2A 4 Fie, teats 2 ; ~~ a4 eee, raed net, Pye wrt TE ot ae ce top et ES ; Senne ees oe fot . 0 oY ‘ > ; 7 4 7 4 ~ aed ee ~— sd ye ot LS Set ee ee See ee : - ee ES ee Ree he tpt y yey beta ‘aa the hglaty 1a oS f plana: oTigt | if i ; pete ‘ coe . : sartseet Stet, Pine eshte t Aare eS ete ee THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY WiuuiaM EF. CastLe Jacques Lors Harvard University The Rockefeller Institute Epmunp B. WILSON Columbia University Epwin G. CoNnkKLIN Princeton University Toomas H. Moraan CHARLES B. DAVENPORT é te Columbia University Carnegie Institution GEORGE H. PARKER HERBERT S. JENNINGS Harvard University Johns Hopkins University RAYMOND PEARL FRANK R. LILuIE Maine Agricultural University of Chicago Experiment Station and Ross G. HARRISON, Yale University Managing Editor VOLUME 26 1918 Mie 6 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. BV 5. 0 y “ ~i CONTENTS No. 1. MAY HeLen Dean Kina. Studies oninbreeding. I. The effects in inbreeding on the growth and variability in the body weight of the albino rat. Four- CASTES WCE) Oi eR, «con ice RS RS i a RP Se 8 re a a 1 P. W. Wurttne anp HeLen Dean Kina. Ruby-eyed dilute gray, a third allelomorph mi the albino series ‘of the rab. 2... 2)... 6.02 ese eae ee ee 55 M. F. Guyer anp E. A. Smirru. Studies on cytolysins. I. Some prenatal Slects of lenswaminnearese: © oot oats o. Gaon he eal Ls choses se mertore 65 Witiram A. Kepner AND ARNOLD Ricu. Reactions of the proboscis of Planaria albissima Vejdovsky. Ten figures...................-...+-.:: 83 S.Hatat. Onseveral effects of feeding small quantities of Sudan III to young aibino rats. ‘Wbree Charts. oo iee 4s 5... ss Lact eect bee ene dere Meee e gaa 101 Cuartes R. StockarD AND GEORGE N. Papanicoutaov. Further studies on the modification of the germ-cells in mammals: The effect of alcohol on treated guinea-pigs and their descendants. Nine tablesand nine figures .. 119 No. 2. JULY Oscar Rippie. A demonstration of the origin of two pairs of female identi- cal twins from two ova of high storage metabolism. Thirteen tables...... 227 B. W. Kunxeu. The effects of the ductless glands on the development of the flesh* lies... -. Sa ere te cece Sep SUG nA anes a oe ole 255 Mario Garcta-Banus. Is the theory of axial gradient in the regeneration of ALFRED C. REDFIELD. The physiology of the melanophores of the horned toad Phrynosoma. Eight text figures and five plates............. ay 745) Heten Dean Kina. Studies on inbreeding. II. The effects of inbreeding on the fertility and on the constitutional vigor of the albino rat. Two CHATUS..: 0. . » - eM itis oe soncals MAPPER Stee we.ge kia c aera Nese rote 335 W. J. Crozier. The amount of bottom material ingested by holothurians (Strehopis). . Migmamemt sees oar 2 ca kaya Steels, oss scomhe ali alee Gio es yet acess 379 No. 3. AUGUST H. H. Newman. Hybrids between Fundulus and mackerel. A study of paternal heredity in heterogenic hybrids. Four figures................ 391 W.C. Auer anv E. R. Stein, Jr. Light reactions and metabolism in May- fly nymphs. I. Reversals of phototaxis and the resistance to potassium cyanide. II. Reversals of phototaxis and carbon dioxide production. Hourncharts: ; : eee tice casts Nazaire SES I asap ee ees isi vaa tee ep It 423 1V CONTENTS ALVALYN EK. Woopwarp. Studies on the physiological significance of cer- tain precipitates from the egg secretions of Arbacia and Asterias. Two charts; andiiires houres.. 2. Suh eee tee ene se . ee 459 S.O. Mast. Effects of chemicals on reversion in orientation to light in the colonial form, Spondylomorum quaternarium......................... 503 A. FRANKLIN SHULL. Relative effectiveness of food, oxygen, and other sub- stances in causing or preventing male-production in Hydatina......... 521 Suinicnt Matsumoto. Contribution to the study of epithelium movement. The corneal epithelium of the frog in tissue culture. Nine figures........ 545 AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE MARCH 2 STUDIES ON INBREEDING I. THE EFFECTS IN INBREEDING ON THE GROWTH AND VARIA- BILITY IN THE BODY WEIGHT OF THE ALBINO RAT HELEN DEAN KING The Wistar Institute of Anatomy and Biology The rapid development of the new science of genetics has opened up many fertile fields of investigation and it has also re- vived interest in the problem of inbreeding which has been dor- mant for many years. Charles Darwin (’75, ’78) considered the subject of inbreeding so important that not only did he collect all available data regarding it, but he himself carried on a series of inbreeding experiments that extended over a period of eleven years, Darwin’s experiments on plants were followed by those of Crampe (’83), of Huth (87), and of Retzima-Bos (93; ’94) on various species of mammals. The conclusions reached by each of these investigators can well be stated in the words of Darwin (78): “The consequences of close interbreeding carried on for too long a time, are, as is generally believed, loss of size, con- stitutional vigor and fertility, sometimes accompanied by a tendency to malformation.”’ Darwin adds, furthermore: ‘‘ That any evil directly follows from the closest interbreeding has been denied by many persons, but rarely by any practical breeder and never, as far as I know, by one who has largely bred ani- mals which propagate their kind quickly.” On account of the almost universal prejudice against inbreed- ing, or because the results of former work seemed conclusive, the problem of inbreeding was practically ignored by scientists after the publication of Retzima-Bos’ results in 1893, and only within the past decade has it again received any serious consid- eration. ‘The recent experiments of Gentry (05) on swine, of Castle et al. (06) and of Moenkhaus (’11) on Drosophila, and 1 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 1 MAY, 1918 2 HELEN DEAN KING of various stockbreeders on horses and cattle (Chapeaurouge, 09; Anderson, ’11) have shown conclusively that there is no general physiological law forbidding inbreeding and that the re- sults obtained depend very largely on the character of the stock that is inbred. As Chapeaurouge has stated: Die Erfolge und Miserfolge nichster Inzucht haingen klar und sicher nicht nur von der Gesundheit und Konstitution der Stammel- tern und den dusseren Verhiltnissen ab, in welchem die Tiere gehalten werden; sondern auch vor allen Dingen von der richtigen Auswahl der Tiere zur Weiterzucht, welche nie die allgemeinen Bedingungen des ziichterischen Zweckes gegeniiber den speziellen aus dem Auge lassen darf. Je mehr die Haltung den natiirlichen Verhiltnissen oder den Bediirfnissen der Zuchttiere entspricht um so weniger sind tible Folgen zu befiirchten. There are a number of questions of economic as well as of scientific import which recent work on inbreeding does not answer: Does close inbreeding, if carried on for a long period . of time, ever lead to degeneration if only the best animals, from sound stock, are used for breeding? If degeneration does fol- low from this kind of mating, does it affect merely the body size, vigor, and fertility, or does it also influence body form and modify the structure and action of the central nervous system? If no evil results appear, can inbreeding be used to improve a race by combining the best of the dominant characters with any de- sirable recessive ones that may appear? Finally, does inbreeding change the normal sex ratio, as Diising (’84) and others have maintained? Taking advantage of the opportunity which the animal colony of The Wistar Institute of Anatomy and Biology afforded for carrying on an investigation that might answer some of the above questions, a series of inbreeding experi- ments on the albino’ rat was started in the spring of 1909. As the work, although still in progress, has already extended over a period of eight years, it seems advisable that some of the results obtained should be published, since they lead to rather definite conclusions. The present paper, the first of a series, gives a detailed account of the investigation, with an analysis of the effects of close inbreeding on the growth and variability EFFECTS OF INBREEDING ON BODY WEIGHT 3 in the body weight of rats belonging to fifteen inbred genera- tions; subsequent papers will deal with the effects of inbreeding on fertility, on constitutional vigor and on the sex ratio. To avoid needless repetition, a complete bibliography will be given in the final paper. I wish to express my great obligations.to the Director of The Wistar Institute, Dr. M. J. Greenman, for placing at my dis- posal ample facilities for carrying on the investigation and for his unfailing interest in the problem. I am much indebted, . also, to Dr. H. H. Donaldson for advice and criticism that have been most helpful when difficult situations have arisen at various stages of the work. 1. MATERIAL AND OUTLINE OF THE