a ae ry r ae LENE pare a Pap’ = Ze, \ pe: ; ee” HEL . Pes ; areas pete Cort sty’ yiateta DE pkvtety cot Paes ty 3 ¢ ae eRe Bi oa! case TAA AL 5 A Be die tS | aie otha ct z eet sc " Pree aot oe bce pt : sf - LEO 7 tetatelts tats ie et? ee a a = + Lh oer dk ee wy ‘« Ren ey eat ee eS Oe THE JOURNAL OF me PHRIMENTAL ZOOLOGY EDITED (BY Wiuuiam E. Caste ‘Jacques Lors Harvard University The Rockefeller Institute Epmunp B. WILSON EpwIn G. CoNKLIN Columbia University Princeton University Tuomas H. Moraan CuarLes B. DAaveNPORT Columbia University Carnegie Institution GEORGE H. PARKER HERBERT 8S. JENNINGS Harvard University Johns Hopkins University > RAYMOND PEARL FrANK R. LILLIE Maine Agricultural University of Chicago Experiment Station and Ross G. Harrison, Yale University Managing Editor VOLUME 27 UN ike!) THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. CONTENTS No. 1. OCTOBER, 1918 Heven Dean Kinc. Studies on inbreeding. III. The effects of inbreeding, with selection, on the sex ratio of the albino rat. One figure.......... Suinicut Marsumoro. Demonstration of epithelial movement by the use of vital staining, with observations on phagocytosis in the corneal epithe- Pini HOUT MO UTOS)o. eccaty ce acne eee tes Ati ney ean RY Pee ee ea WiuiraMm B. KirkHam. Observations on the relation between suckling and therate ol embryonic development in mice: 44... 2542 Jal s bee ee PureR ecamiprT. yAnahiosis:of the earthworms. 2:2)... 2 .9..2 025. -oan eee: H. H. Couuins. Studies of normal moult and of artificially induced regen- eration of pelage in Peromyscus. Fifteen figures...................... No. 2. NOVEMBER, 1918 BE. R. Hoskins anp M. M. Hoskins. The reaction of Selachii to injections of various non-toxic solutions and suspensions (including vital dyes), and to excretory toxins. Twenty-seven figures (six plates)................ EKumer Ropers. Fluctuations in a recessive Mendelian character and selection, — wo plates and three text-tivures, = 50920 s0..4450... 08s at CHARLES Hartan AsBorr. Reactions of land isopods to light. Fourteen TONITE Se pe cepa estate tat Oa es hy cy gi nee CE AB inna Ach Carma FoR SaiP, -) SUCN tae W. J. Croztmr. Assortive mating in a nudibranch Chromodoris zebra Heilprm. "Twenty-three figures and charts: 2) .°..5:.2-.0.0. 2-20 No. 3. JANUARY, 1919 Gary N. Cauxins. Uroleptus mobilis, Engelm. I. History of the nuclei during division and conjugation. Ninety-five text figures............. W. J. Crozier. On the use of the foot in some mollusks. One figure... ... 8S. O. Masr. Reversion in the sense of orientation to light in the colonial forms, Volvox globator and Pandorina morum. ‘Two figures........... ANDREW JOHNSON BrGNey. The effect of adrenin on the pigment migration in the melanophores of the skin and in the pigment cells of the retina of (AVS: TROY? eGo atone epee tetect MOOR Oe BON oes en OR ian, Lr W. W. SwincLe. Studies on the relation of iodin to the thyroid. I. The effects of feeding iodin to normal and thyroidectomized tadpoles... ... W.W.SwiInGie. Studies on the relation of iodin to the thyroid. II. Com- parison of the thyroid glands of iodin-fed and normal frog larvae... ... M.H.Jacops. Acclimatization as a factor affecting the upper thermal death points of orgamismss: ....5.....-- Lg LG PERN Qo etere te ill 195 247 293 359 367 391 397 417 427 1V CONTENTS No. 4. FEBRUARY, 1919 GerrrupeE Marrean Waiter. Association and color discrimination in mud- MUNRO W RAN Gesticklebackss seNemmiloUneS seems r ae ee esea ie eee 443 G. H. Parker. The organization of Renilla. One figure.................. 499 Mary B. Starx. An hereditary tumor. Twelve figures (three plates).... 507 AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, AUGUST 15 STUDIES ON INBREEDING Ill. THE EFFECTS OF INBREEDING, WITH SELECTION, ON THE SEX RATIO OF THE ALBINO RAT HELEN DEAN KING The Wistar Institute of Anatomy and Biology ONE FIGURE During the latter part of the nineteenth century it was gener- ally believed that sex In man and in various animals is determined mainly by the amount of nourishment that the embryos receive; well nourished embryos were supposed to become females; those that were poorly nourished were assumed to develop into males. A considerable amount of evidence in favor of this view was collected by Diising (’83, ’84, ’86), who maintained, furthermore, that close inbreeding interferes with embryonic nutrition, by lessening the vitality of the mother, and so producés a great excess of male young. In the literature of the succeeding twenty years that deals with the subject of sex determination, Diising’s statement regard- ing the effect of inbreeding on the sex ratio was widely quoted and generally credited. Those who challenged the truth of the assertion were, in the main, advocates of the ancient theory, generally ascribed to Hippocrates (460-377 B.C.), that sex is determined in the ovary; eggs from the right ovary producing males and those from the left ovary developing into females. During this period three series of experiments were made that give data regarding the sex-proportions in a closely inbred stock. Huth (87) inbred rabbits, brother and sister, for six generations and found a relatively low sex ratio (78.8 @: 100 2) among the ninety young in which the sex was ascertained; Copeman and Parsons (’04) obtained a similar result in their inbreeding experi- ments with mice. Schultze (’03) concluded that inbreeding 1 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 1 OCTOBER, 1918 2 HELEN DEAN KING has no pronounced tendency to produce an excess of male young, although he found a high sex ratio (110.90": 100 2) among 135 mice that were the offspring of brother and sister matings. The question as to whether inbreeding does or does not alter the sex ratio was not satisfactorily answered by any of these experi- ments, for in each case the number of animals used was small, and there was, apparently, no selection of the best stock for breed- ing or any way of checking the results. Moreover, none of these investigations were continued long enough to give evidence that could be considered as conclusive. The effects of inbreeding on the sex ratio seemed to me to be a problem of sufficient importance to warrant a careful and pro- longed investigation. For if it were possible to swing the sex ratio of any animal in a definite direction by factors that could be controlled, one might hope to gain valuable information regard- ing the nature of sex—a problem that has been a favorite subject of speculation for many centuries and one that modern methods of research have not, as yet, satisfactorily solved. 1. MATERIAL, METHOD, AND SCOPE OF THE INVESTIGATION The albino rat (Mus norvegicus albinus) was the animal used in this investigation, which was begun in 1909. Details regarding the manner in which the experiments were conducted were given in the first paper of this series (King, 718), but it has seemed ad- visable to repeat them here in order to give a clear understanding of the way in which the problem has been approached. The basis of the inbred strain was a litter of four albino rats, two males and two females, taken from the general colony of these animals maintained at The Wistar Institute of Anatomy and Biology in Philadelphia. The litter was selected for the purpose in view solely because of its size, not because of the ancestry or the vigor of the animals. One of the two females in the litter was called ‘A’, and her descendants form the A series of inbreds; the other female was called ‘B’, and her de- scendants are the B series of inbreds. Since the mating of brother and sister from the same litter is the closest form of inbreeding possible in mammals, such matings EFFECTS OF INBREEDING ON THE SEX RATIO 3 would be expected to be more potent than any other kind in producing an alteration in the sex ratio. In these experiments, therefore, brother and sister matings only were used to obtain strictly inbred litters from which all females used for breeding were taken. The plan of breeding that was followed through the first twenty-five generations of these animals was this: Females A and B, as well as all of the females in their respective lines that were subsequently used for breeding, were paired twice with a litter brother and then twice with an unrelated male taken from the stock colony. Sex records for the first two litters produced by any group of females might be expected to show whether inbreeding had any effect on the sex ratio; sex data for the third and for the fourth litters cast by these same females would, it was hoped, indicate whether the male or the female was responsible for the alteration, if any, in the sex ratio. For convenience the litters obtained from the mating of inbred fe- males with stock males are here designated as ‘half-inbred’ litters; no animals from such litters have ever been reared. Emphasis should be placed on the fact that, with few exceptions, the sex data given in this paper were obtained by examining the litters very soon after their birth. The sexes can readily be distinguished at this time, as Jackson (’12) has shown, and if accurate sex data are wanted it is imperative that they be taken as soon as possible, since the young that are stillborn, or those that die soon after birth, are usually eaten by the mother within a few hours. In order to keep track of a large series of animals it was neces- sary to find some way in which the pedigree of any particular individual could be told by a glance at the record card. The scheme of marking devised, which is outlined below, has proved to be very convenient and also most satisfactory for the filing of permanent records. The letter A or B is used to show from which of the two females, A or B, the animal was descended, and thus places the individual in its proper series. The serial letter is preceded in all cases by a number which signifies the generation to which the animal belonged. An index number, 2, 3, or 4, following the serial letter shows in which of the mother’s litters 4 HELEN DEAN KING the animal was born; if no index number is used the rat was a member of its mother’s first litter. The subscript following the serial letter is the number which serves to distinguish each particular rat from the other rats belonging to the same genera- tion and litter group. When it is desired to indicate the sex of the individual its number is inclosed by the sex symbol. An illustra- tion will, perhaps, render the scheme clearer. This symbol denotes a female rat belonging in the seventh generation of the A series of inbreds. She was a member of the second litter cast by her mother, and her individual number in the series of rats belonging to the second litters of the seventh generation was twenty. In the early generations of both inbred series the animals suffered severely from malnutrition which produced a marked effect on their growth, fertility, and longevity, as previous papers in this series have shown (King, ’18, 18a). During this period a considerable proportion of the individuals were sterile, and it was not possible to select animals for breeding; any rats that would breed at all were used to continue the strain. Nutritive conditions were improved at the time that the rats of the fourth inbred generation were approaching maturity, and a decided improvement in the condition of the animals was noted in a very short time: they gained rapidly in weight, the litters cast became larger and sterility almost disappeared. At this stage of the investigation it became possible to attempt to alter the sex ratio by selection within the inbred strain. From the seventh generation on, every female in the A series of inbreds that was used for breeding was taken from a litter that contained an excess of males; breeding females in the B series of inbreds were all taken from litters containing an excess of females. The plan EFFECTS OF INBREEDING ON THE SEX RATIO 5 of pairing a female twice with a litter brother and then twice with an unrelated stock male was continued through the first twenty-five generations of both inbred series. In each series litters having the desired sex ratio were reared as possible breeding stock only when the young were of large size and lusty at birth; all other litters were discarded regardless of their sex ratio. At the time that the animals became sexually mature the largest and most vigorous pairs were the ones taken to continue the strain. Selection of breeding stock, it will be noted, was based primarily on the sex ratio in the litters, not on the size or on the vigor of the young. This means that the animals in one generation that became the progenitors of the succeeding generation were selected because of their parents, tendency to produce young of a certain sex. A pair of rats that produced two litters, each of which had the desired sex ratio, was considered as having an unusually strong tendency to pro- duce unisexual young; individuals from each of these litters were used for breeding when possible. The basis of the selection, therefore, was along the line in which Pearl (12, 12a, 717) has obtained such marked success in increasing egg production in poultry, i.e., according to the ability of the parents to transmit to the offspring the quality desired. In the early part of this investigation the number of breeding females was, of necessity, small, but in the later generations about twenty females in each series were used for breeding, so that at least 1000 rats were obtained in each generation of the inbred strain. Sex records for the first twenty-five generations are given in the present paper; the data comprise 3408 litters con- taining 25,452 individuals. 2. THE NORMAL SEX RATIO IN THE ALBINO RAT. The normal sex ratio in any species can properly be determined only by obtaining the sex data for the total number of offspring produced by many females during the entire period of their reproductive activity. Unfortunately, no such series of data for the albino rat have been recorded, and only two sets of ob- servations regarding the normal sex ratio in this animal have, 6 HELEN DEAN KING as yet, appeared. Cuénot (’99) examined thirty litters of albino rats, containing 255 young, and found a sex ratio of 105.6: 100 ¢ ; data for 1089 litters of stock Albinos, collected by King and Stotsenburg (15), gave a sex ratio of 107.597:1009. Neither of these determinations seemed to furnish a proper standard for comparison with the sex ratios obtained in the inbred strain, even though they differed by less than two points. The number of individuals examined by Cuénot was too small to give results of much statistical value. The sex ratio given by King and Stotsenburg was based on the findings for a relatively large number of animals, but the litters recorded were, for the most part, cast by females that had not reached the height of their reproductive activity. The sex ratio among the offspring of young females could not justly be taken as a norm for the Albino strain in general, since it has been shown that in the albino rat the sex of the young seemingly depends, to a certain extent, on the age of the mother (King, 716 a). In order to ascertain the normal proportion of the sexes in the strain of Albinos from which the inbred animals were taken, I obtained the complete breeding history of a considerable num- ber of stock females during the past four years. As all of these individuals were reared under the same environmental conditions as the inbred rats, the sex ratio among their young would seem to be a suitable standard by which to judge the sex ratios found in various generations of the inbred animals. ‘To make the ratios more strictly comparable, the data for only the first four litters of the stock series were used in computing the sex ratio which was to serve asthe norm. These data, arranged by litter groups, are shown in table 1. TABLE 1 Showing the sex ratios in the first four litters of a series of stock albino rats LITTER SERIES |NUMBER stelae| Pelee tani MALES FEMALES 00 Eee eiRe 1 116 717 385 332 115.9 2 116 843 426 417 102.2 3 103 671 328 343 95.6 4 89 eV) oy 302 285 105.9 424 2818 1441 1377 104.6 EFFECTS OF INBREEDING ON THE SEX RATIO 7 Table 1 shows that there was a relatively large excess of males in the first litters cast by this series of stock females (115.9¢: 100 ¢), and that in succeeding litters the sex ratio tended to fall considerably. A similar change in the sex ratio of successive litters of mice was noted by Copeman and Parsons (’04), and was found also by King and Stotsenburg (715; table 7) in a series of stock albino rats. Large groups of statistics for human births, as summarized by Ahlfeld (’76), by Diising (’83, ’84), by Punnet (03), and by Newcomb (’04), all show that the sex ratio is very high among the first children of young mothers and then tends to fall with succeeding births until the mother is about thirty years old. Whether a similar change in the sex ratio is characteristic of other mammals has not been determined as yet. Among the 2818 individuals comprised in this series of stock litters there were 104.6 males to each 100 females. A sex ratio of 105: 100 2 was, therefore, taken as the norm by which to judge the sex ratios obtained in the various groups of inbred rats. This sex ratio, it will be noted, is very close to that given by Cuénot, and is lower, by over two points, than the sex ratio found in the large group of stock Albinos born in The Wistar Institute colony during the years 1911-1914 (King and Stotsenburg, 715). 3. THE SEX RATIO IN INBRED LITTERS OF ALBINO RATS The A series of inbred rats may be designated as the ‘male line,’ since after the sixth generation all of the breeding females in this series were taken from litters that contained an excess of males. Table 2 gives, by litter groups, the sex data for the 13,116 individuals obtained ‘in the first twenty-five generations of this series. Table 2 is inserted chiefly for reference, and a detailed analysis of the data, as given, will not be attempted. The summary of the data for the various litter groups shows that the sex ratio for the first litters produced was much higher than that for the second, third, and fourth litters. A similar change in the sex ratio was noted in the litter series of stock animals given in table 1. The B series of inbreds is called the ‘female line,’ since, after the sixth generation, all breeding females in this series came from HELEN DEAN KING ee Ee SS ee ee SS y' LIL |€9IT |998T |6cSz | OFS |6'SIT |ISST |ZOST |SSES | SEF 6° FIT |IGAT [890% |6F8E | ESF |O GGT |S9PT |OGST |SSEE | S8F Tiel ish \6S POL Gi 8-660 |Z) 3\8b se Ook 2 |SSIT |S6 |80T |€02 | 9G |€ 9IT 08 86 |EZt 9G GS 6 GGL |€9 LL OL 1G. |S 06" 166 1h8 LL we «IP LIT |26 =|80l |00¢ | SZ |O' GIT |P8 |OOL |P8T GG VG GGéT |29 (08 {crt 61 |9 ZIT \r2 18 I9T Go |S Sct 08 sO |§8T GG \F IST |r GIL |98T. c% &G 6 S8_-\19 GG SII AV AVE ASI ADIL IEE fe sl lee VG |6 OLL {IIL j|§cl PE LG \8 S&T \S8 SIT |§0G LG GG € SOL |c6 [6 281 Go |9 FOL |G0L |FIT \§ce 9Z |S IGT |86 GIT |LI¢ 16 \€ OIL 166 LOL \661 KG IZ IT 8c1 29 EZ OST ZL NG 6GL3\S9 as) ae Ge Ge as OO Tae Si len GG 26 \§ IPL 08 SIl |&6l LG 02 0981 |0S 89 SIT CIT |8 601 |c8 06 GLI GG «|8 GIT jI8 16 S21 &@ |v SST |8$ 68 L¥L 6B 61 Osea ize 9 {000 GI 19 ZOL \62 cS |P9T 61 8 IIT 1&6 FOL |Z61 6 |9 9&1 |TZ 46 ~—«|89T iG SI 6, SIE ey, TOL |82T 6Z |8 POL |FOL j60L j\&Ie Go |0 SOT \6IL |Scl |rrG LG \O GL \¥6 OIL |O1Z KG LT G SEL 92 9¢ c9 OL |2 GET \cS 69 IcI SI |v C&T |IZ F6 GOT IZ |0 OZI |S9 SZ Stl IG 9I 0 O01 |89 89 |9&T 21 |v SOL \€&8 06 €Z1 I@ |F FIT |06 SOL |S6r &G |8 GLE \69 60L |TLT &% GT 6 SIT |€9 ey 9§T Sos asa OTS sir G |6 26 |S6 £6 SST 1G |24 681 |8S I8 6§T IG FI ¢c IP. |S9 C6 L&T 0c |0 OFT |S9 16 9ST TG |6 9OL |2Z8 $6 OST CG |G OL FZ a OS= 0S GG sl 8°ScI |c9 SZ OFT LE WV 26) AZZ, GL 6ST 61 |G &ZI |69 G8 PGT 0G |9 €9T |SS 06 cr OZ ai ¢ TOL |29 89 SEI ST |0 SIT {22 18 POT 02 |9 LST 99 POL jOZT G \& GEL |29 68s PPL 1G IT 9 Pcl |T9 92 LET LD NAGS: a\SSel Oe I9T 02 |0 G6 {08 92 9ST 0G |€ SIT |09 IZ [ST OZ Or 6 76 |6& O92, GL 12 LOL |09 19 ral OT |S 9&1 cS IZ Sol LU | Paros. \7S: Ig COL LT 6 Pv LOL |2¢ 6G 9g 8 |9 GET \EP Lg OOT GI |r Ser |&9 62 Gal AL AS ONES shy p9 GCI dle |=28 6 LIT |8c 6& 19 6 6 GI \Sh |r9 601 GI | PEI \SS PL 661 9T |0 E21 |ZE p9 LOT OT yi GeO |Te= =\c8 = |S9: EL wie GSEeiZe 6P 98 FI |O 211 |v Gg GOL cI |0°88 0S PF 6 cI 9 SiG) | See (S60 [oe OL |I 88 6S Gg ITT cI 18 S8 {02 09 OST SI |2 911 |S £9 A) SI ¢ 9 SCT |8T 8G 9F 9 I 96T |&@ 62 GG 8 |¢°c6 OP LE Lili GI |9 E9T 126 9€ {89 §1 v 0° SZ {cl 6 1 G |0°SZT {cl GI 1G c PF 9ST |IT GI 96 ¢ 3°GG |&§Z al GE L € 0 OOT |Z Ys; FL € He AWS Ibe | ral 61 € 8°89 |9T at 1G P & GGL \6 IT G tS G 0 0¢E |c 2 6 i 9 9 \! 0 GZ IP § L I 0'SL IP § L I I ree e 5 Be me wee g e we Be eee g 5 Be me ree é 5 pe me EE) Be [SE | ge | see) 2 |e | G2) ee | eee) & | ® | ee] ge lebe| & | ® | de] ge O74 wD SA NET sia So a 29 ae | oes D a8 ae ee B a4 cao} a E ° Abe Fo e oe ae g po oes 2 -vuaxao S= Ces Bn oo an a ar Ss 5 (Q4UEGNI-ATVH) HALLIT HOO (QUUaANI-AIVH) UALLIT GUInL (GQ4UaNI) HALLIT GNOOAS (QHUANI) UALLIT LSHId (sdnowb 49974 Wt pabuntio DDp) pauqur fo sariuas py ay} fo suoyviauab aay-Ajyuan) ysuuf oy} fo Yyovd UL SOUND LAS 4) PUD SpDNpLIpUlL fo saqunu ay? Buunoys 6 ATaV EFFECTS OF INBREEDING ON THE SEX RATIO 9 litters containing an excess of females. Reference data showing the proportion of males and females produced in the various generations of this series are given, by litter groups, in table 3. The data comprise a total of 1656 litters containing 12,336 individuals. The summary for each of the four litter groups of the B series (table 3) shows that the sex ratio was at its lowest point in the first litter group, and then tended to rise in each of the subsequent groups. This is a reversed relation of the sex ratios to that shown in the litters of the stock controls (table 1) and in the litter groups of the A series (table 2), and would seem to indicate that some ageney, other than environment or the age of the mother, had influenced the relative proportion of the sexes in this series of animals. In order to compare the sex ratios in the litters sired by inbred males with the sex ratios in the litters sired by stock males, the sex records for the first and second litters produced in each generation of the two series were’ combined, as were also the records for the third and fourth litters. Table 4 shows the combined data for the litter groups of the A series; table 5 shows similar data for the litter groups of the B series. Reference to the data given in table 4 and in table 5 will be made later. To facilitate an analysis of the results obtained in the A series of inbreds, the data, as shown in table 4, were combined in generation groups (table 6). This grouping of the data was purely arbitrary. It seemed useless to compare such large series of records generation by generation, or even to combine the records for two succeeding generations. Since after the sixth generation the selection of breeding animals was made according to a definite plan, it would seem that, logically, the data for the first seven generations should form one group. Such a group, how- ever, was too large for the purpose of ascertaining whether selection produced a varying effect in different generations. It was finally decided to make a total of eight groups, each of which, except the first, should contain the data for three generations. Because of the small number of individuals, records for the first four generations were combined in one group. HELEN DEAN KING i=) om —————— EE ee €°26 |Z9IT JELOT |SEss | LOE 8'€8 |FL |c9 |98T | 02 9°S0T |9¢ |89 jPIT | ZT Ga Mey rel IPSTE = A SEL T°86= 166 <129" |G" |-61 9°08;-—|29 108= (STL =| OF O'OOT |r |hr 88 IT S'S (29 |r& 196 teal 9°88 |6L |0L |6FT | 02 698.119 |cS. \SIL. | ZI 0'OcT |j09 jel |cet | GL 6°SOT 89 i242 {OFT | ZT Vy 82.2 18S 1GELe 80 9°26 IFS |0G |FOT | SI TL |68 (66 {89 a 6601 |I2 {82 |6rT | 81 co COL |9F |L4h |&6 VI O'EIT |&6 |96 \6P OT 6°06 |IT OT 1G € G06 |Ihv |26 {82 Or GPel €& =|Ih COL 8 Orci irs 186 €T 0° O0I |6 6 81 € 00g |9 § 6 T a = sue) EE /Se| ae BES) = Dee | B per] | |e] & Be, B| & (QHUANI-AIVH) UTLLIT ALANA ete |20z1 less jozze | 1? 296 \t6 \ss est | gz 6cor les ioe cat | 92 122 |eg |eo ert | 02 zz6 lett |o6 112 | ¥2 cog |r2 leo |set | SI 6eor \zz los |zet | tz z16 \eot |ee {ger | zz 126 \6s es tat | 02 z'z6 los les leat | tz 299 lptt loz oer | zz c-gor leg igs feat | 02 Teor lz |rs {ror | oz c-zg jog loz oct | 6r cog leg lz |rgt | GI 692 |sz loo [get | SI 196 {re leo |zor | 91 e-ert leg |so leet | 21 p ror lor \sr ire | &I 16s leo or ror | ot o'oor tg |t¢ zor | #1 668 let 12 lost | st 126 lez luz |9¢ | 2 Dsezra se Gere oziis= 6 it V2 euf| | © | ee] 8 Biol ee | MME e a) ay tae ie Soames E9, “BI # (QQUENIVGIVH) AALLIT GUInd €°88 |PI6L |Z69T |909E | c9F F'08 |Z0~ {98 €'249 |€It |92 eres Gilbane | ae G92 |6IT {16 6°6 (86 16 G98 |96 |&8 1°62 |€2I |86 9°48 {26 |&8 We Nay) N24 6°S8 |66 {&8 GoL |08 (89 G'TOT |249 =|89 8°92 |S6 |€2 SOIT |€8 = |c6 ¢ 90T |c9 99 698 |sL 29 8°96 |€6 {06 0 STL |09 {69 6°GOL |I¢ {FS 0° 201 |24g 19 6° L0T 94 |c8 PT OFL |8s Iv PF 98T IE §6ST ry s9 \sf 19 ges| 2 | 2 | e8 pee|e | | es aS ie) 22, “e (QQUANI) YALLIT GNOOGS $104 31] Jo requinN G18 |PLLT |GPPT |61ze | 89% ~ S28 Ol 168 ren | 92 | “92 G98 {TIT |96 |20c | 22 Ses: 166. \28- TSGe Vaaes 4 27 OrO0T \28: -\48 PLT |eOmeel ze Gele= 66 222 /O2T- “seca ate 922 |86 |9L |Pst | 9% | 02 9'r9 |eIT |e |98t | %% | 6 9°78 |26 |92 |sor | 8 | 8I 9°o2 106 (89 |son | ez | ZI F9¢ |LOT |2¢ |ger |v | OT D280, “19s =e 1202 =| ST 129 |98 [49 |@pL | 12 | #1 o'r joor |po |pst | 12 | gt b'06 [eZ (99 eer | 02 | at 8°22 \€9- ler enn | Gl | Tr 468 jog |sp |por | 21 | OT 896 |F9 |29 |9cr | 02 | 6 Zr. |c9 |9F |sor | ST | 8 GGT Gs = Ge |Glie Sheela G°GzI |4b |6g |90r | ST | 9 9°86 [49 |99 ser | 02.) ¢ O'oor jee jee 99 | IT | + SeULL 1eteoai6e 4\oeems eee e ooo jo jo jor |@ | 2 DabOes hee aloe 2 ala oe2| 2 | E | oF | | peel e | ® | ee) °e Bees Es > ak Be Bo ¥ (GQUaNI) UALLIT LSUla eG ee ee a a a ee a re (sdnowb 1aj)4) Ur pabuv.iwv vyDp) paiqui fo sarwas g ay) fo suoypniauab aarf-hyuan) ysuf ay} fo yova U2 SOD! Tas 94} PUD S]DNpLarpur fo saqunu ay; Buinoys € HTAVL EFFECTS OF INBREEDING ON THE SEX RATIO TABLE 4 Showing the sex ratios in the inbred and in the half-inbred litters produced in each of the first twenty-five generations of the A series of inbreds INBRED HALF-INBRED (FIRST AND SECOND LITTERS) | (THIRD AND FOURTH LITTERS) males to 100 dividuals females dividuals ters ters Number of in- Number of lit- Number of in- Number of Number of lit- SIGUA CoE Cine Cor hoy | GENERATIONS ~J — bt ou “I — ~) (e9) em bo Ne) Wo) le) ido) 2) for) a=] oO ie) St (SI (oS) — OO oo — or co © tw (=) Omer — “I ~I bo 106) 682) 368) 314) 114.0) 250} 1612) 856} 756) 113.: ie ~J iS Wo} (JG) 2] as o7e) (0/2) iw nS Ww — = S 88} 651} 336) 315} 106.6} 196! 1430] 748] 682} 109. 115} 904| 491} 413) 118.8) 241) 1857/1031} 826; 124. 59] 121.9} 235) 1809) 999) 810) 123. 119} 955} 516] 439) 117.5] 265} 2099/1144) 955} 119. 129}1020} 534} 486) 109.8] 291} 2290)1232)1058) 116. 118.1) 274; 2019/1106) 913) 121. 8-10/108} 77 11-13/126) 953) 540} 413} 130. 14-16}130} 999) 558) 441) 126.: 17-19) 146/1144) 628) 516) 121. 20—22)162)1270} 698) 572) 122. 23-25) 152 Sy en = fo) (sii 00 e (= is as — 1vS) &> co “Iw © I a — — bt Rs) =P) bo pasg or S or — bo Go ou = bo ~ (oe) co S) i CO i) a (>) eZ) 2. 3678/5230) 2800|2430] 115 .6|1502|11504|6260)5244) 119.3 55 +1.47 | |-#1.36 8-25|824| 6274/5460) 2814 some eight points above the norm, it might appear that inbreed- ing had tended to increase the relative number of males. Such an interpretation of the results is not warranted, however, since the sex ratio in the litters produced by the mating of unrelated parents was higher than that in the litters obtained by the mating of brother and sister, and’since a similar increase in the sex ratio was not found in corresponding litters of the B series (table 7). As the females of the seventh generations that were used for breeding were all taken from litters that contained an excess of males, it is among their offspring that we may look for a possible alteration of the sex ratio as a result of selection. The sex ratio in the inbred litters of the eighth generation of the A series was 118.29°:1009. This sex ratio is very much lower than that found in the inbred litters of the seventh generation (1500: 100 ¢), but it is still 13 points above the norm (105%: 100 @). As examination of the records given in table 4 shows that in 14 HELEN DEAN KING only one generation (the tenth) after the eighth did the sex ratio for the inbred litters fall to norm, in all other generations it was con- siderably above the norm, the highest ratio (145.3 7: 100 2) being found in the litters of the eleventh generation. While the sex ratios for the inbred litters of the eighth to the twenty-fifth generations varied considerably, the variation was much less after the twelfth generation than before (table 4). A part of this variation was doubtless phenotypic, since seasonal changes in temperature seem to alter the sex ratio in the rat (King and Stotsenburg, ’15), and probably also other agencies, such as the age of the mother (King, 716 a), have a similar effect. As all of the sex ratios were relatively high, however, the devia- tions from the norm cannot be ascribed either to environment or to chance, so they must have been due, in part at least, to the manner in which the breeding animals were selected. A most striking uniformity in the sex ratios of the inbred litters belonging in the eighth to the twenty-fifth generations of this series is shown by the grouping of the data as made in table 6. The lowest sex ratio (112.57: 100 2) was found in the first group (eighth to tenth generations) ; the highest sex ratio (130.77: 100 @ ) appeared in the second group (eleventh to thirteenth generations). Between these extremes there was a difference of only 18 points, while in the four following groups of litters the range of variation in the sex ratios was less than 5 points. For the total of 824 inbred litters the sex ratio was 122.347:100¢9. This latter ratio was not due to an abnormal preponderance of males in a few sets of records, but was based on a series of data that in seventeen out of eighteen cases showed an excess of males greater than that considered as normal for the species. The results obtained, therefore, seem to indicate that by selecting breeding animals from litters that contain an excess of males, the sex ratio can be swung in the direction of the selection, although the line is contin- ually inbred, brother and sister. There was in this case, however, no cumulative effect of the selection. The sex ratios were more uniform in the later generations than in the earlier ones, but they were no higher. It is rather an odd coincidence that the sex ratios in the inbred litters of the eighth and of the twenty-fourth generations were exactly the same (118.27: 100 9). EFFECTS OF INBREEDING ON THE SEX RATIO 15 Data given in table 4 show that in the half-inbred litters pro- duced in the eighth to the twenty-fifth generations of the A series the range of variation in the sex ratios was from 99 to 140.8 males for each 100 females, six of these ratios being slightly below the norm. When the data were combined in generation groups (table 6), if was found that not a single group gave a sex ratio as low as the norm. The sex ratios for the litters in the later genera- tion groups were somewhat more uniform than those for the litters in the earlier generation groups, but the uniformity was not as striking as that in the corresponding groups of inbred litters. For the total of 678 half-inbred litters the sex ratio was 115.6 #: 100 2. This ratio was some 11 points above the norm and less than 7 points lower than the sex ratio in the inbred litters belonging to the same group of generations (122.397: 1009). While the litters produced by the mating of inbred females with out- bred stock males thus tended to have a lower sex ratio than did the strictly inbred litters, they did not give the sex ratio that was to be expected according to the current view that chance alone deter- mines whether a male-producing or a female-producting spermato- zoon shall fertilize the egg. Such an hypothesis requires that the sexes shall appear in approximately equal numbers when large series of sex‘data are examined. In the present case the proportion of the sexes among the 5230 individuals obtained was very far from equal. In only one group (ninth generation) out of eighteen was there a nearly equal proportion of the sexes, in all other groups there was a pronounced excess of males. The first twenty-five generations of the A series of inbreds com- prised 1752 litters containing 13,116 individuals, 7116 males and 6000 females. The sex ratio fie this series of animals was 117.4 &:100 2. This ratio was over 12 points above the norm, and since it was based on data for such a large group of animals, it would seem to indicate that in the rat the sex ratio can be altered by selection within a closely inbred line. In this instance the relative number of males was apparently increased by selecting breeding females from litters that contained an excess of males. The sex data for the inbred and for the half-inbred litters of the B series, combined in generation groups, are shown in table 7. 16 HELEN DEAN KING TABLE 7 Showing, by generation groups, the sex ratios in the inbred and in the half-inbred litters of the B series (female line) INBRED HALF-INBRED (FIRST AND SECOND LITTERS) | (THIRD AND FOURTH LITTERS) SAUER ENS QL) UNE EDS of mn a i=) ro) 5 EB alion noe SS is [sz 68 |S | 32 S s : 2 3 2 5 e = i ° & he aS Paste nes ae) poe | es nS ~~ = o2| oc mn Ont lon 4 n m2 nm = Q rel o a |aglo%] , | 2 |a8e |os/SF] , | 3 oss |f5| 52) , | 8 | 88a a ds|/ gt] 8 3 Seg |/sgel aed] & saesijsge2| es BI s ga 8 z BS|eea 2 ra] = o |32| 4a SS Ss © ap Sa — & Sam Pa) SRE SVEh || eh Eis | HERESY erat) Suet ie Seer i) Sha Neri emi eis Wl aie. 6 14 |G Pf ulcsh cs 4Z|4 a ae Z Z el eee 111} 109.9} 15) 109) 53) 56) 94.6) 51) 342).175) 167) 104.8 1-4 | 36) 233] 122 é 5-7 |100} 699} 364} 335) 108.7} 78] 603} 300) 303) 99.0) 178) 1302) 664) 638] 104.1 ~] pan ~) iss) Or co (gy) Or eo) to} oe) (Je) iw) bo Rs) 1-7 |136) 932) 486} 446) 109.0) 93 1644; 839} 805} 104.2 8-10/104| 795) 382) 413) 92.5} 73} 491} 252) 239) 105.4) 177| 1286) 634) 652) 97. 11-13]120| 876} 400} 476} 84.0) 99| 763) 358) 405} 88.4) 219) 1639) 758] 881] 86. 14-16|130} 888) 386} 502] 76.9)116} 928) 450) 478] 94.1] 246) 1816} 836} 980) 85. 3 7 6 7 2 6.0 5.3 17-19/138|1083) 472} 611| 77.3)113) 897} 414) 483) 85.7| 251) 1980) 886)1094; 80.9 20-22/152)1102|) 505) 597| 84.6)109] 845} 401) 444) 90.3) 261) 1947) 906)1041) 87.0 3.3 23-25/150/1149} 506) 643) 78.7/123) 875} 414) 461) 89.8} 273) 2024) 920/1104) 83. 8-25|794|5893|2651|3242| 81.8|633|4799]2289|2510) 91. 1]1427|10692|4940|5752| 85.9 +1.56| +1.74 +1.39 In the B series, as in.the A series, there was a wide range of variation in the sex ratios of the litters produced in the first seven generations (table 5). When the data were combined in genera- tion groups (table 7), the sex ratio in the 136 inbred litters (109 o: 100 2) was found to be above the norm, while that in the half-inbred litters (98.3 o: 100 @) was below the norm. These two ratios so nearly balance each other that for the total of 229 litters the sex ratio was 104.2 o: 100 92, or less than 1 point be- low the norm: in the corresponding litters of the A series the sex ratio was 8 points above the norm (113.2 #: 100 ?). On com- bining the records for the first seven generations of the two in- bred series (A, B), it was found that the total of 479 litters gave a sex ratio of 108.6 #%: 100 @. While this ratio is over 3 points above the norm, it is not sufficiently high to warrant the con- clusion that the normal sex ratio was changed through inbreeding, particularly as the ratio was due in great part to an unusual ex- cess of males in the hal’-inbred litters of the A series (table 4). EFFECTS OF INBREEDING ON THE SEX RATIO 17 As far as can be judged from the results of this part of the inves- tigation, close inbreeding, even when the animals are poorly nourished, does not increase the proportion of male offspring to any extent. The breeding females in the seventh generation of the B series of inbred were all taken from litters that contained an excess of females; among their offspring the sex ratio was 94.3 o: 100 9. In not one of the subsequent generations was the sex ratio in the in- bred litters as high as the norm, the nearest approach to the norm was in the twelfth generation, where the sex ratio was 101.8 0: 100 @ (table 5). In these inbred litters, as in the corresponding ones of the A series, the sex ratios were more uniform in the later than in the earlier generations, but there was no cumulative effect of selection in either case. In the Bseries, after the thir- teenth generation, there was very little change in the relative proportion of the sexes from one generation to the next, and some of the variation found, as stated for the A series, can doubtless be ascribed to environmental action. When the data for the inbred litters of the eighth to the twenty- fifth generations of the B series were combined in generation groups (table 7), 7f was found that the sex ratios for the various groups showed even greater deviations from the norm than did those for corresponding litter groups in the A series, but that this deviation was in the reverse direction, 1.e., the number of females born greatly exceeded the number of males. The highest sex ratio for any group in the B series was 92.5 o& :100 ¢@: for the entire group of 794 litters the sex ratio was 81.8 # :100 @, or 23 points below the norm. ‘This latter ratio is far too low to be considered as a chance variation, and it certainly cannot be attributed to the action of environment. For both series of inbreds were reared simul- taneously under the same environmental conditions, and if one ventured to suggest that environment swung the sex ratio in the B series towards the female side it would be necessary to assume that the same environment acted on the animals of the A series in a reverse direction and so swung the sex ratio towards the male side. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 1 18 HELEN DEAN KING As the sex ratio for the inbred litters of the B series was 23 points below the norm, while that for corresponding litters of the A series was 18 points above the norm, it would appear that the sex ratio in the rat can be swung by selection farther towards the female side than towards the male side. Moenkhaus (’11) obtained a similar result in his inbreeding experiments with Drosophila. The half-inbred litters in the eighth generation of the B series gave a sex ratio nearly 10 points higher than the norm, so here selection was not effective at once in changing the sex ratio. In none of the subsequent generations, however, was the sex ratio in these litters above the norm, most of them were considerably below wt (table 5). When the data were combined in generation groups (table 7), it was found that the sex ratios for all groups except one (eighth to tenth generations) were very low. For the total of 633 litters the sex ratio was 91.1 « :100 92, thus being 14 points below the norm and 9 points higher than the sex ratio for the inbred litters of this series. In each of the inbred series the sex ratios in the half-inbred litters belonging in the eighth to the twenty-fifth generations showed less deviation from the norm than did the sex ratios in the corre- sponding inbred litters, yet ineach case the difference between the sex ratio for the inbred group of litters and that for the half- inbred group was less than the difference between the sex ratio for the half-inbred litters and the norm. The possible signifi- cance of these results will be discussed later. In order to obtain the sex ratios for the various generations of the inbred strain as a whole, the data for the two series (A, B) were combined as shown in table 8. The range of variation in the sex ratios of the litters in the first. four generations of the inbred strain was greater than that among all of the other generation groups (table 8). This result was to be expected, considering the relatively small number of individuals in these generations and the adverse conditions under which the animals lived. When the data were combined, however, the sex ratio obtained (110.8 & : 100 2) was only 5 points above the EFFECTS OF INBREEDING ON THE SEX RATIO 19 TABLE 8 Showing the sex data for each of the first twenty-five generations of the inbred strain (series A, B), also the sex ratios when the data were combined in generation groups NUMBER 00 os P 2 I NUMBER OF MALES TO “Teese li inne cioe a eerie epee IM oak GROUPS 1 5 34 21 13 161.5 2 20 137 65 72 90.3 3 36 180 90 90 100.0 4 68 442 240 202 118.8 110.3 5 132 952 470 482 97.5 6 107 7A5 392 353 111.0 7 111 766 Amal) . 349 119.5 108.0 8 103 772 402 370 108.6 9 129 910 466 444 104.9 10 141 1034 514 520 98.8 103.6 u 154 1140 594 546 | 108.8 ie 147 1127 586 BAI 108.3 13 159 1229 609 620 98.2 104.8 4 161 1215 619 >. ‘596 103.9 15 161 1255 649 606 107.1 16 159 1155 567 588 96.4 102.5 17 184 1459 728 731 99.6 18 166 1307 660 647 102.0 19 166 1313 642 671 95.7 99.1 20 178 1294 671 623 107.7 21 184 1441 717 724 99.0 29 190 1502 750 752 99.7 101.8 23 170 1319 663 656 101.1 4 190 1386 689 697 98.9 25 187 1338 674 664 101.5 100.4 3408 25452 12895 12557 102.7 +1.28 20 HELEN DEAN KING norm. ‘The sex ratios in the litters of the fifth to the twenty- fifth generations varied from 95.7 to 119.5 males to each 100 females. Variation, it will be noted, was around the norm, eight of the twenty-one ratios being at or above the norm, the rest below it. When combined in generation groups the sex data gave a very uniform series of ratios, as the last column of table 8 shows —not one of these ratios varied more than 6 points from the norm. A variation as great as this would doubtless be found in the sex ratios of any other large series of albino rats, regardless of the manner in which the animals were bred. For the 3256 individ- uals comprised in the first seven generations of the inbred strain the sex ratio was 108.6 # : 100 2. This ratio is sufficiently close to the norm, I think, to indicate that, in the rat, inbreed- ing per se does not produce a marked increase in the number of male offspring. The sex ratio in the 22,196 individuals in the re- maining eighteen generations was 101 &# : 100 2 : for the entire series of 25,452 animals in the inbred strain the sex ratio was 102.7 # :100 @. While these last two ratios are slightly below the norm, it is evident that in the inbred strain as a whole the sex ratio was not greatly influenced either by inbreeding or by selection. The very different sex ratios obtained in the two series of the inbred strain seem to show, however, that through selection the one inbred strain was separated into two distinct lines, one line (A) having a tendency to produce an excess of males, the other line (B) tending to produce a preponderance of females. Unfortunately, one cannot predict with certainty what the sex ratio will be in the litters cast by any given inbred female, neither does the sex ratio in the litters cast by one female give a clear indication regarding the proportion of the sexes that will be found among the offspring of asisterrat. Itis only by taking the ~ averages for a large number of litters in a given series that the change in the sex ratio is made manifest. As an illustration of the individual differences in females regarding their tendencies to cast young of a certain sex, four sets of data for litters cast by sister females areshown in table 9. In each case given, sister rats were first paired with the same litter brother and later with the same stock male. EFFECTS OF INBREEDING ON THE SEX RATIO pHi | TABLE 9 Showing the difference between inbred sisters regarding their tendency to produce an excess of male or of female young when mated with the same male NUMBER NUMBER els saps MALES FEMALES SIRE ates aes MALES FEMALES SIRE 1 Bra | By 1 ul 3 ai ly dee he 10 Gye eile tame, 2 ili 5 6 11B7s 2, 11 8 3 11B7s 3 10 2 8 Stock 3 11 Uf 4 Stock 4 9 5 4 Stock 32 10 22 4] 26 16 2 17Bis | 17By5 1 8 3 5 17Bis 1 10 4 6 17Big 2 9 4 5 17Bi6 2 10 3 7 17Bis 3 8 3 5 Stock 3 7 3 4 Stock 4 3 1 2 Stock 4 11 5 6 Stock 28 11 17 38 15 23 3 12A434 12Ai35 1 8 6 2 12Ais¢ 1 8 3 5 12Ai36 2 8 4 4 12Ai36 2 9 4 5 12Ai3¢6 53 a Uf Stock z 5 4 1 Stock 4 9 3 6 Stoek 4 9 5 4 Stock 32 13 19 31 16 15 4, 13A%4, 138A 4 1 9 5 4 13A?, il u 5 2 13A2; 2 10 5 5 oA 2 9 9 13A3, 3 10 5 5 Stock a 12 a 5 Stock 4 10 5 5 Stock 4 8 5 3) Stock Be HELEN DEAN KING The first set of records given in table 9 shows the very great difference in the sex tendencies of two sister rats belonging in the B series. Female 11B;, had cast three litters when she de- veloped pneumonia and had to be killed. Each of these litters contained such a large excess of femalesthat among her thirty-two offspring the sex ratio was only 45.4 7 :100 9: Female 11By, on the other hand, showed a very strong tendency to produce male young, whether she was paired with a brother or with a stock male; among her forty-one offspring the sex ratio was 162.5 « :100 @. As yet no other sister rats have shown such a pronounced difference in their sex tendencies. A very great similarity in the sex tendencies of sister rats is shown by the second set of records in table 9. Each litter cast by 17Bu, and by 17B:s contained an excess of female young, whether the sire of the litter was an inbred or a stock male. In each group of litters the sex ratio was about 65 @ :100 °@. Female 12A,34 produced an excess of male young in each of the two litters sired by her brother, but the two litters sired by a stock male showed a very great excess of female young. Con- versely, while female 12A,;; cast more female than male young when paired with a brother, she showed a strong tendency to pro- duce an excess of male young when mated with a stock male. The last set of records in table 9 shows a case where the total number of offspring produced by each of two sister rats contained a nearly equal proportion of the sexes, but this proportion was attained in very different ways. Female 13A, showed a most pronounced tendency to produce an equal number of male and female young in each of her four litters. In the litters of female 13A, the sexes were very unequally distributed; one litter of nine young consisted entirely of females—a most unusual phe- nomenon in a litter of such size. Numerous other cases, similar to the ones given, could be fur- nished from the records for these inbred rats. The cases cited are sufficient, I think, to show the individual differences in the females regarding their tendencies to cast male or female off- spring. Incidentally, these records show, also, that the female plays a more important réle in determining the sex ratio than is generally believed. EFFECTS OF INBREEDING ON THE SEX RATIO 23 An examination of the sex data for successive litters cast by many hundreds of female rats does not indicate that there is ‘a change of sex tendency from litter to litter’ in the female, as Papanicolau (15) states is the case in guinea-pigs. Such a tendency is not shown in any of the cases given in table 8, and while the sex-proportions in the litters do change inmany cases, the change is not general or striking enough to warrant the con- clusion that there is a definite sex-determining factor involved. 4. THE SEX RATIOS IN THE LITTERS OBTAINED BY THE MATING OF STOCK FEMALES WITH INBRED MALES As a check for the results obtained by the mating of inbred females with stock males, series of stock females were bred to males from various generations of the inbred strain. The num- ber of such experiments was small, considering the scale on which the main series of experiments was conducted, but the results obtained were uniform enough to be significant. The stock females used in these experiments werereared under the same environmental conditions as the inbred rats. When they were about three months old they were paired with males from the A or from the B series that had sired inbred litters. In order to make this series of records more strictly comparable to that obtained in the inbred strain, only four litters from any one female were recorded. The data for the litters obtained by the mating of stock females with inbred males are given in table 10. Stock females paired with males from the fifth generation of the ‘inbred strain produced litters in which the sex ratio was below the norm, whether the sire of the litter belonged to the A or to the B series of inbreds. The litters sired by males from A series, however, had a much higher sex ratio than did those sired by males from the B series, although at the fifth generation there was no selection of breeding animals in the inbred strain according to a definite plan. The eighteen litters in this series gave a sex ratio of 94.7 # : 100 @, or 10 points below the norm. This ratio might seem to indicate that inbred males tended to produce an excess of female offspring, but the number of litters 2 24 HELEN DEAN KING TABLE 10 Showing the sex ratios in litters produced by the mating of outbred stock females with inbred males INBRED GENERATION SERIES TO TO WHICH | NUMBER OF | NUMBER OF NUMBER OF MALES WHICH SIRES SIRES LITTERS INDIVIDUALS Bucwbiite! NE ES To 100 FEMALES BELONGED | BELONGED A 5 10 59 30 29 103.5 B 5 8 52 24 28 85.7 18 111 eA 57 94.7£4.30 A 9 7 51 28 23 121.7 A 12 12 60 30 30 100.0 A 13 5 33 20 13 153.9 ik 15 19 177 79 98 80.6 A 16 u 126 67 59 113.6 A 17 14 117 60 57 105.2 A 18 o 3G 347 177 170 104.1 107 911 461 450 102.3+5.88 B 10 12 97 51 46 110.9 B 12. 11 75 38 37 102.7 B 13 8 49 18 24 75.0 B 15 19 172 87 85 102.3 B 16 29 243 116 127 91.3 B 18 38 313 152 161 94.4 117 942 462 480 96.2+30.14 aIsovteals 4 eas pte 242 1964 977 987 99.1 examined was too small to warrant any general conclusion from the results obtained. The second section of table 10 shows the sex ratios in the vari- ous groups of litters obtained from the matings of stock females with males from the ninth to the eighteenth generations of the A series of inbreds. There was considerable variation in these sex ratios, as was to be expected considering the number of animals involved. The total of 107 litters in this group gave a sex ratio of 102.3 7 :100 2. This ratio, it will be noted, was below the norm, although the sires of the litters were males that, paired with their sister, had fathered litters in which there was, as a rule, a preponderance of male young. EFFECTS OF INBREEDING ON THE SEX RATIO 29 The sex ratios in the various groups of litters obtained by the mating of stock females with males from the tenth to the eigh- teenth generations of the B series of inbreds showed a much nar- rower range of variation than that found in the litters sired by males of the A series of inbreds, although the number of litters produced in the two series was about the same. The 117 litters in this group gave a sex ratio of 96.2 #7 : 100 ¢@, which was 9 points below the norm. Any significance that this ratio might seem to have, when taken alone, is apparently annulled by the fact that the sex ratios for the other litters groups were also below the norm, whether the sires of the litters came from the A or from the B series of inbreds. Moreover, the probable error of the mean, calculated from the averages for the various sets of litters, was so large in every case that the differences between the sex ratios of the various groups were rendered valueless. The 242 litters in this series gave a sex ratio of 99.1 #7 : 100 @. While this ratio was some 6 points below the norm, it differed by only 4.4 points from the sex ratio.found in the 1510 litters ob- tained by the mating of inbred females with-stock males (103.5 7 : 100 2). The results as a whole, therefore, do not indicate that the sex ratio was influenced to any extent by the fact that the sires of the litters were inbred rather than outbred males. The final experiment to be made, the pairing of females from the one inbred series with males from the other inbred series, was not begun until the animals reached the twenty-sixth generation. The number of litters as yet obtained is too small to afford a basis for any general conclusion, but thus far females of the A series (male line) when paired with males from the B series (female line) have produced more male than female young, and, conversely, females of the B series, when paired with males of the A series, have shown a tendency to cast more female than male young. The results of these various series of experiments are summar- ized and discussed in the following section. bo (or) HELEN DEAN KING 5. DISCUSSION As a basis for discussion the results obtained in this investi- gation are briefly summarized as follows: 1. The inbreeding of litter brother and sister for six consecu- tive generations, during which there was no selection of animals for breeding, did not increase the number of male offspring to any extent. The sex ratio in the 3256 young obtained was 108.6 7: 100 ¢°, or less than 4 points above the sex ratio taken as the norm (105 # : 100 9). 2. Beginning with the seventh generation all breeding females in the A series were taken from litters that contained an excess of males. After this time the females in this series tended to produce an excess of male young, whether they were paired with a litter brother or with an unrelated stock male (table 6). The litters sired by inbred males gave a sex ratio of 122.3 7 : 100 9°, or over 17 points above the norm; while the litters sired by stock males showed a sex ratio of 115.6 # : 100 92, or nearly 11 points above the norm. 3. From the time that the breeding females in the B series were selected from litters containing an excess of females (seventh generation), the litters produced showed a reverse proportion of the sexes to that shown by corresponding litters in the A series (table 7). Litters sired by inbred males had a very low sex ratio (81.8 # :100 ¢); in the litters sired by stock males the sex ratio was 9 points higher than that in the inbred litters (91.1 # : 100 2), but it was still significantly lower than the norm. 4. On combining the data for the two inbred series it was found that among the 25,452 individuals comprised in the inbred strain the sex ratio was 102.7 # :100 9°, or less than 3 points below the norm (105.0 # :100 2). It thus appears that through selection the inbred strain was separated into two distinct lines: one (A) showing a high sex ratio, the other (B) a low sex ratio. Selection had the greater influence on the female line, since the sex ratio for the litters of the B series showed greater deviation from the norm than did that for the litters of the A series. 5. Stock females mated with inbred males tended to produce litters in which the sex ratio was below the norm, regardless of EFFECTS OF INBREEDING ON THE SEX RATIO 27 whether the male belonged to the A or to the B series of in- breds. The litters sired by males from the A series showed a higher sex ratio (102.3 & : 100 @), however, than did the litters sired by males from the B series (96.2 o :100 2), but these ratios are not significant, since they differ from each other and from the norm by less than three times the probable error (table 10). Diising’s contention that close inbreeding increases the relative number of male offspring was based mainly on statistics of human births collected from several isolated communities in which there were many consanguineous marriages, and on the supposedly great preponderance of male births among the Jews, who are a clannish race and intermarry more frequently than do other civilized races. The latter evidence is undoubtedly invalid, as Pearl and Salaman (13) have shown that the normal sex ratio among the Jews is the same as that in other races of man (105 @ : 100 °), and that the anomalous sex ratio among them is due to faulty registration, male births being recorded where those of females are not. The high sex ratio in the other cases cited by Diising can doubtless be attributed to a similar cause. The great excess of males found in various strains of thoroughbred dogs Heape (’08) ascribed in part to inbreeding, but in these cases also it is prob- able that the statistics are not reliable, since female pups are commonly discarded from large litters and males are registered more often, as a rule, than females. The inbreeding experiments of Huth (’87) on the rabbit, of Schultze (03) and of Copeman and Parsons (’04) on mice were made with relatively small numbers of animals, and the sex ratios obtained showed no greater deviations from the norm than might have been expected under the conditions of the experiments. Shull (13) found no change in the sex ratio of Hydatina senta asa result of repeated inbreeding, the proportion of male-producers and of female-producers remaining practically constant. In the present series of inbreeding experiments with the albino rat, all of the animals belonging to the earlier generations suffered severely from malnutrition, which Diising (’84) considered as a very potent factor in increasing the number of male offspring, yet among the individuals in the first seven generations the sex ratio was only 28 HELEN DEAN KING slightly higher than the norm. The results of these various series of experiments would seem to indicate that inbreeding per se has little, if any, effect on the sex ratio. Moenkhaus’ (’11) extensive series of inbreeding experiments on Drosophila so closely parallel my own experiments on the rat, both in the manner in which the experiments were conducted and in the results obtained, that a brief résumé of his work must be given here. é In order to obtain the normal sex ratio in Drosophila, Moenk- haus ascertained the sex of 26,933 imagos that developed from eggs laid by wild flies, and found among them a sex ratio of 88.8 & :100 2. Inthis species, therefore, there is normally an excess of females, as other investigators (Rawls, ’13; Hyde, ’14; Warren, 18) have noted. The experiments were conducted in the fol- lowing way: “Two pairs were selected from nature, the one showing a high, the other a low female ratio. These were se- lected as the parents of the two strains to be developed. From among the offspring of each of these two pairs a number of single matings were made. From among these the pair that showed the most favorable ratio in the desired direction was selected to con- tinue the strain. The same process was repeated as often as desired.” ° . In this way Moenkhaus developed two inbred strains in one of which the individuals showed a high sex ratio, in the other a low sexratio. The results of this part of the investigation showed that ‘Gt is possible to develop a strain with a high female ratio much more easily and pronouncedly than a male strain.’””’ Moenkhaus then made reciprocal crosses between the two strains in order to determine, ‘‘first, whether the maternal or the paternal elements had an equal share in the control of this ratio, and second, whether this ratio was determined in the process of fertilization.”’ The experiments showed, in a most decided way, that “the male has little or no influence in determining the sex ratio in this species. In most of the cases the ratio of the offspring falls pretty closely around that of the strain from which the females were taken. It is not certain, however, that the sex ratio is established before fertilization, since the experiments do not with certainty EFFECTS OF INBREEDING ON THE SEX RATIO 29 entirely exclude the male influence.” In his summary Moenkhaus states: ‘‘The sex ratio is one of the qualities that is, like color, an inherent character of this creature, strongly transmissible and amenable to the process of selection. . . . Sex is probably very little, if at all, influenced at fertilization in this species, but it is probably determined much earlier and by the female.’ Moenkhaus’ conclusions regarding the character of the sex ratio and its amenability to selection are as applicable to the rat as they are to Drosophila, judging from the results of my in- breeding experiments on the former species. Neither of these investigations, however, give any information regarding the causes that condition sex, although each seems to indicate that the female takes quite as important a part in this process as does the male. In the inbred strain, after the animals for breeding were selected in each generation according to a definite plan, the two series (A and B) became two separate lines as regards the sex-propor- tions among the young. In the one line (A) the litters contained, as a rule, an excess of males; in the other line (B) there was a cor- responding excess of females. Between these two lines there was no very marked difference as regards the size of the individuals at a given age, their fertility or longevity, as the data given in previous papers have shown (King, 718, ’18 a). Generation after generation, as far as the experiments have been carried, the sex ratios in the inbred lines have remained distinct, and the varia- tions from the norm have been in the same direction in each generation of each series. These results are definite enough, and they are based on data from a sufficiently large number of animals. I think, to warrant the conclusion that in the rat the sex ratio is to a certain extent at least, a character that is amenable to selection. 1 Warren (’18) has recently repeated Moenkhaus’ selection experiments on Drosophila, and concludes that the sex ratio in this form is ‘‘not readily, if at all, modifiable by selection.’’ Warren believes that the modified sex ratios found in two of his three series of experiments were due to ‘chance variation,’ and he attributes the anomalous sex ratios obtained by Moenkhaus to the action of a sex-linked lethal factor—the explanation offered by Morgan (’14 a) to account for the unusually low sex ratios found in several strains of Drosophila. 30 HELEN DEAN KING As the rats had been inbred brother and sister only, they were, according to Fish’s ('14) calculations, 79.687 per cent homozy- gous at the time that the selection of breeding animals began (seventh generation). Selection, if effective at all in changing the sex ratio, should act in one or two generations, unless a con- siderable number of factors were involved. In the latter case selection might produce a gradual change in the sex ratio which would reach its culmination only after a number of generations. In each series, as table 4 and table 5 show, the sex ratio in the inbred litters of the eighth generation was close to the sex ratio that was the average for all of the litters produced in the eighth to the twenty-fifth generations, and the sex ratios in the later generations showed no greater deviation from the norm than did those in the earlier generations, although they were somewhat more uniform. Selection thus produced its maximum effect at once, and could not shift the centre of gravity of the variation in the direction of the selection, as it did in the experiments which Castle and Phillips (14) made with piebald rats. It would appear from these results that very few heritable factors concerned in the production of the sex ratio, possibly not more than a single pair, were acted upon by selection, and that, as Pearl (’17) has stated: ‘selection acts only as a mechanical sorter of existing — diversities in the germ plasm and not as a cause of alteration in "Nie As sister rats show such marked individual difference regarding their sex tendencies, and as both nutritive (Slonaker and Card, 18) and environmental conditions (King and Stotsenberg, 715) seem to influence the sex ratio in the rat, it would seem that the sex ratio may be modified by so many agencies that it would be useless to attempt to determine the number or the nature of the particular factors that were acted upon by selection in the present ease. The factors involved are evidently not of very great po- tency, and their action is clearly shown only when a relatively large number of animals are closely inbred under environmental and nutritive conditions that are as uniform as it Is possible to make them. Whatever their nature, or in whatever manner they may be inherited, I believe that these factors act on the metabo- EFFECTS OF INBREEDING ON THE SEX RATIO ol lism of the ova in such a way as to render the ova more easily fer- tilized by one kind of spermatozoa than by the other. Inthe A series of inbreds, under the conditions given, the ova tended to attract spermatozoa that were ‘male-producing;’ in the B series, the ova tended to attract spermatozoa that were ‘female-pro- ducing.’ In advocating the possibility that fertilization may be selective I am aware that I run counter to the general belief that any egg is capable of fertilization by any spermatozo6n that happens to come in contact with it, and that those whose views have much weight in molding biological opinion believe that this hypothesis is ‘so improbable as almost to invalidate any interpretation into which it enters” (Wilson, 710). Just why this hypothesis is considered so untenable is not clear. It is true that it has not been definitely proved in any case, but neither has it been disproved, nor has any convincing proof been offered, as yet, for the very elaborate hypotheses that have been advanced to account for heredity in general and in specific cases. The burden of proof rests equally upon those who object to this hypothesis as on those who maintain it. We owe to McClung (’02) the suggestion that the accessory chromosome may be a sex-determinant. In discussing the pos- sible action of the accessory chromosome in determining sex, MeClung (’02 a) states: ‘even up to the time of fertilization the female elements are so placed as to react readily to stimuli from the mother. Here they are approached by the wandering male elements from which they may choose—if we may use such a term for what is probably chemical attraction—cither the spermatozoa containing the accessory chromosome. or those from which it is absent. In the female element, therefore, as in the female or- ganism, resides the power to select that which is for the best interest of the species.”’ In advocating selective fertilization as the probable cause of anomalous sex ratios, Heape (’09) says: “it must be remembered that there are an enormous number of spermatozoa available for the fertilization of each ovum, and, moreover, it will be recol- lected there are undoubtedly chemotactic properties associated ae, HELEN DEAN KING with ova which insure that ova of different species floating in the sea shall each be fertilized by spermatozoa of the same species, so that to grant there is still more delicate chemotaxis at work is not an illegitimate but is indeed a reasonable supposition.” Castle (03) also postulated selective fertilization in the elaboration of his Mendelian theory of sex-determination. The one attempt that has been made to test the hypothesis of selective fertilization (Morgan, Payne and Browne, ’10) seemed to indicate that the egg is fertilized by the first spermatozo6n that strikes it ‘head-on,’ but the conditions under which the obser- vations were made were so abnormal that no definite conclusions from them were possible, and even Morgan (11) states that the evidence is ‘admittedly insufficient.’ An earlier experiment that has a bearing in this connection seems to have been overlooked and therefore needs to be noted here. Marshall (’10) injected into the vagina of a pure-bred DandieDin- mont bitch a mixture of seminal fluid taken from a pure-bred dog of the same species and from a mongrel terrier of unknown an- eestry. Fifty-nine days later the bitch littered, producing four pups which were much alike. One of the pups died early, but as the other three developed into mongrels which resembled the terrier sire there was little doubt but that all four puppies were mongrels. Marshall cites another case in which a Dandie Din- mont bitch copulated with a dog of the same breed and two days later with a Scotch terrier. The bitch littered three pups; one was pure Dandie Dinmont, the other two half-bred Scotch terriers. These cases, according to Marshall, were indicative of a ‘selec- tive’ on the part of the ova of the pure-bred bitch to ‘‘conjugate with dissimilar rather than with related spermatozoa.”’ I have recently been making a series of experiments somewhat on the order of those cited by Marshall, and the results obtained indicate a very strong tendency on the part of the ova of the al- bino rat to conjugate with spermatozoa from the wild gray rat rather than with the spermatozoa of the albino rat, although under the conditions of the experiment, details of which will be pub- lished later, the advantage in every case was with the spermato- zoa from the albino male. If fertilization can be selective in such EFFECTS OF INBREEDING ON THE SEX RATIO 33 cases I ean see no valid objection to the assumption that the chemotactic reaction between ova and spermatozoa may be even more delicate and thus, under given conditions, make possible the fertilization of an egg by a spermatozo6n that has one sex potency rather than the other. There is another possible interpretation of the anomalous sex ratios found in the inbred litters of the two series. We might assume that inbreeding had acted on the males in some way so as to render one kind of spermatozo6n more potent than the other in fertilizing the ova, and that this difference in potency came to have an heritable basis in the germ plasm and so could be acted upon by selection. In the A series of inbreds, according to this assumption, the ‘male-producing’ spermatozoa became the more potent; in the B series the ‘female-producing’ spermatozoa came to have the greater potency. Were this assumption correct it should receive confirmation both from the results of the experi- ments in which inbred females were paired with stock males and from the experiments in which stock females were paired with males from different generations of the two inbred series. The litters obtained in the former case should show a nearly equal proportion of the sexes (provided it was merely a matter of chance which kind of spermatozoa fertilized the ova), since the males were outbred and therefore, theoretically, the two kinds of sper- matozoa had equal power to fertilize the ova. In the latter case the litters obtained should show a high sex ratio when the sire came from the A series of inbreds and a low sex ratio when the sire belonged to the B series. As shown in table 4 and in table 5, the half-inbred litters pro- duced by the mating of inbred females with stock males gave sex ratios that were very far from equality. In only one generation of each series was there an approximately equal proportion of the sexes, in all other cases the variation was in a definite direction: in the A series there was an excess of males; in the B series the females predominated. In both series, moreover, the sex ratios in the half-inbred litters were much closer to those in the cor- responding inbred litters than they were to the norm. The uni- formity in the various series of records and the small size of the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 1 34 HELEN DEAN KING probable error of the mean exclude the possibility that the sex ratios could have been produced by chance or by environmental action. The results, therefore, do not support the contention that the male is the chief factor in determining the sex ratio in the rat. The sex ratio in each of the three groups of litters obtained by the mating of stock females with males from various generations of the two inbred series was below the norm, whether the sire of of the litters belonged in the A or the B series. The sex ratio in the group of litters sired by males from the A series (102.3 @ : 100 2) was only 6 points higher than that in the litters sired by males from the B series (96.2 # :100 @). The results in this case, therefore, do not indicate that inbreeding, with selection, influenced the potency of the spermatozoa in any way; they seem rather to signify that the particular stock females used for breed- ing tended to attract spermatozoa that were ‘female-producing’ rather than those that were ‘male-producing.’ The results of the various experiments in which inbred and out- bred animals were paired, taken in connection with those from the experiments in which matings were made between litter brother and sister, seem to show that in the rat, as in Drosophila (Moenkhaus), the female has a greater influence than the male in determining the sex ratio, and that chance alone cannot be the factor that determines whether an egg shall be fertilized by a ‘male-producing’ or by a ‘female-producing’ spermatozoon. The size of the probable error of the mean (tables 6 and 7) in- dicates that in each series the difference between the sex ratio for the group of inbred litters and that for the group of half-inbred litters is a significant one. Apparently, therefore, the chemo- tactic reaction between the ovum and the spermatozo6n is not quite the same where these sexual elements come from unrelated individuals as when they are produced by individuals that are closely inbred. A somewhat analogous ease is found in the her- maphroditic ascidian, Ciona, where normally, as Castle (’96) and Morgan (’04, ’05) have shown, the eggs are not fertilized by sper- matozoa from the same individual, although they are readily fer- tilized by spermatozoa from any other individual, while the EFFECTS OF INBREEDING ON THE SEX RATIO 30 spermatozoa from the first animal are functional when used with ova of another animal. Morgan (14) has suggested that the infertility of the eggs of Ciona to spermatozoa from the same in- dividual may be due to the similarity in the hereditary complex of the germ cells which in some way decreases the chances of their uniting. The selective fertilization experiments made by Marshall (710) with different varieties of dogs and also my own experiments with different varieties of rats show that the ova of these animals have a strong tendency to unite with spermatozoa from individuals belonging to unrelated stock rather than with spermatozoa from individuals of the same ‘blood.’ When my own experiments are completed the results will show, I hope, whether there is a still more delicate chematactic reaction between the ova and the spermatozoa which will lead to the production of more males than females among the hybrid offspring. The ano- malous sex ratios that appear in F, hybrids almost invariably show an excess of males. This suggests that the greater the dif- ference between individuals as regards theis blood relationship the stronger is the attraction between the ova and the ‘male-pro- ducing spermatozoa. If this suggestion proves true, its converse ought also be to true, and in a closely inbred line we would expect that the chemotactic reaction between the ova and the sperma- tozoa would be such that an excess of females would be produced. Such a possibility is not incompatible with the results of the pres- ent investigation, since in the inbred strain, as a whole, the sex ratio was below the norm, while the sex ratios in the litters of the female line (B) showed a greater deviation from the norm than did the sex ratios in the litters of the male line (A). The results of this series of experiments, as a whole, seem to indicate that in the rat, as in the pigeon (Riddle, 14, 716, 717), in Drosophila (Moenkhaus, ’11) and in the guinea-pig (Papani- colau, 715), the female has more influence in determining the sex ratio than has the male. Yet it is not in the differentiation of the ova, nor in the development of the spermatozoa, that the key to the riddle of sex-determination will be found. A knowledge of the interaction of the germ cells, and of the conditions that in- fluence it, must be gained before the final solution of this problem can be attained. 7a5 4 Pe: AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, AUGUST 7 DEMONSTRATION OF EPITHELIAL MOVEMENT BY THE USE OF VITAL STAINING, WITH OBSER- VATIONS ON PHAGOCYTOSIS IN THE CORNEAL EPITHELIUM SHINICHI MATSUMOTO Osborn Zoological Laboratory, Yale University FOUR FIGURES Though the cornea of the adult frog is thin and transparent enough for the observance of epithelial movement, it is not an easy matter to note in detail every part of the process. For this reason certain vital stains were used in the present | experiments. In the amphibians, vital staining can be produced either by introducing the dyes into water in which the animals swim, or by injection, though the latter procedure is not practicable in the case of corneal epithelium, owing to its lack of blood-vessels The stained tissue may then be explanted. Another method is to add the dyestuffs to the culture medium containing the un- stained tissue. Several papers have appeared dealing with the different phases of this subject (Lewis and Lewis, ’15; Russel, ’14). For the purpose of the present work it is necessary that the dyes should not have any deleterious effect upon the epithelial cells in order that they may remain capable of maintaining their normal activity. The staining must be produced easily and last for a considerable time. That the epithelium of amphibians (especially larvae) is tingeable by various dyestuffs without affecting the organism or tissue elements to any marked degree has been shown by various observers (Fischel, ’01). Whether vital staining may be employed in order to demonstrate moving epithelium in the cul- ture is, however, another question. 37 38 SHINICHI MATSUMOTO s A. METHODS OF VITAL STAINING 1. Methods of staining by immersing the whole animal in water containing dyes a. Neutral red. Neutral red, which was first used by Ehrlich for vital staining, was considered as probably being the most suitable for the present purpose, and this proved to be the case. In the majority of the experiments ‘‘neutral red for vital staining (Ehrlich)’’ was used, and the frogs were subjected to various concentrations of the dye for various lengths of time. The experiments may be divided into two groups. In the first the animals were subjected to a relatively concentrated (1: 20,000) neutral red solution for a short time (one to five hours), and in the second they were kept in a more dilute solu- tion (1: 100,000 to 1,000,000) for a much longer period (one-half to four days). The frog was then carefully washed in running water, the tissue removed and prepared in a manner similar to that used in previous experiments (Matsumoto, 718). Both modes of staining are adapted to the study of the movement of epithelium. Though the frog could be stained an intense red without doing it any apparent harm, still the dye proved to be more or less harmful to the cells. In fact, an earlier degeneration of the explanted tissue due to the injury of cells in such preparations was frequently seen. As it was desirable to stain the tissue in such a manner that the granules would be just evident enough for observation, weak solutions (such as 1: 100,000 to 1: 200,000; ten to twenty-four hours) were preferable. The use of the more rapid staining, at times, brought on the disadvantage of having the granules appear unevenly distributed. The arrangement of the neutral red granules shows a more or less marked difference in the several cell layers of the corneal epithelium. Observing from the upper (outer) surface of the corneal tissue in the culture, we see the following: 1. The outermost layer (fig. 1, d) consists of flat polygonal cells with large nucleus and unstained fine refractive granules in the cytoplasm. These cells, which may be present all over the VITAL STAINING OF CORNEAL EPITHELIUM 39 original epithelial surface or may be absent here and there, do not usually exhibit neutral red granules and show no activity. 2. The next layer (fig. 1, ¢), which is the uppermost layer in some preparations, consists of cells which contain in the eyto- plasm the most minute red granules, almost uniform in size. These appear in abundance after prolonged intense staining. The contour of the individual cells is hardly visible at first. In some instances, especially when the rapid method of staining was employed, larger granules appeared in the cells so abun- dantly that they almost filled the cell body, leaving only a clear round or oval space in the center, occupied by the nucleus. 3. The basal cells (fig. 1, a), which are evidently smaller than those of the upper layers, exhibit, as a rule, distinct and much larger granules, not very uniform in size, and often more numer- ous on one side of the nucleus than on the other. 4. Between the two latter layers frequently another layer of cells (fig. 1, 6) is visible. These are intermediate in size between the cells of the basal layer and those of the layer above, andas regards the arrangement of the colored granules show a resem- blance to the cells of the second layer. Furthermore, between the second and third layers there were found a number of deeply red stained bodies, round or irregular in shape, some of which showed a more deeply stained red spot (one of these is shown in fig. 1, a, and another in fig. 2,a). The origin of these bodies is unknown. When the culture is observed several hours or more after the preparation, the cells of the upper and basal layers are easily distinguishable according to the distribution of granules in their eytoplasm, and this becomes more marked as the tissue becomes more translucent. The individual granules were sometimes observed to coalesce gradually, changing in shape. When the tissue is intensely stained, the cells of the connective tissue and of the posterior endothelium also show distinct gran- ules which are usually not so abundant and are more or less dis- tinguishable from those of the epithelium. The nucleus re- mains unstained throughout. 40) SHINICHI MATSUMOTO J _ Cultivated in vitro, the corneal epithelium, when stained with neutral red, showed a condition entirely similar to that seen in unstained preparations. The cells exhibited practically the same degree of activity, though at times they degenerated a little earlier than in the control preparation. By this method, the cell movement along the tissue, which took place, as a rule in the fluid or semifluid culture medium, could be easily detected. That the spreading epithelium was, as a rule, two and sometimes three layers thick, was clearly demonstrated (fig. 2, a), while in an unstained preparation it was rather hard to determine whether they were one or two cells in thickness. As the epithelium spread out and the individual cells became very flat and thin, a change in the arrangement of the granules took place. Then the granules of the basal cells, which were rather irregularly distributed at first, frequently assumed a cres- cent-shaped arrangement (fig. 2, a). If the staining was light, the tiniest red granules in the cells of the surface layer became faint relatively earlier. As a rule, the hyalin processes of the cells on the border did not exhibit any red granules. In general, after all activity of cells ceased, the red granules faded out and fatty granules increased, whereas the nucleus remained unstained. The same was true of the isolated epithelial cell. So far as our observations go, there was in general a direct relation between the cell activity and presence of neutral red granules, although of course the intensity or richness of the gran- ules,was not always parallel with the cell activity. b.* Some other dyes. In the corneal epithelium subjected to ‘Nile blue sulphate,’ the granules were beautifully stained. This dye was found, however, to be injurious to the cells, and even the solution 1: 200,000 in most instances caused more or less injury to the epithelium, though not enough to interfere immedi- ately with the cell activity, since in many cases the cells contin- ued to move. Disintegration of cells took place earlier, and a preparation which showed intense granulation never lived so long as the control culture (unstained or stained with neutral red). In addition to this, in the preparation subjected to the dye, fine violet crystals appeared in the cells or in the medium. Nile blue, VITAL STAINING OF CORNEAL EPITHELIUM 4] therefore, does not seem to be a very favorable dye for the cor- neal epithelium. Neither is ‘methylene blue (rectif.)’ suitable for the purpose, as it fades too easily. So far as our observations go, ‘gentian violet’ did not give a satisfactory result. Even the weakest solution, which stained the cells, caused their death. ‘Brilliant cresyl blue 2 b’ and ‘methyl green’ did not stain the granules clearly, nor did ‘indigo carmin,’ ‘toluidine blue,’ or ‘litmus’ give positive results. 2. Staining of the cornea by administration of dyes in the conjunctival sac Arnold (’00) demonstrated neutral red and methylene blue granules in the frog cornea by conjunctival administration of the dyes in the form of fine powder. In this way typical granules appeared in the cells, often causing injury. No experiments of this kind are recorded in the preserit paper. 3. Staining of excised cornea In some experiments the cornea was first excised from the animal and then stained. The tissue, after removal, was im- mersed in physiological saline solution containing neutral red, freshly prepared, so that the epithelium became intensely red, and could later be cultivated after thorough washing. The tech- nique is simple and is useful in some particular instances. Usu- ally, however, crystals appeared in the culture, when this method was employed. 4. Staining by addition of dyes to the culture medium Both neutral red and Nile blue, freshly dissolved in salt solu- tion, were experimented with. They are easily soluble in pure water, but not so easily if the water contains any salt (NaCl). The appearance of crystals and the uneveness of the staining proved this method to be an unsatisfactory one. Staining of nucleus and granules by the use of gentian violet failed. Methyl- ene blue stained granules, but it could not be employed for the purpose on hand. 42 SHINICHI MATSUMOTO B. PHAGOCYTIGC PHENOMENON OF THE CORNEAL EPITHELIUM Though the experiments on this line have not yet been com- pleted, some interesting phenomena will be described here. It was noted in the course of this work, that not infrequently in the preparations of cornea, where the pigment of the iris or chorioidea was accidentally mixed with the epithelium, the cells of the latter, originally not pigmented, took up the pigment, be- coming gorged with it. This was easily demonstrable, but more markedly when the pigment of the eye was finely teased and put into the culture medium with the corneal epithelium. The cells of the moving border, especially, took up the dark pigment abundantly. That the melanin granules are really taken up in the cytoplasm of cells, was clearly demonstrated in the preparations which were first vitally stained with neutral red (fig. 3). That those pig- mented cells were not the originally pigmented epithelium of eye, accidentally mixed in the preparation, is beyond doubt. The epithelial cells were able to take also melanin which was previ- ously heated or boiled. .The arrangement of the pigment granules in the cytoplasm was characteristic, resembling that of the neutral red granulations. The cells of the basal layer were often abundantly packed with the pigment, and those of upper layers were able to ingest it, too. If in the epithelial cells which are originally not pigmented, abundant melanin is taken up, it is extremely difficult to distin- guish them from the original pigment cells. The possibility of epithelial phagocytosis has been considered by Riehl (84); somewhat later, Rabl (96) also took up the ques- tion and found that the epithelium of adult salamanders (S. maculosa) is able to take up carmin injected subcutaneously. In cultures of carcinoma and sarcoma, Lambert and Hanes (’11) demonstrated that carmin granules are taken up by the cells. Carrel and Burrows (711) observed a similar phenomenon in cul- tures of sarcoma. The melanin problem represents one of the most interesting questions in dermatohistology and experimental dermatology and is a much discussed subject. I should not go, of course, so VITAL STAINING OF CORNEAL EPITHELIUM 43 far as tosay that the so-called ‘EKinschleppungstheorie’ of melanin is generally accepted, though the fact stated above shows the possibility of its correctness. Though the observations here recorded have been made only upon cells in culture, still it may not be impossible that the same occurrence takes place in vivo under certain conditions. At any rate, the phagocytic action toward melanin of epithelium of the adult frog in culture is here definitely shown. Therefore, it is not quite safe, in the culture of pigmented epi- thelium, such as iris, to consider the pigment granules as the peculiar possession of that epithelium, because the original non- pigmented epithelium can ingest pigment granules. The study of a complex and peculiar cell, such as the pigment cell, requires great care in the methods used and in conclusions drawn. Carmin, which was ground into fine powder, was also taken up into the cell bodies, which then showed beautiful carmin granulation. For this purpose the use of tissue, vitally stained with Nile blue, offered great advantage. The arrangement and distribution of the granules was entirely similar to that of the melanin; the particles were arranged with considerable uniformity around the periphery of the cytoplasm. As figure 4 shows, corneal epithelium, cultivated in plasma (or serum), containing both melanin and carmin, exhibits beautiful melanin and carmin granulations which can be preserved safely. Examination of such preparations in serial sections demon- strated the intracellular and perinuclear position of the granules. That they are in the cell bodies may be placed beyond doubt even by the careful observance of fresh preparation. Further- more, such preparations have the advantage of enabling the observer to watch how cells with ingested granules may become detached from the main mass and move off in the culture medium. Fine granules of Chinese ink or minute foreign bodies acci- dentally mixed in the culture were also taken up by the epithelium. In regard to this most interesting phenomenon of epithelial phagocytosis, our present knowledge is very deficient. A strict differentiation of the epithelial movement from that of leucocyte 44 SHINICHI MATSUMOTO may not be possible in the case of the tissues of the frog living in vitro. Not only in their movements, but also in the phago- eytic phenomena (with respect to foreign bodies) do they show a certain degree of resemblance. The question of the said phenomena in the warm-blooded ani- mals, as well as their bearing upon bacteriology and serology, will be considered later. I wish to thank Prof. R. G. Harrison for the direction and support which he has given to this work. C. SUMMARY The epithelium of the cornea of the frog which was vitally stained with neutral red and Nile blue exhibits characteristic eranules in the cytoplasm. If properly stained, the granules exist through the entire period of cell activity, without practically affecting the cells, and facilitate the observance of cell movements. Phagocytic phenomena of the corneal epithelium with refer- ence to melanin, carmin, ete., are definitely demonstrated. The distribution and arrangement of melanin and carmin gran- ules, if finely powdered, show a certain degree of resemblance to that of neutral red and Nile blue. VITAL STAINING OF CORNEAL EPITHELIUM 45 LITERATURE CITED Arnoutp 1900 Granulabilder an der lebenden Hornhaut und Nickhaut. Anat. Anz., 18, 45. CaRREL AND Burrows: 1911 Cultivation in vitro of malignant tumors. Jour. Exp. Med., 13, 571. Fiscoet 1901 Untersuchung iiber vitale Farbung. Anat. Hefte, 16, 417. LAMBERT AND Hanes 1911 Characteristics of growth of sarcoma and carcinoma cultivated in vitro. Jour. Exp. Med., 13, 495. Lewis AND Lewis 1915 Mitochondria and other cytoplasmic structures in tis- sue cultures. Am. Jour. Anat., 17. 339. Lors 1912 Growth of tissue in culture media and its significance for the analy- sis of growth phenomena. Anat. Rec., 6, 109. Matsumoto 1918 Contribution to the study of epithelial movement. The corneal epithelium of the frog in tissue culture. Jour. Exp. Zodl., ~ vol. 26. Opprrt 1912 Causalmorphologische Zellenstudien. V. Mitteilung. Arch. Entw.-Mech., 35, 371. Rast 1896 Untersuchungen iiber die menschliche Oberhaut und ihre Anhangs- gebilde mit besonderer Riicksicht auf die Verhornung. Arch. Mikr. Anat., 430. Rrest 1884 Zur Kenntnis des Pigmentes im menschlichen Haar. Viertelj. Derm. u. Syph. RussEL 1914 The effect of gentian violet on protozoa and on tissues growing in vitro, with special reference to the nucleus. Jour. Exp. Med., 20, 545. Zintonko 1874 Uber die Entwicklung und Proliferation von Epithelien und Endothelien. Arch. Mikr. Anat., 10, 351. PLATE 1 EXPLANATION OF FIGURES 1 Experiment 202. C.11. Neutral red granules in different layers of the corneal epithelium (adult frog) cultivated in plasma, 12 hours old. a, Granulesin the basal cells; b, cells of a layer often recognizable between a and c; c, granules in the cells of upper layer; d, cells of the uppermost layer, which show no activity. Xx 450. 2 Experiment 206. A. 3, showing neutral red granules in the moving epithelium, a, endothelium, 6b, and below, those of connective-tissue cells; c; 24 hours old. Culture in plasma. Movement (from left to right) of the ameboid border of the epithelium, a, on the endothelial surface 6, is demonstrated. That the moving epithelial membrane consists at least of two layers of cells is readily confirmed by the types of granulation. Note the crescent-shaped arrangement of granules in the basal cells. On the right, b, those of the endothelium are represented; the contours of the cells are not visible. Connective-tissue cells (c), the contours of which are visible, show granules in their fine processes, too. As a rule, the hyaline processes of the epithelium do not show any granules (see also figs. 3 and 4). The tissue was intensely stained (four days in the dye 1: 200,000). 250. 3 Experiment 226. A. 2, showing phagocytic phenomenon of the corneal epi- thelium, which are previously stained with neutral red; 24 hours old. Resemblance of arrangement of melanin granules to that of neutral red isnoticeable. a, Drawn from a part of epithelium, moving on the endothelial surface of cornea; b, epithe- lial cells of actively moving border. X 450. 4 Experiment 236.12. Phagocytosis of the epithelium to melanin and car- min; 18 hours old. Culture in autoplasma, in which fine powdered carmin and fresh melnin of iris were added. Drawn from a part of moving epithelial rim. x 450. 4) PLATE 1 VITAL STAINING OF CORNEAL EPITHELIUM SHINICHI MATSUMOTO one ed “nze 47 AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, SEPTEMBER 16 OBSERVATIONS ON THE RELATION BETWEEN SUCK- LING AND THE RATE OF EMBRYONIC DEVELOPMENT IN MICE WILLIAM B. KIRKHAM Osborn Zoological Laboratory, Yale University INTRODUCTION The present paper embodies the results of two separate but simultaneously conducted experiments, both designed to eluci- date further the problems presented by the lengthened gesta- tion period in mice which are pregnant and at the same time suck- - ling a previous litter. Daniel (710), working with such animals, obtained results which showed an almost constant relation of one day added to the gestation period for each animal suckled; the present writer (’16), however, in a somewhat similar series of experiments, obtained results which showed no correlation be- tween the length of the gestation period and either the number of young suckled or the number of embryos carried (cf. also table 3). This work did, however, reveal the fact that when more than two young are being suckled the implantation of embryos instead of occurring on the fifth day after fertilization usually is delayed until the fourteenth day, during which period the blas- tulae lie free in the lumen of the uterus. The cause of this de- layed implantation was tentatively stated in that paper to be due to an inhibition of some sort exerted by the fully activated mam- mary glands upon the uterine mucosa. The testing out of this hypothesis was the purpose of the following experiment. INTERRUPTED SUCKLING AND THE RATE OF EMBRYONIC DEVELOPMENT The program for this experiment was to take healthy female mice which had just given birth to litters, pair them for twenty- four hours with healthy males, and then remove all but one of the 49 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. | a 50 WILLIAM B. KIRKHAM suckling young at intervals varying, with different females, from one to thirteen days. One young animal was left in each case to avoid too violent a reaction upon the lactating organs, and the presence of a solitary suckling animal has been found to have no influence upon the length of the gestation period. The females were all killed on the thirteenth day after the birth of the suck- ling young, and any embryos then present were sectioned and their age determined by reference to a check series of known age from non-suckling females (Kirkham, 716). The results of this experiment are set forth in table 1 and should be studied in two groups. The first group, cases when the full litter was suckled one to six days, embraces a period before the next set of embryos were ready to implant, while the second group, the remainder of the table, covers an interval when the blastulae were being in- TABLE 1 Data of all mice used in an experiment to determine the effect on developing embryos of removing all but one of the suckling young. Unless otherwise stated. the full number of young born were suckled until removed for the purpose of the experiment. Stage of development of embryos is in terms of actual age of similar embryos from non-suckling females where thirteenth day post-partum equals twelfth day of embryonic development. All the females were killed on the thirteenth day after the birth of the suckled young STAGE xomper | sucmuz> | vouna. | Empryos. |, ®MBRYONIC REMARKS days IE, 1 6 6 7 J 21 2 7 6 11 J 26 3 6 1 12 1 young died AS Par 3 6 8 7 J 24 4 G 3 8 I 2s 4 6 6 10 J 18 5 10 7 11 Jaal7 6 7 11 8 J 28 6 5 5 10 J 29 6 4 6 10 J 16 7 8 12 6 1 young died Jel? 8 12 10 7 J 13 9 5) 6 6 J al0 10 7 9 6 Ap til il a i 4+ 4 1 horn uterus lost T 14 13 5 9 4 SUCKLING—-EMBRYONIC DEVELOFMENT IN MICE 51 hibited from implanting. The importance of this grouping is apparent when one considers that in the first group of cases in every instance an interval of from a few hours to several days intervened between the end of suckling by the full litter and any readiness of the embryos to implant. Embryos in the second group, on the contrary, were presumably all prepared, before the removal of the suckling young, to implant themselves, and did so as soon as the inhibition acting upon the uterine mucosa fell below a certain minimum. Analysis of the data, with these facts in mind, reveals the interesting fact that while some sets of embryos (J 18, 21, and 26) are as fully developed as control series when no young were being simultaneously suckled, others show a lag in development of from one to four days. No such diversity of results has been found by the writer in series of embryos from non-suckling fe- males, and furthermore, it is evidently not directly correlated with the number of young suckled (cf. J 26 and J 27), the number of embryos (cf. J 21 and J 22), or with the combination of these two numbers (cf. J 18 and J 27). There remains the explanation that the irregularity is due to an individual variation either in the strength of an inhibitory influence exerted by the mammary glands upon the uterus or in the susceptibility of the uterus to such influence. Evidently some individuals are so constituted metabolically that the inhibition is rapidly and entirely neutral- ized, so that if even a short interval elapses between cessation of full mammary activity and the arrival of the eggs in the uterus, implantation and further embryonic development proceed as though no young had been suckled, while in other individuals an even longer interval still leads to delay in embryonic growth. The possible existence of such an inhibitory influence of the acti- vated mammary glands upon the uterine mucosa has previously been shown by the experiments of Adler (712), who found that repeated injection of extracts of mammary gland into pregnant guinea-pigs and rabbits arrested the development of embryos and often produced abortion. 52 WILLIAM B. KIRKHAM CONSTANT NUMBER SUCKLING AND THE DEVELOPMENT OF EMBRYOS The second method of attacking the problem of prolonged gestation in suckling female mice was to determine whether or not, if the number of young suckled was constant, the stage of embryonic development could at any given time after fertiliza- tion be theoretically determined. Four was chosen as the con- stant number, since it was neither the smallest size of litter known to prolong gestation nor, on the other hand, was it such a large number as to prevent the use of nearly all of the available material. Females who had just given birth to litters of four or more young had the male already present removed, together with any excess number of young. This was done the morning following the birth of the litter, and at varying intervals thereafter the females were killed, the eggs or embryos obtained from these females were sectioned, and their stage of development determined, the as- sembled data being shown in table 2. The first notable feature in table 2 is the delay in implantation, no implanted embryos being found before the twelfth day after ovulation, a confirmation of work published by the writer in a previous paper (716). After the twelfth day of gestation there appears the same lack of correlation, as noted above in con- nection with table 1, between their stage of development and the time of ovulation, and since if degeneration is going to occur it always takes place at the stage of four or five days’ development, this possible factor can be ruled out. In other words, we have here proof of the fact that the irregularities in the rate of development of embryos in females which are simul- taneously pregnant and suckling is due, only in small measure, if at all, to the number of young suckled, for if the number suckled and the rate of embryonic development were correlated the data in table 2 should indicate a regular correlation between stage of development and age of embryos, and such is not the case. The explanation of this failure to correlate the ratio of embry- onic development in suckling females with either the size of the SUCKLING—-EMBRYONIC DEVELOPMENT IN MICE 53 TABLE 2 Data of all mice used in an experiment to determine the effect on developing embryos of the suckling of a fixed number of young. In this experiment all litters within twelve hours of their birth were reduced to four young. Stage of embryonic de- velopment in terms of actual age of similar embryos from non-suckling females AGE OF SUCKLED | STAGE EMBRYONIC SERIAL NUMBER SIZE OF LITTER Saude nae eGeanaT DELAY days days days T75 4 3 2 0 Till 4 4 3 0 T70 4 5 4 0 T 24 4 6 4 1 T 63 8 9 4 4 T 67 8 10 4 5 T 43 4 11 4 6 T 42 4 12 4 7 T 66 5 13 4 8 T 62 8 14 + 9 et 4 15 10 4 T 54 5 16 7 8 T 65 6 17 9 7 T 60 5 18 5. 12 T 56 4 19 6 12 T 68 8 20 14 5 Ais 4 21 18 2 T6l 5 22 9 12 Nays 6 23 13 9 T 59 8 24 11 12 AD TAL 7 26 17 8 2 9 26 19 6 litter or with any other factor is to be sought in the first part of this paper, where the irregular influence of the termination of suckling by a full litter is conclusively shown. Add to this the fact that under normal circumstances the suckling young begin to eat solid food at about the time implantation occurs in females which are simultaneously pregnant, and the evidence here presented all indicates that after full suckling ceases, whether by removal of the litter or weaning matters not, the inhibition to implantation is withdrawn or overcome at a widely different rate in different females. Therefore, even when the number suckled is the same, the sets of embryos show no constant rate of develop- - 54 WILLIAM B. KIRKHAM TABLE 3 Data regarding the length of the gestation period in mice simultaneously pregnant and suckling young. The gestation period in non-suckling females averages twenty days NUMBER SUCKLED rag apne DELAY REMARKS days days 1 20 0 A male was present with 1 20 0 each female when the 2 19 0 young were born, and 2 20 0 in each case the male 3 20 0 was removed 24 hours 3 29 9 later 3 30 10 4 31 Vill 4 30 10 ment, nor, as a consequence, is the period of gestation a matter that can be figured out in advance, except with broad limitations (table 3). SUMMARY 1. In mice simultaneously suckling and pregnant the removal of all but one of the suckling young at any time during the first six days after the birth of the suckling litter leads in some in- stances to implantation of the embryos as soon as they reach the uterus; in other instances the implantation is more or less delayed. These varying results can be correlated neither with the time of removal nor with the number of young taken away. 2. In mice simultaneously suckling and pregnant the removal of all but one of the suckling young at any time from seven to fourteen days after the birth of the suckling litter regularly re- sults in implantation being delayed, but the exact extent of this delay can be correlated neither with the exact time of removal nor with the number of young removed, although removal during the earlier part of the period in question does hasten implanta- tion as compared with the time required if the young had con- tinued to suckle. 3. The facts in the above paragraphs justify the statement that full activity of the mammary glands is the chief cause of SUCKLING—_EMBRYONIC DEVELOPMENT IN MICE 5 Or delayed implantation in the case of mice which are suckling young, and also that this influence of the mammary glands is subject to marked individual variation. 4. The suckling by pregnant female mice of the same number (four in this experiment) of young will not necessarily lead to either synchronous development of embryos or the same length of gestation periods. The only explanation that can be offered at the present time for this lack of uniformity is the individual variation, noted in the preceding paragraph, of the inhibition from the mammary glands or in the strength of the counteracting forces, probably due to metabolic idiosyncrasies. LITERATURE CITED ApuerR, L. 1912 Versuche mit ‘Mammimum Poehl’ betreffend die Function der Brustdriise als innerlich sezernierende Organ. Miinchner Med. Wochenschrift, 59. DanieEL, J. F. 1910 Observations on the period of gestation in white mice. Jour. Exp. Zool., vol. 9, no. 4. KrrxHam, W. B. 1916 The prolonged gestation period in nursing mice. (Ab- stract.) Anat. Rec., vol. 10, no. 3. 1916 The prolonged gestation period in suckling mice. Anat. Ree., vol. 11, no. 2. ee ia One SIRE Sak silat, amebiasis on i ‘) ne PO Sagas itibtn: ites io aman it & sean ade sea evai hen ae 3 rat | “ie me ey cae yy is porenitiAe ie ry, pits ey em Cy t, e “Tegner Laster vine De, Pe ak | eal pg autieen bi «See ay Sema art ewe dh yan ap va ; biaey an a ol, Ulrbee sai pate ait ar to! (ig 198: fem ae ni. 3 . is eee Rie wat LL PER Wire’ », Feaenerns: ae of tae uk hyve Tiel met Pc deh : fol agtahs on Aas oat ET ni math ad SS isehivitisl ov ne yiogeao tings Voc soak, sia “ef on ieee a gs . mass vlc adi ih eaareasds Beene na hh (iF Latinihs oes Ora rai naar thor aul a pada Pie: oe nies Bhp eal if wel nt Bue: . BUN — ay we yt vital nae a death wy you A ae a ] P é , oo, pun Pe a : . » 4 t 4s By Di ant 6” ia a a ay ee Ho “ty aH Wohwiiey’ afi i es aie 44 2. ie - 9 . 4. san ; oe ‘ big oe ce et leat oil, Goaiiovhodt “Tha: me iyi Wri Tht cient aud rs aA, be 3 * ae ay rail ae: Beas? ogy 2 ea palbanst eb 4 iets : ef oe : ? 7 = git Setilye aie pei a4 sana, ty Reh 4 av in ia tice ee } ¢ he stn, aa. Pens Ps Jala ak it At As, i dey ean rok dal, The i - > Lee 7m ‘ii wai ihn oF hyi's wh th tm a oe HY BT, a Vo } (ae t Be it _: Tee Antt’ weak |, finish ; ae oe bit Ges § oo chan faa Pg hen ae a ea - ta ‘ : i ae : y i aa ‘4 Tig! ray : 4 wiley G ae ret) Lav). eae ees A Havilyy * iy ea 4 Pith Ni y ee . aA 3 " 7) eee ‘ a - 2) : if 4 y ae ) es iy. tae a he. ; 4 ae ae * Ve a ah ; me re hee hess & weg ¢ (i “¢ 1? 7) ‘7. RAT GRR) kas Tee vik ity « 7 % o n ; ) i , ‘ ‘ . iw j > ew Vs ¢ an en / Aen 1M os > a as 7 - ae a ‘ + yr Era | hig sid sia at! On ag Ashe obinke fart ae ‘ : ae Oo, , ; Se ns: hes ~ "108 ie hie ¢ : 7 oy, ae? hao 7 te ny ae Le ete Mae Lee ah, ni u he ne “witha us: a th re bq ‘ iv paises raat if . > ao rane i t 1 t Rin ‘ig ." Hel ye) hi f Pea eh tra ya ae iat Sata nate a iota iy , San HLEPeT is , | Ke ae ¥ a oy fois i Lit nn why 9) st AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, OCTOBER 0 ANABIOSIS OF THE EARTHWORM PETER SCHMIDT Agricultural College in Petrograd The remarkable biological phenomenon, now known under the name of ‘anabiosis,’ given it by Preyer (’91), was already observed in 1701 by the first Dutch microscopist, Anton Leeuwenhoek. Studying microscopical animals, he found that some of them, namely, the representatives of the present groups of Tardigrada and Rotatoria, living in the moss of the roofs and in the sand of the roof-gutters, can be completely dried up and retain for a long time their vitality in such desiccated state. If after some weeks or months of preservation one places in water the dried bodies of these animals, they quickly become swollen and in a short time the animals revive. The same observation was made afterwards by Needham (1745) and Baker (1764) on some Nematodes belonging to Anguillu- lidae; Baker revived these microscopical worms after twenty- seven years of preservation in the desiccated state. The famous Italian biologist of the eighteenth century, Abbott Spallanzani (1777), repeated all the experiments on the Tardi- erada and Rotatoria of the previous authors; he came to the same conclusions and cleared up many interesting details of the phenomenon. For a long time these experiments on resuscitation of micro- scopic animals had interested the scientific world rather more as a curiosity. But in the second half of the last century the at- tention of the biologist was called to the fact of the resuscitation in connection with the question of the origin of life and,spon- taneous generation. The property of resuscitation of Rotatoria and Tardigrada was once more carefully studied by Doyére (’42) and the resusci- tation of Anguillulidae by Davaine (’56). Afterwards this 57 58 PETER SCHMIDT question was studied in a more detailed way by Gavarret (’59). In 1860 the ‘Société de Biologie’ in Paris was in such a degree interested in clearing up*this question that it elected a special commission, with Broca as president and Balbiani, Brown- Sequard, Darest, Guilliemint, and Robin as members, for con- trolling the experiments on revivification of Rotatoria and Tardigrada. This commission (Broca, ’60) not only confirmed . the fact of the reviving of these microscopic animals, but also stated their excessive endurance to high temperature. In the dried state they endure a temperature of 100°C. and can remain for very long time in a vacuum without loss of vitality. However, this question has not been reviewed and reéxplored in full measure by modern methods. Some zoologists who have studied the resuscitation of one or another respresentative of Rotatoria and Tardigrada for instance, Fromantel (’77), Plate (86), Zacharias (86) came to negative results, but the others, such as Schulz (15), have not only confirmed the results of previous authors, but have also given new proofs of the re- sistance of exsiccated animals to the absence of oxygen. Ac- cording to experiments of Schulz, the Tardigrada, Rotatoria, and Nematodes can remain exsiccated during one year in an atmosphere of pure hydrogen and not lose their vitality. Summarizing all that is known upon this subject, one may say that most experimenters have succeeded in the revivification of exsiccated Tardigrada, Rotatoria, and Nematodes, which, living in moss and sand, are adapted to such loss of water. The body of these microscopic animals wrinkles up completely by the exsiccation and loses its form, but if placed afterwards in water it swells and regains its natural size and contour. It can be re- garded as proved, also, that such exsiccated animalcules endure easily and without injury high temperatures (100°C. and more; in experiments of Doyére, 140°C.), as well as absence of oxygen. But many points of this remarkable phenomenon remain unelucidated and require further investigation. Firstly, it would be necessary to ascertain how far the exsiccation can go without loss of vitality. It is, a priori, improbable that the animalcules can lose all the water contained in their living tissues, ANABIOSIS OF THE EARTHWORM 59 and it may be that the uncertainty of the results stated by some experimenters depends mostly upon the fact that not all animal- cules lose the same quantity of water by the exsiccation. But it is altogether impossible to determine the percentage of water lost by the exsiccation of anabiotic Tardigrada, Rotatoria, and Nematodes in consequence of the microscopic dimensions of the creatures. A casual observation has directed my attention to another object more appropriate for experiments of this kind and show- ing features completely analogous to the anabiotic Tardigrada, Rotatoria, and Nematodes—that is the earthworm. The Lumbricidae are animals also adapted by the nature of their habitat to exsiccation; they live in the uppermost layers of the soil, which are often more or less dry, and the worms must lose water. Evidently they do not lose it in such degree as the moss inhabitants, but, on the other hand, they have a higher organiza- tion, and the loss of the same percentage of water must be for them of more serious consequence. The comparatively large size of earthworms permits a more detailed study of the process of exsiccation and also the determination, with full precision, of the degree of loss of water. In the first study concerning the earthworms, made by me in collaboration with my pupil, Miss T. V. Stehepkina (Schmidt and Stchepkina, ’16), in the Zoological Laboratory of the Agri- cultural College in Petrograd, we have tried to determine the influence of lower temperatures on earthworms, and we have shown by experiments that the loss of water has no influence on the endurance of earthworms as to low temperature. The normal worms, as well as the partly exsiccated ones, die between —1.6° and —2°C. But in studying this question, we discovered that the earthworms show a very great tolerance of loss of water contained in their bodies. I give here a table of results obtained in 1916 in these experi- ments (table 1). 60 PETER SCHMIDT TABLE 1 NUMBER Laie kr eine ies ibe Loss TIME OF EXSICCATION grams grams per cent 1 1.2720 0.2430 19.1 6 hours 45 minutes 2 0.8815 0.2995 33.4 6 hours 45 minutes 3 0.9055 0.2670 28.3 6 hours 45 minutes 4 0.9255 0. 2650 28.5 6 hours 45 minutes 5 1.1520 0.3255 28.1 6 hours 45 minutes 6 1.2589 0.2583 25.1 _—+| 6 hours 45 minutes The exsiccated earthworms, having lost even 33.4 per cent of the weight, are wrinkled and completely motionless, but placed on moistened filter-paper they quickly revive and regain their normal size. This observation excited my interest, but absence of material in wintertime prevented the continuation of the experiments. : In the summer of 1917 I spent one month in the Experimental Agricultural Station Nikolaievskaya, belonging to the Petrograd Agricultural College. The chemical laboratory of the station is provided with all that was necessary for my study, and I profited by this occasion to continue my experiments. The question that I proposed to solve was to find out the most convenient methods of exsiccation and of revivification of the worms and to determine the percentage of the loss of water which still leaves the possibility to revive. The earthworms, belonging to the species Allolobophora foetida, were dug out in the garden and then kept on moistened filter-paper in a glass dish for two or three days, so that their gut was cleaned from the earth it contained, which otherwise would have disturbed the experiments. Before weighing, the worm was always dried with the filter-paper and then placed in a small glass dish, previously weighed on a chemical balance, and tied up with a bit of muslin. After weighing, the glass dish with the worm was placed in a desiccator with calcium chloride. Twice or three times a day all the glass dishes were weighed and the percentage of loss calculated. The exsiccation was occasionally delayed by the withdrawal of the glass dish from the desiccator. If I considered the exsiccation to be sufficient, I took out the earth- ANABIOSIS OF THE EARTHWORM 61 worm and placed it directly on moistened filter-paper for the revivification or preserved it temporarily in a corked test-tube. Before proceeding to the experiments on exsiccation I made a series of weighings in order to determine the percentage of water in the earthworms. The glass tubes with the .worms were weighed and placed in a steam-bath where they were dried at 100°C. until the weight became constant. Then the worms were dried during a certain time in the desiccator over calcium chloride and weighed once more. The results of thirteen determinations is given in table 2. TABLE 2 SER WEIGHT en tEe LIVING pate RON aes Diet: sortase grams grams per cent 1 0.7122 0.1225 82.8 2 0.8017 0.1340 83.6 3 0.9301 0.1434 84.5 4 0.8101 0.1500 81.5 5 0.8983 0.1379 84.6 6 0.5782 0.0845 85.3 7 0.5105 0.0901 82.3 8 0.6432 0.1395 78.3 9 0.6570 0.1115 83.0 10 0.5165 0.0647 87.4 el: 0.5842 0.0739 87.3 12 0.5798 0.0669 88.4 13 0.7328 0.1099 85.0 INSP Let oe Reged 2 Ee Bae Oe A ee ST aS ER) SO ag ee 84.1 The results of these weighings are very near to those of similar weighings undertaken by Miss Stchepkina and myself in 1916, when we found the earthworms to contain on the average 82.7 per cent of water. EXPERIMENTS, SERIES I The first series of my experiments consisted in the exsiccation of earthworms and their immediate revivification. The results of the exsiccation are given in the following table 3. 62 PETER SCHMIDT TABLE 3 LOSS OF ex j WRC ELD OF DAT N: RE WEIGHT OF NUMBER | srpse weicniNe | THRIVING | "‘wercursa | exsrocanron| THE EMSC. | opm cent OF grams H hours grams 1 | 25VIT17h' | 1.0133 | 29 VIT 20h 99 0.4081 | 47.5 2 | 25VIT17b1 | 0.9194 | 27 VIL11h 42 0.3114 | 66.8 3 | 25VIT17ht | 0.8605 | 27 VIT11h 42 0.2134 | 75.2 4 | 25VIT17ht | 1.2222 | 27 VIT11h 42 0.5304 | 56.2 1For brevity I use the astronomical mode of designation of time. Earthworm no. 1 was exposed to a very gradual drying, it being confined in a narrow glass cylinder. ‘Towards the end of the exsiccation (29 VII 20h) the worm had lost 47.5 per cent of its weight, was strongly contracted, became dark brown, had completely lost its mobility, but retained the elasticity of the body. After drying it was kept for thirty-nine hours (until 31 VII 11h) in a small hermetically corked glass tube and then placed on moistened filter-paper. It showed no signs of mobility, but its body was elastic and not frangible. After one hour I already observed some movements of the body. In the evening 31 VII it had completely revived. After reviving its weight was 0.9026 gram, i.e. 0.1117 gram lower than before the exsiccation. Earthworms nos. 2 and 3 were dried more quickly and lost 66.8 and 75.2 per cent of the weight of the body. After exsiccation they assumed a dark brown color and were covered with a dry, erust-like skin. Four hours after the end of the exsiccation (27 VII 14h 50’) they were placed on moistened filter-paper, but did not revive and showed no signs of life. Earthworm no. 4 lost 56.2 per cent of the weight of the body. Its upper end was overdried and covered with crust-like skin. Placed 27 VII 14h 50’ on moistened filter-paper, it showed at 15h 50’ some weak contractions and movement on the caudal end of the body, but the proximal end was much swollen. The next day this individual died. ; This first series of experiments showed that the worms can revive after the loss of nearly half the weight of their body; that is, a loss of more than 50 per cent of the water contained in the body. ANABIOSIS OF THE EARTHWORM 63 EXPERIMENTS, SERIES II The second series was not so successful as the first, perhaps because I used earthworms of a smaller size: they dried too quickly and lost their elasticity. The conditions of these experi- ments were the same as in the first series; the only difference was that the worms were not exposed to exsiceation in flat glass dishes, but in small cylindrical test-tubes. This seemed to be less advantageous. The results of the exsiccation were as follows (table 4): TABLE 4 LOSS OF WEIGHT OF WEIGHT OF WATER IN NUMBER | inet werenine | THELMVING |?’ Vercaing | exsrccation| "HE BXSIC- | PER CENT OF WORM grams hours grams 1 29 VII 14h 0.3486 | 31 VII 20h 54 0.1113 69.1 2 29 VII 14h 0.2806 | 31 VIL 11h 45 0.1419 49.5 3 29 VII 14h 0. 2425 31 VII 11h 45 0.1074 59.8 4 29 VII 14h 0.2216 | 31 VIL 11h 45 0.0983 56.6 Worm no. 1 was completely overdried, and when placed on moistened filter-paper showed no signs of life. Worms nos. 2 and 3 also did not revive. Worm no. 4, placed on moistened filter-paper 31 VII 11h, displayed after 1 hour some weak move- ments in the caudal end of the body, but its proximal end was dead. I tried to save the caudal end by cutting it off from the proximal half of the body, but notwithstanding it died next day. EXPERIMENTS, SERIES III This series was undertaken with the purpose of finding out whether it were possible to preserve the exsiccated worms and how long at normal temperature. To this end I placed them, after drying in sterilized test-tubes corked with cotton, with a cork and covered with melted paraffin. The results of the exsiccation are given below (table 5): 64 PETER SCHMIDT TABLE 5 WEIGHT OF weicur or | _LOSS OF sounen | Date orm, | musuveca | PATROREEAL | mous or | Tn‘exas | MAREE WEIGHT grams hours grams 1 2 VIII 12h 1.2478 | 4 VIII 11h 47 0.6783 45.6 2 2 VIII 12h 1.0244 | 4 VIII 11h 47 0.5289 48.3 3 2 VIII 12h 0.8028 3 VIII 19h 31 0.3946 50.8 4 2 VIII 12h 0.9067 | 3 VIII 19h 31 0.5007 44.8 5 2 VIII 12h 0.6986 3 VIII 19h 31 0.3552 49.1 6 2 VIII 12h 0.9241 4 VIII18h30’ 544 0.4946 46.4 After preservation in test-tubes for twenty-four hours, worms nos. 1 and 3 were placed 5 VIII 11h on moistened filter-paper. Both had died and did not revive. On the four other worms I detected in the morning 6 VIII many white spots—colonies of bacteria. The earthworms placed on filter-paper had all died undoubtedly by infection with bacteria. This series shows that zt 7s wmpossible to preserve for a long tume the exsiccated earthworms at normal temperature, as it is certainly impossible to sterilize them—their gut and body being full of microorganisms. EXPERIMENTS, SERIES IV The purpose of the fourth series was to establish more precisely the limit to which the earthworm can be exsiccated without loss of vitality. The experiments were arranged in the same manner as in series and II. The worms were dried in glass dishes coy- ered with muslin. The results are given in table 6. TABLE 6 : WEIGHTOF| LOSS OF pate or tHe | WEIGHT OF DATE OF FINAL HOURS OF | am EX- | WATER IN NUMBER | inst WEIGHING get ree WEIGHING pert caren PEs OEE ie grams hours grams 1 7 VIII 13h 1.5272 | 9 VIII 19h 36’ 542 0.5775 61.6 2 7 VIII 13h 0.5798 | 8 VIII 13h 24 0.1539 73.4 3 7 VIII 13h 0.7328 | 8 VIII 16h Dil 0.2825 61.4 4 7 VIII 13h 0.5165 | 8 VIII 13h 24 0.1965 61.9 5 7 VIII 13h 0.5842 | 8 VIII 16h 27 0.1725 70.4 6 7 VIII 18h 0.5782 | 8 VIII 13h 24 0. 2334 59.6 ANABIOSIS OF THE EARTHWORM 65 After the exsiccation was finished the worms were placed at once on moistened filter-paper. The revivification went on in different ways. Worm no. 1 was dried slowly and with an interruption (from 8 VIII 20h to 9 VIII 11h I kept it outside of the desiccator placing the glass dish directly under a glass bell) ; it was complete- ly motionless after the drying, but its body was smooth, elastic, and without crust-like skin. Placed on moistened filter-paper on 9 VIII 19h 40’, it had revived completely by the next morning, had a normal appearance, and displayed very energetic move- ments. Its weight was 1.3067 grams—nearly the same as before the exsiccation. Placed on earth, it dug itself in at once. On 11 VIII 11h it was placed once more on filter-paper, and after some hours I noticed that several of its posterior segments (15 to 20) were twisting themselves away from the body and were lost; certainly, they had suffered from the drying. Nevertheless, the worm was alive, and 12 VIII I used it for the second time for the experiments of exsiccation (ef. series VII). Nos. 2 and 5 were overdried, and placed on moistened filter- paper became only swollen. No. 5 showed some signs of life in the caudal end of the body, but died also. Worm no. 3 retained the cylindrical form of its body and was smooth. Its body was elastic, dark brown, and covered with a hard skin. Placed on moistened filter-paper 8 VIII 16h it became after 40’ swollen, and exhibited slow contractions when touched with the forceps. At 18h 30’ its body displayed spontaneous contractions, but at 19h 305 I noticed that the middle part of the body had become necrotic—the blood was flowing and staining the filter-paper red. Next morning (9 VIII 11h) the upper part of the body moved energetically, but the caudal end died and showed blood effusion. Worm no. 4 after drying was in its proximal part dark, and was covered with a crust-like skin; its caudal end was not so dark and more elastic. Placed on moistened filter-paper on 8 VIII 14h it exhibited at 15h 30’ some slow contractions in the caudal end. At 16h the contractions were noted also in the proximal end. At 18h 30’ the caudal end contracted energetically, the proximal THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 1 66 PETER SCHMIDT one slowly. Next morning (9 VIII 11h) the proximal end was dead, the caudal end showed signs of life and contractions. At 13h the worm died completely. This case demonstrates very clearly, that its death was caused by the overdrying of the proxi- mal end of the body, which became necrotic and, infected by microorganisms, infected also the revived caudal end. Worm no. 6 had left the glass dish during the exsiccation and was overdried in some parts of the body. Being revived it dis- played some movements, but swellings had formed here and there on its body. Its head revived and moved energetically, but in a short time the worm had succumbed. The experiments of this series have shown, that earthworms can revive after the loss of even 61.6 per cent of the weight of the body; i.e. of 73 per cent of the full quantity of water in the organism! But the revivification takes place only under favorable conditions of exsiccation, probably when it proceeds equally and gradually. When the exsiccation was too rapid, some parts of the body were overdried and the skin became crust-like. In this case evidently the blood-vessels of the skin burst during the swelling and the blood flowed out. This blood effusion caused, as it seems, an infection with microérganisms and the overdried part of the body perished, causing the death of the worm. This explains the death of worms nos. 3 and 4, when some parts of the body were revived and showed energetic movements. If the loss is more than 70 per cent of the weight of the body— i.e., more than 83 per cent of the whole quantity of water—the worm does not revive, but sometimes, nevertheless, it showed weak movements, as, for instance, no. 5. If the exsiccation was not equal and uniform in the whole body, the death of the worm was possible even at the loss of less than 60 per cent of the weight of the body (for instance, no. 6). EXPERIMENTS, SERIES V The fifth series was a continuation of the third series. As the worms kept in the test-tubes died evidently through infection with microorganisms, I determined to try whether the dried worms could not be preserved at a low temperature. To this = ANABIOSIS OF THE EARTHWORM 67 end the worms were placed after the exsiccation in test-tubes upon the ice of an ice-house, where the temperature was about +1°C. The results of the drying are given in table 7. TABLE 7 WEIGHT OF HOUBBOM| oe on ee OF weamen | pgrhgicume |[THELTING| — "“Vacaine'” | BXBICEA: | iocanno.