. To prepare the planarians for operation, they were narcotized in a 0.1 per cent solution of chloretone until they ceased to move. Cuts were then made with a sharp scalpel, care being taken to have the cut edges as nearly straight as possible. Triangular pieces were taken from all regions of the body, each triangle having for one of its sides a portion of the origina! uncut right or left margin of the worm, and, for the other two sides, cut edges which intersected near the original median axis of the worm (fig. 2a, 5a, 7a). The two cut edges, intersecting at a point which I shall refer to as the vertex of the triangle, are distinguished in the following account as the anterior and posterior edges. It was only towards the end of experimentation that the im- portance of fairly exact measurements of the lengths of the cut and uncut edges, the angle where the cut edges meet, the distance of the vertex of this angle from the median axis of the worm from which the piece is taken, and the size of the piece, was realized. In the earlier part of the work, no camera drawings were made until the day following the operation. Because of the decided contraction of the pieces at this time and the con- sequent distortion of their original shape, it was possible to estimate only rather roughly their original measurements. Later in the work, however, camera drawings were made immediately after operation while the pieces were still in chloretone, the very slight contraction in this condition being negligible; a second drawing of each piece was made on the day following, when they A STUDY IN POLARITY 159 were in the contracted state. The drawings made while the pieces were still in chloretone formed the basis for classification into groups, according to the relative lengths of the cut ‘edges, the size of the angle at the vertex, etcetera. The drawings made on the day after the operation during the earlier experi- ments were compared with those of the later work and each of the earlier ones was placed in that group which it most resembled. One may fairly assume that pieces which resemble one another on the day after operation would also have been similar immedi- ately after the operation. Thus it was possible to estimate with some degree of accuracy the measurements which the triangular pieces in the earlier work had immediately after operation. In the following account it was thought best, however, to enumer- ate the cases separately; hence the earlier experiments, in which the original measurements are estimated merely, are referred to as Series I, whereas the later ones, in which the pieces were drawn while still in chloretone, are designated as Series II. The mortality of such triangular pieces is very great. Less than one-fifth of them survive the operation and regenerate. Pieces taken from the region of the pharynx (fig. 5a) had the greatest vitality, though regeneration of pieces from other regions of the body, if accomplished, proceeded along exactly the same lines as in the pieces from near the pharynx. Bardeen (’03) found that in Planaria maculata he could more frequently obtain double-headed worms from cross-pieces when they were taken from the pharyngeal region than when from any other region of the body. Morgan (’04) was also more successful in getting pieces from this same region to regenerate, but he remarks, “Whether this is only because shorter pieces are more easily obtained here, or because the very short pieces from this region survive the operation, remains an open question.”’ The latter explanation seems to be the true one, since in many cases In My experiments the same sized pieces were taken from all regions of the body and only those from near the pharynx survived. When, in the operation of cutting, the epidermal layer is broken, a great mass of loose parenchyma cells flows out from the wound, and if the two cuts form a very acute angle, the 160 J. M. D. OLMSTED projecting point on the triangular piece becomes rounded off by loss of material (cf. Morgan, ’98, p. 393). Immediately after the operation there is always a very slight contraction of the cut edges, even though the piece is still immersed in chloretone. This, no doubt, is due to the direct stimulation of the muscle fibers. As soon as the effect of the narcotic is gone, the piece contracts ¢reatly, often assuming the form of a hollow cone, the apex of which lies approximately at the center of the dorsal surface of the piece. Epithelial cells soon cover the wound (Lang, 712, p. 272), and after twenty-four hours new white tissue can be seen along the cut edges. This new material is never evenly distributed along the cut edges, but (figs. 2b, 5c, 6c) a greater amount of it appears near the center of the anterior edge, a less amount along the posterior edge, and very little at the vertex (cf. Morgan, 798, p. 378). cosr eae oon 178 PNIECERIONRODSOLe uate RUNS Some NC Saari. tated ERR ORR. 178 Il, INGACITOMNE tO OlinCuorry Suinomtlbl, oj shcekeoasbbeseconcedcepocousou- 179 Za OrcancEschisithve suOMOOGHexGRaCiSe =... : hese eG acece etre sees - 184 ALOU STOLEN GIA) Manne Ga hwals Ge Low Sens Om Poeiras Ue oo ae artaeone 184 TEAS TOYO Eh ak Sere aie A te ve! 9 ee oi epee iets aa EEL a ie rea a gE 186 IMaVED Sitio) cKO ce Mie ere Rome Nee ote ork iris A Reis Cnet ne nee a 187 COME USTONSNes mepess « vHORES oak Ske Ee RC as HAE ARS ase eyo 191 IBWCHTEOMN Canal ou) ERAitteme eee ame eae sec ono es Neo oom co eae apiee aeone 192 [PORcansesensitivertO LOOd exXtnacts. 4.6. oneness oe al 192 hiemooupmantlesancaventaclesses ss snnn ee ne een Menno reer 193 ‘TPIAVE Sj OlaVONAAN ASLO Sra Hige ac too e Sia oe Ole o Gin acc c Gem tas oo ea ae 194 D. “Mae inincliare Gi Clieenly WOyNel sane coeds ae se ee noel ces coGeeb eb odeoe 196 Sy INn@ @gjalarecligiannse: Goaons econ eons poe Se etic ds babe oe eee 208 PLDSRD IS CUSSION MERE wey shine cts Sane wean cc a teers inte: yen te ens, Sie 216 1. Movements resulting from olfactory stimulations................. 216 Dre AS COHAN OMSINe lll pee ae Pe teers Auch ete oi Ao Mele ciate ole ole ees 221 TW MSU TT aT yar eee ee En EINE ECT TIN SES oe Se scp os sits S Seeds sie ohsts 224 We oiteratinerc ited peewwaa ext ain Ps ae atc eames hye tery oo Aisi s aierele es 227 I. INTRODUCTION It has long been recognized by fishermen and naturalists that many species of marine carnivorous snails are conspicuously successful in finding food. ‘To the former they may be a source of considerable annoyance through their habits of entering lobster pots, eating fish entangled in nets and feeding upon bivalve mollusks, sometimes doing so much damage that like the whelk and oyster drill they become serious pests. The extensive 177 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. 1 178 . MANTON COPELAND literature devoted to mollusks contains numerous references to these gasteropods congregating in great numbers about dead animals, and several observers have described buried snails coming out of the sand when food was placed in the water near them. This ability of the carnivorous snails to find food has quite generally been attributed to a well developed sense of smell, but very little definite information concerning their olfactory reactions and organs appears to have been obtained by experimental methods. Most recent writers in discussing the sense of smell in this group refer to the work of Nagel (’94), whose studies on marine gasteropods however were evidently limited, and who, by stimulating them with strong or irritating substances, used stimuli which were inappropriate for calling forth olfactory reactions, at least so they could be distinguished from other types of chemical responses. Accordingly an experi- mental investigation of olfaction in two species of snails was undertaken in view of determining particularly the sensitiveness of the animals and the characteristics of their responses to olfactory stimuli, whether they are directed to food or find it by random movements, and the location of the receptors con- cerned in the reactions. A part of the work was carried on at the United States Fisher- ies Biological Station at Woods Hole, and I wish to express my gratitude to the resident Director, Dr. P. H. Mitchell, for many favors received during my stay. ¢ II. EXPERIMENTS ALECTRION OBSOLETA An investigation of olfaction in marine gasteropods was begun by a study of the food reactions of the mud snail, Alectrion obsoleta (Say), formerly included under the genus Nassa or Ilyanassa. It is a rather small species, the expended foot measuring from fifteen to eighteen millimeters in length, and about ten millimeters in width in its broadest part. The an- terior border of the foot is extended on both sides so as to pro- duce recurved processes, approximately one millimeter long, OLFACTORY REACTIONS OF MARINE SNAILS 179 provided with sharply pointed tips. The slender tapering tenta- cles, about five. millimeters long, are borne on stout processes projecting in an antero-lateral direction three millimeters from the head, and the eyes are situated at the base of the tentacles on the outer side. The siphon, which runs forward above the head, represents a prolongation of the mantle. The ventral wall is split its entire length, but when the edges are brought together a tube is formed, open at the tip, through which water passes to the gill within the mantle chamber. The organ pro- trudes from ten to fourteen millimeters beyond the anterior border of the shell, and can be swung about in various directions. Alectrion occurs commonly on tide flats, where it is a thorough scavenger, feeding on a variety of food substances, but often to be seen collected in great numbers on some dead animal such as a fish or- a crustacean. Dimon (’05) writes of its food as follows: Among the various food materials upon which Nassa was seen to feed were hen’s egg shells, dead hermit crabs and Squilla, live Nereis, ulva, the alga that grows upon the shell of Nassa itself, the thick black mud of the inner harbor, and the alga that gathers on the glass of aquaria. . . . . Nassa did not usually attack a live uninjured clam, though I have seen the snails collected at the edge of the mantle of a clam that was apparently alive, devouring it and preventing the shell from shutting by pushing themselves into the opening between the valves. Belding (710) has shown that Alectrion occasionally feeds upon the living scallop, gaining access to the soft parts of the mollusk by entering between the open valves. 1. Reactions to olfactory stimula That an olfactory sense plays an important part in the dis- covery of food by Alectrion seemed probable. Dimon (’05) reports finding a Squilla about six inches in length on the mud, with ninety-eight mud-snails crowding about it. It was then taken away from these and put into a pool about eighteen inches in diameter, which was full of quiet mud-snails. These snails immediately became active, but without moving definitely toward the Squilla. When one happened to reach it, it stopped and began to 180 MANTON COPELAND eat, so that in a short time they were once more gathered thickly about the Squilla. The same test was repeated in another pool, with the same aimless wandering and gradual collecting of the snails about the Squilla. In the laboratory the Squilla was put into an agate pan about sixteen inches square, in which were thirty-five mud-snails. The effect was the same as in the pools out of doors. The snails moved about actively at first, and in the course of twenty minutes, twenty of them had collected on the Squilla and the others had quieted down. The Squilla was large and had probably been dead for some time, so that its odor or taste was quickly diffused through the pool. Towards smaller, live creatures, such as Mya arenaria with a cracked shell, the response of Nassa was less rapid, and was not at all definite, unless the snail came very near the dam. The definiteness and promptness of response seemed to vary somewhat with the nearness of the animal to the stimulus, and also with the individual snail. This description gives the impression that the food was scented by the snails, but that its discovery was perhaps largely a matter of chance. In the hope of gaining more detailed information on the behavior of Alectrion in the presence of food, the following experiment was tried: A shallow rectangular glass dish, measuring approximately thirty-eight by twenty-three centimeters, was filled with sea water. At one end of the dish was placed a ball of cheese cloth about three centimeters in diameter, and at the opposite end a cheese cloth bag of the same size Gontaining fresh fish (Fundulus) meat. Both packets were weighted with a pebble to hold them in position. Ten snails were then placed along a line midway between the two packets and their behavior noted for one hour. That the snails scented the fish when some distance from the baited bag was soon made evident. The bottom of the dish was slightly concave at the margins, so that the fish juice slowly drained toward the two corners nearer the baited bag. The snails which moved into the region of the juice extended their probosecides, and worked them over the surface of the glass. This proboscis reaction, which later was carefully studied and found to be characteristic of snails stimulated with dilute food materials, often beginning when the animal was several centi- meters away from the packet, was particularly marked where the juice drained toward the corners of the dish, and sometimes occurred in the corners, which were about ten centimeters from OLFACTORY REACTIONS OF MARINE SNAILS 181 the tish meat. The proboscis was extended farther when the snail approached the source of the stimulating material closely, and two or three of them chimbed upon the bag in their attempts to secure the food. Sometimes the snail’s siphon was brought within one or two millimeters of the bag before the proboscis was protruded. This was especially true when the bag was approached from the center of the dish toward which less Juice was spreading. Eight of the ten individuals came near to the bag and extended their proboscides, the latter reaction always occurring before the siphon touched the cheese cloth. No behavior of this sort was observed at the opposite end of the dish, although six individuals came into contact with the un- baited packet, touching it ten times during the hour. They always moved away from it and no proboscis activity took place. Other experiments of a similar kind, varying somewhat in details, were carried out, and gave essentially the same results. That dilute materials emanating from fish meat stimulate Alectrion and cause the animal to thrust its proboscis, in short, that’ the snail scents distant food, was clearly demonstrated. Whether the final discovery of the food is the result of chance movements, or of some directive influence of olfaction, is a prob- lem which will be discussed later, particularly in connection with the reactions of Busycon. The responses of Alectrion to distant food in a current of water were next investigated. Dimon (’05) describes the activities of the mud snail when in a stream containing food juices which flowed down a beach, and concludes that “if a current flows from the food to a snail, the animal will crawl up toward the food.”” The same author also finds, however, that the snail “shows a tendency either to move against a moderately strong current, or to orient itself when at rest with its head pointing against the current.’ The reader, therefore, is left somewhat in doubt as to the significance of the snail’s movements against the current in the first instance, and may ask: Were they affected by the presence of food up the stream? In order to determine whether food juices added to the water caused Alectrion to move more frequently against the current, 182 MANTON COPELAND the following tests were made: A wooden box open at the top, nearly one hundred and seventy centimeters long and about twelve centimeters wide, was lined with glass, and placed in a horizontal position on a table. A rubber tube for conducting sea water was inserted in one end of the box near the bottom. The water flowed the length of the box and out an opening at the opposite end. The current was moderately strong, and care was taken to keep it as uniform as possible throughout the experiment. A snail whose reactions were to be studied was placed in a position near the middle of the box, and all its movements with or against the current recorded until it passed a line seventy-five centimeters, either down or upstream, from the starting point. Each animal was first tested four times in the current without food. Two, which moved against the current to the seventy- five centimeter line in all their trials, were discarded, as the experiment necessitated the selection of individuals which showed at least some tendency to move in the direction of the current. Five animals were finally selected which answered the requirements. After determining their responses to the current alone, they were tested again with two fish (Fundulus) placed near the head of the stream above the seventy-five centimeter line. The fish were constantly bathed by the water after it left the tube, but were so situated on the right and left margins of the stream as not to interfere with the force of the current. They were cut open, and were turned over or moved from time to time during the tests in order to disperse the juices more freely. In trial one, throughout the experiment, the snail was placed heading across the current, so that its initial impact was on the animal’s left side. It was started in the opposite direction in trial two; in trial three it was headed down stream and in trial four, upstream. The position of the snail at the beginning of the test, however, seemed in no way to affect its later activity and the final result. A single snail was given but one trial a day, and all were kept in good physiological condition by occasional feeding. . Table 1 shows the result of the experiment. OLFACTORY REACTIONS OF MARINE SNAILS 188 TABLE 1 Showing the distances in centimeters moved by mud snails and their arrivals up and down stream, when started midway between two lines 150 cm. apart in a current with and without food juices IN CURRENT WITHOUT FOOD JUICES IN CURRENT WITH FOOD JUICES ANIMAL TRIAL Distances moved Arrivals Distances moved Arrivals NUMBER |NUMBER Up Down Up Down Up Down Up Down stream stream stream stream stream | stream | stream | stream Ags 75 0 + 80 5 a 1 2 0 75 Be 75 0 -- 3 0 75 + 75 0 — 4 0 75 aa 81 6 ~ 1 0 75 a 75 0 a 9 2 90 15 SF ilile/ 42 — 3 120 45 + 119 44 + 4 51 126 se 134 59 + 1 18 93 + i aa + 3 2 116 41 =e 1D 37 os 3 0 75 + 63 138 a 4 7 82 + 173 98 + 1 0 75 + 61 136 a 4 2 75 0 =- 4 79 3 0 75 Se 80 5 a 4 15 90 a 10 85 1 53! 0 = 77 2 a. 5 2 632 0 = 75) | 0 os 3 87 12 + (fa eee ee = bl aa 0 75 + aoe Oy. ee | | —| - Gta 770 1104 6+ | 12 1636 | 736 | 16 4 1 Closed trial with snail resting at 53 em. line. 2 Closed trial with snail resting at 63 em. line after starting it once by touching it. Of the twenty preliminary trials in the current alone, there were eight in. which the snails either arrived at the seventy-five centimeter line upstream, or showed a marked tendency to move against the current. The number of arrivals upstream was doubled, however, when the fish were placed near the head of the current. The five snails traveled 770 centimeters against 184 MANTON COPELAND the current without food and 1636 centimeters with the food present. The distances moved across the current are not recorded. A snail failed in but one case to increase its arrivals upstream when the fish were there (Animal number four). It should be noted, however, that this individual showed a greater tendency to move against the current when food was present than it did in its,absence. The force of the current infrequently caused a snail to lose its foothold and slip a short distance on the glass, but it soon recovered itself and continued its locomotion. The total distance moved as a result of slipping in the forty trials was but little over fifty centimeters, over forty of which were recorded for this same animal in its trials with food juices in the water. The average time taken by the snails in reaching the seventy- five centimeter line up the current without food, in six trials, was: approximately twenty-seven minutes, and with food present, in sixteen trials, twenty-nine minutes. The experiment indicates that snails which are inconstant in their reactions to a current, or more often go downstream, move more frequently against a current when it carries dilute food juices; in truth, they exhibit olfactory responses leading to the discovery of distant food. When the animals tested were al- lowed to continue their progress against the current beyond the seventy-five centimeter line, they usually arrived at the fish and began feeding. 2. Organs sensitive to food extracts Having made certain that Alectrion responds actively to dilute food stimuli, experiments were begun to determine the sensitiveness of the external parts of the snail to odorous material, in the hope of discovering the olfactory receptor. An extract of fish meat was prepared by grinding muscle tissue of Fundulus in sea water and filtering the product. A small amount of dry carmine was added to the filtrate, in order to make it visible in water. The tentacles. The tentacles were first tested. When the fish extract was squirted over the tentacle of a moving snail by OLFACTORY REACTIONS OF MARINE SNAILS 185 means of a finely drawn out pipette, a marked reaction followed. The tip of the tentacle coiled rather violently, the animal stopped, the siphon was swung into the stimulating material and the proboscis extended and worked over the bottom of the glass dish in which the tests were made. Numerous animals were tested, and although considerable individual variation in respect to sensitiveness to the stimulus was noted, the reactions were remarkably constant, as the following responses of five snails will show. Sea water mixed with carmine was always applied with a special pipette to the tentacle before the fish juice in order to make sure that the reactions observed were due to chemi- cal rather-than tactile stimulation. The snail was moving when the test was made, and at least a minute elapsed between each trial with the fish juice. The material was applied five times to the right tentacle, and then five times to the left one, and care was taken to have the snail in water freé from the stimulat- ing substance when the test occurred. In fifty trials, ten with each individual, there were forty-seven reactions as described above. One failure to respond was noted, and two cases where the proboscis was extended but locomotion continued. ‘The animals scented the juice, therefore, forty-nine times in fifty trials. To sea water and carmine the snails usually responded, if at all, by twitching or coiling slightly the tips of their tentacles without cessation of locomotion. With the exception of one doubtful case, the proboscis was protruded but three times in the fifty trials, and then without the animal stopping as it characteristically did when stimulated with fish juice. Another method for comparing the effect of a pure tactile stimulus with that of one accompanied by a chemical stimulus was as follows: A small piece of cotton was rolled into a ball and placed in the open end of a pipette. Some filtered Fundulus extract was then put into the pipette back of the cotton. By exerting slight pressure on the bulb it was possible to flood the cotton with juice, which then could be applied locally to any part of the snail’s body. JUMPROG LIKCUNO) Neo ye Pro aeie abc o Bodao De OD COOaC cnt CPA O ESO co IoC e Ger mOre 229 likeli xternaleappearan cen... verse ere ence racine Gieicionieis Sc eke aes es elles 230 OTIC GL ON yes TC PoE Corea ste miso late Seis ee eis Seas 230 Dee CC OVOTE SL 5 eben RO ae ec hide aeitis Fae aie ote ve 232 3), IM oIMIMPR AKON’ LSE Se ercoctS Hood bbe buced Sma at ers SUED apoE Cab eece 234 ASP Ain phi pod eCOMMVeEn salle everett ey aes n ears even yar oleseot ax oockeh al olen oie 237 ELE Hood and energy supplys- ca. cerreetr sar eters ns ate wave sche ie's state w Slay vietelera 237 LER WiateTACUrrenit crite vat cree eee Te ee eee eT ee ene hal or leverer 238 FAS VOLE HIV EYE Bye Soe Reon bho bts GAs GiB eR ERS CRI GAO backs Bieta 6 Eten es Ee 240 IWarEheemovementsy osc crditas see rer vaio creo okeke ier e ede okey e a aieke ces lets 243 is Biola TerbaOlOSUas esndsens dogccns Uda pbobe Aeon Coes coe oe Coe damone 243 Dees OMANI VEMEMNLG sree crn tarrme: Someries ckarsrocrarste cic ce ete eneyet a eucte 248 SR OPOMUANCOUS HIN OVEIMEN Use aerate s. Muet. eter eee EN eee eiees = 253 Vip Suan nvpyrr artes 2.043 fer PHN ees aes ak edge bigtamsoersy: Feqaeeeias 257 Vile {Bil lyo grap liye en teara an ne oe area oe cue Disb Se wiee Basisc ic + cehieesiseit ers 258 I. INTRODUCTION Ascidia atra is a large tunicate common in Bermuda and the West Indies. It was first described from Guadaloupe in 1823 by Lesueur. Many years later it was found in Bermuda by the Challenger Expedition, and figured by Herdman (’82), who, however, confused it with the European form Ascidia nigra Sav. The distinction between the two is that the European species possesses intermediate papillae on the longitudinal bars of the branchial sac, while the American species does not (VanName, 02). ! Contributions from the Zodlogical Laboratory of the Museum of Compara- tive Zodlogy at Harvard College, No. 303, and contributions from the Bermuda Biological Station for Research, No. 78. 229 230 SELIG HECHT Ascidia atra lives attached to rocks, in most cases well under low water. The attachment of the larger animals is by the posterior edge; in the smaller individuals a portion of the left side is frequently attached as well. The species is well distributed throughout the Bermuda Islands. Although animals for this research were obtained from many regions, the main supply came from localities very near Agar’s Island. Ascidians first became of interest during the last half of the nineteenth century. Their significance in relation to the origin of vertebrates, which was first made apparent by the work of Kowalevsky (’67), resulted in innumerable researches on the anatomy and embryology of the group. Since that burst of activity, sixty years ago, the knowledge of ascidians has not kept pace with the newer points-of-view. As a consequence, little, indeed, is known of the life and activities of these animals. It is with the hope of supplying this deficiency that the present series of papers is presented. The work was done at the Bermuda Biological Station with the assistance of a grant from the Humboldt Fund. I wish to acknowledge the kindness of Prof. E. L. Mark, who made this grant possible, and at whose invitation the experiments were performed in Bermuda. To the Resident Naturalist, Dr. W. J. Crozier, the depth of whose friendship was manifest in many ways, I extend my frank admiration and thanks. My chief indebtedness, however, is to Prof. G. H. Parker. His teachings and researches, which have the “rare merit of combining both anatomical and physiological view-points,”’ have influenced my work and thought. It is a privilege to express my gratitude for the inspiration which he has given me. Il. EXTERNAL APPEARANCE 1. Orientation At the outset of this account of the physiology of Ascidia atra, it is necessary to define the various surfaces and planes of the body. The earlier writers on ascidian anatomy were far from agreed on the application of such terms as anterior, poste- PHYSIOLOGY OF ASCIDIA ATRA LESUEUR Zot rior, dorsal, and ventral. As a result, however, of the embryo- logic investigations of his time, Kupffer (75) applied these desig- nations correctly. The nomenclature suggested by him has been accepted by all of the later workers (Herdman, ’82), and will be used in the description of the present species. Fig. 1 Medium sized specimen of Ascidia atra, life size, showing view of right side. The large opening of the animal is within the rim of the oral siphon, and the smaller one is within that of the atrial siphon. These two openings are at the anterior end, the oral being to- ward the ventral edge, and the atrial toward the dorsal edge. The place of attachment is at the end opposite the siphons, and constitutes the posterior part (fig. 1). Consequently, the side of the animal which shows many creases and irregularities in the test is the right side, and the smooth face forms the left side. DSP, SELIG HECHT According to this scheme, the pharynx, or branchial sac, of Ascidia covers the right side almost completely, whereas the renal body and the intestine lie on the left side. The atrial cavity extends along the dorsal edge and terminates anteriorly in the atrial siphon. Quite unusually for this genus, the heart lies on the right side along the ventral edge. 2. Color On first sight Ascidia atra appears to be of a dead black color. Closer examination shows that it is really a very deep blue. The color is located in the test, on the inner surfaces of the siphons, and on the outer faces of the oral tentacles; it may even extend beyond the tentacles into the anterior part of the branchial sac. In sections of the animal, the blue pigment is shown as a thin rim marking the outer edge of the test. The coloring matter is insoluble in water. Acetone extracts of the test are reddish in appearance, and do not resemble the opaque purplish blue seen in sections of the test. The extract has the properties of an indicator; it is red with acids and green with alkalies, the color change occurring near the neutral point (Crozier, 716). The blue pigment is contained in spherical granules, which are nearly all of the same size: approximately 3 micra in di- ameter. They occur in very compact groups of four to six. Although each group may represent a cell containing the pig- ment granules, it is difficult to make out any cell substance or cell boundaries in thin sections of fresh and fixed tissue. There- fore, it seems improbable that the granules as they are found in the test are within the living substance of a pigment cell. The groupings may, however, represent the remains of meta- morphosed cells whose cytoplasm has disintegrated. On the basis of such an idea, there should be present in the body of Ascidia some living cells which would be the precursors of the pigment groups; and, moreover, it should be possible to find intermediate stages between the two. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR Dae It was not difficult to satisfy the first of these requirements. The blood contains cells whose volume is the same as that occupied by the compact group of pigment bodies. In addition, these blood cells are packed full of spherical granules whose size and number correspond to the pigment granules in the test. Most of the blood cells are of a rich green appearance, the color being resident in the granules. There are also to be found, in much less abundance, similar blood cells whose granules instead of being a transparent green, are an opaque, dark blue, which to all appearances is identical with the color of the test. It was possible to observe the change from green to blue in individual cells in drawn blood under the microscope. I considered it, therefore, extremely likely that the blue blood cells, representing the later stages of the green cells, are the forerunners of the groups of pigment granules in the test. In order to prove this satisfactorily, it was necessary to find a stage between the free, blue cell and the group of pigment gran- ules imbedded in the cellulose of the test. Examination of sections of the test brought out only a few, and these doubtful, instances, indicating that pigment deposition in an adult Ascidia is probably not a very active process. The evidence came when it was found that an animal would regenerate its test. An individual denuded of a portion of tbe test began almost at once to secrete a new one. At the end of one day, a thin layer of cellulose of the characteristic color had been formed over the denuded portion. When this delicate layer was removed and examined with the microscope, there were found hundreds of definitely shaped, blue blood cells imbedded in the cellulose, imparting to it the usual color of the test. Asa result of this, it seems safe to conclude that the blue pigment granules in the test of Ascidia are the remains of the metamorphosed green cells of the blood. In this connection the observations of Caullery (95) are of interest. Botrylloides cyanescens, which in nature is yellowish green, turns blue after remaining in the laboratory. Caullery found that the green color was due to cells which contained a number of colored granules, and that the blue appearance in 234 SELIG HECHT captivity was the result of the change of these granules to a deep blue. The figures which he gives (Caullery, ’95, fig. 52) for these cells resemble the blood cells of Ascidia and of ascidians in general (Cuénot, 91). The color change in Botrylloides is artificial; in Ascidia atra it is in the regular course of events. 3. Formation of the test Freshly collected specimens of A. atra, as well as animals in their natural surroundings, possessed a bright and clean appear- ance, which was often lost in the laboratory in a short time. In confinement, the outer surface of the test soon changed to a dull gray. The gray material was gradually sloughed, coming off in shreds, and resembling a human skin peeling after a sunburn. Although the layers were removed by the movement of the water, more appeared in a short time, and the animals continued to shed the outer portion of the test as long as they remained in the laboratory. The animals were kept in battery jars of about ten liters capacity, into which the seawater flowed in a gentle stream. Under such circumstances, the water surrounding the animals had but little motion. This is quite in contrast to the compara- tively turbulent conditions to which the species is normally subjected. Consequently, it seemed probable that the ap- pearance of the test was merely a superficial laboratory product, and not due to any real effect on the animal. Indeed, individuals kept in smaller jars, in which a more vigorous current was present, showed little sign of this surface change. The sloughing, therefore, merely indicates that Ascidia renews its test continu- ally by secreting fresh test material on the inside, and allowing the outside surface to disintegrate and to be removed by the action of the waves. This conclusion is strengthened by the phenomena which attend the regeneration of the test. Occasionally animals were collected which showed an appearance that can be interpreted only as a regeneration of the test and perhaps of other structures (Hirschler 14). Figure 2 is a sketch of such an individual. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR Zoe The test shows a ragged surface undoubtedly representing the places where it had been torn. Such a test is highly instructive. In cross section (fig. 2) it can be seen that regeneration had not taken place by the mending of the injured edges, but by the growth of the new test around the body tissue. It does not seem as if the injured portion had been specially reconstructed ; but rather as if the test material had been secreted by the sur- face generally, and only incidentally covered the injured part. These appearafices may be duplicated experimentally. Asci- dians present varying degrees of ability to regenerate the test. Some, like Cynthia and Phallusia, seem incapable of surviving Fig. 2 Sketch of regenerated animal. The transverse section of the test shows the mode of regeneration. even partial removal of it (Fol, 08). Ascidia atra, however, not only recovers rapidly after portions of the test have been removed, but it makes good such deficiences in a short time. It has been described how, within a day after the removal of a section of the test, there has already been secreted a thin layer of cellulose. This film is continuous with the innermost surface of the uninjured test, and adheres closely to the soft tissue which has formed it. Animals in this condition, when placed in a sheltered position in the natural environment of the species, continue to thicken the film until it assumes the ordinary dimension of the test. In one animal, for example, a hole several centimeters square was mended in two weeks. Sections of such tests are identical with those of animals like the one in figure 2. 