EXPERIMENT The albino rat (Mus norvegicus albinus) is a domesticated variety that is well adapted to confinement and is easily cared for. It breeds throughout the year, producing litters of rela- tively large size, and under proper environmental conditions from three to four generations can be obtained in a single year. This combination of characteristics makes the rat very useful as a laboratory mammal, and exceptionally favorable material for a study of the problem of inbreeding. The experiments were started with a litter containing four rats, two males and two females, which was born on the 10th of May, 1909. This litter was selected from a number of new-born litters in the general stock colony merely because it happened to contain the number of individuals required for starting the in- vestigation, and not because of the ancestry, the size, or the vigor of the individuals. One of the females in the litter was called ‘A,’ and her descendants form the ‘A series’ of inbreds; the other female was designated as ‘B,’ and her descendants form the ‘B series’ of inbreds. The plan of the experiment required that females A and B, as well as all of the females subsequently used for breeding, should be mated twice with a brother from the same litter and then twice with an unrelated stock male. Since the mating of 4 HELEN DEAN KING brother and sister from the same litter is the closest form of inbreeding possible in mammals, no other form of mating has ever been used to obtain inbred litters. In every generation all females used for breeding have belonged to inbred litters; none of them have ever been taken from ‘half-inbred’ litters obtained by the mating of inbred females with stock males. In the early part of the experiment the number of breeding animals was, of necessity, small. In every generation after the - sixth about twenty females from each series were used for breeding, so that approximately 1000 young were obtained in each generation. . All four of the rats used in starting the experiment were killed when they were no longer wanted for breeding purposes. Each rat was weighed, measured, and carefully dissected. When the various records were compared with the norms for the albino rat (Donaldson, 715) it was found that all of the rats were under the average body weight for their age, but that they were nor- mal in all other respects as far as could be determined by the usual methods of laboratory procedure. The fact that these individuals were sound, healthy animals and normal in all essential respects is a point on which I wish to place special emphasis in order to forestall the possible criticism that the re- sults obtained in this work were due to the use of an exceptional strain of rats. In the earlier generations the inbred rats exhibited all of the defects which are popularly supposed to appear in any closely inbred stock. Many females in both series were sterile, and those that did breed usually produced only one or two litters which were generally of small size. A considerable proportion of the rats were dwarfed, or stunted in their growth, and many of them developed malformations, particularly deformed teeth. The animals showed, also, a steady decline in vitality in suc- ceeding generations and usually died at a relatively early age. If the experiments had been discontinued at this point the re- sults would have been a confirmation of the conclusion reached by Darwin and by several other investigators, that inbreeding invariably leads to sterility and to physical degeneration. EFFECTS OF INBREEDING ON BODY WEIGHT 5) Fortunately for this work, many rats in the general stock colony, in which there was no inbreeding, exhibited the same characteristics as the rats belonging to the inbred strain. It was evident, therefore, that the unfavorable condition of the animals could not justly be attributed to inbreeding alone. On investigation it was found that all of the rats were suffering from malnutrition due to the character of the food that they received. At the time that these experiments were begun the rat was used as a laboratory mammal in only a few of the larger research institutions in the country, and little was known of the environmental and nutritive conditions best suited to its needs. Following the plan of feeding in general use in other animal colonies, the rats were fed chiefly on bread soaked in milk and on corn; meat and vegetables being given only once each week: Such a diet does not furnish the proportion of food elements that the rat requires if it is to be kept in good physical condition: there is too much starch, too little protein. In the spring of 1911 a radical change was made in the rats’ food. Milk and fresh bread were eliminated from the diet and ‘scrap’ food, consisting of carefully sorted. table refuse, was fed once each day; cobcorn being kept in the cages as an extra ration. Such a diet has proved to be a most satisfactory one, and it has been used continually up to the present time, except that dog biscuit has been substituted for cobcorn as extra food supply. This change was made last year following the loss of a consid- erable number of animals through intestinal disturbance caused by the eating of fermented corn. A very marked improvement in the general condition of all of the rats in the colony was noted very soon after the diet was changed. The animals gained in size and in weight, sterility almost disappeared, and the average number of young in the litters was increased. From this time on malformations were no longer common, and not a single instance of deformed teeth has been discovered in the thousands of animals that have been bred in the colony during the past five years. Simply by a change in the food characteristics said to typify the ‘dire effects of inbreeding’ were eliminated, and up to the present time they 6 HELEN DEAN KING - have not reappeared, although the rats have been carried through twenty-eight generations of brother and sister matings. In the inbred colony, up to the sixth generation, very little selection of breeding animals was possible; any females that would breed at all were used to continue the series. The change in food was made at the time that -the animals of the fourth inbred generation were reaching maturity. In the course of the two following generations the effects of malnutrition gradually disappeared, and in the sixth generation most of the rats were of normal size and relatively large litters were being produced. After this time large and vigorous animals were available for breeding. purposes, and it became possible to make a careful selection of the breeding stock. From the seventh generation on the selection of the individ- uals which were to serve as progenitors of the succeeding gener- ation was always made among the newborn young, as the sexes can readily be distinguished at this time (Jackson, 712). In the A series of inbreds, which is called the ‘male line,’ all litters containing an excess of female young were always discarded; in the B series, the ‘female line,’ litters with an excess of male young were never reared. Unless the individuals in the litter were of normal size and vigorous at birth they were killed at once. The young which were retained remained with their mother until they were one month old, when they were again carefully examined, and if they did not come up to the norms for stock animals of like age they were discarded. If the young rats ful- filled all requirements as to body weight and vigor they were returned to the cage to be reared as possible breeding stock. This rigid selection left in each generation, as a rule, at least three times the number of animals that were required for breed- ing. When the rats become sexually mature, at about three months of age, they were again inspected, and any that were be- low normal in any way were rejected. Generally only one female of a litter, the first to breed, was taken to continue the line. If, however, the individuals were unusually large and vigorous, two, very rarely three, breeding females were taken from the same litter. EFFECTS OF INBREEDING ON BODY WEIGHT ff Since the change in the food in 1911 and the removal of the colony to new quarters in 1913, the environmental conditions under which the rats were reared have been as uniform as it was possible to make them. All inbred rats, and also the stock animals used for controls, have been subjected to the same con- ditions of light, of temperature, and of nutrition, and they have been cared for in a similar way. Any differences between the two inbred series, or between inbred and stock animals must, therefore, be ascribed to causes inherent in the individuals; they cannot be attributed to the varying action of environment or nutrition. 