[exicontr on grams hours grams 1 10 VIII 17h 0.5066 | 11 VIII 14h 21 0.1967 61.1 2 10 VIII 17h 0.7472 | 11 VIII 18h 30’ 255 0.3206 57.0 3 10 VIII 17h 0.5003 | 11 VIII 17h 24 0.1677 66.4 4 10 VIII 17h 0.5605 | 11 VIII 14h 21 0.2150 61.6 5 10 VIII 17h 0.5084 | 11 VIII 14h 21 0.2167 57.3 1) 1 WARE st7am 0.6656 | 12 VIII 20h 51 0.2527 62.0 Worms nos. 1, 4 and 5 were placed after drying in test-tubes and carried to the ice-house. Worms nos. 3 and 6 were placed immediately on moistened filter-paper. No. 2 was left in the laboratory at the normal temperature of about 20°C. (in day- time). The worms placed upon the ice (nos. 1, 4, and 5) were kept there for fifty-one hours (till 13 VIII 16h 20’) and seemed then to be completely normal without signs of infection. Nevertheless, the revivification was unsuccessful. Placed on moistened filter-paper 13 VIII 16h 30’ they displayed at 17h weak movements in the proximal end and in the tail. Worm no. 1 at 18h 45’ exhibited energetic contractions of the proximal end, but its tail moved weakly and the middle part of its body was motionless. Worm no. 4 was dead. No. 5 dis- played energetic contractions in the tail and in the head, but the middle part of the body with clitellum had died. Next day (14 VIII 10h 30’) the proximal end of worm no. 1 and its tail contracted very energetically, but the middle part of the body died. Clitellum showed blood effusion and the filter- paper beneath was red. The proximal end and the middle part of the body of worm no. 5 died, but its tail contracted. In the evening 14 VIII both worms were definitely dead. 68 PETER SCHMIDT The ill success of these experiments was caused, I believe, by the fact that I had occasionally used earthworms with clitel- lums. The skin of the clitellum is very delicate and full of blood- vessels, which suffer from the exsiccation and cause the blood effusion and infection. The revivification of the control worms nos. 2, 3, and 6 also yielded nearly negative results. Worm no. 2, preserved in a test-tube at normal temperature during twenty-four hours, was placed on 12 VIII 18h 50’ on moistened filter-paper. It seemed to be in a good state of preservation and showed no signs of infection. But as it was swollen, only some weak contractions of the caudal end were seen, and at the proximal end white swellings appeared, as a sign of infection. On the morning 13 VIII the worm was dead. The worm no. 3 was exsiccated too quickly and overdried. It showed no signs of life and at the proximal end a blood effusion was observed. Worm no. 6 proved to be overdried at the proximal end, evi- dently also because there was a clitellum. The caudal two- thirds of the body was revived and moved energetically, but the proximal third perished. After twenty-four hours the whole worm was dead. This fifth series of experiments has nevertheless shown, that the worms which have lost not more than 61 per cent of the weight of their body, if preserved at low temperature, retain their vitality during forty-eight hours. Their revivification was only partly obtained, but this depended upon secondary causes, which could be avoided. Unfortunately, lack of time prevented me from re- peating this series of experiments with all necessary precautions, needed after the foregoing ill success. Earthworms without cli- tellum should be taken and exsiccated more slowly and gradually. EXPERIMENTS, SERIES VI The sixth series was undertaken with the view of studying the influence of exsiccation effected at low temperature. The worms were put into small glass-tubes and placed into a desiccator that stood on the ice in the ice-house. The small size of the glass tubes caused a very slow exsiccation of the worms. ANABIOSIS OF THE EARTHWORM 69 The results of the exsiccation were as follows (table 8). TABLE 8 | WEIGHT OF LOSS OF wks DATE OF THE Ve eRe ns TOES DATE OF FINAL OES OF THE EX- WATER IN _ NUMBER | iRsT WEIGHING Rate Meaty WEIGHING Hare BIECREED PRE CUNT oe grams hours grams 1 IS Wiiiwtsh 0.63851 | 16 VIII 15h 30’ 155 0.2965 53.3 2 13 VIII 15h 0.4306 | 17 VIII 18h 99 0.1740 09.5 3 (Savina oh 0.4180 | 17 VIII 18h 99 0.1717 59.0 4 13 VIII 13h 0.3294 | 14 VIII 19h 30 0.1196 63.6 Worm no. 1, placed 16 VIII 15h 30’ on moist filter-paper, showed at 16h 40’ energetic contractions of the body, and in the evening (19h) had completely revived. Worms nos. 2 and 3 placed 17 VIII 18h 15’ on moist filter- paper showed at 18h 35’ the first contractions of the proximal part of the body. At 18 VIII 10h both worms were completely revived and moved with great energy. Worm no. 4 placed on moist filter-paper at 14 VIII 19h 30’ showed contractions at 22h, but died next morning (15 VIII 11h). This sixth series shows, that the combination of exsiccation and low temperature gives the most satisfactory results. Owing to the slowness of the exsiccation and the retardation of the activity of microorganisms the treatment affects the vitality of the worms in a very small degree and the loss of water is easily endured. Lack of time has not allowed me to repeat this series of experi- ments and to try the preservation of earthworms exsiccated in this manner at a low temperature. It is possible that the results might be more successful than in the foregoing series. EXPERIMENTS, SERIES VII In this last series I intended to try the exsiccation of worms that had already undergone this operation and had been revived. But I could use for this purpose only worm no. 1 of series IV. This earthworm was weighed on 12 VIII 19h and its weight was 1.2610 gr.—the diminution of its weight as compared with 70 PETER SCHMIDT its weight on 7 VIII 13h (cf. IV series) was caused by the loss of its posterior segments after the first exsiccation (p. 10). Placed in the desiccator on 12 VIII 19h, it was kept there until 14 VIII 19h (i.e., forty-eight hours) and its weight diminished on 0.4679 grams or 62.6 per cent of the weight of its body. At 19h 30’ it was placed on moistened filter-paper and at 22h some contrac- tions and movements of the proximal end and of the tail were observed. But on the next morning (15 VIII 11h) it was found dead. On the proximal end hemorrhage and swellings could be seen. It is possible that the exsiccation had surpassed the lhmit or had gone on too rapidly. This experiment is of course not sufficient for denying the possibility of repeated revivification of earthworms. CONCLUSION All the mentioned experiments, which for lack of time I had no possibility of finishing, prove, as I believe, that the phenom- ena manifested in the exsiccation of earthworms are completely analogous to the results of exsiccation of Tardigrada, Rotatoria, and Nematodes, and with full right may be called ‘anabiosis’ (‘over-life’). Actually by the exsiccation the earthworms lose completely their mobility, their size diminishes to one-half or one-third of their length and volume and they show no manifestations of life.. In the dorsal vessel, sometimes well seen through the skin, no contractions can be detected with a microscope. The seg- ments of the body are also completely motionless. The exsic- cated worm is dark brown, but must retain the elasticity of its body and its skin must be soft if it is to revive. It has the ap- pearance of a corpse ora mummy. In this state the worm can retain the capacity for revivification for thirty-nine hours at normal summer temperature (cf. worm no. 1 of the experiments in series 1) and according to series V and VI, for 48 hours, and perhaps more, at low temperature. It is possible, that at suit- able low temperatures one can preserve the vitality of the exsic- cated worms during a very long time. I shall undertake experi- ments in this direction at the first opportunity. ANABIOSIS OF THE EARTHWORM 71 Thus, only one difference can be stated in the earthworms as compared with the other groups of anabiotic animals—it is that they are not so amenable to continued preservation in the exsic- cated state at normal temperature. This is evidently accounted for by the more complicated organization of worms and the presence of a more highly organized blood system, as well of a large quantity of microorganisms in the gut. The exsiccation of the skin carried on too far destroys its capillary vessels and causes the blood effusion. The microorganisms of the gut and of the surface of the body, multiplying under favorable con- ditions, bring about the death of the worm. While showing a full analogy to the anabiosis of Rotatoria, Tardigrada, and Nematodes, the phenomenon of exsiccation of the worms has the practical advantage of affording the possi- bility of determining the amount of the loss of water contained in the body of the worm. And in this direction I have discovered a fact, which seems to be of great interest: a very large percentage of water can be lost without the complete loss of vitality. As my experiments (cf. series IV, worm no. 1) have evidenced, earthworms can revive and regain the normal state of life after a loss of 61.6 per cent of the weight of the body, or nearly 73 per cent of the weight of the water contained in the body. If we take into consideration that the organization of the earthworm is comparatively infinitely more complicated and therefore more delicate than the organization of Rotatoria, Tardigrada and Nematodes, we can readily admit, that these microscopical animalcules may have the capacity to revive after having lost 80 to 85 per cent of the amount of water in their bodies, or perhaps even more. This consideration throws some light on the seeming mysteri- ousness of the phenomenon of anabiosis that was discovered more than 200 years ago, but till now no known analogy in the higher groups of the animal kingdom. Petrograd, February 1, 1918. 72 PETER SCHMIDT LITERATURE CITED Baker 1764 Employments for the microscope. Broca, P. 1860a Etudes sur les animaux resuscitants. Presse Scientifique, T. 1, pp. 211-222. 1860 b Rapport sur la question soumise 4 la Société de Biologie au sujet de la reviviscence des animaux desséchés. Paris. Mémoires de la Société de la Biologie, T. 2, pp. 1-140. DavaIneE, C. 1856 Recherches expérimentales sur la vitalité des Anguillulides du blé niéllé. Paris. Comp. Rend. Ac. Sc., 48, pp. 148-152; Ann. Nat. Hist., 18, pp. 268-269. Dover 1842 Mémoire sur les Tardigrades. Annales des Se. Naturelles, Paris, ‘II ser. T. 18 (Zool.), pp. 5-35. FROMANTEL, E. pE 1877 Recherches sur le revivification des Rotiféres, des Anguillulides et des Tardigrades. Associat. Frang. p. l’ avane. d. sciences. C.R.dela VI sess. Le Havre, pp. 641-657. GavarreT, M. 1859 Quelques expériences sur les Rotiféres, les Tardigrades et les Anguillulides des mousses des toits. Ann. d. Se. Natur. Paris, IV ser. T 11 (Zool.), pp. 315-330. LEEUWENHOEK, A. 1701 Epistolae ad Societatem Regiam Londinensem, vol. 2, p. 381 ss. Lugd. Bat. Neepuam 1745 New microscopical discoveries. London 8°. Puatre. L. 1885 Beitrige zur Naturgeschichte der Rotatorien. Jenaische Zeit. fee Nate Bde wl Or ape alse Preyer 1891 Uber die Anabiose. Biolog. Centralbl., Bd. 11, no. 1, 2. SPALLANZANI 1777 Opuscules de Physique animale et végétale. Trad. par J. Sennebier, T. 2, p. 224 ss. Scumipt, P., et SrcHePKINA, T. 1917 Sur l’anabiose des vers de terre. Comp. Rend. Soc. Biol., Paris, 1917, pp. 366-368. Scuutz, E. 1915 On some investigations upon anabiosis (in russ.). Isvest. Petrogr. Biol. Labor. (Bulletin du Laborat. Biologique de Petrograd), vol. 15, pp. 81-86. ZacHartas, O. 1887 Kénnen die Rotatorien und ‘Tardigraden nach vollstindi- ger Austrocknung wieder aufleben oder nicht? Biolog. Centralbl., Bd. 6, pp. 230-235. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, SEPTEMBER 16 STUDIES OF NORMAL MOULT AND OF ARTIFICALLY INDUCED REGENERATION OF PELAGE IN PEROMYSCUS H. H. COLLINS Scripps Institution, La Jolla, California FIFTEEN FIGURES CONTENTS NGO CG UIOM ores Senet erties sini cys See Se cs Oe cn fe ret mea arene 73 Dee POSTE EN al MIGENEY ot)... sR es a eas Nook cue eee eee 76 SReeeliot MMOL te ketoe Seyi ee 22 (chy MARI Facle's AM cheese 1 2s Se ae 81 44 Regeneration, of pelazean juvenaltmicer to. ia/45. 6.3m. ociened hese bee iaet 81 Ha nmereneration, Of adult pelages: . .....:.: vmware ts Us)" scaler ie Scams = emdog 86 GREDISCUSS TOTES rene Ey RTA. 3: 0 DORMER RC aie co oliekes stele tee ave oes ated ances er ses 88 Mee SMITNTNT aN Rises SLUTS e «nies tk ie Sta heehee BE Be eonS senate atc ce ops citar 94 1. INTRODUCTION It is a matter of common knowledge among naturalists that many birds and mammals undergo more or less marked ‘changes in appearance as the result of moult. These moults may mark different stages in the life cycle of the individual or they may be seasonal in character. In general the greatest changes in appearance are incidental to the transition from the juvenal to the adult. Especially among birds, there are many species in which the differences are quite striking. To quote Allen (’94a) in regard to Passerine birds: This ‘first’ or ‘nestling’ plumage can usually be recognized by its loose, fluffy texture, as compared with that of adult birds of the same species, even though the coloration may be similar; but generally it differs notably also in color, and often in pattern of markings, from that which immediately succeeds it, or from any plumage which may be afterward acquired. Familiar illustrations are furnished by the robin and the bluebird, where the first plumage is so strikingly unlike, both in color and markings, that of the adult bird of either sex (p. 92). 73 74 H. H. COLLINS In a later paragraph, he says, ‘‘Although this first plumag is particularly interesting and instructive, affording frequently clues to ancestral relationships, it has not until recently attracted the attention it deserves, even among ‘professional’ ornithologists.”’ Whitman (04) in his work with pigeons, appears to have re- garded the comparative study of juvenal plumages as highly important in tracing phylogenetic relationships. On the other hand, it is said that closely allied subspecies sometimes differ more at this early stage than at any later period. Apparently there is still much to be learned concerning the real significance of first plumages. Though the general features of the change of pelage have been described in many of the mammals, the ornithologists seem to have proceeded somewhat farther in the study of the details of the process. Dwight (’00), in a study of the Passerine birds of New York, writes as follows: The plan on which a moult proceeds is a perfectly definite one, al- though often much modified and obscured. Old feathers or rows of feathers tend to remain until the newcomers adjacent have matured sufficiently to assume their function, when the old fall out and their places are taken by the new which develop from the same papillae. The systematic replacement of areas of feathers shows most obviously in the wings where not only do the remiges fall out one after another in definite sequence and almost synchronously from each wing, but the greater coverts are regularly replaced before the fall of the secondaries beneath them, the lesser coverts before the median and even in the rows of the lesser coverts alternation seems to be attempted. . . . On the body the protective sequence is less obvious, but the moult regularly begins at fairly definite points in the feather tracts radiating from them in such manner that the outer rows of feathers where the tracts are widest and the feathers of their extremities are normally the last to be replaced (pp. 88, 84). Furthermore, Dwight found a regular sequence in the de- velopment of the various feather tracts, although in young birds an outbreak of moult in any of the tracts earlier or later was less unusual. Although, as a rule, the moult proceeds so gradually and so simultaneously on opposite sides of the body that the power of MOULT AND REGENERATION OF PELAGE IN MICE 75 flight is not impaired, there are cases in which the process is not so plainly adaptive. The Duck family (Anatidae), among others, according to Coues, drop their wing quills so nearly simultaneously as to be for some time deprived of the power of flight. The details of the process of moult are not so well known in the case of mammals. However, it appears that in general the process is more irregular than in birds. According to Allen (’94): “As a rule, particularly among the Rodentia, the change be- comes first apparent on the feet and about the nose extending gradually up the limbs and over the head and from the base of the tail anteriorly, and from the sides of the body toward the median line.’”? This appears to be the usual method especially in the spring moult, but the process is said to be ‘‘subject to much irregularity, even among individuals of the same species, and it seems to vary somewhat in different groups” (p. 107). In describing the condition in rabbits Nelson (’09) says that “the moults usually begin about the head and feet and proceed more or less irregularly over the body, but there is no absolute rule, and patches of new pelage may appear on any part of the body, especially if the old coat has been thinned by abrasion or other local cause’ (p. 30). However, it appears that certain other students of the genus do not find the process of moult as irregular as described by Nelson. According to Barrett-Hamilton (712), the order of change in the European hares, though not invariable, generally follows a fairly regular sequence. In the autumnal moult, the feet and legs, the gray parts of the ears and parts of the head are first to undergo the change. Then follows the rump, and the white area of the ventral surface gradually creeps upward on the sides until the brown of the summer coat is extinguished or remains as a “‘small island or islands.’’ In the spring the sequence and directions of growth are completely reversed, the new pelage appearing first on the head and median dorsal region, growing downwards. This same reversal is described by Allen in his paper on the changes of pelage of the varying hare (Lepus ameri- canus). 76 He) He: ‘CORLENS Passing to the Muridae, it may be said that although certain species of this group have long been reared in captivity and used extensively in experimental work, very little is known concerning their changes of pelage. Probably the most complete account is that of Osgood (’09) in his monograph of the genus Peromyscus. As stated by Osgood, the mice of this genus pass through three fairly distinct phases due to age—the juvenal (young in first coat), the adolescent, and the adult. According to his descrip- tion of the assumption of the postjuvenal or adolescent pelage, “This [i.e., juvenal] stage is succeeded by the adolescent pelage, which first appears on the middle of the sides. Its growth pro- ceeds rapidly upward on each side until union is effected in the middle of the back, and then incloses the rest of the body, the rump and nape, usually being the last parts to be covered”’ (p. 20). With reference to seasonal changes, he says: The new pelage may be acquired in regular and obvious manner with the fresh coat well distinguished from the old worn one, the growth proceeding from before backward and the middle of the rump being the last part to be invested, or the change may be quite insidious and apparent only upon careful examination. The regular method is fol- lowed in the adults of most species, while the other is more often evident in immature individuals (p. 19). My own observations in this field have been, in the main, con- fined to a few species of this same genus. During the past year and a half I have devoted considerable attention to a study of the seasonal and life-cycle changes in the pelage of several races of California deer-mice, reared in the murarium of the Scripps Institution. It is the purpose of the present paper to discuss, somewhat in detail, the normal process of moult, especially with reference to the assumption of the postjuvenal pelage, and, furthermore, to describe certain modifications of this process experimentally induced. I take this occasion to express my sincere thanks to Dr. F. B. Sumner for many valuable suggestions and criticisms. 2. THE POSTJUVENAL MOULT The description of this moult is based upon an examination at weekly intervals of a series of twenty specimens of the first cage- born generation of Peromyscus maniculatus gambeli (Baird). MOULT AND REGENERATION OF PELAGE IN MICE ae Frequent examinations, somewhat more at random, were also made upon the general stock, including the subspecies sonoriensis (Le Conte).and rubidus Osgood, as well as a few specimens of P. eremicus fraterculus (Miller) and P. californicus insignis Rhoads. By etherizing the animals and parting the fur, it was possible to follow the moult from the time of the first appearance of the new hairs through the skin. At birth the body is devoid of hair and pigment except for the vibrissae and supraorbital cilia. On the second day the upper parts begin to assume a bluish-black coler and the hair may be seen coming through the skin of the pigmented area. A day or two later, the ventral white hair may be observed. At the age of four to five weeks, the young are, as a rule, in full juvenal pelage. There are no further traces of pigment in the skin, which is now flesh color. This pelage, like the later ones, is made up of a fine soft underfur and a thinner coat of much longer and coarser overhair. As is the case in the adult pelages of many other rodents, the hairs of the underfur are banded or ticked (agouti), being of a blackish plumbeous or slate color basally, with a narrow subterminal zone of pallid mouse gray, while the tips are black.1 The overhairs are not of the agouti type, lacking the subterminal band. The general effect on the dorsal surface may be described as between neutral and deep neutral gray. The juvenal pelage of the ventral surface, like that of the dor- sum, is made up of underfur and overhairs. Basally, the color is the same as in that of the dorsal surface, but the distal region is white. The lateral line of demarcation between the dorsal and ventral surfaces is very sharply defined (fig. 5). The microscopic structure of the hairs in the juvenal pelage is essentially the same as described by Sumner (718) for the adult. There is, however, a very evident difference in the proportionate number of the different kinds of hairs. The slender hairs with but a single axial row of pigment bodies, alternating with the air spaces, are much more numerous, while the yellow pigment is much reduced in the subterminal bands. The overhairs are 1 Color descriptions are based on Ridgway’s key. 78 H. H. COLLINS attenuated at the base, being no larger at this level than the hairs of the underfur. Both kinds of hairs are very much flattened, but the larger ones show no local attenuations such as are de- seribed by Barrett-Hamilton (’16) for Mus musculus, though this appearance may be simulated by torsion. In microscopic structure, the vibrissae are markedly different from the hairs just mentioned. Though these are the largest hairs on the body, there is but one axial row of lacunae containing a relatively small amount of pigment. Most of the pigment occurs in the cortex as. small granules arranged in longitudinal striae. This cortical pigment extends to the base, but gradually disappears toward the tip. This is the reverse of the arrange- ment in the body hairs, in which the greater part of the pigment is found in the axial region arranged in from two to four rows of lacunae in all except the smallest hairs. Furthermore, in the body hairs, the cortical pigment, which is restricted mainly to the superficial region of the cortex, is most dense in the terminal zone, gradually disappearing toward the middle region. The lower vibrissae are devoid of pigment almost or quite to the base, but this terminal white region becomes much reduced dorsally. The structure of the two supraorbital cilia and of the hairs of the tail is similar to that of the vibrissae. In certain pelage removal experiments to be described later, it will be noted that vibrissae and body hairs are not regenerated in the same manner. This fact suggests the possibility of some sort of correlation between the morphological and physiological differences. The transition from the juvenal to the postjuvenal pelage usually begins at the age of six weeks and is completed about eight weeks later. The new pelage first appears on the throat near the angle of the jaw, or rarely on the.anterior surface of the forelimb along the lateral line.2 Growth proceeds toward the median ventral line of the head and, at the same time, anteriorly under the eye and ear and posteriorly over the forelimb and shoulder. From these regions, it passes posteriorly above the ventral white to the hind limbs, at the same time creeping up toward the dorsal median line (figs. I to 3). 2 The line of demarcation between the white hair of the ventral surface and the dark hair of the dorsum. MOULT AND REGENERATION OF PELAGE IN MICE 79 Figs. 1 and 2 Diagrams of the dorsal and ventral surfaces, showing directions of growth in the postjuvenal moult of P. maniculatus gambeli. The regions on which moult proceeds more or less independently are shown by the dotted lines. The longer arrows indicate more rapid growth. J.l., lateral line; p., point of origin. Fig. 3 Diagram of lateral surface, showing directions of growth in the post- juvenal moult of P. maniculatus gambeli. Symbols as in figures 1 and 2. SO H. H. COLLINS The moult may be well under way before there are any evi- dences of it on the surface. The details of the process can be learned only by parting the overlying juvenal pelage and obsery- ing the new hair as it comes through the skin. The new post- juvenal pelage is first seen on the surface, usually on the fore- limbs, and somewhat later as triangular areas on the sides. These lateral areas gradually become confluent, first, as a rule, just posterior to the shoulders (figs. 5 and 6). After this ‘saddle phase’ (fig. 6) has been reached, further growth for days or weeks may be limited to the region posterior to the saddle. The direction of growth is posterior and, at the same time, upward on the hind limbs from the lateral line, the region above the base of the tail being the last to undergo the change (figs. 3, 7, 8). On the ventral surface, the moult is regularly completed before that of the dorsum. As shown in figure 2, growth proceeds from the throat posteriorly. In many cases, there may be no super- ficial indications of the change, though in some instances a de- finite moult line? may be observed. The moult is now completed over the whole body surface, except the region extending on the dorsal surface from the tip of the snout to the shoulders. The investment of this area may occur soon after that of the rump, but usually only after an in- active period which in extreme cases may be as long as two months. In this region the postjuvenal pelage first appears, anteriorly, on the tip of the snout, passing posteriorly to the eyes, thence as two diverging strips to the anterior insertions of the ears, the intervening space being filled in by lateral and posterior growth (fig. 1). Posteriorly, the moult line moves from the shoulders forward toward the ears, where the two areas coalesce. Growth in the two directions may occur simultaneously or that of one region may be slightly in advance of the other. The postjuvenal pelage is somewhat longer and coarser, though still shorter than that of the adult, which it closely resembles in color and texture. The general color effect is quite different from that of the juvenal. It»may be described as varying from 3 The line separating the old and new pelages. MOULT AND REGENERATION OF PELAGE IN MICE S81 Saccardo’s umber to sepia, with the dorsal median stripe more or less strongly marked with black. This difference in the general color of the two pelages appears to be due mainly to the increased amount of yellow pigment in the subterminal bands of the post- juvenal hairs. The postjuvenal moult of two other races of California deer- mice (P.m. sonorienses and rubidus) as well as in a ‘yellow’ mutant of gambeli, which appeared about two years ago in the murarium stock is essentially the same as above described. In two other species of Peromyscus (californicus insignis, and eremicus fraterculus) which occur in the vicinity of La Jolla, the process is quite similar, but upon a closer examination there appear to be certain characteristic differences in the points of origin and directions of growth. 3. LATER MOULTS In general, in the assumption of the adult pelage and in the seasonal moults, the process is the same as above described. However, these later moults appear to be somewhat more irregu- lar, and frequent partial moults further complicate the situation. These moults will be treated more at length in a later paper. 4. REGENERATION OF PELAGE IN JUVENAL MICE After having observed the rather marked regularity in points of origin, sequence, and directions of growth, in the assumption of the postjuvenal pelage, it seemed worth while to determine to what extent, if at all, the process might be modified by the arti- ficial induction of regenerative processes. In the series of experiments here deseribed, a total of about forty gambeli were used, varying in age from two and one-half to seven weeks. The mice were etherized and the pelage was re- moved by plucking out with the fingers. The hair is quite loose, especially just previous to a moult, and may be very readily removed in this manner without injury to the skin. Except in those cases where the moult was too far advanced, the following regions were operated on in every individual: THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 1 82 H. H. COLLINS 1. The dorsal median region of the head from the tip of the snout to the base of the skull. 2. The hips and thighs. 3. About 1 sq. em. between and anterior to the forelimbs. 4, About 1 sq. em. in the midventral region, just anterior to the hind limbs. Although individual differences were in some cases quite noticeable, in general the mode of replacement on the depilated areas was much the same. The history of one brood (offspring of 2 106) may be regarded as fairly typical. The three members of this brood were oper- ated upon, as outlined above, at the age of eighteen days. At that time they were in full juvenal pelage. Without exception, the skin of the posterior half of the dorsum was still dark—an indication that the growth of the juvenal hair was still in prog-’ ress. The denuded areas on the head and ventral surface, however, were devoid of pigment. Two days later, the pigment in the skin of the dorsum had almost entirely disappeared, except on the exposed area on the hips. Here a rather striking effect was observed. The skin was as dark as when the pelage was first removed, the line of demarcation between the two areas being very clearly defined. On the seventh day, the exposed skin on the head was begin- ning to darken. That of the hips was somewhat lighter than when last observed, though it was still much darker than the surround- ing skin. On the ventral surface, there was a slight darkening of the skin of the posterior area. Ten days after the operation, the pigment in. the skin of the depilated region on the hips had not wholly disappeared. In the meantime the exposed skin on the head had become much more intensely pigmented. The throat patch remained flesh color, while the posterior ventral area was slightly darkened. When the brood was examined on the seventeenth day, some rather marked individual differences were noted. In one case the depilated area on the head was entirely covered with post- 4 Normally, the pigment disappears and the skin becomes flesh color after the hair attains its full length. MOULT AND REGENERATION OF PELAGE IN MICE 83 juvenal pelage, about half grown out, almost uniform in length, though slightly longer posteriorly. Those of the ventral surface were also wholly covered with pelage of about uniform length. In the meantime the normal postjuvenal moult had appeared on the throat near the point of the jaw, extending under the ear and down on the anterior face of the forelimb entirely outside the depilated region and in typical fashion. At the same time post- juvenal pelage, barely through the skin, was found extending as a narrow band on the lateral line from the base of the tail to the forelimb on one side of the body, and from the hind limb to the . forelimb on the other. Both strips were continuous with the new pelage coming in on the posterior ventral area. Never having observed this condition in a normal moult, I am inclined to regard the premature appearance of postjuvenal pelage in this region as an abnormality resulting from the operation. At this time the depilated area on the hips was still bare. In the other two members of the litter the normal moult had not appeared by the seventeenth day. The depilated regions of the head were covered with a uniform growth of postjuvenal pelage, but the hip regions showed no indications of regeneration. Within a few days postjuvenal pelage appeared on the hips where the juvenal pelage had been removed, and a month after the operation the new pelage on all the depilated areas was fully grown out. At this time both surviving members of the brood were approaching the ‘saddle phase’ of the normal moult (fig. 12). A comparison of the foregoing description with that-of the normal moult will show the marked extent to which the normal process has been modified by artificially induced regeneration. If replacement on the depilated areas were to follow the normal sequence, the order would be as follows: 1) throat; 2) posterior ventral area; 3) dorsal lumbar region; 4) dorsal head region. Moreover, as already noted, the dorsal head region is normally invested some weeks after the appearance of the new pelage on the throat and forelimbs. It will be observed that growth on the depilated regions is much more nearly simultaneous than is normally the case. Fur- thermore, the sequence is not the same. Regeneration on the 84 H. H. COLLINS head precedes replacement on the hips and hind limbs—an in- version of the natural order. Growth is almost simultaneous on the two denuded areas of the ventral surface, while the pelage of the hips was last to be replaced. It will be noted also that the mode of replacement on the head was radically different from the normal process. Typically, erowth proceeds, a) dorsally, from the tip of the snout to the anterior insertions of the ears, or, in some cases, posteriorly to a point midway between the ears; b) from the shoulders, anteriorly over the back of the neck to the ears. In contrast to this con- dition, in regeneration the new hair appears almost simultane- ously over the whole of the depilated area. It is but natural to suppose that the normal process of growth would be less profoundly modified were regeneration to occur immediately before or during the normal moult. But this does not seem to be the case. In mice operated upon at the age of six weeks, with the normal moult well under way, regeneration occurred in essentially the same manner, as regards sequences and directions of growth. Even in an extreme case, in which the entire body except the head and rump were covered with post- juvenal pelage, replacement on the head was somewhat in advance of that on the rump. Replacement on the head was fairly uniform, though the snout below the eyes was last to be invested—an inversion of the normal condition. In exceptional cases the normal order is not so completely dis- guised,.as shown by the history of another brood (that of 2 105), the four members of which were operated upon in the same manner and at the same age as the former brood. The first suggestion of the normal process, such as would occur without operation, appeared on the depilated area of the throat about a week after the hair was removed. Here there was a perceptible darkening at the point of the jaw and along the anterior face of the forelimb, thus outlining the region where the postjuvenal pelage normally first appears. At the same time the rest of the exposed area showed no signs of pigment formation. Within a few days incoming pelage appeared on this pigmented area, the rest of the region being covered some days later. Typi- MOULT AND REGENERATION OF PELAGE IN MICE 85 cally, as has been already indicated,’ replacement after depilation occurs simultaneously over the whole region. A further reminder of the normal process was noted in the mode of regeneration on the head. Instead of the fairly uniform replacement usually observed after depilation, there were two independent centers of growth. Pigment (followed in a few days by the incoming hair) first appeared on the snout just below the level of the eyes, spreading anteriorly to the tip and posteriorly to a definite transverse line connecting the anterior insertions of the ears, the skin posterior to the line remaining unpigmented for several days. In another member of the brood pigment developed at the same time on the snout and back of the ears, leaving a small patch of pigmentless skin between them. The incoming postjuvenal pelage does not ordinarily trans- eress the limits of the depilated region and, as a rule, the mode of change on the rest of the body surface remains unmodified. Occasionally, however, there is apparently some ‘action at a distanee.’ One such ease has already been cited. In another instance, where the juvenal pelage had been removed from the hips, the incoming postjuvenal, passing beyond the denuded region, extended anteriorly to the middle of the dorsum, meeting at this point the normal moult which was proceeding in the opposite direction. In other words, the normal direction of growth had been reversed on a part of the dorsum which had not been operated upon. In all cases where the juvenal pelage was removed it was replaced by the postjuvenal. However, I have noted what at first appeared to be an abortive attempt to regenerate juvenal hairs. During the week following the removal of the juvenal pelage (in those cases only in which the skin was dark at the time) a growth of fine short blackish hairs appeared on the denuded areas. This was succeeded in a few days by the incoming post- juvenal hair. ® See page 83 above. ® See page 83 above. 86 H. H. COLLINS When subjected to microscopical examination, it was found that these hairs represented the basal portions of juvenal hairs, broken off in the skin at the time of the operation. In most cases, the tips plainly show evidences of fracture. It is interesting to note that in many of these hairs the localization of pigment has been modified to a marked extent. Evidences of the segmental arrangement may be partially or wholly obliterated. The amount of pigment appears to be some- what greater than in the normal hair at this level. This increase, however, may be more apparent than real, asa result of the more diffuse arrangement of the granules. 5. REGENERATION OF ADULT PELAGES In addition to the foregoing studies of regeneration of hair in juvenal mice, a number of rather incidental observations were made on adults. In order to facilitate the study of the details of normal moult, the old pelage was removed by clipping close to the skin. In all, seventeen adult gambeli were included in this series. Ten of these mice were clipped over the whole body, including head and limbs, while in the remaining cases the hair was removed from one side only. The mode of replacement was so irregular that the primary object of the experiment was not attained. However, the results from the point of view of regeneration may perhaps be of some interest. Although obscured by various irregularities, there were some vestiges of the normal process. The first evidence of growth was seen, as a rule, on the throat and forelimbs. Replacement on the ventral surface preceded that on the dorsum, while, in general, the moult’ proceeded from before backward. One of the most striking and characteristic departures from the normal process was the appearance of more or less numerous small isolated patches of hair over the posterior half or two- thirds of the body. These ‘hair islands,’ which were usually 7 Since the process includes the shedding of the clipped hair, the term ‘moult’ may properly be used in this connection. MOULT AND REGENERATION OF PELAGE IN MICE 87 more in evidence on the dorsal surface, appeared simultaneously with the new pelage on the throat and forelimbs. Following this partial restoration after the new hair had at- tained its full length there were no further signs of regeneration for a period varying in different individuals from one to four months. Then new hair appeared at the same time on all the bare spots between the ‘islands,’ except on the rump, where, in most cases, replacement was still incomplete when the animals were last examined, five months after the operation. Although there appears to be no subsequent growth of the body hairs, which were fully grown out at the time of the opera- tion, it is worthy of note that the restoration of the vibrissae is accomplished by the elongation of the identical hairs which were cut. Although, as previously pointed out, there are cer- tain structural differences between these two types of hairs, the cause of the difference in the mode of regeneration is not at all evident. In those individuals in which the pelage was removed from one side only, the mode of replacement was essentially the same. Apparently, at least within these limits, there is no correlation between the rate and mode of regeneration and the size of the area operated upon. While engaged in a study of seasonal changes in color due to fading, abrasion, and other causes, it seemed highly desirable to be able to compare old and new pelages on opposite sides of the body of the same individual. With this object in view, a series of twenty adult gambeli were trapped in worn and faded pelage, a few weeks prior to the autumnal moultingseason. The animals were etherized and the old hair was plucked out on one side only, from the dorsal median line to the lateral line and from the tip of the snout posteriorly to the base of the tail. Although this series of experiments will be described more fully in a later paper dealing with color variations, they may be briefly mentioned here in connection with the topic of regeneration. Replacement occurred much more rapidly than in the preceding series. With but one exception, complete restoration was ac- complished one month from the date of the operation (fig. 15). 88 H. H. COLLINS Replacement occurred at about the same time over the whole area, though that on the forelimbs and throat was slightly in advance and the hip and hind limb were covered last. With but few exceptions (probably not observed at the right time), ‘hair islands’ were seen particularly on the posterior part of the dorsum, but they were obliterated within a few days by the appearance of hair on the intervening patches. The differences in the appearance of the two sides of the body are slight and in no case comparable with the contrast in the coloration of individuals representing buff and dark extremes within the species. 6. DISCUSSION A study of the details of moult in living juvenal Peromyscus discloses a greater regularity in the process than appears to be characteristic of adult mammals in general. It is found that the change occurs more or less independently on different parts of the body, suggesting tracts somewhat com- parable to the pterylae in birds. ‘This is most clearly seen in the method of moult on the dorsal surface of the head where growth proceeds from the neck anteriorly, and from the tip of the snout posteriorly to the ears. Then again, moult appears independ- ently, although synchronously on opposite sides of the body. While the marked regularity of replacement of feathers on the wings of birds may be regarded as an adaptation for the preserva- tion of the power of flight, the sequence of moult on the various pterylae, of the body proper, and the similar phenomenon in mice as well could scarcely be interpreted in the same light. In this connection, Dwight (00) writes as follows: The important part that the blood supply plays in this plan ap- pears to have been quite overlooked nor have I had the opportunity to fully investigate it. I may say, however, that the radiation of the moult from given points corresponds very closely to the distribution of the superficial arteries, beginning where the main trunks come to the surface and ending with their ultimate ramifications (p. 84). This same idea is suggested rather indirectly by Schultz (16), who regards the color markings of mammals as due to differen- MOULT AND REGENERATION OF PELAGE IN MICE 89 tials in growth which are due, in turn, to inequalities in the pe- ripheral blood supply. We shal! consider this theory more at length in the discussion of his studies of the regeneration of hair. In the species of Peromyscus studied I find some deviations from the process of moult as described by Osgood (’09) in his monograph of the genus. With reference to the postjuvenal moult he writes: ‘‘This [i.e., juvenal] stage is sueceeded by the adolescent pelage, which first appears on the middle of the sides”’ (p.20). Idonot find this to be the case in any of the three species examined. Another difference is in regard to the regions last invested. Instead of the “rump and nape usually being the last parts to be covered,” the Juvenal pelage normally persists on the head between and just anterior to the ears for days, often for weeks after the complete investment of the rump region. In these regards, the species in quéstion do not appear to be typical of the genus. The precocious appearance of feathers or hair characteristic of a later plumage or pelage has been mentioned by a number of observers. With reference to the varying hares, Allen (’94b) says: In the case of wounds from fighting or other cause, resulting in the violent removal of large bunches of fur, it is interesting to note that in the autumn the new hair comes out white, often weeks in advance of the general change, and that in spring, under similar circumstances, the hair comes out brown, like the summer coat, much in advance of the general change from winter to summer pelage (p. 121). A similar condition has been described by Schultz (15) in the Himalayan rabbit. In this animal, which is a pink-eyed albino with black feet, muzzle, and ears, the black markings do not appear in the juvenal pelage. By plucking out the hair on one ear, Schultz obtained animals in which one ear was black while the other remained unchanged until the next pelage was assumed. In the domestic fowl the secondary sexual feathers which are characteristic only of the adult plumage of the male may be caused to appear prematurely by plucking out the undifferenti- ated body feathers which precede them. According to Pearl and Boring (714), “If the juvenile feather is removed from the follicle the next feather produced by that follicle will be the 90 H. H. COLLINS secondary sexual feather, and not a feather of the juvenile type. After that all further regenerations are of the sexually differen- tiated feather” (p. 144). I have oceasionally noticed the premature appearance of post- juvenal pelage without operation in young mice which had lost a patch of hair before the time of the regular moult, or in places where apparently the juvenal hair had failed to appear. In the course of his rather extensive studies of the regeneration of hair in rabbits, Schultz describes certain phenomena, similar to those which I have found to occur in mice. For example, he found after shaving large patches on the dorsal and ventral surfaces of an adult black and tan rabbit that restoration was accomplished quickly on the ventral surface, while the depilated region on the dorsum, with the exception of a few ‘hair islands,’ remained bare for a year after the operation. On the other hand, when the pelage was plucked out, restoration was found to occur promptly at all seasons of the year and in animals of different ages. As already pointed out in my account of the regeneration of adult pelages,* the conditions are quite similar in the case of Peromyscus. In both animals the activation of the hair follicles is more readily accomplished when the mechanical stimulus is added to the effects of temperature upon the exposed skin. Schultz regards the appearance of ‘hair islands’ as due to differences in the peripheral blood supply. Furthermore, he sees in this phenomenon an evanescent manifestation of the mottled color pattern, as seen, for example, in dappled gray horses. Another of Schultz’ experiments may be briefly mentioned because of its general bearing on his theory of animal coloration. In the Himalayan rabbit, according to his account, when white fur was plucked out it was replaced by black, although this color is normally limited to the feet, ears, and muzzle. The capacity for pigment formation seemed to be general and markings could be produced at will on parts of the body where they never occur in nature. In the: course of his experiments, it became ® See page 87 above. MOULT AND REGENERATION OF PELAGE IN MICE 9] evident, so he believed, that light played an important réle in the production of pigment in the skin of the depilated surfaces. On the margins of the depilated areas, shaded by the surrounding fur, the regenerated hair was found to be white, while that in the partially shaded region was less intensely pigmented than the fully exposed central area. Furthermore, he describes having obtained hairs of the banded or agouti type by exposing the denuded skin to light at certain intervals only. The experiments were repeated on a number of other rodents with negative results. Nevertheless, Schultz suggests that the differential coloration of the dorsal and ventral surfaces char- acteristic of many mammals may be due largely to differences in illumination. While my own investigations have not as yet been carried far enough to warrant the formulation of an alternative hypothesis, it nevertheless appears obvious that the theories advanced by Dwight and Schultz are inadequate to account for some of the phenomena observed in the moults and color patternsof mammals. Dwight’s theory of the correlation between the distribution of peripheral blood-vessels, and the points of origin and the sequence of moult on different parts of the body of birds does not appear to be applicable in the case of mice. We should scarcely expect to find differences in the arrangement of superficial blood-vessels sufficient to account for the differences in points of origin of the moult observed in species of the same genus. But, more than this, the fact that, in regeneration following removal of pelage, the normal sequence is so markedly modified speaks against this hypothesis. In no mammal are the differences in coloration of the dorsal and ventral surfaces more marked, nor are the two regions more sharply delimited than in some of the species of Peromyscus. The sharpest contrast is seen on the tail. Whatever the réle of light may have been in the evolution of this color pattern, it appears to be a negligible factor in its ontogenetic development. Since these animals are mainly erepuscular or nocturnal in habit, the growth of the hair occurs in, diffuse light. Differences in illumination of the dorsal and 9? H. H. COLLINS ventral surfaces are practically nil. Then, too, the dorsal surface of the tail is rarely wholly covered by the median stripe. Here, under identical conditions of illumination, heavily pigmented and pigmentless hairs develop side by side. Furthermore, I have observed no differences in the color of hairs of the shaded margins and of the exposed central portions of depilated areas. The capacity for the production of hairs of the agouti type appears to be quite definitely confined to the dorsum, the position of the lateral line apparently being unaffected by the operation. Recognizing the fact that in a large number of cases the markings of animals obviously cannot be attributed to differences in intensity of illumination, Schultz has recourse to a second theory, namely, that such color markings are due to the inequalities in the peripheral blood supply. The black markings of the Himalayan rabbit are found on the extremities where the blood supply is somewhat reduced. The dorsal median stripe found in many mammals, in Peromyscus for example (figs. 7, 9, 15), is said to be due to the pressure on the skin of the underlying vertebrae, which impedes the circulation. The rings on a cat’s tail overlie vertebral processes, and so on through the category. It must be pointed out in this connection that Schultz appears to be somewhat inconsistent in his application of this theory. In one paragraph we read: Meine Ergebnisse, dass wachsendes Haar besonders fiir Farbstoff- bildung geeignet ist, und zwar um so mehr, je lebhafter die Wachstums- vorgiinge, sind eine Art Nachahmung der von Darwin bemerkten Natur- erscheinung, dass weisse Taubenrassen unbefiedert, dunkle befiedert dem Ei entschliipfen. Die Kaninchenalbinos, die ich hielt, schienen mir bei der Geburt auch so gut wie kahl, die farbig geborenen Rassen aber starker behaart. In vielen Naturmustern finden wir die starksten Farbstoffanhaufun- gen gerade an Stellen, die durch starkstes Haarwachstum gekennzeich- net sind, und an solchen Vorspriingen und Ausbeutelungen der Haut, die zeitweilig starker wachsen miissen als ihre Umgebung, z. B. Mahne, Schweif und Beine der Grauschimmel (p. 161). However, in regard to the black markings of the Himalayan rabbit, he writes: MOULT AND REGENERATION OF PELAGE IN MICE 93 Betrachtet man die Russenkaninchen als Ganzes, so erhélt man den Eindruck, dass an ihnen die mehr innen gelegenen Teile farblos bleiben, daher nicht nur die andern inneren Gewebe, sondern sogar die roten Augen, welche nach innen versenkt sind. Die dem Herzen ferneren, den Schidigungen der Aussenwelt und der schlechteren Durchblutung mehr ausgesetzten Teile, Nase, Schwanz, Ohren, Fiisse, sind der Farb- bung verfallen, uberhaupt neigt daher insbesondere die Haut zur Farbstoffbildung. Die inneren Organe scheinen gerade wegen ihrer besseren Durchblutung, geringeren Schadigung wegen Farblosigkeit zu besitzen (pp. 551 and 552). That is to say, pigment formation is most pronounced where the processes of growth are most active, and at the same time, in regions having a relatively poor blood supply. From which it appears to follow that an adequate supply of blood tends to inhibit growth. It appears that, at best, Schultz’ theory of the relation of blood supply to pigmentation is applicable only in a limited number of eases. The list of exceptions is overwhelmingly large. To cite a specific case, the hair on the feet of Peromyscus is pigmentless. This is characteristic of the genus, whence the name, ‘white-footed mice.’ Furthermore, in many species of small mammals, individuals having white-tipped tails are of frequent occurrence. In the case of Zapus insignis, as cited by Miller (’93), this white-tipped condition has become characteristic of the species. In certain species of Peromyscus, Sumner (718) describes the occasional appearance of the same character, and of pigmentless snouts as well. In the alternative inheritance of many color patterns, we are confronted by another category of facts which are not readily interpreted in the light of Schultz’ hypothesis. In conclusion, we may refer to the interesting and suggestive researches of G. M. Allen. This investigator has found that in general in mammals and birds, pigmentation centers in eleven separate areas, five paired, and one unpaired. Pigmentless markings are said to arise when contiguous areas fail to meet. Each area may vary independently. Further investigations along these lines may go far toward clearing up some of the puzzling problems which one encounters in a study of animal coloration. 94 H. H. COLLINS 7. SUMMARY 1. The process of normal moult has been followed in a large series of living mice representing several species of Peromyscus. 2. In this study of the living material, the process of moult is found to be, ina measure, comparable inregularity of sequence and directions of growth with the moults of birds. 3. In the postjuvenal moult, growth occurs more or less in- dependently on certain regions of the body, suggesting the mode of moult in the pterylae of birds. 4. The moults of adults are generally of a more irregular character. 5. In young mice the change of pelage is quite obvious, but in adults it may be quite insidious and evident only upon close examination. 6. In general, the process of moult is quite similar in different species, but in some instances there appear to be certain minor differences. 7. By plucking out juvenal hair, the precocious appearance of the postjuvenal pelage may be induced. 8. Under certain conditions, the appearance of this postjuvenal pelage, after artificial removal of the Juvenal, is preceded by the outgrowth of an aberrant type of hair which persists only for a short time. Within these hairs the localization of pigment is abnormal. 9. The normal sequence of the incoming hair is profoundly modified by artificially induced regeneration. 10. Restoration of pelage in adults oecurs irrespective of sea- son, after the plucking out or clipping of the old hair. 11. This restoration is accomplished by the outgrowth of new hairs, except in ease of the vibrissae, which are replaced by the elongation of the cut hairs. 12. Restoration is much more rapid when the hairs are plucked out than when merely cut. 13. The differences in coloration of the old and the new pel- ages as seen on opposite sides of the body of adult gambeli are MOULT AND REGENERATION OF PELAGE IN MICE 95 slight, never approaching the differences between individuals representing light and dark extremes within the species. 14. Light appears to be a negligible factor in the development of the differential coloration of the dorsal and ventral surfaces. BIBLIOGRAPHY ALLEN, G. M. 1914 Pattern development in mammals and birds. Amer. Nat., vol. 48, no. 571, July, pp. 385-412 no. 572, Aug., pp. 467-484; no. 573, Sept., pp. 550-566. Auten, J. A. 1894a First plumages. Auk, vol. 11, no. 2, Apr. pp. 91-93. 1894b On the seasonal changes of color in the varying hare (Lepus americanus Erxl.). Bull. Am. Mus. Nat. Hist., vol. 6, pp. 107-128. BarrettT-HamILTon, GERALD E. H. 1912 A history of Britishmammals. Part XII. London: Gurney and Jackson, pp. 304-306. 1916 25 ncn shod Epa votercpay tees ore eaters So epoarnabels BpLe 108 SPBIELEGOUY CORUNA CT ooh ae oes octane CEP ees oc thi fuecad gy ooeeeilteae doom ole eB 115 WD ISCUISSIOM A ee Tre te ere nities sale od Are eM ones sorte eons Misi andere ons eg 124 Mee rupitoOrimy lana Ae pe ls VI SR hd OS ECL On) Be 124 PO PLESVSIRTSS SU hy 9 ages Te SRE fe Ce ra ES eae Scot © OCA Os reemA RA SF SO ie OSE 126 SY. TEER WCTE URS LS SRG Sa EY tn Re eR rei ict SAT Cpe PRS OS» So NR fe 131 A MEEREI UCIT ren Ey 2 IO, Chea a) ce a lcigis aly ateaeeet ta cin SONS haisteyelst eda clare, rege 133 PRIMO R ANH ee oti ee Siaes Aaa bs eels MRI Se ayhiat wei Me arc Makes 4 ae 134 GMM COMNACIE rere Beet cey Michie se AU Sees ele aha. Sek, AMD MEMOS E Yeeeah a 135 7 STS AE ORR NE Be Ae en Rte Sag naan «OM OREN EY ek SER ane 135 SPENCE maynred pe Mes es soe Mec ys eee ere eae ade o Se eR harte tm ae pa hone i emer OMAVASCU ATER USEC S cee regia cke cis thes ore Nee arcane cra etnfsohieonatersencte, fe ole 137 Sealant ite) CONCLUSION: , 512). hoa Oh ON. 8 iy apdohianle Unt taheeegee » SEEMS 138 tend ITE IE EG iio hier Sek Selick 8 Ay Ghana & days hegaraters Later nia enter SNe Sees anise Aare 141 INTRODUCTION The present study grew out of an attempt to determine the function of the digitiform gland, the development of which has recently been described (Hoskins, 717). This gland is peculiar to selachians, being found in no other group of animals, al- ‘though in Chimaera, there is present in the wall of the posterior portion of the intestine a group of cells which may possibly correspond to it (Disselhorst, ’04). 