236 SELIG HECHT The edges of the old test gape at the injured region, and the new test which is formed underneath them is continuous with the old test where the two meet close to the mantle tissue of the animal. It was impossible to secure complete regeneration in an animal whose entire test had been removed. ‘This was due solely to a deficiency in technic, and not to an incapacity on the part of the animal. The seawater supplied to the laboratory contains almost no plankton organisms, and since thesé furnish the food supply of Ascidia, even perfectly normal animals died after a week or more of laboratory confinement. I was unable to devise a method of keeping animals without a test out in the open water, because they could not be attached to anything. Consequently it was possible to observe them only for a few days in the labora- tory. Here they showed the usual beginnings of regeneration. After a day, a thin layer of test material was formed on the right side of the body and on the entire surface of the siphons, but not on the surface of the renal body. Soon the cellulose began to extend over to the left side, making the bare region smaller and smaller. Undoubtedly, under better conditions, a complete test would have been regenerated. The newly formed test material is pigmented in the usual way. The pigmentation is not dependent on the presence of light. Animals whose tests had been removed at night, were kept in complete darkness for several days. They regenerated as usual, and the fresh test contained the blue pigment. Moreover, a new, pigmented test will form on the right face of the animal under the intact, opaque, old one, when the latter has been accidentally separated from the ectodermal surface which secretes it. Therefore the formation of a pigmented test is not the result of a photic stimulus only. These regeneration experiments, as well as the phenomenon of sloughing, indicate that normally there is a continuous addi- tion to the thickness of the test, in order to compensate for the disintegration of the exterior, and for the changing size of the animal. PHYSIOLOGY OF ASCIDIA ATRA LESUBUR ae “ 4. Amphipod commensal One of the striking things associated with the external ap- pearance of Ascidia is the frequent presence of the young of a species of Orchestia within the cavity of the oral siphon. The occurrence of crustacean messmates in ascidians has long been the subject of comment, and many species of copepods have been described for ascidians all over the world (Scott, 07). However, with the probable exception of Verril’s (’70) report of an amphipod in the ‘interior’ of Ascidia callosa, I have found no previous record of the free association of an amphipod in the branchial cavity of an ascidian. The species in A. atra is a pretty animal varying in size from two millimeters to nearly a centimeter. It possesses bright red eyes and a dark band across the middle of the back, both struc- tures showing conspicuously against the whiteness of the body. In the oral cavity of an Ascidia which has not been disturbed for a time, the amphipods are arranged near the rim of the siphon with the anterior end facing outward. Frequently as many as ten may be found in this position in a single siphon. It is a startling sight, when the blackness of the interior of the siphon is illumined, to see the brilliant red eyes of the creatures arranged in a circle a few millimeters within the cavity. The amphipods are capable of rapid locomotion when forced to leave their host, and may perhaps be free living at times. Their position in the oral siphon of Ascidia, however, is of dis- tinct advantage to them. The water current entering the oral siphon brings with it a host of small organisms to serve as food for Ascidia. The amphipods share this with their host, and, therefore, furnish an example of real commensalism. III. FOOD AND ENERGY SUPPLY The only source of metabolic and growth materials which is available to Ascidia is the surrounding seawater with its sus- pended and dissolved content. In order to utilize this supply, the animals perform certain activities whose function it is to furnish quantities of fresh seawater continuously, and to remove 238 SELIG HECHT therefrom the substances necessary for the existence of the spe- cies. Both of these processes are accomplished in the branchial sac. ‘The enormous development of this structure testifies to the necessity of working on a large scale in order to abstract the relatively meager proportion of food and energy contained in the seawater. 1. Water current A study of the water current (Hecht, 716) has already shown that this form of activity has the following properties. The current is produced by the cilia of the branchial sac. It is maintained under a low pressure of 1.7 mm. of seawater. The quantities of water moved are large; in a medium sized individual, 173 liters of seawater are transported in a day. The volume of water moved per unit body weight varies inversely as the size of the animal. Since the water enters by way of the oral siphon, and leaves through the atrial siphon, it is of primary importance to Ascidia to avoid a mixing of the incoming and outgoing currents. In the open water the movements of the sea undoubtedly change the water immediately surrounding an individual, so that a fresh supply of seawater is frequently available. Ascidia, however, does not rely on such a chance renewal of its food and energy supply, because even in very quiet water, such as that in a large dish in the laboratory, the two water currents are definitely iso- lated from each other. If, in such a dish, particles of carmine are floated near the atrial and oral siphons, it is at once apparent that the outgoing current is considerably stronger than the incoming current. Figure 3 shows a drawing of a small specimen of Ascidia life size. The arrows near the oral siphon indicate the range of its activity, that is, the distance from the opening within which a particle of carmine was sucked into the cavity. For this specimen the distance was at most 5 millimeters. The arrow pointing away from the atrial siphon represents the distance within which a particle was deflected by the outgoing current. This range was ca. 65 millimeters. Other individuals showed similar relations PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 239 for the range of activity of the two currents. In columns 2 and 3 of table 1 are given the values obtained for two addition&l animals. Record VI.20.3 is of an individual about twice, and VI.22.1 of an individual about four times, the size of the speci- men in figure 3. g 202 Fig. 3 Range of the incoming and outgoing streams of the water current. ‘ The maintenance of such a contrast in the force of the two branches of the water current is accomplished by a difference in the size of the openings of the oral and atrial siphons. The cilia furnish the motive power, and the same quantity of water which they move passes through both openings in the same time. 240 SELIG HECHT TABLE 1 RANGE DIAMETER ANGLE ANIMAL BETWEEN Incoming Outgoing Oral Atrial EOIN mm. mm. mm. mm. ViAE2082 5 65 4.0 220 43° V1I.20.3 HZ, 110 ono 3:5 65° WALPPAI 15 230 13.0 6.0 35° The oral orifice is large, whereas the atrial is small. Therefore, the velocity, and consequently the momentum, of the water in the atrial current is greater than in the oral current. The di- ameters of the siphon rims of the three animals mentioned are given in columns 4 and 5 of table 1. The figures are not very accurate, because of the difficulty of maintaining the living animal in a constant state of expansion. They show unmistakably, however, that the difference in the force of the two currents depends, in the main, on the size of the siphon orifices. A second significant factor concerned with the separation of the two currents is the angle formed by the diverging axes of the two expanded siphons. In the last column of table 1, this angle is recorded for the same three animals. The individual variation in the extent of the divergence of the siphons is sur- prising; the net result, however, is that the currents are prevented from mixing. The combination of a difference in range with a difference in direction of the two streams of the water current makes Ascidia independent of the fortuitous movements of the surrounding sea. 2. Feeding The stream of seawater which passes through the branchial sac of Ascidia brings with it a supply of solid food in the shape of plankton organisms. The exact method which is used in the collection of these organisms has been the subject of conflicting statements. Earlier writers, such as Roule (’84), described it as follows. The mucus secreted by the endostyle is spread over the inside of the branchial sac by the ciliary activity of the gill bars. This PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 241 mucus catches the food particles which come in with the water, and the mixture of food and mucus is transported across the face of the branchial sac, dorsally and posteriorly into the oesophagus. In recent textbooks (Herdman, ’99) the process is described in a totally different way, somewhat like the fol- lowing. The mucus from the endostyle passes anteriorly to the peripharyngeal grooves. Here the food particles are caught at their very entrance to the branchial sac, and carried by the mucus on its way along the dorsal lamina to the oesophagus. Delage et Herouard (’98, p. 144) point out the differences in these descriptions, but cautiously avoid anything but a general- ized account of feeding. The matter has been recently investigated on many transparent ascidians by Orton (713), who has proved very clearly that the earlier accounts are correct. I have examined the process of food collection in Ascidia and in the transparent Ecteinascidia tur- binata, and my observations are in complete agreement with those of Orton. Occasionally specimens of A. atra are found which are quite translucent. By feeding carmine to such animals, it is possible to see the red band of mucus-entangled material swept along the branchial sac, upward and backward into the oesophagus. The mechanism by means of which the transportation is accomplished deserves a closer scrutiny. At regular intervais along the junctions of the transverse and longitudinal vessels of the branchial sac, there are present small papillae which pro- ject into the cavity of the sac. A papilla is really the wall of a blood sinus, and in A. atra its ventral wall is composed of a ciliated epithelium (fig. 4, A). At its junction with the inter- secting vessels, I have always found a flat semicircle of what seems to be smooth muscle cells (fig. 4, B). The location of the ciliary surface of the papilla and the muscle at its base are intimately concerned with the collection and movement of the food. By removing a part of the test and branchial sac, it is easy to observe with a binocular microscope the function of the papillae. Food particles which are filtered by the meshes of THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 1 242 SELIG HECHT the branchial sac are rapidly lashed to the tips of the papillae by their ventral cilia. Here they are caught by the mucus, and incorporated into the thread of food which is passing across the branchial sac. This cord of mucus and food is transported by the papillae. Waves of contraction bring two rows of papillae together, and by the action of the cilia the food cord is passed from one row to the next, until it reaches the oesophagus. Fig. 4 Papilla of the branchial sac. A, median section; B, section at the base of the papilla. The mechanism for these papillary movements is probably local, because touching a papilla with a glass rod causes a con- traction to appear. This would indicate that the waves are the result of a series of stimulations of the papillae by the contact of the food mass. The food as it enters the oesophagus is in the shape of a cord, and in this manner it is passed along the digestive tube. With the food also goes the mucus. Although the food is digested and absorbed, the mucus is probably not affected at all. When PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 243 the feces come out in neat, flat, oblong packets, they are incased in a thin layer of gelatinous material, which is probably the mucus. The presence of the gelatinous covering of the feces is best seen in animals which have been in the laboratory for some days. In such cases the feces contain but little excrement, and are composed mainly of a transparent mass of mucus. The seawater in the laboratory is coarsely filtered, and con- tains very few organisms. In this way the food supply of Ascidia is cut off; consequently, it does not live long in confinement. This shows that the dissolved organic content of seawater, to which Piitter (’07) has attributed such great importance, is of little significance in maintaining the metabolic balance of Ascidia. The species has developed an elaborate mechanism for capturing the organisms in seawater, and without them it slowly starves to death. IV. THE MOVEMENTS OF ASCIDIA Ascidia atra, though permanently attached to the rock, is capable of moving not only certain of its structures, but also of bending and contracting its body in relation to its base of attach- ment. The siphon rims can close and open, the body can con- tract along the dorso-ventral axis, and the entire animal can bend with a surprising degree of vigor. When arranged in certain combinations and sequence, these activities form the reflexes with which the animal responds to stimulation. From such an aspect they will be considered in the description of the sensory reactions of the species. At present, however, my object is to present the physiology of these movements in themselves by examining the factors which are concerned in their production. 1. Stphon rim closure Ascidia is usually described as possessing eight lobes on the rim of the oral siphon and six on the atrial. These are shown in figure 1. All such photographs and descriptions are of dead animals and tell only a partial truth. In the normal, living animal under water, these lobes are not shrunken and collapsed, 244 SELIG HECHT but stand out expanded on the siphon in the form of thin lappets (fig. 5). Under special circumstances it is possible to secure a local contraction of the region near an individual lappet. Ordinarily, however, the entire siphon rim shuts as a unit. This closure is Fig. 5 Sketch of living, expanded Ascidia, to show the cheeks on the right side and the protruding lappets on the siphon rims. conditioned by the presence of well-defined ridges and folds in the test, along which the contraction takes place. An end-on view of a nearly closed oral siphon (fig. 6) shows that the alternation of folds and ridges depends on a surprisingly accurate pattern, which involves thick and thin portions of the supporting test. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 245 The closing of the siphon rim, however, is more than a mere puckering together of its edge due to the action of circular muscle fibers. The rim is not only pulled together, but is also drawn down toward the body of the siphon. This is due to the action of longitudinal muscle fibers which, in the.siphon, lie nearer the cavity of the siphon than do the circular muscles. The siphon rim in Ascidia is so opaque that it was impossible actually to observe the action of the two sets of muscles. A transparent species Ecteinascidia turbinata, furnished the desired opportunity. Individuals two or three days old, measuring three to four millimeters in length, can readily be examined Fig. 6 End-on view of a partially closed oral siphon, showing the geometric arrangement of the folds in the test. The left side of the body is uppermost. with the low power of the microscope. These animals show the two factors of siphon closure beautifully. At first the circular (sphincter) muscles contract and partially close the rim. This is followed by a contraction of the longitudinal fibers, which results in a drawing in of the rim, thereby completing the closure. In these young individuals I have frequently observed the longitudinal muscles of the oral siphon contract so vigorously that the upper portion of the siphon was completely inverted and tucked into the branchial cavity. In Ascidia this retraction is provided for by a sudden decrease in the thickness of the test near the rim (fig. 10). The combined action of the two sets of muscles results in a closure which is really complete. No trace of a water current can be demonstrated after the siphons have been shut. 246 SELIG HECHT Fig. 7 Apparatus to record the movements of Ascidia. In order to analyze the movement of the siphon rims, I secured kymographic records of their activity. The apparatus which was employed is represented in all essentials in figure 7. The drawing needs little explanation. The long, light, aluminum lever was nearly, but not quite, balanced by the weight on the short arm. ‘This slight excess on the long arm kept the pendant vertical rod in continual contact with the right edge of the siphon rim. Such a procedure proved more effective and less disturb- ing to the animal than actually attaching the lever to the tissue. In addition the vertical arm was curved so as to fit into the cavity of the siphon. To keep the animal in place, it was fixed to a piece of plate glass, which was heavy enough so that even a vigorous movement of the entire animal did not change its position. Fig. 8 Record of siphon rim contraction, and record of body bending. The base line marks one minute intervals. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 247 The smaller contraction in figure 8 represents a siphon rim movement; all the other records are almost identical with it. Curve 7 in figure 11 is another example, and represents in addi- tion an analysis of the movement. The record is divisible into four distinct phases. The first two are the phases of contraction; the second two those of recovery. The first phase is an almost straight line, and represents, therefore, a comparatively rapid contraction of the siphon rim. Although of short duration, lasting about a second, this por- tion of the movement accomplishes nearly three-fourths of the total contraction. The remaining closure is made at a much slower rate, as is shown by the amplitude and duration of the second phase. The condition of maximum contraction is reached in approximately four seconds. Recovery begins almost im- mediately, and in the beginning is comparatively rapid. The third phase of the record may be defined as that portion which lies between the position of maximum contraction and the point where the curvature of the line changes so as to be convex to the time axis. It lasts nearly twice as long as the second phase, and includes about one-half the recovery of the siphon rim. The last phase of the movement is the longest, occupying nearly three-fourths of a minute. At the end of it the siphon rim has assumed the diameter which it had at the beginning of the contraction. An analysis of the time relations of the phases of two separate records of the same siphon rim is given in the accompanying table (table 2). The two movements were made under the same conditions within a few minutes of each other, and were produced by the same intensity of mechanical stimulus. The similarity in the resulting records is very evident. TABLE 2 Siphon rim closure. Exp. VI.23.1 DURATION OF PHASES, SECONDS 1 2 3 4 I ee} 3.4 5.3 36.8 II 1.5 3.4 9.2 | 40.5 248 SELIG HECHT It will be noticed that the general shape of the curves produced by the closing and opening of the siphon rims resembles that of the contraction and recovery of smooth muscle (Winkler, ’98). The effective agent in the closure is, indeed, the sphincter of smooth-muscle cells in the siphon working against the elasticity of the tissues and the test. Although the presence of the test undoubtedly helps in the opening of the rim, the recovery from the contracted condition can occur without the test. Animals Fig. 9 Right side of animal with the test removed. h., heart; /.m., longitudinal muscles; ¢.m., transverse muscles. from which the test had been removed were still capable of closing and opening the siphons. It seems reasonable to suppose that the elasticity of the tissues which is responsible for this recovery is due to the presence of spaces filled with blood under pressure. 2. Body movemenis The bending of the body on its long axis occurs so that the right side of the animal always forms the concave surface of the bend. This is accomplished by the contraction of longitudinal muscle strands which lie on the body wall of the right side only PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 249 (fig. 9). In the living animal this side of the body adheres firmly to the test. The left side, however, which includes the renal organ, parts of the intestine, etc., is free from the test, the two being connected only by a single blood vessel. Consequently any contraction of the long muscle strands on the right side will result in a bending of the body and test in that direction. This is facilitated by the fact that the right side of the test is much thinner than the left (fig. 10). The bending of the body on its long axis is always associated with a movement of the siphon rims. This becomes clear when Fig. 10 Longitudinal section of test, passing through oral siphon. its kymographic record is examined (fig. 8). The kink in the very first part of the curve denotes a siphon rim closure. It will be seen that the movement of the siphon precedes the vigorous activity of the body as a whole. The curve made by the bending of the body (fig. 11, curve 6) ‘resembles in all essential features the one made by the siphon rim contraction (curve 7). It may be divided into four phases, the durations of which are relatively the same as in the activity of the siphon. The first phase is short, and accomplishes the main extent of the contraction. During the longer, second, phase a slower activity brings about the maximum point in the curve. The resumption of nearly the normal position of the body is accomplished in the third phase, whereas the complete relaxation 250 SELIG HECHT L25/ Curve 6 STW DUTES Fig. 11 Analysis of records. Curve 7, siphon rim closure; curve 6, body bending. takes place slowly during the much longer period of the fourth phase. The actual time occupied by the different periods is given in the accompanying table. In table 3 there is represented an analysis of three records of the movements of one animal produced under the same conditions and intensity of stimulation. The appearance and duration of the second and third phases of the record depend in the main on the intensity of the stimulus which produces the body contraction. The more intense the stimulus, the longer do the two phases last, their greater duration showing itself in the maintenance of the condition of maximum contraction. Compare, for example, the values in table 3 (also fig. 8) with curve 6 of figure 11. Although both were TABLE 3 Body Contraction. Exp. VI.28 .1 DURATION OF PHASES, SECONDS _ bo oo cs — —_ oan cow _— ~J Co mm bo COW wm > CO _ — ie) on III PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 251 obtained from the same individual, curve 6 is the result of a stimulus twice as intense as those whose records are given in the table, and consequently, the two phases last considerably longer. Whereas the actual bending of the body involves merely the action of the longitudinal muscles, the recovery to the normal shape depends upon several factors. One of these is the activity of a set of muscles situated on the left side of the siphons. Al- though these muscles are not well developed, and extend for a short distance only, they act probably like extensor muscles, and tend to antagonize the action of the long muscle strands of the right side. A second factor is the elasticity of the soft tissue of the body. That these two agencies alone are capable of bringing an animal back to normal shape is shown by the com- plete, though slower, recovery of animals with the test entirely removed. The test, however, is a structure of considerable significance in the resumption of the normal form after the bending of the animal. Although apparently homogeneous, the cellulose ma- terial of ascidians has been shown to possess a fibrillar structure visible in polarized light (Schulze, ’63). Probably this hetero- eeneity, which is also to be seen in some stained preparations of A. atra, is of importance for the elastic properties of the test. Its peculiarities resemble those of a viscous solid. To a sudden distortion, the test will respond in a manner comparable to most elastic bodies. If, however, it is subjected to a slow distortion, it will partially accommodate itself to the new form, and never return to the original shape. In the bending of Ascidia the activity is sufficiently rapid to cause an immediate elastic recoil on the part of the test. Sec- tions of the body show that the resilience of the test is utilized | to good advantage (figs. 10 and 12). The left side is noticeably thicker than the right and, consequently, serves as an elastic back which antagonizes the muscles of the right side. Although controlled by the relaxation of these muscles, the elastic rebound of the left side probably serves in a large measure to straighten the curved body. 252 SELIG HECHT The difference in the elastic response of the test to strains of short and long duration may be the explanation of the distorted appearance of many laboratory and museum specimens. The collection and transportation of living animals involve a continued stimuiation. This results in the maintenance of the curved condition for a long time, until finally the elastic limit is passed, and the animal remains permanently abnormal in appearance. The elasticity of the test is further made use of in the third type of movement of which Ascidia is capable. This is a con- traction of the body along its dorso-ventral axis, in such a way that the right side forms the concavity. The muscles which are concerned are the transverse fibers on the right side (fig. 9). In order that they may exert their influence, the test is thinned & a c Fig. 12 Transverse sections of test. a, near tip of oral siphon; b, near base of oral siphon; c, through middle of body. out along a narrow line, in the middle of the right side, running parallel to the long axis. The contraction of the transverse muscles bends the test along this line, with the consequent forma- tion of two prominent cheeks (figs. 5 and 12c). The contraction of the body on its dorso-ventral axis extends in some cases well into the oral siphon. ‘There are circular mus- cles present in the siphon, which by their contraction can de- crease its diameter. Here also the test shows an arrangement of thick and thin portions on the right side, whereas the left side still maintains its uniform thickness in order to aid in the recovery (fig. 12b). It is consequently obvious that the test is intimately concerned in the contraction of the body on its short axis. The same is true for the siphon rim movements and especially for the bending of the body on its long axis. Definite structures in the test go PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 253 with definite sets of muscles. Therefore, besides serving as an excellent covering for the soft internal parts, the test in addi- tion functions as an exoskeleton, on which depends the proper execution of the movements of the animals (Fol. ’08). 3. Spontaneous movements In the continued observation of Ascidia under all sorts of con- ditions, it became evident that complete movements of the body and siphons often occurred when no apparent external stimulus was present. The animal is extremely sensitive to mechanical stimulation, and at first I was inclined to attribute these aberrant contractions to very sight movements of my body or of other people in the laboratory. Such an explanation was, however, abandoned when the same movements occurred under con- ditions which precluded this source of stimulation. Somewhat similar spontaneous contractions have been de- seribed for the cirri and oral hood of Amphioxus as the result of the mechanical stimulation of the cirri by the accumulation of particles of sand (Parker, ’08, p. 481). This explanation does not hold in the case of Ascidia. When animals which had been carefully washed were placed in filtered seawater, they continued to perform spontaneous movements. Moreover, animals with the test entirely removed and with the greater part of the siphon cut away, and consequently deprived of their sensory apparatus, still exhibited frequent contractions. The factors for their produc- tion, therefore, rest within the organism itself. The solution of the difficulty came when a record was kept of the appearances of the spontaneous contractions under con- ditions which excluded external stimulation. Several animals were placed in individual battery jars containing about five liters of filtered seawater. The jars rested on a heavy table placed on the concrete floor of an isolated house built directly on the rock of Agar’s Island. The animals were observed con- tinuously for an hour, and the time of each spontaneous con- traction was noted. Figure 13 gives a graphic account of two such animals. It is very evident that there is a rhythmic occurrence of the spontaneous movements. 254 SELIG HECHT By means of the apparatus which has been previously de- seribed (fig. 7), kymograph records were made of this rhyth- micity. Animals were allowed to register their activity at times of the day and night when Agar’s Island was deserted except for the presence of two people in a house more than a hundred yards from the laboratory. The curves in figures 14 and 15 show the movements of two animals which are entirely typical of all the others. There can, therefore, be no doubt of the - rhythmic character of the spontaneous contractions exhibited by Ascidia. The rhythmicity of the movements possesses a peculiarity which resemble the refractory properties of the vertebrate heart (Woodworth, ’02). It is well known that immediately after a contraction of the ventricle, it fails to respond to stimula- tion. After this refractory period, an external stimulus will cause a pulsation even before the expected rhythmic contraction is due. Similarly in Ascidia a stimulus which is so slight that it causes merely a siphon rim movement will, when applied regularly at intervals of a minute, call forth complete body movements at approximately the periods when they are expected to occur rhythmically. The following record is chosen as an example because the spontaneous contractions of this animal have already been recorded (figs. 13 and 14). Exp. VI.27.3. Animal in a nine liter battery jar. Stimulated every minute by the impact against the jar of a pendulum bob swinging from a distance of five centimeters. 12:21 Siphon rim movement 22 “ “cc “ec 93 “cc “ce ““ 94 “ce iT9 “ce 25 Complete body movement 26 Siphon rim movement 2h “ “cc “ 28 “ “ “ 29 Complete body movement 30 Siphon rim movement 31 “ce “ec “cc 32 “cc “ “ 33 Complete body movement PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 255 I-30 11-40 W-50 12.00 12.10 12:20 12:30 Fig. 13 Graphic record of occurrence of spontaneous movements. Comparison with the other records (figs. 18 and 14) shows that the complete body movements occur at intervals similar to their rhythmic appearance spontaneously. The significance of this type of experiment I believe to be as follows. After the Fig. 14 Records of rhythmic, spontaneous movements. The base line marks five minute intervals. ls EN a a Ds a OU ed a a SG la Se Be Re aE i a TA INS Fe eH 24.) C WRA/A Fig. 15. Records of rhythmic, spontaneous movements. The base line marks five minute intervals. 256 SELIG HECHT performance of a rhythmic contraction, Ascidia passes through a period during which a certain strength of stimulus fails to call forth a movement of the body. However, as this refractory period passes by and the time approaches for the culmination of the next rhythmic movement, this same sub-liminal stimulus will call forth not only the usual siphon contraction, but also the body contraction in advance of its scheduled time. The function of these rhythmic movements is by no means clear. Their accomplishment of a partial discharge of the water in the branchial sac would be intelligible if something else be- sides the water were also expelled. Ascidia, however, possesses an effective mechanism for avoiding just such a necessity. The tentacles in the oral siphon screen out all but the smallest particles which come in with the water current. Everything that passes beyond them is incorporated into the food cord in the branchial sac. The particles which are large enough to touch the tentacles, set off a reaction that drives the water and the particles out of the cavity of the siphon. Substances in suspension either get into the branchial sac and stay there, or they are forced out at once. Therefore, the function commonly attributed to rhythmic movements in other animals (Redfield, ’17) cannot apply in the case of Ascidia atra. To call the movements a respiratory rhythm would also fail to explain their existence. Ascidia has a highly efficient respira- tory mechanism which moves large quantities of seawater. The renewal of the contents of the branchial sac by the rhythmic discharges would be of no significance compared to the continuous stream of water produced by the cilia. The relative infrequency of the discharge would also argue against a respiratory rhythm. The spontaneous movements in Ascidia are not an isolated instance. I have observed them in a colonial species, Ectein- ascidia turbinata, the individuals of which are about two centi- meters long and quite light in weight. When the animals are attached, the effect of the body contraction is solely to discharge the water. If, however, an individual is removed from its attach- ment and placed in a large jar of seawater, it Jerks itself along the bottom in a manner that vividly recalls the behavior of Salpa PHYSIOLOGY OF ASCIDIA ATRA LESUEUR Zot under similar conditions. The frequency with which Salpa pulsates is several hundred times as great as those with which Ascidia and Ecteinascidia perform their spontaneous contractions. These facts as well as the apparent lack of function of the rhythmic movements have led me to suggest that perhaps the rhythm is the degenerate remains of a once vigorous activity. The ascidians are generally supposed to have originated from the free swimming appendicularians. These possess no tentacles, and most probably the earliest ascidians did not possess them either. It is, therefore, entirely intelligible that the rhythmic discharge of the water from the branchial sac of these ancestral ascidians was of considerable value as a cleansing process. More- over, the salpas are derived from the early ascidians. With their specialization for a pelagic existence, the rhythmic move- ments were developed into a mechanism for respiration, feeding and locomotion. The stem line of ascidians, however, soon developed tentacles, and the rhythmic discharge of the bran- chial sac contents, therefore, decreased in importance. The frequency with which it occurred probably also decreased. On the basis of this hypothesis, there exist at present two divergent lines of development of the spontaneous movements. One of these constitutes the salpas, whose frequency of contraction is several hundred times greater than that of the ascidians, which constitute the other line. V. SUMMARY 1. The blue-black color of Ascidia is: due to the presence of spherical pigment granules, which are the metamorphosed re- mains of the green blood cells. Before becoming imbedded in the test, the green cells turn blue, and may be found as such in the blood stream. 2. Ascidia is capable of regenerating its test. The process of regeneration and the normal sloughing of the test show that there is a continuous secretion of material on to the inner face of the test. 3. A species of amphipod lives commensally in the branchial sac of Ascidia. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 1 258 SELIG HECHT 4. Ascidia maintains a seawater current of large volume and low pressure. The volume of the water moved per unit weight decreases as the size of the animal increases. The difference in intensity and direction between the incoming and outgoing currents enables Ascidia to secure a continuous supply of fresh seawater independent of the movements of the sea. 5. Food collection is accomplished in the branchial sae with the aid of the papillae and the mucus secreted by the endostyle. The food particles are transported in a mass across the face of the branchial sac, dorsally and posteriorly into the oesophagus. 6. Ascidia cannot survive solely on the dissolved organic contents of the seawater. It must be furnished the suspended contents as well. 7. The species is capable of three kinds of movement: the siphon rims can close and open, the body can contract along the dorso-ventral axis, and the entire animal can bend toward its base of attachment. The movements are accomplished by several sets of smooth muscle, which depend for their proper action on the function of the test as an exoskeleton. 8. Ascidia performs rhythmic movements to discharge the contents of the branchial sac. No function can be ascribed to this rhythmic occurrence, and a suggestion is made that it may represent the degenerate remains of an activity homologous with the rhythmic pulsation of the salpas. VI. BIBLIOGRAPHY Cauutery, M. 1895 Contributions a l’étude des ascidies composées. Bull. sci. France et Belg., T. 27, p. 1-158. Crozier, W. J. 1916 Some indicators from animal tissues. Jour. Biol. Chem., vol. 24, p. 443-445. Cuénort, L. 1891 Etudes sur le sang et les glandes lymphatiques dans la série animale. Arch. Zool. Expér., Sér. 2, T. 9, p. 138-90. Detace, Y., ET Hrovarp, E. 1898 Les Procordés. Traite de Zoologie concréte, T. 8, 379 pp. For, A. 1908 Note sur la régénération de la tunique chez les Tuniciers. Bull. Soc. Zool. France, T. 33, p. 79-81. Hecut, 8. 1916 The water current produced by Ascidia atra Lesueur. Jour. Exp. Zodl., vol. 20, p. 429-434. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 259 Hrerpman, W. A. 1882 Report on the Tunicata collected during the voyage of H. M. 8. Challenger during the years 1873-76. Part I. Ascidiae simplices, 293 pp. 1899 Ascidia. Liverpool Mar. Biol. Committee Memoirs, I. Liver- pool, 52 pp. Hirscuuer, J. 1914 Ueber die Restitutions- und Involutionsvorginge bei operierten Exemplaren von Ciona intestinalis Flem. nebst Bemer- kungen ueber den Wert des Negativen fiir das Potenzproblem. Arch. mikr. Anat., Bd. 85, p. 205-227. Kowatevsky, A. O. 1867 Entwickelungsgeschichte der einfachen Ascidien. Mem. Acad. Sci. St. Pétersb., T. 10, 19 pp. Kuprrer, C. 1875 Tunicata. Jahresb. Comm. wiss. Untersuchung deutschen Meere in Kiel., Bde. 2 u. 3, p. 197-228. Lesueur, C. A. 1823 Descriptions of several new species of Ascidia. Jour. Acad. Nat. Sei. Phila., vol. 3, pt. 1, p. 2-8. Orton, J. H. 1913 The ciliary mechanisms on the gill, and the mode of feeding in Amphioxus, Ascidians, and Solenomya togata. Jour. Mar. Biol. Assoc., N.S8., vol. 10, p. 19-49. Parker, G. H. 1908 The sensory reactions of Amphioxus. Proc. Amer. Acad. Arts. Sci., vol. 43, p. 415-455. Pirrer, A. 1907 Die Ernihrung der Wassertiere. Zeit. allg. Physiol., Bd. 7, p. 283-320. RepFieLp, E. 8. P. 1917 The rhythmic contractions in the mantle of Lamelli- branchs. Jour. Exp. Zodél., vol. 22, p. 231-239. Rovtez, L. 1884 Récherches sur les Ascidies simples de cétes de Provence. Ann. Mus. Hist. nat. Marseille., Zool. T. 2, p. 1-270. Scuutze, F. E. 1863 Ueber die Struktur des Tunikatenmantels und sein Verhalten im polarisierten Lichte. Zeit. wiss. Zool., Bd. 12, p. 175- 188. Scorr, T. 1907 Observations on some copepoda that live-as messmates or commensals with ascidians. Trans. Edinb. Field Nat. and Micr. Soe. vol. 5, p. 357-372. Van Name, W.G. 1902 The Ascidians of the Bermuda Islands. Trans. Conn. Acad. Arts Sci., vol. 11, p. 325-412. VerRRILL, A. E. 1870 Parasites of Ascidians. Amer. Nat., vol. 3, p. 383. Winker, H. 1898 Kin Beitrag zur Physiologie der glatten Muskeln. Arch. ges. Physiol., Bd. 71, p. 357-398. Woopworti, R.S. 1902 Maximal contraction, staircase contraction, refractory period, and compensatory pause of the heart. Amer. Jour. Physiol., vol. 8, p. 2138-249. AUTHOR'S ABSTRACT OF THIS PAPER ISSUHD BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 22 THE PHYSIOLOGY OF ASCIDIA ATRA LESUEUR! II. SENSORY PHYSIOLOGY SELIG HECHT TWO FIGURES CONTENTS I. The responses of Ascidia and their nervous relations.................. 261 i Descripbion Ol TEACtiONn sn Jo5s onan sae oo Re sich e wale wast cue bee 262 Zoe NCEVOUS NElAblONSt erry ce dike eed Nhe 6 Sick fo aes Der ene 266 ireMechanicalistunulationvey |.) a actssa fa oo ac as oe ee cere tet eee nce 268 Is AMI Mee cree De ace Reet ates | Ce ne eae ean ee ee a a 268 PINUS Op AE COSD US whe OW A Bis Bes 2), ine ae a RI SRT Sten ees pe Bad 270 Des VepeA ears timation s 408 eink td S.A tes Sold Mae ERIS REEE oat ee, 273 TESST SINE eee eyes een Terao ah eee oe oe tik Sy AS eR ge Oe a 277 L/S cANEEd) eT EDUAUIR Se noo Oa eae Ail a ame Re 2 Or a A a Re De a 279 PEP HE TNAOSERSTCAVEL Yi ie «ier codes acy jothoa cee cen On Pei 279 PVA SOLO M SOUSILEVELY asics a pees ciara ited Ae Oe a ea oe 281 Wane henaeales tintul tran ys) 80925 4: lesan tale: teas. ect eet. 283 ie Eresenceromelemicall Sense). .4 4.58 250. 6orac sa anaes nde eeiadadal «a 283 Zs ANUETA RYO lao on crc ke RE aR a ee ee OR 284 So DISiIOUEOM,OL SENSIUIVIGY) .. 0.01 2 -s22 +s feel nasies,« See). dudes eee 286 Am OS OUICHPRESBUTES. 458), 32 2). 1, atid t Acdeen tants 22 cd 0h lak he ga ese 287 dep VesMlGseneMbEeheMmuCAl a: 5 2. 0.04. 2sky! oie c oso sacheleds anieheaaeek « 289 Gre NIECE GHEE SONSE: OLPANS) ! 0 a2 cos 6 oe dk k jaye wlewsysns son ode Sle einseje <2 292 MMIC SITE DIT OT 3 ced 5 3 5 ASAE SORE a RIE ET ee naa ee ee ne 296 Vil Bibliopraplivermnmreae tye. cic tke ee. eC Sek ay at 298 I. THE RESPONSES OF ASCIDIA AND THEIR NERVOUS RELATIONS The assumption of a sessile mode of life involves a sacrifice in the number and kinds of responses of which an animal is capable. The comparatively few reactions exhibited by the sessile tunicates are undoubtedly accountable for the almost complete absence of our knowledge of their sensory physiology. The common European ascidian, Ciona intestinalis, is the only one in which anything is known of the behavior under stimulation. ‘Contributions from the Bermuda Biological Station for Research, No. 79, and contributions from the Zodlogical Laboratory of the Museum of Compara- tive Zodlogy at Harvard College, No. 304. 261 262 SELIG HECHT Even here, however, the data are meager and scattered, and con- sist largely of incidental observations. This lack of knowledge and the abundant presence of Ascidia atra at Bermuda served as incentives for the following investigation of the sensory re- actions of this species. 1. Description of reactions Jordan (’07) on the basis of his observations has called the ascidian Ciona, an animal poorinreflexes (reflexarmes Tier). With- out subscribing to any of his theoretical generalizations, which Baglioni (713) has justly criticized, I have no hesitation in simi- larly describing the behavior of Ascidia atra. Tests with a variety of conditions of stimulation have revealed very definite activities, by means of which the animals respond to changes in the environment. The number of these activities, however, is small. The structures and movements which are involved in their execution have already been described (Hecht, 717). It remains to explain their relation to one another and to the source of stimulation. The presence of the open siphons and of the water current makes it possible for Ascidia to receive indica'tions of changes of the environment not only on its exterior, but also on its interior surfaces. This distinction is of fundamental im- portance, because the place of reception of the stimulus deter- mines the kind of movement which the animal executes. As a result there are manifested two qualitatively distinct groups of reactions. Each of the groups consists of three responses, which involve the use of different combinations of muscles. The group of direct responses depends for its origin on a source of stimulation which affects the external surface of the animal. This group of responses is concerned mainly with mechanical stimuli. Although the three reactions included under this head result from different intensities of the same outside disturbance, the reactions themselves involve an activity of different muscles, and not a different degree of activity of the same effectors. 1) If the test of Ascidia be touched very lightly, the siphon nearer the point of stimulation will contract. The extent of the PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 263 resulting closure depends on the intensity of the stimulus and on its distance from the siphon. After a short interval the siphon rim opens and the animal is normal again. 2) If, however, the stimulus has been stronger, not only does the siphon rim nearer the stimulated area close, but the other siphon rim also closes. A new set of muscles has been called into play. This reaction is to be differentiated from the one in which both siphons are stimulated, such as when a drop of water is allowed to fall on the surface of the water in the aquarium. In this case each siphon is independently stimulated by the same disturbance. Such a reaction persists when all nervous connections between the two siphons have been cut (cf. Loeb, ’92 and Magnus, ’02). It is otherwise with the response which I have described. Nor- mally, if one siphon is touched so carefully that the animal is not jarred, both siphons will close provided the proper intensity of stimulus is used. When, however, the nervous connections between the two siphons is severed, only the stimulated siphon rim contracts. 3) In response to an ordinarily vigorous mechanical stimulus, A. atra reacts by the employment of still an additional set of effectors, the longitudinal muscles of the body. Not only do both siphons close, but the body bends on its long axis toward the right side. This bending toward a structurely determined side is of signifi- cance in the ecology of Ascidia. Most individuals of the species are attached with the body projecting at any angle, but mainly in a nearly horizontal plane. All such animals which I examined, were found with the left side of the body uppermost. Conse- quently, the curving toward the right side results in bringing the siphons into such a position, that a disturbing body on the outside will roll off, and one on the inside will fall out. In the previous work on ascidians the reaction which involves the bending of the body has been the only one which has received any adequate attention. It has been generally regarded as the only reflex of which this group of animals is capable, and there- 264 SELIG HECHT e fore called ‘the reflex’ (Loeb, ’02). Jordan (’07) has more appro- priately called it the protective reflex (Schutzreflex). The reactions which are comprised in this direct group show individual variations depending upon the intensity of the stimulus which sets them off. They can, however, be very definitely separated from one another when the animal is observed with any degree of care, or when graphic methods are employed. (See for example, figure 8 of the first paper of this series: Hecht, ’17.) All these reactions are to be kept apart from those which are in the group of crossed responses. The stimuli which result in the reactions of this second type are all localized on the interior surfaces of the siphons, of the atrial cavity, and of the branchial sac. They include changes in the environment not only of a mechanical nature, but of a thermal, photic, and chemical kind as well. Although they may be produced by the same kind of stimulus varying in intensity, the three reactions included in this group are, nevertheless, the result of different combinations of effectors. As in the case of the direct reflexes, they must, therefore, be sharply distinguished one from another. To make this clearer, I may refer to the behavior of a human being suddenly exposed to a bright light. The person will reflexly close his eyes. If, how- ever, the light be made excessively bright, he will not only close his eyes but also place his hand over them. The nature of the stimulus is the same, but the greater intensity of the stimulus brings forward a new activity superimposed upon the simple eye closure. The same is true of the following reactions of Ascidia. 1) An exceedingly delicate stimulus on the inside of one siphon results in a closure of the other siphon. The stimulated siphon remains wide open, while the sphincter of the other siphon is called into play. This kind of response can be secured only under very carefully controlled conditions. The animal must. not be jarred and the stimulus must be a delicate one. It is best to use large animals because they are not as sensitive as the smaller ones. 2) An increase in the intensity of the stimulus produces a re- action which does more than merely stop the water current; in PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 265 addition, it brings about a discharge of the water present in the branchial sac. The stimulated siphon remains open, the other siphon closes tightly, and the animal contracts vigorously along its dorso-ventral axis, resulting in a sudden decrease in the ca- pacity of the respiratory chamber. Occasionally the stimulated siphon may also contract partially so as to decrease the size of its opening. This gives the ejected water a greater momentum. 3) The last reaction of this group combines a bending of the body on its long axis with the movements of the previous response. This is the usual reaction which A. atra gives under ordinary conditions of stimulation of its internal surfaces. The last two reactions probably correspond to what Jordan (07) has described in Ciona as the ‘Ejektionsreflex’: “Closure of one siphon, rapid contraction of all muscles, other siphon (most frequently, but not always, the anal siphon) remaining open” (’07, p. 98). This description is repeated: by Polimanti (11), who, however, added nothing to it. Jordan did not study this reflex at all, but contented himself with the statement that it serves to throw out foreign bodies, and that the causes for its appearance are not clear. In Ascidia there is no doubt about the nature of the stimulus which will produce any of these three crossed reactions. It is always a disturbance on the interidr surfaces of the body. I have observed the same ‘Hjektionsreflex’ in the common Ecteinascidia turbinata of Bermuda under the same conditions of stimulation as in Ascidia atra. Jordan’s statement of its function is correct; it must, however, be broadened to include not only the ejection of foreign particles, but also the response to any internal irrita- tion, such as strong light or chemicals. The point of special significance is the crossed behavior of the siphon rims. Stimulation of the outside of a siphon causes that siphon rim to close. Stimulation of the inside of a siphon results in that siphon remaining open while the other siphon rim con- tracts. This points to the presence of a complexity of innervation in ascidians of which there has previously been no suspicion. The one factor which the six reactions of Ascidia possess in common is their negative character. A source of stimulation 266 SELIG HECHT is either excluded by the closing of the entrances to the body, or it is thrown out by a discharge of water. I have never observed any positive response to a stimulus in this species. This is not unexpected from its mode of existence. The animals are entirely dependent for their supply of energy on what is brought in by the water current, and they merely exercise a choice by rejecting anything which acts as a stimulus. 2. Nervous relations Among higher animals, the tunicates are peculiar in the concentration of the entire central nervous system into a single inter-siphonal ganglion. In Ascidia atra, according to Hilton ('13),? this is a roughly cylindrical mass, on one side of which is to be found a rather unusual neural gland (Metealf, ’00). It gives off many more nerve trunks than are usually described for this genus of ascidians. From the oral end there arise three large nerves, which go to the region of the oral siphon. Several nerves leave the atrial end, while from the middle of the ganglion there emerge three large nerves, four smaller ones, and many minute ones. It is significant that all the nerves contain both afferent and efferent fibers (Hilton, 713, p. 116). Practically nothing is known of the nerve endings in ascidians. The same may bé said of the presence of sense cells. Hilton describes the fibers of the oral nerves as ending in the oral tenta- cles, but fails to state whether they form free nerve terminations or arise from sense cells. Lorleberg (’07), after prolonged in- vestigation of the nervous system of Styelopsis, concludes that there is a complete lack of sense cells, but that there are un- doubted free nerve terminations present. In relation to the reactions of ascidians, one point Is clear: the only demonstrated means of direct nervous communication between the siphons is by way of the ganglion. The ganglion, however, has more than the mere conducting function supposed 2 This author refers to the species as Tunica nigra. I have it from Professor Mark that Hilton’s work was done on Ascidia atra. Moreover, his description of the species as the ‘‘ascidian very abundant on Agar’s Island’’ leaves no doubt as to its identity. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 267 by Loeb (92). Although some of the results of Fréhlich (’03) on the removal of the ganglion of Ciona have been questioned by later authors (Jordan ’07, and Kinoshita 710), the combined work of all the investigators on Ciona proves that ganglion re- moval affects at least the threshold sensitivity, the tonus, and the rate of recovery after stimulation. Ascidia does not remain normal in the laboratory long enough to permit of a study of the quantitative effects of ganglion re- moval. I had, therefore, to content myself with a determination of the qualitative results produced by the mere nervous isolation of the two siphon regions from each other. This was accom- plished by means of a rapid incision into the test and mantle so directed as to result in the severing of the nervous mass into two parts. The animal recovered from this slight operation in a few minutes. The behavior of individuals under such nervous conditions was very instructive. Of the group of direct reactions, the first persisted, and seemed, qualitatively at least, to be normal. The second reaction, that is, the closure of both siphons, disappeared at once. As long as the whole animal was not jarred, no amount of contraction of one siphon called forth a similar response of the other siphon. The reaction involving the body flexure depends mainly on the bending of the oral siphon. Therefore when this siphon was stimulated the bending occurred, but the atrial siphon still remained unaffected. The essential element of the group of crossed responses is the closure of the siphon which is not stimulated. This element . completely disappears after the operation. Stimulation of the inside of the oral siphon, frequently even when strong enough to involve the dorso-ventral contraction and the body bending, fails to affect the atrial siphon, and only causes a partial contraction of the oral one. Irritation of the inside of the atrial siphon brings about no change at all in the oral. These experiments leave no doubt of the ability of each portion of the animal to perform its part of a reaction even though it is isolated nervously from the rest. The reaction of the animal as a whole, however, depends on its nervous system being intact. 268 SELIG. HECHT Il. MECHANICAL STIMULATION 1. Touch Ascidia atra is an animal that under normal conditions is stimulated preéminently by mechanical means. This is the only variety of stimulus which is capable of calling forth all the possible responses of the species. The selection of its food—if mere ex- clusion may be called selection—is made on the basis of size, and rejection depends on the mechanical stimulation by the larger particles. The remarkable sensitivity to touch was known to even the oldest zodlogists who concerned themselves with the study of the large monascidians. Its very delicacy in Ascidia atra was a stumbling block to locating precisely the sensitive regions. | The presence of a heavy cellulose test would suggest an in- sensitivity of the exterior to any stimulation. Yet, even a gentle touch on the surface of the body results in a reaction of the direct type. Careful experimentation has convinced me that this is not due mainly to an irritability of the test to mechanical stimula- tion. An individual normally attached to a rock, and removed to the laboratory with its attachment intact, serves best for this type of experimentation. Moreover, if the substrate be securely clamped in the aquarium, the accidental jarring of the animal may be almost completely eliminated. Under these conditions a gentle touch with a glass rod on the test surface leaves the animal undisturbed. A coarser application at once stimulates. I am not prepared to deny the presence of touch receptors on the surface of the test. But I am convinced that most of the results of mechanical stimulation of the test are not due to sense organs within it, but to the passage of the stimulus through the elastic material to the more sensitive region of the siphon rim. In favor of this view is the lack of any demonstrable nerve con- nection between the test and the tissue underneath it. More- over, sources of stimulation, such as light, heat, and chemicals, which cannot easily be transmitted along the test substance, fail to be effective when applied to the outside of the test; whereas in all other regions, they are just as effective as touch. PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 269 The reaction to mechanical stimulation of the test is not due to an irritation of the underlying mantle tissue. Individuals whose tests have been removed from a portion of the body show that the mantle is insensitive to touch. It is interesting to ex- plore the sensitivity of such an animal. Even vigorous poking of the mantle (the animal must be rigidly clamped, of course) is followed by no effect. One may approach to within one millimeter of the cut test and produce no stimulation. But once the test is touched, the animal immediately gives its char- acteristic respopse. An animal wholly denuded of its test is insensitive to touch on the outside except near the rims of the siphons. In the normal animal as one approaches the region of the siphons the sensitivity to mechanical stimulation rises rapidly, and at the rim of the siphons the irritability is very great. The rim of the oral siphon is usually divided into eight lobes, and the atrial into six lobes. These thin lobes are the most sensitive portions of the outside of the body. By using large animals that have been a few days in the laboratory, and stimulating the individual lobes with a fine glass rod, I have secured local contractions of the portion of the rim contiguous to the stimu- lated lobe. The folds between the lobes are only slightly less sensitive than the lobes themselves. On theinside of the siphons below the lobes, a similar degree of sensitivity exists.» Inside the atrial siphon the irritability is greatest near the rim, but the entire atrial cavity is also sensitive to touch. Within the oral siphon the surface is extremely sen- sitive, and remains so as far down as the ring of oral tentacles. Beyond this the sensitivity falls off rapidly. Of the surfaces which produce the group of crossed reactions, the tentacles are probably the most sensitive. The prettiest automatic response of Ascidia results from their stimulation. By illuminating the inside of the oral siphon it is possible to touch a single tentacle with a fine glass rod. If a delicate stimulus be applied carefully, it is most interesting to see the atrial rim close quietly while the oral siphon remains undisturbed. If the stim- ulus is more intense the ‘Ejektionsreflex’ is produced. When a 270 SELIG HECHT small particle of sand is dropped carefully upon the tentacles, the slight back pressure produced by the closing of the atrial rim at once squirts the particle out of the oral siphon. In view of the certainty and ease with which these reactions may be demonstrated, not only in A. atra, but also in Ecteinas- cidia turbinata and in another unidentified species, it is difficult to understand why some authors have reported that the tenta- cles are practically insensitive to mechanical stimulation. Thus Roule (’84, p. 37), who studied Phallusia, and Lacaze-Duthiers et Delage (99), who observed Cynthia, state that no noteworthy reaction occurs when the tentacles are touched in this way. This is all the more strange because it is precisely here that See- higer (1893-11, p. 323) has found most of the bristle cells to which he rather doubtfully ascribed the réle of touch receptors. The perception of mechanical irritation by the internal surface of the atrial siphon is of significance in the daily routine of the species. A decidedly sensitive area is at the bottom of the atrial cavity near the anus. The feces are discharged into this cavity. Here they furnish the mechanical stimulus for a reflex of the crossed type: the oral siphon closes and the body contracts, squirting the water and the feces out through the atrial siphon. To one unacquainted with the presence of the group of crossed reflexes, the defecation of Ascidia seems almost a conscious pro- cedure. It ‘tries’ to force out the feces, and if a piece becomes caught in the siphon rim or in the atrial cavity, it ‘tries’ again to dislodge it by means of the ejection reflex, until finally it succeeds. The whole process can, however, be called forth by placing a glass bead ora pebble in the atrial cavity, or by repeat- edly stimulating it with a glass rod. 2. Vibration The extreme sensitivity of Ascidia to mechanical stimulation is manifested in its ability to respond to vibrations (compare Marage, ’05). Ascidia lives in shallow water, and if the rocks within two or three meters of an individual are stamped upon with even a modicum of vigor, it closes its siphons. NH.>Na This parallels the stimulating strengths of these cations found by Cole (10, p. 607) for the common chemical sense in the frog. In order to determine the effects of a group of anions, the following salts were used: KCl, KBr, KNO;, KI, CH;COOK, and KSCN. The first experiments were made on small, and consequently very sensitive, animals. By this means large differences in stimulating power became evident; this is typified by Exp. VIII.3 of which the following table is a summary (table 6). Later, in ordef to separate KCl, CH;COOK and KSCN, larger, and therefore less sensitive, animals were used. Exp. VIII.4 was of this type and gave the results shown in table 7. THE JOURNAL OF EXPERIMENTAL ZOOLOGY. VOL. 25. No. 1 290 SELIG HECHT TABLE 6 Liminal concentrations of a series of potassium salts SALT CONCENTRATION KCl : 0.075 N KBr 0.050 N KI 0.010 N KNO; 0.15 N CH;COOK 0.075 N KSCN 0.075 N TABLE 7 SALT CONCENTRATION KCl 0.20 N CH;,COOK 0.15 N KSCN 0.10 N A combination of the two tables gives an anion series of stimulating power as follows: Br SCN =CEACOO > ClS=NO; Excepting SCN and NOs, which are not in the usual positions, this order agrees with the familiar Hofmeister series (Hober, 714, p. 309). An absolutely complete agreement is hardly to be expected, because my tests were made in seawater. Hd6ber (’14, p. 323) has constructed ‘Uebergangsreihen,’ in which he has been able to change the position of some members of this lyotropic series by altering the milieu in which the experiments were per- formed. Analogous to this is Cole’s (10) observation for the stimulation of the frog foot, in which the positions of NH, and K were reversed by an increase in the concentration of the solutions. Acids. , Seawater to which acid is added, gradually returns to its normal hydrogen-ion concentration. Therefore, the solu- tions to be tested were freshly made up immediately before being applied to the animal. This was accomplished by having a stock 0.1 N solution made up in rain water, and diluting it to the desired concentrations with seawater. The effect of the dilution of the seawater is insignificant. Three acids, hydrochloric, formic and acetic, were tested. The following table gives the values which were obtained in Exp. VIII.9, typical of the others (table 8) - PHYSIOLOGY OF ASCIDJA ATRA LESUEUR 291 TABLE 8 Liminal strengths of acids for the stimulation of Ascidia ACID CONCENTRATION HCI 9.0016 N HCOOH 0.0018 N CH,COOH 0.010 N The order of the stimulating efficiency of the acids is, therefore, HCl > formic > acetic Bases. As representatives of this group of substances, I used NH.OH and NaOH. Exp. VIII.9.1 gave the liminal values shown in table 9. This places them in the order, NaOH >NH,OH TABLE 9 BASE CONCENTRATION NaOH 0.010 N NH,OH 0.015 N Sugars. Both glycerin and sucrose did not stimulate until they reached a concentration of 1 M. This quantity of solute, plus the salts of the seawater in which these substances were dissolved, brought the concentration of the stimulating solution to just that equivalent of concentrated seawater which irritated Ascidia osmotically. We can, therefore, conclude that Ascidia is not sensitive to these two substances. This has been found to be generally true for aquatic animals (Parker, 12). Crozier (15a), however, has shown that glycerin and maltose can stimulate Holothuria. Alkaloids. The sulphates of quinine, strychnine and morphine were tested. The order of their effectiveness, strychnine > quinine >morphine was found in Exp. VIII.1, the results of which are given in table 10. These values show a surprising sensitivity of the species to alkaloids. A bitter taste in man may be secured from 0.00004 M quinine sulphate. This amounts to one-tenth of the uncorrected concentration to which Ascidia reacts. 292 SELIG HECHT TABLE 10 ALKALOID CONCENTRATION Strychnine 0.00005 M Quinine 0.0004 M Morphine 0.001 M Anesthetics. Ether, chloral hydrated, ethyl aleohol and amy] aleohol all caused reactions which were very pronounced. The order of their effectiveness, taken from the values obtained in Exp. VIUI.1.1 and VIII.2.2 and given in the accompanying table (table 11), is amyl ale. >chloral, ether >ethyl] ale. TABLE 11 ANESTHETIC CONCENTRATION - Ether 0.02 M Chloral hydrated 0.02 M Ethyl alcohol 0.75 M Amyl alcohol | 0.001 M 6. Nature of the sense organs The morphological nature of the chemical, and indeed of any other kind of receptors in ascidians, is practically unknown. Seeliger (93-11, p. 323) has described, rather doubtfully and with much reserve, the presence of bristle cells on the tentacles of Ciona. Lorleberg (07), however, failed to secure any trace of such structures in Styelopsis; although he found many regions richly supplied with nerve endings. It may then be that the organs of chemical sense in Ascidia are similar to those which underlie the common chemical sense of vertebrates (Parker, ’12). The physiology of the receptors would seem to favor such an assumption. The problem of the physiological nature of the chemical sense organs is simplified in Ascidia atra by the mo- notony of response to all classes of substances. This negative reaction of the crossed type and its variation with the intensity of the stimulus have already been made clear. We are, there- fore, dealing apparently with an automatic reflex, of which the receptor and effector mechanisms are all set, and the conduction PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 293 provided for. The application of the stimulating substance to the sense organ merely starts a prearranged process of response. In order to understand the nature of the process set up in the receptor, it will be necessary to consider more closely the phys- iological effects of the substances used in the stimulation of Ascidia. It was on the basis of the action of the salts of the alkali metals that Héber (’14) first pointed out the relation between irritability and colloidal constitution of the plasma membrane. Since then the ubiquity of the cation and anion series has been demonstrated for such diverse processes as melanophore con- traction (Spaeth, ’13), hemolysis (Héber, *14) and rhythmic pulsation (Crozier, 16a). The presence of these ionic series in the sensory stimulation of Ascidia indicates that the significant process which underlies it, resembles, if it is not identical with, the determining reactions of the other physiological phenomena. The acids have already received attention in regard to their sensory effects (Richards, ’00; Kahlbaum, ’00). The anomalies which are exhibited by the acid taste in man are typified in the behavior of the three acids which were used in these experiments. Although HCl is more effective than formic acid, the difference — between them is not great. They both, however, are much more powerful than acetic. In the penetration of cells by acids (Crozier, ’16b), we find the same order of effectiveness. The anomalies which were referred to are as follows. When the dis- sociation constants of the acids are taken into account, it is found that the same effect is produced by acetic acid with a lesser quantity of hydrogen ions than by formic acid; and less in turn by formic than by hydrochloric acid. In Ascidia the liminal concentrations of the acids contain the following quantities of hydrogen ions: acetic, 4.1x10-‘N; formic, 6.0X10-!N; and hydrochloric, 1.6 10-N. An analogous difficulty exists in the effects of NaOH and NH,OH. Experiments on penetration have shown that NH,OH enters tissue rapidly, whereas NaOH may hardly be said to pene- trate living tissue at all. Still, NaOH is more toxic than NH,OH (Harvey, 713). Similarly it is a more effective sensory stimulant than NH,OH. 294. SELIG HECHT The physiological inertness of the sugars is known only too well to require more than mention. Their ineffectiveness has made their use possible in experiments where the effect of osmotic pressure only is desired (Hoéber, ’14, p. 496). It is therefore altogether in keeping with the parallelism between genera! phys- iological activity and sensory stimulation that Ascidia fails to be stimulated by even high concentrations of glycerin and sucrose. It has been suggested that the sensory inactivity of the sugars may be due to the lack of these substances in an aquatic environ- ment (Parker, ’12). The improbability of the occurrence of saccharin in the seawater, however, does not prevent its chemical. stimulation of Ascidia. The liminal concentration of a commer- cial preparation was 0.025 M, to which the usual negative re- sponse was given. It is necessary, similarly, to look in a different direction for the explanation of the sensitivity of Ascidia to alkaloids and anes- thetics. The minute quantities of alkaloids which are effective in stimulation find their counterpart in the extremely low con- centrations in which they penetrate cells (Overton, ’97). As a consequence of these results there can be no doubt of the essential similarity between the general physioiogical reactions of chemical substances and their effects on the sensory processes in Ascidia. This indicates that the action of the stimulating agent on the sense organ involves an effect of the same nature as the action of these substances on other cells and tissues. Moreover, it shows that the effect of a chemical on the receptor concerns that structure primarily as a cell, and only secondarily as an organ for receiving stimuli. It must be emphasized that these generalizations are not intended for the sense of taste in vertebrates, but solely for the sensitivity of animals, like Ascidia and Holothuria, which possess a general chemical sense. This type of irritability corresponds in many ways to the common chemical sense of vertebrates (Parcer, 712), although the two need not necessarily be homolo- gous. ‘Lhe problems involved in the higher organs of taste, par- ticularly the sweet taste, do not concern us here. They represent specialization for certain needs; and in the present condition PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 295 of our knowledge it is futile to attempt an explanation of their physiology. It has been tacitly assumed that chemical sense organs are capable of detecting substances in concentrations which fail to affect the ordinary cells of the body. This is largely because the effects on the sense organs become evident through certain effectors, whereas the action on other tissues must be noted by special, indirect means on the cells themselves. When, however other tissues are studied, it is seen that they are influenced by concentrations of the same magnitude as sense cells. The effect of minute changes.of the hydrogen and hydroxyl ions on the permeability of eggs and blood corpuscles need only be men- tioned. Acids and bases enter cells in concentrations like those which stimulate animals. The poisoning effects of extremely low concentrations of alkaloids are also familiar. The modifications produced by these various substances are more or less the same for all cells and tissues: witness the simi- arity of effects produced on egg cells, sperm cells, fronds of algae, blood corpuscles, chromatophores, hearts, medusa bells and a host of others too numerous to mention. The concepts of ionic antagonism and salt balance apply not only to these tissues, but to sensory stimulation as well (Crozier, 715 b). It is therefore clear that chemical sensitivity is merely one of a large numer of similar manifestations of the fundamental nature of cells The explanation which seems to me to account for all the phenomena o this sensory activity, in Aszidia at least, is that the factor which primarily converts a group of cells nto chemica! sense organs 1s not any special modification of their structure or sensitivity, but rather their connection, directly or indirectly, with an effector system. In this way the problem of the chemical sensé@ of such aquatic forms is linked with the general problems of the physical chemis- try of cells and tissues. Our present knowledge, in this respect, of the chemical senses is, however, extremely meager. The time is therefore not ripe-for any adequate explanation of the proces: in the receptor cell which results from the contact with a sa>- stance in solution. 