2. THE GROWTH IN BODY WEIGHT OF INBRED RATS In view of the results that earlier investigators (Crampe, Ritzema-Bos) obtained in their inbreeding experiments with rats, little attention was paid to the fact that the body weights of the animals in the earlier inbred generations were consider- ably less than the norms for stock albino rats of like age. When the individuals of the sixth inbred generation became mature, it was noted that many of them were much larger than stock animals of the same age. This fact was so at variance with the generally accepted belief regarding the effects of close inbreed- ing on body size that it seemed desirable to make a study of the weight increase with age of individuals in the later generations of the two inbred series. From the seventh generation on from three to five litters of each inbred series were weighed, first when the animals were thirteen days old, again when they wére weaned at thirty days of age, and thereafter at intervals of one month until they were fifteen months old. At the thirteen- and _ thirty-day periods animals of the same sex were weighed together and the average body weight for the group recorded, as at these ages individual differences in body weight are, as a rule, too small to make separate weighings necessary; at all other ages individual records were taken. The litters that were used for a study of growth in body weight were all selected at birth on the same basis as the litters 8 HELEN DEAN KING that were to be reared for breeding purposes. There was no culling of the less desirable individuals, however, and all mem- bers of every litter were reared and weighed at the ages noted. When the animals became mature the largest and most vigorous pair in each litter was usually used for breeding. As the growth of very young rats depends largely on the amount of nourishment that they receive from their mother, litters of medium size, containing from five to eight young, were, as a rule, those selected for weighing: Such litters, more- over, represent the general run of individuals in a colony more fairly than do very large or very small litters in which animals with extreme body weights are often found. It was not always possible to weigh adult animals at exactly the ages designated in the various tables, but the weighing of a litter was omitted if it could not be taken within one week of the time specified. In weighing experiments of this kind there is always an un- avoidable error due to the presence of a greater or a less amount of undigested food in the alimentary tract. To obviate this source of error as far as possible the rats were weighed in the morning before they had received their daily food ration. Animals that were obviously ill and females known to be preg- nant were never weighed, while the weight of suckling mothers was not recorded if it was below the previous record. In the rat pregnancy cannot be detected with certainty until about the thirteenth day, so undoubtedly many gravid females were weighed unknowingly during the course of this investigation. The increase in the body weight of a female as the result of pregnancy cannot be very great up to the thirteenth day, how- ever, since Stotsenburg (715) has shown that the weight of a fetus at this time is only 0.04 grams. Errors in the records due to the inclusion of the weights of pregnant females were doubtless balanced by the weights for animals that were in early stages of pneumonia when there was no external evidence of the disease. The present paper gives data showing the weight increase with age in 333 males and in 306 females belonging to the first fifteen generations of the inbred group. Altogether these gen- erations comprised a total of 1601 litters, containing 11,657 EFFECTS OF INBREEDING ON BODY WEIGHT ) individuals, of which the majority were the progeny of brother and sister matings, the others were the offspring of inbred fe- males and stock males. The number of animals for which weight records were taken is, it is hoped, large enough to be fairly representative of the inbred colony as a whole and to give results that have some statistical value. No animals belonging to the first six inbred generations were weighed at regular intervals, but fortunately a series of weight records is available that will give some idea of the size of these rats after they became adult. In order to ascertain the effects of close inbreeding on the weight of the central nervous system, one of my colleagues, Dr. 8. Hatai, made careful autopsies of the four rats used in starting the experiment and he also exam- ined a large number of their descendants. From the record ecards for these animals I was able to obtain the body weights of a considerable number of rats belonging to the first six gen- erations of the two inbred series. The body weight data for only the best of the animals reared during this period were used in the present case; records for all individuals that were noted as having deformed teeth or other malformations were excluded, as well as the records for those animals that were ob- viously runts. Table 1 shows the body weight records for 92 males and for 73 females belonging in the first six generations of the A series of inbreds: table 2 gives similar data for 85 males and for 64 females of the B series. Only one record for each individual was obtained, i.e., the body weight at the time that the animal was killed. _ The ‘mean age in days,’ as given in the first column of table 1 and of table 2, shows the median points in thirty periods that correspond to the ages at which the rats of the later inbred generations were weighed. For example, under the mean age ‘151 days’ are grouped the body weights of all of the animals that were killed when they were from 136 to 165 days of age. As no records were taken of animals that were less than 105 days old the first age group given is that for 120 days. TABLE 1 Showing the increase in the weight of the body with age for rats belonging in the jirst six generations of the A series of inbreds MALES FEMALES MEAN AGE Body weight in grams amber Body weight in grams Number of indi- of indi- Average | Highest | Lowest | Viduals | 4 verage Highest | Lowest | Viduals days 120 200.1 252 134 ill 142.8 193 106 13 151 188.2 267 133 24 127.5 165 103 15 182 218.2 317 135 1% 182.2 199 105 5 iD, 183.6 227, 138 5 123.7 153 102 8 243 239.8 293 182 11 iis 178 eis 2 273 225.0 292 1155; 11 146.4 218 13 5 304 246.2 273 203 5 137.6 143 128 3 334 274.0 312 225 3 189.0 225 162 5 365 310 1s 164.8 215 143 5 395 273.0 294 269 3 Wel? sal 191 151 6 425 166 ili 455 164 8 180 152 5 92 73 * Record not used in constructing graph. TABLE 2 Showing the increase in the weight of the body with age for rats belonging in the first six generations of the B series of inbreds ‘ MALES FEMALES MEAN AGE Body weight in grams Naamber Body weight in grams Number of indi- of indi- Average | Highest Lowest viduals Average | Highest Lowest viduals days 120 205.8 266 142 24 149.7 170 136 4 151 210.6 200 141 16 149.4 211 113 11 182 241.6 345 146 15 156.0 185 109 5 212 235.7 315 157 12 119.2 173 100 + 243 210.5 304 238 + 169.7 189 149 4 273 21506) 241 185 6 161.0] 193 131 4 304 233.7 287 174 + 161.7 189 138 12 334 . 249 te ey 1916)! +) 218 161 6 365 223.0 257 202 3 163.6 185 144 5 395 167.8 193 139 5 425 455 172.2 213 145 4 85 64 * Record not used in constructing graph. 10 EFFECTS OF INBREEDING ON BODY WEIGHT 11 A comparison of the corresponding data for the individuals of the two series of inbreds is best made through the graphs in fig- ure 1 constructed from the data for the average body weights at different ages as given in table 1 and in table 2. In figure 1 the graphs for the body growth of the males are considerably higher than those for the females, since the male rat, after reaching maturity, is normally a much heavier animal than the female of like age. Considering that only one record Growth in body weight. Albino Rat. 320 EEE 4 300k#4+44 PEER EEE SaSn5005 i cht faint aaa 280 Body weight in grams 5 M cu Br T A T I a | folate aes Series A Cer 260 PH EEE EEE FH + | | seers secs H | 240 Set = Hi eco ey gee ts | : ; 220-4 sane | af 4 Ht i - is HEE cH m= |Series B a soneatistfoeatestas BSGED/ cteatasestanteatane | 200/-- EH ‘Seual +H HHH i | | l | i | i Fy ] DBE Females 180} Ho t 5 | se a8 Series Bi CeCe eee | +} Tos ea See eeeeae CoH H t is Series A i epee i TT tact ial 140) ! | t - : } ae Bee aH 120 7 ee ae im i rH HHH 100 - - Ege saceneeR Sasnouu eo tet : if sasneus! It | tH. 608+4+44--- t 4 4 t 40-4 EEOC SeeEReaan! 20 4 Cer Age in days 0 20 40 60 80 100 120 140 160 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 Fig. 1 Graphs showing the increase in the weight of the body with age for males and females belonging in the first six generations of the two series of inbred rats (data in table 1 and in table 2). was taken for each animal the corresponding graphs for the two series, especially those for the females, run remarkably close together. It is evident, therefore, that there was no significant difference between the two series as regards the growth of the individuals during the first six generations of inbreeding. Table 3 gives the-body weight data for all of the individuals of the first six inbred generations for which records were taken (a combination of the data in table 1 and in table 2). 