101 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 2 NOVEMBER, 1918 102 E. R. HOSKINS AND M. M. HOSKINS Nearly every reasonable function has been ascribed to the digitiform gland, but usually from morphological or hypothetical rather than physiological evidence (Hoskins, 717). The gland is a compound tubular structure with a large central lumen or duct which empties into the intestine about half way between the spiral valve and the cloaca. It is suspended in the mesen- tery above the intestine. Cytologically, the gland is not unlike the kidney tubules, and this, together with the fact that its secretion is discharged into the intestine posterior to the region where digestion occurs, leads to the expectation that it has an excretory function. The fact that the gland appears to be the same in both sexes argues against the probability that it is accessory to the sexual apparatus. Another part of the work is a study of the function of the kidney (mesonephros) compared with that of mammals (meta- nephros). It is common knowledge that the former is a less efficient excretory organ than the latter, but the subject has never been completely studied experimentally. A study is made also of the excretory function of the liver, which is of considerable importance in selachians, on account of the ineffi- ciency of the kidney. In selachians the liver undergoes fatty metamorphosis to such extent that nearly every cell appears filled with fat, yet the organ is able to excrete solutions and particles freely and in large amounts. Denis (713) has shown that dogfish are able to withstand large doses of excretory toxins, but this author did not study the problem histologiéally. The question of vital staining is con- sidered here only incidentally and no attempt is made to describe completely the reaction of the dogfish to vital stains. These substances were used primarily as an aid to the study of the excretory function. Most of the experiments were performed at the Marine Bio- logical Laboratory at Woods’ Hole, Massachusetts, in quarters provided us through the kindness of the Department of Zoology of Yale University and of Dr. Frank Lillie. The accompanying photomicrographs were made at the Osborn Zoological Labora- tory of Yale University. Two experiments were done at the REACTION OF SELACHII TO INJECTIONS 103 Aquarium in New York. We desire here to express our thanks to the Zoology Department of Yale University, to Dr. Lillie, and to Dr. C. H. Townsend, of the Aquarium. MATERIAL AND METHODS The animal used in most of the experiments is the dogfish Mustelus canis, which is fairly abundant at Wood’s Hole. In the two experiments performed at the New York Aquarium we used Acanthias. In all about seventy-five animals were made use of. These were of both sexes and of various sizes, but most of them were adult males. The study was divided into three phases; a) the reaction to non-toxie solutions; 6) the reaction to suspended particles, and, c) the reaction to excretory toxins. The solutions used are indigo-carmine, potassium: iodide, dextrose, sucrose, pheno- sulphonepthalein, and Weed’s potassium ferrocyanide-iron am- monium citrate solution. The supensions are carmine, Congo red, neutral red, and trypan blue. The toxins are potassium chromate, tartaric acid, and uranium nitrate. All were pre- pared with sea-water, which is isotonic with the blood of dogfish. In preparing Weed’s solution, sea-water, diluted one-third, was used. Injections were made either into the muscles of the tail and back or intravascularly. In the latter method the large vessels of the pectoral fins were used. A small wooden peg may be used for a hemostat and is very convenient as it can be easily removed for repeated injections. Intravascular injections may be made, also, directly into the sinuses of the head. A simple device for holding the animals during the experi- ment may be made from a board 4 inches wide and 2 feet long. A long hole is made near one end for the dorsal fins and a row of nails driven partly in along each side of the board. The ani- mal will be securely held on this if a cord is laced from the nails back and forth across its body. A tube of running water in- serted into the animal’s mouth will provide for respiration dur- ing an operation. ‘The entire apparatus with the fish in normal position may be immersed in the aquarium if it is necessary that 104 E. R. HOSKINS AND M. M. HOSKINS the animal be kept immobilized. In collecting urine from an immersed animal the receptacle can be sealed tight and weighted to keep it under the water. A long glass tube leading into it will permit the air to excape as the urine accumulates. If along rubber tube is used to connect the cannula with the collecting bottle the animal can even be permitted to swim about in the aquarium while the urine is being collected. Denis (12) also describes a good method for collecting urine. The duct of the digitiform gland enters the intestine at a peculiar angle, being completely bent on itself. In order to insert a cannula into the duct it is best to reach into the intes- tine with a pair of blunt forceps and turn the gut inside out through the cloaca. This straightens the duct and a cannula ean then be introduced directly into it. No anaesthetic is required in operations as the animal will lie quietly in the ap- paratus described above and seems insensible to pain. Anti- septics were also found to be unnecessary in operations. EXPERIMENTS 1. Solutions A. Indigo-carmine. 1. An adult male dogfish was injected intramuscularly with 5 ce. of 0.5 per cent indigo-carmine. This appeared in the urine in fifteen minutes. It did not appear in the stomach, spiral-valve or digitiform gland contents, at least in detectable quantities. The bile was bluer than normal. 2. The experiment was repeated with a female, and similar results obtained. 3. Two animals, a male and a female, each received 5 ce. of the above solution intravenously. It was recovered from the urine of each after eight minutes. 4. A female was injected as in ‘3’ and permitted to live five hours, at which time the urine was still dark blue in color. All five of the above-mentioned animals were autopsied at various times after injections, varying from a half-hour to five hours. All secretions examined gave the same reactions as in REACTION OF SELACHII TO INJECTIONS 105 ‘1.’ Also in every case the anterior portion of the kidney and liver was of a light blue color. The posterior portion of the kid- ney was dark blue and a dissection of the liver showed the bile | ducts to be colored blue rather than the usual green. The other organs of the body cavity were not distinctly blue in color. Glands were fixed in absolute aleohol, but the dye soon dis- appeared either by fading or going into solution. Sections were made and examined microscopically, but no blue granules were seen even though the gross specimens were blue. The dye was probably impure and hence dissolved out in the alcohol, as was found to be true in some eases by Heidenhain (’83) working with other animals. B. Potassium iodide. An adult male fish received 10 ec. of a3 per cent solution of potassium iodide intravenously. After three and a half hours tests were made with nitric acid and starch (Hawk, 716), and the iodide was found to be present in the urine in the gastric and intestinal fluids, the bile, and the secretion of the digitiform gland. C. Dextrose. An adult female fish received intravenously 15 cc. of 25 per cent dextrose solution. At the end of one hour the urine was negative to Benedict’s sugar test (Hawk,’16), but was positive after four hours and negative again after nine hours. At the latter time the bile gave a positive reaction, but the secretion of the digitiform gland reacted negatively. D. Sucrose. The animal received intravenously 15 cc. of 30 per cent sucrose, and after six and one-half hours was autop- sied. The reaction to Benedict’s test was positive in the case of the urine, bile, and blood serum, but negative with the gastric and spiral-valve contents. The digitiform gland secretion was not tested. E. Phenosulphonepthalein (Rowntree and Geraghty, ’12). 1. One and two-tenths ce. of this pthalein was injected intra- muscularly. It appeared in the urine in fourteen minutes. At autopsy it was not present in detectable amounts in the stom- ach, intestines, or digitiform gland. 2. An animal received intravenously 1 ce. of pthalem. Au- topsy was performed after one hour. At this time the pthalein 106 E. R. HOSKINS AND M. M. HOSKINS was present in the urine, bile, and in the stomach and spiral- valve contents. The secretion of the digitiform gland and an extract of the spleen were negative. 3. An adult fish received 1.4 ec. of pthalein. The urine was collected for five and a half hours and the bile in the gall-bladder was removed at the end of that time. The urine, bile, and 1.4 ec. of pthalein in graduated cylinders were each diluted with water, acidified and all compared (Rowntree and Geraghty, 712). By this comparison it was shown that approximately one-half of the injected pthalein had been recovered, and of this, about one-fourth was in the urine and three-fourths in the bile. The bile pigments interfere slightly with the test. A trace of ptha- lein was found in the serum, digitiform gland secretion, and in the stomach and spiral-valve contents. That in the latter two may have come from excreted bile. 4. An adult received 1 cc. of pthalein. A cannula was in- serted to collect the urine and the animal was then permitted to swim about in the aquarium for four and a half hours in order that the excretion might not in any way be interfered with by the unnatural position of the animal in the holding apparatus that was used in the first three experiments. In this specimen as in the others more pthalein was recovered in the bile than in the urine. Pthalein was also found in the spiral valve, but not in the stomach. F. Weea’s solution (potassium ferrocyanide and iron ammo- nium citrate, Weed, 712). 1. A young male (0.8 kg.) received intra muscularly 7.5 cc. of Weed’s solution and was autopsied after four hours. The urine and bile gave the Prussian-blue reaction with acid. The serum and the stomach and spiral- valve contents were negative to the test. 2. A young animal (0.5 kg.) received intramuscularly 9 ce. of Weed’s solution, followed after five hours by 9 ec., and after twenty-two hours by 8 cc. Twenty-five hours after the first injection the Prussian-blue reaction was obtained with the urine, bile, and serum. The liver turned blue on standing. 3. A young female (0.5 kg.) received, as above, 10 ce. of the solution, followed by 5 ec. after four hours, and 5 ec. after seven REACTION OF SELACHII TO INJECTIONS 107 hours. Autopsy was performed eight hours after the first injec- tion. The Prussian-blue reaction was obtained from the serum and urine, but not from the bile, or stomach and spiral-valve contents. Pieces of organs were fixed in 20 per cent formal- dehyde acidified to 1 per cent with hydrochloric acid. 4, An adult (Acanthias) received four hourly injections of 6 ec. each of Weed’s solution, and was autopsied four and a half hours after the first dose. Organs were fixed in absolute alcohol acidified as above. 5. An adult (Acanthias) received three injections of 9 ce. each of Weed’s solution at hour intervals and was autopsied forty-five minutes after the last injection. , Forty per cent for- maldehyde acidified to 1 per cent was used as a fixative. Mi- croscopic examinations were made of various organs from the animals, ‘3,’ ‘4,’ and ‘5’ with the following results: a. Digitiform gland. In all the specimens examined there is a central, distinctly blue zone, the width of about one-half the radius of the gland as seen in transverse sections. Microscopi- cally, the blue granules are seen almost entirely within the blood-vessels, although a very few are found in the cells of the tubules, but none in the lumen. The granules are present in the capillaries and central venous sinuses in greater concentration than in the vessels of any other organ. b. Kidney. The gross specimens are dark blue in color. Sections show many Prussian-blue granules in the blood-vessels and tubules. A very few only are present in the glomerular cavities. They are found in the cytoplasm of both the secretory and excretory tubules. Blue granules are found also in the con- nective tissue of the kidney. c. Liver. The gross specimens are all light blue in color. In a few hepatic cells blue granules are seen and they are present in greater number in the ducts. This indicates either that the solution passed directly into the ducts, as was thought to be the case of trypan blue in bony fishes (Wislocki, 717), or else that the solution passed through the epithelium quickly and out of the ducts more slowly, thus becoming concentrated in the ducts. In the liver of animal ‘3’ the endothelium and the leucocytes of 108 E. R. HOSKINS AND M. M. HOSKINS the hepatic sinuses have a deposit of blue granules on or in them. This condition is not seen in other organs of the same animat nor in the liver of other animals. d. Spleen. Blue granules are scattered throughout the spleen among, but not within the cells. e. Spiral-valve, stomach, gills, and muscle. All these have blue granules in the blood-vessels and scattered through the connective tissue, but none within any cells. .In gross free- hand sections the epithelium of the stomach and spiral-valve appears white, in striking contrast with the dark blue connec- tive-tissue. 2. Suspensions A. Carmine. Powdered carmine was used for coarse granules, although some of it seems to go into solution in sea-water. It will not all settle out on standing, cannot be separated out with ordinary filter-paper, and will even pass through a celloidin membrane. After passing through ecelloidin no granules can be seen with the ordinary microscope. This may not be in true solution, but if not it is practically so- 1. Filtered carmine after settling for several days was inipeled intravenously into an adult male dogfish. The animal received four daily doses of 12 ce. each and was, autopsied eight hours after the last injection. Microscopie examination of the various organs failed to show the presence of any carmine granules, although the spleen appeared distinctly red after fixation. The granules were either too small and too much scattered to be seen or else the carmine was in solution and washed out in pre- paring the sections. The urine was not collected. 2. Ten adult animals were injected intravenously each with 10 ec. of a heavy suspension of powdered carmine in sea-water and autopsied after various periods as follows: 13, 63, 21, 24, 50, and 96 hours. ‘This carmine was not filtered. Pieces of the various organs and tissues were fixed in formalin, sectioned in paraffin, and stained with haematoxylin, or examined unstained. Urine was collected from several of the animals, but in only one specimen of it was any carmine found, and in that but one REACTION OF SELACHII TO INJECTIONS 109 granule. This sample of urine also contained a few blood cor- puscles and hence the carmine probably entered with the blood. Microscopical examination. a. Digitiform gland. Carmine is present in the endothelium and leucocytes of the blood-vessels in this gland in all stages, but very little in the first (1 hours), and less in all the sections than in the endothelium of the kid- ney, liver, spleen, and gills. In the 21-, 24-, and 50-hour speci- mens there is an occasional carmine granule seen in the cyto- plasm of the tubules. These few granules may possibly have been dragged into the cells in the cutting of the sections and in any case are too few in number to be of any importance. None was seen in the lumen of the tubules. b. Kidney. In the first two stages there is considerable amount of free carmine in the capillaries, much of it in large ‘clumps filling the entire vessel, and a few leucocytes and endo- thelial cells contain carmine granules. In later stages (21 to 96 hours) there is a gradual decrease in the amount of free carmine in the capillaries and there are no large masses of it present. There is a decreasing amount seen in the leucocytes and endothelium. A few sections have ecar- mine granules in the connective-tissue cells. In only three tubules of the hundreds of sections examined are carmine gran- ules to be seen in the cytoplasm. These few granules may have been carried in, in the preparation, and as in the digitiform gland are considered to be of no practical significance. Each of two tubules of one section contained a granule in the lumen, but one also finds blood corpuscles occasionally in the tubules and these carmine granules may have entered accidentally. c. Liver. Carmine granules are seen in all stages after the first, in the leucocytes and endothelial cells of the sinusoids, in the hepatic cells, and in the normal pigment accumulations, which are very abundant in the liver of the dogfish. Carmine was found in the bile in all stages after the first, as free granules, in cells resembling leucocytes and mixed with the normal pigment. d. Spleen. Carmine is present in all the sections of spleen examined and in decreasing amounts in later stages (fig. 1). 110 E. R. HOSKINS AND M. M. HOSKINS In all stages carmine may be seen as granules free in the sinuses, and ingested by the endothelium of the sinuses and by splenic cells. The insert in figure 1 shows an endothelial cell that has extended a pseudopodial process and engulfed a very large granule. The figure also shows an endothelial cell filled with carmine and apparently about to be liberated as a free phago- cyte (macrophage?). In many places the ingested carmine seems to be enclosed in vacuoles. There is a decided increase in the number of mitotic figures seen in the spleen beginning with the 21-hour stage, indicating a leucocytosis. e. Spiral-valve and stomach. In these organs there is an occasional phagocyte in the capillaries that contains carmine. There is no carmine outside the vessels. f. Gills. The large sinuses which are interposed between the branchial arterial arches and the capillaries in the filaments are lined with an endothelium which is very phagocytic. This will be described in detail under ‘Trypan blue.’ In the earlier stages the gills were not examined, but sections of later stages (50 and 96 hours) show the endothelial cells of the sinuses to be en- gorged with carmine granules. Many of these cells have also been ‘budded off’ and lie in the channel as free phagocytes. This endothelium is doubtless one of the: sources of circulating phagocytes in the dogsfish. Even the arterial arches are lined with phagocytic endothelium especially along the side next to the sinuses. g. Heart. The 14-hour stage shows a few cells of the endo- eardium that are phagocytic to carmine. The heart in other stages was not examined. ? h. Blood. In all stages the blood contains free carmine and granules ingested by phagocytes. The number of free granules decreases in the later stages. B. Neutral red. Two series of experiments were conducted with neutral red and to some extent with different results. In 1916 the sample of neutral red used seemed to form a true suspen- sion, as it could all be separated from the water with ordinary filter-paper. , REACTION OF SELACHII TO INJECTIONS qi tT 1. An adult received intravenously. 10 cc. of suspension of neutral red and was permitted to swim about in the aquarium. At that time the animal was immobilized, received an addi- tional 18 ec. of the suspension, and its urine was collected for five hours without any of the neutral red appearing in it. Au- topsy showed the bile to be reddish brown. ‘This turned orange with alkali and dark lilac with acid. Serum gave a similar reaction, but the gastric fluid was negative. The suspension was prepared by stirring neutral red in sea-water and allowing the mixture to settle. 2. An adult was injected intravenously with 15 ce. of satu- rated neutral red suspension. The urine was collected for six hours and was at all times free of the dye. Autopsy showed neutral red in the bile in a considerable amount. 3. A female received intravenously, 10 ce. of the neutral red suspension and the urine was collected for eight and a half hours without any of the dye appearing in it. As above, it was present in the bile. The two following experiments were carried out a year later than the former, with a different sample of neutral red. This could not all be removed by filtering and part of it even passed through a celloidin sac, so that it probably was partly in solution. 4. Ten ce. of filtered neutral red in sea-water was injected intravenously into an adult fish. Four hours later, examination showed the dye to be in both the bile and urine, but not in the stomach or spiral-valve contents. . 5. Hight ee. of the dye in the above form was injected intra- venously into an adult dogfish. There was 24 ec. of bright red urine collected in twenty-four hours, and at autopsy 2.5 cc of reddish-brown bile was removed from the gall-bladder. The 24 ec. of urine was diluted to 100 ec. with water and the 2.5 ce. of bile to 200 ce. Maximum redness was produced by neutral- izing, and the color of the two were then of about the same intensity. Thus the total excretion of the dye by the liver was more than eighty times as much as that excreted by the kidney. The total amount of dye removed by the liver was not collected 112 E. R. HOSKINS AND M. M. HOSKINS because some escaped with bile into the intestine. The neutral red in the urine appeared to be in solution when examined under the microscope, whereas some, at least, of that in the bile was in granular form. Before the filtered neutral red was injected it appeared to have only very minute granules, smaller than those in the bile, so that in passing through the liver, granules must have collected together. Denis (12) found the urine acid to litmus, but neutral red excreted in it retained its red color. C. Congo red. The Congo red used could easily be removed by filtration or settling and would not pass through a celloidin membrane. | 1. Two adult dogfish received intravenously each 5 ce. of a heavy suspension of Congo red in sea-water. After twelve hours the urine was still free from the dye, as was the stomach and spiral-valve contents. A considerable amount was present in the bile. 2. A young fish (0.25 kg.) received doses of Congo red of 12 ce. and 18 cc., respectively, on two successive days and was autopsied forty-eight hours after the first injection. The liver and spleen were stained deeply with the dye, but the other organs of the body cavity appeared practically normal in color. The gills were not examined. A large amount of the dye was present in the bile, a small amount in the spiral valve, and a trace in the stomach. The dye in the spiral valve entered with the bile probably entirely and that in the stomach probably was regurgitated from the spiral-valve. | D. Trypan blue. Trypan blue was used primarily in studying the excretory function. Its action as a vital stain, although not exhaustively studied, will be considered also from the data we have, because the reaction in selachians seems to be in some ways different from that in teleosts as described recently by Wislocki (717). A brief review of our results with this dye has been published (Hoskins and Hoskins, 718). 1. An adult male was injected intravenously with 9 cc. of a 2 per cent suspension of filtered trypan blue in sea-water. Four hours later the animal received 6 ce. more. Fifteen hours later another injection of 6 cc. was made. Autopsy was performed REACTION OF SELACHII TO INJECTIONS 113 thirty hours after the first Injection. The urine was examined several times and was free of dye until nearly the time of au- topsy, when a trace of the red portion (Wislocki, ’17) was pres- ent. The bile contained a considerable amount of the whole dye. It was also recovered from the serum, but not from the stomach or spiral-valve contents. The liver and spleen were of a slightly darker blue color than the other organs, but all of the organs as well as the skin and peritoneum appeared slightly colored by the dye. Microscopical examinations were made of the digitiform gland, kidney, liver, spleen, spiral valve, stomach, skin, and mus- culature. All of these retained their blue color after fixation as seen in gross specimens, but blue granules are found under the microscope, definitely, in only the endothelium of the liver and | spleen. The stain in the other organs and tissues is either too diffuse to be seen microscopically or else was all free in the blood- vessels, and hence washed out in preparing the sections. 2. A young fish (0.25 kg.) was given four daily intravenous injections of 3 ce. each of the above-mentioned trypan blue, and was autopsied five days after the first injection. The dye was found in considerable amount in the bile and serum and was also present in the spiral-valve, but not in the stomach contents. The liver and spleen were very dark blue in color as were the gills at the base of the filaments. The skin was dark blue before fixation, but light blue afterwards. The digitiform gland, spiral valve, post valvular intestine, peritoneum, and fascia of the muscles were light blue in color, the stomach was still lighter, and the kidney, except the peritoneal covering and connective tissue, was practically normal in color. All these organs except the first three were of a lighter shade after fixation than before, owing doubtless to loss of blood which contained trypan blue. Microscopical examination. a. Digitiform gland. Most of the dye is too diffuse to be located microscopically. A few leucocytes and endothelial cells of the capillaries of the paren- chyma and serosa contain blue granules and one cell of a tubule contains eight such granules in a vacuole, but the amount of 114 E. R. HOSKINS AND M. M. HOSKINS the dye detectable in this organ is too small to be of practical importance. b. Kidney. Microscopically, the dye can be seen in only a few leucocytes and endothelial cells of the capillaries. c. Liver. The liver contains most of the fixed trypan blue in the entire body. This organ is relatively large and in almost every cell in a large number of sections examined there are from five to twenty blue granules of different sizes (fig. 2). In places the granules appear to be in vacuoles. A rather large number of endothelial cells of the sinusoids likewise are seen to contain the dye. In figure 2 one such is seen nearly separated from the wall of the vessel and so full of stain that the nucleus is barely visible. In the liver of the dogfish there is a relatively large amount of normal pigment both in the sinusoids and in the parenchyma, collected into large masses. These black granules are often found in phagocytes. The large mass shown in a sinusoid in figure 2 is made up both of this black pigment and trypan blue eranules. These black pigment masses may be found in the bile normally, and in the present experiment both black masses and masses of both blue and black granules were seen. There were found also in the bile many blue granules free and in es- caped cells which are present also in normal bile. All the granules shown in figure 2 in the hepatic and endothelial cells and in the leucocyte are trypan blue. The mass at x is mostly normal black pigment. d. Spleen. See oak eS 47.01 Hein senaceeeeees wet 240 K90 51.74 ak eon eee eee lt 252800 49.45 Ecos eee el ea Necaes ah oem als Ley 50.02 pe nets eae ees A Oa 48.13 AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, OCTOBER 8 REACTIONS OF LAND ISOPODS TO LIGHT CHARLES HARLAN ABBOTT Biological Laboratory, Brown University CONTENTS if.) IGTRAROYS (RA OAR ane ae COS ApEn otteais COE BOR OE Tiber EEE OR OO ACES orcm aaiaine 194 PERS LOTTO eer eye tes cro care ihe iiara, Siopatecdeaayd grapen'sserelt eb-ahicese grein deere eae 195 ie MVintorialensae eee ING s see acces oes sls acters avates dese Meee sabe cea de nee 196 ViaeNormals behavdorser -pecioo. wacelo sito nants Set tera ta. Chelacebeici ame ea ae 197 FAMSTLCACULONS eUOMMLO IN ee se Baas perio JB gh Norse crate gatajen ohare ih Shegorauareraeter antl 197 EMMIS OBO GNC SHS ra Ria crer ea As bier, oUt Sispsinin oni Na shave ara repeat evel eteveags 197 Dem E HOCO lamas Neri. eee re MS Se oes « views Sere ere eile ae oe 197 SAL VILSTONEE eer sneer es ee ined aude Shay TD ate a tetas 198 Bt Relationiof contaetyand wislone ew c2e «aah om n:ds steye Saepyeain se << 198 Ve Pxpermments with Girective: light). 9 czis 800. observer may have to infer an interpretation along the lines of preconceived ideas of what the process should be. I will try to avoid the latter pitfall by describing some actual spindles as they appear under the highest lens system at my command. The most conspicuous and the most frequent stage is illustrated in figure 55, a. In this stage the chromosomes are densely stained and appear as rods occupying the two central quarters 318 GARY N. CALKINS ‘ of the spindle. The number of rods is four but each rod is double. How these double chromosomes are formed from the eight single ones I am unable to decide. Another, and a less fre- quent stage, is illustrated in figure 55, c. Here there are eight $3 RM, oe ws (2) Gee ies @ 9 ox Fig. 48 First maturation spindles, second type, with definite nuclear plates. x 800. Fig. 49 Pair in prophase stage of first maturation division, with degenerating parachute nuclei. 800. Figs. 50 to 53 Division of granular mass of chromatin in degeneration of first maturation nuclei into two parts, and vesicular swelling of the nuclear membrane. x 3200. distinct chromosomes arranged in groups of four. If the former stage is a metaphase, is the latter an anaphase or a prophase stage subsequent to that shown in figure’54, d? In no case found do the chromosomes lie with their long axes at right angles to the THE NUCLEI OF UROLEPTUS MOBILIS 319 long axis of the spindle, and in no case is there any evidence of _V’s or Y’s. In some cases the rods appear to run unbroken from pole to pole of the spindle (figs. 55, 6). In other cases there may be more than four distinct chromosomes, sometimes five or six or seven, in the nuclear plate. In still other cases there are four definite chromosomes at the poles of the anaphase stage (fig. 55, d). Until I can be convinced of the exact method of chromosome division by further study of this phase, I will not offer an inter- pretation of this puzzling problem. Let it suffice here to state Figs. 54 and 55 Prophase, metaphase, anaphase, and telophase stages of second maturation mitosis where four double chromosomes are reduced to four single ones. Bouin fixation, lron-haematoxylin stain. 3200. that the number of chromosomes is reduced from eight to four by this second division. The four chromosomes of the anaphase stage (fig. 55, d) lose their identity in the chromatin mass of the late telophase (fig. 55, e). The final division of the nuclei is rather abrupt, and there is no evidence of long connecting strands between the daughter products. c. The third diwision. The products of the second division are small granular micronuclei, usually four in number, which are ready for the third division with no extensive intervening resting stage. In some individuals all four of them undergo this third division (fig. 57, left); in some only three, while in others only 320 GARY N. CALKINS two divide (figs. 56 and 60). In all individuals the functional pronuclei are only two in number and both come from the same one of the four original nuclei. The others disappear by ab- sorption, although they may persist until some time after the union of pronuclei. Fig. 56 Pair, each with two nuclei in the third maturation division. X 800. Fig. 57 Telophase of third division; four pairs of nuclei in one, three pairs in the other individual. Macronuclei fragmented. X 800. Figs. 58 and 59 Nuclei in metaphase and anaphase of third maturation divi- sion. X 3200. The early spindles of the third division are smaller (2 » wide and 6 » long), but are similar in type to those of the second mat- uration division (figs. 58 and 59). There are, again, four bars of chromatin, or four chromosomes, and four are present in the daughter nuclei after division (fig. 59), which is apparently transverse. THE NUCLEI OF UROLEPTUS MOBILIS a2 The third division figures are easily distinguished from those of the first and second divisions by the greatly elongated con- necting body between the daughter nuclei. In some cases these connecting fibers attain a length of 30 yu, and different stages of division are frequently found in the same individual or in the pair. The nuclear membrane plays the chief rdle in this con- Fig. 60 Late telophase of third division and formation of pronuclei (on right). x 800. : Fig. 61 Unusual phase in which eight daughter nuclei of the third division and two undivided nuclei are present without pronuclei formation. X 800. necting fiber, the chromatin apparently forming no part of it (figs. 62 and 63). As the nuclei separate the walls come together until they appear like a single connecting line (fig. 60). The chromatin now separates into granules which become distributed throughout the nuclear vesicle; the membrane closes and the pronucleus is formed. Bye GARY N. CALKINS This history is not followed by the nuclei which degenerate. With them no vesicle is formed, but the daughter nuclei, if they do not degenerate first, form homogeneous massive nuclei which gradually lose their staining capacity and ultimately disappear by absorption. Figs. 62 and 63 Intermediate telophase stages of the third division. > 3200. Fig. 64 Stage of interchange. The two migrating pronuclei with their attrac- tion spheres,’ are passing one another in the anterior fused region. One undi- vided third division nucleus is present in each cell. X 800. d. The pronuclei and the interchange. ‘There is no difference in type, although there may be a slight difference in size without significance, perceptible in the two pronuclei formed by. this third division. The stationary pronucleus is nearly always in the central region of the cell where it is left after the third divi- sion. The migrating pronucleus, formed at the other pole of the dividing nucleus, is left some 30 u nearer the anterior end. While ® THE NUCLEI OF UROLEPTUS MOBILIS Bye there is no internal structural characteristic to distinguish it from the stationary pronucleus, a peculiar external or cytoplasmic structure now appears which is decidedly characteristic and which is absent from the stationary nucleus. This structure consists of a clearly defined homogeneous mass of very fine & Figs. 65, 66,and67 Three pairs with migrating pronuclei in different positions. The attraction sphere precedes the pronucleus. 800. granules which replace the ordinary alveolar make-up of the cytoplasm. It resembles an attraction sphere and behaves as such in the later history (figs. 64 to 66 and 71 to 75). The wandering pronucleus in each cell, preceded by this homo- geneous granular mass, moves towards the anterior end where cell fusion has taken place (figs. 64, 65, 66, 71 to 75). The two pass THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 3 ® 324. GARY N. CALKINS one another in the posterior part of the connecting bridge of pro- toplasm (fig. 64) and then each continues its migration in the foreign cytoplasm (fig. 66) down to the stationary micronucleus Figs. 68 antl 69 Two pairs with pronuclei close together. The ‘attraction spheres’ have disappeared. X 800. Fig. 70 Two pronuclei about to fuse. The difference in size is without sig- nificance. > 3200. (figs. 68 and 69). Here the cytoplasmic guide can no longer be made out. The two spherical pronuclei lie close to one an- other, then elongate to form two ellipsoidal nuclei (figs. 70, 76, 77 and 78), and finally fuse. THE NUCLEI OF UROLEPTUS MOBILIS Be e. The amphinucleus. The two pronuclei unite in a region of the body slightly anterior to the geometrical center. There is Figs. 71 and 72 Enlargements of the wandering pronuclei and ‘attraction spheres.’ > 1500. some evidence that the stationary nucleus advances to meet the migrating nucleus, since the former is formed either in the center, or posterior to the center, of the cell (figs. 64, 68, 69 and 76). 326 GARY N. CALKINS Each is spherical and vesicular (figs. 70 and 77) at first, but both elongate and become spindle-formed before fusion occurs. This double spindle becomes the first cleavage spindle in which eight large, homogeneous chromosomes are present. In figure 81 the chromosomes of one side have divided before those of the other side. Its division results in two nuclei of similar size and char- acter; one becomes the first micronucleus, the other (after a Fig. 73 Relative positions of wandering and stationary pronuclei and of ‘at- traction spheres’ about to pass one another. X 1500. later division), the first macronucleus of the ex-conjugant (figs. ~82to85). They are both small, and their chromatin is condensed into the homogeneous massive type of micronuclei. One is an- terior, the other central in position. The former divides to form two micronuclei which condense to form the vegetative micro- nuclei, the other forms a spindle (fig. 82), which gives rise to two products of dissimilar fate, one becomes swollen and vesicular THE NUCLEI OF UROLEPTUS MOBILIS VATE with its chromatin collected in granules arranged about the periphery: (figs. 83 to 85), the other degenerates. The vesicular phase of the former is preliminary to the formation of large, Figs. 74 and 75 Wandering pronuclei passing and after passing one another. <21500: poorly staining granules which fill the vesicular nucleus (figs. 86 and 87). As these granules are formed, the peripheral chro- matin bodies disappear, their material evidently changing into 328 GARY N. CALKINS the larger, pale granules. With this new nuclear complex, viz., one ghost-like new macronucleus (‘placenta’), one degenerating micronucleus, and two functional micronuclei, the conjugating individuals separate (figs. 79, 80, and 86). Fig. 76 Pair with pronuclei in each individual about to fuse. X 800. Fig. 77 Initial stage of fusion of the pronuclei. > 1500. Fig. 78 Elongation to spindle form of the fusing pronuclei. X 1500. IV. THE EX-CONJUGANT Separation of the conjugating individuals begins at the pos- terior part of the fused regions, progress in separation being marked by the increasing length of the free peristomes (figs. 79 and 80). Finally, a connection remains only at the frontal mar- gins, and the two cells pull apart. For a few seconds, only, they remain quiet, but within a few minutes their activities are fully restored. THE NUCLEI OF UROLEPTUS MOBILIS 329 Unlike most of the hypotrichous ciliates in culture, the great majority of Uroleptus individuals continue to live after conjuga- tion. Of thirty pairs isolated, no less than fifty-three ex-conju- gants continued to live and to divide. When we consider Bait- Figs. 79 and 80 Early stages in the development of the new maeronuclei. X 800. sell’s (12) experience with Stylonychia and my own experience with the heterotrich Blepharisma, in which 100 per cent of the ex-conjugants died, this result is remarkable. Even Paramecium has a much larger post-conjugation mortality. There is no a priori reason why mortality in one genus should be higher than 330 GARY N. CALKINS Fig. 81 First division metaphase of the amphinucleus; the chromosomes on | one side are partly divided. X 3200. Fig. 82 Anaphase of the second division after fertilization with two sets of eight chromosomes, and resting nucleus, the other product of the first division. X 3200. Figs. 83, 84, and 85 Late telophage stages and final separation of the macro- nucleus-forming second division after fertilization. > 3200. Figs. 86 to 90 Stages in the development of the new macronucleus, concen- tration of the fragments of the old macronuclei and their ultimate disappearance, and stages in the concentration of the two new micronuclei. Figure 86 one hour old, the others from 48 hours to 96 hours old. X 800. THE NUCLEI OF UROLEPTUS MOBILIS aol that in another, and certainly no theoretical interpretation of the purpose of conjugation, admits a mortality of 100 per cent. The explanation of such mortality must be sought in the environ- mental conditions, and particularly in the culture medium used. The continued life of Uroleptus ex-conjugants indicates a normal condition of environment and a culture medium that satisfies all conditions of vitality. The living ex-conjugants can be distinguished at a glance from individuals in any other phase of vitality. They are shorter than ordinary vegetative forms and larger in diameter. From dividing forms, or individuals immediately after division, they are distin- guished by the clear, non-refractile, vesicular macronucleus which appears, vacuole-like, for two or three days’ after conjugation. The young ex-conjugant contains, in addition to the new mac- ronucleus and two new micronuclei, the granular remains of the eight original macronuclei (fig. 86, one hour old). In some in- dividuals, the typical linear arrangement of the original macro- nuclei is retained (fig. 86). In others there is some confusion in arrangement with a tendency to mass near the center (fig. 87). In all cases such massing eventually occurs preliminary to final absorption and always in the posterior half of the cell. During the first forty-eight hours after separation, each of the eight macronuclei which had developed increasingly large granules during the conjugation stages again forms a compact, spherical, homogeneous mass of chromatin which takes an intense nuclear stain (fig. 87). These slowly disappear by absorption in the protoplasm. First one, then another, loses its staining capacity and fades away leaving no trace. In some cases, however, they again undergo granular fragmentation before being absorbed in the cytoplasm (figs. 88 and 89). Ultimately, four to five days after separation, not a trace remains of the old macronuclei. In the meantime a new macronucleus is developing. The pale ghost-like character of this cell organ is retained for at least three days, but it becomes much larger and ellipsoidal in form during this period (figs. 89 to 91). The granules within, and solidly filling it, now assume a definite spherical form and begin Son GARY N. CALKINS to stain more intensely. Each elongates to form a chromatin rod, and these arrange themselves in lines running from end to end of the nucleus. This ultimately divides by simple constriction with dumb-bell formation; the two daughter nuclei divide again, and these once again, until eight new macronuclei replace those that were absorbed (figs. 91 to 95). Figs. 91t095 Stages in the division of macronuclei and micronuclei leading to the formation of the typical nuclear complex of the vegetative forms, 96 to 120: hours after separation of the conjugants. X 800. The micronuclei also divide during this period until from six to eight are formed. Some of these degenerate, leaving the typical number four to six, in the young individual (figs. 93 to 95). The formation of the nuclear clefts in the macronuclei of the young individuals involves some processes which strongly sug- gest the entrance of supernumerary micronuclei. This will be left for further investigation, however, and will be dealt with in a separate paper on the formation and significance of the nuclear cleft. THE NUCLEI OF UROLEPTUS MOBILIS ooo V. COMPARISONS A. Multiple nuclei in ciliates -In many ciliates a multiple number of nuclei, both macro- nuclei and micronuclei, is the rule. Balbiani (’60, ’61), in his earlier work at least, held that the number of micronuclei is always the same as the number of macronuclei, or in beaded forms, as many as there are segments of the macronucleus. Maupas (83), Gruber (’87), Biitschli (88), and many later observers, disproved this view and demonstrated that, in some cases, the numbers are the same (as in the majority of Oxytrichidae); in other cases, e.g., Trachelius ovum, Condylostoma patens, Stentor, ete., the micronuclei outnumber the macronuclei; and in still other cases the macronuclei outnumber the micronuclei (e.g., Uronychia transfuga, Gonostomum pediculiforme, Holosticha lacazei, Loxophyllum meleagris, Spirostomum ambiguum, ete.). The relative numbers of the two types of nuclei, furthermore, may differ in different individuals of the same species. This is the case in Uroleptus mobilis, where also the number of micro- nuclei is inferior to the number of macronuclei. a. Fusion of macronuclei during division. It is quite probable that the multinucleate condition of macronuclei is only an ad- vanced stage in physiological adaptation to the needs of the cell. At one extreme, the more generalized condition, we find the typical uninucleate cells of the majority of the infusoria. At the other extreme are forms like Dileptus gigas, in which the endo- plasm is filled with minute chromatin bodies each of which be- haves like a granular nucleus. Between these two extremes lie the ciliates with band-form, branched, beaded, or multiple macronuclei. The single nucleus of Spathidium spathula is. drawn out as a rod; in Euplotes, Aspidisea, Didinium, Vorticella, etc., the rod becomes horseshoe shape; in many species of Suc- toria it becomes much branched; in Stentor, Spirostomum ambi- guum, Bursaria, and others the rod becomes more or less con- stricted at intervals to form the characteristic beaded nuclei. All of these conditions are modifications of the uninucleate type. 334 GARY N. CALKINS The transition to the multinucleate type is by no means abrupt. In the genus Stentor, for example, Johnson (’93) shows that in S. polymorpha the connecting strands (commissures) are relatively thick and conspicuous, while ip 8. pyriformis and in S. igneus no trace of such commissures could befound. Inthese two species, therefore, the multinucleate condition results from a fragmen- tation of the beaded rod form. A similar fragmentation of a homogeneous rod results in the formation of ten to fourteen nuclei of Uronychia transfuga, but in this case a very delicate connecting membrane can be made out. An extreme case, again, is the multiple fragmentation observed by Lebedew (’08) in Trachelocerca. Balbiani, Biitschli (88), Maupas (’83), and others have main- tained that, with few exceptions, in all multinucleate forms, the separate nuclei, or fragments, are held together by similar com- missures and are in effect, therefore, single nuclei. This conclu- sion was based in part upon the obvious connecting strands in many forms and in part upon the fusion, prior to division, of all the parts, or fragments, to form a single homogeneous dividing nucleus. The fact of fusion was regarded as proof of the fact of invisible commissures. Against this view, which is unquestionably too general, are the facts associated with those forms in which the multinucleate condition is a result, not of fragmentation, but of consecutive nuclear division. In most of the binucleate Oxytrichidae the two nuclei arise by equal division, the division taking place just prior to cell division. Here there is, indeed, no fusion, and the effect is that of a single nucleus which has divided precociously for the following cell division. Gruber (’87) showed that a similar fusion of entirely discon- nected small nuclei of Holosticha scutellum occurs during prep- aration for division, and that these nuclei arise from the repeated independent divisions of a single nucleus during and immediately subsequent to cell division. Exactly the same phenomenon oc- curs in Uroleptus mobilis; the disconnected nuclei, both after division and after conjugation, arise by repeated consecutive divisions of a single nucleus. THE NUCLEI OF UROLEPTUS MOBILIS ooo It is obvious, in these cases, that the multinucleate phase is an adaptation from the uninucleate condition, and the conclu- sion may be drawn that the independent nuclei, by fusion, return to a more primitive and more generalized uninucleate condition prior to division. In some forms, finally, the independent nuclei do not fuse again prior to division and each divides independently, the prog- eny passing to the daughter cell in which they happen to be (Trachelocerca, Dileptus gigas, Ichthyophthirius, Opalina). b. Absence of fusion of micronucler during division. The num- ber of micronuclei in Uroleptus mobilis varies between two and six. At no time during vegetative life, division, or conjugation is the number reduced to one, except for the very brief period during conjugation, when the functional pronuclei fuse, and before the first division of the synkaryon. Even at this time, however, there are still from three to five degenerating pronuclei in the cytoplasm of each individual. During the five days required for reorganization of the cell after conjugation, there are two micronuclei, and they begin to divide in the same period as the new macronucleus. As a result of their division, the fully reor- ganized cell has six to seven micronuclei. During the early phases of division, viz., during the fusion of the macronuclei, these micronuclei are found in full mitosis (figs. 4 to 14). I have never seen as many as six In mitosis at one time, but have fre- quently found individuals with five and four at this period of macronuclear concentration. When the macronucleus is ready to divide I have found individuals with two and with four micro- nuclei in mitosis (figs. 10 and 11), but never with more. If two is the initial number as indicated by ex-conjugants, then it is reasonable to infer that the number becomes reduced to two before division of the macronucleus. If this is the case, how has the reduction taken place? On this point I have no certain evidence. Balbiani (’60, ’61) at first believed that multiple micronuclei are bound together in a common pouch with fine connectives like those between multiple macronuclei. No later observer has confirmed the hypothesis, and Balbiani himself retracted this 336 GARY N. CALKINS view in 1881. Gruber (’87) observed a single micronucleus in the dividing stage of Stichotricha scutellum and followed it through two or three divisions, after which he was unable to detect any trace of micronuclei. He inferred that they become progres- sively reduced in size until they could not be identified among the many macronuclei. He also inferred that the conspicuous micronucleus at the time of division is a product of the fusion of all the minute nuclei distributed throughout the cell, just as the single macronucleus is the fusion product of all the macronuclei. At best this was only an hypothesis. Other cases of fusion have been suspected either during so-called autogamous fertilization (Ichthyophthirius, Neresheimer, ‘08; Buschkiel, ’11) or during encystment (Stylonychia, Fermor, 712). In no case, except at fertilization, has the fusion of micronuclei been established. Nor in Uroleptus mobilis have I any evidence that the reduc- tion of micronuclei at division is due to fusion. The two micro- nuclei at this period are no larger than the five nuclei at an earlier stage (figs. 10 and 11). All that I can say is that some micronuclei, after attaining a condition of full mitosis, simply disappear. B. Conjugation There are approximately 1800 species of ciliated protozoa known to science. In 98 per cent of these the conjugation proc- esses are entirely unknown, while the facts regarding conjuga- tion in the remainder are in many cases only fragmentary and the full history is largely conjectural. In a few instances the story is fairly complete and, verified by different observers, has come to be regarded as the characteristic history of conjugation in all ciliates. How far this is true can only be determined by independent study of conjugation in different species and a study uninfluenced by preconceived notions of what the process should be. The widely diverging accounts of the details in forms like Paramecium (Hertwig, Maupas, Hamburger, Calkins and Cull), Anoplophrya (Collin), Trachelocerea (Lebedew), Ichthyophthirius (Nerescheimer, Buschkiel) show that no uniformity characterizes the phenomena and that, until many more data are obtained, generalizations have little value. THE NUCLEI OF UROLEPTUS MOBILIS Sar Very few complete studies have been made of conjugation in hypotrichs. Apart from, and since, the classical, epoch-making work of Maupas, who included the full history of conjugation in the hypotrich Onychodromus grandis and a somewhat less com- plete history of Stylonychia pustulata, of Euplotes patella, and E. charon, there has been only one work on hypotrichs, that of Prowazek on Stylonychia pustulata in 1899. As is well known, Maupas (’88) described eight (A to H) phases common to all types in the process of conjugation. Phase A, the initial phase, is characterized by the growth and early changes of the micronucleus; the second phase, B, by the first divis on of the micronucleus. In modern terms this phase is called the first maturation division or first meiotic division. The third phase, C, is the period of the second division, now called the second maturation or second meiotic division. The fourth phase, D, is the period of the third division resulting in the formation of the pronuclei. The fifth phase, E, is the period of interchange and fusion of the pronuclei; the sixth, F, and seventh, G, are the stages of the first and second divisions, respectively, of the fertilization nucleus or amphinucleus. The last phase, H, finally is the period between the second division of the amphinucleus and the first fission of the cell after conjugation. The keen powers of observation and generalization which Maupas pos- sessed are well shown by this recognition of the successive stages in conjugation, particularly in connection with the distinction between the first and second divisions of the micronucleus, and at a time when the significance of these divisions in maturation phenomena was quite unknown. Subsequent investigations have shown that, with minor variations, these successive phases hold good for the great majority of cases. a. The preparaiory stage (phase A) of Uroleptus mobilis. The difficulties in working out the initial stages of conjugation in Uroleptus are increased because of the multinucleate condition of the conjugating organisms. The number of micronuclei in vegetative stages varies from four to six and the same variable numbers appear in the conjugating individuals. Analogous con- ditions are found in Paramecium aurelia (Hertwig, Maupas), 338 GARY N. CALKINS Onychodromus grandis (Maupas), Stylonychia pustulata (Mau- pas, Prowazek, 99), each having two micronuclei; in Didinium nasutum, with two or three (Prandtl, ’06), in Blepharisma undu- lans, with from four to five micronuclei (Calkins, ’12) and in Bursaria truncatella with from sixteen to eighteen: micronuclei (Prowazek, 99). Maupas (’88) described an anomalous, additional, or prelim- inary division in the case of Euplotes patella and in E. charon, where it is found in both conjugants. In other cases where such a division occurs (in the peritrichida) it is limited to the mi- crogamete (Vorticella monilata, V. nebulifera (Maupas, ’88), Carchesium polypinum (Maupas, ’88; Popoff, ’08), Ophrydium versatile (Kaltenbach, °15), and Opercularia coarctata (En- riques, ’07). The prophases of the first maturation spindle in ciliates are fre- quently highly characteristic. In Loxophyllum meleagris (Mau- pas, ’88), Spirostomum teres (Maupas, ’88), Euplotes patella (Maupas, ’88), Colpidium colpoda (Hoyer, ’99), and in Ble- pharisma undulans (Calkins, ’12) however, there appears to be no typical prophase stages beyond the swelling of the nucleus, fragmentation of the homogeneous chromatin of the micronu- cleus and formation of the chromosomes. Hoyer (’99) describes a typical spireme in the case of Colpidium colpoda, but this is very exceptional in ciliates and needs confirmation. In Paramecium (aurelia, bursaria, and caudatum) a very typi- cal prophase stage occurs in the form of a crescent, derived from the homogeneous micronucleus which first draws out in the form of a long cylinder which later forms the characteristic crescent. A modification of the crescent occurs in Chilodon uncinatus (Maupas, ’88; Enriques, ’08), where the chromatin is drawn out in the form of an elongate comma-shaped band. This is still further modified in Cryptochilum-nigricans (Maupas, ’88), Vorti- cella monilata and VY. nebulifera (Maupas, ’88), and in Oper- cularia coarctata (Enriques, ’07), where a long chromatin rod extends, in some cases, the entire length of the cell. Still another type of prophase, and a type to which Uroleptus belongs, is found in Onychodromus grandis (Maupas, ’88), Bur- saria truncatella (Prowazek, ’99), Didinium nasutum (Prandtl, THE NUCLEI OF UROLEPTUS MOBILIS 339 706), and Anoplophrya branchiarum (Collin, ’09). Here the nu- clear membrane first swells up to form a nucleus two or three times the diameter of the original micronucleus, while the com- pact mass of chromatin, placed either centrally or peripherally, fragments into numerous chromatin granules. In Uroleptus mo- bilis there is an intranuclear centrosome which divides, one-half passing to the periphery of the nucleus at the pole opposite the chromatin mass, while the other half remains in this mass. The two halves remain connected by a deeply staining strand (prim- ary centrodesmus) throughout the prophase period, but none of. these structures can be demonstrated in the fully formed spindle (figs. 28 to 34). The distal centrosome is the focal point of spindle fibers which spread out from it to the fragmenting chromatin mass and forms one pole of the mitotic figure. It was a similar stage in Anoplophrya that suggested the term ‘cande- labra’ to Collin. I have ealled it the parachute nucleus. So far as I am aware, this is the first demonstration of the centronucleus type of nucleus in ciliates. The majority of investigators show signs of embarrassment when it comes to the description of the formation of the second pole, or the full spindle, of the first maturation nucleus. In the _ transformation of the crescent type, Maupas, Hertwig, and Ham- burger all agree that the spindle is formed by the shortening of the long axis of the crescent and that the tips of the crescent form the poles of the spindle. Calkins and Cull, however, find that the division center or kinoplasmic mass which forms the poles of the spindle, migrates from its terminal position in the crescent to the center of the convex side. This new position becomes the first pole of the spindle. In the candelabra or parachute type, the second pole is formed by the outgrowth from the chromatin mass, of a second pole similar to the first, the chromatin granules thus being left in the nuclear plate position or center of the spindle figure. In Uroleptus mobilis this second pole is formed by the migration of the proxi- mal centrosome from the chromatin mass while the connecting strand between the two centrosomes becomes thinner and im- possible to distinguish, in the later stages, from the spindle fibers. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 3 340 GARY N. CALKINS b. First maturation spindle and the chromosome problem. In the conjugation of ciliates with multiple micronuclei the usual result of the prophase activities is the formation of one first maturation spindle for each of the micronuclei present. In Onychodromus grandis each of the two micronuclei forms a spindle (Maupas, 88). The same is true of Stylonychia pustulata (Maupas, 788; Prowazek, ’99), and Paramecium aurelia (Hert- wig, 89). In Didinium nasutum (Prandtl, ’06), and in Blephar- isma undulans (Calkins, 712), if more than two micronuclei are present, only two of them form spindles; in Bursaria truncatella, on the other hand, all of the sixteen to eighteen micronuclei are active at this stage (Prowazek, ’99). In Uroleptus mobilis there may be two, three, or four of these primary spindles (figs. 45 to 47). The chief interest of these first maturation nuclei lies in the so-called chromosomes, especially in the methods of their for- mation and division. Unfortunately, we have but little exact knowledge on these points owing to the frequent large numbers and small size of the chromatin elements. Maupas (’88) made no attempt to enumerate the chromosomes; nor did he describe their formation beyond the brief account of the fragmentation of the homogeneous chromatin masses of the micronuclei. Hertwig (’89) believed that there were eight or nine chromosomes in Paramecium aurelia, basing his view not on the chromosomes, but on the number of fibers which he could distinguish in the connecting strand between the two daughter nuclei. Later observers find that the number, in all species of Paramecium, is very much greater than this (more than 100 (Hamburger, Calkins, and Cull) ). In some more favorable types than Paramecium the number of chromosomes has been made out with some degree of accuracy. Prandtl (’06) finds sixteen in Didinium nasutum, a number which I have been able to confirm (unpublished). Prowazek (’99) was a little in doubt whether there were twelve or thirteen in the nu- clei of Bursaria truncatella, but describes six chromosomes in Stylonychia pustulata. Stevens (’10) describes four chromosomes in Boveria subeylindriea, var. concharum, but gives no details as to THE NUCLEI OF UROLEPTUS MOBILIS 341 their formation or reduction; Enriques (’08) finds the same num- ber in Chilodon uncinatus. Popoff (08) enumerates sixteen chromosomes in Carchesium polypinum in the first maturation spindle, and Enriques (’07) finds the same number in Opercularia coarctata, while Kaltenbach ('15), somewhat doubtfully, attrib- utes twenty chromosomes to Ophrydium versatile. Collin (09), finally describes six chromosomal elements in the first spindle of Anoplophrya branchiarum. Hamburger (’04) is a bit hazy in her description of the origin of the chromosomes in Paramecium bursaria. The late stage of the crescent is regarded as a spireme from which the chromosomes are formed as short, curved, or V-shaped rods. Calkins and Cull derive the chromosomes of Paramecium caudatum from a synezesis stage which precedes the crescent, and from which the chromosomes emerge as double rods which are fully divided by the time the first spindle is completed. According to this ac- count, the metaphase stage occurs during the metamorphosis of the crescent into the spindle, so that the latter when formed is always in the early anaphase stage. The process of chromosome formation in Paramecium caudatum may be regarded as a linear fragmentation of the dense chromatin of the micronucleus. In the other forms in which the chromatin history has been followed, the process may be interpreted as a granular fragmentation, and the problem is to construct the chromosomes from these granules. The parachute nucleus, which is characteristic of most of the ciliates with a small number of chromosomes, may be compared with the Paramecium nucleus at the time when the division center has migrated from the apex of the crescent to the middle of the convex side. The chromatin content of such parachute nuclei consists of separated granules of chromatin, of which the number is difficult to count. In Uroleptus mobilis when diffusion of the granules has apparently reached its limit, I find from twenty-four to twenty-eight such granules (figs. 35 to 37). Prandtl’s figures show that there are approximately thirty-two in Didinium nasutum. Enriques (’08) and Collin (09) describe a similar fragmentation of the comma- form chomatin rod of Chilodon uncinatus and of the homogeneous 342 GARY N. CALKINS chromatin mass of Anoplophrya branchiarum, the granules of chromatin collecting in the center of the first maturation spindle. In Didinium, Chilodon and Anoplophrya, these granules fuse until a definite number of chromosomes result, sixteen in Didin- ium, four in Chilodon and six in Anoplophrya. In Uroleptus there may be such a fusion of granules to form eight chromosomes (fig. 40), or the nucleus may divide without such fusion, the eranules being distributed in two equal groups in the daughter nuclei (figs. 38 and 39). It is tempting to assume that the former mode is typical and that eight chromosomes represent the nor al condition of the first maturation spindle. This assumption would bring Uroleptus in line with the maturation phenomena of higher animals. But the fact of the matter is that I find just as many spindles of one type as of the other, and both types are, apparently, equally potent. In no ciliate, in which the number of chromosomes can be counted, does this first division result in a reduction to one-half. Furthermore, there is much uncertainty as to whether the divi- sion of the chromosomes is longitudinal or transverse. Prandtl (06) could not determine which it is in Didinium nasutum; Collin (09) was equally uncertain; indeed, he seems to have had some doubts as to whether these granular elements on the first maturation spindle were even to be regarded as chromosomes. Enriques (’08), while admitting the difficulties of interpretation, is convinced that the chromosomes of Chilodon uncinatus are transversely divided. Popoff (08) did not observe the fact, but assumes, apparently from analogy with conditions in many meta- zoa, that this first division in Carchesium polypinum is equational. In Paramecium the chromosomes are not granules, but rods which are too numerous to count. With this form it should not be difficult to determine whether the division of each rod is trans- verse or longitudinal. Owing to the early views in regard to the origin of the first spindle from the crescent, it was generally assumed that division was transverse. Calkins and Cull (07), however, found that the rods are double when formed and that the pairs are each separated in this first maturation division. Such a longitudinal division would correspond, therefore, with a THE NUCLEI OF UROLEPTUS MOBILIS 343 reducing division if the conditions in protozoa are the same as in metazoa. This, however, cannot be granted offhand for the second division is also longitudinal, and it was impossible to tell whether the double chromosomes. were likewise longitudinally divided. ~Waldeyer (’88) defines chromosomes as ‘‘the deeply staining bodies into which the chromatic nuclear net-work resolves itself during mitotic cell-division” (quoted from Wilson, ’99).. Apart from the chromatic nuclear net-work from which they originate, this definition would include the chromosomes of the ciliate micronucleus. But it would also include the granules of chro- matin which elongate and divide in the typical macronucleus during division. The number of chromosomes, furthermore, is regarded as fixed for the same species, but the number of dividing chromatin granules in the macronucleus is immeasurably greater than the number of chromosomes of the micronucleus. It follows that either these dividing granules of chromatin are not all chro- mosomes or that the number of chromosomes is not fixed for the same species in all cases. It isobvious that, according to the definition, all chromatin bodies which arise from the chromatin of the resting nucleus are not chromosomes. The difficulty in most cases might be avoided by a strict interpretation of the definition which involves ‘mitotic’ cell division, although in some cases, e.g., Spirochona gemmipara (Hertwig, ’77), the macro- nucleus divides by mitosis. The suggestion made by Rabl that the chromosomes retain their individuality throughout life of the organism has grown to a strong conviction in later years, es- pecially through the work of Conklin, Morgan and recently Richards. I doubt very much, however, if anyone can be found hardy enough to apply this conception to the chromosomes of protozoa. Nevertheless, I believe that we are justified in re- garding chromosomes in ciliates as equivalent structures to chromosomes of the metazoa. In the prophases of the first maturation spindle in ciliates we find parallel stages with those of the first maturation division in higher animals. The synezesis stage is represented by the chromatic net-work which arises from the homogeneous chromatin in the parachute type of nucleus 344 GARY N. CALKINS (figs. 28 and 29) and the fusion of granules to form the chromo- somes parallels their origin in higher forms. c. The second maturation division. In all ciliates in which the number of chromosomes has been accurately counted, the reduc- tion in number occurs during the second maturation division. This was first made out in the case of Didintum nasutum by Prandtl (06), who found that the sixteen chromosomes resulting from the first maturation division, were separated into two groups of eight each in the second division. This method of reduction appears to be characteristic for ciliates and agrees with what Goldschmidt (05) designates the ‘first type of reduction’ exemplified among metazoa by Zoogonus mirus. Prior to Prandtl’s work there were no conclusive observations on reduction in ciliates. Maupas (’88) and Hertwig (’89) made no attempt to follow the chromosome history. Prowazek (’99), while mentioning twelve to thirteen chromosomes in the first maturation division of Bursaria truncatella and six in Stylony- chia pustulata, does not give their fate in the second division. Subsequent to Prandtl’s work, we find a number of well-defined eases of chromosome reduction. In Opercularia coarctata En- riques (’07) finds a reduction from sixteen chromosomes to eight, the process agreeing with that in Didinium. In Chilodon uncinatus (Enriques, 08) the same observer describes the four chomosomes resulting from the first maturation division as fusing to form two pairs in the telophase stage. In the second divi- sion each pair divides, the daughter nuclei receiving two chro- mosomes each. A similar fusion of two pairs of chromosomes in vegetative mitosis was observed by Stevens (10) in Boveria subcylindrica, but she was unable to trace the history of the four chromosomes in the maturation processes. In Carchesium polypinum, Popoff (08) there are, again, sixteen chromosomes in the products of the first maturation division, which are sep- arated into two groups of eight each by the second division. In Anoplophrya branchiarum, Collin (’09) describes six chromo- somes in the first maturation spindle, each of which is equally divided. The second maturation spindles were not identified, but this excellent and equally candid observer finds only three THE NUCLEI OF UROLEPTUS MOBILIS 345 chromosomes in the nuclei undergoing the third division, and concludes that the reduction must have taken place during the second maturation division. In Uroleptus mobilis, finally, we find a process of reduction conforming to these processes in the other ciliates described. The eight chromosomes of the first maturation mitosis (or the many granules of chromatin in aber- rant types, fig. 40), fuse, after division, to form eight chromosomes of the metaphase of the second maturation division (figs. 54 and 55). But these eight chromosomes become partially fused to form four pairs. This phenomenon is similar to that de- scribed by Enriques (’08) as occurring in Chilodon uncinatus; but in Uroleptus the fusion is not complete, the two distinct chromosomes of each pair simply lie in close contact (fig. 55, a). The early anaphase stages of this second division furnish some evidence indicating that the two members of each pair possibly slide apart in opposite directions, so that four single chromosomes finally collect at each pole (fig. 55). The observations on chromosomes and reduction in ciliates may be summarized in tabular form as follows: Lag Lag EEE [EEE GENUS, SPECIES, AND ORDER OBSERVER . a 5 a S ae z assn/25 =a BOuPIZORE = BaAlS aNA Anoplophrya branchiarum (Holotrich)......... Collins, ’09 6 3 Boveria subcylindrica (Holotrich)..............] Stevens, ’10 4 is Bursaria truncatella (Heterotrich)..............] Prowazek, ’99 12-13 ? Carchesium polypinum (Peritrich)............. Popoff, ’08 16 8 Chilodon uncinatus (Holotrich)................ Enriques, ’08 16 8 Didinium nasutum (Holotrich).................] Prandtl, ’06 16 8 . Opercularia coarctata (Peritrich).............. Enriques, ’07 16 8 Ophrydium versatile (Peritrich)................| Kaltenbach, ’15 | 20+ te Stylonychia pustulata (Hypotrich).............| Prowazek, ’98 6 ? Uroleptus mobilis (Hypotrich).................| Calkins, 719 8 4 As to the number of nuclei participating in the second matura- tion or reducing division, we find many variations. In ciliates having but one micronucleus in the vegetative stages the numer- 346 GARY N. CALKINS ical relations are fairly uniform, two spindles in the second maturation division being the rule. There are, however, some exceptions. ‘Thus in Paramecium bursaria, according to Ham- burger (04), one of the nuclei formed by the first maturation divi- sion degenerates without forming a spindle, so that only one nucleus undergoes the second maturation division. A second exception is found in Euplotes patella and in all vorticellidae examined up to the present time. Here the micronucleus under- goes one preliminary mitosis prior to the first maturation divi- sion (p. 338). The resulting two nuclei then undergo the first maturation division and the four resulting nuclei form eight by the second maturation division. In vorticellidae this unusual division occurs only in the microgamete. The micronucleus of the macrogametes, on the other hand, does not undergo a pre- liminary division and the usual history of uninucleate forms is followed. In ciliates with two micronuclei, both undergo the first matu- — ration division. According to Prowazek (’99), the four resulting nuclei in Stylonychia pustulata divide again, thus forming eight products of the second division. According to Maupas (’88), however, two of these first four nuclei of Stiylonychia pustulata, and of Gre, chodromus grandis as well, degenerate so that aay -two nuclei divide in the second maturation division. In ciliates with many micronuclei in the vegetative stage there seems to be no general rule as to the number which undergo the second maturation division, unless, indeed, it be variability. Prandtl (06) found a variable number in Didinium nasutum; Prowazek, (’99) a large number in Bursaria truncatella, and I find a variable number in Uroleptus mobilis. The usual number of spindles in the second maturation division here is two, although three are frequently found, while individuals with one or with four have not been seen. d. The third division. The third, or pronuclei-forming divi- sion, is a peculiarity apparently almost universal in the Infusoria. The spindles are frequently heteropolar (Didinium, Paramecium), and the telophase stage is often characterized by long connecting strands of nuclear substances (Paramecium Blepharisma, Uro- THE NUCLEI OF UROLEPTUS MOBILIS 347 leptis, etc.). There is no uniformity in regard to the number of nuclei to undergo this third division, although only one of the dividing nuclei provides the functional pronuclei. In Anoplo- phrya, Paramecium, Chilodon, Colpidium, Leucophrys, Glau- coma, Loxophyllum, Spirostomum, Bursaria, Blepharisma, Bo- veria, Lionotus, and in the vorticellidae only one spindle is formed for this third division. In Onychodromus, Stylonychia, and Euplotes, according to Maupas (’88), two spindles are pres- ent and four pronuclei are formed, two of which degenerate and disappear (in Stylonychia, according to Prowazek (’99), only one third division spindle is present). Uroleptus mobilis differs from the majority of other ciliates in having from two to four nuclei undergoing this third division at the same time, and as many as eight products of this division may be present in the cell before any two are metamorphosed into pronuclei (fig. 61). e. The pronuclet. Prandtl (’06) was the first to note a differ- ence in size between the wandering and the stationary pronuclei of Didinium nasutum. Calkins and Cull (07) noted a similar difference in pronuclei of Paramecium caudatum, and were able to trace this difference back to a heteropolar third-division spin- dle. Other observers have failed to find any constant differ- ences, and this is the case in Uroleptus mobilis, where the size relations of the fully formed pronuclei, as carefully measured in ten individuals, are as s follows: PAIR INDIVIDUAL WANDERING PRONUCLEUS STATIONARY 1 A 23 X Su 3X Su B 3 xX Su 25 xX 25 by 2 A 24x Qheu B56 3) jw B Bie oy IN 2ix3u 3 A Bs yy 2ix3u B SOK ai 0 2i x 27 uw 4 A 24x 2h 23 X 3 B By ea) |b 24 X 3p 5 A by X 3s UL 3) ue 8) | B Sys al po 4x 4iu 1 Pairs 1 to 4 were fixed in Bouin’s fluid; pair 5 in sublimate acetic. 348 GARY N. CALKINS These differences are too uncertain to permit any conclusion as to the size relations between wandering and stationary pro- nuclei. In some cases the wandering pronucleus is smaller, in other cases larger than the stationary pronucleus. Apart from size differences, we occasionally find structural differences between the wandering and the stationary pronuclei. Maupas (’88) was the first to note the presence of a dense aggre- gate of cytoplasmic granules at the forward end of the advancing pronucleus of Euplotes patella, while no such aggregate was seen in connection with the stationary pronucleus. Hoyer (’99) ob- served ‘astral rays’ about each of the pronuclei of Colpidium colpoda, but without distinctive differences; similar radiations, more pronounced about the wandering pronucleus, were de- scribed by Prandtl for Didinium nasutum. In Uroleptus mobilis there are no radiations such as occur in Didinium, but the granu- lar aggregate at the forward end of the wandering pronucleus is highly characteristic and its presence confirmed in material fixed in sublimate acetic, in Flemming’s fluid, in Bouin’s fluid and in Schaudinn’s fluid. It appears to be a directive center, possibly analogous to the centrosphere of Noctiluca. It cannot be inter- preted as a collection of granules due to pressure of an advancing solid for the advanced end of the mass curves around the anterior end in a definite way (figs. 71 to 75). With the approach and union of the pronuclei this mass disappears. f. Fusion of the pronuclei. The outcome of the interchange of pronuclei is the fusion of migrating and stationary pronuclei. In all ciliates which have been carefully examined, with the ex- ception of the vorticellidae, this interchange and fusion is mutual. Hoyer (’99), however, holds a different view. He, like Maupas before him, failed to find evidence of the union of pronuclei in Colpidium colpoda, and concluded that no fusion occurs, the foreign pronucleus in each individual forming the functional micronucleus. In the face of the overwhelming evidence in other, and probably more favorable ciliates, this peculiar view cannot be admitted. In the vorticellidae the microgamete fuses with the macrogamete and loses its identity in the protoplasm of the larger conjugant. Here one of the two pronuclei.formed by THE NUCLEI OF UROLEPTUS MOBILIS 349 each conjugant degenerates, and only a single fertilization nu- cleus is formed. At the time of fusion the pronuclei may be either in the form of spherical and vesicular nuclei or drawn out into more or less pointed spindle form. We find the former type in Anoplophrya branchiarum, Didinium nasutum, Spirostomum teres, Blepharisma undulans, and Opercularia coarctata, while union of pronuclei in the spindle form is reported in all species of Paramecium, Chilodon uncinatus, Leucophrys patula, Onychodromus grandis, Stylony- chia pustulata, Vorticella nebulifera, Ophrydium versatile and Lionotus fasciola. In some cases it seems to be immaterial whether the pronuclei are vesicular or spindle form at the time of fusion (Loxophyllum meleagris, Carchesium polypinum). Uroleptus mobilis appears to be an intermediate type, for here the pronuclei approach and meet while in the spherical and vesicular form, but before the fusion membrane dissolves both pronuclei elongate and assume the spindle shape (figs. 77 and 78). The chromatin, however, is not aggregated into chromo- somes at this period, but lies in diffuse granular form exactly as in Paramecium caudatum. While the general form is that of a nucleus in division, these fusing pronuclei can no more be re- garded as mitotic spindles than can the fusing spherical nuclei; in both types we have merely the antecedent phases of mitosis. g. Reconstruction of the vegetative nucler. The first division of the fertilization nucleus occurs very quickly after fusion and is not often seen. In Uroleptus I have found only one spindle that can be identified as the first division spindle (fig. 81). In this spindle the two sides are not symmetrical, indicating the di- vision of chromosomes from one pronucleus earlier than in the other. The further history of the first two nuclei after conjuga- tion differs in different cases, and for some species the accounts by different observers do not agree. According to Enriques (’08), in Chilodon uncinatus one of these first two forms the new macro- nucleus and the other the new micronucleus. This, perhaps, is the simplest case on record. In all of the other cases described by different observers the macronucleus is not differentiated from the micronucleus until at least four nuclei have been formed 350 GARY N. CALKINS by a second division of the first two. One or two of these four may degenerate; one or two may form macronuclei. Or the four may divide again, forming eight nuclei before differentiation be- gins. In Anoplophrya branchiarum and in Euplotes patella two of the four nuclei degenerate, the other two form one macronu- cleus and one micronucleus. In Colpidium colpoda (Hoyer, 99), Stylonychia pustulata (Maupas, ’88) and in Lionotus fas- ciola (Prowazek, ,’09), only one of the four degenerates, two be- come functiénal micronuclei, and one becomes the macronucleus. In Didinium nasutum (Prandtl, ’06), Paramecium bursaria (Hamburger, ’04), Glaucoma scintillans (Maupas, ’88), Leuco- phrys patula (Maupas), Spirostomum teres (Maupas), and in Stylonychia pustulata (Maupas), two of the four nuclei become macronuclei and two become micronuclei, none of the nuclei degenerating. In Blepharisma undulans (Calkins, 712) all four of this stage become macronuclei enclosing the micronuclei. In another group the first four nuclei divide once again before the nuclei are differentiated. Here we find Paramecium cauda- tum, Paramecium putrinum, Cryptochilum nigricans, Carchesium polypinum, Vorticella nebulifera, and Ophrydium versatile. In the last three mentioned, seven of the eight nuclei form macro- nuclei. These fuse to fom one in Cryptochilum (Maupas), but in the other three forms they remain separated and are distrib- uted to the daughter cells by unequal division until each cell has one (Maupas, Kaltenbach, Popoff). In Paramecium cauda- tum and in P. putrinum, four of the eight nuclei form macronuclel, while the fate of the other four is differently interpreted. Mau- pas (’88) and Doflein (11) hold that three of these degenerate, leaving one functional micronucleus. Calkins and Cull (07) find that all four are functional. An exceptional history is shown in the reorganization of Bur- saria truncatella. Here no differentiation occurs until sixteen nuclei are formed (Prowazek, ’99). Two to five of them form macronuclei, three or more form micronuclei, and the remainder degenerate. In Uroleptus mobile difteréntiation of the nuclei occurs with the second division of the fertilization nucleus. One of the two nuclei formed by the first division divides unequally into a large vesicu- THE NUCLEI OF UROLEPTUS MOBILIS 345 | lar nucleus, the beginning of the macronuclei, and a small nucleus which degenerates. The chromosomes are eight in’ number in the first two divisions of the fertilization nucleus where they can be counted without much difficulty. After the differentiating second division there is a long pause, during which the young macronucleus undergoes its metamorphosis. During this period, which lasts for approximately 72 hours, the two functional mi- cronuclei become more compact and smaller, and when they finally divide, the spindles are so minute that no sure observation can be made as to the number of chromosomes contained. In mitoses from ordinary vegetative divisions, however, the number is eight, the chromosomes frequently lying in four pairs similar to the condition in the second maturation division but occa- sionally one or more pairs are not united so that five, six, or seven may be counted. h. Fate of the old macronucleus. All observers, with the ex- ception of Prowazek (’99, ’09), are agreed as to the fate of the old macronucleus. Prowazek, among the recent workers at least, appears to be alone in concluding that the old macronucleus is cast out of the cell as waste material. His conclusion is drawn from the premise that nucleins cannot be digested and must therefore be eliminated. As Collin (09) points out, Prowazek describes the loss of staining power of the old macronucleus frag- ments, and to this extent at least admits that the nuclei are di- gested. As he offers no evidence of such elimination from direct observation of the reorganization processes of Bursaria, Stylon- ychia pustulata, and Lionotus fasciola, his conclusion may be dismissed as highly improbable. The majority of observers have traced the gradual reduction in size of the old macronucleus or its fragments until nothing re- mains. This certainly is its history in Uroleptus mobilis, where the eight old macronuclei become greatly reduced in size and then fade away, one by one, until all are absorbed in the cytoplasm. The following table is a summary of the points made by dif- ferent observers on different species of ciliates and is useful for purposes of comparison. In most cases where two or more ob- servers have worked on the same species only the most important papers are cited. —_—s—O_.evwU’__-_ OO LLLLLPRPRLLLL—L—————<<< it & 7 fe tel att vin i Je) Ta eee) bee ee eer See ih Vialts ares) > : p a re bs Rabe . Pity’, emer Shon e win! TE Ne ihisil F Lo lites wae ; parwice adie Lin atm Cmgieeu dole tints ier ht “in 7 cl | Mruliob NY) ae lal epdir Ail Pa ON a PAS RTA eh Dee to, abe Sy Shenae a et a RA Hi Nour T ne a rekitinn . iy i cS RRR. Aaoteye halos: Ate at Beale well 4 Bisse Sale ae ROSEY emir, trl latent ary pected, . : Bieheatds: “ob io Buas Alin, , wiry! sail ri pt uly) Sen ‘ame oy th Paarios 2 wi toy, Wy tat eae! ites oie CS), Saliees ei Ag hice we fal ‘ an BAT a 4s aneip io 3 Luv aee | Gt pci dy Corry oe in ee a 7 lal) paint: yard ul ed a a eis) sey fate cL) i q ay i sATe:s inks lyaviciah (seers rid, ah Oh ian ys (May, on oa i Wee } Giphar, Mon HERS UNE? MS ol ee aa Cae Bato al lat vate gat Me ee ee a | penny te A RWI Hh i Vani ies sah Ast nee e hat Dy He aie eae as ee ae HM tiped)s Ab A En : * Srabbitd ae ud — | pele ag x Bias 2). Whe “ee oe Resumido por el autor, 8. O. Mast. Reversi6n en la orientacién hacia la luz en las formas coloniales Volvox globator y Pandorina morum. 1. Volvox y Pandorina son generalmente positivos en la luz débil y negativos en la luz intensa, si estan adaptados a la oscuri- dad; pero si estan adaptados a la luz, lo contrario sucede a veces. 2. Silas colonias adaptadas a la oscuridad se exponen a una ilumi- nacion constante, al principio se comportan como neutrales, de- spués se transforman en positivas, mds tarde en negativas y final- mente en positivas otra vez. Cudnto mas intensa es la ilumina- cién mds corto es el tiempo necesario para pasar por todos estos estados, pero en tales intensidades es necesaria mucha mas energia para producir los cambios en la orientacién que en una ilumina- cion menos intensa. 3. La reversién depende en cierto grado de la cantidad de energia recibida, pero bajo ciertas condiciones pa- rece depender principalmente de la duracién del cambio de ilumi- nacion. 4. La reversién no estd regida por la fotosintesis. La luz roja y la amarilla, en las cuales la fotosintesis se verifica con relativa intensidad, producen pocos efectos sobre la reversi6n, mientras que en la luz verde y en la azul, en las cuales la fotosin- tesis es relativamente débil, son casi tan efectivas como la luz blanea. 5. Los rayos que poseen la mayor eficiencia estimulante (azul y verde) son los mas potentes en la produccién de la rever- sion. 6. El sentido de la orientacién depende del estado fisio- légico de las colonias, asi como de la constitucién del medio de cultivo. Depende también de la edad de las colonias. Las colonias jOvenes son mas aptas para ser negativas que las de mas edad. 7. La reversiOén esta’ probablemente asociada con cam- bios en la permeabilidad. Translation by Dr. José F. Nonidez, Columbia University. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 REVERSION IN THE SENSE, OF ORIENTATION TO LIGHT IN THE COLONIAL FORMS, VOLVOX GLOBATOR AND PANDORINA MORUM S. O. MAST Marine Biological Laboratory, Woods Hole, and the Zoological Laboratory of the Johns Hopkins University TWO FIGURES CONTENTS MER OGUG GLO Me act are no) oi ncn tte enh a 8 arene cle ant i hn aS oe e 367 WMaterialsxamal metnodsesere sci sce her cnc eee Anbar e een eee 368 Relation between illumination and reversion in orientation................ 369 Relation between reversion in orientation and the quantity of light energy 375 Relation between reversion in orientation and the time rate of change in TNT ST OTS. Meee mR aS chek too ty oot ee REG oslo ey ekagecebckamt “a Sea a ey ee ae 377 Hiectrombheage on thercolonies on reversiOMes..-..4 se sesso ude so acee aur 378 Hitech om photosynthesissonsreversiOmaaseacim ahs aes veo ae neta eee ee 379 [Bik OH USMY OSIRAARUIAS Oiol, MENWEMEMONG sodcooos oh aomedbooooononesodensudbe cOGE 382 Hite ct oOrmchemirces] Stonereversvonssanie Geo sa a) ager. seas 4. mae annem 383 DISCUSSION etree ters oh Me eT eOR SAE os Sed: EOE SROs ar SPIEL Ld SEU RRAM EY areal tie aS 385 SHUI OTD OT EW ANY sche en ec ror Cie OS el Coane ee Sits iota Rh Re We 388 GET A IEC RCTCCO cerca eees AERA rE aN Toe eae oe A oe aE eM tals Brno 390 INTRODUCTION The literature on reversion in the sense of orientation was briefly reviewed in a preceding paper which dealt primarily with the effect of chemicals on the sense of orientation to light in Spondylomorum (Mast, 718). The results presented in that paper indicate that reduction in alkalinity, increase in some anes- thetics, increase in temperature and decrease in illumination all have the same effect on the sense of orientation, causing photo- negative specimens to become photopositive. They also indi- cate that the same change in the sense of orientation may occur without any appreciable change in the ‘environment. On the basis of these results it was concluded that reversion in the sense 367 368 Ss. O. MAST of orientation is probably due to some specific change or state in the physiological processes of the organisms which can be induced by any one of the factors mentioned, i.e., alkalis, anesthetics, temperature, or light. In this paper we shall deal primarily with the effect of illumi- nation on the sense of orientation, but we shall also briefly con- sider some other factors. MATERIALS AND METHODS The specimens used in this investigation were all collected at Woods Hole in a fresh-water pond. Volvox appeared early in July and continued for nearly a month. It was found only at the edge of the pond in a few small puddles, but in these it was very abundant most of the time. Pandorina appeared early in August, shortly after Volvox had disappeared, and it continued nearly to the end of the month. It was not found in all parts of the pond, but was much more widely distributed than Volvox and equally abundant. Pandorina thrived much better in the laboratory than Volvox, but neither lived more than a week or two. Most of the obser- vations were consequently made on specimens within a few days after they had been collected. The observations were all made in a large basement dark-room in which there was remarkably little variation in temperature throughout the season and practically none during the time occu- pied by any given experiment. The dark-room was so fitted up that either natural or artificial light could be used. Aside from ordinary electric lamps, there were two gas-filled stereopticon lamps, one 250 and the other 1,000 watt. These two lamps were mounted in an adjoining room from which light was admitted to the dark-room through a hole in the wall, which could be varied in size (fig. 4). Thus a horizontal beam of light of the desired size was produced. This extended through the dark-room parallel with the surface of a series of black tables 7 m. long The two lamps could readily be interchanged in position. By this means and by varying the distance between the organisms REVERSION IN ORIENTATION TO LIGHT 369 and the lamps it was possible to obtain quickly a wide range in illuminations. The beam of light, before it entered the dark- room, passed through a heat-screen consisting of a Pfeifer warm- ing-stage containing distilled water. This screen was so ar- ranged that a stream of water continuously flowed through it so as to prevent excessive heating. The observations were prac- tically all made in six rectangular glass aquaria which were des- ignated observation aquaria. These aquaria were 2.7 cm. wide, 2.7 cm. long and about 1 em. deep. They were constructed from pieces of the best-quality microscope slides and Khotinsky sealing-wax. Fig. 1 Diagram representing arrangement of apparatus. a, rectangular ob- servation aquaria; /, gas-filled steriopticon lamp; b, beam of light; w, wall be- tween dark-room and room containing lamp; h, heat-screen containing running distilled water or saturated solution of chlorophyl in 96 per cent alcohol; ¢, dead black table, 7 m. long; s, light-sereens. RELATION BETWEEN ILLUMINATION AND REVERSION IN ORIENTATION Both Volvox and Pandorina are very sensitive to light and they orient fairly precisely. Like a considerable number of other similar organisms, they have ordinarily been found to be posi- tive in weak and negative in strong light. This has often been observed in previous work (Mast, ’07, ’11, 718) and it was re- peatedly observed in the experiments performed in eonnection with this work. The following detailed description of a typical series of observations clearly illustrates the relation between this change in the sense of orientation and the intensity of illumination. On August 8, 4.40 p.m., cdlonies of Pandorina which had been in darkness for nearly four days were exposed in strong direct sunlight. At first they were very inactive and there was no indi- THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 3 370 . S. O. MAST cation whatever of orientation. But after approximately one minute it could be seen clearly that they were swimming slowly from the light. Gradually they became more and more active, and as they became more active they became more strongly negative, until at 4.50 p.m. they were swimming rapidly and fairly directly from the source of light. At 4.55 p.M., they were moved 4 m. from the window and placed in diffuse light. Here they promptly became strongly positive. They were then re- turned to direct sunlight near the window, when they again promptly became strongly negative. These changes in illumi- nation were repeated several times and the same results were obtained each time. It is thus evident that strong illumination tends to make these organisms negative and that weak. illumination tends to make them positive. Under certain conditions, however, just the opposite holds. In previous work this was observed only once (Mast, ’07, p. 165). In the observations now under considera- tion it was, however, repeatedly observed both in Volvox and in Pandorina, and the conditions under which it occurs have been discovered. This is demonstrated by the results obtained in the observations described below: On July 18, Volvox colonies were collected at 7.30 A.M., in a pool fully exposed to direct sunlight. They were taken to the laboratory, and put into two finger-bowls. One bowl was put into total darkness, the other was exposed in strong diffuse day- light. Thus one group of colonies remained light-adapted, while the other group became dark-adapted. At 3 p.m., light- adapted colonies and dark-adapted colonies were put, respectively, | into each of six rectangular observation aquaria. ‘These were then exposed in the dark-room in light from the 250-watt lamp, one aquarium containing dark- and one containing light-adapted colonies side by side in each of the following illuminations, 62.5 m.c., 250 m.c., and 4,000 m.c. (fig. 1). The dark-adapted colonies were strongly positive in all the illuminations and the light- adapted were negative in all, but they were clearly more strongly negative in the lowest intensity than they were in the highest. In 4,000 m.c. they were only slightly negative; in 250 m.c. REVERSION IN ORIENTATION TO LIGHT Bil they were clearly more strongly negative. One of the aquaria was then moved nearer to the lamp into an illumination of 16,000 m.c. The colonies immediately became positive, although not strongly. The dark-adapted colonies were not tested in this intensity, but judging from results obtained in numerous other experiments, one of which is described in detail in the preceding pages, they probably would have been negative. These results seem to indicate that long exposure in strong light produces changes in Volvox which make it negative in weak and positive in strong light. The following results ob- tained in observations on Pandorina support this conclusion. August 4 was a bright clear day. At 11.30 a.m., Pandorina was taken in abundance from a small puddle at the edge of the collecting pond. This puddle had been fully exposed to direct sunlight all forenoon. The specimens collected were taken to the laboratory and some of them were immediately exposed to direct sunlight in an observation aquarium, surrounded by water which was changed from time to time so as to prevent excessive rise in temperature. They were strongly positive and remained so. At 3 p.m. the aquarium was placed on the stage of a microscope in direct sunlight, and a beam of direct sunlight, concentrated by means of the concave mirror, was thrown up through the aquarium. The colonies promptly aggregated in this extremely intense illumination forming a dense mass. At 5 p.m. many colo- nies were lying on the bottom inactive, but all the rest were still fairly strongly positive. These results indicate clearly that under certain conditions Pandorina cannot be made negative by exposure to intense light. The specimens used in this experi- ment were unfortunately not tested in weaker light, but judging from the results of other observations they probably would have been negative. For example, at 3 p.m., August 11, colonies adapted to strong diffuse light were exposed in an illumination of 16,000 m.c. They were inactive for a few moments and then became definitely positive. The. illumination was then considerably reduced. The colonies continued to proceed toward the light, but only for a few momenis, when practically all of them turned through 180 degrees and proceeded from the light. oe Ss. O. MAST This was repeated many times and the same results were per- sistently obtained. Thus itisclear that, under certain conditions, Pandorina is definitely positive in strong and definitely negative in weak light. What are the conditions under which this occurs? It was repeatedly observed that colonies exposed in strong diffuse light, in which they were strongly positive, usually become negative in the evening at the approach of twilight. This sug- gests that the phenomena may be associated with the night and day variations in illumination and that it may be analogous to the so-called sleep movements in higher plants. The fact, how- ever, that it does not occur in dark-adapted colonies and in colo- nies which have been exposed to relatively weak light does not support this contention. The physiological condition in which the colonies are positive in strong and negative in weak light is, in all probability, de- pendent upon the amount of light energy received in the immedi- ate past as the following results indicate. At 8 a.m., August 23, dark-adapted specimens in observation aquaria were exposed in illuminations of 16,000 m.c., of 4,000 m.c. and in lower intensities. ‘Twenty minutes later, 8.20 A.M., a few of the small colonies in 16,000 m.c. were negative, the rest were all strongly positive. The aquarium was left in this illumination in the dark-room and observations were made every hour. The temperature remained practically constant through- out the day. At 9 and 10 a.m., the reactions were practically the same as they had been at 8.20 a.m. At 11 a.m. practically all of the small colonies and a few of the large ones were negative. More and more continued to become negative, until at 3 P.M., practically all were negative. At 4 p.m. a few of the colonies were scattered, all the rest were negative. From this time on gradually more and more became scattered. At 7 P.M. many of the colonies were scattered, some were negative and some were distinctly positive. At 8 p.m. there were more positive colonies, and at 9 p.m., when the experiment was closed, there was a large positive collection. In the aquarium in 4,000 m.c. the colonies were at first all positive. An hour later, 9 a.mM., a few were negative. From REVERSION IN ORIENTATION TO LIGHT ale this time on gradually more became negative, until at 6 P.M. nearly all were negative. At 9 p.m., when the experiment was closed, more of the colonies were scattered than earlier, but there were no positive colonies. In 62.5 and 250 m.c. many of the colonies became negative, but some of them remained positive throughout the experiment. In all intensities lower than 62.5 m.c., none of the colonies became negative at all. This experiment demonstrates that a certain intensity of light is necessary to induce reversion in the sense of orientation from positive to negative, that the time required depends upon the intensity, and that in strong illumination the colonies, after having become negative, become positive again if they are ex- posed long enough. Whether or not they continue positive in- definitely after the last reversion was not ascertained, nor was it ascertained in this experiment whether or not they would have been negative in low illumination after they had become positive in high. But in all probability these colonies would have re- mained positive in strong and would have been negative in weak illumination indefinitely, for negative orientation in weak light was never observed in dark-adapted colonies; it was observed only after long exposure to intense light, and continued exposure to such light failed to make the colonies negative after they had once become positive in this light. The changes that occur in the reactions of Pandorina as indi- cated by the results described above may be visualized by means of a graph presented in figure 2. This graph indicates that dark-adapted colonies of Pandorina when first exposed to light are neutral for a short time, i.e., they do not orient (fig. 2, a—b), that they then become positive, increasing rapidly to a maximum (b-c) then decreasing slowly to a minimum (c—d), after which they become negative, passing through a maximum at e,and that they finally become positive again (f-g). It is probably during this last period that the colonies are positive in strong and nega- tive in weak light. As previously stated, this reaction is found only in light- adapted colonies, never in dark-adapted ones. What, now, is the difference in different illuminations in the reaction of colonies in these two states? 374 Ss. O.. MAST Numerous observations were made to ascertain this. In all of these observations dark-adapted colonies were put into one rectangular aquarium and light-adapted ones into another. These two aquaria were then placed side by side at the desired distance from the source of light and the reactions of the colo- nies compared. Frequently three sets of such aquaria were exposed in different illuminations at the same time. The results obtained appear to be extremely contradictory. Usually the light-adapted colonies were found to be more strongly e€ 8.20 8.40 3.002 8.00 a a.m p.m p.m Fig. 2 Graph representing changes in the sense of orientation in dark-adapted Pandorina exposed to constant illumination of high intensity. The horizontal axis represents time, the vertical axis degree of orientation, above the base line positive, below the base line negative. The graph indicates that Pandorina was first neutral (a—b) then became positive, increasing to a maximum (b-c) then de- creasing to a minimum (c—d), after which it became negative, increasing to a maxi- mum (d-e), then decreasing to a minimum (e-f), after which it again became positive (f-g). After the colonies reach the last stage they are probably in a state such that they are positive in strong and negative in weak illumination. positive in strong and less strongly positive in weak light than the dark-adapted ones, but in many instances precisely the oppo- site was found to be true and in some there was no appreciable difference in the reaction. Moreover, in a given intensity, e.g., 4,000 m.c., the light-adapted were in some instances more strongly positive and in others less strongly positive than the dark- adapted colonies. How can these puzzling results be explained? Dark-adapted specimens, as we have seen, are frequently neu- tral when first exposed to light, then they become strongly posi- tive and later negative. If they are in the strongly positive stage when their reactions are compared with those of light- REVERSION IN ORIENTATION TO LIGHT 375s adapted specimens, they are likely to be more strongly positive than the light-adapted specimens; whereas if they are in the negative stage they will be found to be less strongly positive. Other contradictory features in these results can be similarly explained. The cause of the passage through the various stages mentioned above, especially the return to positive. orientation after having been negative, is not known. It may be associated with what is ordinarily called acclimatization or adjustment or with some sort of a periodicity. However this may be, reversion is un- doubtedly associated with physiological processes or states and these processes are dependent both upon time and intensity of illumination. This the evidence presented clearly indicates. How are these processes related to the amount of light energy received? RELATION BETWEEN REVERSION IN ORIENTATION AND THE QUANTITY OF LIGHT ENERGY The relation between reversion and the amount of light energy . received was ascertained as follows: Colonies of Pandorina adapted to darkness or to weak light were put into each of seven observation aquaria. These were then exposed in various intensities of light from the 1,000-watt lamp as indicated in table 1. Observations were then made, first at intervals of thirty minutes and later at intervals of one hour. After each observation the position of the colonies in each aquarium was recorded. From these records the approximate time at which reversion occurred in each intensity was ascertained. The re- sults thus obtained appear in table 1. When first exposed the colonies in all of the aquaria were neutral and relatively inactive. Gradually they became more and more active, and as they became more active they began to swim toward the side of the aquarium nearest the source of light, where in the course of several minutes practically all had col- lected in a dense mass. This occurred first in the aquarium in the highest and last in that in the lowest illumination. Later 376 Ss. O. MAST the colonies began to scatter, evidently becoming neutral again. Thus they remained for some time, then they became negative, and finally collected along the side of the aquarium farthest from the source of light as definitely as they had formerly col- lected on the opposite side. This again occurred first in the highest and last in the lowest illumination. In all of the aquaria the smaller colonies invariably became negative before the larger ones. Thus there was often found in TABLE 1 Relation between time of exposure, intensity of illumination and reversion in orientation ENERGY IN rennvarey or | me or nar | Tuahcrnas” [Tubes METER-CANDLES OCCURRED RIE ASGIN nee te REVERSION August 22. Specimens 16,000 -0 3.45 P.M 5 80 000 which had been in weak 4,000 .0 4.45 6 24,000 light several days 1,000.0 5.45 7 7,000 250.0 6.45 8 2,000 HO 6.45 8 888 hath 6.45 8 444 20.4 (?) u 160(?) August 23 Same speci- | 16,000.0 3 7 112,000 mens after having been 4,000 .0 4 8 32,000 in darkness overnight 1,000.0 Ul 10 10,000 250.0 8 11 2,750 IO 9(?) 13(?) 1,443(?) OF Ul 9 No reversion 20.0 9 No reversion aquaria a very definite positive collection and at the same time an equally definite negative collection. Later, however, prac- tically all of the colonies became negative, and the record in the table indicates the time when this occurred in each aquarium. The table mentioned contains the results obtained in two ex- periments made on two successive days with the same organisms. By referring to this table it will be seen at once that reversion from positive to negative orientation depends upon the intensity of illumination and upon the time of exposure, but that the energy REVERSION IN ORIENTATION TO LIGHT Oa. required to. induce reversion varies with the intensity of the illumination and with the condition of the colonies. In oné experiment it required in all intensities much more energy than it did in the same intensities in the other and in both experiments it required much more energy in the higher than in the lower intensities. In one experiment no reversion was observed in illu- minations below 250 m.c., and longer exposure probably would not have induced it, for at the close of the experiment, 9 P.M., the colonies were inactive. In the other experiment reversion was observed, to some extent, in the lowest illumination tested, 20.4 m.c. The question as to whether or not it occurs in all intensities that induce positive orientation is consequently not definitely settled. Why it requires more energy to induce reversion from positive to negative orientation in strong than it does in weak light is not clear. Itis, however, well known that in Euglena and Volvox heat and light energy have opposite effects on reversion (Mast, "11, p. 300). The same is true for Pandorina as we shall dem- onstrate later. Now, a certain amount of light which is absorbed is doubtless transformed into heat. This probably occurs in all illuminations in the same proportion, but in the lower illumina- tions it occurs so slowly that radiation may have relatively a much greater effect than in the higher. Consequently, the heat produced would be more effective in the higher illuminations than in the lower, and since heat tends to make the colonies positive, it would require more light energy to overcome its effect in the higher than in the lower illuminations. The value of this sug- gestion could doubtless be ascertained by studying the effect of different regions in the spectrum on reversion. RELATION BETWEEN REVERSION IN ORIENTATION AND THE TIME RATE OF CHANGE IN ILLUMINATION In the experiments just discussed, it required to induce rever- sion, sufficient time to indicate that it is dependent upon the quantity of energy received. Under certain circumstances, re- version is, however, of such a nature that it appears to be asso- 378 S. O. MAST ciated with the time-rate of change in illumination rather than with the quantity of light. For example, on August 8 colonies of Pandorina which had been in darkness four days were exposed to direct sunlight at 4.40 p.m. They were inactive for about one minute, then they began to swim, slowly at first and gradu- ally more and more actively, and in practically every case from the light. They were definitely negative. The observation aqua- rium was now moved 4 m. from the window. Here the colo- nies were strongly positive. This change in illumination was repeated several times with the same results. The aquarium was then exposed in direct sunlight near the window a few centi- meters from the diffuse light. The colonies soon began to swim rapidly from the lght. The microscope was then carefully moved into the diffuse light without changing the distance from the window. ‘The colonies immediately became strongly positive, but after they had proceeded toward the window about 1.5 e.m., they became neutral, and approximately one minute later they were strongly negative again. The aquarium was now re- turned to direct sunlight. The colonies remained negative. They were then again moved into diffuse light where they im- mediately again became strongly positive, than neutral and later negative. This change in illumination was repeated many times and the same results were always obtained. Now, the reversion from negative to positive orientation ob- served in this experiment, after sudden reduction of intensity, was in all probability dependent upon the time-rate of change in illumination, for without any further change the colonies, in the course of about one minute, again became negative. This seems clearly to indicate that if the reduction had consumed sufficient time there would have been no reversion. EFFECT OF THE AGE OF THE COLONIES ON REVERSION In the preceding section it was pointed out that in the experi- ments on the relation between the intensity of the illumination the smaller colonies of Pandorina invariably became negative before the larger ones did. No observations in reference to this REVERSION IN ORIENTATION TO LIGHT 379 were made on Volvox, but it was repeatedly observed in various cultures of Pandorina, especial'y in those left uncovered for some days, permitting evaporation. Whether or not the younger colonies in dark-adapted cultures exposed to light became active and positive before the older ones was not ascertained. EFFECT OF PHOTOSYNTHESIS ON REVERSION IN THE SENSE OF ORIENTATION It is well known that acids added to the culture solution tend to make Volvox and similar organisms positive in their reactions to light and that under certain conditions low illuminations also tend to make them positive, while high illuminations tend to make them negative. In low illuminations there is, owing to limited photosynthesis and continued respiration, a tendency toward an accumulation of carbon dioxid, result ng in an increase in acidity, while n high illumination there is, owing to rapid photosynthesis, a tendency toward a reduction in carbon dioxid. This seems to ndicate that the tendency toward positive orientation n low and negative orientation in high illumination may be dependent upon photosynthesis. If this is true, then red and yellow light in which photosynthesis is relatively strong should be more effective in producing reversion in the sense of orientation than blue and green in which photosynthesis is relatively weak. Numerous observations on both Volvox and Pandorina adapted to the colors mentioned were made as follows: Some of the colonies to be tested were put into jars in each of four black light-tight boxes containing a large window made of blue, green, yellow, and red glass, respectively. Others were put into jars in absolute darkness and still others into jars in strong diffuse sunlight. After having been in these conditions eight hours or longer specimens were taken from each jar and put respectively into six observation aquaria. These aquaria were then placed side by side in the same illumination in the dark-room, others similarly treated were placed in other illumina- tions. In this way the reactions in different illuminations of the 380 S. O. MAST colonies adapted to the various colors could readily and accu- rately be compared with each other and with those of the colonies adapted to darkness or to strong diffuse light. Without going into details regarding the results obtained, it may be said that in practically every test the red- or yellow- adapted colonies responded essentially like dark-adapted colonies and the blue- or green-adapted ones responded essentially like light-adapted colonies. ‘The red- and yellow-adapted colonies were usually negative in strong and positive in weak illumination, never the reverse; while the green- or blue-adapted colonies, like light-adapted ones, were frequently positive in strong and negative in weak illuminations.: The colors to which these colonies were adapted were not spectroscopicaily tested and the illumination was not measured. The question, consequently, arises as to whether or not, under the conditions of the experiments, photosynthesis in the red and the yellow was actually greater than in the blue and the green. This question was answered as follows: A given amount of pond-water taken from a vessel containing colonies equally distributed was put into each of six 100-cce. wide- mouthed bottles. One of these bottles was now placed in each of the four boxes mentioned above, i.e., in the red, the yellow, the green, and the blue light used in the preceding experiments; one was put into darkness and the remaining one into strong diffuse light. After having been in these illuminations one or more days a given amount of solution was removed from each bottle and put into a test-tube. A drop of neutral red was now added to the solution in each tube. This solution was found to be distinctly akaline in every case. Hydrochloric acid was then added to each tube until all were practically neutral and the same in color. The solution in the tube which required the greatest amount of hydrochloric acid was, of course, the most strongly alkaline, and in this solution photosynthesis had been most rapid, for carbon dioxid is consumed in the process of pho- tosynthesis and the alkalinity is dependent upon the amount of carbon dioxid present. REVERSION IN ORIENTATION TO LIGHT 381 There was considerable difference in the results obtained in the different tests, but taken as a whole they show conclusively that photosynthesis was most rapid in the diffuse white light and less rapid in the other illuminations in the following order: yellow, red, blue, green, darkness. These results, consequently, demonstrate that photosynthesis was more rapid in the red and the yellow light used in these experiments than in the blue and the green. And since the blue and the green were more effective in producing reversion in the sense of orientation than the red and the yellow, it is evident that the effect of light on reversion is not determined by photo- synthesis unless the photosynthesis which occurred during the exposure to white light in making the tests in the dark-room is involved. This, however, does not seem probable since all of the aquaria were exposed to the same illuminations. Moreover, essentially the same results were obtained in observations in which the aquaria were exposed to green light produced by means of passing the beam of light in the dark-room through a satu- rated solution of chlorophyl in 95 per cent alcohol. Now, since photosynthesis is reduced to a minimum in green light produced in this way and since the reactions of the colonies in the different aquaria were essentially the same as in white light, it 1s evident that photosynthesis during exposure in the dark-room is of no practical consequence. The conclusions, therefore, that rever- sion in the sense of orientation is not determined by photosynthe- sis appears to be valid. The region in the spectrum of maximum stimulating efficiency lies in the green near wave-length 524 wu for Pandorina (Mast, 717, p. 509) and in the blue-green near wave-length 494 yy for Volvox (Laurens, 18). The results of the experiments described above indicate that the regions of maximum efficiency in producing re- version in the sense of orientation in Pandorina and Volvox prob- ably have the same location as the regions of maximum stimu- lating efficiency. If this is true, it is probable, although by no means certain, that the processes involved in stimulation are also involved in reversion. 382 Ss. O. MAST EFFECT OF TEMPERATURE ON REVERSION In nearly all of the organisms that have been tested increase in temperature tends to induce positive and decrease in tempera- ture negative orientation to ight There are only a few forms in which changes in temperature appear to have the opposite effect, but there are a considerable number in which changes in temperature have no effect on the sense of orientation (Mast, a. pp: 272-279). The sense of orientation is probably not specifically related to the temperature in any of the forms studied. For example, the results obtained in observations on Euglena (Mast, ’11, pp. 274— 277) indicate. that this form may be, under certain conditions, negative or positive in practically all temperatures in which it orients at all. The effect of changes in temperature on the sense of orienta- tion in Volvox and Pandorina was ascertained as follows: The colonies were mounted on a Pfeifer warming-stage under a bin- ocular. This stage was so arranged that hot or cold water could be passed through it at any rate desired. The whole apparatus was then exposed to constant illumination of the desired intensity. The results obtained are in harmony with those obtained in the study of Euglena. The reactions of Volvox were, however, only superficially studied. Without going into details, it may be said that the results clearly show that arise in temperature tends to produce positive, and a fall in temperature negative photic ori- entations, and that the sense of orientation is not directly de- pendent upon the temperature. For example, in one experiment it was found that Pandorina continuously exposed in a given illumination was, at 10.43 a.M., negative in 16 degrees, neutral in 17 degrees, and positive in 18 degrees; at 11.19 a.m., the same colonies were negative in 13.5 degrees, neutral in 14 degrees, and positive in higher temperature, and at 11.28 a.m. they were positive in 14 degrees. Thus the point at which they became positive changed from 18*degrees at 10.43 a.m. to 14 degrees at 11.28 a.m. REVERSION IN ORIENTATION TO LIGHT 383 EFFECT OF CHEMICALS ON REVERSION It is well known that acids and some narcotics tend to make many organisms that orient to light, photopositive. Salts and alkalis, on the other hand, rarely have any effect on the sense of orientation. In a recent paper on Spondylomorum (718) it was fairly clearly demonstrated that the effect of the addition of acids is due to: the reduction in the alkalinity of the culture medium and not to the acids as such. The results obtained in experiments on Volvox and Pandorina support this contention and they show that the response of the organism is not specifically dependent upon the chemical constitution of the surrounding medium. , The effect of acid on the sense of orientation in Volvox and Pandorina is in all essentials precisely the same as it is in Spon- dylomorum (Mast, 718). If a trace of acid is added to a solution containing negative colonies they become strongly positive, re- main so a few moments and then become negative again. If more acid is now added they again become positive and later negative, just as they did after the first addition of acid. Thus they continue to become positive and negative after each addition of acid until the solution becomes fatal. The water in the pond in which the Volvex and the Pandorina used in these experiments appeared gave, in every instance, a very definite alkaline reaction with neutral red, and when acid was added the sense of orientation was reversed long before the alkalinity was neutralized. In fact, sufficient acid to give even the slightest acid reaction invariably proved fatal. This seems to indicate that reversion in these forms, just as in spondylo- morum, is due to a reduction in alkalinity, and not to the effect of acid as such. It also indicates that the sense of orientation is.not directly related with the concentration of the alkalis. The amount of reduction in alkalinity required to produce reversion varies with the concentration of the solution and the physiological state of the organisms. It is usually very small. _ For example, in one experiment titration against HCl showed . that the water in which the colonies (Pandorina) lived was 0.0019 384 Ss. O. MAST N alkaline. In this solution, under the conditions of the experi- ment, the colonies were strongly negative. After the addition of sufficient acid to make the colonies positive, the solution was 0.0015 N alkaline, and in a few other tests the reduction necessary was even less. In some it was, however, considerably more. This shows that the condition of the organism is involved inthe process. In Spondylomorum it was found that reduction in alkalinity produced by the addition of distilled water had little, if any, effect on the sense of orientation (Mast, 718, p. 512). Similar results were obtained in Volvox and Pandorina. Both of these forms live for several days in chemically pure water and respond nor- mally, but only in relatively few tests was there any indication whatever of reversion due to the dilution of pond water with pure water, and in these tests the effect of the dilution was very sight. But that there was actually an effect was shown by the fact that in diluted pond water it required considerably less acid to induce positive reactions than’ it did in normal pond water. For example, in a given test in a solution consisting of one part of pond water and nine parts of pure water, it required only one- third as much acid to induce reversion as it did in the pond water. The dilution consequently seems to tend to make the colonies positive, but not in proportion to the degree of dilution. If a reduction in alkalinity produces reversion from negative to positive orientation, it seems reasonable to expect the reverse if the alkalinity is increased by means of adding alkalis. This, however, does not appear to occur. I repeatedly added sodium hydrate to solutions containing positive colonies. of Volvox or Pandorina, but never obtained any indication of reversion except in case the alkali was added immediately after the colonies had been made positive by the addition of acids, and in such cases there was always some question as to the actual effect of the alkali. However, an increase in the concentration of the solu- tion due to slow evaporation clearly tends to make the colonies negative. These results are in full harmony with those obtained on Spondylomorum (Mast, 718, p. 512). REVERSION IN ORIENTATION TO LIGHT 385 Chloroform has the same effect on reversion that acids have, but ether and alcohol have very little, if any effect. With ether, no clear case of reversion was obtained at all, and with aleohol reversion occurred only in colonies that were almost neutral. The concentration of alcohol necessary to kill these colonies, especially Volvox, is very surprising. In equal parts of pond water and 96 per cent alcohol they live for several hours. The cause of the effect of anesthetics on reversion is not known, but it is certainly not dependent upon reduction in alkalinity, for the chloroform used was clearly slightly alkaline. This question will be discussed in the following section. The evidence presented thus far indicates that the sense of orientation is dependent upon the constitution of the culture medium. This contention is further supported by the fact that ordinarily if colonies are negative in one jar and positive in another in the same illumination, as often happens, the solu- tion in the former is more strongly alkaline than that in the latter. This is, however, by no means always true, and in some instances in which it was found to be true it was also found that the colonies retained their sense of orientation after they were interchanged. That is, the colonies which had been positive in the weaker solution were now positive in the stronger, and those which had been negative in the stronger were now negative in the weaker solution. This demonstrates conclusively that the state of the colonies may determine the sense of orientation. DISCUSSION It has been demonstrated in the preceding pages that decrease in illumination, decrease in alkalinity, increase in temperature, increase in anesthetics, and increase in the age of the colonies all tend to make them positive. It has also been demonstrated that light-adapted colonies are, under certain conditions, positive in strong and negative in weak light, and that reversion at times depends upon the time-rate of change in illumination. It has, moreover, been demonstrated that the region in the spectrum of maximum efficiency in producing reversion probably coin- cides with that for maximum stimulating efficiency. Reversion THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 3 386 S. O. MAST in the sense of orientation, consequently, can be induced by a number of different factors, and this seems to indicate, as con- cluded in the preceding paper in this series (18, p. 518) that it is due to some specific physiological change which can be pro- duced by any one of the factors referred to. The fact that the region of maximum efficiency for reversion probably coincides with that for stimulation indicates that the physiological phenomena associated with these two processes may be the same. Stimulation, however, is usually if not al- ways accompanied with an increase in permeability and a de- crease in electrical potential. Reversion, then, if our contention is correct, should also be accompanied with changes in permea- bility and all of the factors which produce positive orientation should induce changes in one direction while all those which produce negative orientation should induce changes in the oppo- site direction. As previously stated, decrease in illumination, decrease in alkalinity, increase in anesthetics, increase in tem- perature, and increase in age all tend to produce positive orienta- tion. What effect have these factors on permeability? Krabbe (96) found in stems of Helianthus annuus that a rise in temperature greatly increases permeability, and Rysselberghe found the same in experiments on Tradescantia and Spirogyra (Trondle, 710, p. 173). He maintains that the permeability for water, potassium nitrate, glycerine and urea increases slowly from 0 to 6 degrees, more rapidly from 6 to 20 degrees, and again more slowly from 20 to 30 degrees, and that it is eight times as rapid at 30 degrees as it is at zero. This shows clearly that increase in temperature produces increase in permeability. We have demonstrated, as stated above, that increase in temperature causés positive photic orientation. Therefore, if reversion is due to change in permeability, positive orientation must be associated with increase in permeability, and if this is true then all of the factors which produce positive orientation should cause increase in permeability. The work of Lillie (’09, p. 248) indicates that this is true for decrease in alkalinity; Tréndle maintains that moderate illumi- nation, such as favors positive orientation, causes increase in REVERSION IN ORIENTATION TO LIGHT 387 permeability, and that strong illumination, in which negative ori- entation is usually found, causes decrease in permeability; and according to McClendon (17, p. 141), a considerable number of investigators hold that anesthetics in low concentration stimu- late. The concentrations of anesthetics that produce reversion, therefore, probably cause increase in permeability, although Os- terhout (13) in his ingenious experiments did not discover any such effect. We thus have considerable evidence in favor of the idea that positive orientation is dependent upon increase in per- meability. However, if this is true, then permeability ought, under certain conditions, depend upon the time-rate of change in illumination; it ought to be greater in old than in young colo- nies and it ought, under certain conditions, be greater in intense than in moderate illumination. Moreover, all conditions which - cause increase in permeability, e.g., NaCl, ought to produce posi- tive orientation, while all those which cause decrease in permea- bility ought to produce negative orientation. With these ques- tions, concerning which there is at present no trustworthy evidence, I hope to deal in the following paper in this series. Some investigators maintain that in many of the unicellular and colonial forms orientation is due to a series of shock-reactions; others maintain that there is no definite relation between shock- reactions and orientation. However this may be, it is certain that Euglena and Gonium and probably all other similar organ- isms usually, if not always, respond with the shock-reaction to a sudden increase in illumination if they are negative and to a sudden decrease if they are positive. That is, the same reaction may be induced either by a sudden increase or by a sudden de- crease in illumination, depending upon whether the organisms sare negative or positive. If shock-reactions are due to increase in permeability then increase in permeability must be due, in negative specimens, to increase and, in positive specimens, to decrease in illumination. And if orientation is due to shock-reactions, then reversion in orientation must be due to the phenomena which produce this change in the cause of increase in permeability. In relatively high temperature or moderate illumination and in solutions rela- 388 . S. O. MAST tively weak in alkalinity or relatively strong in narcotics, in all of which orientation is positive and permeability relatively high, a sudden decrease in illumination must produce an increase in permeability, and in relatively low temperature or intense light and in solutions strong in alkalinity or anesthetics, all of which tend to produce negative orientation and low permeability, a sudden increase in illumination must produce increase in permea- bility. There is at present no evidence which bears on these problems. I have stated them in order to emphasize the fact, that whether orientation depends upon shock-reactions or not, any theory that accounts for reversion in orientation should account for the changes in the cause of the shock-reactions which accompany it. SUMMARY 1. The reactions to light in Volvox and Pandorina are prac- tically the same. Both forms orient fairly precisely and both may be either negative or positive. 