296 SELIG HECHT One such attempt has been made. On the basis of his work on echinoderm eggs and Arenicola larvae, Lillie (11) has proposed an explanation for general irritability. It is, that sensory stimula- tion means an Increase in the permeability of the irritable element. Lillie’s explanation is based on the assumption that the de- marcation current and kindred phenomena are functions of the differential permeability of the cell membrane to certain sub- stances, notably H and OH ions. The work of Loeb and Beutner (14) has, however, shown that this bioelectric potential is due on the contrary to the presence of certain lipoid materials in the protoplasm. It is still uncertain to what extent differential solubility and the effect of interphase’ boundaries are concerned in the interpretation of these results. It ismuch to be regretted that the experiments were discontinued. There is, moreover, another and more significant objection to Lillie’s idea. All the substances which increase permeability undoubtedly do stimulate. But many substances, like Ca and the anesthetics in general, all of which have a decreasing action on permeability (Osterhout, ’16), also serve as vigorous stimu- lants to Ascidia and other aquatic organisms. The theory in its present form can therefore not be accepted as an adequate explanation. However, the attempt at an inter- pretation along the lines of permeability and similar concepts is entirely in the right direction. VI. SUMMARY 1. Ascidia possesses six distinct reactions to stimuli, all of them negative in character. They may be divided into two groups of three each: the direct refiexes, which depend upon a stimulation of the exterior of the body, and the crossed reflexes, which depend upon a stimulation of the interior of the body. 2. Theintersiphonal ganglion connects the two siphons. Sev- ering this nervous mass completely abolishes the crossed reactions, and interferes with the direct ones. Nevertheless, each portion of the animal is able to perform its part of a reaction, even though nervously isolated from the rest. : PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 297 3. Ascidia is sensitive to tactile stimulation. The regions of greatest sensitivity are the siphon rims and the oral tentacles. 4. Vibrations through solid and liquid media affect Ascidia, although transmission through the seawater is the normal method of stimulation. The receptors are located in the lobes of the siphon rims. 5. The records of the amplitude of contraction to regularly repeated mechanical stimulation show that the cessation of response after a time is due mainly to a fatigue of the receptor mechanism. 6. The ‘ocelli’ of Ascidia are not organs for photo-reception. The animals are sensitive to light of very high intensity only, and the sense organs are located within the siphon near the oral tentacles. 7. Ascidia is thermosensitive. It reacts to temperatures above 32°C. and below 20°C. 8. Its test is insensitive to light, heat and chemicals. 9. The animals react to large changes in osmotic pressure, and to the presence of the following classes of substances in solu- tion: salts, acids, bases, alkaloids and anesthetics. Solutions of sugars do not stimulate, but saccharin gives a decided reaction. 10. The liminal concentrations and the relative effectiveness of all these stimulating substances are very similar to those which have been demonstrated for other physiological activities. It is, therefore, suggested that the primary factor which converts a group of cells of Ascidia into chemical sense organs is their con- nection with an effector system. 298 SELIG HECHT _ VI. BIBLIOGRAPHY Baauiont, S. 19138 Die Grundlagen der vergleichenden Physiologie des Nerven - systems und der Sinnesorgane. Winterstein’s Handbuch vergl. Physiol., Bd. 4, p. 1-450. Coin, L. W. 1910 Reactions of frogs to chlorides of ammonium, potassium, sodium and lithium. Jour. Comp. Neur., vol. 20, p. 601-614. Crozier, W. J. 1915 a Thesensory reactions of Hologhutia surinamensis Ludw. Zool. Jahrb., Allg. Zool., Bd. 35, p. 233-297. 1915 b Ionic anipeenisn in sensory stimulation. Amer. Jour. Phys- iol., vol. 39, p. 297-302. 1916. a The rhythmic pulsation of the cloaca of Holothurians. Jour. Exp. Zoél., vol. 20, p. 297-356. 1916 b Cell penetration by acids. Jour. Biol. Chem., vol. 24, p. 255- 279. Cusuny, A. R. 1916 On the analysis of living matter through its reactions to poisons. Science, N.S., vol. 44, p. 482-488. Frouuicu, A. 1903 Beitrige zur Frage der Bedeutung des Zentralganglions bei Ciona intestinalis. Arch. ges. Physiol., Bd. 95, p. 609-615. Harvey, E. N. 1913 A criticism of the indicator method of determining cell permeability for alkalies. Amer. Jour. Physiol., vol. 31, p. 335-342. Hecut, 8. 1917 The physiology of Ascidia atra Lesueur. I. General Physiol- ogy. Jour. Exp. Zoél., vol. 25, p. 229-259. HerpMan, W. A. 1904 Ascidians and Amphioxus. Cambridge Natural His- tory, vol. 7, p. 33-138. Hinton, W.A. 1913 The central nervous system of Tunicanigra. Zool. Jahrb., Anat. Abt., Bd. 37, p. 113-130. Hoser, R. 1914 Bieialeehe Chemie der Zelle und der Gewebe. cyeabee und Berlin, xvii + 808 pp. Howetz, W.H. 1912 Text Book of Physiology. Phila., 1018 pp. Jorpan, H. 1907 Ueber reflexarme Tiere. Zeit. allg. Physiol., Bd. 7, p. 86-135. Kanipaum, L. 1900 The relation of the taste of acid salts to their degree of dissociation, II. Jour. Phys. Chem., vol. 4, p. 533-537. Krnosuita, T. 1910 Ueber den Einfluss mehrerer aufeinanderfolgender wirk- samer Reize auf den Ablauf der Reaktionsbewegungen bei Wirbellosen. I. Versuche an Tunicaten. Arch. ges. Physiol., Bd. 134, p. 501-530. 1911 Ueber den, ete. III Mitteilung. Arch. ges. Physiol., Bd. 140, p. 198-208. Lacaze-Dururers, H. pr, et Devacn, Y. 1899 Etude sur les Ascidies des cétes de France. Mém. Acad. Sci. Inst. France, T. 45, p. 1-323, 20 pl. Linturn, R. 8. 1911 The relation of stimulation and conduction in irritable tissues to changes in the permeability of the limiting membranes. Amer. Jour. Physiol., vol. 28, p. 197-222. Lors, J. 1892 Untersuchungen zur physiologischen Morphologie der Tiere. II. Organbildung und Wachsthum. Wiirzburg, 82 pp. 1902 Comparative physiology of the brain and comparative psychol- ogy. N. Y., x + 309 pp. ' PHYSIOLOGY OF ASCIDIA ATRA LESUEUR 299 Lors, J., AND BeutrNer, R. 1914 Ueber die Bedeutung der Lipoide fiir die Entstehung von Potentialunterschieden an der Oberfliche tierischer Organe. Biochem. Zeit., Bd. 59, p. 195-201. LorLeBerG, O. 1907 Untersuchungen ueber den feineren Bau des Nerven- systems der Ascidien. Zeit. wiss. Zool., Bd. 88, p. 212-248. Maaenus, R. 1902 Die Bedeutung des Ganglion bei Ciona intestinalis. Mitt. zool., Stat. Neapel, Bd. 15, p. 483-486. Maracas, M. 1905 Pourquoi certains sourd-muets entendent mieux les sons graves que les sons aigus. C. R. Acad. Sci., T. 141, p. 780-781. Mercatr, M.M. 1900 Notes on the morphology of the Tunicata. Zool. Jahrb., Anat. Abt., Bd. 13, p. 495-602. Nace, W. A. 1894a Ergebnisse vergleichend-physiologischer und anatomis- cher Untersuchungen ueber den Geruch- und Geschmacksinn und ihre Organe. Biol. Centralbl., Bd. 14, p. 543-555. 1894 b V Bealcichendipbeiolocisctie und anatomische Untersuchungen ueber den Geruch- und Geschmacksinn und ihre Organe. Bibl. Zool., Bd. 7, Heft 18, viii + 207 pp. 1896 Der Lichtsinn augenloser Tiere. Jena, 120 pp. OstrerHOUT, W. J. V. 1916 The decrease of permeability produced by anes- thetics. Bot. Gaz., vol. 61, p. 148-158. Overton, E. 1897 Ueber die osmotischen Eigenschaften ae Zelle in ihrer Bedeutung fiir die Toxicologie und Pharmakologie. Zeit. physik. Chem., Bd. 22, p. 189-209. Parker, G.H. 1908 The sensory reactions of Amphioxus. Proc. Amer. Acad. Arts and Sci., vol. 43, p. 415-455. 1912 The relation of smell, taste, and the common chemical sense in vertebrates. Jour. Acad. Nat. Sci., Phila., vol. 15, Ser. 2, p. 221-234. 1917 Actinian behavior. Jour. Exp. Zoél., vol. 22, p. 193-229. Pouimmanti, O. 1911 Beitriige zur Physiologie des Nervensystems und der Bewegung bei niederen Tieren. II. Arch. Anat. Physiol., Physiol. Abt., Suppl. 1910, p. 39-152. Ricuarps, T. W. 1900 The relation of the taste of acids to their degree of dis- sociation. Jour. Phys. Chem., vol. 4, p. 207-211. Rove, L. 1884 Recherches sur les Ascidies simples de cétes de Provence. Ann. Musée Hist. Nat. Marseille, Zool., T. 2, p. 1-270. SEELIGER, O. 1893-1911 Tunicata. Bronn’s Klassen und Ordnungen des Tier-Reichs, Bd. 3 (Suppl.), Abt. 1, p. 1-1280. Spagtu, R. A. 1913 The physiology of the chromatophores of fishes. Jour. Exp. Zodl., vol. 15, p. 527-585. el ee iA pt fine (end until vaeene , ky dirornial iwati vied Adal: lovee «cl mr of ok STUART Rt ee) ete ahr td oe siti nites leat - Laphl quwlin i) ec igash tal Aut ob OSOR GEL ¢ Mite. Thm ARR MY EP ena: felblane ] a Biiba dein dhweryces Ge inna? by ite eL"MOP AA > é eo ty) mt vel MD backs deine alae on Miler Ae! Phy on _— F i itil VC AP ey ofodiom of} 40 MoV OBL” TAT ae nae: \ oe eed eae Bn eet yu ths _ Vi iieeis© loigyigtate dodatdnde ae uh” a MELA eve we A Ae - Lay RUMOR” bares err. pais on ; = i u EF ek 7 e ; Se ee ed a ee : j nea . Reet deb Bh ik as loehg ojo medplapede ea? ck ial inind y deri fey “i : ; : > ti4 : = om 4 fi : A sbi’. \ _ 38 oe Deh Ay xD | phi i ape Har ati 1 i’ was : : icin dad ' alienate “eT y ! NY ; a 17 ry “ee wal: § > fee] oti ia ; - ; iP RAE Aa a if anh eh ough thee tore oe) 20 : fe : fi ie lool eget dl Ba coiqalenaats oils ah j sind ashi , t iv _ i es £48, thy ita 4 f MGT os eas Sal Os ent o ry t Hiveegialilld holt Ay 7) a r - : : Mf APS ee ER ou iva : Fal beth ee agar) ie ae ia peg | ut ONS iapreat “uel | iced tat aii ia oid SP pbk Lot ( tee ie. : : . Ma ~~ es i wt A ee ie ate rae es lt ASCH ‘iin =f) efuhIOme re - i ai ee. "Ct atin viudh qty Vint: pide a: be | abe oe heb bibein iabii Phila chr: wana bh) Why sf i, # bEAE tha hoe gai moh . _ i Ja Aero gh RK t aoe 3 « Y. ‘Kae Dy a widen % ‘f wane Ae a Ni-2G be. OF (aay oy nfl ivi ky ag fda iF : ‘ine ce ; 4 as AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 2 CHANGES IN THE RELATIVE WEIGHTS OF THE VARI- OUS PARTS, SYSTEMS AND ORGANS OF YOUNG ALBINO RATS UNDERFED FOR VARIOUS PERIODS CHESTER A. STEWART Institute of Anatomy, University of Minnesota, Minneapolis ONE FIGURE AND FOUR TABLES CONTENTS Iwisyrerriell ginal (neiN Oy GREE oom Gee oe cee otc an a aoe emnnoS Ener heme eee ee 302 At OLObetalwlengtin COMO: IENLtM: coe sis:- 01 cts - «certs wate wile 2) stciele ate oie axe. o- ts 318 15 (G10 ke Sead ete ee ae eee, Semen Ri Asse oe Ee A Oe oe Sea ee a Cee Se 319 Eimemiipies andetron kya tata. 7s eek see eat ee oe ee Ce ee 320 ANGE ULTTN ENb 2 hy cre tea ay cia, RP IAI Re eS YE ose PoE BSR Ss oe oe Slates ocvaee 321 SLD SLGOL OT Rae A ctl Goce ee ee EE TT Oe GT CR Tee 7, CT 322 1 RUA THT CoA ee OC ORS are IETS 6 OCDE DUR OEE oe nets Ae ee ee 326 WALSIOST GREE sows dhnah ete SR Sica Ae PED: AA ee MNO A. heen RD, 2 327 SUOMI Chae PEERS ART cis» sR ree teeta oe ee aa EA oe ee 327 BEAU dhe ced ear ef ar Cll... coed ers 60 19°0 (FF9 0-809 0)SZ9'0 (FS0° 1-68 0)¢26°0 0¢9'0 (FE9 0-22 0)609 0 sucpib 92 °ST (122° €1-F€ T1)09 21 (82 1-19 ZI) 26 ZI (88 OI-ST 6)0Z OT (EI OI-92 '6)22°6 (0% TI-ZI'6)06'6 (§&' OI-8Z 6)Z8'6 SULDL6 (0° L1-F F1)S ST 9ST (6 91-0 S1)6 ST (6 FI-0'§1)8 SI CECE G &L (6°SI-I'€1)% FI Qt (9° SI-G'S1)9'ST (G'TI-G'6)¢ 01 6°6 (801-0 01)¢ ‘01 (G II-G'6)Z OT F'6 (Z II-0 01)9 OT supnsb (0° 68-0 '€8)8°98 0°S2 (0° 62-0'€2)%°¢2 (0° 08-0'92)8° 82 012 O12 (0° 28-0'22)8°Z8 OTL (0°Z2-0'02)0° TZ (0°92-0'99)8° 69 0°¢9 (0'99-0'°29) 2° #9 (0°69-0'99)0°89 0°¢9 (0° 29-0'#9)¢°¢9 “ULUL "8p (72-479) 69 WW 489.L 9 BBP IBISTAA “BP(SI-ID)EL WW 104}00D 2 “SP GP A IAL € BYVP IBISTAA Or] Se | LORI UL “8p (6S-ZF) ZF WW SAL 9 YUP IVISTAA ‘VP(LI-SD FI IW 194}U0D € ‘BP (8Z-1Z)ZS A I89AL 8 ByeVp IBIST MA ‘ep 2 qf [O1gUOD g§ "BP (EZ-0Z) 1Z IN 982L 6 BYBP IVIST AA “Bp(8-2)E°L IN [049U0D F duoo 'IVNIids NIVUd LHSIAM AGO LAN DTHSIEAM ACOH SsoUuo HISNGAT AGO GSV GNv xXas ‘UGA WON ‘SLVY JO NOLLdINOSaa SY9OM oaIY} JO ose woIy Sdnois XIs 4sey fyQATG WOIy poyzopun O19M SANOIG XIS SI OY} UT S}ZBI SO], (wostupdwoos fo sispg ay, si abp snwhy) ay? fo aspo uf) "870.4709. hue 07 burpuodsaitoa yib6ua] fipog fo sjou sof (g], ‘uosp}pUog) sa7qn} wopsr.yy ay? WoLf DIP YZUN ‘suvb.1o ay} fo yybran fo (abun. puv) abnuaap ¢ HITaV.L 311 WEIGHTS OF UNDERFED YOUNG ALBINO RATS 9698" TPs" a (99% 0-98 '0)STF’ See’ (1S&' 0-F62 0) 228" (S9F 0-228 0)80F 183° (2820-926 0) 282" ($F 0-768 0) 20F° G9" (208° 0-962 0) 008" 60F T9G" GIG GSC" GST" 861° ($Z%' O-S8T 0)202° ail’ (EE O-E0T 0) TZT oo o i=) i=) —=) i) =) oo (929° I-FSF 1) 06S" (O81 I-¥80'T)L1T vL6— (980° 1-086 '0)9T0° a 0 be (€° 92-1 89)0%' 29 G8" 29 (Z' €8-F £9) 62° 69 (F €£-G'0L)96' TL (6° €C-8 Sh) TF’ 0G (6° 1S-@ Sh)S¢' 8h (6° 64-6 0F)Z0' 9F (2° 0S-" LF) 01 6F £9 68 66 68 19°82 GS L1G (29° SI-0% 81) 98 FI (2° ST-¥8 ST) IZ FT (Z'96-0'29)9' 18 Gg G9 0°82 (G° TOI-8'99)9 LL 1°99 (0'22-0°92)¢° 92 (¢' 96-0 6h) 1 FE I 6F (¢'9¢-0'09)8 'S¢ (F'ZS-I ¥F)9 SP aad (8'Zg-G 0¢)9' TS b CP G CF Gane, 0°08 Z 9% 9 6 (1° LI-0'FI)T 91 € FI (Z 91-0 'ST)9 ST (0' 21-0 9F1)0 6FT 0 L&T 0 281 (0' LST-0' 0ST)8 SST 0 IFT (0° T¥I-0 OFT)G OFT (0 SFI-0'SS1)¢ S&T 0 S31 (0'FZI-0 IZ1)E Zor (0 OFI-0 ZEL)€ FEI 0121 (0 IZI-0 0Z1)¢ O21 0-01 0 '9IT 0°9IT 0'8IT 0°86 0°86 (0'88-0'18)6 F8 0 $2 (0'92-0'FL)E SL ‘SP(QIE-ITE) PLE A I89L 3 BYBpP IBISTAA “Bp EC J [O1QUOD T “Bp (9TE-E2Z) 162 IW 18°L F BYCpP IBISTAA "GP (SP-OF) LP IN 104900) & "@P(SZF-LLE)Z6E A SAL F Byep BIST ‘ep (SE-EP) FF A [OUND Z "BP (SZF-FLE)ZIF IN 189.L F Byep ABST “Bp (LE-GE)9E IN 191700D Z “Sp OT§ A I89L T BYBp IBISTAA "ep LE F,[04U0D T "ep 6&1 IN AL I Byep IBVYSTM “Bp ZZ IN [019U0D T “@p(Z2-L9)69 d SAL L BYVP TBISTA “ep(SI-IDeT A 1049U0H 2 ponuruop—e ATAVL A. STEWART CHESTER 312 (Ol I-#2°0)&6'0 86° ‘0 (29° 0-09'0)9¢ 0 (00° I-69 '0)Z8'0 16° 0 (89° 0-LF'0)2¢ (68°0-82'0)€8' (960-9 0) (¢9°0-28 0) SF SP (IF 0-SE'0)8e° (FF 0-98 '0)0F’ CP (ZF 0-88 '°0)28° Supib 0 0 0 0 =) (89% '0-ZET'0)S9T 0 6910 (€9% 0-912 '0)88z'0 ($62 O-IFT 0)€8T'0 €91'0 (ZL2 0-202 '0)0€2'0 (291 O-ZET 0) FFT’ PST G61 0 oo (691 0-921 0) 6F1° SPI (SEZ 0-261 0)012 0 =) (1210-9210) IFT’ 181 (G0Z 0-821 0) 1610 SS) (E91 0-621 0) 2ZFT’ hay (SZ 0-ZL1 0) 1020 suc S=) (6FL 0-901 '0)FZ21'°0 660°0 (860° 0-800) 680° 0 (IT 0-60 '0)601'0 ¢60'0 (OLT' 0-€80'0)#60°0 (POL 0-060’ 0)¢60°0 280°0 #80'0 (ZIL 0-280 '0)260 0 €80°0 (060° 0-20 °0)T80°0 (060° 0-£90' 0)#20°0 6900 (F80 0-90 '0)€20°0 (080'0-8¢0'0)990°0 990°0 (980 0-£90'0)#20°0 suns (810° 0-000 0) T10'0 6200 (290° 0-1#0'0)8F0'0 (010° 0-1200'0)Z800'0 6600 (920 0-6F0'0)840'0 (&Z0'0-Z10'0)8T0'°0 620° 0 180°0 (0Z0' 0-0600'0)Z10'0 1800 (0F0' 0-180 '°0)20°'0 (620° 0-110°0)210°0 020°0 (€Z0 0-020 0)£20'0 (220° 0-6900°0)&10'0 020°0 (120 0-810’ 0)£Z0'0 suipsb (¥900' 0-1800'0)&F00'0 6000 (2900 0-ZE00' 0)0F00 0 (6F00' 0-100’ 0)8800°0 1£00°0 (9¢00° 0-8Z00 0) 0F00°0 (0900 0-€€00 0) ZF00°0 #8000 9F00 0 1(0G00' 0-8100 0) 1TF00 0 £2000 ($900 ' 0-200 0)ZF00 0 (€¢00° 0-L100'0)000°0 8200 °0 (6200 ° 0-9200' 0)2200°' 0 (98000-0200 0)0800°0 ; 9200 °0 (¥£00 0-9200 '0)0E00'0 swupsd 2(T91 0-881‘ 0)091 0 820°0 (860° 0-280 0)€60'0 (LST O-TST 0) 6FT 0 910°0 (160° 0-880’ 0) €60°0 (FZ1'O-O1T'O)8IT'0 620°0 2600 (221 0-901 O)8IL'0 690'0 (160° 0-80 0)280°0 (860 0-€80 0) #60°0 190°0 (290° 0-290 0)#90°0 (260° 0-180 °0)060°0 8c0'0 (020° 0-80 '0)¢90°0 s1uDsb UGAIT BONDT LUV aH BOWAHL daIousAHL STIVEGA penunu0p—¢ HIdViL 313 WEIGHTS OF UNDERFED YOUNG ALBINO RATS (99°€-F8'T)S2'Z | «(289° O-SFS'0)9T9'0 Ih CSh'0 qI'e ELF 0 (G02 0-LFF 0)6¢¢'0 Z6Pr 0 (S6F 0-9¢F O)FLF 0 (9F $-E0'Z)9S'°S LY'P (18 ¥-29° §) 61°F (CL $-F9' 19'S |+(Z0F 0-S6E°0) TOF 0 ego's £680 (F6°S-G9 °Z)6L'°% | (628°0-79E'0)TLE 0 (g68 0-0F8'0)998 0 PPE 0 PPE 0 (OF 0-218’ 0)0¢E'0 6FE 0 (16g 0-69 0)08E'0 (ILE 0-8&Z 0)98z'0 9120 (6% 0-8&Z '0)922'0 (09°2-99' T)F0'% | e(GE8' 0-299" 0) 629 0 (922 0-ZFZ '0)€9Z0°0 928 (ST %-26'1)90°% 99¢°0 (S8h O-F2F 0) FEF 0 OF T #969 T ro €S¢'0 26 T SPE 0 80°T » 69¢6°0 G0'% #96 0 92 'T G2o°0 99 0 (gee 0-€88 0)¢Es 0 662 0 9FG 0 I1z 0 (880° 0-€80'0)280°0 Z61'0 82e 0 (680 0-120‘ 0)8SZ0'0 0910 (662 0-0&2'0)¢2z'0 (280° 0-010'0)2Z0°0 Fl 0 (221 0-11 0)6TT ‘0 (F9TC 0-8200'0)1T0°0 FOL 0 (ZE% O-I8T 0) L060 480] 601 0 aa 800° 0 1¢0'0 €90°0 s$L00 0 L€10'0 0200 °0 ¢(0600'0-9200'0)¢800°0 6€10 0 (8210 0-9010'0)ZF10 0 (8010 0-¥F00'0)#200°0 OTI10°0 (2400°0-1200'0)*200°0 (1800 0-F¢00' 0) 2900°0 1010 0 (2°10 O-ST10'0)2Z810°0 0¢00 0 2600 °0 ZL00 0 T¥00°0 99000 F900 0 sol COT PST +(826 0-220) L160" 99T (T8T'O-S9T O)EZT (166 0-GE% 0) 926° OFT (ILI O-SET 0)F9T” (FIE 0-Z9Z 0) L8z° OFT’ (Z9T O-19T 0) T9T” SEG 9§T- 6ST O&G 601° FET So o oo oo ponurzuop—% ATAV.L STEWART CHESTER A. (8ZE0-1FZ 0) 1920 11Z'0 (061 0-891 0)62Z1'0 (99% 0-G0Z'0)¢Ez'0 £020 (20Z 0-19T'0)08T 0 (ZZ 0-€02 '0)01Z'0 8810 Lvl 0 (SFE O-98T 0)E22'0 6210 (S91 0-221 '0)8ST ‘0 (22 O-9E1 0) 29T° IS (SET 0-121 0)081 0 i) (FOL O-OZ1'O)ZFT a (SEI O-IIT'0)&Zr 0 Su4n sb oo ¢(8Z10'0-9200 '0)2600°0 79000 (FS00 0-000 0) €F00'0 (#600 0-0200'0)Z800°0 6000 (8000-0800 0) 1F00°0 (F200 '0-F900'0)0200°0 ¢¢00'0 9F00'0 (26000-9900 0) #200 °0 Z200'0 (0S00'0-ZE00'0)ZF00 0 (0200 0-1F00'0)0¢00°0 FF00'0 (0€00' 0-#Z00 '0)2Z00°0 (S00 0-ZE00 0)8800°0 1F00°0 (Z£00 ' 0-ZZ00'0)2z00°0 sun ib (FZ I-68°0)ZI'T 180 (¢9'0-TF 0)z¢'0 (ZI I-¥9'0)26'0 6) ~ i) (89° 0-IF'0)TS'0 (20 I-92'0)88' 690 6¢° —) i) (SO I-02 '0)€8°0 G¢'0 (9¢ 0-2F'0)z¢ 0 (F9' 0-8 '0)9F 0 1g°0 (9F 0-ZE'0)88'0 (§9'0-GE'0)&F 0 62 '0 (IF 0-62 '0)¥£'0 sueplb (299° 2-069 (PIS I-80" (FOG Z-SP2 (299° I-80" (000 Z-SEs" (866 €-9FL” (€9F I-2E8° (91 1-269 (SPL 1-20" (S86 0-899" (#8 I-STO° sundb (621 0-6Z0°0)T20'0 CPO 0 (990° 0-260’ 0)€S0°0 (¢60°0-F0 0)190°0 rag) (€20' 0-FF0 '0)FS0'0 (290° 0-S¥0'0)8¢0'0 6800 1F00 (SOL 0-F£0'°0)290°0 9¢0 0 (SF0'0-0F0 0) 8F0'0 (680 0-810’ 0)9Z0°0 6200 (690° 0-ZF0'0)8F0 0 (ZE0' 0-020 '0)9z0'°0 120°0 (F20°0-ZE0 0) S00 suenub SAGNCIN STIVNGU Vudos 314 Ay dur gy Ua f SANILSGALNI GNVY HOVWOLS penurywoO—?s HIAVL N@aATdS 315 WEIGHTS OF UNDERFED YOUNG ALBINO RATS (099° 0-629 °0)6¢9'0 LL9°0 6820 (E82 0-9F9'0)929°0 2890 (968° 0-192 '0)F08"°0 (026 0-9TF 0)E8¢ 0 cre 0 (ZFS 0-029 ' 0) TES" co) (E99 0-P2F 0) FGF 0 Z0S 0 (889° 0-Z8S'°0)019'0 919°0 €8h 0 OFS 0 pLE'0 8&e 0 che 0 (9820'0-G210°0)¢0z0'°0 10z0°0 Z810°0 (0910 0-FFI0'0)FS10*0 * ¢210°0 (S810 '0-9610°0)Z210°0 (F680 0-Z210' 0) 88200 — 6ST0°0 (9S10'0-0&10'0)EF10'0 (8FZ0'0-G810°0)F0Z0'0 1810°0 (2210 0-¥810°0)9¢10'0 1SZ0°0 - OFTO'O T9TO°0 T9100 86000 9800°0 (80°-S¢'Z)Z8°% £9'F G6°S (60° €-0Z'%)99'°Z OL'F (Sl F-¥6 '£)t0' F (SZ 2-68 1) 10'S is (62° 8-8 'Z)08'¢ (Ih 2-68 '1)90°% Irs (Z2Z°=L8 1626'S 88'T Looe 02°% - rr oo O oD aN 1D (086 22-8EF' 11602" LI LOI'8 (0Z8 0Z-00L'¢) 2@F OL (ZF 6-9CL" 1)669'8 (96F 8-066 °€)622°¢ (1h 8-LF9° 2)080°8 (189° S-618" €)899 F (ZL S-400'F)€98'F StI 80L°9 CFO 869 '§ (622 0-861 0) 112° O61’ C6 (1@Z'O-LLT 0) 161° £61 (98F O-S8T 0) Ze" (GZS 0-LZIT 0) 8&3" OFT” (620-208 0) EE" (628 0-981 0)ZI1Z" std (ZST 0-21 0) ZET TOT LOT T0G° o 0 0 i =) oo A. STEWART CHESTER Z100°0 Z100°0 (01000-9000 0) #8000 '0 (8000°0-9000 0) 69000 °0 (01000-9000 0)Z8000 0 (60000-9000 0)62000°0 (8000°0-9000 0) 2000°0 8000°0 (80000-9000 °0)€2000°0 (01000-8000 °0)28000°0 (0100°0-9000°0)92000°0 (80000-20000) 22000 '0 (60000-9000 °0)62000°0 (01000-2000 °0)#8000 0 suv. 0Z00°0 8100'0 Gc00 0 ($Z00° 0-2100° 0)0Z00'0 Z100'0 (0Z00' 0-€100'0)9T00°0 (22000-61000) 12000 Z100'0 (S100 0-F100'0)2100'°0 (61000-2100 '0)8100°0 TT00°0 91000 (6100 0-100’ 0)ZT00'°0 T100°0 (8100' 0-100 0)9100 0 (0200 °0-0100°0)9100°0 6000 °0 (Z100° 0-100 '0)Z100'0 (6100 0-#100'0)9100'0 600°0 (F100'0-1100'0)Z100'0 swmpsb (Z900' 0-200 0) #00°0 €F00'°0 (0F00' 0-6100' 0)8Z00'0 (82000-9200 0)¢900'0 6€00 0 FE00 0 (2600° 0-6100'0)€00°0 1g00 0 (9200 0-1100 0)8100°0 swmpsb (S10 0-210’ 0)F10'0 (6100-010 '0)¢10°0 (#6Z0' 0-O00T0'0)210°0 (910'0-€10'0)¢10'0 (020 '0-ZI0'0)¢T0°0 (S600 0-8900'0)8200°0 sudub (020° 0-SF0'0)660'0 090°0 (FF 0-€80'0)6E0°0 (880° 0-2L80°0)890°0 #900 (1E0'0-€€0'0)ZF0'°0 (FS0' 0-680 0)9F0'0 IF0'0 (S10 0-F10 0)910'°0 supnsb ) 31¢ AdGOd IWANId SISAHdOdAH SaIUVAO INACICGIda SaLsad € e penuyuoj—?Z ATAVL YOUNG ALBINO RATS Sty UNDERFED OF WEIGHTS (6000 °0-2000°0)8000°0 8000°0 (80000-2000 0)¢2000°0 (100° 0-0100 0) T100°0 (Z100'0-0100'0) T1000 (€100°0-€100°0)¢100°0 (€100'0-1100°0)Z100'0 (€100°0-1100'0)%T00°0 0100°0 6100°0 (S00 0-00" 0)0F00°0 8&00°0 9€00°0 (6&00° 0-8Z00°0)£&00'0 #£00°0 (28000-1800 0)ZF00'0 (1000-6200 0)#£00'0 8200 '0 (100' 0-000 ° 0) Tg00°0 (0800 ' 0-9200°0)8Z00°0 92000 (F£00' 0-100 0) £g00°0 0€00°0 $600 '0 0€00°0 (0110 0-Z200 0) 1600°0 ¢600'0 FI10'0 (SZL0’°0-€800 0) 0100 €800'0 (00T0° 0-0800'0)0600°0 "]BOISO[ONIB » ‘[BNPTAIPUL T JO 9FBIOAY g ‘S[BNPTIAIPUL Z JO VHBIBAY » ‘S[UNPIAIPUI & JO BDBIBAY ¢ ‘S[BNPIAIPUL 9 JO BBBIDAY z "S[ENPIATIPUT G JO OFBIOAY |; (Str 0-28% 0)ETE 0 (1ZT' 0-€60°0)801 0 (€20' 0-120 '0)0#0°0 (F90°0-€20'0)6S0°0 (660' 1-¥86'0)E¢0'T POLO (IFT 1-296 '0)9F0'T (262 0-6I1T 0)822°0 TZ€°0 (TIF 0-828 0)F6E 0 VOL. EXPERIMENTAL ZOOLOGY, THE JOURNAL OF 318 CHESTER A. STEWART RATIO OF TAIL LENGTH TO BODY LENGTH Immediately after death each rat was laid on its back and gently extended. The distance from the tip of the nose to the anus, and from the anus to the tip of the tail was carefully meas- ured. From the measurements thus obtained individual ratios between tail length and body length were computed. The average of these tail ratios expressed as percentage of the nose- anus length is given in table 1 for each group of rats. The data in table 1 show that with advancing age, the tail ratio increases in both the control and test rats. However, as com- pared with that in younger controls of corresponding body weight, the tail ratio, without exception, is much higher in the test rats. At ten weeks of age the ratio for the underfed individuals averages approximately 0.78 (73.9 per cent in the males and 82.0 per cent in the females), as compared with an average of 0.55 (53.7 per cent in the males and 56.6 per cent in the females) for the younger controls of corresponding weight. The differences at other ages, though somewhat less, are also very pronounced. It is there- fore evident that in the underfed rats, the tail continues to grow in length more rapidly than the body. Thus the tail ratio in the underfed rats at ten weeks (0.78) approaches, but does not quite reach, the normal ratio of 0.88 (Jackson ’15 b) for (larger) rats of corresponding age. In the rats underfed for very long periods, however, the data (table 1) indicate an elongation of the tail (ratio 0.89-0.93) even beyond the normal ratio for corresponding age. 10. gate, the fourth being considerably behind the third in its de- velopment (fig. 9). Viewed from above, the arm at this time is EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 421 seen to be bowed toward the body. Further changes in the form of the limb are concerned largely with the lengthening of the various segments, notably the digits, and the more distinct demarcation of the arm, fore arm, and manus. _ Rotation takes place at the shoulder, the arm pointing more laterally and ventrally, so that the tip of the first digit rests on the bottom. Further rotation at this joint, coupled with flexion at the elbow, brings the manus much further forward beneath the gills, and the animal now rests upon two digits of each limb. The balancers, which serve to support the larva on its belly, are not lost till this stage is reached. T he first muscular move- ments take place at the shoulder before this period, and later, movement at the elbow and wrist joints begins; the limb is then used in crawling, the positions Just described being those of normal rest. These changes are completed just about the time the yolk is entirely gone and the larva begins to feed. SIMPLE EXTIRPATION OF THE LIMB BUD Mode of operation The first experiment which will be considered consisted in the simple extirpation of the body wall of the limb region. To per- form this operation the scissors are inserted through the outer layers of the embryo at the anterior (cephalic) border of the re- gion to be removed and then the embryo is turned while a cir- cular incision of the desired size is cut. The part of the body wall thus isolated may be readily lifted from the underlying tissue and removed in entirety. Care has to be taken at the upper border of the wound to disengage the limb rudiment from the pronephros, if the latter is to be saved intact. When the limb dise is lifted, both somatopleure and splanchnopleure come away with the overlying ectoderm, since the mesoderm has not split’ at that time. The operations were mostly done in the stage shown in figure 2, though the age of the embryos used varied from the stage shown in figure 1 to that in figure 3. No difference in the results, which may be ascribed to the difference in age within these limits, has as yet been noted. In the younger 422 ROSS G. HARRISON specimens the layers of the embryo are more readily separated, though in the older ones the tissues are of a firmer consistency and can be handled more satisfactorily. After the dise is excised the wound at first gapes, but a few minutes later it contracts. Then in the course of the next twenty-four or forty-eight hours the ectoderm stretches itself over the wound bed and covers it entirely. There is much variation in the time required to complete this process, and in some cases the wound has been found still partly open four days after the operation. Sections show that the wound is first covered by ectoderm and that the mesoderm creeps in soon after between that layer and the yolk. The first problem is to determine whether by an operation of this kind the development of the limb can be prevented, and, if so, how the size of the wound affects the outcome of the experi- ment. Two kinds of operations were done. In the first the limb dise was simply lifted and the wound left without further treat- ment. In the second the wound bed was afterwards carefully cleaned of all mesoderm cells. In some cases of each kind the pronephros was left intact and in others it was removed. Ex- tirpation of this organ facilitates the cleaning of the wound, but since many cases of non-regeneration with intact pronephros occurred as well as some cases of regeneration without its pres- ence, its influence, if any, upon the development of the limb must be unessential. Relation of regeneration to size of wound and completeness of removal of mesoderm | The smallest wounds were three somites in diameter, encom- passing in all but two cases the region ventral to myotomes 3, 4, and 5. In the two exceptional cases the wound, which was not cleaned, included the area below somites 4, 5, and 6; one of these regenerated and one died. The largest were 43 somites in diameter, extending from the boundary between the second and third to the vertical line dividing the seventh myotome in half. EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 423 TABLE 1 “Showing results of extirpation of circular areas of body wall of given diameter; wound not covered WOUND NOT CLEANED WOUND CLEANED és o a) dx % d 2 2 DIAMETER OF g eS co} ee = 28 ce) 3 EXTIRPATED AREA 2 25 by Total ae 2 2 by Total a z as | ga | Ze “a | 9s | 82| ge os ae 2 Ep & 5 5 bo aS ah | 85 B Sp Oo ‘) a Ay 6) 6) A oo Not recorded...| 11 LOT 4a e252) al 5224) 0 0 ee) 3 somites......- 19 1 4 24 (20) | 95.0) 13 12 9 |34 (25)| 52.0 31 somites.....| 1 0| 0 my @ 1 0 Gah °@) 34 somites..... 111 93| 81 |215 (184) | 82.8 3 18 ti 1S2 (20) T4253 4 somites...... 14 2 2 189 16) |F8ie5 2 3 a Sees) 42 somites..... 2 0| O Dimas (2) Motaleence-s 158 | 36] 91 |285 (194) | 81.4) 19 oD 34 186 (52)| 36.5 «Ym calculating the percentages the number surviving and not the whole numbers of operations has been used. Percentages are given only in those elasses where the number of cases is sufficiently large to be of significance. The standard wound, which was made in the largest number of cases, was 34 somites in diameter, including the area ventral to the third, fourth, fifth, and half of the sixth myotomes. The» wounds of four segments in diameter took in usually the region extending from the third to the sixth somites, inclusive, though in one case the place of the wound was shifted one segment, and in ten other cases a half segment toward the head, thus including in the latter that portion of the body wall between the posterior half of the second somite and the anterior half of the sixth inclusive. As shown in table 1, the result of the experiment depends in a certain measure both upon the size of the extirpated area and the completeness of the removal of tissue. Taking all of the experiments without reference to size, 81.4 per cent of the cases with ordinary wound regenerated limbs while only 36.5 with cleaned wounds did so. In both sets of experiments, increase in the size of the wound over three somites considerably reduces the proportion of regenerated limbs. High mortality of the large clean-wound class has considerably diminished the number of 424 ROSS G. HARRISON M eases of this kind available. In these cases the healing is usually very slow. The yolk, being long exposed, begins to disintegrate after a time, and the embryo rarely recovers if this process sets in. It is possible that when a few mesoderm cells are left in, they cover the yolk and facilitate the overgrowth of the ectoderm. Although the regenerative capacity is dependent to some extent on the size of the wound, there are experiments with wounds of each of the several sizes in which no regeneration took place. Within the limits of these operations it is there- fore impossible to say that wounds beyond a certain size pre- clude regeneration altogether. This can only be said of those cases in which the wound is covered with indifferent skin, as described in another section (p. 432). It is reasonably certain from a study of the normal develop- ment that the cells which ordinarily give rise to the limb bud take origin in the region below the fourth and the neighboring parts of the third and fifth somites. In this region, as the limb bud becomes more prominent, numerous mitoses are found, while the rest of the somatopleure is almost devoid of dividing cells. In those cases in which the region corresponding to the whole of these three somites, or even more, is removed the cells which give rise to the regenerated appendage are, therefore, probably such as in normal development do not participate in its development. Miss Byrnes (’98) has emphasized this in her paper on the hind limbs of Rana. Relation of regeneration to age of embryo at time of operation There seems to be no definite correlation between the age of the embryo at the time of operation and the occurrence of regenera- tion. The tabulation here given (Table 2 a), which includes both covered and uncovered wounds shows that when cases are re- corded in significant numbers, both positive and negative results are obtained after operation in both older and younger stages. The number of cases operated in the extreme stages is small and no statistical value can be placed upon the figures there. EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 420 TABLE 2a Showing frequency of regeneration after operations in the several stages WOUND NOT CLEANED WOUND CLEANED ALL EXPERIMENTS COMBINED STAGE CG Cases Per Cc Cases | Per WDerd | Per pepe | ote | See [eae repens | cenere or dis-| Total eases | Cont erated | grated | erated erated erated | erated carded erated 25 4 0 0 1 PN (5) 26 3 0 2 3 0| 8 (8)| 27 8 2 80.0) 3 1 ETE: 4|28 (24)| 45.8 28 20 On| 10020) . Qin LL. | tee 20 | 53 = (33)| 66.7 29 65 | 12 84-4) 10 -| 25 | 28.6 Al |153 (112), 66.1 30 26 5 83.9) 5 9 | 35.7 28 | 73 (45)| 68.9 31 12 3 80.0) 2 3 28 | 48 (20)| 70.0 32 “4 4 2 0 4|14 (10)| 60.0 33 14510 BSeo maa! 0 4|29 (25)| 60.0 Miscellaneous* 4 5 0 2 Gr a a) ovale eas: 160 | 41 27 65 137 |430 (293) * Stage not recorded or else outside above limits. TABLE 28 Showing frequency of regeneration after operation in the younger and older groups of stages, respectively ALL EXPERIMENTS w ND ANED WOUND AD ou? NOT CLEAN OUND CLEANED COMBINED STAGE Cases Per Cases Per Cases Per wee not cent ate not cent oe not cent Aisi || Seleae || ee emited! tooo es bn eas | exated eats || ieee erated | erated erated | erated erated | erated Stage 29 or under...| 100 | 14 Siataeied 5 525.0!) 11) 6S | e£-3 Over stage 29....... BErh 2p hides.) «10 12 | 45.5 66 | 34 | 66.0 One point stands out, however, which may possibly be of sig- nificance, though no altogether satisfactory explanation is ap- parent. In the group with wounds not cleaned a higher per- centage of regenerating cases is found among those operated in the younger stages, while in the group with cleaned wounds the higher percentage is at the opposite end of the series. This is brought out more clearly when the experiments are divided into but two classes, comprising respectively the cases operated when older than the standard operating stage and those in or below it (Table 2b). The lower percentage of regeneration in 426 ROSS G. HARRISON TABLE 3 Showing time factor for regeneration in the different classes of experiments OLDEST averace | AVERAGE |NUMBER OF|NUMBER OF] STAGE res) Sara peed ON CASES CASES NEGATIVE NUMBER | OPERATION | OPERATION | REGENER- | REGENER- | OBSERVA- CHARACTER OF OPERATION | 5, GiGEs ee LAST ATION FIRST| ATION FIRST TION Tat AGONIST PREVIOUS | OBSERVED | OBSERVED | FOLLOWED RISD | OBSERVA- |AT 10 Days) aT 15 DAYS BY . | TION OR LATER | OR LATER | REGENER- ATION days days | days Wound not covered: Not cleaned...... 158 6.8 2.8 21 =| 4 10 Gleaned.......... 19s | 9:2 5.4 7 2 15 Wound covered: aa ¢ ~ | ~— Not cleaned...... 