12 HELEN DEAN KING TABLE 3 Showing the increase in the weight of the body with age for rats belonging in the first six generations of the inbred series. A combination of the data in table 1 and in table 2 MALES FEMALES MEAN AGE Body weight in grams Number Body weight in grams Namber of indi- of indi- Average | Highest Lowest viduals Average Highest | Lowest viduals days 120 204.0 266 134 35 144.4 193 106 17 151 197.2 267 133 40 136.6 211 103 26 182 229.1 317 135 32 139.1 199 OSes 10 212 218.3 315 138 18 22, 173 100 12 243 248 .0 304 182 15 W223 189 149 6 273 Mpa Of 292 155 ike 152.8 218 10 9 304 240.6 287 174 9 156.9 189 128 15 334 267.7 312 225 4 190.4 225 161 1 365 244.7 310 202 4 164.2 215 148 10 395 273.0 294 269 3 172.9 193 139 i131 425 166 1 455 168.1 213 145 9 177 137 * Record not used in constructing graph. A comparison of the data in table 3 with body weight data for a series of stock albino rats reared as controls for the inbred series after the change to ‘scrap’ diet had been made (table 13) will show to what extent malnutrition decreased the body size of the individuals in the first six inbred generations. Graphi- cally this result is shown in ‘figure 11 and in figure 12 (compare graph D with graph A and graph C). Body weight data were taken, at the intervals stated, for 99 males and for 76 females belonging in the seventh to the fifteenth generation of the A series of inbreds. The average body weights of the males at different age periods are shown, by generations, in table 4: corresponding data for the females are given in table 5. Data for 57 males and for 93 females belonging in the seventh to the fifteenth generations of the B series of inbreds are shown according to generations in table 6 and in table 7. TABLE 4 Showing, by generations, the average body weight at different ages of 99 males belong- ing in the seventh to the fifteenth generations of the A series of inbred rats GENERATIONS AGE 7 8 9 10 11 12 13 14 15 days grams | grams | grams | grams | grams | grams | grams | grams | grams 13 18 18 18 18 16 19 19 18 18 30 43 45 46 41 44 44 49 47 42 60 150 a0 134 129 102 127 134 123 PB} 90 201 151 199 214 W7/e 184 186 193 189 120 242 196 249 264 227 229 244 233 204 151 Aa lecoone 28 || 289 lezoralis 249n eo TSaly 259235 182 306 286 308 315 279 273 291 273 256 ZI Ble || Bilis |] SEXO) | Ag) |) ASS} | SBT Stilts | ASS Pe 243 329 | 341 | 342] 318] 302) 295] 317) 293 | 286 273 305) | ebys || ate4b |) Stoke |) Sil jie Sle |) Bee) SEN BiOt0) 304 357 345 334 310 314 319 333 302 303 334 368/368) 376 || 305. | “327/316 | 339") 307 | 307 365 SAMO | OOM OOO) | sasOulms24. Er ooon |) soll iol 395 SOMO OMsSOnl SOM SOOM eol oni S400 le roloe|p molt 425 ANE soe) 4045 930%) S54 3345) 355) |) 330) | 9321 455 424 | 3838 | 403 | 324) 348] 3389] 344) 322) 320 Number of rats WEIOMEG eee ae. 6 6 9 14 12 10 15 15) 12 TABLE 5 Showing, by generations, the average body weight at different ages of 76 females belonging in the seventh to the fifteenth generations of the A series of inbred rats AGE days | 13 30 60 90 ANE 151 182 212 243 2a 304 334 365 395 425 455 Number of rats welgihed).... 4... * One record only. 7 grams a6 40 126 160 172 190 190 187 189 187 8 grams “I GENERATIONS 10 11 12 13 14 15 grams | grams | grams | grams | grams | grams ifs) 15 18 7 107/ 18 37 41 4] 45 45 40 107 90 102 110 108 105 184 139 145 157 162 144 184 181 168 W722 171 163 189 182 187 192 185 183 208 189 195 195 204 203 193 184 218 215 209 211 202 200 203 216 214 218 223 192 226 218 PAE 222 236 209 239 PPA 222 OP, 242 222 244 218 223 227 256 223 238 220 226 226 253 225 253 Depall 228 224 244 260*| 268*| 230 | °226 219 PART 257 243 224 219 217 8 10 8 9 10 11 13 TABLE 6 Showing, by generations, the average body weight at different ages of 67 males be- longing in the seventh to the fifteenth generations of the B series of inbred rats AGE GENERATIONS 7 days 13 30 60 90 120 151 182 212 243 273 304 334 365 395 425 455 Number weighed of rats grams 21 49 170 244 277 322 344 376 372 454* Ai is * One record only. 9 10 11 12 13 14 15 grams | grams | grams | grams | grams | grams | grams 20 18 19 19 19 20 20 53 49 40 47 50 52 49 140 172 149 140 144 | 132] 145 211 2M) “220; |) 212°) 1935), Sono S201 258 | 265 | 262 | 240) 225) 255) 233 288 | 291 300 | 264] 274; 279} 267 340 | 3384] 330] 278] 283) 300) 288 360 | 340 | 353] 290| 297] 305] 303 367 | 3386} 368 | 286] 303) 302) 308 371*| 316 | 367 | 3806] 309] 311 311 380*| 335 408} 313 | 3138) 322) 383i B02" 381 337 | 303*| 330) 321 3457! * 333))) ada |) 328 |) 326415 o62)|eao 365*| 365 | 336*| 334*| 351 339* 353 357*| 339* 347* 343*| 330* 3 5 9 8 6 7 9 TABLE 7 Showing, by generations, the average body weight at different ages of 93 females belonging in the seventh to the fifteenth generations of the B series of inbred rats AGE Number of rats “J ual GENERATIONS 9 10 11 12 13 14 15 grams grams |" grams grams grams grams grams 17 17 18 18 18 18 18 BO 48 | 39 39 46 48 | 45 (2917125 | 113 | A20 04 104s ioe 168° 0470 | 165 | 1499) 147 | Wé6ia aot 180) 2174 | 182 | 167-| 169.) aIsaa) aanze 200 | 199 | 191 | 186| 195 | 196] 191 205 | 208 | 200] 195:| 196] 214} 202 199 | 217 | 202] 204} 204] 210] 206 S11 | 217 | 225 | 212) S14 ie 20d FSi 212%| 236 | 208.) 2207 et99|" 214 P17 1 239'| 243 | 207°) Sapa B23; |) iz 239 255 | 211 | 243] 237 | 224 242 | 262*] 249 | 215.| 238| 241] 229 9A1 |) 275*| 241 | 2128) Raa") 280.) 229 261*| 238 | 207 239 | 234 280*| 244 2310) 235 5 11 12 10 9 12 16 * One recérd only. 14 EFFECTS OF INBREEDING ON BODY WEIGHT 15 Tables 4 to 7 are inserted chiefly for reference, although they bring out two important facts more clearly, perhaps, than do any of the other tables. The data, as given in these tables, show that rats belonging to the earlier generations of the inbred series did not live as long, as a rule, as did the individuals in the later gen- erations. This was, particularly noticeable in the individuals of the Bseries. Up to the twelfth generation only three rats in the B series (two females and one male) lived to the age of 455 days; in subsequent generations many individuals lived for the entire weighing period of fifteen months, and some of them were kept until they were nearly two years old. One who be- lieves with Crampe and Ritzema-Bos that continued inbreeding necessarily lessens vitality and so shortens the life of the indi- vidual meets here with the seemingly paradoxical fact that the animals that belonged to the later inbred generations outlived those that belonged to the earlier generations. In this experiment _ the use of only the most vigorous animals for breeding purposes has seemingly overcome any tendency that inbreeding might have to shorten the life of the individuals. The second point of interest brought out by tables 4 to 7 is that in each generation of the two inbred series the average body weight of the males exceeded that of the females at every age for which records were taken. At birth the male albino rat is slightly heavier than the female, whether the animals belong to a stock or to an inbred strain (King, ’15b). Data for the growth in body weight of the albino rat, as recorded by Donaldson (’06), show that as early as the seventh day after birth the growth of the female is more vigorous than that of the male, and that the female is, as a rule, a relatively heavier animal than the male up to about fifty-five days of age. Ferry’s (13) growth data for the albino rat (Donaldson, ’15; table 65) confirm Donaldson’s findings. In Jackson’s (18) data for the albino rat, ‘‘the excess of average weight was invariably in favor of the male at birth, and also in the majority of cases at all succeeding ages;’’ while the records obtained by Hoskins (’16) show that the albino female is a heavier animal than the male only at the age of about six weeks. In a series of stock albino 16 HELEN DEAN KING rats reared as controls for the inbred series (King, 715 a) the ma- jority of the males exceeded the females in body weight at each weighing period. As no records were taken when the animals were just six weeks old, the relative size of the two sexes at this age was not determined. At thirteen days of age, as tables 4 to 7 show, the average body weight of the inbred males was only one gram more than that of the females. Such a slight difference as this would be negligible, considering the possible error in the weighings due to the varying amount of food in the alimentary tract, save for the fact that it is found in every generation group. In some litters the females, were, on the