2. Dark-adapted colonies are usually positive in weak and negative in strong illumination, never the opposite. Light- adapted colonies are sometimes positive in strong and negative in weak illumination. 3. If dark-adapted colonies are exposed to continuous illumi- nation they are neutral for a short time, then they become posi- ‘tive, passing through a maximum, after which they become neutral again, then they become negative, passing through a maxi- mum, after which they again become neutral and finally positive again. After they have reached this final state they remain posi- tive no matter how intense the light may be, and they probably are negative in weak light. 4. The time required to pass through these various stages de- pends upon the intensity of the light. The higher the illumina- tion, the shorter the time. Reversion is, therefore, dependent upon the time of exposure as well as upon the intensity of the illumination. But it requires much more energy to induce re- version in high that it does in low illumination. REVERSION IN ORIENTATION TO LIGHT 389 5. Under certain conditions, sudden decrease in illumination makes negative colonies momentarily positive. This change in the sense of orientation is dependent upon the time-rate of change in the intensity of the illumination. 6. Reversion is not primarily dependent upon photosynthesis. Red and yellow light in which photosynthesis is relatively strong have little effect on reversion, while green and blue, in which photosynthesis is relatively weak, are nearly as effective as white light. 7. The rays of light which have the greatest stimulating efficiency (green and blue) are the most potent in producing reversion in the sense of orientation. 8. Increase in temperature causes negative specimens to be- come positive and decrease causes the opposite, but neither the degree nor the extent of change in the temperature is specific in its effect. Under certain conditions, the colonies may be nega- tive or positive in practically all temperatures in which they orient at all. 9. Alkalis have little, if any effect on reversion. Acids and some anesthetics, especially chloroform, cause negative colonies to become strongly positive, but reversion is not specifically de- pendent upon the concentration of the chemicals. Colonies which are positive in a solution having a given chemical concen- tration may be negative in the same solution or even in a weaker solution. The effect of acids is probably due to the accompany- ing reduction in the alkalinity of the cultural solution. 10. The sense of orientation is dependent upon the physio- logical state of the colonies as well as upon the constitution of the culture medium. 11. The sense of orientation is dependent upon the age of the colonies. Young colonies are more likely to be negative than old ones. In a given solution the young specimens frequently collect at the side of the dish farthest from the light while the old ones collect at the opposite side. 12. Reversion in orientation is probably associated with changes in permeability, positive orientation being associated with an increase and negative orientation with a decrease in permeability. 390 Ss. O. MAST LITERATURE CITED Krasse, G. 1896 Uber den Einfluss der Temparatur auf die osmotischen Pro- zesse lebender Zellen. Jahrb. f. wiss. Bot., Bd. 29, 8. 441-498. Laurens, H, anp Hooxrr, H. D. 1918 The relative sensitivity of Volvox to spectral lights of equal radiant energy content. Anat. Rec., vol. 14, pp. 97-98. Litiif, R.S. 1909 On the connection between stimulation and changes in the permeabilty of the plasma membranes of the irritable elements. Sci., vol. 30, pp. 245-249. Mast, 8. O. 1907 Light reactions in lower organisms. II. Volvox. Jour. Comp. Neur. and Psy., vol. 17, pp. 99-180. 1911 Light and the behavior of organisms. New York, 386 pp. 1917 The relation between spectral color and stimulation in the lower organisms. Jour. Exp. Zodél., vol. 22, pp. 471-528. 1918 Effect of chemicals on reversion in orientation to light in the colonial form, Spondylomorum quarternarium. Jour. Exp. Zodl., vol. 26, pp. 503-520. McCuenpon, J. F. 1917 Physical chemistry of vital phenomena. Princeton, 240 pp. OstrEerHOUT, W. J. V. 1913 The effect of anesthetics upon permeability. Sci., vol. 37, pp. 111-112. RYSSELBERGHE, Fr. vAN 1902 Influence de la température sur la perméabilité du protoplasme vivant pour l’eau et les substances dissoutes. Ree. de l’Inst. bot., Univ. de Brux., T. 5, p. 207. TrONDLE, A. 1910 Der Einfluss des Lichtes auf die Permeabilitit der Plasma- haut. Jahrb. f. wiss. Bot., Bd. 48, S. 170-282. Numerous other references on this subject may be found in Washburn, Animal Mind, 917, pp. 200-214; Holmes, Studies in Animal Behavior, 1916, pp. 116-119, and Mast, Light and the behavior of organisms, 1911, pp 264-287. Resumido por el autor, Andrew Johnson Bigney. Los efectos de la adrenina sobre la emigracién del pigmento en los melandforos de la piel y en las células pigmentarias de la retina de la rana. El presente estudio se ha efectuado sobre Rana pipiens, com- parando los resultados obtenidos con los producidos en otras especies. El autor ha comprobado los efectos de la luz y la oscuri- dad ‘sobre la rana normal. Los efectos de la adrenina sobre el pigmento de las células de la retina fueron estudiados en ranas sometidas a la accién de la luz y en otras colocadas en la oscuridad durante varias horas. En las ranas sometidas a la accién de la luz, el pigmento aparece difundido, como sucede normalmente en estas condiciones, pero en las ranas que han permanecido algun tiempo en la oscuridad también aparecia difundido, demostrando esto que la adrenina produce la difusién de dicha substancia. Puesto que los efectos de la adrenina y la luz son los mismos no pueden ser estudiados en las ranas colocadas a la luz, pero en las que han permanecido cierto tiempo en la oscuridad, que normal- mente produce la contraccién del pigmento, la influencia de la adrenina se hace notar; aleanza el maximum de intensidad a los siete minutos y sus efectos no cesan hasta pasadas unas cinco horas. Después de comprobar los puntos mencionados, el autor emple6 diversas concentraciones y ha encontrado que con solu- ciones tan diluidas como la de 1: 5,000,000, la influencia de la adrenina sobre el pigmento puede comprobarse. El pigmento de los melandforos de la piel fué examinado bajo las mismas condi- ciones que el de la retina y se pudo comprobar que la adrenina produce su contraccién a la luz, en vez de provocar su difusi6n. Translation by Dr. José F. Nonidez, Columbia University. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 THE EFFECT OF ADRENIN ON THE PIGMENT MIGRATION IN THE MELANOPHORES OF THE SKIN AND IN THE PIGMENT CELLS OF THE RETINA OF THE FROG! ANDREW JOHNSON BIGNEY The purpose of this paper is to discuss the influence of adrenin on the pigment migration in the dermal melanophores and the retinal pigment cells of the frog. It seems to be fairly well es- tablished by the work of a number of investigators that adrenin does induce a migration of the pigment granules from the cell processes into the body of the cell of the dermal melanophores of fishes, amphibians and reptiles. This has been proved in the frog by Corona e Moroni (’98) and by Lieben (’06). According to Klett (08), the retinal pigment of the frog also migrates into the body of the cell under the influence of adrenin, but Fujita (11) states that this drug produces just the opposite effect on this pigment. In order to clear up this uncertainty and to de- termine the relation of the retinal migration to that of the skin, the present investigation was undertaken. The work was sug- gested by that of Redfield (17) and was done under the direction of Prof. G. H. Parker, to whom I make due acknowledgment for his many valuable suggestions. The frogs used in these experiments were almost entirely Rana pipiens Schreber, though a few of Rana clamitans Latreille were used, but there was no difference seen in the migration of the pigment in either the skin or the retina of these two species. The adrenin employed was that prepared by Parke, Davis & Co. and sold under the name of “Adrenalin Chloride in strength 1 to 1000.” 1 Contributions from the Zodlogical Laboratory of the Museum of Compara- tive Zodlogy at Harvard College. No. 314. 391 392 ANDREW JOHNSON BIGNEY In testing the effects of this drug on the melanophores of the skin, preliminary experiments were made on the effect of light and darkness on these cells. This was merely to confirm the work of previous investigators that in the light the pigment | expands and in the dark it contracts. This action takes place in an hour or so; but to be sure that there was a complete migra- tion, the frogs were kept either in the light or in the dark for six hours. . To determine the influence of adrenin on this pigment, two frogs were placed in strong, diffused daylight for six hours, after which 0.06 cc. of a solution of adrenin one part in a thousand was injected into the dorsal lymph spaces of each animal. The frogs were then kept a quarter of an hour in the light and killed and a small portion of the skin from the hip was removed and prepared for microscopic examination by fixing it in Perenyi’s fluid and mounting it, unstained, by the usual method. In both instances the melanophores were found to be strongly retracted, which was opposite to the state induced by light. Two more frogs were next treated in the same manner, but they were kept in the dark, and upon examining their skin the melanophores were found retracted, thus showing that the adre- nin produces no other effect in the dark, These results harmo- nize with the investigations of previous workers. To determine the strength of the adrenin necessary to produce these reactions, a number of frogs were kept in the light the usual time, then injected with the adrenin in concentrations, one part in ten thousand, one in fifty thousand, and one in five hundred thousand. The frogs were killed at the end of an hour. Those with one part in ten thousand had the pigment al- most completely retracted, while those with one in fifty thousand had slight retraction, and those with one in five hundred thousand showed no effect. The same concentrations were used on frogs kept in the dark and in all instances the melanophores remained retracted, as was to have been expected. To determine how long the influence of the adrenin lasted, the usual amount, 0.06 ec. of a solution one in a thousand was in- EFFECT OF ADRENIN ON PIGMENT IN FROG 393 jected into the lymph spaces of a number of frogs kept in the light and they were killed at varying intervals from seven min- utes to five hours. Complete retraction was found to last for about two hours; at three hours, moderate contraction was noted, while in four to five hours retraction was replaced by full expan- sion, thus showing that the influence of the adrenin had com- pletely passed off. The migration of the retinal pigment under the influence of light and darkness is well known to be outward from the cell in light and into the cell in the dark. To test the influence of adre- nin on these pigment cells, eight frogs that had been kept in the light for five hours were injected each with 0.06 cc. of adrenin solution one in a thousand and were killed in pairs, the first pair after fifteen minutes’ exposure to the drug, the second after thirty minutes, the third after forty-five minutes, and the fourth at the end of an hour. The eyes after removal were fixed in Perenyi’s fluid, cut into sections, and mounted unstained. In all instances the retinal pigment was found to be fully expanded. The experiment was then reversed, the frogs being kept in the dark five hours and subjected to the adrenin while still in the dark. They were also killed in pairs at quarter-hour intervals. To my surprise, the retinal pigment in every case was fully expanded. To avoid possible error, I repeated the experiment, with exactly the same results. Control frogs injected with phys- iological salt solution, instead of adrenin, were made to accom- pany the others. The retinal pigment in all of these remained ‘retracted as in the regular dark condition. It is therefore cer- tain that the adrenin acts upon the retinal pigment in the same way as light, as maintained by Fujita (11). It thus appears that the action of adrenin on the retinal pigment cells is the op- posite of what it is on the melanophores of the skin. My results, therefore, are opposed to those of Klett (08), who maintained that adrenin caused a retraction of the retinal pigment. His re- sults were obtained by an intra-ocular application and not by injecting it into the blood stream. He could get no results by the latter method. In all his experiments the frogs were in direct, strong light or in diffused light, but never in the dark, 394 ANDREW JOHNSON BIGNEY and the concentration of adrenin was strong enough to be poison- ous to the animals. It is, therefore, not surprising that he failed to observe the real action of the drug. Furthermore, since the adrenin and light act in the same way, the real action of the drug could not be noticed in the light condition. Had Klett carried on his tests with frogs kept in the dark instead of the light, I am convinced that his results would have agreed with Fujita’s and mine. To determine the concentration necessary to produce the mi- gration of the retinal pigment, frogs that had been kept in the dark were injected with adrenin in strengths varying from one part in a thousand to one in one hundred million. The frogs were subjected to the influence of the drug for one hour and then killed. At concentration one to one thousand there was com- plete expansion of the pigment, at one to ten thousand almost complete, at one to fifty thousand there was less expansion and less regularity in the condition of the preparation, and at one to one million or one to five million the results were not very uni- form. While the influence of the adrenin could still be detected at these dilutions, there was an irregularity which recalled the variation often noticed in normal frogs. The action of the drug on the retinal pigment is possible in even greater dilutions. This sensitiveness of the retina is in marked contrast with that of the skin, which is not nearly so responsive. ‘There seems to be no sharp ending in its influence either on the skin or the re- tina, but a gradual diminution. To determine how long the influence of the adrenin on the retinal pigment lasted, a number of frogs kept in the dark were injected with a solution of adrenin one in one thousand and the eyes were prepared at hour intervals from one to six. After an hour the pigment showed complete expansion; after two hours somewhat reduced expansion; in three to four hours still further reduction, and in five to six hours the retraction was complete, thus showing that the influence passes off in about four hours. It is clear from these experiments that adrenin causes a con- traction of the pigment in the dermal melanophores and an ex- pansion of that in the retinal cells—processes precisely the oppo- site of each other. | EFFECT OF ADRENIN ON PIGMENT IN FROG 395 Another interesting question presents itself. How does this drug act in causing the above results? Is it through the nerves or directly upon the cells by being carried in the blood? To answer these questions, frogs in which the optic nerve of one eye had been cut very close to the brain, the other optic nerve not having been disturbed, were injected with adrenin under the usual conditions. .In this experiment the retinal pigment was found to be expanded in both eyes, thus showing that the optic nerve is not concerned in this action, but that it is very probably due to the drug carried in the blood. Another experiment was performed in which the eyes of several frogs were removed and treated directly by injecting the adrenin into the eyeball of one side and physiological salt solution into the eyeball of the other side in each frog. Irregular results were obtained, the pigment sometimes being retracted, but most gen- erally expanded in all the eyes. These irregularities were evi- dent in the control frogs as in the experimental set. This led to the suspicion that the condition of the frogs was not satisfactory owing to the season of the year. The earlier part of this work was performed in the autumn and winter and the results were strikingly uniform and consistent, but the later part of it was done in the spring. It was, therefore, suspected that the advent of the breeding season had something to do with the results. It is not improbable that this irregular condition is dependent, as Fuchs (’06) has suggested, on an especially excited state of the animals whereby adrenin is secreted naturally and in considerable quantities by the frog itself. This suggestion, which seems plausible, is nevertheless purely hypothetical. It is intended to continue this part of the work with the view of a final determina- tion as to the real cause of this irregularity. 396 ANDREW JOHNSON BIGNEY BIBLIOGRAPHY Corona, A., E Moroni, A. 1898 Contributo allo studio dell’ estratto di cap- suli surrenal. La Riforma Medica, anno 14. (Cited from van Ryn- berk, 1906.) Fucus, R. F. 1906 Zur Physiologie der Pigmentzellen. Biol. Centralbl., Bd. 26, S. 863-879, 888-910. Fusita, H. 1911 Pigmentbewegung und Zapfenkontraktion im Dunkelauge des Frosches bei Einwirkung verschiedener Reize. Arch vergl. Ophthalm., Jahrg. 2, 8. 164-179, : Kuerr 1908 Zur Beeinflussung der phototropen Epithelreaktion in der Froschretina durch Adrenalin. Arch. Anat. Physiol., Jahrg. 1908, Physiol. Abt., Suppl. Bd., S. 213-218. Lizsen, S. 1906 Uber die Wirkung von Extrakten chromaffinen Gewebes (Adrenalin) auf die Pigmentzellen. Zentralbl. Physiol., Bd. 20, S. 108-117. REDFIELD, A. C. 1917 The coédrdination of the melanophore reactions of the horned toad. Proceed. Nat. Acad. Sci., vol. 3, pp. 204-205. vAN Rynperx, G. 1906 Uber den durch Chromatophoren bedingten Farben- wechsel der Tiere. Ergeb. Physiol., Jahrg. 5, 8. 347-571. Resumido por el autor, W. W. Swingle. Estudios sobre la relacién del iodo con la tiroides. I. Los efectos de la alimentacién i1odada sobre los renacuajos normales y tiroidectomizados. El presente estudio trata del problema entre la relaci6n del iodo y sus compuestos con la actividad y funcién de la tiroides, deter- minada por los efectos que siguen a la ingestién de estas substan- cias por las larvas de rana, normales y desprovistas de tiroides. Cuando se alimentan renacuajos normales con cristales de iodo, los cambios propios de la metamorfosis aparecen pocos dias después y dicho proceso se lleva a cabo en corto tiempo si se alimentan con algun cuidado. La ingestioén del iodoformo pro- duce resultados semejantes pero sus efectos no son tan rapidos; con el ioduro potdsico se obtienen los mismos resultados. El iodato potdsico no parece producir efecto alguno sobre la meta- morfosis. En animales privados de las glandulas tiroides cuando median 4 mm. de longitud, alimentados después con iodo, la metamorfosis se llev6é a cabo rapidamente, a pesar de que en las larvas desprovistas de tiroides y alimentadas con los alimentos adecuados nunca aparecen cambios metamorficos, sino que, por el contrario, crecen hasta convertirse en renacuajos gigantes. El examen microscopico de estas larvas no pudo revelar indicacién alguna de tejido tirofdeo. La comparaci6n entre la rapidez de la metamorfosis en las larvas alimentadas con iodo y las alimen- tadas con extracto de la tiroides o con tejido tiroideo demuestra que el iodo es mas potente en la producci6én de los cambios, meta- morficos. El autor consigna experimentos describiendo el efecto del iodo sobre el canal alimenticio y otros érganos y hace notar que el iodo acttia dentro-de los tejidos como un verdadero hor- mon, sin intermedio de la glandula tiroides, que funciona princi- palmente como un reservorio del iodo. Translation by Dr. José F. Nonidez, Columbia University. AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 23 STUDIES ON THE RELATION OF IODIN TO THE THYROID! I. THE EFFECTS OF FEEDING IODIN TO NORMAL AND THYROIDECTOMIZED TADPOLES W. W. SWINGLE Fellow in Biology, Princeton University CONTENTS PET OCUIC ELON SEE eeE wae ish sk eats oe content ch il ssregtaya SI pe atS or 397 JE TRIGTER AONE ot OE SSIS OS 8 GO RRP VOS DIOR EG DIS Gre ro Rian Cicer ean 398 Meatentalean darn eG hod sperreprinc. oor. Srvc rias ss. cS ARES aye Aelchs cory tvetetoen eras 400 Observations (Feeding of iodin compounds to normal larvae).............. 401 (Pe Otassimmplodidery peer ys. s Loreen ees a ees NAGS Heer ea Gp arST sk 20-9 401 Z2eelOCINECEY Stal Sos reese ae oes Ys Mae Fis haere LR SO 405 SLOG OLOTIA Metre oe ee I nisrer ae bard etene wievas Se arnt eae 409 Heedinesiodinetosthyro1dectomized larvae menances oc satis. 7 eee 411 Summanye nde con customer eta ys 6 Acres sate a cia on otig aS Se + oy more ete ae 414 INTRODUCTION Early in the spring, while carrying on a series of feeding ex- periments with tadpoles, the writer suggested to Mr. A. C. EKitzen, a student in the Medical School of the University of Kansas, that he undertake the problem of feeding inorganic iodin, and various organic compounds of this substance to frog larvae. The object was to note the macroscopic effect upon growth and metamorphosis and the histological effects on the thyroid and germ glands. Most of the investigators who have dealt wth the problem of the relation of iodin and its compounds to the activity of the thyroid have used mammals as experimental material, forms which, in the author’s opinion, offer no such sure and certain ! The experimental work for this paper was done while the writer was in- structor in zoology at the University of Kansas. Acknowledgment is made to this laboratory. 397 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 3 398 - W. W. SWINGLE criteria for judging the effect of iodin administration and thyroid activity upon the organism as do frog larvae. The original work of Gudernatsch (since followed by that of Morse, Lenhart, and Swingle) showed that any increased stimulation of thyroid activity, as for instance by feeding thyroid extract, was at once indicated by metamorphic changes in the tadpoles. The present problem was to utilize these somatic changes as a e¢ri- terion of hyperthyroid function and to gauge the effects of iodin upon the gland by feeding it to immature larvae. Mr. Eitzen followed the suggestion, and, in conjunction with the writer, started the work, but owing to induction into military service shortly afterward gave up the problem. Consequently if fell to the writer to work out his own suggestion. The results obtained contribute some additional information regarding the relation of iodin to the physiological activity of the thyroid, to the probable role of this gland in the economy of the organism, and to the more general problem of amphibian metamorphosis. LITERATURE The active principle of the thyroid gland is at present un- known, or at any rate much in dispute. The relation of iodin to this gland has been for many years a matter of great interest to clinicians and investigators, and it has long been recognized that iodin holds an important place among the constituents of the thyroid. Since 1820 iodin has been used more or less in the treatment of thyroid diseases, but it was not until 1895 that it became known that the glands usually contain iodin. Baumann (95) contributed this important information to our knowledge of the gland. This investigator isolated a specific iodin compound from the gland which he named iodothyrin. It apparently has many of the specific properties of the gland tissue itself. Oswald (’02) sueceeded in isolating the protein with which the iodin in the gland is combined. He called this protein thyro- globulin; it was found to constitute one-third to one-half of the | weight of the dry gland. The iodin content varied from zero to 0.86 per cent or more. Baumann’s iodothyrin could be obtained from it by hydrolysis. RELATION OF IODIN TO THYROID 399 Kendall (14) has reported the separation of the physiologi- cally important thyroid constituents into two fractions: an A or alpha fraction, containing ten times the percentage of iodin of the original thyroid. This alpha fraction is toxie and its ad- ministration produces typical symptoms of excess thyroid feeding. The B fraction contains less iodin and is non-toxic. The relation of: iodin to the physiological activity of the thyroid is very close; in fact, the activity of the gland appears to depend upon its iodin content. One investigator, Roos (99), working with dogs, found that more nitrogen was excreted after the administration of thyroid ' rich in iodin than after that of thyroid containing little iodin. Marine and Williams (08) observed that thyroid containing a larger percentage of iodin caused a greater loss of weight in dogs than did a preparation containing a smaller percentage. Hunt and Hunt and Seidell (07 and ’08), in an extended series of experiments in which the effects of thyroid upon the resistance of animals to certain poisons was determined, found that the physiological activity of the thyroid depended upon its iodin content. Morse (’14) fed the larvae of Rana pipiens on various iodin compounds, but obtained negative results with all inorganic iodin. He was able to accelerate metamorphosis in these animals by feeding iodized amino-acids (3-5-di-iodo-thyrosin, C, H;sOHI,CH..CH,.CHNH:COOH). This amino-acid was de- rived from the thyroid tissue by acid hydrolysis. Lenhart (15) fed thyroid tissue to tadpoles and observed that the higher the iodin content of the gland fed, the more rapid the body metabolism. This review is by no means exhaustive. Besides the investi- gators quoted, there are equally as many, if not more, who claim that iodin has no relation to thyroid activity and that it does not function within the organism. 400 WwW. W. SWINGLE MATERIAL AND METHODS Late in March (1918) several bunches of Rana pipiens eggs in late segmentation stages were collected and brought to the lab- oratory to develop. All of the animals used in a single experi- ment were taken from cultures, the larvae of which came orig- inally from the same bunch of eggs, hence were of the same age. The animals were separated into lots of fifty each when the free feeding stage was reached and kept in large glass con- tainers under identical environmental conditions. Several groups of fifty tadpoles each were fed inorganic iodin crystals finely ground, mixed with wheat flour in the proportions of 1 to 100, using the same procedure described by the writer in a previous paper (18) for administration of thyroid extract. Another lot of larvae was fed iodoform mixed with flour as described. Two other cultures were fed potassium iodide, also in flour, using the same proportions as before. Solutions of potassium iodide were prepared and an attempt made to rear the tadpoles in them, but the method proved unsatisfactory and so was abandoned. The larvae were fed each day, care being taken not to overfeed ; the water was changed daily, and with the advent of warm weather twice daily. Very early in the work it was observed that unless the iodin was mixed with food in some way, the larvae refused to eat it. This was especially true of the iodin crystals. Great mortality results if the crystals are merely thrown into the containers among the tadpoles. The quickest and most effective way of rendering the iodin palatable was to mix finely ground crystals with wheat flour (1 to 100) stir until the flour was a delicate brownish hue and then feed the dry mixture. Small bits of algae were fed the animals along with the iodin. The various mixtures of iodin and flour prepared by the method described by the writer in a previous paper (718) for thyroid administration keep well, with the exception of the iodin crystals and flour. The mixture appears to lose strength after about two weeks. Regarding the chemical nature of the sub- stances formed by the mixture of iodin and its compounds with RELATION OF IODIN TO THYROID 401 flour, no mention will be made here. This phase of the work together with certain histological data gathered by examination of normal and thyroidectomized iodine-fed larvae, will be the subject-matter of part II of this paper. The results obtained with feeding normal larvae potassium iodine will be discussed first. OBSERVATIONS I. Experiments with feeding potassium vodide to normal tadpoles The tadpoles were first fed potassium iodide and flour when they averaged 10 mm. in length; none of the animals showed indications of limb buds. April 11th, eight days after the first administation of potassium iodide, the animals were measured and examined with a microscope for limb buds. At this time some of the larvae appeared somewhat emaciated and about one-fourth of the animals had hind limb buds, though the buds were small. The controls were somewhat larger than the experimental larvae, but no limb buds were found. ‘Table 1 gives the total length of twenty of the larvae from each culture, in millimeters and indicates those animals with limb buds. April 18th. When examined on this date the potassium- iodide-fed animals were lighter in color than their controls but of about the same size. All of the experimental animals had hind limb buds whereas none of the controls had developed them (table 2). The emaciated appearance of the larvae, noted at the previous examination, had disappeared. April 26th. The potassium-iodide-fed larvae had by this time outgrown their controls and were considerably lighter in color. All of the experimental animals had well-developed hind limbs showing differentiation into the two primary divisions with toe points. Only three of the controls had limb buds, none showed any differentiation of parts. The difference between the two cultures of larvae in regard to limb development was striking (table 3). May 6th. The animals fed on the iodine compound were now distinctly larger than the algae-fed larvae. The light color of Ahese animals was marked. Their limbs were well developed— 402 W. W. SWINGLE TABLE 1 April 11 KI+FrED LARVAE CONTROLS FOR KI-FED LARVAE Length Limb buds Length Limb buds mm. mm. 16.0 ae 15.0 ~ 15.5 ar 16.5 -- e345) AF 15) 05) — 14.5 = 17.0 — 15.0 =F 18.5 _ 14.0 aE 17.0 — 16.0 = GES _ 14.0 + 15.0 _— TSO AF 19.5 — 14.0 = 18.5 _ ee = 20 .0 _ 14.0 = 19.0 — 13.0 = 18.5 — 113), = 16.0 _ 16.0 = 20.0 _ 14.5 = 18.5 — 14.0 + 1145) 95) — 15.5 = 17.0 — 13.0 = 19.5 —_ 14.5 + 15.5 _ Average ; iW) length. .14.5 much in advance of the controls (table 4). No mortality had occurred in the cultures of either group. May 18th. The experimental animals were much larger than the controls, more sluggish, lighter in color, limbs much more highly developed and much larger. No other differences were noted (table 5). The animals of this culture were not measured again, owing to the pressure of other work at this time. They were fed and tended as usual, however. It was observed sometime later that the potassium-iodide-fed animals appeared to be slowly falling behind in growth rate. They remained much lighter in color and had limbs much longer than the controls. The cul- tures were kept until the 1st of June. When it became impos- RELATION OF IODIN TO THYROID 403 TABLE 2 April 18 KI-FED LARVAE CONTROLS FOR KI-FED LARVAE Length Limbs Length Limbs mm. mm. 21.0 + 23.0 = 18.5 AF 18.5 _ 17.0 =F 22.0 =- 20.0, + 21.5 _ PAY 5) ar 20.0 -- 19.5 ar 22.0 — 18.0 ae 19.0 — 21.0 + 20.5 — 2220 + PAP) 5 _ 1 ORS =e 18.5 — 20.0 + 19.0 = 23.0 =F 21.0 _ 21.5 =F 19.5 _ g/d) se 23 .0 _ 22.0 =e 18.0 — ZORO + 24.0 — 19.0 ar 20.5 — 23.5 + 22.0 — 22.0 + PAB) 5) — 20.5 + 23.0 — 20.37 21.07 sible to prolong the experiment, the animals were killed and preserved for microscopic examination. Discussion of potassium vodide feeding experiment. The results of this experiment, while not so interesting perhaps as those obtained by administration of iodin crystals, seemed worth recording in detail. The table of measurements shows that the erowth capacities of the larvae receiving potassium iodide were much increased. It is interesting to note in this connection that Adler (13), in a brief paper dealing with the effects of iodin upon the germ glands of amphibians and mammals, observed that iodin compounds appeared to stimulate the growth rate of the amphibian larvae with which he worked. He records no measurements, however, and made no observations regarding metamorphic changes. 404 W. W. SWINGLE TABLE 3 April 26 KI-FED LARVAE CONTROLS FoR KI-FED LARVAE Length Limbs Length Limbs mm. mm. 28.5 + 26.5 27 .0 -- 28 .0 PLAS, + 26.0 _ 30.0 ain Pia) _ 29.5 + 25)..5 — 28 .0 + 30.0 _— SD + PAL Go) + 29.5 + 26 .0 + 3220 + 23.0 -- 32) + 26.5 — 33.0 ae 29.0 + 30.0 =e 21.5 _ 33.0 + 24.5 + 32.5 + 26.0 _ 29.5 + 2.0 — 31.0 + 31.0 _ 29.0 + 25.5 _ 33.49 + 25.0 = 30.5 ++ 27.0 _ 32.0 + 28.5 — 30.5 Par AVS Potassium iodide in flour clearly stimulates the growth and differentiation of limbs. Morse (14) was unable to produce metamorphic change by administration of iodin compounds to tadpoles, with the exception of iodized blood albumin. In the light of the results obtained by the writer with iodin crystals, it is difficult to understand why this investigator failed to get positive results. His method of feeding the iodin probably accounts for the discrepancy in our results. Why feeding potassium iodide stimulated the growth of tadpoles is not clear. Feeding other iodin compounds gives just the reverse results, 1.e., growth ceases. RELATION OF IODIN TO THYROID 405 TABLE 4 May 6 KI-rED LARVAE CONTROLS FOR KI-FED LARVAE Length Limbs Length Limbs ee a a mm. 34.0 + 28.5 a Stn) + 26.0 _ 30.0 + 29.5 4 29.5 + ZOEO — . 32.9 + 27 .0 -- 36.0 == 30.0 ae 39.0- a Dies = 32hON + 29.0 ae 34.0 a 31.5 er 33.5 + 28.0 sit 41.0 + 28.5 a 38.0 +° 30.0 a 36.5 + 29.0 + 44.0 ae 31.5 ar 40.0 + PERS — Bao) + 30.0 + 36.0 af 26.5 = 39.0 + 30.0 + 41.5 =F 31.0 + 38 .0 + 28.5 = 36.2 28.7 2. Experiments with feeding iodin crystals to normal larvae This experiment was conducted in the: same manner as de- scribed for the potassium iodide experiment. The larvae averaged 10.5 mm. in length when the iodin was first fed, April 4th. None of the animals of either control or experimental cultures showed any indications of limb development. Seven days after the first feeding, the animals fed on the iodin had limb buds ‘and showed other bodily changes not.found in the controls (table 6). The larvae appeared emaciated; head much elongated (this change is only apparent and has been shown by the writer to be due to atrophy of the coiled gut); body thin; eyes bulging; pigmentation light. One or two of the animals showed signs of tail involution. All of the larvae were extremely sluggish and 406 WwW. W. SWINGLE TABLE 5 May 18 KI-FED LARVAE - CONTROL OF KI-FED LARVAE Length Limbs Length Limbs mm. mm. , 44.0 Large 32.5 Very small 46.0 Large 31.0 Very small 45.5 Large 34.0 Very small 44.0 Large 31.0 Very small . 39.5 Large 34.0 Very small 46.0 Large 33.5 Very small 46.5 Large \ 29.5 Very small 44.0 Large 35.0 Very small 47 .0 Large 305 Very small 38.5 Large 30.0 Very small 45.0 Large 33.5 Very small 42.5 Large 31.0 Very small 40.0 Large 34.0 Very small 43.5 Large 33.5 Very small 39.0 Large 30.0 Very small 41.5 Large 33.0 - Very small 42.0 Large 30.5 Very small 38.0 Large 29.5 Very small 40.5 Large 34.0 Very small 44.0 Large 30.5 Very small 42.3 32.05 remained near the surface of the containers. The hind limbs were small, and as yet they had not differentiated into their two primary divisions with toe points. These animals revealed clearly all of the symptoms of hyper- thyroidism, the reaction to which is characteristic in this species. Examination of the controls showed none of . these changes. None of the animals had limb buds. April 17. On this date the differences between iodin-fed and control larvae were marked. All of the body changes noted on April 11th were now much more obvious. ‘The tails of the iodin- fed animals were undergoing atrophy; the hind limbs were plainly visible and had differentiated toes. The total length of the ani- mals had decreased owing to tail resorption (table 7). There was great mortality among the larvae of the iodin culture. Out of RELATION OF IODIN TO THYROID 407 TABLE 6 April 11 IODIN-FED LARVAE CONTROL FOR IODIN-FED Length Limbs Length Limbs mm. mm. 11.0 a Way 0) — 12.0 a5 16.5 _— TL ay ae 15.0 — 10.5 ar 18.5 — 12.0 ar iio — 11.0 =F 16.0 _ 12.5 a= 20.0 _ 13.0 ap 18.0 — 11.0 SF 15.5 — 15 ar 19.5 -- They Rt5) ar 17.0 - 10.0 =F 1725 — 12.0 ar 20.0 _ 10.5 se 19.0 LO AF 15.0 — 12.5 se M5 _ 10.0 ar 17.0 — 1305 =P 19.0 _ 11.0 AF 21.0 _ 12.0 ar 18.5 =— 11.6 + 17 6 -- several cultures containing fifty larvae each, only thirty animals remained alive on this date. The controls had increased in size during this interval, but none showed signs of limb development. A new series of cultures of iodin-fed animals was started. The animals averaged 13 mm. in length when first fed iodin. Eight days later the first indications of body change appeared, and ten days from the date of first feeding the limb buds of the experi- mental larvae averaged 0.8 mm. in length. The controls showed no ‘signs of limb development. At this time the animals were taken off the iodin diet and fed alge in order to prolong the ex- periment, as the mortality rate Was abnormally high. The con- trol animals at this time averaged 19 mm. in length; the iodin-fed larvae, 14.5 mm. For several days both sets of tadpoles were fed only algae and then measured at the end of the period. The 408 W. W. SWINGLE TABLE 7 : April 17 IODIN-FED LARVAE CONTROL OF IODIN-FED Length Limbs Length Limbs mm, mm. mm. 12.0 1G) 20.5 None 10.0 1.0 24.0 None 9.0 1.4 PALA) None 9.5 0.5 235 None 9.0 1.0 22.0 None 10.5 ils 24.5 None 10.0 1.0 20.0 None 25 20) 19.0 None 13.0 1S PAW) None 10.5 0.5 Pia) None 9.0 ORS 20.0 None 9.5 145 20.5 None 10.0 1.0 21.0 None 10.5 16) 24.5 None 9.0 1.0 PP) {0 None WL 5} 2.0 20.0 None 1220) 5 19.5 None 125 1.0 21.0 None 9.0 0.5 2320 None 9.0 1.0 20.5 None 10.4 We PAL controls averaged 23.5 mm.; the iodin fed, 16.5 mm. The latter had well-developed hind limbs. JIodin was again administered for eight days and the animals measured. Table 8 indicates the differences between the two sets of larvae. The fore limbs of the iodin-fed larvae appeared shortly after- wards. Six of these animals were reared to metamorphosis before death occurred. One of the lot was the smallest frog the writer has ever seen. The controls for this lot developed nor- mally, but have not at the present writing undergone metamor- phosis. The fore limbs have not yet appeared. In a previous paper the writer had found that when thyroid extract is adminstered to frog larvae, the alimentary tract under- goes a remarkable shortening process in a brief space of time. RELATION OF IODIN TO THYROID 409 TABLE 8 IODIN-FED LARVAE CONTROL OF IODIN-FED Length Limbs Length Limbs mm. mm. mm. mm. L748) 6.0 27 .0 0.3 16.0 4.5 28.5 0.5 15.5 5.0 30.0 0.5 15.0 6.5 20.5 0.6 16.5 6.0 23.0 0.4 1Gso; 2 5 4.0 29.5 0.2 14.5 5.5 32.0 0.3 18.0 7.0 26.5 0.6 17.0 6.0 28.0 0.4 16.5 3.5 28.5 0.3 15.0 4.0 31.0 0.5 16.5 9.9 29.5 0.6 15.0 5.0 26.0 0.2 14.5 4.5 28 .0 0.3 14.0 @0 30.0 0.5 18.5 5.0 33.0 0.6 17.0 4.0 205 0.4 17.5 8.0 29.0 0.2 14.0 7.0 30.5 0.3 16.0 4.0 26.5 0.4 15.0 5.4 28 .6 0.405 An attempt was made to see if feeding iodin had the same effect upon the gut. A series of larvae appropriately controlled was started upon the iodin diet to test this matter. The animals were fed iodin for twenty days. Table 9 shows the effect upon the gut at the end of this interval. 3. Experiments with feeding iodoform to normal larvae An experiment similar in all respects to the potassium iodide and iodin feeding was carried out in which iodoform and flour was used as food. The proportions of flour and iodoform used was the same as described for the other experiments. Iodoform appears to have a marked effect in accelerating metamorphic changes in frog larvae, although the reaction is not so rapid as when iodin crystals are used. The iodoform is toxic for the 410 W. W. SWINGLE TABLE 9 IODIN-FED LARVAE CONTROL FOR IODIN-FED LARVAE Body length Length of gut Body length Length of gut mm. mm. mm. mm. ZORS 33.0 25 .0 107.0 20.0 35.0 23.0 102.0 19.5 ay 10) ZOx0 121.0 21.0 36.0 26.0 LO 19.5 38.0 2525 111.0 20.0 32.0 27.0 119.0 20.0 31.0 2029 97 .0 16.5 36.0 24.5 123°:0 2225 24.0 30.0 99 .0 19.0 28 .0 27.0 116.0 20.0 31.0 25 .0 102.0 Ia Ss) 0) 28.5 97.0 20.0 34.0 26 .0 114.0 18.0 29 .0 23.9 118.0 PND 24.0 28 .0 99.0 18.5 27.0 22.0 105.0 20.0 30.0 24.5 122.0 18.5 25.0 29 .0 96.0 19.0 28.0 2529 98 .0 WE 30.0 26.0 110.0 19.45 30.9 25 .85 108 .6 animals and the mortality rate is very high when this compound is fed for any length of time. Several cultures were kept long enough to show the accelerating effect of this substance upon metamorphosis. One culture, the animals of which averaged 12.5 mm. when first fed the iodoform, was kept for ten days. During this interval the animals ceased to grow. and showed the reaction characteristic of iodin or thyroid feeding. . Limb buds were found on all of the experimental animals, but none on the controls. This phase of the work was not carried out in detail because of the toxicity of the iodoform mixture. . Discussion of the experiments with feeding wodin to normal larvae. The results obtained by feeding iodin crystals and iodoform are identical with those obtained when extract of the thyroid gland is used as food. The changes typical of hyper- RELATION OF IODIN TO THYROID 411 thyroidism appear a few days after administration of the sub- stance. Overfeeding with iodin has the same result as over- feeding with thyroid extract—death of the organisms from a too rapid rate of metabolism. Jodin, if carefully administered, will stimulate metamorphosis in a shorter time than the fresh gland tissue. ‘Two years ago the writer fed fresh thyroid gland to frog larvae and kept a record of the rate of body change and the time required for such changes to occur when the larvae were fed fresh tissue and the powdered extract. A comparison of these time intervals with those of the iodin-fed animals of the present experiment show clearly that the fresh gland tissue is not so effective in inducing metamorphic change as the inorganic iodin. The animals used in both cases were of the same species. EXPERIMENTS WITH FEEDING THYROIDECTOMIZED LARVAE IODIN Since normal larvae were found to react to iodin feeding by marked metamorphic changes, it was considered worth while to carry out the same experiment upon animals whose thyroids had been removed during early embryonic life. The work of Allen (18) has shown that thyroidectomized larvae fail to undergo metamorphosis, but instead permanently retain their larval characters. If such thyroidless animals could be induced to metamorphose under the stimulus of iodin feeding, new light would be shed upon the iodin-thyroid problem and upon the causes of amphibian metamorphosis. Kighty very young toad tadpoles, the thyroid glands i which had been removed at the 4-mm. stage by Prof. B. M. Allen, of the University of Kansas, were obtained through the generosity of this investigator when the larvae averaged 6.5 mm. in length. The animals were fed agae until they were 10 mm. long and then fed iodin crystals and flour. This series was controlled by both normal and other thyroidectomized larvae of the same age. The iodin mixture was first administered the 27th of May. Ten days later, examination of the culture revealed well-developed limbs on all of the thyroidless larvae. The limbs were visible without the aid of a lens and had differentiated toe points. The 412 W. W. SWINGLE animals appeared emaciated and showed other symptoms charac- teristic of hyperthyroidism. The controls had increased in size somewhat, and microscopic examination showed tiny limb buds. The iodin feeding was continued until the time of metamor- phosis, which took place in a normal manner and was success- fully completed by all with the exception of nine larvae which died during early metamorphic change. The thyroidless animals under the stimulus of the iodin completed metamorphosis in a much shorter time than the normal control animals with thyroid glands intact. The thyroidectomized animals used as controls for the series are at the present writing very large and show no signs of limb buds. About this time twenty extremely large thyroidless - toad larvae were obtained from Miss Mary Larson, of the University of Kansas. The glands of these animals had been removed several weeks before; they were much larger than normal toad tadpoles at metamorphosis; they were, in fact, typical giant thyroidectomized larvae like those described by Allen. These animals showed no indications of limb buds when started on the iodin diet, but are at the present writing (June 12th) undergoing metamorphosis. The fore and hind legs of the animals are well developed; mouth is changing from larval to adult form; tails almost completely resorbed. Four of these giant thyroidless larvae used as controls for the iodin-fed culture are larger than at the beginning of the experi- ment, but show no signs of limb development. A number of both large and small thyroidectomized tadpoles were killed at different stages of the experiment and preserved for microscopic examination. Careful search made for vestiges of the thyroid gland have yielded only negative results. No indications of accessory thyroid glands have been found in any of the thyroidectomized animals examined by me in connection with this work. An experiment to test the ability of thyroidectomized larvae to withstand the deleterious effects of iodin in different con- centrations was attempted. Jodin is only very slightly soluble in water, but if the crystals are finely ground, a certain pro- RELATION OF IODIN TO THYROID 413 portion is dissolved. It was soon found that both normal and thyroidless animals are quickly killed by very weak dilutions. Two ce. of such a solution of iodin in 500 cc. of water suffices to kill both kinds of animals in a few hours. No conclusions could be drawn from the results, save, perhaps, that flour when mixed with iodin must in some way protect the tissues of the larvae from the latter. Discussion of effects of feeding iodin to ihyroidectomized larvae. The results of iodin feeding Just recorded should in some measure serve to clarify the conflicting views regarding the relation of iodin to the physiological activity of the thyroid and inciden- tally throw some light upon the causes underlying amphibian metamorphosis. The various views regarding the relation of iodin to the thyroid held by investigators to-day may be briefly summarized under three heads: 1. Some are of the opinion that the activity of the thyroid depends upon its iodin content and that thyroid free of iodin has no physiological activity. 2. Another group of writers take the view that there is no relation whatever between the physiological activity of the thyroid and its iodin content; i.e., that the iodin usually present has no importance in the economy of the organism. 3. Still other investigators admit the parallelism between physiological activity and iodin content, but deny that iodin is the causal agent, believing rather it is simply associated accidentally with the active principle of the gland. The effects of iodin feeding to normal tadpoles give additional confirmation to the first of these views. But especially interesting in this connection are the results with feeding iodin to thyroid- less larvae. If animals without the vestige of a thyroid gland are stimulated to complete metamorphosis in an abnormally short time by iodin, it would appear that iodin functions within ~ the organism as a hormone itself and that the gland functions chiefly for storage purposes. The evidence from the thyroidec- tomized larvae indicates that the animal body is capable of utilizing iodin directly without the intermediation of the gland. The fact that the thyroid gland of man and animals does not invariably contain iodin (shown by Miwa and Stoeltzner, Roos THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO.3 414 W. W. SWINGLE Topfer and Jolin) and yet such individuals remain healthy may be explained by assuming that the tissues of such animals assimi- late directly the iodin taken into the body, leaving no surplus to be collected by the thyroid. Since the tissues of thyroidec- tomized tadpoles can take up iodin in large quantities and use it, there seems nothing inherently improbable in the suggestion. According to Voegtlin and Strouse, iodized amino-acid when fed to tadpoles accelerates the metamorphosis of the animals, but fails to replace the function of the thyroid in pathological cases where there is a deficiency of thyroid function. The work of these authors, in so far as the acceleration of metamorphosis is concerned, agrees with the results obtained in the present experiments. The function of the thyroid in patho- logical cases, however, may be an entirely different question. SUMMARY AND CONCLUSION 1. Iodin and its compounds when fed to the larvae of Rana pipiens and Bufo lentiginosus stimulate metamorphosis in these animals very rapidly. 2. Inorganic iodin when fed to thyroidless larvae of Bufo lentiginosus brings about metamorphosis in an abnormally short time. 3. Iodin appears to function within the organism as a hormone itself without the intermediation of the gland. 4. The suggestion is made that the extraction of iodin from the blood and its storage is the chief function of the thyroid gland. RELATION OF IODIN TO THYROID 415 LITERATURE CITED Auten, B. M. 1918 Jour. Exp. Zodél., vol. 24, no. 3. BastncER 1916 Arch. Int. Med., vol. 17, p. 260. Baumann, E. 1895 Zeits. f. physiol. Chem., Bd. 21, [> oil). Brevi, ARTHUR 1913 The internal secretory organs. Hunt, R. 1907 J. Am. Med. Assn., vol. 49. p. 1325. Hunt, R., AND SEIDELL, A. 1908 Bull. No. 47. Hyg. Lab, U.S. Pub. Health and Mar. Hosp. Serv., Wash., D. C. Hutcutnson, R. 1896 Jour. Physiol. 20. 8. 494. 1898 Jour. Physiol., vol. 23, p. 181. _ Jouin, S. 1896 Festschrift, O. Hammarsten, Upsala. KeEnpatu, E.C. 1914 Jour. of Biol. Chem., vol. 20, p. 501. 1917 Jour. Amer. Med. Assn., vol. 66, p. 811. Lenuart, C. H. 1915 Jour. Exp. Med., vol. 22, p. 739. Menpet, L. B.. 1900 Am. J. Physiol., vol. 3, p. 290. Marine, D., anp WituiaMs, W. W., 1908, Arch. Int. Med., vol. 1, p. 378 Miwa, S8., AND STOELTZNER, W. 1897 Jahrb. f. Kinderh, Bd. 45, S. 83. Morss, M. W. 1914 Jour. Biol. Chem., vol. 19, no. 3, p. 421. 1918 Biol. Bull., vol. 34, no. 3. OswaLp, A. 1902 Arch. f. Path. Anat., Bd. 169, S. 461. 1902 Beitr. z. Chem. Physiol. u. Path., Bd. 2, S. 555. Swinele, W.W. 1918 Jour. Exp. Zoél., vol. 24, no. 3, p. 521. 1918 Jour. Exp. Zodl., vol. 24, no. 8, p. 545. 1917 Biol. Bull., vol. 33, no. 2, p. 116. Toprer 1896, Wien, klin. Wehnschr., Bd. 9, 8S. 141 (quoted by Hunt and Seidell). Resumido por el autor, W. W. Swingle. Estudios sobre la relacién del iodo con la tiroides. II. Comparaci6n entre la tiroides de las larvas normales de rana y la de las larvas alimentadas con iodo. En este trabajo el autor demuestra que la ingestién del iodo y sus compuestos, 1lodoformo y ioduro potasico, por las larvas de rana, produce rapidamente la metamorfosis, estimulando de este modo la aeccién del tejido o del extracto tiroideos. El] exdmen microsc6pico de la tiroides de larvas alimentadas con iodo y el de la misma glandula de larvas normales de la misma edad, man- tenidas con el mismo tamafio que los animales sujetos al experi- mento por medio de una nutricién deficiente, demuestra que las glindulas de aquellas son mayores que las de las larvas normales y contienen mas substancia iodada. La comparacién entre la tiroides de larvas de una longitud media de 10.5 mm., alimenta- das con iodo, y la de larvas normales de la misma edad pero de una longitud media de 21.5 mm., producida por una alimentacién abundante, demuestra que las glindulas de ambos grupos de animales son aparentemente del mismo tamano. El autor de- scribe experimentos en los que se ha comparado la rapidez de accion de varios compuestos iodados sobre la iniciacién de la metamorfosis. El iodo inorgdnico result6 ser el mas eficiente y en segundo y tercer lugar el iodoformo y ioduro potdsico, respec- tivamente. Los experimentos llevados a cabo para determinar la solubilidad del iodo en el suero sanguineo normal demuestran que el del conejo a 37°C. actua como disolvente de los cristales de iodo en una proporcion de .00075 gramos de esta substancia por centimetro ctibico. El poder disolvente del suero de la rana es algo menor que el del conejo, pero considerablemente mayor que el del agua. El] autor incluye en el trabajo una discusién sobre la relacioén del iodo con la metamorfosis de los anfibios y la funcién tiroidea. Translation by Dr. José F. Nonidez, Columbia University. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 23 STUDIES ON THE RELATION OF IODIN TO THE TERY ROLE? II. COMPARISON’ OF THE THYROID GLANDS OF IODIN-FED AND NORMAL FROG LARVAE W. W. SWINGLE Fellow in Biology, Princeton University INTRODUCTION In the foregoing paper (part I of these studies) the writer described the effects of feeding iodin and various of its com- pounds to normal and thyroidless frog larvae. It was found that the administration of iodin, iodoform, and potassium iodide greatly accelerated metamorphosis in these animals, despite the fact that normally thyroidless tadpoles never assume the adult characters. The results obtained with feeding iodin to thyroidectomized larvae had led the writer to advance the view that iodin functions within the organism as a hormone itself, without the inter- mediation of the gland. Furthermore, it was suggested that the chief function of the thyroid appears to be that of iodin storage, and not, as the current view would have us believe, the elabora- tion of internal secretion. The present paper is a presentation of the results of a comparative study of the thyroid glands of iodin-fed and normal animals of the same age together with some additional data gathered since the publication of part I of this series on iodin feeding and amphibian metamorphosis. MATERIAL AND METHODS Some of the material used in the present work was obtained from the iodin-fed and control cultures described in part I. The The experimental work for this paper was done while the, writer was instructor in zoology at the University of Kansas. 417 418 WwW. W. SWINGLE material taken from these earlier cultures consisted entirely of Rana pipiens larvae. The remainder of the material was taken from cultures of iodin-fed Bufo larvae. All animals used were appropriately controlled with animals of the same age and reared under the same environmental conditions. The iodin-fed larvae were killed and preserved for microscopic examination when they showed marked indications of hyperthyroidism. The fixing fluid used was potassium-bichromate-acetic (Tellye- snicky’s). Only the lower jaw and heart region were preserved. After washing thoroughly, the tissue was placed in toto in alum- cochineal for thirty hours; washed in distilled water and run up through the alcohols to absolute. The tissue was cleared and kept in oil of wintergreen until ready for use. All measure- ments recorded for the glands were made with an eye-piece micrometer, Bausch & Lomb microscope, 16-mm. objective, ocular 1. The greatest length and width of the gland only were measured. CULTURE I. IODIN-FED AND CONTROL LARVAE The animals of this culture averaged 10.5 mm. in length when first started on the iodin diet; none of the animals revealed any indications of limb buds. Thirteen days from the date of first iodin administration all of the experimental animals showed marked symptoms of hyperthyroidism, the indications of which are characteristic in this species (Rana pipiens). All growth of the larvae had ceased with the first iodin feeding, and from then on had in many of the animals actually decreased, owing to tail resorbtion; the larvae were emaciated; tail atrophy was apparent; movement sluggish; all iodin-fed animals had _ well-developed hind limbs. The controls for this culture showed none of the changes enumerated, but had increased considerably in size. None of the controls had limb buds. Table 1 gives the length of the animals of both experimental and control groups and indicates the condition of the limbs thirteen days from the beginning of the experiment. ‘Two sets of controls were used for each iodin-fed culture, one set was fed beef and large quantities of algae each day, the other set RELATION OF IODIN TO THYROID 419 was fed very little in order to hold the growth of the animals in check so as to keep them approximately of the same body length as the animals of the iodin-fed culture. The length of the well-fed group of controls only is indicated in the table; the animals of the underfed culture measured 10.5 mm. TABLE 1 IODIN-FED LARVAE CONTROLS FOR IODIN-FED LARVAE Total length Limbs Total length Limbs mm. mm. 12.0 + 20.5 — 10.0 + 24.0 — 9.0 ae 21.0 —_ 9.5 + 23.9 o 9.0 + 22.0 -- 10.5 + 24.5 _ 125 + 20.0 — 10.0 ae 19.0 = 13.0 + 21.0 — 10.5 AF 23.5 = 9.0 + 20.0 — 9.5 4 20).5 10.0: + 21.0 — 10.5 + 24.5 — 9.0 ar 22.0 — IMA at 20.0 — 12.0 + 19.5 — 12.5 + 21.0 _ 9.0 AF 23.0 — 9.0 ae 20.5 _ 10.36 PM 595) Microscopic examination of the thyroid glands of this series of tadpoles revealed a rather interesting condition; the glands of the iodin-fed larvae were approximately of equal size with those of the controls. This, despite the fact that in regard to body size, the iodin-fed animals were only half the size of the controls as a glance at table 1 shows. When the glands of the iodin-fed animals were compared with those of normal animals of the same age held at 10.5 mm. by underfeeding, it was found that the iodin-fed larvae had considerably larger glands. The average 420 W. W. SWINGLE length and width of twenty thyroid glands of these underfed animals compared with those of iodin-fed and overfed animals is shown in table 2. The measurements given are of the right glands; the left glands were also measured, but as they showed nothing unusual the measurement for them are not recorded in the table. TABLE 2 Measurements of the thyroid IODIN-FED LARVAE CONTROL (UNDERFED) CONTROL (WELL-FED) Length Width Length Width Length Width mm. mm. mm. mm. mm. mm. 0.2727 0.0909 0.1818 0.0545 0.3636 0.1270 0.3090 0.1090 0.1727 0.0545 0.3181 0.1181 0. 