2 13-5" |) “2tot 1 1 4 Cleaned eee eee 8 9.4 6.0 3 1 11 * The figures in this class have no significance since in one of the two cases no observation was recorded between the fourth and the twenty-third day. the older embryos in the not-cleaned series may be due to the fact that the limb bud can be lifted out more cleanly than in the younger stages, so that after-cleaning is less needed. The higher percentage of regeneration after operations in the older stages in the cases with cleaned wounds remains, however, anomalous. Delay in development after extirpation Removal of the limb rudiment, if it does not arrest develop- ment entirely, naturally retards it to a considerable degree, the delay being on the average greater in the case of the cleaned wounds than in the others (Table 3). In a few cases outward signs of regeneration were noted as early as four days after oper- ation, though usually not until later. Since it was not practica- ble to observe the embryos each day, it cannot be stated ex- actly when regeneration did begin in each case. One hundred and fifty-eight cases with ordinary wounds show that the re- generating limb was first noted on the average 6.8 days after the operation, the last previous observation having been made on the average at 2.8 days. In the case of the cleaned wounds, the first observation of regeneration averaged 9.2 days after operation, the last previous observation having been at 5.4 = EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 427 days. This is an average of nineteen cases. Daily observa- tion would undoubtedly have revealed a greater difference be- tween the two classes of cases in time of appearanec of the re- generating limb bud. In a total of 177 cases there are but twenty-eight in which regeneration was first noted as late as ten days after the operation, and but six when the period was fifteen days or over. In only four of the twenty-eight cases was the last previous observation made as late as ten days after the operation, and only once was a negative observation, with subsequent regeneration, recorded as late as fifteen days. It is safe to conclude, then, that when regeneration has not be- come visible from the outside two weeks after the operation it will not occur at all. Nearly all of the cases have been kept alive at least three weeks, many for four, and some as long as twelve (fig. 14), and no instances of regeneration in the later periods of development have been observed. The experiments described in this section show that it is not possible to prevent the regeneration of the fore limb with certainty when the wound is left uncovered, even if the circular area extirpated has a diameter of 4 somites. It may be said in general, however, that the larger the area removed and the more carefully the wound is cleaned of mesoderm the less likely will regeneration occur. This can only mean that the meso- derm cells of the region surrounding the limb bud have, in a gradually diminishing degree as the distance from the limb increases, the potency to form a limb. Their prospective po- tency is, therefore, greater than their prospective significance. Regeneration, when it takes place, is usually perfect, though subject to delay in a varying degree. The process of regeneration The actual process of regeneration has not been followed in detail, though some observations may be recorded here. The earliest case (R. E. 127—) which need be considered was preserved six days after extirpation of the limb rudiment. The regenerating limb appears on the surface (fig. 10) as a 428 ROSS G. HARRISON small nodule ventral and posterior to the pronephros. In sec- tions it is seen to consist of closely packed cells which have approximately the same yolk content as those of the normal limb on the opposite side and which show numerous mitoses. This indicates that after the defect is covered up by inwandering of peripheral cells (p. 422), the process of regeneration, like the original development of the limb, is dependent upon multiplica- Figs. 10 and 11 Embryos with regenerating limb buds, ventral view. x ay Sr ISOMMECH sere oor 0 15 | 4 |19 (15)} 00.0 ASSOMMUCS aE AA eer 0 Bi] Pe Nag) (G4) PRO talc o,sin5 ere | 2 5 3) 1) 8 32 | 9 |49 (40) 20.0 1 1 * These three cases were classed as negative in the tabulation previously pub- lished (Harrison, ’15). They are excluded here because they were not kept under observation for a sufficient length of time. in diameter, regeneration is altogether blocked, eighteen cases all having given a negative result. To these might be added the five cases given in Table 8, which differ from the present experiments only as to the region from which the covering ectoderm is taken (p. 448). When the skin is grafted to the wound it soon sticks to the underlying tissue just as a transplanted limb bud does, and often in the course of several hours the wound becomes com- pletely healed. The wandering of the ectoderm over the de- nuded area, which takes place in the uncovered wounds, is blocked, and probably also the movement of the mesoderm cells. Prevention of the surrounding cells from reaching the proper position by the substitution of other tissues which do not have the potency to produce a limb thus effectually prevents the regeneration of the appendage. The cells immediately surrounding the limb-producing area are evidently unable to form a limb unless they can reach the proper position. On the other hand, indifferent cells in this position cannot produce alimb. It is possible that when no mesoderm is grafted with the skin, the cells of this layer wandering in from the host may in 434 ROSS G. HARRISON some cases give rise to a limb. Some of the individual dis- crepancies in this series of experiments may be due to uncon- scious varying of this factor. In experiments in which another limb bud is transplanted into the place of ari extirpated normal bud, the grafted tissue must act like a piece of body wall from an indifferent origin in so far as its effect upon the movement of the cells of the host is concerned. It must prevent these cells from wandering into the proper position to form a limb, and hence when transplantations of the limb bud are undertaken with proper precautions as to size of wound and thorough cleaning of the mesoderm, it is safe to assume that the limb that does develop arises from the trans- planted material and not from the tissues of the host. The exact determination of the size and character of the wound neces- sary to prevent, regeneration is therefore important for the proper interpretation of any experiments in which the normal limb bud is replaced by a grafted one. EFFECT OF REMOVAL OF PORTIONS OF THE LIMB BUD It was scarcely to be expected that an organ having such marked regenerative capacity as the limb rudiment would show any distinct localized effect of the removal of definite portions. A number of experiments have nevertheless been made to test the prospective potency of its parts. The procedure was as follows: The limb area was first bisected by a vertical or a horizontal incision and half of the dise—anterior or posterior, dorsal or ventral—was removed. Some of the wounds were left without further treatment; in others the mesoderm was carefully cleaned off and the wound left to heal; and in still others the denuded area was covered with ectoderm from the flank of another embryo just as in the experiments with whole limb buds. In a smaller number of cases, a small circular area, 1 or 14 somites in diameter, was removed from the center of the limb rudiment. The first experiments, which were referred to in the preliminary paper (Harrison, 715), were few in number and were made without special cleaning of the wound. They resulted in the development of normal or only slightly defective EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 435 TABLE 5 Summary of results of removal of portions of the limb bud. Wounds cleaned a RESULTING LIMB PART REMOVED z Normal | Redupli- | efective| Abortive| Dead | Total Henicene INMRAOLOI, socaccenec 3 6 0 1 4 {14 (10)} 30.0 IPOSISIAIGIO 5.50004 2 1 3 4 3 13 (10)} 20.0 Dorsale see es 1 0 1 10 2 14 (12) 8.3 Vientralieeeees seer 8 0 0 2 4 14 (10)} 80.0 Centraleyeerees 4 1 1 0 1 eG) OOM Notas erence: 18 8 5 1L7/ 14 62° 48)| 3725 TABLE 6 : Summary of results of removal of portions of the limb bud. Wounds not cleaned RESULTING LIMB PART REMOVED 3 Normal Rete Defective| Abortive| Dead sDotad a | coco PNTIKETAOVEs penecnan oe 12 0 1 0) 4 ie (LS) ee 923 IROStELION eee eee 6 0 4 1 4 15 (11)) 54.5 IDOTSa leet eee 5 0 iff 4 1 LE GO) 5020 Went eee 7 0 1 0 2 10 (8)| 87.5 @entralh-oeeeeae ee 1 0 1 0 0 2 (2)| 50.0 RO sete 31 0 8 5 11 55 (44)} 70.5 limbs, but cannot be regarded as a fair sample of what the oper- ation may effect. During the season of 1917 the experiments have been greatly augmented in number, and with results which seem rather different. There has been a large proportion of cases with defective or completely inhibited development of the operated limbs, although no specific correlation between the part re- moved and the character of the defect has been observed. As in the experiments with whole limb buds, a much higher per- centage of perfect limbs has resulted when the wound is not cleaned. The results of the experiments in the two classes of cases are given in Tables 5 and 6. Those with cleaned wounds (Table 5), including all which were covered with grafted skin, will be considered first. In this 436 ROSS G. HARRISON group it is obvious that the proportion of normal limbs varies very greatly among the different operations. Thus when the ventral half of the limb bud-was removed, eight out of ten re- sulted in perfect limbs; but when the dorsal half was taken out, only one out of twelve gave rise to a normal appendage. The anterior and posterior halves occupy an intermediate position between the two extremes. These discrepancies are largely, if not entirely, due to the difficulty of exactly halving the ma- terial that is to form the limb. In bisecting it vertically the myotomes were used as a guide, and in most cases the incision was made below the middle of the fourth somite, leaving one and a half somites in front of the incision and two behind it. This seems to divide the material more nearly in half than when the incision is made a quarter somite further back. In the case of the horizontal incision it is more difficult to divide the rudi- ment accurately, because there is no sharply defined landmark. The attempt was made to cut a little below the pronephric swelling. The results show that more of the limb material lies above the cut than below. In other words, if the circular area, centering just below the pronephros and extending from the boundary between the second and third somites to half way through the sixth, is bisected vertically and horizontally, more limb-forming material lies dorsal to the line than ventral, and more anterior than posterior. The lines shown in figure 2 desig- nate more nearly the exact quartering of the material. The fact that normal appendages resulted in some cases after each kind of operation shows that there is no localization of prospective potencies in the operating stage. Examination of the character of the defects that do arise confirms this conclusion. The large proportion of the latter are defects of the whole limb, ‘which remains a nodule or becomes entirely resorbed. They must be ascribed to the general effect of the operation and not to the removal of any specific material. Seventeen cases out of a total of forty-eight operations (not ineluding cases that died) with cleaned wounds resulted in this way. Of the partial defects, listed under the caption ‘defective,’ all five affected the hand, especially the digits. The most marked case was one in EXPBPRIMENTS ON THE FORE LIMB OF AMBLYSTOMA 437 which the dorsal half of the limb bud was removed (Rem. E. 27). This was a club-shaped appendage less than half length and without hand (fig. 15). The next most marked defect followed removal of the posterior half of the bud (H. R. E. 12-); here the forearm was slender and but one whole digit was developed, Rete we 16 20 Figs. 15 to 20 Outline views of defective limbs which developed after re- moval of portions of the limb bud. X 17. 1, 2, 3, 4, ordinal number of the digits; d., reduplicating digit. Figure 15. Experiment Rem. E. 27 (dorsal half re- moved); lateral view right side, considerably foreshortened; twenty-six days after operation. Figure 16. Experiment H. R. E. 12— (posterior half removed) ; ventral view; twenty-six days after operation; the arrow points headward. Figure 17. Experiment Rem. E. 13 (posterior half removed) ; limb amputated and preserved twenty-seven days after operation; ulnar half of hand is very defective, but the limb which regenerated after amputation wasnormal. Figure 18. Ex- periment Rem. E. 29 (posterior half removed) ; lateral view right side, arm much foreshortened; first digit (1) abortive and syndactylous with second, which has a reduplicating branch (2D); thirty-one days after operation. Figure 19. Experi- ment Rem. E. 21 (central portion of limb bud removed); limb amputated and preserved twenty-six days after operation; right limb lateral view. Figure 20. Experiment H. E. 21 — (anterior half removed); right side ventral view, eighteen days after operation; the arrow points headward. which, however, had a reduplicating branch (fig. 16). In an- other case of removal of the posterior half (Rem. E. 13) only a single long digit and the stump of a second were developed, the hand being quite defective (fig. 17). Amputation of this in- complete limb was followed, however, by the regeneration of a normally formed one. The two remaining cases showed defects 438 ROSS G. HARRISON in the first digit. In one (Rem. E. 29), where the posterior half of the bud had been extirpated, the first digit is a mere stump and is syndactylous with the second, which, however, has a small reduplicating branch on the ulnar side (fig. 18); in the other (Rem. E. 21), from which the central portion had been ex- cised, the first digit is short (fig. 19). . The cases with ordinary (not cleaned) wounds (Table 6), like the foregoing, show the highest proportion of defects after re- moval of the dorsal half of the limb bud. However, the lowest proportion of defectives occurred after removal of the anterior half, a result for which there is no apparent explanation. Re- moval of the dorsal portion gives a relatively large number of cases of complete suppression of the limb. Of the partial de- fects, the most marked case (H. E. 21—) has the hand totally lacking, the arm ending as a rounded stump (fig. 20); here the anterior portion had been removed. Another case (H. HE. 13-, posterior half removed) appears similar though less clear. In four of the remaining, two of which followed extirpation of the posterior (Rem. E. 2 and 7), one of the ventral (Rem. E. 1) and one of the central (Rm. E. 9) portion of the bud, the defect in- volves only the first digit, which is either absent or short (figs. 21 B-24 B), while the other two cases (H. E. 18-, removal of dorsal half, and H. R. E. 33-, removal of posterior half) have the second digit short (figs. 25 B and 26). One of the cases (Rem. E. 9) shows the same defect on both the operated and the unoperated sides (figs. 24 A and B). In a few of the cases, webbing of the first two fingers occurred (figs. 27 B and 28). This has at times been found in other experiments and even in unoperated limbs, and since the deviation is slight these cases have been classed among the normal. Turning to the reduplications, we find them concentrated among the cases in which the anterior half of the limb bud had been removed, six out of the eight falling within this group. They include a variety of forms, such as two separate and nearly complete limbs (fig. 34), a single normal limb with a spur at- tached to the upper arm (fig. 33), and a single limb with merely a branched or double digit (fig. 35). Three of them (Rem. E. EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 459 Figs. 21 to 28 Effect of removal of portions of the limb bud (continued). In the double figures A represents the limb of the unoperated and B that of the operated side. In all cases lateral view. X17. Figure 21. Experiment Rem. E. 2 (removal of posterior half); upper arm unoperated limb (A) much fore- shortened; operated limb (B) generally less advanced in development, and with abortive first digit (1); seventeen days after operation. Figure 22. Experi- ment Rem. E. 7 (removal of posterior half); digits of normal limb (A) foreshort- ened in part; first digit of operated limb abortive; twenty-six days after opera- tion. Figure 23. Experiment Rem. E. 1 (removal of ventral half) ; unoperated (A) limb not so long as operated (B); which has only one long digit; twenty-nine days after operation. Figure 24. Experiment Rem. E. 9 (central portion re- moved); both limbs foreshortened; unoperated limb further advanced, but both show defect of first digit; twenty-six days after operation. Figure 25. Experi- ment H. E. 18 —(dorsal half removed); limb of operated side has short second digit (2); eighteen days after operation. Figure 26. Experimen: H. R. E. 33 — (removal of posterior half); second digit (2) is abortive and fused with the first (1); third digit well developed; twenty-one days after operation. Figure 27. Experiment Rem. E. 3 (anterior half removed); operated limb somewhat less developed than normal; syndactyly of digits 1 and 2; twenty-seven days after operation. Figure 28. Experiment H. E. 28 — (removal of posterior half) ; well- developed limb with syndactylous first and second digits; twenty-eight days after operation. 440) ROSS G. HARRISON Figs. 29 to 37 Effect of removal of portions of the limb bud (continued). Lateral view of limbs in all figures except figure 30. X 17. Figure 29. Experi- ment Rem. E. 48 (removal of anterior half); arm foreshortened, but hand not; radial reduplication of hand; twenty-seven days after operation. Figure 30. Experiment H. E. 6 — (anterior half removed) ; ventral view; radial reduplication of hand more complete than in last case; thirty-nine days after operation; the arrow points headward. Figure 31. Experiment Rem. E. 52 (removal of ante- rior half); A, normal tetradactylous limb; B, operated limb, with double hand, considerably foreshortened; twenty-seven days after operation. Figure 32. Experiment Rem. E. 28 (anterior half removed); arm considerably foreshort- ened, probably reduplicated internally; first digit double (1D); N, nodule at oo BXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 441 48, Rem. E, 52, and H. E. 6—) show more or less complete sym- metrical recuplication of the hand (figs. 29, 30, and 31). An- other (Rem. E. 28) has only the first digit doubled externally (fig. 32), though the unusual thickness of the limb indicates partial internal reduplication. Near the shoulder there is a distinct nodule. The case of the single limb with spur (Rem. KE. 17) is interesting in that the spur developed out of a bud which erew posterior to the wound scar and which at first seemed to be the main limb. Anterior to the wound a second bud appeared, first noticed eight days after the operation. On the twelfth day it was still but a nodule, and not till the eighteenth day did it look like a regenerating bud. It finally developed into a normal limb of the proper laterality, the posterior bud remaining attached to it as a spur (fig. 33). The case with two independent limbs (H. R. E. 10—) is fundamentally similar to the foregoing. The main limb bud developed posterior to the wound. Later there appeared anterior to this a second bud. In this case, however, the two buds remained permanently separate. The anterior one gained over the posterior and became a normal limb of proper laterality with good function (fig. 34). The posterior one remained somewhat defective (second digit short and ulnar digits undeveloped), and when last examined alive, thirty-eight days after operation, it showed no movement. The remarkable feature of this case is that both limbs are right-handed, as was probably true of the spur case also, though the spur is too de- shoulder; thirty-one days after operation. Figure 33. Experiment Rem. E. 17 (anterior half removed); hand foreshortened dorso-ventrally; S, spur repre- senting a duplicate limb; forty-three days after operation. Figure 34. Experi- ment H. R. E. 10 —(removal of anterior half); two left limbs; ANT., anterior member, developed secondarily from anterior border of wound; POST ., posterior member developed from the remaining half of the limb bud; forty days after op- eration. Figure 35. Experiment Rem. E. 18 (posterior half removed); radial digit reduplicated (D); limb amputated and preserved twenty-seven days after operation, followed by regeneration of normal limb (fig. 37 b). Figure 36. Ex- periment Rem. E. 16 (central portion removed) ; reduplicated second digit (2D); twenty days after operation. Figure 37. Experiment Rem. E. 18 (posterior half removed); A, normal left limb, B, normal right limb, regenerated after amputation of abnormal limb shown in figure 35; thirty-three days after ampu- tation. 442 ROSS G. HARRISON fective to reveal its laterality. This case differs from the three described above in which the hand is symmetrically reuuplicated, one member being a right and one a left, following the rule of mirroring. It is probably of a fundamentally different nature from the others in that the posterior member obviously arose from the remaining half of the limb bud after operation, while the an- terior one regenerated from the anterior border of the wound, the two remaining far enough apart not to influence one another. In the case of the true reduplications, the two members presum- ably arise from a single center which later doubles symmetri- cally. The other two cases of reduplication are not important. One (Rem. E. 18) involved the first digit only (fig. 35), and the other (Rem. E. 16) the second digit (fig. 36). In three cases the abnormal appendages which developed were amputated between the shoulder and elbow. One of them (Rem. E. 13, with a very defective hand with only one digit (fig. 17) ) and another (R. E. 18 with a reduplicated and a de- fective digit (fig. 35) ) regenerated normal limbs (fig. 37). The other failed to regenerate. These experiments show that such anomalies can hardly be due to the removal of specific organ- forming tissues from the rudiment. The anomalies are summarized in table 7. From this tabula- tion it is seen a) that defectiveness of the first digit may occur after removal of the posterior or the ventral halves or the cen- tral portion of the limb rudiment; 6) that defectiveness of the whole hand may arise after removal of the anterior, posterior, or dorsal halves; and c) that abortive limbs may occur after removal of any of the four halves. As for the reduplications, those of major degree are confined to operations on the anterior half of the limb bud. Minor reduplications, affecting the digits only, occurred in one case after each of three different operations. It would require a number of experiments many times that in- cluded in the above table to give statistical value to the num- bers in the several categories, and it is not likely that these can be done in the near future. © Possibly the repetition of the ex- periments on a large scale might show, for instance, a relatively high proportion of defects in digits after removal of the posterior EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 443 TABLE 7 Showing distribution of anomalies among the several experiments in removing portions of the limb bud PART REMOVED Anterior Posterior Dorsal Ventral Central CHARACTER OF ANOMALY Not cleaned Not cleaned Cleaned Total Total | Not cleaned | Cleaned Not cleaned Total Cleaned Not cleaned Total Cleaned Total | Cleaned ee Syndactyly of first two digits*. Wy || Wy \\ 2 First digit defect- ive or absent.... 1 De |B Second digit de- WCii@evces 220ec 1 1 1 1 Hand defective.... 1 1 Forearm and hand defective....---- 1 1 1 High degree of defectiveness— hand absent..... 1 1 1 1 Reduplicated Gulati sooBaOee 1 th Weil 1 Reduplicated hand mirrored. .| 3 3 Limb with redu- plicating spur... 1 1 Two distinct limbs not mir- Whole limb abor- tive or resorbed.| 1 1 4 1 Ee Wo | 4 |f4 \2 2 Total anomalies...| 7 1 SHS: eos jl" 1d GN. 24 1 3 Normal excl. syn- dactyly:. sagas ee 3) lO ES see ieee ee, tok || ane! 7 {15 ee eae Per cent nor- ne esate < 30.0|92.3]65.2)20.0/54.5)42.9 83/50 .0|27.3'80.0\87.5/83.3/66.7/50.0 62. | | jen eee | * Ag in tables 5 and 6 syndactyly is classified with the normal. 444 ROSS G. HARRISON half, though the small figures given in the table cannot be deemed significant in this direction. Moreover, this defect, -.ke that of syndactyly, has been found in cases which had not been oper- ated upon at all, and is probably to be regarded as due to slight general disturbances of growth. , On the other hand, the removal of the anterior portion seems to have a definite tendency to bring about reduplication, which is probably due to a more or less complete separation of the re- maining portion of the limb rudiment from a regenerative center in front of the wound scar. Notwithstanding these anomalies, the experiments speak as a whole for the equipotentiality of the four quadrants of the limb bud, at least in a qualitative sense. Quantitatively, the lines which divide the limb-forming tissue equally are anterior, and dorsal to the vertical and horizontal diameters, respectively, of the limb bud as defined by the experiments (fig. 2). These statements are valid for the free extremity only, and must be held in abeyance with respect to the shoulder girdle, where it is known that localization has taken place in the em- bryo at the time of operation (p. 429). While the prospective potency of the limb-forming cells is the same as regards the topographic divisions of the limb, the experi- ments give no evidence regarding histogenetic potencies. Whether certain cells at the operating stage are already differentiated into cartilage, bone, connective tissue, or muscle-forming ele- ments cannot be determined either by direct observation or by any of the experiments yet devised. What the prospective significance of the cells of each of the four quadrants of the limb bud is, 1.e., what part of the free ex- tremity is formed in normal development out of each, has not yet been determined, though the study of normal embryos points to the view that the distal part of the limb is developed more par- ticularly from cells lying in the posterior half, the ulnar half aris- ing from the dorsal and the radial from the ventral quadrant. It is difficult to devise experiments to test this hypothesis. Graft- ing of tissue colored by vital stains, such as neutral red and Nile blue sulphate, is not feasible because the stain is all de- EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 4495 posited in the ectoderm. A few experiments have been made by grafting markers in the form of pieces of notochord, which could readily be recognized, in the mesoderm of the different quadrants of the limb bud, in the hope that they might be lo- cated after the limb was developed. As yet these have resulted negatively. ATTEMPTED SPLITTING OF THE LIMB BUD Some years ago, in experimenting with larvae of the anuran, Pelobates, Tornier (’05) found that by making a deep incision through the hind-limb rudiment and the base of the tail he could produce double appendages. A few experiments have been made for the same purpose in connection with the present study. The material has necessi- tated, however, a rather different mode of operation, and the results have proved to be different. : The limb bud was deeply incised through the middle, either dorso-ventrally or antero-posteriorly, and a narrow strip of tissue including both ectoderm and mesoderm was removed. Thirteen experiments were made. Six were lost by accident eleven days after the operation, but all of them had at that time normal limbs on the operated side. The other cases were kept for sixteen or eighteen days and again all had normal limbs, though in four cases development was somewhat retarded. Nine of the embryos were in the oldest stage used (fig. 3) at the time of operation. The wound usually left a distinct scar or groove running across the limb bud, which, however, was obliterated after several days. In no ease did the operation result in redu- plications. The difference between these results and those of Tornier may be ascribed to the fact that in the case of the latter the operation was more radical and done upon older embryos, so that the chance of the divided limb rudiment healing together completely was much less. EFFECT OF SUPERIMPOSING LIMB BUDS In experiments upon early embryonic stages the most usual test of equipotentiality of the parts has been the development THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, No. 2 446 ROSS G. HARRISON of a whole organism out of any part of sufficient size. Another test, more difficult and less frequently tried, is the rearing of a single organism from two eggs or embryos which have been made to fuse together." The experiments described in the foregoing section have shown that any half of the limb bud can give rise to a whole limb. Those to be taken up now demonstrate that two limb buds fused together will develop into a normal single limb. The operation of superimposition or fusing together of two limb buds is carried out as follows: A circular incision of the proper size (83 segments in diameter) is made through the ecto- derm of the fore-limb region, care being taken to injure the underlying mesoderm as little as possible. The ectoderm may then be readily stripped from the middle layer by inserting the points of the scissors or needle at the dorsal part of the cut and tearing the ectoderm away. Often a few mesoderm cells, especially from the ventro-posterior quadrant, come off with the ectoderm, but the greater part of the cells composing the limb bud remain in place and not infrequently every cell is left intact. An entire limb bud from another embryo is then grafted in the usual way over the mesoderm of the limb thus exposed, and such grafts heal in very readily. The results of these experiments differ according to the orientation of the grafted bud, in har- mony with the rules of laterality (Harrison, ’17). At present only the cases in which the grafted bud has its normal orienta- tion will be considered. Five such experiments were made, all of the embryos surviving. and giving the same results. Normal limbs developed which at first showed difference in size, but this difference was after a time obliterated. The greater massiveness of the double bud was usually apparent the day after operation and was most marked about three or four days later. In two cases it is recorded as persisting for twelve days, though the difference from the nor- mal gradually diminishes, disappearing entirely by the time the yolk is entirely gone, i.e., about eighteen days. 11 Cf. Morgan, 795; Driesch, ’00, ’10; Goldfarh, ’14. EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 447 It is clear from these experiments, even though small in number, that a single perfectly normal limb will regularly develop out of two limb buds fused together. EFFECT OF REMOVING THE ECTODERM ALONE In the edrly days of the study of the limbs the question whether the first sign of their appearance was in ectoderm or mesoderm was much discussed as one of phylogenetic significance. What- ever our attitude toward such questions may how be, it is im- portant from the standpoint of developmental mechanics to in- quire whether the factor which determines the development of the limb is located in the outer or the middle germ layer.” Comparison of the experiments with cleaned and with un- cleaned wounds, in which the ectoderm is treated the same way in both series while the mesoderm is differently handled, already suggests what the answer will be. The experiments described +n this and the following sections are intended to add further and more direct evidence that the power to produce the limb rests primarily in the mesoderm. It is not always possible to avoid injuring the mesoderm to a slight extent in the operation of removing the ectoderm. There- fore the experiments 1n which it was attempted to remove the ectoderm alone differ only in degree from the cases of simple extirpation with non-cleaned wounds. Five operations of this kind were performed, and in all five cases limbs developed promptly, though there was some delay as compared with the normal. The wound was not covered with grafted skin in any of the cases. 2 Braus (09) has touched upon this question in connection with his study of the origin of the skleroblasts of the limb. Onp. 165 he says, ‘‘In einer spiteren Arbeit werde ich nachweisen koénnen, dass in ailteren Stadien ebenfalls das Ecto- derm nicht fiir das Zustandekommen eines typischen Skelets notwendig ist; denn dasselbe bildet sich eben so gut unter einem ortsfremden Ectoderm, welches experimentell gegen die typische Nachbarepidermis ausgetauscht wird, und in grossen Entfernung von jeglicher Haut wie in der ungestorter Entwicklung.” ~The paper dealing especially with this work (Braus, ’08) is unfortunately not accessible to me at present. 448 ROSS G. HARRISON EFFECT OF REMOVAL OF MESODERM ALONE In order to remove the mesoderm alone, the ectoderm covering the limb region is first incised around three-fourths of its cireum- ference. It is carefully lifted from the underlying mesoderm and left hanging by its ventral border. The mesoderm is then removed from the region below the pronephros, all loose cells being cleaned off, as in experiments already described, and then the covering layer is finally stretched back into place and held for a short time by a glass bar. The ectoderm contracts con- siderably while the wound is being cleaned, but with the aid of a fine needle it can usually be drawn over the wound. In some cases perfect healing*was obtained in less than an hour. In others small areas of yolk were found still uncovered on the day after the operation. The quickness of the healing seems, how- ever, to have no effect on the result, for in the three cases in which regeneration occurred, the healing was characterized as good, fair, and bad, respectively, while the cases of non-regen- eration followed both good and fair healing. Twelve experiments were made, in two of which the embryos died. In the ten remaining cases the wounds were of different sizes, varying from three to four segments in diameter, bounded as in the simple extirpations (p. 422). The results are given in ¢ table 8. In five cases the cleaned area was of the smallest size, ex- tending from the line between the second and third somites to that between the fifth and sixth. Three of these gave rise to regenerated limbs, while the other two did not. None of the other cases, which had larger wounds, regenerated.'* In six of the cases which showed no regeneration the pronephros was left intact and in only one case was it removed. These experiments differ from those described in the third section (p. 482) only with respect to the region from which the ectoderm covering the wound is taken. The results are in full agreement, and the corresponding figures given in table 4 and table 8 could with propriety be combined. 13 In my preliminary note (Harrison, ’15), on the fifth line from the bottom of p. 542, the words ‘over three’ should be substituted for ‘four.’ EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 449 TABLE 8 Showing effect of removal of limb mesderm CASES REGENERATED CASES NOT REGENERATED SIZE OF WOUND Healing of wound Healing of wound Total Total Good Fair Poor Good Fair Poor SISOMMICES* eae et ot es 1 1 1 3 2 2 3E* SOMIGES He er ee oe ee 0 1 1 Se SOMMLESH nae acters ee ees 0 1 1 2 ASSOMMGES Hae nee eee s ones 0 2 SING Geller ete x6 Perey cesta 1 1 1 3 3 4 7 * In the table previously published (Harrison, ’15, p. 540) in the first column fifth line read 32 instead of 33. While the number of cases is not large, it is beyond doubt that the presence of the normal ectoderm over the denuded limb re- gion no more incites the development of a limb than does the presence of ectoderm from a distant region. TRANSPLANTATION OF THE MESODERM Transplantation of the whole limb bud, ectoderm and meso- derm, results, as is well known, in the development of a limb in the new position. Transplantation of one or the other of the two layers should afford additional evidence, more cogent than that already given, as to the potency of the several layers in determining the development of the appendage. Only one of these experiments, the transplantation of the mesoderm, has been tried; the negative results recorded with reference to the ectoderm in the previous section are thought to be sufficient evidence from that side. In order to transplant or inoculate the mesoderm into some other region of the body, a pocket is first made under the skin of the embryo by sticking the points of a pair of fine scissors obliquely through the ectoderm and slightly opening them. The position chosen for this pocket was in most cases the flank of the embryo at the lower border of the muscle plates. In four eases, all of which resulted negatively, it was made above the 450 ROSS G. HARRISON eye. After the first embryo has been prepared, the ectoderm is removed from the limb region of another specimen, with as little injury as possible to the mesoderm, and the latter is cut out from below and behind the pronephros and transferred as a single piece to the pocket. It is often difficult to get this small mass of tissue inserted, because it is very sticky and is liable to be pulled out when the instrument is withdrawn. Having a small hole in the distal side of the pocket facilitates a deep insertion and consequently the retention of the transplanted cells. Healing of wounds of this character is rapid and without secondary complications. ‘The mesoderm cells are, however, rather loosely held together and a considerable amount of disintegration may occur—more than when the limb bud is transplanted in toto. The results of the experiments were as follows: Bmbryogdiedsorematureliys. ye. i cies ete Oe ee 4 Resorptionyotstransplanted tissue. \:0s.4.