average, heavier than the males at thirteen and also at thirty days of age, and in a few instances the weight of the largest female surpassed that of the smallest male even when the animals were sixty days old. Although the albino female, whether stock or inbred, has a relatively smaller birth weight than the male, she soon comes to have a body weight that is very nearly equal to that of the males in all cases, and often exceeds it. Even though the absolute body weights are less at any given age, therefore, the female grows more vigorously than the male during the first few weeks of postnatal life. An early acceleration in the growth of the fe- male also occurs in the guinea-pig (Minot, ’91; Castle, ’16), and it finds a parallel in man, as Donaldson (’12) has pointed out, since during a certain phase of development, when children are from thirteen to sixteen years old, the average weight of girls is greater than that of boys; while at all other ages boys, as a rule, are the heavier. The relative growth in body weight of males and of females in corresponding generations of the two inbred series, or in suc- ceeding generations of the same series, can be determined by referring to the data: in tables 4to 7. Fora comparative study of the body growth of the individuals in the two series it seemed advisable to combine the data for three succeeding generations. Data for the A series of inbreds are given in table 8. The growth graphs shown in figure 2 were constructed from the data for the males of the A series, as given in table 1 and EFFECTS OF INBREEDING ON BODY WEIGHT 7 TABLE 8 Showing the average body weights at different ages of inbred rats of the A series separated into three groups according to the generation to which the individuals belonged MALES FEMALES ae Generations | Generations | Generations | Generations | Generations | Generations — 0-12 13-15 = —-12 13-15 days grams grams grams grams grams grams 13 18.1 Anja 18.4 50 16.0 Le ot8) 30 + 49 .4 43.4 43 .6 42.5 40.0 43.2 60 122.3 118.3 127.0 98.9 98.5 107.6 90 186.0 193.2 188.9 Meo 157.9 153.6 120 230.6 242.5 228.1 160.5 178.5 168.3 151 269.5 267 .9 256.9 181.5 191.0 186.2 182 304.3 290.4 216.0 192.2 195.3 201.9 212 320.2 293 .2 287.1 189.8 196.2 Dilelyae 243 337.1 305.9 298.3 205.5 201.7 216.0 2a 357.4 312.5 307.7 214.7 215.1 219.4 304 349.5 314.5 Bleer 221 219.4 PANE T. 334 369.0 315.4 318.4 22a 234.9 223.5 365 379.2 318.1 3201 252.2 235.2 224.7 395 375.6 323.6 323.6 232.0 239.5 226.4 425 399.4 327.0 332.0 PAPA) 254.0 223.8 455 403.5 336.0 333.2 241 .0* 251.6 219.0 * One record only. in table 8. In this, as in some of the other figures, the space between the graphs was slightly widened, where properly the lines would run very close together or overlap, in order that the course of each graph might be clearly followed. Figure 2 shows that the body growth of the males in the three generation groups of the A series of inbreds progressed at about the same average rate until the animals were 150 days of age, as is indicated by the position of graphs A to C. At this point there was a marked acceleration in the growth of the males belonging to the group comprising the seventh to the ninth generation (graph A) which continued until the end of the weigh- ing period. Graph D, representing the growth of the males of the A series during the first six generations, begins at the 120- day period, as no younger rats belonging to these generations were weighed. It seems almost incredible that this graph can THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 26, NO. 1 18 HELEN DEAN KING represent the adult body size of the progenitors of the animals whose growth is indicated by graph A, since rarely do we find a group of mammals having in the adult state a body size so much greater than that of its immediate ancestors. 3 Growth in body weight. Albino Rat/t11 eet | T £00 Fit t Series A Males 20 1 co t : = gaenuee tot | 380) | 5 Ht HEH cE PEE c 1 ay 3 aeectentt . FEE EEE ee Ire 360) fs EEE $i aque cH H 3401 CI E aa Body weight in grams 4 al. 320 v Ht 4 i. 300 a 280) fe E iB} 260 wa 240) a = a8 ye 220 rH iE it see 200 Ht HEH ct AAA 180 a trot pecuen 7 HH 160 Heo # eonee | } 140 + fe iH : se SEE EEE E 120) SG0BE 4 EI a Oe sil PAL aA 304 380.5 300Re 316.5 225.6 222.4 ile 334 363.1 359.0 323.1 230.6 228 .6 230.8 365 360.4 342.4 344.2 238 .4 229.3 234.3 395 369.0 BD oT 342.2 261.3 236.4 232.6 425 365 .0* 353 .0 348 .0 4} OF 230.4 236.4 455 347 .0* 331.5 sil 256.3 ZaaR2 * One record only. TABLE 10 Showing the average body weights at different ages of inbred rats of the combined series (A, B) separated into three groups according to the generation to which the individuals belonged MALES FEMALES 7-9 10-12 13-15 7-9 10-12 13-15 days grams grams grams grams grams grams 13 19.2 18.1 18.9 16.8 16.9 17.8 30 49.6 44.9 47.6 43.0 42.8 45.1 60 135.5 131.8 131.7 112.7 ORT 106.4 90 195.6 200.9 193.1 154.1 160.7 > 153.6 120 248 .6 250.5 231.4 172.1 177.2 171.8 151 281.1 274.3 262.2 191.2 190.7 190.2 182 315.6 298 .0 279.5 199.2 198.5 203.5 212 328.7 305.7 291.6 203 .4 202.2 208.3 243 343.9 311.4 300.0 208.6 209. 4 213.3 273 361.6 316.1 308.5 217.4 215.8 218.2 304 359.8 322.2 314.2 222.8 220.8 221.5 334 367.3 324.1 319.3 228.7 232.8 226.6 365 374.0 325.5 329.2 243.7 232.4 228.6 395 374.4 331.4 325.3 246.6 238.3 229.0 425 396.2 330.5 333.4 276.0 240.8 229.6 455 403.5 336.7 333.0 279.0 253.3 224.6 EFFECTS OF INBREEDING ON BODY WEIGHT 21 were quite as inferior to their descendants in body size as were the males in the corresponding generations of the A series (graph D). The growth in body weight of various groups of females be- longing in the B series is shown graphically in figure 5 (data in table 2 and in table 9). Graphs A to C in figure 5 were drawn so that the lines are distinct. These graphs should lie very close together and Growth in body weight. Albino Rat. is Ht 2ueSeELL a eries B Males Taal one | cH | fam H +f tt t HE 380} mapae iets Ty 4 : GEO Re a! alte iat a ai q | 360 : on a gests 5 | rH oanP eT i an Peet Ceerr B 340 HHA Foe admbe ues Ft cant t 7 1 Body weight in grams + Ceo hai 3 | { Cr 320 HE oe a0 300 T 280} 260 13 240s i 220 rep | Et t ‘7 a Goat 00 t t i 200F 3 cacy Asia ore TH BN 180-4 t T | HH = | - t - poataiuts 160}+ rH} | woeeuaeess TE ch t TI t aa tate i. ie! it Tm + eee eet t ia Coot 8 f sees | 3a T ies eeenes eeaee a A SHH seeeeeees fate tt a oct eee : at +H} t Et Aisases Et { afer Benross PEEEEEEEEEHT (Age in days SHSHE 20 40 60 80 100 120 140 160 180.200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 Fig. 4 Graphs showing the increase in the weight of the body with age for males belonging to various generation groups of the B series of inbreds (data in table 2 and in table 9: lettering as in figure 2). overlap in many places, since the data in table 9 shows that the actual differences between the corresponding body weights of the various groups are very small. The position of graph D indicates that adult females of the first six generations were very much smaller animals than their descendants of like age, as was the case in the A series also. - The data given in table 8 and in table 9 have been combined in table 10. Ze HELEN DEAN KING Figure 6 shows graphs for the weight increase with age in all of the inbred males for which weight records were taken (data.in table 3 and in table 10). As the position of the graphs in figure 6 show, males of the seventh to the ninth inbred generations (graph A) were animals of unusually large size and they were considerably heavier, after reaching maturity, than their own progeny of like age. At the 365 day period the space between graph A and graph B in- dicates a difference of about 50 grams in the average body 320 Se See Rees eeeee| F ; : eae SeSeee SB! oo [ a ' Po Growth in body weight Albino Rat. +++ tt PEEEEE { inmee + i ttt A —- it it COO poenes B Females 11 i Peet EEE 300 oe enced EEE ia “EERE ae ro EEC FREE Ht aoe ae 280 Body weight in grams[4 | TH co REE EERE Eee coo ra . i Po Cot i -t i SeSeEeaeneeeaeenen 260k 4 eecceecEesseee PARE EEE LEE BES H deue mi 240 rH PE ot oe roa 220 PS eer tt ttt i Ht HHH a Ft oH — > 200 ae tt | a8 i eesreceeeeaaaeee 180 Eee ++ te aa: wees 160) EEEEEEEEE aa t PEt Po ssasgesesas / / aun jit . + a - i | 100 t (aatts Tt ov / inee 80 ye eusey/ / jasuesees! Seay /) / Suesseceens 40g Att eeeeeeess // daneeea eae : +H HH 2 acaaned i eaeeeae a | rt Age in days EERE ECE ooo rH cH cH aa 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 Fig. 5 Graphs showing the increase in the weight of the body with age for females belonging to various generation groups of the B series of inbreds (data in table 2 and in table 9: lettering as in figure 2). . weights of the two groups of animals. From the tenth genera- tion on the course of growth in body weight’ was practically the same in all inbred males, as is shown by the fact that graph B and graph C run very close together throughout their entire length. Figure 7 shows graphs for the body weight increase in the series of inbred females (data in table 3 and in table 10). The relative position of the graphs in figure 7 show that, with the exception of the animals in the first six inbred genera- Growth in body weight. Albino Rat y Series AB Males 3 A 380} 360) 340 Body weight in grams : ? Cc 320 300 280 260 240 220 200 180} 160} 140} 120) 100) 80) 60 40 20) | Age in days Q 20 40 60 80 100 120 140 160 180 200 220 240 260 280.300 320 340 360 380 400 420 440 460 480 Fig. 6 Graphs showing the increase in the weight of the body with age for males belonging to various generation groups of the two series combined (A, B). Data in table 3 and in table 10: lettering as in figure 2. eon ee H(Growth in body weight. Albino Rat (HOE te segs ttt i r ia aE i 300 Series AB__ Females + + cage Soke , i TH 280) ic o cot = A; 260/14 Body weight in grams Coe Hy te 3. HHH = cI ee To 240 ‘sae aa t Pe _— t i 220) fo) co lez BL = HE Segseses 200 EEEeEere | t iS 180) sanee : SH EEE EEE Ht Ht Ht = . 