2363 0.0818 0.1727 0.0636 0.2363 0.0909 0.3726 0.1636 0.1545 0.0545 0.2999 0.1090 0.3181 0.1270 0.1818 0.0818 0.2727 0.0999 0.2908 0.1090 0.1636 0.0727 0. 2636 0.0999 0.2727 0.0999 0.1818 0.0727 0.3272 0.1270 0.2817 0.1181 0.1727 0.0636 0. 2636 0.0909 0. 2727 0.1090 0.1454 0.0818 0.2727 0.1090 0.2908 0.0909 0.1636 0.0454 0.3090 0.1181 0. 2545 0.0909 0.1363 — 0.0545 0.2545 0.0909 0.2727 0.0818 0.1818 0.0727 0.2999 0.1270 0. 2636 0.1181 0.1636 0.0727 0.2545 0.0999 0.2999 0.0999 Oe ie7 0.0545 0.2272 0.0818 0.2817 0.0999 0.1454 0.0454 0.2999 0.1090 0. 2454 0.0909 0.1545 0.0818 0. 2454 0.1090 0. 2636 0.0818 0.1636 0.0636 0.2636 0.1181 0.2727 0.0818 0.1908 0.0909 0.3181 0.1270 0. 2545 0.1090 0.1454 0.0545 0.2908 0.1181 0.2999 0.1090 0.1818 0.0818 0. 2363 0.0818 0.2812 0.1031 0.1663 0.0658 0.2808 0.1076 It is obvious from these figures that the thyroid glands of iodin-fed larvae are larger than those of normal larvae of the same age and body size (held at 10.5 mm. length by under- feeding), though they are not larger than the glands of normal animals of the same age presenting marked size differences due to overfeeding. As is well known, the thyroids of frog larvae increase in size with the growth of the organism as a whole. When the fact is taken into consideration that the iodin-fed tadpoles RELATION OF IODIN TO THYROID. 421 are just half the size of the animals of the well-fed culture of controls, it is clear the iodin-fed animals have relatively much the larger glands. Microscopic examination of the colloid content of the glands of the experimental and the two control cultures of larvae, shows-a marked difference in the amount of colloid visible in the follicles. The glands of the iodin-fed animals were packed with this substance, whereas the glands of the controls showed a rather scanty amount. Since the completion of part I of these studies, the writer has carried out several more iodin-feeding experiments in order to test various points left untouched in the earlier work. One of these points barely touched upon was the comparative rapidity of action of the various iodin compounds in accelerating meta- morphosis in normal and thyroidless tadpoles. A detailed ac- count of the experiment will not be given here, as it was for the most part a repetition of the experiments described in the earlier paper. Suffice to state here that iodin crystals when fed to frog or toad larvae with and without thyroid glands bring about metamorphic changes in the larvae much more rapidly than any of the compounds used; iodoform is somewhat slower in its action, but is much more rapid than potassium iodide. Three feeding experiments were carried out in which potassium iodate was used as food, but as only negative results were obtained the conclusion is justified that this compound has no accelerating effect upon metamorphosis. The larvae eat the substance, but apparently are unable to break it down sufficiently to release free iodin. While engaged in the, experimental work which formed the subject-matter of the previous paper, the writer was under the impression that perhaps the results obtained from feeding iodin to tadpoles were due to the mixture of flour, iodin, and water used, and not entirely to the iodin itself. This erroneous idea was due to the fact that in several earlier experiments made to test this point it was observed that frog larvae die very quickly if placed in containers with inorganic iodin crystals or in weak solutions of this substance. However, further work along this 422 W. W. SWINGLE line has shown that both normal or thyroidless frog or toad larvae will undergo metamorphosis very quickly if placed in extremely weak solutions of iodin. The defect in the earlier work was that the solutions were too strong. Just a trace of iodin in the water is sufficient to produce results if the solution is kept fresh. Cultures of tadpoles fed on wheat-flower paste showed no changes whatever when compared with beef-fed or algae-fed controls. This experiment shows clearly that iodin is the active principle of the mixture of flour, iodin, and water, fed in the previous work, and that the flour has nothing to do with the results obtained. The fact that thyroidless tadpoles readily undergo meta- morphosis when fed iodin led the writer to suggest that the function of this gland is chiefly that of iodin storage, rather than the elaboration of a specific hormone, and, moreover, that the tissues of animals are capable of utilizing iodin directly without the intermediation of the gland. In this connection the results of tests made to determine the solubility of iodin in normal blood serum may be of interest. The serum of amphibians and mammals was used; the amphibian serum was obtained from adult Rana pipiens, the mammal serum from rabbits. The serum of the latter at 37°C. acts as a solvent for finely ground iodin crystals to the extent of 0.00075 gram per cubic centi- meter when stirred vigorously. The solvent powder of Rana - pipiens serum is somewhat less than that of rabbits, though considerably more than that of water. DISCUSSION The iodin-feeding experiments described in this and the pre- ceding paper should prove of interest to students of amphibian metamorphosis, as they give a clue as to the nature of one of the underlying causes of this phenomenon. It has been assumed, and probably correctly so, that one of the prime requisites’ of the change from the larval to adult condition in Anurans is a heightened metabolism. The work of Gudernatsch with feeding thyroid to tadpoles showed that this substance accelerated meta- morphosis, and it is generally agreed that the effect of thyroid RELATION OF IODIN TO THYROID 423 extract on tadpoles is accomplished chiefly by greatly accelerating catabolic activities. The writer in 1915-1916 (results published in 1918) in an experiment to test the effects of inmanition upon the development of the germ cells and germ glands of frog larvae, found that starvation totally inhibits all body growth and differentiation, the animals consequently never assuming the adult condition. Such prolonged starvation undoubtedly acts as a depressor of metabolism. Later Allen observéd that thyroidless tadpoles do not undergo metamorphosis, but instead grow abnormally large. In this case the absence of the thyroid function had led to a prolonging of the anabolic phase of the metabolic activities. In part I of this series of iodin studies it was shown that iodin accelerates metamorphosis in both normal and thyroidless tadpoles. The iodin effect like that of the thyroid tissue or extract (and indeed iodin seems to be the active principle of the thyroid) is in heightening catabolism. In these four experiments there is fairly good evidence for the view that amphibian metamorphosis was due, in part at least, to heightened metabolism of the catabolic type. In a state of nature the metamorphosis of frog larvae, is under normal con- ditions, effected by none of the experimental agencies mentioned, except iodin. This substance is found in many plants (though perhaps accidentally present, as some authors believe); it is present in the soil in combination with other substances and present in the thyroids of most animals. Jodin in some form or other may be said to constitute a normal environmental factor of amphibians, a factor which, when considered! in connection with the hereditary factors governing growth processes in larval amphibians, gives a rationale of the factors involved in the metamorphosis of these animals. As pointed out by Morse (718), it is impossible to bring about complete metamorphosis in extremely young larvae, when an attempt to do so is made by feeding large quantities of thyroid or iodin, the animals die. Marked metamorphic changes appear, but death usually supervenes before complete transformation takes place. A certain cycle of events must take place before metamorphosis, and this normal cycle is in all probability de- 424 W. W. SWINGLE termined by the hereditary constitution of the organism. Bufo metamorphoses after a few weeks, Rana pipiens requires at least two months; Rana catesbiana is said to require two, three, and sometimes four seasons for complete metamorphosis. The difference in the time required by these amphibians to meta- morphose is very great, and yet at certain stages in their life history they may all be found living together in the same pool. It is obvious that the hereditary factors controlling the growth processes play a very great role in determining the time of metamorphosis, and that iodin is simply an initiator of the process. SUMMARY AND CONCLUSION 1. The thyroid glands of iodin-fed frog larvae are larger than the glands of control animals held at the same body length as the animals of the iodin-fed culture by underfeeding. 2. The follicles of the glands of such iodin-fed larvae contain much greater colloid mass than the follicles of the controls. 3. Solutions of iodin will bring about metamorphosis in both. normal and thyroidless tadpoles in a short time. 4. Iodin is much more active in accelerating metamorphosis than any of its compounds. Next in order of activity are iodo-- form, potassium iodide. Potassium iodate appears to have no effect. 5. The suggestion is made that amphibian metamorphosis is a result of the interaction of environmental agencies, such as iodin and its compounds, with the hereditary factors controlling the growth processes. RELATION OF IODIN TO THYROID 425 LITERATURE CITED ALLEN, B. M. 1918 The results of thyroid removal in the larvae of Rana pipiens. Jour. Exp. Zodél., vol. 24, no. 3. GupERNATSCH, J. F. 1914 Feeding experiments with tadpoles; a further con- tribution to the knowledge of organs with internal secretions. Am. Jour. Anat., vol. 15. 1916 Studies on internal secretions. IV. Treatment of tadpoles with thyroid .and thymus extracts. Reported at the thirty-third session of the Am. Asso. of Anatomists. Morss, M. 1914 The effective principle in thyroid accelerating involution in frog larvae. Jour. Biol. Chem., vol. 19, p. 421. 1918 Factors involved in the atrophy of the organs of the larval frog. Biol. Bull., vol. 34, no. 3. Swinete, W. W. 1918 The acceleration of metamorphosis in frog larvae by thyroid feeding and the effects upon the alimentary tract and sex glands. Jour. Exp. Zo6dl., vol. 24, no. 3. 1918 The effect of inanition upon the development of the germ cells and germ glands of frog larvae. Jour. Exp. ZoGl., vol. 24, no. 3. Resumido por el autor, Merkel Henry Jacobs. La aclimatacién como factor capaz de afectar el punto en que acaece la muerte de los organismos sometidos a temperaturas elevadas Si se conoce el tiempo necesario para producir la muerte de las larvas de estrella de mar sometidas a una cierta temperatura elevada, el tiempo necesario para producir el mismo efecto a cual- quier otra temperatura puede determinarse con un grado muy aproximado de exactitud. En general, el coeficiente de tempera- tura es proximamente 2 para cada elevacion térmica de 1°C. El co- eficiente de temperatura de Paramoecium bajo condiciones seme- jantes esta’ mucho mas sujeto a variacién, pudiendo-ser 3 0 mas algunas veces, en otras ocasiones menos de 2. Durante la ele- vacion gradual de la temperatura en las larvas de estrella de mar, se suman los efectos nocivos de todas las temperaturas por las cuales han pasado durante dicha elevacién y mueren casi en el mismo momento en que se aleanza el punto en que deben morir tedricamente como resultado de la suma de todos los efectos noci- vos. El punto en que tiene lugar la muerte del animal es tanto mas bajo cuanto mas lenta es la elevaci6én de la temperatura. En Paramecium, por el contrario, el punto en que el animal muere es tanto mas alto cuanto mas lenta es la elevacién de la tempera- tura. En esta especie la aclimatacién modifica las relaciones tan simples que se presentan en la larva de la estrella de mar. El grado de aclimatacién en una forma determinada puede estimarse cuantitativamente determinando el ‘‘exceso de resistencia” (sur- plus resistance) mediante el método que el autor describe en el trabajo. Medido en términos de cantidad de efectos nocivos necesarios para producir un resultado fatal, este ‘‘exceso de resis- tencia’”’ es préximo a cero en las larvas de estrella de mar, mien- ~ tras que en Paramecium llega a ser 45. Estas cifras dan una medida aproximada de la capacidad de adaptacién a tempera- turas mds y mids elevadas en las dos especies mencionadas. Translation by Dr. José F. Nonidez, Columbia University. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 ACCLIMATIZATION AS A FACTOR AFFECTING THE UPPER THERMAL DEATH POINTS OF ORGANISMS M. H. JACOBS University of Pennsylvania 1. INTRODUCTION The question of the effects produced on organisms by high temperatures is one which has received the attention of biologists for many years. The older workers were interested chiefly in the determination of the so-called ‘upper thermal death points.’ A résumé of their observations is given by Davenport (’97). In more recent times, the importance of the time factor, overlooked in this earlier work, has been recognized, and modern investi- gators have been more concerned with ‘temperature coefficients’ (Pitter, 914; Kanitz, 715) and with the possible causes of injury at the elevated temperatures. In most of the recent, and in practically all of the older work, however, a factor not sufficiently taken into account is the method by which the temperatures used in the experiments have been attained. There are, in general, three chief ways of bringing an organism to a given high temperature. 1) The change may be practically instantaneous, as, for example, if a minute animal in a small quantity of water is suddenly expelled from a pipette into a large volume of water at the required temperature. 2) The change may be gradual, but uniform, as, for example, if the animal is placed in a vessel of water at room temperature, and heat applied in such a way that the rise per minute remains constant until the desired point is reached. 3) The change may be gradual, but at a constantly decreasing rate, as, for example, if the animal is placed in a test-tube containing water at room temperature, and the test-tube is then plunged into a large 427 428 M. H. JACOBS vessel of water which is kept: at the final temperature. In this case, the rise is at first rapid, becoming progressively slower and slower—being represented, in fact, by the logarithmic curve of Newton’s well-known ‘law of cooling bodies,’ except that the temperature in this case is increasing instead of decreasing. Of the three methods mentioned, the third is generally unde- sirable, partly because of the greater difficulty of making allow- ance for a constantly changing rate of temperature increase, partly on account of the difficulty of determining the exact time when the desired temperature has been reached, and chiefly because of the great length of exposure (due to the slow rate of change as the end point is approached) to temperatures whose effects are almost as great as those of the one finally attained. Method 1, in exact work, is applicable only to very small organ- isms, since any attempt to apply it to large ones results in secur- ing essentially the effects of method 3, with the additional disad- vantage that on account of the slow conduction of heat, different parts of the body reach the final temperature at different times. Method 2, therefore, which involves raising the temperature at a known rate until the desired point has been reached, is the most suitable one for all except very small organisms and has been most frequently employed. But in the past very little effort has been made to take into account the influence on the final result of the rate at which the temperature is raised. That this factor is probably of importance is obvious, but to predict in advance in what direction and to what extent it will operate is not always an easy matter. Suppose, for example, that in a certain experiment a lot of organisms are brought in the course of ten minutes from room temperature to 40°C. and kept at the latter point until death occurs. Would the time required to cause death at 40° be greater, or less, if in another similar experiment the preliminary rise were allowed to occupy thirty minutes instead of ten? It might be argued, on the one. hand, that the slower rate would be less favorable to the organisms than the rapid one because of the greater length of exposure to temperatures below 40°, but still sufficiently high to produce in the aggregate considerable ’ UPPER THERMAL DEATH POINTS 429 injury before the final temperature had’ been attained. The possibility exists, however, on the other hand, that the slower rate would be more favorable than the rapid one in giving greater opportunity for adjustment or acclimatization to occur. Which of these two alternatives is the correct one for a given form can, as a matter of fact, be decided only by experiment. In the present paper a method is suggested for determining this point and for dealing quantitatively with certain other aspects of the general problem of acclimatization. The writer wishes to express his indebtedness to Prof. F. R. Lillie for kindly placing at his disposal on several occasions the facilities of the Marine Biological Laboratory at Woods Hole and to Mr. Francis H. Adler for assistance in making certain of the observations on which the paper is based. 2. MATERIAL AND APPARATUS In the experiments to be described, the use of both methods | and 2 was necessary for the quantitative estimation of the extent to which acclimatization occurs. For this reason, only small organisms were employed, starfish larvae eighteen to forty-eight hours old and Paramecium caudatum being the ones chosen. The medium in which they were heated was for the starfish larvae fresh sea-water and for Paramecium, in most cases, the natural culture fluid filtered to remove all animals. It was recognized that the complex nature of the culture fluid might introduce undesirable complicating factors, but preliminary experiments showed that, as a matter of fact, distilled water, which would naturally have been preferred on account of its uniform composition and in which the animals lived normally at room temperature for days, was quite unsuitable for sudden exposures to high temperatures, the animals dying far more quickly in it than in their own culture medium, and the results obtained being markedly irregular. That at least part of the effect of the distilled water was of an osmotic nature was shown by the fact that the addition to the same water of slight amounts of neutral salts or even of cane sugar made it considerably less injurious. It THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 3 430 M. H. JACOBS is perhaps possible also that the absence of appreciable amounts of ‘buffer substances’ may have been another of the factors con- cerned, since there is some evidence of the production of abnormal amounts of acids at elevated temperatures. At any rate, it was found that apparently the most reliable results could be ob- tained when normal culture fluid was used, although very similar results were also secured in some cases with pond-water when the animals had been kept in it for at least twelve hours previous to the experiments. The apparatus employed was of a simple nature. It consisted of a 2-liter beaker, used as a water-bath, supported on a stand and heated from below by an alcohol lamp whose position could be altered to furnish much or little heat as desired. In the beaker were placed a number of test-tubes containing enough water or culture fluid to make them float upright. The trans- parency of the whole apparatus was found to be of advantage, not only in favoring such, manipulations of the material as were necessary, but in making it possible to observe the visible effects of the high temperature on, for example, the movements of the animals. When a sudden exposure was desired, the water in the water- bath and in the test-tubes was first allowed to assume the proper temperature, and then a considerable number of the organisms were taken in the smallest possible quantity of water in a capillary pipette (this being very easy in the case of both of the animals used on account of their habit of collecting in a dense ring around the edges of the culture jar) and suddenly forced into one of the test-tubes in such a way as to insure thorough mixing. The quantity of water used was so small as practically not to affect the temperature of the water in the test-tube, calculation showing that the momentary lowering of its temperature, which was not even indicated on an ordinary mercurial thermometer, could not have been, as a rule, more than 0.1°C. After this sudden intro- duction to the temperature of the experiment the organisms were either all allowed to remain in the test-tube for the required length of time and then suddenly poured into sufficient cool water to bring them back immediately to within their normal UPPER THERMAL DEATH POINTS 431 range of temperature, or they were removed, a few at a time, at the proper intervals with a capillary pipette. Where a gradual rate of temperature increase was desired, the same general methods were employed except that the animals, instead of being introduced suddenly into the test-tubes, were placed in them at room temperature and the whole apparatus was heated at the desired rate, samples of the animals being removed, usually at half-degree intervals. The animals, whether suddenly or gradually exposed, were kept under observation after removal in Syracuse watch-glasses until they had either died or recovered, which in some cases required as much as twenty-four hours, although as a rule their behavior when first examined left little doubt as to the ultimate outcome of the experiment. 3. METHOD OF ESTIMATING ACCLIMATIZATION The extent to which acclimatization occurs during a slow rise of temperature may theoretically be estimated by first finding by method 1 (where there is no opportunity for preliminary acclimatization to occur) the amount of injury inflicted in unit time at the various temperatures passed through during the rise, and in them adding together these separate injuries, be- ginning with the lowest. temperature, and taking into account the duration of each, until a total just sufficient theoretically to cause death is arrived at. The point at which this total is reached is compared with the observed death point in the case of the gradual rise. If the two points practically coincide, it may be said that there is no evidence of acclimatization. If, on the other hand, the observed death point is higher than the calculated one, the presumption is that acclimatization has occurred and the amount of the latter can be estimated, roughly at least, in quantitative form. It must be remembered, of course, that the calculated death point cannot be determined by merely adding the theoretical amounts of injury at, for example, 34°, 35°, 36°, ete., since the rise does not proceed by a series of sudden jumps, but continuously. Since, however, the relation of 432 M. H. JACOBS the amount of injury inflicted in unit time at any one tempera- ture to that inflicted in the same time at any other temperature is, for a certain range, in the forms studied, governed by a simple mathematical law, it is possible by making observations at a few selected temperatures, and thus determining the necessary constants, to calculate the theoretical effect of a continuous change of temperature at any desired rate. ‘Fhe details of the’ method will be made clearer in the following sections where the actual experiments are discussed. 4, EXPERIMENTS ON STARFISH LARVAE Since the results obtained with starfish larvae are simpler than those with Paramecium, they may be considered first. In table 1 are given the lengths of exposure found to cause death when the animals were suddenly subjected by method 1 to the temperatures in question. The fatal exposure in each case is TABLE 1 Times required to kill approximately one-half of the individuals of starfish larvae when suddenly subjected to various temperatures 5 : JUNE 16, 18 HOURS OLD. JUNE 21, 48 HOURS OLD | JUNE 22, 24 HOURS OLD gece aries a : Fatal exposure Qi Fatal exposure Qi Fatal exposure Qi Fatal exposure deg.C 40 8 seconds 10 seconds 1.9 1.8 39 | 15 seconds 18 seconds 1L& 7 38 | 23 seconds 30 seconds Dee rh 37 | 50 seconds 50 seconds 45 seconds 35 seconds 5 19 2.0 36 | 1.25 minutes 95 seconds 1.5 minutes DED, DED, Ded 35 | 2.75 minutes 3.5 minutes 4 minutes 2.5 minutes DAD 2.3 2.0 34 7 minutes 8 minutes 8 minutes Dial Pin) 33 | 15 minutes 20 minutes 13 minutes D2) 32 45 minutes 30 minutes UPPER THERMAL DEATH POINTS 433 taken somewhat arbitrarily, as the time required to produce injuries from which approximately half of the individuals failed to recover. The actual death of the animals, according to the temperatures employed, may occur in from a few minutes to twenty-four hours or more after restoration to normal conditions. Undoubtedly such differences are in certain respects significant, but for purposes of immediate comparison they may be disre- garded and the degree of injury which is just sufficient ultimately to lead to death, regardless of the time required, may be accepted as the most satisfactory available criterion. Itmay be mentioned that the starfish gastrulae obtained from a single lot, of eggs show little individual variation; the fatal exposure is very nearly the same for all. In this respect they differ from Paramecium in which the individual differences are large. It will be noticed in table 1 that the time required to produce fatal injury at any temperature bears a fairly definite relation to the time required at other temperatures. Thus, at 34°, for example, about twice as long a time is required as at 35° and about one-half as long a time as at 33°. Expressed in mathe- matical symbols, 5 Meee. where L denotes the length of life, 6 the temperature, and Q, the temperature coefficient for a change of one degree. Values of Q; are given in alternate columns of table 1. The general average of all of the values of Q; is 2.1. This value agrees closely with that found, for example, by Loeb (’08) for the eggs of Strongylocentrotus and Moore (710) for Tubularia crocea. These results may also be stated in another form (table 2). The amount of injury (Ig) inflicted at the temperature @ by an exposure of unit time (one minute) can be expressed in quanti- tative form by taking as unity the amount of injury just sufficient to produce death. Thus at 38°, in the series selected for table 2, where an exposure of twenty-three seconds is necessary to cause death, Isg-=2.6; in the same way I3.-=0.14; and the other values are given in column 2 of table 2. The mathematical 434 M. H. JACOBS TABLE 2 Theoretical amounts of injury inflicted on starfish larvae at various temperatures. The fatal exposures are taken from column 1 of table 1 TEMPERATURE TOTAL EXPOSURE INJURY IN UNIT TIME deg. C. seconds 40 8 OD 39 115 4.0 38 23 2.6 oT 50 iy 36 75 0.8 35 165 0.36 34 420 0.14 a8 900 0.07 relation that exists between the amounts of injury inflicted in unit time at different temperatures is Ts = I,Q? where I, is the amount of injury inflicted at the temperature chosen as the standard for comparison, and I, the amount inflicted at any other temperature separated from the first one by p degrees. If the second temperature is lower than the first, p, of course, has a negative sign. Of the two constants in the above expression, Qi may in general be taken for starfish larvae with sufficient accuracy as equal approximately to 2, and I, may be determined experimentally for a given lot of organisms for any convenient temperature, preferably a rather low one, as the percentage of error is then less. Having these two constants, it is possible to calculate not only the amount of injury that would be inflicted by an exposure of any length to any temperature to which the above equation applies, but like- wise the amount of injury that ought theoretically to be in- flicted during a gradual rise from room temperature to any desired temperature. In the latter case, the amount of injury would be represented graphically by the area of the curve: y =aQi UPPER THERMAL DEATH POINTS 435 between the limits x = — b (room temperature) and x = p (the highest temperature attained) ; b and p, of course, being measured from the temperature selected as the point of comparison. The constant a is the amount of injury, Io, inflicted in unit time at this temperature. The area of the curve, A, which represents the total injury, I, therefore is: p p —b Rei ma'| pdx =a(_% a ) ena’ igaheh Hoa) Since b (the number of degrees the room temperature lies below the temperature chosen for comparison) is relatively large, the second half of the expression within the parentheses becomes negligibly small and may be disregarded. This is equivalent to calculating the injury that would be inflicted in a rise from an infinitely low temperature instead of from room temperature, but this amounts to practically the same thing, since even considerably above room temperature the injury inflicted in any ordinary time has ceased to be appreciable. If instead of raising the temperature at the rate of one degree per minute as implied in the calculation Just given, the rate had been slower, say one degree in t minutes, the right-hand side of the equation would have to be multiplied by t. The general expression therefore for the area, A, which represents the total injury, I, inflicted up to the temperature p° when the rate of rise is one degree in t minutes becomes (when a is replaced by its equivalent I,): In the case of starfish larvae where Q; is equal to approximately 2.0 and log. Q: therefore to approximately 0.7 we have finally: 2 p L=the-os S In case it is desired to know how high the temperature would have to be raised to inflict just fatal injury, it is only necessary A36 M. H: JACOBS to substitute for I the numerical vale 1.0 and solve the equation for p. In such eases (taking log) 2 = 0.3) wa gee mG, 0.3, The results of applying this method to a gradual and regular rate of temperature increase in the case of starfish larvae are shown in table 3. In the first column is given the rate at which the temperature was raised, in the second the observed death point, and in the third the point at which death ought theo- retically to have occurred (i.e., the point at which the area enclosed by the curve becomes unity) when the value of the constant, Q,, was taken as equal to 2 (this value holding approxi- mately for all of the starfish larvae studied). The value of I), which varies somewhat for different lots of larvae according to age, etc., was determined especially for the animals used in these experiments, and was found to be 0.05 at the temperature (83°C.) chosen for comparison. It must be recognized, of course, that for the portion of the curve between room temperature and 32°, no exact observations are available, and it is uncertain to what extent the above equation applies to it. But the area of this portion of the curve is in any case so small as compared with that above 32° that the final result would be little affected even if a different relation were TABLE 3 Temperatures at which death of starfish larvae occurred after varying rates of tem- perature increase from a starting point of approximately 20°C. THEORETI- THEORETI- CAL DEATH CAL DEATH RATE OF TEMPERATURE TEMPERA- TEMPERA- INCREASE IN OBSERVED DEATH TEMPERATURE TURE TURE DEGREES PER MINUTE CALCULATED | CALCUI.ATED FROM FROM Qi = 2.0 Or 222 1°C. in 1.8 minutes About one-third dead at 36.0° 36.0° 35.8° 1°C. in 4 minutes S5n00 S4eSe 34.8° 1°C. in 5 minutes 34, 5° 34.5° Buk ye 1°C. in 8 minutes All living at 33.5°, all dead at 34.0° Sa.07 33.9° UPPER THERMAL DEATH POINTS 437 shown to hold in this region. It may also be noticed that for the region from 32° to the point of death in these particular experi- ments the value of Q; is in general higher than 2.0, the approx- imate average value for the whole range of temperature studied. In the last column of table 3, the theoretical death temperature is therefore calculated for comparison from the value, Q,=2.2. It will be noticed that in either case the calculated death tem- perature lies within a few tenths of a degree of that actually observed, and the amount of acclimatization that has occurred, if any, is consequently extremely small. In similar experiments on Paramecium, immédiately to be described, the difference may amount to several degrees. 5. EXPERIMENTS ON PARAMECIUM CAUDATUM In the case of Paramecium, the results are more complicated. In the first place, there is considerably more cultural and racial variation in the length of life at any given temperature (de- termined by method 1) than in the case of starfish larvae where the results, on the whole, seem to be remarkably uniform. This is shown in table 4 where some of the results obtained with this form are summarized. The figures in columns 6 and 7 and probably in column 1 are for the three-vacuoled race described by Hance (’15, 717), which TABLE 4 Times required to kill approximately one-half of the individuals of Paramecium caudatum of different races at different temperatures Sie an RACE | (?) RACE 2 | RACE 3 RACE 4 RACE 5 RACE 6 RACE 6 deg. C. 43 30.0 see. 15 see. | 50 sec. 42 1.5 min.| 15.0 sec. | 20.0 sec. | 20.0 sec. | 20 sec.| 1 min.| 2 min. 41 4.5 min. | 45.0 sec. 45 see. 1.0 min. | 1 min.| 8 min.| 4 min. 40 13.0 min. | 2.5 min.| 2.5 min.| 2.5 min.} 5min.| 20min.| 7 min. 39 18.0 min. | 3.0 min. | 3.0 min.| 4.0 min. | 9 min. 18 min. 38 3.5 min. | 7.0 min. 37 4.0 min. 36 6.0 min. 438 M. H. JACOBS in these, as well as in other experiments, has shown itself to be remarkably resistant to high temperatures as compared with the ordinary races. In the second place, the temperature coeffi- cient, Q;, in a given set of experiments is subject to far more variation than in the case of starfish larvae, being as a rule much higher (usually approximately 3) in the region above 40° than in that below this temperature, and being subject at all times to considerable and sometimes inexplicable fluctuations. For this reason, calculations made by the method described are not so exact as in the case of the starfish larvae, but fortunately this is not necessary since Paramecium shows such a high degree of acclimatization that the error due to the simplifying assumption that the value, Q:=38, applies to all temperatures is not able to disguise this fact. In other words, the error that arises from taking this value of Q, for the entire range, while, as a matter of fact, it is considerably less at lower temperatures, is of such a nature as simply to make the difference between the calculated and observed death points less striking than it would otherwise have been. If acclimatization is shown when such a simplifying assumption is made, it would a fortiori be indicated if more exact calculations had been made. This point will be made clearer by an actual example. It was found in one set of experiments that for the three-vacuoled race, Q, between 40° and 43° was equal to almost exactly 3.0. The length of life after a sudden exposure to 41° was found to be 4.5 minutes, i.e., I) (taking this temperature as the standard of comparison) was equal to 0.22. It was also found that when the animals were heated gradually, at the rate of 1° in eight minutes, the observed death point was very close to 44°. The calculated death point, on the assumption that Q: is equal to 3.0 for all temperatures is approximately 40.6°, a difference of 3.4°, indicating a very considerable amount of acclimatization. If instead of assuming that below 40° the injury at any temperature in unit time is only one-third as great as that at the temperature one degree higher (as implied by the value, Q; =3), we had taken into account the lower values of the temperature coefficient which usually apply to this region, it is clear that the amount of UPPER THERMAL DEATH POINTS 439 injury inflicted at the temperatures below 40° would not drop off so rapidly, or, in other words, that the animals would be more injured during their gradual rise by the time they had reached 40°, and that consequently the calculated death tem- perature would be even lower than before. But this would only make stronger the evidence of acclimatization already obtained by the rough calculation. . It is of some interest not merely to show that acclimatization occurs, but to attempt to express its extent in quantitative form. This can be done in an approximate fashion by calculating the area enclosed by the given curve up to 6=44° and comparing this area with that which represents unit injury, i.e., death. Such results, of course, are only rough approximations, but have nevertheless a considerable interest. Using the formula already given: Q? = Gill ° ; ; log. Q: the area in the case just mentioned proves to be 44 units of injury. In other’ words, before death occurred, during the gradual rise, the animals had withstood about forty-four times the usual fatal injury. It is suggested that in this and in similar cases the excess in area enclosed by the curve of injury, up to the point of death, over the area (taken as unity) which repre- sents an amount of injury just fatal when the change is sudden, may be called the surplus resistance, and be used as a rough quantitative measure of the extent of acclimatization. In this case the surplus resistance is equal to 43. The animals have, in other words, added to their normal lives, so to speak, forty- three additional lives by their ability to adjust themselves to the changing environment. The favorable effect of a slow as compared with a rapid rise of temperature on Paramecium is shown by another experiment in which method 3 was combined with method 2. In this case, a tube of very small caliber (8 mm.) with extremely thin walls was prepared by drawing out the lower portion of a thin-walled test-tube in a flame to a considerable length and sealing the small 440 M. H. JACOBS end. Drops of water containing Paramecium could be placed in it and removed with a capillary pipette. It was found by the insertion of a thermocouple in such a drop of water and another in a vessel of water at 41°C. into which the tube was plunged, that the water in the tube reached approximately the tempera- ture of that in the surrounding vessel in about twenty seconds. This fact being known, a number of animals were placed in it and plunged into water at 41° for two minutes and twenty . seconds (giving therefore an exposure of two minutes at 41°). On removal, it was found that all were fatally injured. Another lot were placed in the same tube, but they were brought at a uniform rate from room temperature to 41° in two minutes and then kept at exactly 41° for two minutes longer. In this case about one-quarter of the individuals recovered. A third lot were treated in the-same way except that in this case the rise to 41° occupied twelve minutes. About one-half of these animals recovered. Doubtless a slower increase of temperature would have given even a higher percentage of recoveries. With starfish larvae, it may be mentioned, that experiments made in the same manner showed in every case exactly the reverse effect, i.e., the slower the rate of temperature increase, the higher the mortality. A considerable number of other experiments were tried with Paramecium with the same general results as those mentioned. The degree to which a slow change increased the final resistance varied considerably with different races and under different ex- perimental conditions, but in all cases it was very appreciable. It is apparent, therefore, that in the case of this organism, at least, the upper thermal death points that will be obtained in different experiments by the methods usually employed may be expected to be subject to considerable variations, which in many cases will be difficult to predict. To what extent the same prin- ciple will be found to apply in the case of other organisms can be determined only by further observations. In any event, however, the author believes that to be of value, data on the upper thermal death points of organisms must include not only the length of UPPER THERMAL DEATH POINTS 441 exposure to the particular fatal temperature under considera- tion, but in addition, an exact statement of the manner in which this temperature was reached. SUMMARY 1. The length of life of starfish larvae eighteen to forty-eight hours old at temperatures between 32° and 40°C. is governed with a very fair degree of accuracy by the relation: lg = Q, = approximately 2. Lo+1 For Paramecium, the value of Q; is subject to considerably more variation; above 40°C. it is frequently in the vicinity of 3; below 40° often less than 2. 2. Knowing the above relation and the length of life after a sudden exposure to one or more selected temperatures, it is possible to calculate by. the method given in the body of the paper the point at which death ought theoretically to occur when the temperature is raised uniformly at any given rate. By comparing this point with the observed death point under the same conditions, it can usually be determined whether or not acclimatization has occurred. 3. With the different rates of temperature increase employed in these experiments the observed death points of starfish larvae agree very closely with the calculated ones, indicating prac- tically no acclimatization. With Paramecium caudatum the observed death point may be much higher than the calculated one, indicating that acclimatization occurs even in experiments of short duration. 4. If desired, the ‘surplus resistance’ in a given experiment may be obtained in quantitative form by determining the excess in area enclosed by the curve of injury up to the point of death, over the area (taken as unity) which represents an amount of injury just fatal when the temperature is changed suddenly. 5. As determined in this way, the ‘surplus resistance’ for star- fish larvae for a number of different rates of increase of tempera- 442 M. H. JACOBS ture was found to be in the vicinity of zero. For Paramecium caudatum, on the other hand, a ‘surplus resistance’ as high as 43 has been found. 6. In general, the slower rates of temperature increase are more favorable for Paramecium and more unfavorable for starfish larvae than the more rapid ones. 7. It is suggested that future data on upper thermal death points, ete., shall include not only the times of exposure to the temperatures in question, but exact statements as to the methods by which these temperatures have been reached. LITERATURE CITED Davenport, C. B. 1897 Experimental Morphology. New York. Vol. 1. Hance, R. T. 1915 The inheritance of extra contractile vacuoles in an unusual race of Paramecium caudatum. Science, N.S., 42. 1917 Studies on a race of Paramecium caudatum possessing extra contractile vacuoles. Jour Exp. Zodl., vol. 23. Kantrz, A. 1915 Temperatur und Lebensvorginge. Berlin. Lors, J. i908 Uber den Temperaturkoeffizienten fiir die Lebensdauer kalt- bliitiger Tiere und iiber die Ursache des natiirlichen Todes. Pfliiger’s Archiv, 124. : Moorg, A. R. 1910 The temperature coefficient of the duration of life in Tubularia crocea. Arch. Entwick., Bd. 29. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 4 FEBRUARY, 1919 Resumido por el autor, G. M. White. La asociacién y distincién de los colores por los peces Umbra limi y Eucalia inconstans. Los peces de agua dulee Umbra limi y Eucalia inconstans fueron ensefiados a asociar los alimentos con un cierto color y al mismo tiempo a asociar substancias insfpidas, tales como el papel, con otro color. Los individuos de Umbra distinguian entre si las siguientes luces monocromaticas: roja y verde, roja y azul, amarilla y verde. La variacién de intensidad de las luces roja y verde desde 1.4 cm. hasta 4.9 em. no impide la distincién del color por parte de los peces, indicando esto que su reaccién en tales cireunstancias se debe mas bien al color de la luz que a su intensidad. Mientras que los individuos de Eucalia distinguian entre las luces roja y verde, asociandolas con el alimento y el papel, nunca pudieron aprender a diferen- ciar el azul del amarillo. Ambas clases de peces fueron también sometidos a la accion de la luz transmitida a través de placas fotogrdficas veladas con diferentes tonos de gris, para probar si podian asociar con ellas los alimentos y substancias insipidas, conforme habian hecho con las luces de otros colores. Tal asociacion no se llevé a cabo, lo cual prueba atin mejor que la distincién de los colores por parte de estos animales se debe mas bien a sus longitudes de onda que a su intensidad. Los ex- perimentos efectuados con ambos peces para comprobar si pueden percibir diversos dibujos dieron tan solo resultados negativos, indicando esto que su distincién asi como la de las diferencias en los fondos no son de gran importancia en la busca del alimento. La percepcién del color y la del movimiento parecen ser de la mayor importancia para este fin. Translation by Dr. José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 3 ASSOCIATION AND COLOR DISCRIMINATION IN MUDMINNOWS AND STICKLEBACKS GERTRUDE MAREAN WHITE Department of Zoology, University of Wisconsin TEN FIGURES CONTENTS > TemibAOVSNOVOTAOLa eps oo 6:05. 0.6 a SRR OO OIE HAO Oa AA 2 BASEN Bier ee Nee Oe errno 444 Review of literature on the behavior of fishes.........-..............000000- 444 General review of literature on color vision in fishes................... 446 Conditions,of experimentation’. 305 -.Ayhows yo fas Cleo lt Ga ee. Jando oe 452 MS CMTE TIVE IMOVEMEMUSMee ee nt 2... aR ede oes Crete a. sate te TO Tee unos oe ae Pie 453 Periodic activities and daily rhythm of the mudminnow.................... 453 Sensesoncans used imeseekime) foods...3 ...\soei ass. a. 08.. qo Pas tsa oan Bae he 455 Golo mechs Cid a bi Orn eee Se: = ohne eum ae A ER oe alt Be, Se i a ae 456 Experiments performed upon the mudminnow......................+++ 458 (Coloredhia DEUS ares ite wos cn. eke re Oe cuenta ee eee 458 Migingyelavroyanrenine Inielate inlets, ASenscac co coos cme AwoeMeobosgenveocounmec 462 Constant and varied intensity of illumination...................... 466 Monochromatic. green and yellow filters........................000- 469 shorted photograpialesplatese:s oF Mens wkend ete cnroe Mon 474 * Summary of experiments in regard to color vision of mudminnows..... AT5 Hxperments performed upom the sticklebacks)-)..2......-45-..0..---- 475 Monochromatic blue’ andiyellow filters; !...02.0. .40. 02.05.8008. 25 481 Monochromatic red and green filters.................... er eae 481 NER OPES ECG Cl ae edict ye.) oe ee noe einer stele > Aiea Aaron & 483 PHOSFeUY PAGLOSTAPUIC PIBGES 5/0 nists eet MS aise Siz eae or os eens 485 Summary of results in regard to color vision in fishes.................. 485 Expenimentssonmhe-discrimination ob patterns... 2....... 0. ...8.2. ade eed ele 488 ASROCIADLONS ELORIMed Maa ts MESS M54... «sy uvac Meee seo 5 As cialis WR taiwan 489 Mey PESTO MASS OCIMUION SR ae ae icy. 3 ciao eR EAMES, oc Leanna 4 Grd Ae hs bbe encase 489 WelicaeyaOmassOClahlOns et ae... ie eee Oe a cis dec om ee nents 491 Sam PlexdtyrOLAsSSOGlMONs. May sl f.).rd lage ed Wt, Sahn eos Mh eats tea BR 491 RenmranencenolsassOelatlOnertes so. «fie See ae rtsenisis © - ene petictone 492 Minniireaiion.ol ASSGCIAmOHS: 2.0 cer... 2 Jul) enna meus ile 4 »/s0e)cete 4 wa. cle es 492 GEMCEMeCIS CUSSION BIG COMCIISTONSS i... ac Seek tert ape co a sisln'e ape arto alstelemnae 493 SHELTON & cao bd OR Oil enh che Sis Cd ea sn ae Ce ne RE Rs 494 444 GERTRUDE MAREAN WHITE The behavior of fishes has been a subject of special interest to a large number of scientific investigators. While observations have been made upon a number of species, insufficient evidence has thus far been brought together to furnish a basis for the comparison of various aspects of the psychology of fishes with those of other vertebrates. In connection with this general problem, a series of experiments on the behavior of the mud- minnow, Umbra limi (Kirt!and), and the stickleback, Eucalia inconstans (Kirtland), were undertaken to determine their ability to form associations and to discriminate colors and pat- terns. These experiments were performed in the Zoological Laboratories of the University of Wisconsin. The problem was suggested by Prof. A. S. Pearse, from whom valuable criticism and assistance were received. . REVIEW OF THE LITERATURE ON THE BEHAVIOR OF FISHES The gustatory and olfactory senses of fishes have been studied by Parker, Copeland, and Herrick. Herrick (’03) concludes that the bullhead, the shiner, and the spotted sucker perceive their food through gustatory sense organs. They detect food from a distance and exhibit a ‘seeking reaction.’ The gadoid fishes (pollock, hake, tomcod) are stimulated to take food by the gus- tatory sense, which is located on the fins as well as about the mouth. The tactile sense is also used in finding food, combined with the gustatory sense in these gadoid fishes. Parker (’10, “11, 12, ’13) shows that a true sense of smell, distinct from taste, exists in the bullhead, the dogfish, and the killifish. He defines smell in water as the perception of very dilute substances ema- nating from a distance; taste as the perception of substances near at hand and present in comparatively large amount. Cope- land (’12) shows that the puffer possesses a sense of smell by which it is able to discover hidden food. That fishes are able to hear is a tradition universally accepted . by fishermen. Yet Bateson (’89’—-90) asserts that fishes are not desturbed by sounds made in the air, and that on the whole shocks and concussions do not play much part in the activities of fishes. a ASSOCIATION AND COLOR DISCRIMINATION 445 Parker (03, ’08, *10, 711) believes that certain sounds of low vibration, like the discharge of a gun, are heard and that the sacculus is the chief organ of hearing. According to Parker (’04), the lateral line organs are geneti- cally related to the ear, and are not stimulated by light, heat, salinity of the surrounding water, food, carbon dioxide, oxygen, foulness of the water, currents, or sound, but by vibrations of low frequency—about six per second. Since these organs appear to be affected by such stimuli as disturbances caused by winds and by bodies falling into the water, they may be of significance in orientation, but take no more part in equilibration than the skin, and are less important in this connection than the eye and ear. The sense of touch is well developed in fishes. Bateson (’89- 90) states that the sole appears to use this sense in discovering its food. Herrick (’03) finds that the gadoid fishes which he observed detect their food by means of the tactile sense com- bined with gustatory. Shelford and Allee (’18, 714) made elaborate experiments showing that fishes may react in various ways to gradients of dissolved gases, but no study was made of the sense organs con- cerned. The resistance of fishes to different concentrations of oxygen and carbon dioxide is discussed by Wells (715). In a delicate and ingenious series of experiments, Lyon (’04) proved that the reaction to current in the killifish, the scup, the stickleback, and the butterfish is an optical reflex—as the fish is carried down stream by current, the bottom of the stream ap pears to move in the opposite direction; the fish has a tendency to follow its passing field of vision, and consequently swims against the current. Experiments to show that goldfish and Fundulus can find their way through mazes are reported by Churchhill (’16) and Thorndike (711), respectively. Eigenmann (’00) shows that the integumentary nerves of the blind fishes, Chologaster and Amblyopis, are sensitive to light. According to Parker (’05, ’09), the same is true of ammocoetes; but no salt-water fishes which he observed possessed such photo- ‘ 446 GERTRUDE MAREAN WHITE receptors on the skin. T’schermak (’15) gives a good survey of ° vision in fishes, in which he discusses conditions of vision in water, the absorption of light by water, the formation of an im- age in the fish eye, accommodation, and bifocal vision. LITERATURE ON COLOR VISION IN FISHES The question of color perception in fishes has been a matter of considerable dispute and evidence concerning it has been accumulated from various sources. Adaptation to background by the pigment cells of the skin. The expansion and contraction of pigment cells in such a way as to conform the color and pattern of the skin to the background against which the animal rests have been observed in many fishes. If such changes in the pigment cells are brought about by stimulation received through the eyes and central nervous system, they may serve as evidence of color vision. Lowe (17) states that the cells of the brook-trout begin to react to back- ground after the yolk-sac is absorbed. When the trout are placed in a dark dish, the pigment cells expand, making the fishes appear dark; but when the fishes are in a light dish, the cells are contracted, giving the skin a pale appearance. Frisch (12, ’14) finds that Crenilabrus roissali changes in color when subjected to red, green, and blue light. Adaptation to green and blue light is by means of the contraction of the pigment cells and also by an increase of the blue-green coloring matter. He argues that the adaptation is to color rather than to luminosity, for if the colors used are arranged as they would appear according to their intensity to a color-blind person— yellow, green, blue, red—the fish is reacting to the brightest by con- traction and to the darkest by expansion. This is contrary to all other experiments on the reaction of pigment cells. Frisch also finds that Phoxinus laevis adapts itself to green, blue, and violet by the contraction of its pigment cells which produces a lighter color, and to red and yellow by expansion of the red and yellow pigment cells; the color patterns remaining unchanged for months under constant stimulation. The color markings , ASSOCIATION AND COLOR DISCRIMINATION 447 of the skin conform to the general background rather than to some particular part. In males the reactions are more pro- nounced than in females. Frisch concludes that the stimulus is received through the eye, since blinded fishes show no adap- tive changes. Sumner (’11) states that the skin of the flatfishes, Rhom- boidichthys podas, Phombas laevis, and Lophopsetta maculata, shows adaptation in pattern, shade, and color. They react to black, brown, and gray, but not to red and yellow. Such changes take place irrespective of the intensity of illumination. Mast (714) finds that the flounders, Paralichthys and Ancylop- setta, simulate their background in shade, color, and pattern, exhibiting a remarkable ability to mimic blue, green, yellow, orange, pink, and brown backgrounds. Production of color changes is regulated by stimulation through the eye and de- pends upon the length of the light waves. That this indicates color vision is supported by the fact that flounders adapted to blue and green, when allowed a choice of backgrounds of different colors, prefer the background with which they harmonize in shade and color. _In his observations on coral-reef fishes, Longley (14, ’15, 717) finds that color changes in the pigmentation of the skin are com- mon among even the most brightly tinted fishes, and that the colors have a tendency to resemble those of the surroundings. In general, these observations seem to indicate that fishes of various genera exhibit adaptive reactions in their pigment cells to backgrounds of various colors, that the stimulus is received by the eye and transmitted through the central nervous system. Mating colors. Bright colors, particularly reds and yellows, appear on the ventral side of the males of many species at the time of spawning. The amount of light at the place of spawning must be sufficient for the female to perceive the colors if they have any recognition value. Frisch (12) cites examples of fishes possessing such colors and spawning in shallow water by day light and also of fishes spawning in deep water or at night which do not exhibit decorative coloration. 448 GERTRUDE MAREAN WHITE Warning coloration. ‘There seems to be no valid case of warn- ing coloration in any animal fed upon by fishes. Reighard (08) tried to discover whether the conspicuousness of coral-reef fishes might not have this significance. He found that the gray snapper could be taught to avoid snapping at red fishes when they were treated so.as to be unpalatable, but that when none were artificially treated the gray snapper devoured all species of coral-reef fishes with the same avidity. Longly (17) rejects the hypothesis of warning coloration as accounting for the bright colors of coral-reef fishes. Choice of colored lights and backgrounds. 'The method has been tried of illuminating different parts of an aquarium con- taining fishes with lights of different colors, or of placing va- riously colored papers or glasses under or around the aquarium. The fishes are allowed to choose the part of the tank they prefer. The chief difficulty with this type of experiment is that unless spectral light is used the colors are not pure, and it is extremely difficult to obtain lights of various wave-lengths which have the same luminosity. Pigmented papers have not been made which will reflect light of a single color. The first experiments relating to color discrimination were performed. by Graber (’84), who used glass slides of different colors. Such screens were not ‘pure,’ but allowed light of va- rious wave-lengths to pass through. Although the fishes Cob- itus barbatula and Alburnus spec. showed decided preferences for red, there is little to indicate that the choice was necessarily due to the length of the light waves. Bauer (’10) reports a difference between light and dark adapted fishes. Light adapted Charax puntazzo and Atherina hepsetus avoided a light shining through a red filter 680 to 710u, but when ‘dark adapted’ these fishes prefered red to blue, which Bauer considered to be of the same intensity. These observa- tions were interpreted to mean that fishes are able to recognize colors, but that when they are ‘dark adapted’ color perception of the red end of the spectrum ceases much sooner than for nor- mal human eyes. ASSOCIATION AND COLOR DISCRIMINATION 449 Goldsmith (12) arranged aquaria with colored bottoms. Young plaice preferred red to any other color and came to rest there, while gobies avoided red, refusing to go through a red passageway until it had been sanded. Red, yellow, green, and blue were chosen in order by Gasterosteus. This author thinks that her results show a color sense, but the possibility that the fishes were reacting to brightness does not seem to be eliminated. Feeding experiments. In such experiments a motive is intro- duced. Discrimination on the part of the fish is indicated by an attempt to take food. Zolotnitzky (01) placed pieces of wool of different colors having the shape and size of chironomid larvae on the wall of an aquarium containing Macropodes. The fish snapped often at the red wool, less frequently at the yellow, and left the other colors untouched. Frisch (714) placed bits of food on colored paper of various shades. Fishes that had been accustomed to eat yellow food snapped at it in preference to other colors and red food was taken by fishes trained to eat red. Red food was taken on a black background, but black food was not taken from a black surface nor gray from a gray background. Red and yellow were often confused with each other and with purplish red, but not with gray, green, or blue. Minkiewicz (12) taught a Julis vulgaris to seek food dropped into the tank through a blue tube and at the same time to dis- regard a piece of thread dropped in through a yellow tube. Goldsmith (12) reports that plaice and gobies learned to take food from colored forceps thrust into the water, and later chose the color which had been associated with food even when the relative position of the forceps was changed. The fishes per- sisted in examining the same forceps after an interval of four days even when no food was present. Washburn and Bently (06) were able to establish an associ- ation involving ‘color’ discrimination in the creek club. This fish was fed from forceps to which red sticks were attached. When a similar pair of forceps attached to green sticks was offered simultaneously, the fish preferred the red forceps, even when 450 GERTRUDE MAREAN WHITE neither fork contained food. The probability that the dis- crimination was based upon luminosity was said to be lessened by using a much lighter red in some of the tests. Blue forceps were distinguished from red in the same way. Food was later placed in the green forceps, and the fish learned to go to the ereen first. According to Reighard (’08), the gray snapper can distinguish colors. White Atherinas were taken in preference to those that had been stained blue. Though blue, light green, and dark green Atherinas were taken indiscriminately, red Atherinas were snapped at least often. Red Atherinas with the tentacles of the medusa, Cassiopea xamachana, sewed into their mouths, were refused, and very soon all red Atherinas were refused whether tentacled or not. This association of unpalatability with red lasted without any further practice from July 19 to August 8. The feeding experiments taken as a whole offer strong evi- dence of a color sense in the fishes observed. The same criticism may be made of all of them, however; pure colors were not used, and the possibility exists that the fishes were reacting to brightness. Hess (’09, 712, °13) is the chief opponent of color vision in fishes, maintaining that fishes see colors only as shades of gray— as a totally color-blind person perceives them. He offers the following reasons for this view: In spectral light young Atherina hepsetus and young Squalus cephalus congregated in the yellow-green and green, 1.e., in the brightest part of the spectrum of the dark adapted totally color- blind human eye. Some were found in the green-blue, very few in the red. By moving a black card along the spectrum and intercepting certain rays, the fishes could be driven into the blue or even into the violet, though they always returned when the card was removed. The longer light waves in the red pro- duced no more effect upon the fishes than complete darkness. The luminosity values of different parts of the spectrum were worked out. This was done by lighting one-half of the tank with monochromatic light and the other with equivalent mixed ASSOCIATION AND COLOR DISCRIMINATION 451 light which could be varied in intensity. When the fishes were arranged evenly in both parts of the tank, Hess considered that the two lights were equally bright. By measuring the intensity of white lights which had the same effect on the fishes as various colored lights, he obtained a luminosity curve which agreed very well with that of the totally color-blind human eye. Hess offered the fishes imitation baits of various colors on different backgrounds. If the brightness of the bait corre- sponded with that of the background, it was not taken. Contrary to the observations of Frisch, Hess found no adap- tation to brightness in the pigment cells of the epithelium of Crenilabrus. For fishes living any distance below the surface, Hess insists that a color sense would be useless because of the increasing absorption of light by the water as greater depth is reached, particularly the longer waves. Hence mating colors are value- less. Hess admits the possibility, however, that fishes living in shallow water may possess the ability to discriminate colors. Hess also attempts to account for the reactions of Daphnia and some other animals by assuming that they react to the lumi- nosity of the spectrum as it appears to the color-blind human eye. Loeb and Wasteneys (’15) criticise this assumption on the ground that there is no proof that the heliotropic effects of light in lower animals are accompanied or determined by sen- sations of brightness, and, furthermore, that color-blind human beings do not show any positive heliotropism. Hess’ further contention that animals and plants are sensitive to different parts of the spectrum—all animals to the yellowish-green, and all plants to the blue—is shown to be incorrect. From the sources discussed considerable evidence has been accumulated to show that fishes perceive colors. Adaptive changes in the pigment cells of the skin of various species in re- lation to backgrounds of different colors have been observed, and when allowed a choice, fishes show preferences for back- grounds of particular colors. Mating colors may furnish further evidence of color discrimination. Thus far there seems to be no valid case of warning coloration in animals serving as food 452 GERTRUDE MAREAN WHITE for fishes. The results of feeding experiments are strongly in- dicative of a color sense. On the other hand, Hess contends that fishes see colors as shades of gray, as a totally color-blind human being perceives them. CONDITIONS OF EXPERIMENTATION The mudminnows and sticklebacks used in the experiments to be described in this paper were obtained from Lake Wingra and Lake Mendota, near Madison, Wisconsin. They were not kept in running water, since it was found that they thrived equally well without it, provided the water was changed from one to three times a week. City water was used because -t is drawn from an artesian well and contains few organisms inju- rions to fishes. Individuals were kept in separate jars and care- fully observed, as many experiments as possible being per- _ formed upon each fish. One mudminnow which is still alive has been under observation over three years. When first introduced into aquaria the fishes were left undis- turbed for something like a week, except for changing the water. They commonly refused to eat during this period of adjustment ; usually some died at first, but thereafter few were lost. By the end of a week the survivors were likely to be hungry and suff- ciently accustomed to the presence of persons in the room to eat bits of food dropped into the water. Shortly after this they could generally be induced to take food at the surface, and after several days to jump out of the water for food. It was neces- sary to estimate carefully the amount of food to be given, as the ~ fishes have a tendency to consume large quantities at one time and then fast for several days. In most cases feeding was car- ried on in the laboratory from one to three months before ex- periments were begun, or until the fishes had formed the habit of taking food daily. Although the following observations were made especially on the mudminnow, they apparently apply to the stickleback as well. In making selection for training experiments, marked differences in the adaptability of fishes of different ages were found, Whereas mature mudminnows spend most of the time ASSOCIATION AND COLOR DISCRIMINATION 453 lying quietly in the bottom of the aquarium and are wary about coming to the surface for food, the younger fishes swim about actively, are less easily disturbed by jars and movements, and take food more eagerly. But fry under one and a half inches in length are undesirable for such work, as their movements are irregular and unsteady. INSTINCTIVE MOVEMENTS Instinctive movements may be defined as locomotor responses exhibited by an animal without previous training. Taken as a whole, these responses make up what Jennings has defined as the ‘action’ system, and they nearly always determine how any animal is going to react under a given set of conditions. The mudminnow and the stickleback have the same types of in- stinctive movements, namely, swimming, leaping out of water, and flopping about on land. The mudminnow swims rather deliberately with a smooth motion. When accustomed to lab- oratory conditions, it sometimes springs up and seizes objects outside of the water. Before doing so, it usually hesitates a short distance from the surface, switching its tail in an agitated manner. If the tank containing the mudminnows is left un- covered, the fishes are liable to leap entirely out of the recepta- cle.