- ances eens ee eee 5 Small) nodulesdevelopedar vers .> a) MAN Ee Nk Ae Neel hale oye 3 Longrappendageswathout digits) 1:22. .) cess clean dace oe nee 1 Limb of approximately full size with digits, usually showing redupli- CALIOMeercperpenee earn tetas Ht ke nk CEN ele tyes thei asoue-al Meloc ayn ate 6 Wo) iW eA isle oc, i ae i eat Pe ae ee OR na WR 19 These results are not essentially different from those obtained when the whole limb bud is transplanted, except that the limbs which do develop when the mesoderm is taken alone are more likely to show deformities. This was to have been expected in view of the difficulties of handling the mesoderm without the firm ectodermal covering to hold it together. Since the indi- vidual cases are of interest their histories will be presented separately. Experiment Tr. Mes. 1. May 9, 1912. Mesoderm from left limb bud transplanted to left side. Some oozing from wound three hours afterward, indicating loss of tissue. May 12. Small lump in region of graft. May 15. Transplanted limb is getting much longer and is not much below the normal one in size. May 18. Beginning to show digitations. May 21. Larva has marked spinal curvature, but seems otherwise EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA A451 healthy. Transplanted limb has two well-marked digits and begin- ning of third. Limb not so long or so far advanced as the normal one. Good circulation. May 25. The specimen was preserved on account of the deformity of its back, which rendered its existence precarious (fig. 38). Transpanted limb is a left, having preserved its original laterality. It has a distinct third digit and the beginning of fourth, and on the radial side there is a reduplication of digits. The whole hand is, there- fore, nearly symmetrical. ‘Transverse sections show a small coracoid and a very shallow glenoid fossa, but the scapular portion of the girdle is not developed. Differentiated muscle fibers are present in the limb, but no nerves have been found. Experiment Tr. Mes. 4. May 11, 1912. Right limb mesoderm to right side. May 15. Grafted tissue not very prominent. May 21. Limb bud ‘points’ anteriorly. May 27. Digitations beginning; dorso-ventral doubling. June 7. Specimen preserved. The arm as a whole is a left, ie., its laterality has been reversed (figs. 39 and 40). The radial digit is reduplicated on the radial border. There is another reduplication consisting of a long digit and a nodule, mirrored from an ulno-palmar plane. In the normal limb on this side the first two digits are syndactylous, the first being short. Experiment Tr. Mes. 5. May 11, 1912. Mesoderm from right limb bud implanted on right flank. May 12. Wound still slightly open; transplanted tissue a good hump. May 21. Transplanted tissue has grown and points more distinctly anteriorly. May 27. Good circulation; two digits show; limb looks to be of normal form. June 7. Specimen preserved. Three well-marked digits are present with trace of fourth. The first two digits are webbed; otherwise the limb is normal (fig. 41). It is clearly a left, its original laterality having been reversed. There was no evidence of motility of the implanted limb before killing. The preserved specimen was cut into frontal sections, examination of which shows that the shoulder-girdle cartilage is fairly well formed; the ventral (coracoid) portion is more extensive than the scapula, which is only slightly developed. Pronephric tubules, seen near the base of the limb, indicate that part of the pronephros was transplanted with the limb cells. Muscle tissue is well developed in the limb, though no nerve fibers seem to be present. Experiment Tr. Mes. 16. May 18, 1914. Left limb mesoderm transplanted to right side. May 14. Perfectly healed; small nodule caused by transplanted tissue. ROSS G. HARRISON 452 = = EXPERIMENTS ON THE FORE LIMB OF AMBLYSTOMA 4053 May 20. Transplanted tissue growing well, ‘points’ anteriorly. May 22. Growth considerable; reduplication beginning near base. May 25. Limb consists of two almost equal parts branching near base; anterior member is bidigitate. June 1. i P \ s ' a ‘ ‘ . = T « ’ - ‘..3" ay 1 * i } 2 if j < 4 5 f o oft - ; ssi -_ - 2 bi é ~< i ay ome a4 Sei fe 7 a \ if, ‘f 4 it r. Al . } Pat] i ; pine ies) La M ® bfigh , 1! Oe ' a ‘ . ‘ ee a Tad eet Aa Ay Tan met Ub) Toe, aL viey eit asi ae ee ie ita 4th ba ie) ‘ ays hy } | : i i ® i aA Rie) me fie ay SrihMaaA La i" ’ il : tte — : Diver. cz } : n ff . re iy pri? cy) PEA eee. 4 Ge Pin «- i i lwastaet ies coi ht waar a” eS Phy ° a y o “a t! : a p ’ ‘s aa | aber; - i : - a Ve ‘ et 1 _ , , Hl a . , i | a =a r AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 30 EXPERIMENTS ON THE DEVELOPMENT OF THE. SHOULDER GIRDLE AND THE ANTERIOR LIMB OF AMBLYSTOMA PUNCTATUM S. R. DETWILER From the Osborn Zoélogical Laboratory, Yale University THIRTY-THREE FIGURES 1. INTRODUCTION Experiments on the transplantation of limbs have been car- ried on in connection with various problems: a) on the develop- ment of nerves (Banchi, ’05; Braus, ’05, and Harrison, ’07); b) on the question of orientation and laterality (Harrison, ’17) ; c) on the question of the development of the shoulder girdle (Braus, ’09), the rudiment of which is intimately bound up with that of the limb. It is with these last experiments that we are particularly concerned, since they have a more specific bearing on the results of the investigations set forth in this paper. Although the intimacy of these two systems—shoulder girdle and limb—during development and differentiation led Wieder- sheim (92) to conclude that girdle formation is possible only under the formative influence of the free extremity, the experi- ments of Braus showed, in part, the lack .of interdependence of these systems. Braus found that the removal of the fore-limb buds of Bombinator included the tissue from which the central — or glenoid portion of the girdle develops, and that only the distal parts—suprascapula and epicoracoid—were formed fol- lowing such an operation. The differentiation of these isolated girdle elements from unremoved blastema in the absence of the developing appendage demonstrated their independence of the free extremity. Confirmation of this observation was made on Amblystoma by Harrison (’18), who found, however, that as development proceeded these two separate distal elements gradu- 499 500 S. R. DETWILER ally approximated each other until, in a larva which was kept alive eighty-five days after operation, they had become united into a single cartilage. The formation of the suprascapula and coracoid in the absence -of the glenoid portion of the girdle demonstrates that their rudi- ments are already determined at the time of the operation, and that, while they eventually grow together, their unremoved rudi- ments are nevertheless not capable of restoring the missing parts, viz., the scapula, all or only a portion of the procoracoid, and the shoulder joint, the rudiments of which are removed in a typical limb-bud extirpation. Braus (09) further found that when a limb bud is transplanted to a heterotopic position, a complete shoulder girdle of one-third to two-thirds the size of the normal develops at the place of implantation. From this he concluded (page 271) that the shoulder-girdle rudiment constitutes an equipotential restitu- tion system. According to this conclusion, totipotency is restricted to those girdle-forming cells which become implanted along with the limb bud, for, as has already been pointed out, the unremoved blas- tema can develop only into those parts the rudiments of which are already determined at the time of operation. The forma- tion of a reduced girdle, with all its components, from cells which, in the normal environment, give rise to only the more central parts, would show that in their normal surroundings their prospective potency is greater than their prospective significance. The results of the experiments set forth in this paper seem to necessitate for Amblystoma, however, an interpretation dif- ferent from that which Braus placed on the results of his experiments. This investigation was taken up at the suggestion of Prof. R. G. Harrison. It gives me pleasure to express here my thanks to Dr. Harrison for the guidance he has given me during its completion. SHOULDER GIRDLE AND ANTERIOR LIMB 501 2. NORMAL ANATOMY In order that the experiments may be more fully understood, a description of the normal girdle will first be given. Chondri- fication of the girdle is practically complete in a larva about twenty days after the closure of the medullary folds. The girdle then consists of a cartilaginous structure lying within the body wall and extending from the lateral aspect of the third myo- tome almost to the mid-ventral line (fig. 6). It is made up of the following components: a) the suprascapula, which con- sists of a rod-shaped element lying external to the pronephros and constituting the greater portion of the dorsal zone (fig. 23, s.sc); b) the scapula, which lies just dorsal and anterior to the glenoid cavity and which, in the cartilaginous state, is continu- _ ous with the suprascapula (fig. 23, sc); ¢) the procoracoid lying immediately anterior and slightly ventral to the glenoid cavity (fig. 23, p.cor), and d) the coracoid, a relatively broad expanse of cartilage, constituting nearly the entire ventral zone of the girdle and reaching close to the mid-ventral line (fig. 7, cor. and fig. 23,, cor.). The scapula, procoracoid, and coracoid are con- tinuous proximally and enter into the formation of the glenoid cavity which receives the head of the humerus (fig. 23, gc). Chondrification Chondrification of the girdle proceeds gradually from the cen- tral portion towards the periphery. There are three centers, one for the scapula, one for the coracoid, and one for the procoracoid. The center for the scapula is first to appear. This is followed by the center for the coracoid and finally by the procoracoid center. This was found by Wiedersheim (’89) to be the case with Triton, Siredon, and Salamandra. The same observation was also made by Braus (09) on Bombinator. The union of these three centers completes the chondrification of the central portion of the girdle. The suprascapula has no separate center and chondrification of this element proceeds gradually from the region of the scapula in a dorsal direction. 502 S. R. DETWILER While the central part of the girdle is chondrified before the first two digits of the fore limb are fully formed and before the elbow joint: becomes visible, the dorsal portion of the suprascapula is not entirely chondrified until the fourth digit makes its ap- pearance. The chondrification of the coracoid likewise proceeds gradually from its center towards the mid-ventral line. These observations agree with those of Braus (’09) on Bombi- nator. In this form there are no chondrification centers for the suprascapula and the epicoracoid. The epicoracoid of Bom- binator is homologous with the ventral portion of the coracoid in Amblystoma, which, as has been pointed out, chondrifies gradu- ally from the proximal part towards the periphery. The cartilage center for the humerus appears somewhat earlier than do those for the girdle. Considering the Amphibia as a whole, it can be said that in most cases this is true (Wiedersheim, ’89, ’90, ’92; Lignitz, ’97, and Braus, 709). From Strasser’s (’79) description of Triton, one would assume, however, that in this form initial chondrification of the humerus and the girdle takes place simultaneously. The centers for the ulna and radius appear slightly later than those for the girdle, but they are completely chondrified before chondrification of the suprascapula and the coracoid have been completed. The greater part of the girdle remains cartilaginous throughout life, but the entire scapula and those portions of the procoracoid and coracoid which enter into the formation of the glenoid cavity become ossified. ‘The cartilaginous suprascapula which, in the larva, is a long slender rod-shaped structure, elongates in an antero-posterior direction so as to become a broad flat plate. The procoracoid grows out in an antero-ventral direction and becomes a structure very similar in shape to the procoracoid of Necturus. The coracoid, which comprises the greater part of the ventral zone of the girdle, is a large flat rounded plate of cartilage lying ventral and posterior to the procoracoid. The two coracoids overlap in the mid-ventral line. The shape of the ventral portion of the adult girdle is very similar to that figured by Furbringer (73) for Salamandra maculata. SHOULDER GIRDLE AND ANTERIOR LIMB 503 No description of the shoulder muscles of Amblystoma could be found in the literature. The musculature, however, so far as has been studied, closely resembles that of Salamandra macu- lata, a European tailed Amphibian described by Fiirbringer (op. cit.). In referring to the musculature, Fiirbringer’s nomencla- ture will be employed. 3. EXPERIMENTAL The experiments were carried out upon embryos in two dif- ferent stages: a) the so-called tail-bud stage, and b) the stage of open medullary folds. Fig. 1 Camera-lucida drawing of an embryo of Amblystoma in the tail-bud stage. The larger of the two circles represents the typical limb dise. X 15. pn, = pronephros. 1. EXPERIMENTS ON EMBRYOS IN THE TAIL-BUD STAGE (STAGE 29) A. Extirpation experiments 1. Removal of the suprascapula rudiment. As has already been pointed out (Harrison, 715), the fore-limb rudiment of an embryo in the tail-bud stage consists of a somatopleural thick- ening just ventral to the pronephros, centering in the region of the fourth myotome and extending over into that of the third and fifth. The formation of a suprascapula following the extirpation of the limb rudiment in this stage shows that its rudiment is not included with the limb mesoderm. Although there is no visible suprascapula rudiment, nevertheless, extirpation of the region a-e X 1-38 (text fig. 1), including the outer or cutis layer of the 504 S. R. DETWILER TABLE 1 Showing the results following the removal of the area a-e X 1-3 (text fig. 1) including the outer cell layer of the somites and the pronephros Leos CONDITION OF THE GIRDLE AND THE EXTREMITY INDIVIDUAL} AFTER eile Suprascapula Scapula Procoracoid Coracoid Humerus 1h Seacooe 26 absent present present present present Ba 10....| 24 absent* present present present present IRA chr 26 absent present present present present BG aeees 26 absent present present present present * Only dorsal half wanting. ventral halves of the somites and the pronephros, results in the formation of a girdle without a suprascapula (fig. 24 and table 1). The removal of this area has practically no effect on the de- velopment of the extremity itself and of the remainder of the girdle. Abnormalities in the limb sometimes occur if its rudi- ment is disturbed to any great extent during the operation. 2. Removal of the suprascapula rudiment and of the myotomes. This type of experiment consisted in the entire removal of the third, fourth, and fifth somites and the pronephros. In these experiments the wound was usually covered with ectoderm taken from a second embryo, since complete exposure of the notochord and the medullary tube frequently results in a disintegration of the embryo. Removal of the entire somites produces the same effect on the girdle as that described for the first type of experi- ment, viz., the formation of a girdle without a suprascapula (figs. 8 and 25 and table 2). In the absence of the somites, the limb and the girdle undergo a dorsal shifting (fig. 8). This is apparently due to a release of pressure from that direction. Complete removal of the third, fourth, and fifth somites, while leading to no defects in the limb musculature, brings about marked deficiencies in the ventro-lateral musculature, ob- servations corroborating those previously made by Miss Byrnes (98) and Lewis (10). Miss Byrnes (op. cit.) produced the first experimental evidence to show that the musculature of the an- terior limb of Amblystoma develops in situ and that it is in no way derived from the myotomes or their ventral processes. SHOULDER GIRDLE AND ANTERIOR LIMB 905 TABLE 2 Showing the results following complete removal of the third, fourth, and fifth somites and the pronephros as CONDITION OF THE GIRDLE AND THE EXTREMITY INDIVIDUAL} AFTER | Pao Suprascapula Scapula Procoracoid Coracoid Humerus Ba 13....|' 23 | absent present present present present Ba 11: 23 | absent present present present undifferen- tiated Exel 26 absent present present present absent Di (eee ae 27 | dorsal half} present present present present absent The absence of the limb in case Ex 1 is no doubt due to the destruction of its rudiment in the removal of the pronephros. Lewis (op. cit.), however, in addition to demonstrating this same fact, showed that definite defects in the ventro-lateral muscu- lature follow the removal of the myotomes of the limb region. An examination of figure 8 will show that, in this case, there is complete absence of the ventro-lateral musculature on the oper- ated side. This is, however, not always the case, for in others in which these same myotomes were completely excised, the ventro- lateral musculature is partly filled in by a compensatory elonga- tion of the derivatives of intact myotomes. This same observa- _tion was also made by Lewis. In the absence of the suprascapula, the m. dorsalis scapulae, which normally runs from the proximal end of the humerus over the external surface of the suprascapula (fig. 7, m.dsc), now attaches to the scapula (fig. 8). It is essential, in order to remove successfully the supra- scapula rudiment to include the pronephros. Its removal fa- cilitates cleaning of the wound and thus affects indirectly the results of the experiment even though it exerts no direct influ- ence. ‘This applies as well to removal of the limb bud as a whole (Harrison, 715). Since removal of the outer portions of the third, fourth, and fifth somites suppresses development of the suprascapula and since there is‘no visible rudiment for this element at the time of the operation, the experiments show that the suprascapula is 506 Ss. R. DETWILER derived from tissue other than that which gives rise to the limb and the remainder of the girdle. Sections of older embryos show that dorsal to the pronephros scattered mesenchyme cells gradually appear and become continuous, external to the pro- nephros, with the limb-forming cells. It is evident that this mesenchyme, which later forms the suprascapula, is segregated from the outer or cutis layer of the somites in this region and that the suprascapula is formed in situ. Such a conclusion is strengthened by the fact that after complete removal of the limb rudiment and the pronephros, the suprascapula develops in its normal place, provided the third, fourth, and fifth somites are undisturbed. Not only do these experiments show that the suprascapula is already determined at the time of the operation and that it is formed in situ, but they demonstrate as well the impotency of - the unremoved girdle tissue to replace the missing part. 3. Removal of the dorsal zone rudiment of the girdle and the limb mesoderm. ‘This series of experiments consisted of the re- moval of the area a-e X 1-5 (text fig. 1). This included the outer portion of the ventral halves of the somites, the pronephros, and the limb mesoderm. The wounds were cleaned and covered. The removal of this area suppresses development of the supra- scapula, scapula, and the free extremity, and only the ventral half of the girdle develops, no glenoid cavity being formed in any of these cases (figs. 9 and 26 and table 3). The formation in situ of the procoracoid and coracoid is evidence that they, too, are already determined at the time of the operation and are not dependent for differentiation on the remainder of the girdle and the limb, the rudiments of which were removed in this type of experiment. Although the myotomes proper were left intact, several larvae showed slight defects in the ventro-lateral mus- culature. This is no doubt due to a partial injury of the ventral portions which furnish the muscle buds. The rudiments of practically all the shoulder musculature are included in these extirpations. In one case a few partially de- veloped muscle fibers were present just external to the unre- moved girdle elements. They probably represent the m. supra- SHOULDER GIRDLE AND ANTERIOR LIMB 507 TABLE 3 ; Showing the results following the removal of the area a-e X 1-5 (text fig. 1) including the outer portion of the somites, the pronephros, and the limb mesoderm CONDITION OF THE GIRDLE AND THE LIMB INDIVIDUAL AGE Suprascapula Scapula Procoracoid Coracoid Humerus days RSG ae 26 | absent absent present present absent 1) Bestove 26 | absent absent present present absent Ries 26 | dorsal por- absent present present absent tion present 1 WO. cope 28 | absent absent present present absent xed fase 28 | absent absent present present absent Rade ce a 26 | small nod- absent present present absent ule of cartilage Incomplete absence of the suprascapula in cases R 1 and R 2 is apparently due to imperfect removal of the rudiment. coracoideus, a muscle which normally runs from the proximal end of the humerus to the external surface of the coracoid (fig. 7, m.spc). 4. Limb-bud extirpations. The effects of the removal of a typical limb disc on the girdle (text fig. 1) are in accord with those described by Harrison (’18). In individual H 2 sectioned twenty-two days after the operation, only the suprascapula, a very small procoracoid, and the coracoid were present (fig. 10 and table 4). In another, H 5, only the dorsal part of the suprascapula was present in addition to a fragmentary procor- acoid and the coracoid (fig. 27 and table 4). In these cases the pronephros was removed with the limb bud, and the incomplete- ness of the suprascapula in the second case is no doubt due to a partial destruction of its rudiment in the removal of the pro- nephros. Since no limbs developed in experiment 3 after the removal of the area a-e X 1-5, it is obvious that the ventral portion of a typical limb dise (Harrison, 715 and ’18) which is shown in text figure 1 contains only girdle-forming cells. The formation of only a portion of the ventral zone of the 508 S. R. DETWILER TABLE 4 Showing the effects on the girdle of removal of the limb disc CONDITION OF THE GIRDLE INDIVIDUAL AGE Suprascapula Scapula Procoracoid Coracoid Shoulder joint days Ee eee 20 present absent |present present absent Ee OReecee 20 |* present* absent |fragmentary| present absent * Dorsal half only. girdle following a typical limb-bud extirpation not only indicates that part of its rudiment is removed with the limb bud, but that the part which is unremoved is already determined at the time of the operation. Further, when a limb bud is transplanted to a heterotopic position, the development of a girdle with a ventral zone of reduced size (fig. 28) also serves to indicate that only a portion of the rudiment is transferred with the limb cells. If localization of the cells which are to form the ventral zone of the girdle is complete at the time of the operation and they can be successfully extirpated, then only the dorsal zone should develop following their excision. =e aells > eis eee i 542 C. Solid yellow and yellow-spotting.............-.-.5+----+++2seees 544 Designee sexcilltagiOmbenee cee ere rae eee ie sack cyniceen apteccae tereteree, ae aeuere 549 By artic ovsann he (ke kota ons 8, eae ote k= ata oe wane se a hae Spee 550 a. Statement of factorial differences and description of char- BUC LOS Loy a ANON WN eee eee a ers Sh aia ae eS eee cae OD Doub perimental yd vai! s i) eee cis nse so acege selon ae 554 e. The aumber of loci mvolveds......>. -<..: s0.tatete so sens 22 OOS d. Physiology of color-production.............0..+++.++++++++-+ 558 Ill. The origin of color varieties of the cat................+.-.-.- Aa hee 560 LAY ES Priest ty ec = Aaa ea ree Se ee i ers cane Sear se Are 562 Viniterature Clie cde. mene cise Cte Mais ney cg eons Sicreuet ct ain, crale asters ererbeckelepere le 564 I. INTRODUCTION: THE COLOR FACTORS OF DOMESTIC CATS In a series of experiments begun at the University of Penn- sylvania in the autumn of 1914 and extending up to the present time, the inheritance of coat-color in cats has been investigated. Although the number of litters obtained has not been large, it has been found possible to determine several points in regard to the mechanism of heredity by means of critical crosses. This has been largely due to the fact that the characters studied seg- regate for the most part cleanly from each other so that it has been easy to classify the animals. My thanks are due to Dr. McClung and to Dr. Colton for the kindly interest which they have taken in the work and to the 939 540 P. W. WHITING University of Pennsylvania for the expense of the experiments. I also wish to thank the Zoological Society of Philadelphia for the opportunity of crossing my cats with the Caffre cat. Before presenting the data and discussing the inheritance of the various characters in detail it may be well to name and to define briefly the factorial differences involved. The ticking or agouti series contains at least two and probably more different factors. Black, non-agouti, or uniformity, a, is a recessive allelomorph to agouti, A. An extreme amount of ticking is frequently observed, and this represents, in all proba- bility, a third allelomorph in the series. Results are not con- elusive, but are consistent with such a hypothesis. This ex- treme ticking may tentatively be called A’. The banding or tabby series contains at least two and probably three different allelomorphs. They may be called lined, B’, striped, 8, and blotched, 6. The allelomorphism of these is again merely tentatively assumed, but I believe there is good reason to consider them as a triple allelomorphic series. The agouti and the banding factors will be discussed fully later in this paper. It may merely be stated here that all cats are either lined, striped, or blotched and that there are no in- tergrades. The character of the banding is, however, easily recognized only in the presence of ticking. In the black or non-agouti series banding exists as ‘ghost patterns’ recognizable usually but not always in the fur of kittens, and only occasionally distinguishable in full-grown cats. Intense pigmentation, MV, is a simple dominant over dilute or maltese, m. White, W, is a simple dominant over color, w. Yellow, Y, is allelomorphie with black, y. This locus is sex- linked and shows the male to be digametic. The heterozygous female is wsually yellow-spotted or tortoiseshell, but may range in color to solid black or solid yellow. White-spotting is very irregular and probably depends upon many factors. It will not be considered in detail in the present paper. It appears to segregate independently of the factors above mentioned. INHERITANCE OF COAT-COLOR IN CATS 541 Albinism, as we have it in the rodents, is apparently un’ nown in eats. It is possible that Siamese dilution represents an ap- proach toward this condition. Variations in the tone of coloration are extensive, but appar- ently not clearly segregating. Silvers represent a reduction of yellow pigment and also of black. Smokes are very dark sil- vers. The lighter bands of tabbies are straw- or cream-colored, varying to white in silver tabbies and brown in brown tabbies. Occasionally the brown varies to a rusty red. Silver creams are yellow cats in which the yellow pigment is reduced to a mini- mum so that the hair sometimes appears almost white. Accord- ing to fanciers, silver tabbies bred together occasionally throw brown tabbies. Il. PRESENTATION AND DISCUSSION OF DATA A. Maltese dilution Maltese dilution appears to be a simple Mendelian recessive. It apparently exists in combination with all other factorial dif- ferences, but I have not as yet seen its representative in the lined or narrow type of banding. It is always sharply distin- guishable from black, but shows considerable variation in its own intensity. It is to be compared to slaty-blue in the mouse, the rabbit, and the dog. No corresponding color is known in the rat or in the guinea-pig. Cream or dull yellow is its corre- sponding color in the yellow series; blue and cream, in the tortoiseshell. Five litters from dilute by dilute gave sixteen dilutes—eight males and eight females. | Six litters from intense by intense gave twenty-four intense— sixteen males and eight females. Five litters from intense males by dilute females gave five intense males, nine intense females, and one dilute male. Kight litters from dilute males by intense females gave nine intense males, three intense females, nine dilute males, and eleven dilute females. These data show merely that maltese dilution is not sex- linked. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. 2 542 P. W. WHITING B. White and white-spotting White-spotting in cats is exceedingly irregular in amount and distribution, but tends to appear more commonly on the under parts. There appears to be no regularity in dominance and probably many factors are involved. The degree of white- spotting in the parents tends to appear again in the offspring, although wide segregation occurs. I have usually selected cats with relatively small amount of white and have obtained kittens of the same general character. Solid white appears to be a complete dominant over color whether the color is self or spotted with white. It is possible that it may be allelomorphie with one or more of the white- spotting factors, but my data are not conclusive on this point. Castle (16) regards it as possibly an extreme form of white- spotting. A cream male (8)! with slight amount of white crossed to a blue-eyed deaf white female (7) sired (6) three pure white kit- tens; one male and two females. The kittens had normal hear- ing and developed yellow eyes. The male had a slight smutti- ness of the hairs on top of the head which appeared when he was two weeks old and then shortly disappeared. The females never developed any pigment in the hair. The same pair of cats mated again and produced two solid white kittens which were not reared. | The cream male (8) was later mated (32) to a yellow-eyed white with normal hearing (22). Two whites, a male and a female, were produced, as well as two females that were entirely self color, a blue and cream and a cream. I was unable to de- tect any white hairs on these two pigmented kittens. It would appear, then, that the white mother carried factors-which domi- nated the shght white marking of the father, and thus produced totally self-colored kittens. ' Individual numbers have been inserted after mention of any animal that is referred to more than once in this paper. Matings have been numbered simi- larly. It will thus be possible for the reader to check up the genetic composition of any animal by its progenies from various matings. INHERITANCE OF COAT-COLOR IN CATS 543 Dr. Little reports a mating (43) of a yellow-eyed white male by a cream and white female. A single yellow-eyed white female was produced. A mating (35) of a blue-eyed white Angora male (16) by a yellow-eyed short-haired white female produced three solid whites and a black that was self-colored except for very small white spots between the legs. A mating (36) of the same male by a maltese female produced a solid white and a near self black. Besides these I have three records of blue-eyed white females which produced both colored and white offspring. The male parents were unknown. One produced a solid white yellow-eyed female and a solid black male. One produced in two litters six solid white and three almost entirely black. The third produced a solid white and an intense striped tabby with belly, nose, breast, and feet white. Davenport (05) reports a mating of a black Manx male by a blue-eyed, deaf, white female, which produced two whites, one black, one tiger, and one maltese. The interesting correlation of blue eyes and deafness with white coat is not yet satisfactorily explained. Dominance of eye color seems very irregular. I am informed by breeders of white cats that yellow-eyed by yellow-eyed may produce blue-eyed, and also that blue by blue may produce yellow. Odd-eyed eats also frequently appear in these crosses. - Przibram (’08) reports experiments with odd-eyed white An- gora cats. Results were very irregular. The cats bred true to whiteness. As regards correlation of deafness and blue eyes, he Says, Von Interesse ist es, dass, soweit eine Priifung des Gehors unter- nommen werden konnte, die blau-blauen erwachsenen Katzen alle vollig taub waren, was ja mit friiheren Beobachtungen von Darwin und Rawitz iibereinstimmt; die asymmetrischen Augenfarben scheinen dem ganz entsprechend mit einer halbseitigen, die Seite des blauen Auges betreffenden Taubheit betroffen zu sein. Die Correlation zwischen blauen Augen und Taubheit bleibt also auch bei der asym- metrischen Vererbung bestehen. Dabei bleibt die Correlation zwischen ‘blauen Augen’ der Katzen und ‘Taubheit’ auch fir die Kérperhalfe bestehen. 