1 160H+4 t jt a D 140 1a @ im rie T ae { ag | i T aoe / t 120 Ht Soee g 4 i Poth 10! cH t a 4 t Bot 7 { 1}. i = { we Sh ad Les 1 caeeer / EEE Faenen// a tH s sseseccesete EEeEEee Ht +h, ae HH H o it 1 Si we 206 Y Fett 16 BeBe rrr a “| Age in days’| OEBEREe 4 abe te eepeiestiastt | aRseetaettotts O 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 Fig. 7 Graphs showing the increase in the weight of the body with age for females belonging to various generation groups of the two series combined (A, B). Data in table 3 and in table 10: lettering as in figure 2. 23 24 HELEN DEAN KING tions (graph D), the body increase with age was about the same in all the various generation groups of inbred females up to the time that the animals were one year old. After this age the females in the group comprising the seventh to the ninth genera- tions were somewhat heavier than the other females, but their excess in body weight was very much less than that in the cor- responding groups of males (fig. 6). On referring to the data in table 4 to table7 it is found that in both series of inbreds the average body weights of the animals in the seventh, eighth and ninth generations were greater than those of animals in the tenth and subsequent generations. In the seventh generation, particularly, females as well as males were exceptionally heavy animals at all ages for which records were taken. The largest males yet obtained belonged in the seventh generation of the A series of inbreds; these rats weighed at their maximum 460, 482, and 511 grams respectively. The largest female in the series was a member of the seventh genera- tion of the B series of inbreds, and she weighed 323 grams when she was 425 days old. The probable cause for the unusual weight of the animals in the seventh to the ninth inbred genera- tions will be considered later. Data showing the range in variation and the averages for the body weights at different ages of the weighed individuals in the A series of inbreds (seventh to fifteenth generations only) are given in table 11. Similar data for the rats of the B series are shown in table 12. The graphs in figure 8 were drawn to show the relative growth in body weight of the males belonging to the two inbred series (data in table 11 and in table 12). B The males of the B series of inbreds had a heavier average body weight than the males of the A series up to the last weigh- ing period (455 days), as the position of the graphs in figure 8 show. The crossing of the graphs at the end has no significance, since the largest males of the B series died before the final weighing. Graphs showing the growth in body weight of the females in the two inbred series are shown in figure 9: the data for these graphs are given in table 11 and in table 12. TABLE 11 Showing the increase tn the weight of the body with age for 99 male and for 76 female rats belonging in the seventh to the fifteenth generations of the A series of inbreds MALES FEMALES AGE Body weight in grams Number Body weight in grams Number individ- Pes Average | Highest | Lowest uals Average | Highest Lowest uals days 13 17.9 21 15 99 16.1 22 14 76 30 45.6 82 35 99 41.8 67 33 76 60 123.0 176 72 95 -102.3 159 (fil 74 90 189.8 248 120 99 150.3 193 93 63 120 233.9 305 174 94 170.0 212 1338 63 151 263 .4 360 204 94 187.6 224 157 65 182 287.4 367 216 91 197.5 236 167 62 212 296.8 420 238 85 202.8 248 158 59 243 309.5 419 234 84 209.9 245 171 56 273 322.6 435 265 74 217.0 251 167 53 304 321. 1 415 267 70 221.0 257 181 42 334 328.9 415 268 63 PPA GP 268 196 37 365 336.2 428 289 59 229.8 282 198 35 395 335.9 433 283 51 230.3 274 206 31 425 345.9 448 281 44 2a2—2 268 206 PA 455 361.3 473 306 25 228.3 257 182 20 TABLE 12 Showing the increase in the weight of the body with age for 57 male and for 93 female rats belonging in the seventh to the fifteenth generations of the B series of inbreds MALES FEMALES AGE Body weight in grams Number Body weight in grams Number individ- individ- Average | Highest Lowest uals Average | Highest Lowest uals days 13 19.8 27 16 57 18.0 25 14 93 30 48.6 75 35 57 45.1 72 30 93 60 148.5 190 110 57 115.2 147 80 . 93 90 210.1 266 149 57 161.1 197 1L?/ 68 120 254 1a |) aad 218 55 176.8 | 198 134 | 73 151 282.9 365 232 x5) 193.4 229 156 67 182 307.5 407 253 52 213.7 241 168 70 212 319.4 410 254 44 208 .2 247 174 60 243 S2abl! 408 259 39 212.5 261 179) |= 54 273 330.7 454 275 32 213.3 278 177 Ay - 304 344.3 477 290, 28 221.1 281 179 40 334 344.4 385 299 18 230.3 290 179 26 355 348.2 376 298 lg 233.7 276 194 29 395 300.1 374° 333 11 237.2 301 206 23 425 353.2 365 339 5 239.4 323 210 19 455 336.6 347 320 3 245 .0 317 212 13 25 a Growth in body weight. Albino Rat EEE Males fs an a is Series A AB: [siete _<- Leal oO oO 1 a aa i i eo i oath SNERERETEE BaENGSEee HET INST TTT i A 5 |e | oe i] aaa eae 10 20 ~~ 30 DAYS Chart III. Showing the growth of the albino rats which received a small quantity of the olive oil in the normal ration contrasted with the rats which received in addition a small quantity of Sudan III. O---- oO Control (oil fed) eo————-e__ Experimented begun. The feces are hard and black in color and ether extrac- tion reveals the presence of the dye in a considerable amount in them. ‘The rats are eager for the experimental ration for the first two or three days. but afterwards they consume but very small quantities. The skin around the urethral opening is usu- ally wet and stained pink, owing possibly to lipuria. I have not determined the maximum length of time during which the rat can survive on this Sudan III ration. 108 S. HATAI TABLE 2 Showing the growth of the organs of the albino rat as affected by feeding with Sudan III plus oti. Dose 10 to 15 mgms. Younger series INITIAL MEASURE- SUDAN III AND OLIVE OIL FEEDING CONTROLS LENGTH = MM. MENT WEIGHT = GRAMS BP +5 SPS olsaete laa rie liet2e Bodyitlenothats. -oerae....: 94.8 94.7 94.5 96.3 117.3 ADEM NEE cao goose onecc ome bell Con! 78.5 79.3 79.8 96.0 Bodygwetvht. >... .c 54 her « «cen eeee 21.5 2} il 24.3 44.2 Brae. 22. ict ees Sees 3) ee es Oo 1.332 1.383 1.331 1.463 KG OMS Sis. os eri pen ener 0.319 0.292 0.323 0.354 0.566 LIDS o5 ee a tn, 3 SRE I 2 1.260 pot 1.987 2.324 3.102 Spleens: .. cu see en. |. Se BOGE 0.051 0.065 0.065 0.218 aI ChE AS 4Ane Bie ee tees nets beam 0.166 0.200 0.226 0.248 0.449 Mb ymius:). wee Rec | eae 0.048 0.012 0.012 0.007 0.127 SUpraren al Saaeeeer see ee eee 0:009 0.009 0.010 0.010 0.014 UN. OFC) «eee rae c= cules « eee 0.005 0.005 0.005 0.006 0.008 Ey pophiysis wae eee ee 0.0014 | 0.0012 | 0.0013 0.0014 0.0020 MOSGeS: Vata rent ie oaeteee oe: ae 0.130 0.098 0.087 0.087 0.157 Ovariessarnee a5 rite ee RUE O0865), 0005 0.0038 0.0084 Agee (days) Me. Posen eee ae 28 33 35 38 38 GROWTH OF ORGANS AS AFFECTED BY THE FEEDING OF SUDAN III In order to determine the effect of Sudan III on the growth of the organs, young rats just weaned were treated in the following manner. One or two rats were measured and the weights of the organs determined at the beginning of the experiment. The re- maining rats were examined at different intervals in order to compare the organ weights at later ages with those at the begin- ning of the experiment. At the end of the experiment the rats which had been kept as controls and which were receiving the normal ration were examined together with the last test rats. The object of this final procedure was to show the amount of growth the organs should have made had the rats not been sub- jected to the Sudan III diet. The last test group was also util- ized to determine the percentage of water and the lipoids in the organs. ‘The results of these examinations are given in table 2 for the younger, and table 3 for the older series. FEEDING SUDAN III TO YOUNG ALBINO RATS 109 TABLE 3 Showing the growth of the organs of the albino rat as affected by the feeding of Sudan III. Dose 8 to 9 mgms. Older series INITIAL MEASURE- SUDAN III AND OLIVE OIL FEEDING CONTROL LENGTH = MM. MENT WEIGHT = GRAMS = : {3 go +69 i dave iedays ; sp Gayee 2974309 ROCK IGmGhllesoagsscooecdor Gane 108.1 113.6 116.3 119.5 155.2 Mba Meret las 5 $c ae Segoe = 86.2 94.8 99.0 109.0 139.4 Bodvaweisiite,.