| Mudminnows bury themselves in the mud to tide over dry seasons. The stickleback swims in a jerky, nervous manner, never going _far in one direction, but darting hither and thither. It leaps out of water, but usually not so far as the mudminnow; generally it comes to the surface and bobs up and down, thrusting out only its nose. It is less timid, though always wary. PERIODIC ACTIVITIES AND DAILY RHYTHM OF THE MUDMINNOW Seasonal variations in the activities of fishes are not readily subject to laboratory observation. In almost all cases, however, the best results in the training experiments to be described were 1 It is reported by Mast (’15) that Fundulus leaps from tide pools and succeeds in transporting itself by flopping along on land across a barrier of dry ground more than 3 m. wide and 10 em. high. 454 GERTRUDE MAREAN WHITE obtained in July, August, January, February, March, and April, since during the summer and winter months the fishes were less restless and came for food more readily. That better results were not always recorded for experiments during the fall might be due to the fact that fishes were usually obtained in Sep- tember and early October and were not yet adjusted to labora- tory conditions. In the middle of April, May, and early June a marked rest- lessness was noticed—often the fishes ate erratically and were unfit for experimental work. It hardly seems probable that this was due to a rise in temperature, as the fishes had been kept in a room at ordinary temperature during the winter, and a series of experiments had been carried on successfully during the hot- test weeks of the summer. It was probably due to breeding activities. The breeding season of the mudminnow comes dur- ing the spring after the ice leaves the creeks and ponds where they live. The fishes which have been confined in the laboratory during the winter rarely, if ever, mature eggs, but there is prob- ably a change in the gonads at this season. Two fishes which had been accustomed to eat daily from forceps during the winter months became restless in the spring, eating very irregularly, but in July and August were again available for experiment. The mudminnow was the subject of a series of observations: on daily rhythm of activity. Across one corner of a room was hung a large opaque window shade. In this a hole was cut three inches square and about three feet from the floor. A table was placed in front of the curtain in such a position that vessels containing fishes could be seen by the observer who was seated behind the curtain. The space behind the curtain was dark- ened, and the observer kept as quiet as possible so that her movements might not disturb the fishes. The vessels con- taining the fishes were always allowed to remain on the table several hours before observations were made, During the hours of the day and night, the movements of individual fishes within the tanks were carefully mapped out on paper with as much minuteness as possible, and the amount of time spent in each position recorded with a stop-watch. ASSOCIATION AND COLOR DISCRIMINATION 455 Twenty fishes (all young fry—1# to 1{ inches—except one) were observed during an entire night in November from 10:30 p.M. until 7:30 a.m. The vessels containing them were placed on tables where there was only enough illumination to faintly discern the outlines of the fishes. The position of each fish was noted every half hour, also whether it was active or at rest. Of 201 observations made, the fishes were in 124 instances near the surface, and in 77 instances the fishes rested on or near the bottom. At daybreak nearly all the fishes were moving about actively. Likewise, nineteen fishes were observed at half-hour intervals during one afternoon and evening. From all these observations it was concluded that mudminnows, while only slightly less active at night than during the day, exhibit some- what greater activity at daybreak. Although very young in- dividuals came to rest anywhere in the dish, they spent rela- tively more time at the surface than on the bottom, whereas older individuals evinced a preference for the botton. Protec- tive adaptation would seem to cause the young fry to stay at the surface, since the large fishes which are their natural enemies are usually found near the bottom. SENSE ORGANS USED IN SEEKING FOOD In these fishes the senses of sight and smell are most used in seeking food. The stickleback displays more alertness in using both senses and much higher degree of acuteness of the olfactory sense. The method used by Parker (’11) in testing the olfactory sense of fishes was tried with both mudminnows and stickle- backs. Cloth packets, one of which contained meat and the other cotton, were suspended at opposite ends of the aquarium. The mudminnows did not show that they perceived either packet though they swam in close proximity to both. The sticklebacks behaved differently. The appearance of the packets attracted them at once. Those fishes which went to- wards the packet containing meat darted furiously upon it and pulled at it with great excitement, but those which swam in the direction of the packet of cotton in most cases stopped about 4 em. away and turned off sharply in another direction. Only 456 GERTRUDE MAREAN WHITE once or twice did they actually snap at the cotton packet. Then perceiving the struggles of the rest of the fishes with the other packet, they swam over and joined them. Repetition of this experiment gave similar results. At Woods Hole, Massachu- setts, the same test was performed upon Fundulus heteroclitus for comparison, since this species had been found by Parker to discriminate between the packets. The sticklebacks reacted fully as well to the stimulus of concealed food as did Fundulus. In the use of the sense of sight the mudminnow compares more favorably with the stickleback, though the latter reacts more quickly. Both species pursue moving objects without edor, such as bits of paper, or objects above the surface of the water; both exhibit skioptic reactions, and are stimulated by an increase in the amount of illumination. COLOR DISCRIMINATION Notwithstanding numerous investigations, the question as to whether fishes possess color vision is still somewhat unsettled. Fishes live in a medium where, except at the surface, the light is rather dim. The permeability of water to lght depends largely upon its depth and clearness. There is an unequal ab- sorption of different parts of the spectrum as rays of light pene- trate; the red end is absorbed more rapidly than the blue, so that at a depth of several meters in clear water fishes see as if through a blue-green glass. This fact might lead one to the view that for fishes living at any great depth, the ability to dis- criminate colors would be valueless. There is the possibility, however, that the eye of such fishes is better adapted to perceive minute differences in shade and color of objects in a dim light than is our own—a consideration which should not be rejected because fishes are rather unspecialized in structure compared with other vertebrates. Even though fishes, which live at a depth where red, orange, and yellow rays of light do not penetrate, may have no use for a color sense, this cannot be true of fishes living and feeding at the surface, where rays of light are only slightly refracted and absorbed by the water. In dealing with this problem there are certain questions to be considered: ASSOCIATION AND COLOR DISCRIMINATION 457 1. Do the fishes studied discriminate colors? 2. If so, is the discrimination due to wave-length or to in- tensity? In other words, do fishes see colors as such, or as shades of gray? 3. If the eyes of fishes are affected by differences in wave- length, is their color vision like that of a normal human being? 4. Can fishes form associations with colors? 5. Would such associations be of value to them in their struggle for existence? 6. How do the results agree with present theories of color vision? The following series of experiments were aimed to answer these questions and to supplement the evidence furnished by other workers. They were planned with a view to training fishes to secure food in particular ways, and extended through three years. The general problem presented to the fishes was that of learning to associate food with a certain color and at the same time asso- ciate unpalatable substances such as paper with another color. The mudminnow and the stickleback are both shallow-water fishes which discover their food largely by sense of sight. Va- rious methods of presenting food were tried. Considerable time was consumed at the beginning by using colored electrodes. These were thrust into the water at the sides of the dish and food was offered on forceps in the water. The fishes were given a mild electric shock when they snapped at bait on the wrong color. This method had to be abandoned because one shock often caused a fish to refuse food for several days. Instead of this, the fishes were fed on only one color, while on the color with which unpalatability was to be associated, they were offered balls of paper closely matching the food in appearance and color. In order that there might be no chance to smell the food, the bait was not dropped into the water, but the fishes were taught to leap out of water regularly and take it from forceps. Minced snail meat was the most attractive bait in the long run, although this was varied at times with chopped earthworms, slugs, and liver. Repeated trials deter- . THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, No. 4 458 GERTRUDE MAREAN WHITE mined that the fishes were not able to distinguish between the imitation baits and the food when both were offered out of water under the same conditions. The use of this method usually caused some delay on account of the necessary training of a week or two at the beginning of a set of experiments. The proportion of mudminnows successfully trained was only about two out of every five, but nearly all the sticklebacks could be taught to take food as desired. EXPERIMENTS WITH MUDMINNOWS Experiments with colored papers _ The first set of experiments with mudminnows was carried on during the winter of 1914 to 1915. Five were placed in sep- arate bowls to be trained, but only two of these could be induced to eat with regularity. The experiments were performed by day before a large window. Discs of colored? papers were cut 7.3 cm. in diameter and stiffened with cardboard. An aperture was made in the center of each large enough to allow the discs to be slipped down over the ends of the forceps from which the fishes were fed. The fishes were at first somewhat disturbed by the paper discs and required a little time to become accustomed to them. In a week this timidity vanished; the appearance of a colored disc became a signal for the fishes to dart to the surface and spring out of the water after food. When this association with one color seemed to be thoroughly established, a disc of another color was substituted with paper closely resembling the food in color and appearance in the forceps. The same pair of forceps was never used for both food and paper so that there should be no possibility that the taste of food might be trans-. ferred to the paper. The colored discs slipped down over the forceps were offered alternately. This furnished a severer test of the power of association than did the experiments of Wash- burn and Bently (’06) in which the forceps of two different colors with only one holding food were thrust into the water -simultaneously. | 2 The colors correspond with the following numbers in Klingsiek and Valette’s Code de Coleur: red: rouge no. 2, blue: blue no. 431, violet: blue-violet, no. 481. 459 ASSOCIATION AND COLOR DISCRIMINATION ‘9 [dy 03 9z Areniqa,y ‘sep [Ff “yuouttsedxe oy Jo uoyeing ‘1eded poyvorpur MoyjaA pu’ poo; Jo oouoseid poyvorpul uoods ‘syYST] MOT]OA puv UVEIH ‘Q9 ‘OU MOUUTUUpN “4 “LT ABI OF & Youvyy ‘sAup Gy “yuowTIedxo oy4 jo uolving ‘“todud pozvorpul yoTOIA puv poos Jo souoseid poyworpul onjq ‘Sspxvod yo[OIA pue onjg “ZZ ‘ou MOUUTUIpN ‘C JUST] WIp G SUTATS pousyvoM Sot10}}Vq “v—p “TT 4YSnsny 07 6 A[nE ‘sXvp FE “QuowtIodxo oy} Jo UoTYwanq ‘“aodud pozuosoadoa pos puv pooy pozyuosoddod ontq :7YSTTYsVy Vv JO puo oy} UT S19}[Y UTZB]Os oTyvMIOIYOOUOW po puv ONT_ “2Z “OU MOUUTUIPN, *O 1g Aceniqayq 04 TT Aveniqay ‘sAvp JT SVM JuowT1edxe oy} Jo UOTYBINq, ~“aodud AVIS PoPVOTPUL pod puUB POO}; Jo oouasoid pozvorpur onyjq ‘A]o}VUAO}[B poyuosoid spaxvo por puv onj_ “ZZ ‘ou MouUTUpNyY “g ‘Zo ABI 07 § Youvyy ‘sAvp G1 SUA JUOUITIOdxe ayy Jo uOTyRINq, ‘odd Avs poPVoIpul pot puv poo} Jo aouesoid poyVorpur onyq /A[Snoouvzjnus poyussead spavo pot puv onfg “EZ “ou MouUTUpNyY “Vy ‘poloyo JOU SEM POOF 4VYyy SoyRoTpUI oovds yuR]q VW “yk 0} pasnjor ynq ‘pooy poloyoO suM YSY vB 4BVY} SozVoIpUT oUT] po}}Op VW ‘CO 4B SI OUT] oY} Yoojaod st Aup ULBIL9N BV IOF PLOIII OY} USA “BSSTOSG” oY} UO YUoUITIodxe oY} Jo SAvpP oY} PUB o}vUIPIO oY} UO UMOYS st Avp VATSsodONS Yovo UO SLOI9 JO LOQGUINU OY], “SMOUUTWIPNUL JO UOTZVUIUITIOSTp 1OJOO UO syuoUTIedxe JO S}[NSet oY} Suryuosordoa soAIND, | “Sle 4 Ik £9 Lg og g 98 6é kt ST. 8 T | | i | | 1 ' i] i] ' : We RNS MAND R SOADeow | DANS : 460 GERTRUDE MAREAN WHITE The first test was successfully accomplished by only one fish, no. 27. The colored discs offered were blue as a signal that the forceps held food and red to represent paper. The forceps with the colored discs attached were presented singly. The duration of the experiment was seventeen days (Feburary 11 to February 27). The results are shown in figure 1, B. On the first day three errors were made, i.e., the fish snapped at the paper under the red disc three times. On the second and third days the fish refused to eat on red, but took food on the appearance of the blue disc. ‘Two mistakes were made on the fourth day, after which the fish no longer attempted to obtain food on red. In two instances a day intervened when the fish was not fed, in one instance two consecutive days, after which intervals the fish made a perfect record. The curves for all the experiments to be described are plotted in the same way. The days of the experiments are shown in the abscissa and the number of errors recorded éach day on the ordinate. When no mistakes were made the curve is at 0, but the number of correct tests is not designated in the curves. The results of all the experiments with the mudminnows and stick- lebacks are summarized in tables 1 and 3, respectively, showing the duration of each experiment, the total number of trials, the number of errors and of correct records, and the percentage of errors and of successes in each case. After mudminnow no. 27 had learned to distinguish between the red and the blue discs, violet was substituted for the blue disc. (Fig. 1, D, and table 1.) The entire experiment con- TABLE 1 Results of experiments in which mudminnows were taught to take food offered with certain colored papers or with light passed through monochromatic gelatin filters, and to refuse paper resembling food in appearance when offered with certain other colored papers or other monochromatic lights. ‘Lights from flashlight? means that the source of illumination was a hand electric flashlight and that the light passed through a monochromatic gelatin filter before reaching the fish; ‘lights dim’ means that the dry cell in the flashlight ran down so that the light had a low intensity. ‘Lights’ means that an incandescent electric lamp was used and that the light was passed through monochromatic gelatin filters. ‘Intensity varied’ means that differ- ent known intensities of light were passed through the monochromatic gelatin filters. ‘Colors reversed’ means that the color which in the first part of an experi- ment had indicated the presence of food was changed to mean paper, and that food was now given on the color with which paper had formerly been associated. The letters indicate the colors; blue 1s represented by B, green by G, red by R, violet by V and yellow by Y. ‘Gray plates’ means that ‘fogged’ photographic plates were used and that the source of the light shining through them was an incandescent lamp. “Square and dots’ means that, instead of monochromatic gelatin filters, glass plates with these designs on them were used with the light from an incandescent lamp passed through them. The appearance of the square was accompanied by food and the dots by paper. af ‘ ee = = 6 6 8 6 & 8 2 n FISH TYPE OF EXPERIMENT g z = 4 ce Sane g ei ma a 8 dh | 82 [G28] 22 | 028) 28 BEG ie are oi eae) Bees eae (MR andi@epapers. /.5 Sia ee 79 | 108 | 94 12 | 88.89} 11.11 || R and G lights from flash- No. 25 Hic hiiseeeemeere:,5c Ateeecioe 12 33 21 12 57.14) 42 83 Weights Gliese: ). ha eee 20 1 4.9 0: Mian chaede ease" 2s 0 oem ane 4.5 0 4.9 0: Minne hig, A sec). 22. oo $5 ote Pals 3 4.9 1 4.9 0: i, 4.9 1 4.9 0: INEST Ola Lig tel ERIE at ores Coc i 25 1 Misr te Gite cte.3 3.5) eee Pa) 2 4.9 0 SGI NE ee care 4.9 oF 4.9 0 IMioreln, Aad See Le di 4.9 2, P25) 1 Visine hts epee tert shee ets: See 4.9 2 2.5 0 IM (ental OS AR Pepe ee ceeeteie ctor 4.9 D 4.9 0) 4.9 1 4.9 O Mie chwilio e asses 2s { eA 1 f 4.9 1 4.9 0 Wier a 1a 2 eeeiacks sao oe \ 2 5 1 Manchel Sieg. arene ao 4.9 1 4.9 0 4.9 il 2.5 0 Marchi ae see ict cipecs ces: et 1 | Dae I 4.9 0 March: Lb Bent aactse cate 4.9 1 [ 1.4 1 25 1 4.9 0, WianchllGus- 2. 34 5seee eres 14 1 Mirenall(i...t00pe3.cte Jee ees yo) i 4.9 0 Virnchwe Sit 25.) Me oe) eee 4.9 0 4.9 0 Witainic lng (O eye) ens en sp ee 4.9 0 4.9 0 Marchese psn. ot «acta aoe 4.9 2 25 0 Wier chez eee. <: sony. toe 4.9 2 PAE) 0 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 27, NO. 4 , 474. GERTRUDE MAREAN WHITE light (2.5 e.m.) was frequently used with the yellow filter (4.9 c.m. being used for the green). The experiment lasted forty- one days (February 25 to April 6). On each of nine days one mistake was made, but errors were never recorded on two con- secutive days. The fish never made a perfect record for solong a period as did the fishes in the experiments previously described, but in two instances a perfect record was shown for five succes- sive days, and in two instances for four successive days. It seems safe to conclude that mudminnows are able to distinguish green from yellow, but that these colors are not so readily dis- criminated as are green and red, and blue and red (fig. 1, E). Experiments with gray filters In order to check up Hess’ theory that fishes perceive colors as shades of gray, as a totally color-blind human being perceives them, further experiments were performed. Two photographic plates were exposed to the light one second and four seconds, respectively, producing on one a light and on the other a dark gray surface. These plates were cut to the size of the gelatin filters and the gelatin surface protected by glass so that they could be inserted without scratching in the box shown in figures 4 and 5, and the experiment performed under the same condi- tions as the color tests. If the fishes were reacting to shades of eray, they ought to be able to distinguish between the two pho- tographic plates just as they had done with the colored filters, Mudminnow no. 60, which had distinguished green and yellow, did not show the slightest sign that it could perceive any differ- ence between the two photographic plates during fifty-one dasy (April 21 to June 10). It took food a little irregularly, as some- times happens in the spring, but the results were consistent throughout, since the fish attempted in the same manner to take whatever was offered with both plates (fig. 6, B, and table 1). Mudminnow no. 55, which had learned the red-green combi- nation, was also used with the two grays. The fish also gave no evidence of a perception of difference between the light and dark plates (fig. 6, A, and table 1). ASSOCIATION AND COLOR DISCRIMINATION 475 Summary of experiments vn regard to color vision of mudminnows 1. Mudminnows are able to distinguish between the following wave lengths of light: red, 6004 to 730u and green 510u to 570u; red 600u to 730u and blue 420u to-480u; yellow 580u to 630u, 660u to 710u, and green 510u to 570u, as is shown by the forma- tion of associations of paper and food with these colored lights. 2. Red and blue, and red and violet papers are distinguished in the same way. 3. Varying the relative intensities of the colored lights from 1.4 e.m. to 4.9 e.m. does not affect the result, which indicates that the reaction is to color rather than to intensity. 4, This conclusion is further supported by the fact that fishes (no. 55 and no. 60) which had previously shown that they per- ceived the difference between monochromatic gelatin filters (mudminnow no. 55 has shown a nearly perfect record with red and green, fig. 3, D) could not distinguish between photographic plates which had been ‘fogged’ to different shades of gray. EXPERIMENTS WITH STICKLEBACKS The tests applied to the sticklebacks were similar to those made upon the mudminnow. Experiments were performed using the electrical apparatus attached to the rheostat previously de- scribed (fig. 5). Two sets of monochromatic filters were pre- sented to the fishes; these were yellow no. 73 with blue no. 76, and red no. 71 with green no. 74. After repeated tests with the red and green filters, the light intensity was varied, as had been done with the mudminnows. Observations were also.made to test the discrimination of shades of gray. EXPLANATION OF FIGURES ON FOLDER Fig. 6 Curves representing the results of experiments on the discrimina- tion of light shining through photographie plates ‘fogged’ to different shades of gray. Construction of curves same as in figure 1. A. Mudminnow no. 55. Dark gray plate indicating presence of food and lighter gray plate indicating presence of paper. Duration of the experiment, 30 days, May 12 to June 10. B. Mudminnow no. 60. Dark gray plate indicated presence of food and light gray plate indicated presence of paper. Duration of the experiment, 51 days, April 21 to June 10. There was no day without one or more errors. 476 GERTRUDE MAREAN WHITE Fig. 7 Curves representing the results of experiments on the discrimination of colors and light intensities by sticklebacks. Construction of curves same as in preceding figures. A. Stickleback no. 65. Blue and yellow lights, blue indicated presence of food and yellow indicated paper. Duration of experiment, 18,days, April 28 to May 15. B. Stickleback no. 66. Blue and yellow lights, blue indicated presence of food and yellow indicated paper. Duration of the experiment, 16 days, April 26 to May 13. C. Stickleback no. 65. Green and red lights; green indicated presence of food and red indicated paper. Duration of the experiment, 21 days, May 16 to June 5. ; D. Stickleback no. 66. Green and red lights, green indicated presence of food and red indicated paper. Duration of the experiment, 17 days, May 16 to June 1. E. Stickleback no. 65. Light shining through photographie plates ‘fogged’ different shades of gray, the darker plate (exposed to light 4 seconds) indicated presence of food and the lighter plate (exposed to light 1 second) indicated paper. Duration of the experiment, 32 days, March 27 to April 27. F. Stickleback no. 66. Light shining through photographic plates ‘fogged’ different shades of gray, the darker plate (exposed to light 4 seconds) indicated presence of food and the lighter plate (exposed to light 1 second) indicated paper. Duration of the experiment, 32 days, March 27 to April 27. Fig. 8 Curves representing the results of experiments on the discrimination of colored lights by sticklebacks. Construction of curves and lettering same as in figure 1. A. Stickleback no. 46. Green and red lights. During the first 34 days (b-c) green indicated presence of food and red indicated paper; during the remaining 56 days, red indicated presence of food and green indicated paper. The duration of the entire experiment was 90 days, December 5 to April 3. B. Stickleback no. 50. Green and red lights. During the first 32 days (b-c) green indicated presence of food and red indicated paper; during the remaining 92 days (c-d) red indicated food and green indicated paper. During the last 20 days the lights were varied in intensity (v-v). Duration of the entire ex- periment, 124 days, January 6 to May 8. Fig. 9 Curve representing the results of experiment on the discrimination of patterns by mudminnow no. 27. Construction of curves same as in preceding figures. A square of black paper 2.7 em. indicated presence of food and four black dots 1 em. in diameter indicated paper: Duration of the experiment, 38 days, February 27 to April 5. Fig. 10 Curves representing the results of experiments on the discrimina- tion of calves’ liver and gray paper offered alternately to stickleback no. 57. As in figure 1, only the errors of each day’s record are shown in the curves. A. Duration of the experiment, 35 days, October 24 to November 28. B. Repetition of experiment represented by curve A. This was begun 44 days after the first experiment had been discontinued. Duration of the experi- ment, 30 days, January 12 to February 10. vd iS gy | ! [o>) es I iS} | | | ee — | | : | > wx \ iz = | “a il | oe i> | 3 | 11 | 3 | | | bead | i ‘SNS VAS eT) { Nx a | Ped ! } : i | KJ © : | 3 | | 5 | | | > 5 | : | | 2 I | I i> eel Ss ! r=) | | of | > ! I BY & f 0 I | 4 | | | Cs] | | | io bel “BIAS YOANN S XS AVS IIS CS) Ve o2hKOowWHANS AAA eSoOM Slo Wess x Q S i) ‘S) By >) i) ap - ssi fe = <7 _ = oa . ae .. _ > — i a 5 NO cme = NER A me ati 7 . ASSOCIATION AND COLOR DISCRIMINATION 481 Experiments with monochromatic blue and: yellow filters Tests with blue and yellow filters were applied to two stickle- backs, no. 65 and no. 66. The lights were flashed alternately upon the fishes, the blue being associated with food and the yellow with paper. ‘The yellow light seemed to the human eye considerably brighter than the blue, and the blue plate was much more opaque than the green plate which was used in combination with the yellow in the test made upon mudminnow no. 60. If intensity of light was the determining factor in these discrimi- nations by the fishes, the yellow-blue combination offered the greatest contrast of any experiment tried and might have been expected to secure the most positive results. But, strange to say, the records of this experiment fail to show any clear evi- dence of discrimination. Only 22.22 per cent of the total number of trials proved to be correct for stickleback no. 65, and 16.12 per cent in the case of stickleback no. 66. It was demonstrated in a later experiment that the failure was not due to the complex- ity of the test since these fishes learned to discriminate red and green lights (fig. 7, A, B; table 3). Experiments with monochromatic red and green filters Red and green lights were presented alternately to four sticklebacks including the two which had given negative results in the experiment with blue and yellow lights just described. The experiments on no. 65 and no. 66 were among the last tests made, but an account will be given of them here so that these results may be compared with the yellow-blue experiment. The fishes very soon demonstrated that they detected qualitative differences in the green (no. 74) and red (no. 71) lights which were flashed upon them, reacting differently to each light. After several trials the fishes exhibited hesitation in approach- ing the forceps in red light or refused to do so altogether, while in every case they snapped at the forceps in green illumination. The test lasted only seventeen days for stickleback no. 66 and twenty days with stickleback no. 65, since it was the purpose of this experiment to observe whether they could discriminate 482 GERTRUDE MAREAN WHITE TABLE 3 Results of experiments in which sticklebacks were taught to take food offered with light passed through monochromatic gelatin filters, and to refuse paper resembling food in appearance, when offered with other monochromatic lights. ‘Lights,’ ‘colovs reversed,’ ‘intensity varied,’ ‘gray plates,’ have the same meaning as in table 1 and the colors are as before represented by B for blue, G for green, R for red, and Y for yellow. ‘Food and paper’ means that calves’ liver and gray paper quite different in appearance were offered alternately without colored illumination NUM- PER ‘ DURA | xuM- |BER OF| NUM- |CENT OF| PER FISH TYPE OF EXPERIMENT open [BER OF| TIMES |BER OF] TIMES |CENT OF | IMENT | TRIALS COR- |ERRORS COR- |ERRORS RECT RECT days [ (Geancleelicihitisenmemneeenenicn 34 66 54 12 | 81.82) 18.18 Nor46" 4) |) Colorsmeversed=.. 5) r ee -: 28 | 108 8 | 100 8.00} 92.00 | 28 34 24 10 | 70.59) 29.41 (Gr Binh IR) Wied MAS, oo eda oo sent eal) eo 58 45 13.) 77.59) 22741 20 83 15 68 | 18.07) 81.98 INora0” ~~ |(@olorsimeversedis a... 445. 4| 20 50 34 16 | 68.00} 32.00 [| 20 52 41 TI | Zeehststay) AL. 115} Imbtensitys varniedeesascse de eul) 25 91 84 bh \ 92232) e68 Grayaplatesae — aes kes. Slee 39 1 38 2.56) 97.44 Nor 65) 4) Biandeye lights... Walnie HSE" 27 DA 2920018 Vi dRvameGelrehitsis.. oot eS OE 22 1 Bel evel 2eenee (Gray platesmee, hve eaneee 32. | 34 | 00 | 34 | 00.Q0|100.00 NoW6G.4 | Bland Y, lightsite. +: -lscsineu 16 | 31 5 26) | 16-12) 83.118 liietuemd (Gye hitss Nios st eee meal AL/ lo skn 11 6 | 64.72] 35.28 No. 37 f HoodPand papel ene aac eo | 39 33 6 | 84.62) 15.38 aal Boodvand'paper: 241 ae 30 | 41 34 7, 82.93\ 70% between the colored lights, rather than whether a permanent association could be formed (fig. 7, C, D; table 3). Experiments of longer duration were performed on stickle- back no. 46 and stickleback no. 50. No. 46 was first trained to come for food in green light, and to inhibit the impulse when paper closely resembling the food in color was offered in red light. This test continued thirty-four days, January 5 to Febru- ary 7. On only seven days do the records show errors, and dur- ing sixteen successive days a perfect record was maintained. On the thirty-fifth day the colors were reversed. The permanency ASSOCIATION AND COLOR DISCRIMINATION 483 of the acquired association was demonstrated by the fact that during the twenty days following the fish persisted in attempting to take the paper under the green light which had previously shone upon its food. The reversal of the colors seemed to con- fuse the fish, and an errorless record with the new conditions was never shown as in the previous test. It required a much longer period to form the habit of taking food in red light and refusing paper in green light than to form the first association. The entire experiment lasted ninety days, January 5 to April 3 (fig. 8, A; table 3). . Stickleback no. 50 learned to react negatively to paper offered in red light and positively to food in green light in a series of tests continuing thirty-two days, at the end of which there was a perfect record of fifteen days. Stickleback no. 50 reacted in the same manner to the reversal of colors as had stickleback no. 46. The association of food with green light was overcome with great difficulty, but after seventy-one days of experiment had elapsed, the fish was tested on ten successive days without errors being made. Although the light intensity was varied for the last twenty-three days of the experiment, the oscillations of successes and failures which took place seemed not to be corre- lated with the relative intensities of the lights, and it may be noted in the tabulated results (table 2) that the highest percent- age of successes for the whole experiment was recorded during this period (fig. 8, B). Like the mudminnows, the sticklebacks tested showed that they were able to distinguish between red (600 to 730u) and green (510u to 570u), but the experiments indicated that with blue (420, to 480u) and yellow (580u to 6304, 6604 to 710.) there was no such discrimination. Experiment with an aquarium of sticklebacks An interesting piece of evidence was obtained from an aqua- rium containing fourteen sticklebacks. These fishes were kept under observation for several months, during which they were regularly fed and became very tame. Calves’ liver was given to 484 GERTRUDE MAREAN WHITE them nearly every day from forceps. It was very amusing to see all fourteen of them dart to the top at a slight movement of anyone near them and begin sticking their noses out of the water in anticipation of food. When food was held a slight distance out of water, they would with one accord leap out after it, and at times hang on so tightly that they could be lifted several inches out of the water before letting go their hold. On one occasion, after the sticklebacks had been given a small piece of calves’ liver, the forceps were held out to them empty. None of the sticklebacks approached the forceps, but the merest bit of dark red liver was sufficient to attract them, A bit of rather bright red paper rolled into a ball and substi- tuted for the food was at once attacked. Tan-yellow presented in the same way ellcited no positive response. Lavender which had a pinkish tinge was snapped at twice, while dark blue, gray, yellow, and green failed to attract. When dark red paper. which most closely resembled the color of the liver, appeared the fishes darted towards it from all directions seizing it voraciously. Reddish purple was snapped at several times. These papers were compared with Klingsiek and Valette’s Code de Coleur and were found to resemble most closely the following numbers: Dark red—rouge no. 3 Red—rouge no. 6 Tan-yellow—orange no. 146 Yellow—orange-jaune no. 166 Green—vert no. 301 Blue—bleu violet no. 452 Gray—bleu violet no. 460 Lavender—violet no. 528B Purple—violet rouge no. 571 This experiment indicates that the color of the food which sticklebacks take habitually makes an impression difficult to eradicate. Red no. 6 seemed to the eye to be very bright; red no. 3 was quite as dark as blue no. 452, and not so dark as gray no. 460 which was rejected. ASSOCIATION AND COLOR DISCRIMINATION 485 Experiments with gray filters Gray filters were presented alternately to the two sticklebacks no. 65 and no. 66 exactly as they had been offered to mudmin- nows no. 55 and no. 60. It may be noted that these were the same sticklebacks which had been subjected to the experiment with vellow and blue lights and to the experiment with red and green; both had manifested the power to discriminate between red and green. In contrast the experiment with the gray filters was very striking, for on no occasion does the record show a day without error for either fish. Neither of the curves touches the point at O at any time during the experiments (ig (oH, He table 3). Summary of results in regard to color vision of sticklebacks 1. Sticklebacks were not able to distinguish between blue and yellow lights of the following wave-lengths: 420u to 480u and 580x- to 630u and 660p to 710u. 2. Red 600u to 730u and green 510u to 570u were distinguished even when the relative intensities of these lights are varied from 1.4 cm. to 4.9 cm. 3. Photographic plates ‘fogged’ to a light and a dark gray were not distinguished. 4. Sticklebacks form decided associations respecting the color of the food which they habitually eat. GENERAL CONCLUSIONS ON COLOR VISION IN FISHES The experiments described in this paper and the work of other investigators furnish evidence that some species of fishes perceive differences in colors, and that this discrimination is based upon the wave-length of the light, not on intensity. It seems rather unlikely that the color vision of fishes approximates in character that of human beings. It would be of interest to know what range of colors fishes react to. The stickleback, at least, seems unable to discriminate between blue (420 to 480u) and yellow (580u to 630u, 660u to 710u). Whether these colors are con- fused by mudminnows also has not been tested. 486 GERTRUDE MAREAN WHITE Color perception seems to be of some importance in the lives of fishes, since color associations are formed and persist for a considerable time. Such associations may be formed even when the colors are not present simultaneously. The value of such associations to fishes which seek their food in shallow water would be obvious. Food of a particular color once found to be desirable may be singled out and pursued and undesirable sub- stances more easily avoided. It is somewhat premature to discuss at length the theory of color vision which the results of these experiments would seem to support. The duplicity theory of von Kries assumes that achromatic scotopic (dark-adapted) human vision is carried out through the mediation of the rods alone, the cones being the organ of photopic (light-adapted) vision. This conclusion ‘is derived from the fact that the fovea of the eye consists entirely of cones, while the extreme periphery contains only rods; the remainder of the retina had both rods and cones. The visual purple is located in the rods. For the light-adapted eye, the fovea is the point of keenest vision, and the brightest part of the spectrum is in the yellow. When the eye is dark-adapted, on the other hand, the fovea is no longer the seat of keenest vis- ion, but instead objects are seen more clearly when focused at points away from the center of the eye. Green appears to the dark-adapted eye as the brightest area of the spectrum and it is in green light that visual purple is most strongly bleached. The theory accounts very well for cases of total color-blindness where the fovea of, the eye is affected and the eye is unable to focus strongly on any object. In such instances bright light hinders vision. Night blindness might be explained on the basis of a defect in the rods’ interfering with vision in low illu- mination, but leaving vision in light of higher intensity unim- paired. The researches of Hess on the comparative physiology of vision in the lower animals, however, give little support to the duplicity theory, since the eyes of all classes of vertebrates show adaptation to light and darkness, even including tortoises which have neither rods nor visual purple. The results of the experi- ments with the sticklebacks described in this paper seem to have little or no bearing upon this theory. ASSOCIATION AND COLOR DISCRIMINATION 487 Hering bases his theory of opponent colors on the sensations of complementary colors. When certain wave-lengths of red and green are presented to the human eye simultaneously, the result is a colorless sensation. The simultaneous perception of yellow and blue of certain wave-lengths produces the same effect. These four color sensations, red, green, blue, and yellow, are regarded by Hering as the primary ones from mixtures or dilutions of which all other color sensations are. derived. To these are added the primary sensations, black and white, which on mixture result in sensations of gray. This theory maintains that there exists in the eye three distinct visual sub- stances affected by pairs of antagonistic physiological processes in such a way as to produce the six primary sensations, white- black, red-green, and yellow-blue. In favor of this theory is claimed for consideration the color sensitivity of the part of the normal human retina lying outside of the fovea. The extreme periphery of the retina is color-blind. Within this totally color- blind zone lies an area which is red-green blind, but is sensitive to blue and yellow stimuli, while the center of the retina is sen- sitive to all colors. The most common form of color-blindness is that in which red and green are not discriminated, and cases are known in which yellow and blue are confused. Either type of color-blindness might be explained by the absence of one of the physiological substances. The same explanation might be applied to account for the failure of the sticklebacks in the ex- periments described to discriminate between yellow and blue. The observation that on mixture red, green, and blue produce white is the basis for the Young-Helmholtz theory, according to which there are three primary colors instead of four from which all other colors are derived. The wave-lengths of the red and the blue are closely identical with the fundamental colors chosen by the upholders of the four-color theory, while the green lies on the yellow side of the green of Hering’s theory. This theory explains very well certain cases of partial color-blindness by as- suming the absence or diminution of one of the three theoret- ical components. If this theory accounts for the inability of sticklebacks in these experiments to discriminate between blue ASS GERTRUDE MAREAN WHITE and yellow, the blue component must be assumed to be lacking. These results seem to be explained equally well according to the four-color theory of Hering and the three-color theory of Young- Helmholtz. EXPERIMENTS IN THE DISCRIMINATION OF PATTERNS Mast (14) and Sumner (’11) have shown that certain flat fishes adapt the color markings of their skin to the background against which they rest. Such an adaptation would be frequently of protective value. According to Mast, such stimulation is re- ceived through the eye. One naturally inquires whether rec- ognition of differences in the configuration of their environment is of use to fishes in seeking their food. Experiments with a view to ascertaining whether this is the case were carried out on the stickleback and the mudminnow. Mudminnows no. 27 and no. 40 which had given positive re- sults in the color experiments were tested in several ways to discover whether they could form associations of food and paper with various patterns. On the center of a round piece of card- board 7 cm. in diameter was pasted a five-pointed star of black paper. A ring of black paper 5.3 em. in diameter and 1 cm. wide was pasted to a similar white card. When the star was held above the tank containing the fish, forceps containing food were presented, while the circle signified gray paper which matched the food. As in the color tests, the fishes were made to leap out of water to obtain the bait. Mudminnow no. 40 when sub- jected to this test exhibited during twelve days no signs of form- ing an association with the patterns. Dots and stripes were the test next applied. Black stripes 1 cm. wide were pasted on white cardboard 1 em. apart. On another white cardboard were pasted black dots 1.8 cm. in di- ameter and the same distance apart. Except for the designs on the cards the test was in all respects similar to the previous one, and the results were similar, for mudminnows no. 27 and no. 40 to which it was applied attempted to take food in the same manner when either card was held above them. ASSOCIATION AND COLOR DISCRIMINATION A489 A eirele of black paper 4.5 cm. in diameter pasted on a white card was not distinguished from a black square 4 cm. across, when they were presented under the same conditions as those used in the tests described above, by mudminnow no. 39 during thirty-seven days. In order as nearly as possible to duplicate the conditions of the experiments with colored lights, two pieces of glass were cut the size of the gelatin filters. Upon the center of one was placed a square of black paper 2.7 cm. across and upon the other four black circles 1 em. in diameter. These plates of glass were in- serted in the tin box and the light flashed through them upon the fishes. The appearence of the square was accompanied by food and the black dots were to suggest paper. At the end of thirty-eight days (February 27 to April 5) mudminnow no. 27 showed no signs of discrimination (fig. 9). Two sticklebacks, no. 63 and no. 64, were subjected to the following experiment. Glass plates fitting into the tin box were used. On one was a square 2 cm. in size and on the other a black circle 2.5 em. in diameter. Neither fish showed that it perceived the difference between them, though the experiment continued sixty days. While these experiments do not absolutely prove that differ- ences in pattern are not perceived by mudminnows and stickle- backs, they suggest that the discrimination of patterns and differences in backgrounds does not have a very important function in their search for food. No associations appeared to be formed with the patterns used. These results are in sharp contrast with those of the tests with colors. ASSOCIATIONS FORMED IN FISHES Types of associations The analysis of the psychology of a fish, like that of any other animal, can only be made by interpreting its immediate reactions to stimuli of various sorts. The nature of the sense organs and the reacting organism must determine the types of associations formed, whether they are merely associations of particular 490 GERTRUDE MAREAN WHITE muscular reactions with special stimuli without apparent con- sciousness of their purpose on the part of the animal, or psy- chological associations of a higher type. The formation of associations with color stimuli has already been shown to exist, but under similar conditions associations with patterns were not formed in the tests made. Associations with objects were found to occur. A large live dobson-larva. was dropped into a tank contaming a mudminnow. At first the larva was repeatedly attacked, but when, after a dozen or more trials, the fish was unable to devour it, the larva was completely ignored, although living Crustacea, worms, and other baits were taken. At other times forceps containing no food attracted mudminnows. Moving objects and shadows nearly always induced reaction. Sudden movements caused the fishes to swim about rapidly as if frightened and in search of cover. After the fishes had become accustomed to being fed at the top of the water, the approach of anyone caused the fishes to swim to the surface. Short, jerky. wriggling motions as those of a worm attracted and agitated them. Jarring the tank produced the same result, probably owing to the fact that it was customary to lift the vessel containing a fish about to be tested and set it on the front of the table. While moving a receptacle, it was necessary to keep it covered because the fishes were likely to leap out of the water, sometimes land- ing outside of the vessel. Fishes freshly brought into the lab- oratory or those which had not been experimented upon could be carried about in small vessels without showing any inclination to jump out. 4Proventricular ZZ Proventriculus PLATE 2 PLATE 3 EXPLANATION OF FIGURES 9 A natural tumor in proventriculus ganglion. 10 Pigmented bar on the ventral surface of the last abdominal segment of a larva with a tumor. 11 Section through same, showing pigment in hypodermal cells. 12 Same, showing relation to the other tissues and the tumor. on 2) o iS AN HEREDITARY TUMOR PLATE MARY B. STARK EO wey ronney Deagege care et Ld —_ yee ESOT COE gy Se UUEC Mn y ye Se ate are ene i aera | 929 INDEX BBOTT, Cuartes HARLAN. Reactions of land isopods to light................- Acclimatization as a factor affecting the upper thermal death points of organisms....... Adrenin on the pigment migration in the melanophores of the skin and in the pig- ment cells of the retina of the frog. The GikHiniel ian ig oO ROE CSo Daa oD ataaoe Anabiosis of the earthworm.............----- Association and color discrimination in mud- minnows and sticklebacks................ Assortive mating in a nudibranch Chromo- doris)zebra, HHeilprin'-. sacle citer cer IGNEY, AnpDrew Jounson. The effect of adrenin on the pigment migration in the melanophores of the skin and in the pig- ment cells of the retina of the frog........ ALKINS, Gary N. Uroleptus mobilis, ( Engelm. I. History of the nuclei dur- ing division and conjugation........... Cells of the retina of the frog. The effect of adrenin on the pigment migration in the melanophores of the skin and in the pig- IN pra O SUBD dC OUE & COMI OO Geto Character and selection. Fluctuations in a TECeEsSive) Mend elianencss-eeem as «< -eeiiaes Chromodoris zebra Heilprin. Assortive mat- hae rhebs anol ore so55 po oonmoeeeancocanS Coxurns, H. H. Studies of normal moult and of artificially induced regeneration of pelage in Peromy SCUSHERE eee ect Color discrimination in mudminnows and sticklebacks. Association and............ Comparison of the thyroid glands of iodin-fed and normal frog larve., Studies on the relation of iodin to the thyroid. II...... Conjugation. Uroleptus mobilis, Engelm. I. History of the nuclei during division Corneal epithelium. Demonstration of epi- thelial movement by the use of vital staining, with observations on phagocy- TOSTSHMsb New eet wetter eee crete Sar Crozier, W. J. Assortive mating in a nudi- 193 443 247 391 293 391 157 247 73 443 417 branch Chromodoris zebra Heilprin..... 247 ——. Onthe use of the foot in some mollusks EATH points of organisms. Acclimati- zation as a factor affecting the upper Pheu ales roe tscr oe osten che vera ee alee 42 Development in mice. Observations on the relation between suckling and the rate of Gian] oy AON Ohya Hat cane Ome cee een tec Discrimination in mudminnows and stickle- backs. Association and color............. d Division and conjugation. Uroleptus mo- bilis, Engelm. I. History of the nuclei Gira tiVink: Gate een see rnsa oonmonterice Dyes), and to excretory toxins. The reac- tion of Selachii to injections of various non-toxic solutions and suspensions (in- cludinesyataleeencehercnaacceeeiorrec a. Meee) Anabiosis of the......... Embryonic development in mice. Obser- vations on the relation between suckling angithe:rateOlnsn. coer lee reeer eae: Engelm. I. History of the nuclei during division and conjugation. Uroleptus 1400) OT Re ao oie ie enn ge CO nico ns na ae Epithelial movement by the use of vital staining, with observations on phagocyto- sis in the corneal epithelium. Demon- Stration! Of 5) chekiy acc ac sresane ie» Seer ee 293 101 57 49 293 Excretory toxins. The reaction of Selachii to injections to various non-toxic solutions and suspensions (including vital dyes), SN GEGCOE. Mets Ate stare aes ote hic ae isialcitereatonets 101 LUCTUATIONS in a recessive Mende- lian character and selection............. 157 a Foot in some mollusks. On the use of the... 359 Frog larve. Studies on the relation of iodin to the thyroid. II. Comparison of the thyroid glands of iodin-fed and normal.. 417 ——. The effect of adrenin on the pigment migration in the melanophores of the skin and in the pigment cells of the retina ofithersas aes te seen rience tee 391 Ges of iodin-fed and normal frog larvae. Studies on the relation of iodin to the thyroid. II. Comparison of the GY TOMS Asacc seateionciamaoah manent ene 417 |e (carga grea inde YT osadnooooMons 507 Hoskins, E. R., AND Hoskins, M. M. The reaction of Selachii to injections of vari- ous non-toxic solutions and suspensions (including vital dyes), and to excretory CORINS Se See oy aE teas oe Tae ee 101 ——, M. M., Hoskins, E. R., ann. The re- action of Selachii to injections of various non-toxic solutions and suspensions (in- cluding vital dyes), and to excretory GORANI Ws pan sete c rola oladatneie eae tte areostcben tsetse: 101 NBREEDING. III. The effects of in- breeding, with selection, on the sex ratio of the albino rat. Studies on.............: 1 Injections of various non-toxic solutions and suspensions (including vital dyes), and to excretory toxins. The reaction of Selachii Le eR re ea ie eee tee ater act grtaelertreteren eta 101 Todin-fed and normal frog larvae. Studies on the relation of iodin to the thyroid. Il. Comparison of the thyroid glands of. 417 ——tothethyroid. I. The effects of feeding iodin to normal and thyroidectomized tadpoles. Studies on the relation of..... 397 —— to the thyroid. II. Comparison of the thyroid glands of iodin-fed and normal frog larvae. Studies on the relation of.. 417 Tsopods to light. Reactions of land......... 193 Fl (aeeeaeer M. H. Acclimatization as a fac- tor affecting the upper thermal death MOMts Of Organisms ines. ee see ee 427 ING, Heten Dean. Studies on in- breeding. III. The effects of inbreed- ing, with selection, on the sex ratio of thelalbino nsitera eee on eee eee 1 KirkKHAmM, Witt1AmM B. Observations on the relation between suckling and the rate of embryonic development in mice......... 49 ARVAE. Studies on the relation of iodin to the thyroid. II. Comparison of the thyroid glands of iodin-fed and normal LEG OS AU RIEMe iach nud Ses Be RAD Palco niet 417 Light in the colonial forms, Volvox glabator and Pandorina morum. Reversion in the SCNSELOLOLIENtatlOm Olas eee cee ccl sees 367 ——. Reactions of land isopods to.......... 193 Me S. O. Reversion in the sense of orientation to light in the colonial forms, Volvox globator and Pandorina IBOLUNG eee eee ene sap ae oer felsislabacstete 367 Mating in a nudibranch Chromodoris zebra Heilprine eASSOLtLVe nic mere. cee en errs le 247 532 Matsemoro, SxsrtnicHt. Demonstration of epithelial movement by the use of vital staining, with observations on phagocyto- sis in the corneal epithelium.............. 37 Melanophores of the skin and in the pigment cells of the retina of the frog. The effect of adrenin on the pigment migration in [HIVS) BS Ocha eae GAO aS nO nan ons ois 391 Mendelian character and selection. Fluctua- TIONSHIP ALLCCEASLV.eL ates eet penis oniettee none 157 Mice. Observations on the relation between suckling and the rate of embryonic devel- OPMeniAM! See, qoutes piso se hoe eee ae ae 49 Migration in the melanophores of the skin and in the pigment cells of the retina of the frog. The effect of adrenin on the pig- 391 during division and conjugation. Uro- leptsseritaiins ses vases tee BE SoLOgS OA OE 293 Mollusks. On the use of the foot insome.... 359 Moult and of artificially induced regenera- tion of pelage in Peromyscus. Studies of MOTTA ee ees Ae ie ee ee ee 73 Mudminnows and sticklebacks. Association and color discrimination in............... 443 UCLEI during division and conjugation. Uroleptus mobilis, Engelm. I. History Oletihe 4 ere ree ace eee ce ee OS Nudibranch Chromodoris zebra Heilprin. IASSOLLUVE Mahi esIT Meant «lee oceilas eon mee RGANISMS. Acclimatization as a factor affecting the upper thermal death WOMMULS LOLs Sees epee tan eee eee 4 Organization of Renilla. 4 Orientation to light in the colonial forms. Volvox globator and Pandorina morum. Reversion in the sense of................- 367 ANDORINA morum. Reversion in the sense of orientation to light in the colonial forms, Volvox globator and.... 367 Parker, G. H. The organization of Renilla. 499 Pelage in Peromyscus. Studies of normal moult and of artificially induced regen- erabionioiatenc.o tc ee one ee eae 73 Peromyscus. Studies of normal moult and of artificially induced regeneration of DELA Dem yoe te ne) oP Mee ae 73 Phagocytosis in the corneal epithelium. Demonstration of epithelial movement by the use of vital staining, with observa- CONS OU een ee eee occ, Set ae 37 Pigment migration in the melanophores of the skin and in the pigment cells of the retina of the frog. The effect of adrenin on the. AT. Studies on inbreeding. III. The effects of inbreeding, with selection, on the sex ratio of the albino.............. 1 Reaction of Selachii to injections of various non-toxic solutions and suspensions (in- cluding vital dyes), and to excretory tox- rbavejaoes 4 Be \cy, 5 Sapte Shae peer ae 5 ty See —=— of land ssopods;to light.. 328.702. °. ; Regeneration of pelage in Peromyscus. Studies of normal moult and of artificially INGUCEG mee racet aeons eee ee 73 telation of iodin to the thyroid. I. The ef- fects of feeding iodin to normal and thy- roidectomized tadpoles. Studies on the. 397 Comparison of the thyroid glands of iodin-fed and normal frog larvae. Studiesvonttheeuseep eas ee 417 Renilla. The organization of................. 499 Retina of the frog. The effect of adrenin on the pigment migration in the melano- phores of the skin and in the pigment cellsiohithee ate tsk ae ee ee 391 Reversion in the sense of orientation to light in the colonial forms, Volvox globator and Eandorinaanorimices ss eee nomen 367 247 INDEX Roserts, Ermer. Fluctuations in a reces- sive Mendelian character and selection... CHMIDT,Perrer. Anabiosis of the earth- NO) 160 Reon i ae ee on tide wo san Selachii to injections of various non-toxic solutions and suspensions (including vital dyes), and to excretory toxins. The re- ACtLOM Of hyn eee en eee Selection. Fluctuations in a recessive Men- deliant chanactenand eee nee ener —, on the sex ratio of the albino rat. Studies on inbreeding. III. The effects ofinbreeding with: = pee eee eee Sex ratio of the albino rat. Studies on in- breeding. III. The effects of inbreeding, withiselection. on the- one sseene salen Skin and in the pigment cells of the retina of the frog. The effect of adrenin on the pigment migration in the melanophores Staining, with observations on phagocytosis in the corneal epithelium. Demonstra- tion of epithelial movement by the use of ACHE a eno, 1) SERENE a eM PS Rn danteah oth Man Stark, Mary B. An hereditary tumor..... Sticklebacks. Association and color dis- crimination in mudminnows and........ Suckling and the rate of embryonic develop- ment in mice. Observations on the rela- tiom between:.6¢ 5... ene ero Erne SwinGue, W. W. Studies on the relation of iodin to the thyroid. I. The effects of feeding iodin to normal and thyroidecto- MuIZedhuacpoles: res oe eee eee bine Studies on the relation of iodin to the thyroid. Il. Comparison of the thyroid glands of iodin-fed and normal frog LAT VBONe cS c\crccros crac ere ea See ADPOLES. Studies on the relation of iodin to the thyroid. I. The effects. of feeding iodin to normal and thyroidec- COMIZEO 528 acs See ee Oe eee Thyroid. I. The effects of feeding iodin to normal and thyroidectomized tadpoles. Studies on the relation of iodin to the.... Il. Comparison of the thyroid glands of iodin-fed and normal frog larvae. Studies on the relation of iodin to the.... Thyroidectomized tadpoles. Studies on the relation of iodin to the thyroid. I. The effects of feeding iodin to normal and.... Toxins. The reaction of Selachii to injec- tions of various non-toxic solutions and suspensions (including vital dyes), and to EX CLOEOR Visita rece cue rae eran cie-= eheveuatoe tl meteors Tumors) An hereditary...0c720 +e set tees ROLEPTUS mobilis, Engelm. I. His- tory of the nuclei during division and ConjUugationse ae tere len case Renee ITAL dyes), and to exeretory toxins. The reaction of Selachii to injections of various non-toxic solutions and sus- pensions (neludine ese see eee eee —— staining, with observations on phagocy- tosis in the corneal epithelium. Demon- stration of epithelial movement by the usc 0 Volvox globator and Pandorina morum. Reversion in the sense of orientation to light in the colonial forms............... 3 HITE, Gerrrupe MAREAN. Associa- tion and color discrimination in mud- minnows and sticklebacks............ EBRA Heilprin. Assortive mating in a nudibranch Chromodoris............... 2 391 397 397 417 397 293 101 pot? 2 of CR ~— <<¥ ie Ae no *s " stetess* a ; e,' Seer, x ye A oh tas 3 a 2 Sas ates. ete teteste ay ete ; hated otelatate tetatate ttt h ; tte ° os BAAR ee boe4 Soe a eta tgsara tts, a“ fea atte ¥ a ey os eet ¥ ang % #44) he oe tart PASE ee ates gs