044. P. W. WHITING Dr. C. C. Little showed me a black and white cat with odd eyes. The hair surrounding the blue eye was white, while that about the yellow eye was black. Blue eyes in pigmented cats are rare, except of course in the case of the Siamese. I would suggest, therefore, as a working hypothesis that the in- cidence of white-spotting in connection with the dominant white factor, produces the blue eye, or in other words a ‘white spot’ about the eye of a white cat makes the eye blue, while a ‘pig- mented spot’ about the eye of a solid white cat makes the eye yellow. It may be also that a ‘white spot’ in the ear of a white cat makes it deaf. This would explain why it is so difficult to get blue-eyed white cats with normal hearing as it would be difficult to localize the ‘white spot’ upon the eye and to keep it away from the ear. This may also explain why odd-eyed cats are frequently defective in hearing only on the side having the blue eye, as noted by Przibram. It would not be a difficult matter to test this hypothesis. C. Solid yellow and yellow-spotting The tortoiseshell cat has been the subject of much interest and discussion in genetic literature dealing with sex-correlated phenomena. Doncaster (’05) considered the problem and tried to explain the peculiarities of inheritance by variations in dominance. Little (12) suggested the hypothesis of a single sex-linked pair of allelomorphs with the male digametic. He used the term ‘sex-limited character,’ which has since been restricted to simple Mendelian heredity in which sex reverses the dominance of the allelomorphs. Doneaster (712) accepted Little’s suggestion as in general satisfactory, but pointed out that occasionally black females are produced from matings of black female by yellow male. According to Little’s hypothesis, the females should always be tortoiseshell from the reciprocals of black by yellow and the males should be like the mother, disregarding of course dilution, tabby, ete. Doncaster suggests an occasional break in sex-linkage to explain these anomalous blacks, as also the INHERITANCE OF COAT-COLOR IN CATS 545 occurrence of the rare tortoiseshell male. In 1913 he gave a further discussion of the subject and an excellent summary of data collected from fancy breeders. In 1914 he suggested non- disjunction of the sex-chromosomes in oégenesis to explain the matroclinous black females. These explanations are all more or less unsatisfactory for one reason or another, as admitted by Doncaster and by Little. I have pointed out (1915) that the hypothesis of simple sex- linkage first suggested by Little may be sufficient to account for the conditions if it be considered that the heterozygotes, which must be females, vary from black through various degrees of yellow-spotting to solid yellow. In the male, presumably, con- ditions are much more stable, as it is impossible to have a het- erozygote. Thus I have suggested that a gametically yellow male (YX —) may become tortoiseshell by extreme selection of black extension factors, while a gametically black male (yX —) may become tortoiseshell by an extreme selection of yellow ex- tension factors. The possibility is of course not excluded that there may be a single factor or particular combination of factors that produces yellow-spotting in the male. Ibsen (716) has suggested close coupling of two pairs of sex- linked allelomorphs, and attempts to explain anomalies by cross- ing-over. This does not account for all the results, however, as he himself points out. The data concerning the tortoiseshell problem which I have gained from my experiments are as follows: A long-haired cream male (8) (a.b.m.Y)? was crossed (12) with an intense strippd tabby (25) (A’.B.M.y). There resulted one maltese (a.B.m.y) and two intense striped tabby (A’.B.M.y) males and one blue and cream (a.B.m.Yy) female. The same male was crossed twice (9 and 30) to a black (20) (a.B.M.y). There resulted three black (a.B.M.y) and two mal- tese (a.B.m.y) males and two orange and black (a.B.M.Yy) and three blue and cream (a.B.m.Yy) females. 2 The genetic formula includes only factors known to be present. The duplex condition is not expressed except in the case of the sex-linked pair Y—y in which dominance is variable. 546 P. W. WHITING The same male was crossed (33) to an intense striped tabby (3) (A.B.M.y). There resulted one intense striped tabby (A.B.M.y) male and one blue and cream (a.b.m.Yy) female. The same male was crossed (13) to a maltese (17) (a.b.m.y). There resulted one maltese (a.b.m.y) male and two blue and cream (a.b.m.Yy) females. These blue and cream females are of interest as they are almost anomalous blacks. They were men- tioned by me in a previous paper (1915, p. 519). They are almost entirely dilute blacks except for a very restricted amount of white on the under parts. One has a very small amount of cream bordering the white at a few points, and a small cream spot on the back. The other is entirely without cream except for a few cream hairs on one leg and at the tip of the tail. These kittens then represent a very close approximation to the anum- alous black. A short-haired cream male (24) (A’.b.m.Y) was crossed twice (34 and 49) with an intense striped tabby (3) (A.B.M.y). There resulted five males—three maltese (2 a.B.m.y and 1 a.b.m.y), one black (a.B.M.y), and one intense striped tabby (A’.B.M.y) ; and four females—two blue and cream (a.b.m.Yy), one orange and blotched (A.b.M.Yy), and one blue and cream striped (A’.B.m.Yy). Summarizing matings of ‘yellow’ male (YX —) by ‘black’ female (yX yX), we have seven matings giving fifteen ‘black’ males (yX —) and thirteen ‘tortoiseshell’ females (YX yX). Crosses of yellow male by tortoiseshell female are as follows: A cross (37) of an orange striped male (B.M.Y) by a blue and cream female (26) (a.b.m.Yy) produced two orange-striped females (B.M.Y). A short-haired cream male (24) (A’.b.m.Y) was crossed (44) with an orange and black female (28) (a.M.Yy). There re- sulted one cream (m.Y) male and three females—two cream (m.YY) and one blue and cream (a.m.Yy). The same male was crossed (45) with another orange and black female (30) (B.M.Yy). There resulted three males—one black (a.M.y), one blotched maltese (A’.b.m.y), and one cream (Ban. Y ). INHERITANCE OF COAT-COLOR IN CATS 547 The same male was crossed (47) with a blotched blue and cream (13) (A.b.m.Yy). There resulted two males—one maltese (a.b.m.y) and one cream (b.m.Y); and three blotched blue and cream (A.b.m.Yy) females. Summarizing crosses of ‘yellow’ males (YX —) by ‘tortoise- shell’ females (YX yX), we have four matings giving three ‘black’ males (yX —), three ‘yellow’ males (YX —), four ‘yellow’ females (YX YX), and four ‘tortoiseshell’ females (YOST DO The long-haired cream male (8) (a.b.m.Y) mentioned above was crossed (32) to a yellow-eyed white cat (22). There were produced one white male, one white female, one cream female (B.m.YY), and one blue and cream female (a.B.m.Yy). Since a cream kitten as well as a blue and cream was produced, it is probable that the yellow-eyed white was gametically a tortoise- shell. Dr. C. C. Little has very kindly supplied me with data in regard to an anomalous cream female (23) which breeds like a tortoiseshell. This female he has given to me along with three of its offspring. She has, while in my possession, produced four kittens by her cream son. They are two cream females, one cream male, and one maltese male. The maltese male would of course not be expected from a mating of two yellows. The un- der parts of the anomalous female are white. The upper parts are entirely cream and show the blotched pattern very plainly. No trace of black pigment can be found, although I have exam- ined samples of the hair from various parts of the body under the microscope. The following gives in detail the offspring from this anoma- lous cream female. The unexpected progeny are recorded in italics. When crossed with (42) a dilute blotched male (A.b.m.y), she produced two males—one blotched maltese (A.b.m.y) and one maltese (a.m.y); and one blotched blue and cream female (A.b.m.Yy). When crossed with (39 and 40) an intense striped male {A.B.M.y), she produced in two litters two males—one orange 548 P. W. WHITING (M.Y) and one cream (24) (A’.b.m.Y); and two females—one black (a.M.yy) and one orange and black (a.M.Yy). When crossed with (41, 46, and 50) her cream son (24) (A’.b.m.Y) she produced in three litters three males—one blotched maltese (A.b.m.y), one maltese (a.b.m.y), and one cream (m.Y); and two cream females (m.YY). When crossed with (38) an orange-striped male (B.M.Y), she produced one black male (a.M.y) and three females—two striped orange (B.M.YY) and one striped orange and black (a.B.M.Yy). Summarizing the matings of this cream and white female, we find that when bred to ‘black’ males (yX —) she produced two ‘black’ males (yX —), two ‘yellow’ males (YX —), two tortoise- shell’ females (YX yX), and one black female (yX yX). When bred to ‘yellow’ males (YX —), she produced three ‘black’ males (yX —), one ‘yellow’ male (YX —), four ‘yellow’ females (YX YX), and one tortoiseshell female (YX yX). Of these matings Dr. Little says: ‘‘ The dilute yellow and white female is interesting because she forms gametes carrying black and breeds exactly like a dilute tortoiseshell and white animal, although there is no trace of black pigment anywhere on her.’ She is then an anomalous yellow. Dr. Little further states: “‘ Dilute yellow, like the same color in mice, does not depend upon the depth of color, but is essentially a dull yellow ranging anywhere from intense pigmentation to dilute cream color.” It is of course relatively not as intense as the orange. It is possible that this variation in cream color is due to the same factors which produce the variations towards silver in tabbies and others. Yellow-spotting in cats may be compared essentially to the same condition in guinea-pigs. In the latter there is great range of variability asin the former. In cats, however, one of the allelomorphic pairs determining black or yellow extension is much more potent than the others and is sex-linked. The het- erozygous female (YX yX) represents a much more unstable condition as regards spotting than either of the homozygous females or than either of the haploid males, for in the heterozy- gote the factors yellow, Y, and black, y, are balanced against each other. INHERITANCE OF COAT-COLOR IN CATS 549 The sterility of the tortoiseshell tom has frequently been re- marked upon. Cutler and Doncaster (15) discuss this question and show drawings of sections of the testis of a sterile cat of this sort. Normal reproductive cells are altogether lacking. In summarizing the data on sterility of male tortoiseshells, they find that one was certainly fertile, two completely sterile, one almost if not quite sterile, and two doubtful. It appears, then, that sterility may be highly correlated with yellow-spotting in the male. The black-yellow allelomorphic pair in cats is of particular interest, as it is the only case of sex-linkage known in mam- mals, other than the sex-linked defects of man. D. Siamese dilution Bateson (’13) says of the Siamese cats: ‘“‘ These animals, which breed perfectly true, were introduced from Siam, where they have been kept for an indefinite period as pets of the royal household. Like the Himalayan rabbit, Siamese cats are born almost white, but the fur becomes a curious fawn with darker chocolate points on the ears and extremities.’’ Crosses of Sia- mese by other cats are cited by Weir (89). Quoting from a Mr. Young, he says (p. 76), “‘They invariably showed the Siamese cross in the ground color.” But Lady Dorothy Nevill says, ‘“None showed any trace of the Siamese, being all tabby.” Two pregnant females of common cats brought into the lab- oratory produced kittens of a peculiar ashy color with darker extremities. The kittens resembled very closely adult Siamese cats. One of the pregnant females, a maltese (5) (a.m.) produced (15) two females which were ashy, with nose, ears, feet, and tail slightly darker, and two females and two males which were ashy with black extremities. A record taken fifty days after birth showed that the lighter kittens had become maltese, while the kittens with black extremities had become steel colored or almost black. They later became completely black. Ghost patterns were seen on four of the kittens, but unfortunately a critical 550 P. W. WHITING examination was not made of the other two. Ghost blotched was very much accentuated by the ashiness, but ghost striped did not appear especially so. It is probable for this reason that the two doubtful ones were ghost striped. Of the other four the maltese was ghost striped, and the three steels were ghost blotched. The other pregnant female (2) was a black and white (a.M). She produced (4) four kittens—two black females upon which no record of ghost pattern was made, an ashy female with black extremities, and an ashy male with dark but not black extremi- ties. Sixty-six days after birth the ashy female had developed into a steel black which clearly showed ghost-blotched pattern, and eighty days after birth the ashy male was maltese with ghost-blotched pattern very evident. Unfortunately, the inheritance of this peculiar ashy color could not be followed out at the time the kittens were on hand. I am, therefore, unable to say whether it represents the hetero- zygote for Siamese dilution. E. Banding and ticking a. Statement of factorial differences and description of characters. Ticking or agouti in cats, as in rodents, is characterized by yel- low bands on the hairs. It increases with age so that kittens are relatively less ticked than cats. I have tentatively considered the agouti factors as a series of triple allelomorphs—A’, much ticked, A, little ticked, and a, non-ticked, with dominance of A’ over A and a, and of A over a. The banding factors, I have also represented as triple allelo- morphs—B’, lined, B, striped, and b, blotched. These factors affect the formation of yellow pigment, in a yellow cat (A or a.B.M.Y) forming bands of straw color alternating with orange. In a tortoiseshell tabby cat (A.B.M.Yy) the orange bands in the ‘yellow spots’ are continuous with the black bands in the ‘black spots,’ while the straw-colored bands are continuous through both regions. In a tortoiseshell (a.B.M.Yy) alternate banding of straw and orange shows clearly in the ‘yellow spots,’ while the INHERITANCE OF COAT-COLOR IN CATS 551 ‘black spots’ are uniform black. The same condition obtains in the case of maltese dilution, but the contrast in the bands is not as obvious and there is general reduction in the amount of yellow pigment. Uniformity or lack of banding in yellow cats is apparently due, as has been pointed out to me by Dr. Sewall Wright, to some other condition than the lack of the agouti factor. As regards the existence of such cats, Mrs. Leslie Williams (’08) writes: “The self-orange Persian is more of an ideal than a reality, for it is actually a red tabby without the tabby markings, and at present it is a case of ‘more or less,’ the upshot being that the least marked cat in the class takes the prize.” Silvering is a general reduction in the amount of yellow pig- ment. The straw bands of tabbies then become white. Figure A shows a silver-striped tabby skin. Black stripes alternate with white. In the skin shown in figure B, on the other hand, there is a considerable amount of yellow pigment. A striped tabby (9) brought into the laboratory pregnant had lighter bands of a decidedly reddish color. This apparently represents the opposite extreme of variation from silvering. Intense black stripes alternated with rusty red. She gave birth (22) to three male kittens—one striped with black and red; one blotched with black and red, and one striped with black and straw color. Here, then, is an indication that the extreme reddish tone is hereditary. For an understanding of banding we may first consider figure B. The skin shown here is from a striped tabby male (7.2) forty-two days old. It may be seen that the bands run longi- tudinally along the back and are most easily seen near the mid- dorsal line posteriorly. On the sides the bands are transverse and tend to be broken into spots. We may think of this condi- tion as having been produced by longitudinal and transverse waves of pigment-forming metabolic activity. The longitudinal waves form transverse bands. The areas of greatest activity form orange bands in yellow cats, while in tabbies these bands are black. The areas of less activity form, of course, the lighter bands. The transverse waves appear to originate at the mid- dorsal line. They form longitudinal bands on the back. As 552 P. W. WHITING they pass outward and down the sides, the areas of greater ac- tivity tend to thicken the transverse bands. In the areas of less activity the transverse bands are often evanescent. It thus ap- pears that black or orange spots, in tabbies or yellows, respec- tively, are produced in the regions of greatest metabolic activity. The ticking and the banding factors appear to act in the same regions, and thus the ticking reveals the straw-colored rather than the orange bands. Agouti is, however, in all probability uniform over the body surface in cats as in rodents. This mat- ter will be discussed in detail after the presentation of data. The skin shown in figure B had a high degree of ticking, and thus shows the longitudinal bands clearly. The cat shown in figure A is less ticked and the increased amount of black pigment on the back obscures the longitudinal bands. Figure D shows a very dark-striped tabby. While the bands on the sides are clearly seen, the longitudinal bands are obliterated by the black pigment. The cats shown in figures B and D are from the same litter and represent extreme segregation of ticking. As has been said, ticking increases in cats as in rodents with maturity. The same kitten may, therefore, show different de- grees of it at different ages. It is thus necessary to consider age in making comparisons with respect to this character. Ticking always segregates sharply from black. Various degrees of tick- ing ranging from that shown in figure B to that shown in figure D, however, occur. I have classified the extremes tentatively as A’ and A, but their allelomorphism with a is uncertain. There may be intermediate allelomorphs or the variations may be due to modifiers. The blotched pattern is shown in figures F and H. Figure F is from a kitten extremely ticked at birth. Such a kitten de- velops into a cat that has yellow in all of its hairs. The black bands of the kitten become ticked in the adult. The lighter bands become entirely straw-colored. We have in this extreme ticking an approach toward the sooty yellow, as in the mouse. The skin shown in figure H is from a kitten one week old. Nevertheless, it is much darker than that shown in figure F. Such a kitten develops into a dark-blotched adult. The ticking ad INHERITANCE OF COAT-COLOR IN CATS 5d3 increases with age until the cat appears much like the kitten shown in figure F. For a discussion of the blotched pattern in comparison with the striped, the degree of ticking shown in figure F is most favor- able. The bands shown here are broad and consequently not as numerous as in the striped. A median dark longitudinal band down the back is cut just behind the shoulders by dark and light transverse bands. The alternation of dark and light bands is not as obvious in the blotched pattern as in the striped, since the bands are relatively wide and the longitudinal and transverse bands interfere with each other. The tendency of the bands to become broken into spots or blotches may be explained in the blotched, as in the striped, by a conflict of longitudinal and transverse waves. The lined or narrow-banded pattern is rarely seen in cats in this country. The bands are extremely narrow and frequent and are best seen when the hair is very short and the ticking is of just the proper degree. Figure C shows a rather dark- lined cat (28.3) forty-five days old. The narrow banding shows clearly about the edge of the skin and to some extent on the sides. Such a cat becomes somewhat lighter when it grows older. It is very dark when young and appears black and tan like figure G, which is from a lined kitten (19.3) one week old. Narrow bands are seen in the tan areas of the latter and the back and sides show narrow bands when the skin is turned in certain relations to the line of vision. These narrow bands are really a ‘ehost pattern’ comparable to the ‘ghost patterns’ of striped and blotched seen in fully black cats. They may be seen in the fur running trans- versely down the sides. On the skin they may be seen running in the same way and also longitudinally down the back. They are much narrower and more numerous than the bands of striped eats. Lined cats occur in Africa and to some extent in Europe. They are known as African, Caffre, or Abyssinian cats. In black and maltese kittens ‘ghost patterns’ are seen clearly in the skin and are not difficult to recognize in the fur. As the kittens become older the ghost patterns sometimes show more clearly in the fur for a time, although they disappear from the 554 P. W. WHITING skin. In adult cats ghost patterns are occasionally seen. I have been able to classify all black or maltese kittens as either striped or blotched. A lined cat lacking agouti has not yet been ob- tained, but this I am hoping to do in time by the proper crosses. Figure E is from the skin of a lined kitten at birth. It is an extremely ticked example and would probably have grown to a sooty yellow adult. The back is black, but well scattered with ticked hairs, thus differing from the skin shown in figure G. The transverse bands are shown about the edge of the skin at the sides and about the tail. The longitudinal bands are suggested by two ticked spots at the back of the neck. Just posterior to these spots are two parallel ticked lines. On the body near the tail may also be seen longitudinal bands. Fundamentally, then, the lined, the striped, and the blotched patterns are comparable, differing only in the width of the bands. A pair of lined cats is owned by the Zoological Society of Phila- delphia. The male is dark while the female is much lighter. 1350 ne ——— 4 4 WEIGHT IN GRAMS AGE IN PAYS may not be made with absolute justice because, as will be shown later, the litters in the advanced service groups tend to be smaller. A rough comparison of all progeny may be made in this way, however. Charts 1 to 12, inclusive, present the results in a form that may be easily grasped by the reader, but there are a few points revealed by a study of these graphs that require some discussion. With but few exceptions, the 20th-service graph lies above all other graphs. This is a striking and surprising result and the question at once arises as to the cause of the almost uniform heavier birth weight and more rapid growth of the 15th- and 20th-service litters compared with litters of the same size from less advanced services. The results are in direct contrast to what, according to the traditions of breeders, would be expected. On their face they actually show that the heavier the service of the male, the more thrifty the offspring. It seems best to here consider the possible factors that may play a part in causing the superiority of these advanced service litters over litters fromthe Ist, 5th, and 10th service. THE JOURNAL OF PXPERIMENTAL ZOOLOGY, VOL. 25, NO. 2 590 FRANK A. HAYS During the production of the majority of the Ist- and 5th- service litters the breeding animals were housed in somewhat cramped quarters. Conditions there were not conducive to the most rapid growth of the young and were not as favorable for the breeding females because of small space and rather poor ventila- tion and poor light. Furthermore, the progeny were crowded into rather limited exercising pens, and probably for this reason they did not develop at so rapid a rate as would have been the case under the more favorable quarters used later. The ma- jority of the 10th-service litters, on the other hand, and about half of the 15th-service litters were produced while the stock was housed in more ample quarters where the space was large, the ventilation good, and everything was conducive to health and thriftiness. In fact, the quarters used at that time were prac- tically as good as the present permanent and excellent quarters where the 20th-service litters were produced. The superior en- vironment of the advanced service litters is no doubt partly re- sponsible for the greater growth of the advanced service litters compared with the moderate service litters; but environment cannot be entirely the cause of the superiority of the 20th-service litters over the 15th, and the 15th-service litters over the 10th- service litters. Let us therefore seek a further explanation. Parentage may be an important factor affecting the weight. As has been previously noted, the variability of the female breed- ing stock is considerable, the range of weight was from 2500 to 3250 grams, averaging 3050 grams, but the females have been so distributed among the three breeding males as to make three groups of practically uniform weight and variability in size. Nevertheless, lack of uniform weights in the progeny may still be partly due to variability of the female breeding stock. The size of the sire may also be a factor in controlling individual mean weight. The three sires used were quite different in weight; their weights are as follows: No. 1, 2850 grams; No. 3, 2575 grams, and No. 4, 2225 grams in ordinary breeding condition. Male No. 1 sired eleven of the seventeen litters included in the 20th- service group. He, being the largest of the three males, would be expected to sire the heaviest offspring at birth, and such offspring SEXUAL ACTIVITY OF MALE RABBITS 091 TABLE 1 Average birth weight of litters sired by the three different males used, by service groups Ist 5TH 10TH 15TH 20TH MALE MEMBER Num- | weignt | NY™- | weient | Nw" | Weight) “err | Weight Num- | Weight 1 6 46.7; 6 45.2) 7 45.2) 7 52.5} 11 59.7 3 12 50.9) 8 45.0) 2 45.0) 2 42.7| 00 00 4 8 42.4, 6 58.8) 4 58.8, 4 49.6) 6 53.3 Weight at ninety days 1 5 |1209.5| 3 |1236.6) 7 |1238.1| 7 |1262.0) 6 |1299-3 3 11 |1170.0| 6 {1217.8} 7 {1095.3} 2 |1270.7| 00 4 7 |1054.9| 6 |1096.1) 5 |1424.3} 2 |1199.7) 3 {1009.0 could be expected to keep ahead of the other classes of offspring at least for ninety days. This way of explaining the position of the 20th-service graph above the others is called in questions by chart 3 and also by table 1. The graph of the 20th-service litter lies below the others. This graph represents the growth of a single 20th-service litter (after the first weight) also by Male No. 1 and out of the heaviest female in the breeding stock (No. 15). Therefore, the fact that this litter lies below 5th- ‘and 10th-service litters on this chart cannot be explained as the result of small ancestry. Table 1 shows that the size of the male ancestor is not a very important factor in relation to the size of the young at birth. At the age of ninety days, however, the effect of the heavier sire becomes more important, but nevertheless is probably not as important as some other factors concerned as will be pointed out later. When we consider the 15th-service group, we find that seven litters were sired by Male No. 1, two by No. 3, and four by No. 4. Again we should expect a more uniformly heavy progeny than if all males had contributed an equal number of litters to the data. Chart 8 shows the superiority of the 10th-service group over the 15th-service group up to the fifty-fifth day, after which time the graph rises above all others. 592 FRANK A. HAYS Thus far we have attempted to account for the heavier weights and the greater rate of growth of the advanced service litters as due entirely to factors other than the nature of the spermatozoa and not to any inherited superiority. The effect of such factors does not seem adequate to explain the apparent superiority of the advanced service litters, therefore there is good evidence that a real superiority exists among the advanced service litters as compared with the light service litters. The female ancestors in both service groups were practically equal in weight. One of the 15th-service litters represented in chart 9 was sired by Male No. 1, the other by No. 3. Two of the four litters combined in the 10th-service graph were sired by No. 1 and two by No. 3. The smaller weights of the 15th-service litters during the early part of the observations cannot for the above reasons be explained by male ancestry of different weights. One other hypothesis may be proposed to account for the prob- able superiority of the advanced service progeny over those from very moderate service. Pearl (717, p. 296) treated both cocks and hens with ethyl aleohol, methyl alcohol, and ether at different times during the breeding season in order to study the effects on their progeny. He found the offspring from treated parents in every way superior to those from untreated parents. Pearl assumes that alcohol and other poisons act as selective agents upon the germ cells of treated animals. It is possible that se- lective action might be brought about by heavy sexual service of the male. We have previously shown that heavy sexual service induces the liberation of sperm which often show no pro- gressive motion and are short-lived. Some few of the sperm from these advanced services do exhibit the physical properties that indicate high vital force. The possibility exists then that what few spermatozoa do take part in fertilization are superior to the average in the light service groups because the bulk of the spermatozoa in the advanced service groups are not equipped to take part in fertilization, while this is probably not true in the light service groups. Such a hypothesis as the above will thus account for the superiority of the advanced service progeny. SEXUAL ACTIVITY OF MALE RABBITS 593 TABLE 2 Number of litters included in graphs of charts 1 to 12, inclusive, and the male : ancestry SERVICE GROUP MALE MEMBER Ist 5th 10th 15th 20th 1 6 6 7 7 11 3 12 8 7 2 + 8 6 6 + 6 Concerning the graphs for the 10th-, 5th-, and Ist-service litters, we note that as a rule the Ist-service litters are inferior in weight to either the 5th- or 10th-service litters and that the 10th-service litters are for the most part superior to the 5th- service litters. As previously noted, less favorable environment and greater immaturity of some of the female animals are thought to be the chief factors entering here. The male ancestry is almost uniformly distributed among the three males. Below we note from the table just how the ancestry is distributed. Table 2 shows us that the three males are about equally dis- tributed in the progeny groups from the 5th and 10th services. In the Ist-service group, however, No. 3 has sired twice as many litters as No. 1 and 50 per cent more than No. 4. Since Male No. 3 is a smaller animal than No. 1, we have here a partial explanation for the apparent inferiority of the Ist-service litters over all others. In the 15th- and 20th-service groups the prog- eny of Male No. 1 predominate, and Male No. 3 sired no litters in the 20th-service group. A word of explanation in regard to a few remarkable features of some of the charts may be of value at this point. On chart 3 the depression in the 5th-service graph at sixty days is due to a failure to obtain data on the heavier of the two litters making this graph. This particular litter was unintentionally over- looked for four weighings. On chart 5, the drop in the 10th- service graph at sixty-five days is due to the incomplete record on one litter at the time the graphs were constructed and this litter was made up of very heavy individuals. 594 FRANK A. HAYS Chart 13 represents the grand average growth of all litters in the five service groups as explained on page 582. Each graph thus represents the individual mean for the combined litters in each service group. ‘These composite service group graphs bear out the general deductions that we have madefrom a study of the graphs taken one by one comparing litters of a given number with each other in the five service groups. There is one out- standing objection to the use of such graphs as are shown on chart 13. There is a perceptible negative correlation between number of services of the sire and the number of offspring in litters resulting (Lloyd-Jones and Hays, p.492). In other words, heavy service does reduce the size of litters, especially in the two most advanced service groups used here. Consequently the greater supply of nutrients furnished by the mother in utero as well as the greater supply of milk available after birth will enable the advanced service litters to outstrip the other litters during the periods of observation in this experiment. This con- dition would hold if all litters were equally fit genetically; and we have no evidence that any class of offspring is rendered less fit by heavy service of their sire. To recapitulate, certain errors have been introduced into the growth studies in body weight, chief among which are environ- mental factors, the age and weight of the dam and the weight of the sire. These errors have been partially corrected, and the conclusion seems justified that there is no evidence in this data to show that the amount of sexual service that the male has been required to perform in any way affects the rate of growth of his offspring in body weight for the first ninety days of postnatal life. 2. Litter coefficient of variability The coefficients of variability in table 3 presented below were obtained in the following manner: The coefficient of variability . for each litter in each of the five service groups was determined at birth, at thirty days, and at ninety days by the formula: Standard deviation of each litter. Mean of the litter. SEXUAL ACTIVITY OF MALE RABBITS 595 TABLE 3 2. Coefficient of variability of individuals within the litters at three different periods SERVICE 1st 5th 10th 15th 20th AGE _— '_— '— lez _— ro) ro) S ) iS a2 a2 a2 52 Em BS Per cent 3 Per cent ree Percent [22 Percent | 9| Percent = = P= Ep} p= Z Z Z Z Z days Birth | 26)10.81+1.01) 2010.73+1.14) 19,12.52+1.37| 11/11.43+1.64! 16/8.80=1.05 30 23/10.72+1.07| 17; 8.27+0.96} 19.10.05+1.10) 11) 9.56+1.29) 11\7.89+1.11 90 23/10.10+1.02) 14 6.77+0.86) 17) 7.55+0.87| 810.70+1.80| 9)8.94+1.42 Average. 10.55 8.82 10.13 10.55 8.56 The coefficients of variability for all 1st-service litters at birth were then added together and this sum was divided by the number of litters concerned to secure the coefficient as given in table 3. Likewise the coefficients of variability for all 1st-serv- ice litters at thirty days were added together and this sum divided by the number of litters concerned to obtain the co- efficient as given in table 3. This method was used on the weights at ninety days to get the coefficient, and a similar procedure used on the weights in the other four service groups to obtain their respective coefficients. For the information of the reader the number of litters concerned in each case is presented in the table. MacDowell (14, p. 44) shows in studies on weight of adult rabbits that there is less variability within the litters than between individuals of different litters. For this reason and be- cause we wish to compare progeny of different ancestry, the method of expressing the coefficient of variation of the popula- tions as the average of the individual litter coefficients of that population is considered accurate. Table 3 shows that the coefficient of variation in rabbits is greater at birth than at any other time during our observations. This fact holds good in all service groups. While the coefficient on the average is small, it serves to indicate that prenatal nutri- tion must be subject to wide variations, otherwise greater uni- 596 FRANK A. HAYS formity in weight at birth should be expected. The thirty-day period is the weaning time for all of the litters studied in this experiment. We note from the table that the coefficient of variation falls below what it was at birth in all service groups. Here again there is no evidence of an increased percentage of ‘weak’ offspring in advanced service groups, for if such were the ease we should expect the coefficient to increase when the ani- mals were thrown into competition for nutrition during the first thirty days of postnatal life, and even one inferior individual would alter the coefficient for the litter. At the ninety-day period there is again a decrease in the coefficient of variation in all service groups, except the 15th- and 20th-service groups. The large size of the probable error here indicates that the 15th-and 20th-service groups cannot safely be assumed to be exceptions. Taking up a comparison of the coefficients for the different service groups, there appears to be slightly less variability in the offspring as the number of services increases, but this de- crease is not universal. Since the probable error is rather large, this difference is no way significant. As has been previously stated, there is also a slight reduction in the number in the litters in the same direction. Our data show us further that there is less variability in the smaller than in the larger litters. This fact affords us an explanation for the sight reduction in the coefficient of variation as the number of services increases. In table 3 a further fact seems apparent ‘that occasional genetically weak offspring do not occur in any one of the service group more frequently than in any other service group. The table also shows us that for the first ninety days of postnatal growth there is a tendency for individuals of the same litter to approach nearer to a mean weight than was the case either at birth or at thirty days of age. Fetal nutrition is thus more vari- able than either the nutrition furnished by the mother during the first thirty days after birth or the ordinary food supply fur- nished from thirty days to ninety days. SEXUAL ACTIVITY OF MALE RABBITS 597 TABLE 4 Service-group coefficients of variability at three different periods SERVICE 1st 5th 10th 15th 20th sieimechal % % % = ~ H ~ ~ ge oe se ge PES 29 Per cent = Per cent 3) Per cent a2 Percent |22| £=Percent gS| ES ES a= ES = Z Z Z 7, | N days Birth | 26/17.62+1.65 98. 10+3.10| 20'23.03+2.46| 13)15.59+3.39| 16)21.94+2.62 30 93/19.70+1.96| 18|42.65+4.80), 21/35.53+3.69 13/52.13+6.91| 11,46.37+6.66 90 93/24.25+2.41| 17\23.89+2.76| 21/19.63+2.04 11|24.97+3.53 9 30.91+4.92 Ly) i=) Weighted average.. |24.40 31.56 26.11 34.73 31.65 3. Service group coefficients of variability In tablv 4 are presented the service-group coefficients of vari- ability for all of the progeny studied in the experiment. These coefficients are obtained in the following manner: The sum of the mean individual weights of each litter in the Ist-service eroup was divided by the number of litters, to get an average at birth, at thirty days, and at ninety days. The standard deviation of this average was then calculated and the coefficient of variability (e) obtained by the formula: Standard deviation of the average _ C Average tf The same method was used for all five service groups. The service-group coefficient of variability differs from the litter coefficient of variability given in table 3 in that the former measures the range in weight between the individual litter means of the different service groups, while the latter is a measure of the range in weight between individuals of the same litter. The service-group coefficient of variability is valuable in study- ing the effects of heavy service of males upon the growth in body 598 FRANK A. HAYS weight of their offspring because it will bring to light occasional litters in which every individual is inferior. For example, table 3 shows that there is not an occasional inferior individual in the advanced service progeny. ‘This fact does not remove the pos- sibility of some entire litters being inferior because it is possible to conceive that at one time a male rabbit might sire an excep- tionally good litter on the 15th or the 20th service because of extra high reserve, but the majority of his progeny might be inferior in growth as entire litters. By table 4 we shall attempt to discover if litters as a whole are inclined to be more variable in any particular service group. Table 4 shows that at birth there is less variability in the Ist- service progeny than in any other progeny. This implies that the individual mean weight of the 1st-service litters more nearly represents the mean of every litter in the service group than is the case in any of the other four service groups. There appears to be little tendency for variability to increase as the amount of service increases as shown in the other four service groups at birth. Concerning the variability between litters at thirty days of age, practically the same relationship exists between the progeny of the different service groups as has been already noted in con- sidering the progeny at birth. The table shows us one additional fact at the thirty-day age; namely, that the greatest variability in weight during the ninety days of the observation exists at weaning time or thirty days. This fact is additional evidence that the nutrition furnished by the mother while suckling the young may vary in absolute amount or may be distributed in limited quantities because of the large number of individuals that she may suckle. At the age of ninety days there is a striking uniformity in the coefficients for all five service groups. Only in the case of the 20th-service group is there any noticeable digression, and this is probably due to the small number of litters concerned. Table 4 as a whole does not in any way indicate that inferior litters exist more frequently in any one service group than in any other, and the fact has already been pointed out in connection SEXUAL ACTIVITY OF MALE RABBITS 599 with the study of the weight graphs than in average body weight the advanced service litters are equal and in some cases superior to that of the litters in the light service groups. The fact that variability within litters is small compared with the variability in service groups is well illustrated by a comparison of the coeffi- cients in tables 3 and 4. 4. Growth by measurements as related to frequency of copulation Charts 14 to 21 are presented to show the growth in the mean dimension as obtained on forty-five litters. The method of mak- ing measurement and the determination of the mean dimension have been already explained, pp. 581-582. Each graph repre- sents averages of the mean dimension for all litters of the same size in the respective service groups. The mean dimension for a litter is obtained by adding all head measurements to all meas- urements of ilial extremes and dividing the sum by the total number of readings included in the sum. The expression thus obtained is the average individual mean dimension for the re- spective litters and may be compared with the average individual weights used in the previous charts. These charts of body development show that there is a maxi- mum increase in the mean dimension up to about the twentieth day, after which there is a very noticeable flattening of the graphs. From about the twentieth day on to the end of the observations at ninety days the progressive increase in the mean dimension is about constant. The increase in the mean dimen- sion is thus in marked contrast to the increase in body weight’ previously illustrated by charts 1 to 13. Body weight has been shown to make a rather constant increase up to the end of ninety days, and this is well illustrated by the fact that the weight graphs show little if any tendency to flatten out. Though the number of litters making up a mean dimension graph is in most cases small, they serve to illustrate the same principle as the weight graphs, namely, that the advanced serv- ice progeny are fully equal to the Ist- or 5th-service progeny at all times during the ninety days of the observation. On FRANK A. HAYS LT ass a SRL ML AT aero oT VA r oh Timea MEAN DIMENSION—CENTIMETLAS Cuart No. X1V MEASUREMENT GuRVES iN SERVICE GRoursS Zs Litter Dize 2 Zi y =e e 2 ic: 2 ye - ye {EPI TIE: ves peri Of = sO eee am Dee ee SSS eee 36 5-3 30 17) 4 24 AGE IN DAYS 4 Uf Py ACEC tas eas Se Era ae fee Seren hE are ea pe EE pe pee ee es a 10 73 2 ee Spe SSE Ow SS OSS mm bOmmECS 7 OMAR Ce MEAN UIMEINSION — CENTIMETERS Crarr No XM Measurement Curves IN SERVICE G ROVPS Litter Size is) D ies er AGE IN DAYS SEXUAL ACTIVITY OF MALE RABBITS 601 CHart No XVID Measurement Curvesin Service Garours ~Litter Size 6 MEAN D)MENSION—CENTIMETERS AGE IN DAYS CHarr No XVIL Measurement Gurves 1N Service Groves Litter Size 7 | | | | | a w- AW ---—----—--=— 4 eye } fe ioe: 4 KF = . 4FES 4 | z PSE aa = sea {aT Zz wt 37L= 4 Q z 3 ju + = | 3 3-0 a? 2 AGE IN DAYS ARE eo 37 MEAN DIMENSION— CENTIMETERS FRANK A. HAYS T T T T v Lu T T T T 1 =r ae ar CHart No XVOD : 4 Measurement Guaves in Service Groves A Litter Size 8 . | ls + BI jo? 3 eZ a ) ye 202 ee Oo 2 ZA 36 . 33 3-0 2-9 2 + AGE IN DAYS 21€ (a) Sp V2 ER VO SSRIS Sr Lae ZEA ee See OMLOS9 20] i In OM OSC T T Hh UT 7 as T * i : aa T ig = TLE = T i) 63, CHart No Xt = i | 6-oL MEASUREMENT CURVES IN GERVICE GrRoUPS le : : vA ot Litter Gize 9 BY acid: gu2 | oe sT oa cae | ——————— ek et wy S aS = ein 7/8 MCAN DIMENSION—CENTI METERS AGE IN DAYS Te Oa Ss Ol me 0 en Tm SEXUAL ACTIVITY OF MALE RABBITS ——w sl ce T T -) a fj a T T ica T T T ram! ae CHart No. XX 60 MeasuREMENT GuRveESiIN Service GRours 57 Litter Size 10 WL _-ft MEAN DIMENS! ON—CENTI METERS AGE iN PAYS CHART No. XXL Measurement Guaves in Geavice Groves | Litter Size // 2 oT. lhe a eee Po LE EL OTe eee Z u 45¢ | 4 z 2 44-9 4 w z 7-2 z Pm 3-6F = AGE /N DAYS JS Sit D ea Sa Sot SOM AS FO Sa. oO 603 604 FRANK A. HAYS 63 Cuaat No XX Pp Grano Averact Measure nent Curves g 5 =a F Au Service Groves ZA oe mane A > #1 5” => ee ee cay Gre 2 ee a Oe, ste epee © = a Z 43 ee = Z KE Z hs FZ Y 4D Ky is VY 34 fe Z, 72 Uy, £ 335 J 3 ole y 27 24 Vj AGE IN DAYS M aa LOmmms Oma? j D a ar a 2 =F ; ID charts where but a single litter makes up a graph a rather sudden break may sometimes be noted in the graph. This, in our opinion, is the result of error in measurement, and for this reason the graphs made up of several litters will be less influenced by minor errors and hence should be more representative of actual dimensions. In chart 22 are presented grand average graphs made up as the average of twenty-one Ist-service litters, fifteen 5th-service litters, eight 10th-service litters, and one 15th-service litter. Here the coincidence of the Ist-, 5th-, and 10th-service graphs is very striking. This fact bears out our previous conclusions from body weight studies that heavy sexual service of the male has no effect upon the growth of his offspring. Our evidence in studying the increase in the mean dimension does not show any effect on the progeny, from the heavy service of the male. The 15th-service graph is made up of but one litter of two individuals sired by Male No. 1 and out of an average sized female. The fact that this litter is few in numbers and has as a sire the larg- est of the males will probably account for their larger mean dimension. SEXUAL ACTIVITY OF MALE RABBITS 605 TABLE 5 Percentage mortality in offspring during the first five days of life and between the fifth and the ninetieth day of life SERVICE Ist sth | 10th | 15th | 20th Number of animals born............----+-++-- 180 {119 {139 84 77 Number dying first five days.......------+---- | 16 15 16 ily 7 Per cent dying first five days......-..-.------ 8.89| 12.61) 11.51) 13.09) 9.09 Number dying between 5 and 90 days........- 21 36 17 9 19 Per cent dying between 5 and 90 days........- 11.67| 30.25) 12.23) 10.71) 24.68 Summarizing the results of the measurement studies, we note that there is very close proximity of the graphs for the different service groups. This points very strikingly to the probable fact that heavy service of males has no effect upon the growth of their offspring in the length of head and in the breadth of ilial expanse. In table 5 the progeny are grouped by services and the num- ber and the percentage mortality is given for each service group. Under the row marked ‘‘Number dying first five days” are in- cluded all animals dead at birth as well as those that died dur- ing the first five days of life. The other row of the table in- cludes only animals actually dying between the fifth and the ninetieth day of postnatal life. / The percentage of mortality during the first five days shows a slight increase as the number of services increases up to the 15th-service group. Comparing the Ist-service group with the 20th-service group, we note that the percentage mortality in the first five days is practically the same in both groups. Since the environment has been more favorable for the 20th-service litters than for the 1st-service litters, as previously pointed out, there is no indication that twenty copulations by a male do in any way tend to reduce the percentage of his progeny that will survive the first five days of postnatal life. The table shows practically the same percentage of mortality during the first five days in both the 5th- and the 15th-service groups. The explanation for the rather high percentage of mortality in the 5th-service group 1s that two litters were destroyed outright by the mother and a THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 25, NO. 2 606 FRANK A. HAYS number of the other 5th-service litters were born during ex- tremely hot weather when the mortality was very high even among the older animals. The 10th-service group shows a higher death rate than the Ist-service group. In all these cases the percentage of mortality during the first five days does not seem to depend upon the number of services that the male is required to make. Table 5 shows that there is very little consistency between the mortality percentages as revealed in the first part of the table and between the percentages of deaths that occurred between five and ninety days. The first five days is a very critical time in the life of the young rabbit and very slight exposure may bring disaster. When this period is over the deaths usually re- sult from bowel disorders or from septicaemia. Bowel disorders are most common during the very hot weather of summer in the stock, and it is very unfortunate that a large number of the ani- mals in the 5th-service groups should have been so attacked. The 10th-service progeny also show a higher death rate than the Ist-service progeny, even though these 10th-service litters were housed under more favorable conditions than were the majority of the 1st-service litters. The mortality percentage of the 15th- service offspring is the highest of any of the service groups dur- ing the first five days of life, but it falls below that of all other service groups between the age of five and ninety days. Practi- cally one-fourth of the 20th-service rabbits died between the fifth and the ninetieth day of postnatal life. An outbreak of septicaemia happened to occur among a number of these litters. This being the case, we are inclined to believe that this sudden outbreak of disease rather than any inherent weakness of the progeny resulting from heavy sexual service of the sire is here operating to cause the high percentage of mortalities. 6. Relation of number of services made to sex of offspring A study of the relation of sex of the offspring to the amount of sexual service the male is required to perform is important because such data will show if either male or female producing SEXUAL ACTIVITY OF MALE RABBITS 607 TABLE 6 Sex ratios in service groups. Males to. 100 females SERVICE | Ist | 5th | 10th | 15th | 20th Number of individuals concerned.........-.-- | TSW GOO LEE il SES hae TS See he ies Ug oe ee | 129] 77 | 80| 53 | 28 sperm (Bachhuber, ’16) is weakened by excessive functioning of the male reproductive organs. Table 6 presented below shows the sex ratio of the offspring in the different service groups. Table 6 shows that in the Ist-service group there are 129 males to every 100 females. After the lst-service group there is a regular decline in the number of males produced, with the ex- ception of the 10th-service group. There is apparently some underlying cause to bring about the high percentage of females to males in all the advanced service groups, and there is a direct relation between the amount of service previously performed by the male and the proportion of female offspring that he will sire. The properties of the spermatozoa are perceptibly modified by heavy sexual service of males (Lloyd-Jones and Hays, 17), there being a larger percentage of weak sperm in the advanced service sperm. Two possibilities exist: either female-producing spermatozoa are formed more largely than male-producing spermatozoa as the amount of service of the male increases or the male-producing sperm are in themselves weaker than the female-producing sperm and consequently fewer of them survive to take part in fertilization. On the first point there is no evidence available. Concerning the second point, Stockard (’13) offers the hypothesis that in the case of guinea-pigs the larger female-producing sperm are more affected by aleoholization of the male than the smaller male-sperm producing. In the case of excessive sexual service, however, the large female sperm may be more vigorous because of their size or their greater chromatin content and thus out- distance the male-producing sperm in the struggle of fertiliza- tion, thus giving a higher percentage of female progeny in the heavy service groups as compared with the light service groups. 608 FRANK A. HAYS TABLE 7 Sex as related to mortality. Percentage mortality of the sexes SERVICE PERIOD SEX 5th 10th 15th 20th Rirstehivierday sere ce oe eeeen Piers On 8.69 7.14 0 9.09 2 6.85 | 11.90 | 17.74 9.09 Between five and ninety days............| & 6.52 8.93 9.09 | 27.27 | Ss 12.33 | 15.47 | 11:29 | 24.24 In table 6 we considered the relation of sexual service to the sex of the offspring and found that a predominance of females to males is the rule in the heavy service groups. In table 7 we shall consider sex of the offspring dying before the close of the observation period at ninety days. Table 7 shows that up to the 15th-service group there is a higher death rate among the female offspring than among the male offspring. In the 20th-service group, however, the fact will be noted that females are just as likely to survive as males for the first five days of postnatal life. Between the fifth and the ninetieth day there is a slightly lower death rate of females than males in the 20th-service group. These facts seem to indicate that in comparison with males of the same class, female offspring from the 20th-service are in respect of their ability so survive superior to ordinary offspring from the less advanced service groups. The fact still seems evident that these female ° offspring in the 20th-service group are slightly more likely to die than ordinary offspring. SUMMARY OF FACTS 1. Body weight of the rabbit is a measure of growth that is subject to considerable variations largely brought about by slight changes in the environment. 2. The rate of increase in body weight continues at a uniformly rapid rate for the first ninety days of the rabbit’s life. 3. The factors that appear to govern the weight of the young at birth are age of mother, state of health of mother, weight of SEXUAL ACTIVITY OF MALE RABBITS 609 mother, weight of sire, character of food of mother, and num- ber of individuals born in the litter. 4. The factors that govern the rate of postnatal growth of the young ‘or the first ninety days are weight at birth, number in litter, milk supply furnished by the mother, and, after weaning, the character of the food supplied to the young and general character of the quarters. 5. No inferiority in the offspring from the heavy service groups is revealed by comparing the body weights with those of the light service groups. 6. The average litter coefficient of variability in body weight at birth at thirty days and at ninety days is no greater in the progeny in the heavy service groups than in the light service groups. Greater variability might be expected if a part of the offspring are made genetically inferior by inferiority of the male element in the advanced service groups. 7. The service group coefficients of variability indicate greater variability in the weight of the general population than within the litters, but do not indicate that heavy service produces ‘weak’ litters. 8. Body development seems to progress at the maximum rate during the first twenty days of postnatal life, after which time there is a rather marked decline in the rate of increase in head length and breadth of ilial expanse. 9. No inferiority in the offspring from the advanced services is revealed from a study of body growth by measurement. 10. Offspring in the more advanced service groups do not show a significantly higher percentage of mortality during the first five days of life than do the offspring in the light service groups. 11. A higher mortality does not seem to exist in offspring from the advanced service groups as compared with the light service groups between the ages of five and ninety days. 12. Heavy sexual service of males gives a decrease in the proportion of male to female offspring that is very perceptible. 13. Female offspring are to some degree more likely to suc- cumb than male offspring in all service groups except the twentieth. 610 FRANK A. HAYS 14. The high percentage of deaths of female progeny is largely due to the predominance of females to males in the litters. 15. By no means thus far used has any inferiority of progeny from heavy sexual service been discovered. They are fully equal if not superior to progeny from very light service of male. DISCUSSION The amount of sexual service that the male performs has a marked effect upon the physical properties of his spermatozoa (Lloyd-Jones and Hays, 717); the whole basis of this work is to discover if these effects are in any way made manifest in the offspring. Growth in body weight must be assumed to be due to a com- plex of stimuli acting upon every living cell of the organism. If it were possible to modify the contribution of growth stimuli from the male germ cell by extreme sexual use of the male, an effect should be produced upon every cell of the body in his offspring and a reduction of these stimuli would thus result in a decreased body growth. The sum total of the body increase in the offspring from the heavy service series is fully equal and even superior to the increase in the offspring in the light service groups. This apparent superiority has been attributed to vari- ous factors, largely environmental and possibly to superior male reproductive cell. After these factors are corrected for, which we have found impossible to do, we believe that the rate of growth in body weight would be identical in all five service groups. A study of body weight as reported here will only reveal the char- acter of the total population and will not reveal the occurrence of an occasional inferior individual. The coefficient of variability of litters, on the other hand, is valuable in that it will reveal the occasional inferior individual in the litter. If only a part of the offspring in the heavy serv- ice groups are inferior as far as rate of growth is concerned, there should be a greater coefficient of variability in the litters from heavy service than among the light service litters. No such evidence appears in our data, and this fact we feel warrants the assumption that not even a part of the offspring in the SEXUAL ACTIVITY OF MALE RABBITS 611 heavy service group are more inferior as far as ability to in- crease in body weight is concerned than the offspring in the light service groups. The service group coefficient of variability does not reveal that any inferiority of entire litters is brought about by heavy sexual service of males. This coefficient does show that the largest coefficient of the first ninety days of postnatal life is found just at the close of the suckling period at thirty days. The coefficient further shows that the variability in weight of the general population is much greater than within the litters. Body measurements furnish us with further material for the study of the offspring in the different service groups. These data do not reveal any new facts to indicate any greater inferiority of offspring in any one of the five service groups. Here again the same modifying factors have been in operation that have af- fected the body-weight data, and a correction, if possible, for these we think would show that the offspring in all five of the service groups are identical in body dimensions. Concerning the question of rate of mortality in progeny from light and heavy service, we have no evidence that there is a higher death rate in the advanced service groups over that ob- served in the light service groups. A direct relation apparently exists between the amount of sexual service of males and the percentage of females that they will sire. The ratio of males to females is highest in the Ist- service group and progressively decreases up to the 20th-service group. There is a possibility that heavy service exerts a selec- tive action upon the sperm cells and may eliminate from fertili- zation the majority of the male-producing spermatozoa. T he large female-producing sperm cells may show a greater rate of motility, greater endurance, or for some other cause out-distance the male-producing spermatozoa, thus resulting in a prepon- derance of female offspring in advanced service. groups. A possible explanation for the high percentage of deaths among females lies in evidence showing that the percentage of female offspring is increased by heavy service of the male as shown on page 607. The weight (Minot, Jackson, King) of 612 FRANK A. HAYS female offspring in multiparous animals at birth is slightly less than that of the males. If this is true for the rabbit, it may render the females less able to compete with the male offspring for nourishment during their early life when food supply is of such vital importance in determining the survival of the young. The fact that the great majority of the offspring dying in early life have been females seems to warrant the assumption that females are actually less able to compete with the males during the early part of life. The data do not justify the conclusion that there is any higher rate of mortality in the advanced service groups than in the lighter service groups after the first five days of postnatal life. If inferiority of offspring exists in the advanced service groups because of the predominance of females, which we may assume under all ordinary conditions are less able to survive than males, it is apparent that no real inferiority exists, but that the mortality is greater because the percentage of females is greater in the heavy service groups. In conclusion, it may be noted 1) that the methods used for measuring the character of offspring from different degrees of sexual service of sires fail to show that any inferiority of the offspring can be induced by using a male excessively; 2) that the male in heavy sexual service furnishes germ cells that are fully the equal in their contribution to his offspring of those elaborated by a male in very moderate sexual service. ACKNOWLEDGMENTS The writer wishes to express. his high appreciation to Dr. Orren Lloyd-Jones for his constant codperation and helpful ad- vice, to Dr. H. S. Murphey for assistance in making a study of the male and female genitalia, and to Prof. G. M. Turpin and Prof. H. D. Hughes for furnishing quarters for this work for a time. SEXUAL ACTIVITY OF MALE RABBITS 613 BIBLIOGRAPHY BacuuuBer, L. J. 1916 The behavior of the accessory chromosomes and of the chromatoid body in the spermatogenesis of the rabbit. Biol. Bul., April. Day, C. E. 1913 Productive swine husbandry. Philadelphia. Hatar, S. 1908 Studies on the variation and correlation of skull measurements in both sexes of mature albino rats. Amer. Jour. Anat., vol. 7. Jackson, C. M. 1913 Postnatal growth in the albino rat. Amer. Jour. Anat., vol. 15, No. 1. Kine, H. D. 1916 On the postnatal growth of the body and of the central nervous system of albino rats that are undersized at birth. Anat. Rec., vol. 11. Luioyp-JONES, ORREN, AND Hays, F. A. 1918 The effects of frequency of copu- lation on the properties of the seminal discharge. Journ. Exp. Zodl., vol. 25. MacDowe tt, E. C. 1914 Size inheritance in rabbits. Carnegie Inst., Pub. No. 196. MacDowett, FE. C. 1914 Multiple factors in mendelian inheritance. Jour. Exp. Zoél., vol. 16. Minot, C.S. 1891 Senescence and rejuvenation. Jour. Phys., vol. 12. Minot, C.S. 1908 Age, growth, and death. New York. “PeaRL, RAymMonp. 1917 The experimental modification of germ cells. Jour, Exp. Zo6l., vol. 22, No. 2. Puscu, G. 1915 Algemeine Tierzucht. Dritte Auflage, Stuttgart. QuETELET, A. 1871 Anthropometrie, On Measure des differentes Facultes de Vhomme. StrockarD, C. R., AnD Papanicotaou, G. 1916 A further analysis of the he- reditary transmission of degeneracy and deformities by the descend- ants of aleoholized mammals. Amer. Nat., vol. 50. VauaHaN, H.W. 1916 Pedigree studies. (Unpublished.) Wrieut, L. The illustrated poultry book. ya OIE i) Ly TAT oe SUBJECT AND AUTHOR INDEX jae of male rabbits: I. On the properties of the seminal discharge. The influence of excessive Sexuale.----'- Activity of male rabbits. II. On the nature of their offspring. The influence of exces- sive sexual.....-.--------es3crr ttt once, BL! Albino rats underfed for various periods. Changes in the relative weights of the various parts, systems and organs of < ROE igen wer nob poe Pano ane argo 301 Alectrion obsoleta (Say) and Busycon cana- liculatum (Linn.). tions of the marine snails.......------ aooc Hil Amblystoma, a self-differentiating equipo- tentialsystem. Experiments on the devel- opment of the fore limb of... -.- Se cocoa 413 Amblystoma_ punctatum. Experiments on the development of the shoulder girdle and the anterior limb of....-.-----.--- er 499 Ascidia atra Lesueur. I. General physiol- ogy. The physiology Gio ase 229 Ascidia atra Lesueur. II. Sensory physi- ology. The physiology Ole eee oe saat 261 IRDS. Sex studies. XI. Hermaphro- ARivtes -o sagan a aapenoae decogaad Op Sesh OooD 1 Blattidae. Experiments on the physiology of digestion in the......-.---------++---+7" 355 Bortne, AticeE M., aND PEARL, RayYMOND. Sex studies. XI. Hermaphrodite birds... 1 Busycon canaliculatum (Linn.). The olfac- tory reactions of the marine snails Alec- trion obsoleta (Say) and.......------+-++-- 177 ANALICULATUM (Linn.). The olfac- tory reactions of the marine snails Alectrion obsoleta (Say) and Busycon. . 177 Cats. Inheritance of coat-color in......-.---- 539 Coat-color in cats. Inheritance of........---- 539 Color in cats. Inheritance of coat......-.---- 539 Coretanp, Manton. The olfactory reac- tions of the marine snails Alectrion obso- leta (Say) and Busycon eanaliculatum (Uiarini) eee tr seer aete an epi 177 | Deters S. R. Experiments on the development of the shoulder girdle and the anterior limb of Amblystoma punc- SEE iro SER OREOE CORR eh Ot CO CET ODE 499 Development, metamorphosis and growth due to a specific action of that gland? Is the influence of thymus feeding upon......... Development of the fore limb of Ambly- stoma, a self-differentiating equipotential system. Experiments on the............. Development of the shoulder girdle and the anterior limb of Amblystoma punctatum. _ Experiments on hosts. css eons 499 Diemyctylus viridescens. Some experiments _ on regeneration after exarticulation in... 10 Digestion in the Blattidae. Experiments on the physiology ER earch wader ars 355 Drosophila and its mutants. The reactions to light and to gravity in...... Wr So ticie cate ond 615 QUIPOTENTIAL system. Experiments on the development of the fore limb of Amblystoma, a self-differentiating..... 413 Exarticulation in Diemyctylus viridescens. Some experiments on regeneration after... 107 ees upon development, metamor- phosis and growth due to a specific action of that gland? Is the influence Of thymus... 2-2 ---2- seen ein 135 Fore limb of Amblystoma, a self-differentiat- ing equipotential system. Experiments on the development of the.....-.--------- 41 rere and the anterior limb of Ambly- stoma punctatum. Experiments on the development of the shoulders se. a--—eee 499 Gravity in Drosophila and its mutants. reactions to light and to......-----+-+-+-- |S ere Ross G. Experiments on the development of the fore limb of Amblystoma, a self-differentiating equi- potential system. ......------.-+:7-+727: 413 Hays, FRanK A. The influence of excessive sexual activity of male rabbits. ° II. On the nature of their offspring....-.--.--;+:- 571 Hecut, Sevic. The physiology of Ascidia atra Lesueur. I. General physiology..... 229 Hecut, Sevic. The physiology of Ascidia atra Lesueur. II. Sensory physiology.... 261 Hermaphrodite birds. Sex studies. XI...... 1 eS of coat-color in cats........ 539 IGHT and to gravity in Drosophila and its mutants. The reactions to......-.-- 49 Limb of Amblystoma, a self-differentiating equipotential system. Experiments on the development of the fore......-..-----: 413 Limb of Amblystoma punctatum. Experi- ments on the development of the shoulder girdle and the anterior......--.-----.--+;° 499 Luoyp-JoNES, ORREN AND Hays, ; The influence of excessive sexual activity of male rabbits. I. On the properties of the seminaldischarge......-------++--- 463 Meare Rospert STANLEY. The_reac- tions to light and to gravity in Droso- phila and its mutants.......---.0--++-" 49 Maculata. A study in polarity. The re- generation of triangular pieces of Planaria. 157 Male rabbits. II. On the nature of their off- spring. The influence of excessive sexua BY Ch AUNT ARE OS Obi SIS Ie eg it 571 Metamorphosis and growth due to a specific action of that gland? Is the influence of thymus feeding upon development...... 189 Morritt, C. V. Some experiments on Tre- generation after exarticulation in Di- emyctylus viridescens......----+---5:+;° 107 Mutants. The reactions to light and to gravity in Drosophila and itS< ck <<. suces 49 616 INDEX Orne (Say) and Busycon canalicu- latum (Linn.). The olfactory reactions of the marine snails Alectrion.......... 177 Offspring. The influence of excessive sexual activity of male rabbits. II. On the na- ETITS LOLEL HORT Fee te ss See rs ice 571 Olfactory reactions of the marine snails Alec- trion obsoleta (Say) and Busycon canali- culatum a@iinn,) he) see... = 17 OtmstepD, J. M. D. The regeneration of tri- angular pieces of Planaria maculata. A StuUGYsINGPOlATIGYicn oe Se eng eel cee 157 Organs of young albino rats underfed for various periods. Changes in the relative weights of the various parts, systems Fo Dye RU cr See oIG Gini Ces ISAC be 301 HYSIOLOGY of Ascidia atra Lesueur. I. General physiology. The........... 229 Physiology of Ascidia atra Lesueur. II. Sen- sory physiology. o Phece. ieee. ce it ote 261 Physiology of digestion in the Blattidae. Hixperimentss on! thee. 4-26--.-- eee 355 Planaria maculata. A study in poplarity. The regeneration of triangular pieces of.. 157 Polarity. The regeneration of triangular pieces of Planaria maculata. A study in. 157 Punctatum. Experiments on the develop- ment of the shoulder girdle and the an- terior limb of Amblystoma.............. 499 ABBITS. I. On the properties of the seminal discharge. The influence of ex- cessive sexual activity of male......... 463 Rabbits. II. On the nature of their off- spring. The influence of excessive sexual activity (Ol male. sso see. as oe Be on 571 Rats underfed for various periods. Changes in the relative weights of the various parts, systems and organs of young albinoscst ess ents coe aes Lees Reactions of the marine snails Alectrion obso- leta (Say) and Busycon canaliculatum (@iainn)= ) Mhe“olfactomy,.7 2. 2--he oe = Sakis Reactions to light and to gravity in Drosophila andwtssmutants.) sbhes=s-to- 4. yest 49 Regeneration after exarticulation in Di- emyctylus viridescens. Some _ experi- THLETNES MON hic eeie eee sree a le = ote arora sis « c 107 Regeneration of triangular pieces of naria maculata. A study in polarity. TDs aeparyaier Riss bis a mei ore OnEiee oe Gosee sae 157 ee Expon W. Experiments on the physiology of digestion in the BIStHIGRE Se daseesine cm cscs esemecnes «sie 355 Self-differentiating equipotential system. Experiments on the development of the fore limb of Amblystoma,a.............. 413 Seminal discharge. The influence of excessive sexual activity of male rabbits. I. On the jpropertiesol these saree ee tee nates 463 Sensory physiology. The physiology of ARCIGIA Jatra Ibesieur elles sec see ee Sex studies. XI. Hermaphrodite birds..... 1 Sexual activity of male rabbits. I. On the properties of the seminal discharge. The influence of excessive................ 463 Sexual activity of male rabbits. II. On the nature of their offspring. The influ- GUCEVOLMOXCERBIVE sc slate ise - onlicictiss tacts Rie 571 Shoulder girdle and the anterior limb of Am- blystoma punctatum. Experiments on the development of the.......:.!.....1.. 499 Snails alectrion obsoleta (Say) and Busycon canaliculatum (Linn.). The olfactory reactions of the marine.................. 177 Stewart, CuHeEster A. Changes in the relative weight of the various parts, sys- tems and organs of young albino rats un- derfed for various periods............... 301 System. Experiments on the development of the fore limb of Amblystoma, a self- differentiating equipotential............. 413 Systems and organs of young albino rats un- derfed for various periods. Changes in the relative weights of the various parts.... 301 HYMUS feeding upon development, me- tamorphosis and growth due to a specific action of that gland? Is the influence HLENHUTH, Epwarp. Is the influence of thymus feeding upon development, metamorphosis and growth due to a spe- Underfed for various periods. Changes in the relative weights of the various parts, systems and organs of young albino rats.. 301 TIRIDESCENS. Some experiments on / regeneration after exarticulation in Diem yetylisi.< saioa-Goece seeeee 107 EIGHTS of the various parts, systems and organs of young albino rats un- derfed for various periods. Changes ins be TelALLV.C oes c.< eee eee ee Oe eee 301 Wuitinc, P. W. Inheritance of coat-color 121, | CALBS 5,