«.i%2.2 sudan |||, 8-9 Soll 40.1 44.7 98 .4 TBAT aa 15 Or ce Glee EEE core ioe 1.379 1.412 1.412 1.438 1.598 IGIGINEV Sh Aarts fae Met eae sa 0.461 0.473 0.482 0.446 0.794 IN OTA ere tk. sk acts Me ade 2.063 2.577 2.887 2.615 3.861 Pee ee st ois - 0.142 0.151 0.222 0.273 0.550 aM CReaS 2 te fais cs cae ae 0.280 0.302 0.341 0.421 0.703 Moa S is bt ok oe ae eee 0.107 0.073 0.067 0.031 0.273 Suprarenals: +. 5... se -gee ae oe 0.014 0.013 0.014 0.013 0.023 Muh tOld. = 2055 £)..< see oe 0.008 0.014 0.015 0.018 0.020 Ey POpliysisan: +. - ate fees oe 0.0016} 0.0018; 0.0019) 0.0019} 0.0040 PRESTCR Ss stets cee sk Rae ee a ee 0.202 0.199 0.0229 1.262 Owarteste: tee. sates eS ss 0.0107} 0.0084 0.0065} 0.0152 Als eu (dais) bts soo eae ote on 29 36 43 64 64 From table 2 we note that despite the fact of almost no growth of the entire body of the test animals, such organs as the liver and pancreas made a striking increase in weight throughout the entire period of the Sudan III feeding. On the other hand, the thymus gland was greatly atrophied, and the sex glands also showed a marked diminution inweight. Noneof the other organs showed noticeable loss when compared with those of the rats examined at the beginning of the experiment. These alterations just noted appear also in the second series of the experiments (table 3), where a much smaller dose of Sudan III (8-9 milligrams) has been given. I may add here that al- though on the average the kidney weights of the Sudan-fed rats do not differ much from their initial weights, nevertheless the individual records show that in the majority of cases the kidneys of the test rats have suffered a slight atrophy. I am thus inclined to believe that the kidneys also undergo a slight atrophy as the 110 S. HATAI TABLE 4 The weights of the organs determined in both control and Sudan III fed rats at the end of the experiment SERIES IV SERIES VII AND VIII guRIES IX AND X LENGRe nee Dose Dose Dose VERSIE SS EERIE Control Bade aad Control eadeceae ee Sudan 1¢+19 | oil fed for 22 -| oil fed for biscuit oil fed for 21 days 28 days 2c 28 days 2741929 29 2c Bodyalenathe.... -... 129.0 ODEs 146.5 110.0 142.5 109.5 Manlglencuhe eee .see 113.0 SB. | si w 105.0 130.0 101.0 Body weight......... 52.0 26.7 79.8 36.8 71.9 33.9 BAUM es: cee eeoe as 1.490 1.338 1.516 1.483 1.604 1.478 Heart 0.372 0.190 0.371 0.191 TkGiG boVehten 6 GWeA oe OF 0.473 0.274 0.721 0.394 0.641 0.423 EGVeLSA. ss Seer 2.219 1.992 3.073 2.046 2.588 2.174 Lungs 0.516 0.344 0.488 A SPLCeI ae aerate 0.255 OP23 0.167 0.079 0.148 0.074 IRAN Che as aceschiete: «kee 0.425 0.268 0.686 0.365 0.591 0.348 IMRAN Se. Ges aoe be 0.146 0.015 0.256 0.030 0.206 0.012 Suprarenals.......... 0.015 0.009 0.025 0.010 0.019 0.011 hiv Told epee oe 0.009 0.010 0.020 0.010 0.0138 0.008 ELypophysises..+ 3... \: 0.0023) 0.0013) 0.0029} 0.0018} 0.0029) 0.0019 Mestesrrnerr. sacs 0.341 0.071 0.818 0.201 OWariesae ates ie ee 0.0097) 0.0070) 0.0149} 0.0052 Ace tidays) itis. 48 48 50 58 57 57 result of feeding Sudan III. Some of the organs, however, in- crease in weight. The increase of the liver in weight seems to be due to an accumulation of fat, as this can be shown histologi- cally, as well as by the amount of lipoids extracted (page 113). The increased weight shown by the pancreas may be due to the hyperactivity of this organ in connection with the metabolism of fat. The atrophy of the thymus gland as the result of feeding Sudan ITI is most noticeable. The rate of thymus atrophy seems to depend on the amount of Sudan III administered. A larger dose appears to reduce the thymus much more quickly than a smaller dose. Since the thymus is highly sensitive to almost any abnormal physiological condition, such as an exposure to Roentgen rays, inanition, etc., we may assume that the present FEEDING SUDAN III TO YOUNG ALBINO RATS Lit TABLE 5 Showing the percentage of water in the various organs of Sudan III fed rats con- trasted with those of control rats eee | Amor | ted Control Sudan Control Sudan Control ou ol ead Bede 19'+18 I+19 17'1+29 2419 29 20 pica We Body length...... 129.0) || 102.7 || 153.0) | 120.2 | 146.5 | 1100) 142.5 | 10985 Body weight Loe oes 52.0 2007, 99.0 46.4 79.8 36.8 71.9 33.9 Age, days.........| 48 48 63 63 58 58 57 57 BlOOdeaee ee eee. | ee oleGale S3n00 a S0e4 I sS3- Onl SOe 9.) SOL88|) 816), S2e Brann eer vena TOEOM COLON, ZOLOFT F826) | e9eON | 7820s 7 9ee|) 786 iHearte.s: USD NW GE HAW | CBs) 80683 |) Alar iKedneyseeneaces. - fas) || COO | Hans 15. 4 iGe3 75.3 162051 9 7650 VET Eten S 6x >: 3.1 67.1 KOLO IN GS. 2a Reon (Gds2/ tle lee | ceed. Spleen...... CT eOA OOM Miho ie COLT) || MESmle le MOkOo ln aneoul el Oek ILUtVESi. Se ote pENeeE (8.2) 80.2.) 7952) “80.0 | 79.4 |) 479.8 Panereas)...... C224 OnOE | Gene ahon 69.9 74.6 | 69.2] 76.0 aishiryanSpeeeetas se -- 18208" 80.9) W9ee |) SEO Gee |) esis instance of atrophy may be due either to the abnormal nutri- tional condition or possibly to a direct toxic action of Sudan III on the thymus. I have given in table 4 (see also the last two columns in tables 2 and 3) the weights of the organs determined in both control and test rats at the end of the several experiments. The organs given in table 4 were used for the determination of water and lipoid content, as well as for the demonstration of the presence or absence of Sudan III in various organs, by both microscopic and extraction methods. As we should expect, the organs of the test rats are smaller than those of the controls. It is interesting to note. also that even the liver and pancreas, although they have made a continuous growth during the ex- perimental period, are yet far behind those of the control rats. I wish to eall special attention to the case of the control rats which received the same amount of olive oil as the test rats (series [X and X). In this case we notice that the control and test rats show the same degree of difference in the organ weights as in 142 S. HATAI TABLE 6 The percentage of water in the brain of Sudan III fed rats contrasted with that in the brain of the control rats BRAIN WEIGHT, BRAIN WEIGHT, GRAMS. GRAMS PERCENTAGE PERCENTAGE AGE NO. SEX OF WATER AGE NO. SEX OF WATER Control eee Control Sican days | days 35 C=107+429 1 Adgiaieesiesy |) (C= 1c 1.492) 1.422 S* = 207 +19 80 .070|79 .820 SS) = lee 79.160)78.700 35 Oy aet! Se le 1.414) 1.324) 57 CG = ile Ib (led) Al .a3hs5 S=1c7°+19 80.560/79 . 960 Ss) = ile 79 .000|78 . 500 36 C=2¢7 +19 1.440) 1.377| 60 C= ie 1.598) 1.470 S =2¢ +19 80.240. 80.070 Ss) = 112 78 .730)78 .330 37 C= ie 1.483} 1-415) 61 GS ile 1.561) 1.528 Si = Ile 79 920/79 .720 'S) = lee 78 .910/78 .620 48 C=107+19 1.490) 1.338) 61 C= iL © 1.750) 1.534 S=2¢7 +19 79 450/79 .030 = 1 © 79 .060)78 .930 55 C= ie 1.434 1.397, 68 C= ile 1.596) 1.455 S = ie 79 .170|78. 780) S=1¢0°++12 79 .000|78 .370 *S. Sudan III fed. the other cases where their controls did not receive any oil at all. This shows that the olive oil alone, in the amounts here given, does not produce any noticeable effects on the organ weights. The percentages of water in the various organs are given in table 5. From table 5 we see a considerable difference in the water content between the control and test rats. The organs which give a diminished percentage of water are the liver, spleen, kidneys, and brain, while those which give an increased percentage are the blood, pancreas, and lungs. The greatest difference between controls and test rats is shown in the blood, pancreas, and liver. Although the brain gave a very small difference in the water content, nevertheless this difference is highly constant. With a view to illustrating the constancy of difference in the percentage of water in the brain, I have compiled table 6. FEEDING SUDAN III TO YOUNG ALBINO RATS ie In this table the means from all the twelve series of experi- ments are given. We notice that in every instance the per- centage of water in the Sudan III brain is less, and thus the dif- ference in the water content between the control and the test rats cannot be doubted. The full meaning of these alterations in the percentage of water in the several organs I am unable to explain; nevertheless, it seems evident that in the case of the liver the diminution of water is mainly due to the accumulation of fat. From two determinations I have obtained the following results for the amount of fat in the liver: CONTROL SUDAN IIt per cent per cent Wasenle rn gs tsiess es OL OEO2 35.52. | Cold ether only used. Caserta soe os hoe eOROA 36.95 | Ether and alcohol extraction. The histological examination shows clearly that the greater part of the liver was infiltrated by fat. In the case of the brain I found a significant increase in the amount of lipoid in the test rat, as will be seen from the following: CONTROL SUDAN III per cent per cent ‘Alcohol-ethen extracteee eee ee ee 42.15 43.94 This increase in lipoid may account for a slight diminution of water content in the test rat, but this requires further study. A slight diminution of the water content in the case of the kidneys and the spleen may also be due to a slight increase of ] pin content as in the case of the brain, but I have no data to prove this. The increased water content noted in the lungs and pancreas I am also unable to explain at this moment. It seems, however, highly probable that in the case of the pancreas, hyperfunction in connection with the metabolism of fat may account for this. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 1 114 S. HATAI As to the high water content of the blood of the rats fed with Sudan III we may attribute it to the condition of profound anaemia. The blood examination kindly made by Dr. Rivas (University of Pennsylvania) shows a loss of 12 to 40 per cent in the number of erythrocytes. I have made an alcohol-ether extraction of the dry blood and obtained the following results: CONTROL SUDAN III ao per cent per cent Alicohol-etherextract:........ sc seer o. 1.56 3.66 This higher content of fat in the blood may indicate that the elimination of fat from the blood has been much disturbed. GENERAL REMARKS Various alterations produced as the result of feeding Sudan III to the growing albino rat have been presented. We found that the alteration is not limited to a mere retardation of the growth of the body and organs, but that the composition of these organs is also modified. It is strange that most previous investigators who fed Sudan III to rats, as well as to several other animals, failed to notice any toxic effect of this dyestuff. This failure was perhaps due to the fact that most investigators used adult animals for their investigation, without very careful testing, while Sudan III is strongly toxic only to the growing animals. Riddle, who fed Sudan III to young chicks, noticed a retarda- — tion of the growth of the body, as well as defective feathers. This investigation by Riddle, on one hand, and my own, on the other, shows clearly that Sudan III is injurious to young animals at least. I have already mentioned the high resistance of the rats whose bodies weigh more than 50 grams, but I have not tested the reaction of the fully grown adult rats to Sudan III. Ihave, how- ever, tried feeding Sudan III to female rats which were nursing young. In these cases a toxic action of Sudan III was found without exception. The lactating glands react promptly to the dye, and failure of the secretion of milk is evident within a day FEEDING SUDAN III TO YOUNG ALBINO RATS 15) or two. Thus in many instances I was obliged to give the mother food free from Sudan III in order to keep alive the young rats which were depending on her. So far as these nursing fe- male rats were. concerned, the effect of Sudan III is similar to that in young rats which have just been weaned; that is, it arrests growth and causes emaciation, curvature of the spine, rough hair, etc... No examination of the organs in the suckling young has yet been made. It seems clear from this that Sudan III is toxic to nursing females, as well as to their suckling young. It is of interest to note that extraction with ether shows a small trace of Sudan III in the following organs: liver, pancreas, lungs, and kidneys, but fails to show even a trace of Sudan III in the brain, spleen, or heart. Corper (’12) demonstrated the pres- ence of Sudan III in the liver, and often in the lungs, of guinea- pigs, but always failed to find it in the brain, spleen, heart, testes, or adrenals. Mendel and Daniels (12, ’13) found Sudan III in the liver (four out of five cases), but never in the kidneys of the rat. This discrepancy as to the presence of Sudan III in the kidneys might be due to the fact that the rats used by the pres- ent writer were young and were also heavily dosed. Highly interesting are the questions how Sudan III produces such profound alterations in the growth of the body and organs and how we are to explain the chemical alterations found in the organs. Riddle, who noticed a retardation of growth and the formation of defective feathers in the chicks fed with Sudan III, considers that the stained fat becomes less easily available to the organism and thus Sudan III produces effects similar to simple starvation. This question of the availability of stained fat was taken up by Mendel and Daniels, who, however, combat Riddle’s conclusions and think that the stained fat is just as readily available as un- stained. Although I have not made a special study on this point, nevertheless it seems clear that the alterations found in the test rats cannot be explained, as merely the result of simple inanition, as proposed by Riddle. A high degree of anaemia, fatty infil- tration of the liver, as well as the nephritic condition, point to pathological alterations rather than to the phenomenon of simple inanition. 116 S. HATAI It appears to me, on the other hand, that in the test rats the normal metabolism of fat must have been much disturbed, as can be seen from the presence of a large quantity of fat both in the liver and blood, and thus we may consider that the stained fat also is less readily utilized. Another interesting question is the origin of the lipoids con- tained in the organs. We found that Sudan III is practically absent, even after five weeks of continuous feeding, in nearly all the organs so far examined. Particular interest attaches to its absence from the brain in which nearly 50 per cent of the solids are represented by lipoids. Moreover, the amount of Sudan III to be recovered from the liver, pancreas, lungs, and kidneys is almost negligible. Microscopical examination fails completely to reveal any trace of Sudan III or stained fat in the organs even when the extraction mass yields a trace of it. Since Sudan III is so intimately combined in fat, and wherever fat appears Sudan IIT is always found, and furthermore since Sudan III stains not only fat or oils, but fatty acids also, it seems prob- able that if the cell elements should absorb the fat or fatty acids for building up the organ lipoids, we might expect Sudan III to be absorbed aiso. There is no a priori reason to suppose that Sudan ITI alone is rejected, while the fat or fatty acids containing it are capable of being absorbed. This result suggests that the lipoids found in the organs may not be formed by the fat or fatty acids furnished from without, but are elaborated by the organ itself; in other words, the organ lipoids are endogenous and not exogenous in origin. CONCLUSIONS 1. Albino rats 27 to 33 days old were fed with Sudan III mixed in olive oil. “The amount of Sudan III given daily was about 8 or 9 milligrams in one series and 10 to 15 milligrams in the other. Borray’ (soils) TNS 4PW | epes-ec )8 “19h xe LS-|LI Ypnpoad ay |8s'0L1 yonpoad “vony BLE sa) >Bvarny | OF-T sr >bvsay es VU “ure lZryJi ~SEeEeTt [%ls-0¢ ~%ITbE |y4ql-7 wor |4%ebsi %CL4T IYGty * ~Rele bell ~%syou | %or Yor — “~ ses, — —— i ae a a “~ SAS — “— aS a, GS 8t BLol 8l|g¢ orsalos 8 |g te 96 97E tI | O & tL tH SI | GS te gol Bs L |G ot zy” Oo 9) LE gi 1 | O vl +z BI g9 tb Gla Ren at ae ge ll ee ea | oe sa. zaolh [hor asta Na oe ae a hh Gia een es 6. Il Dyoyooje | Dijoyooje |W! ouosje | Dijoyooje | D}0Yoo/e | DIjoYyoo}e | pepest} | popesip | popeoiy SUgied YOR |.e4,OLU AJ UC] oUFEF JU |spUeied Yjog] 1ayjoul AiUO} -vaype fAjUO|sjud1ed Yog}o4f,OUL AuQ}oupef Aue +6S sor wl id (Pepn|oul js-lif SUOlpe.18Ues] |e) Suaiorsues eae yea wu Ser Gsene polepng‘, uadsep D'OYOd|e yoyo De fo SJCUNUB [IY] fo spusied ym sjeuUe DIOYUOIV DIAOHOOIY SLNAYYd HLOgd HO SOITOHOONY MAHLOW ‘OITIOHOO1V S| MEHLW4 AINO NSHM ANSA9OUd AHL VO S1O4sdda GHL 4O SISKIVNY NY We Sala shy ae (uoltye.rauas ysalf) spUded POPE} YyIM S|EUIUE DOGO 170 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU Table 3 is an arrangement of the records of animals on the basis of paternal and maternal aleoholism. The first group of animals are those from parents treated directly and having no other alcoholic history. Thus the total 153 differs from the total 186 animals with treated parents in the third column of table 2, since in thirty-three cases the former group had not only treated parents, but also treated ancestors. The second group contains all the animals with parents of alcoholic descent, but not di- rectly treated, the total number is 408; these are the same ani- mals that compose the fourth column of table 2. The third or last group contains all of the 594 non-inbred animals of the alcoholic lines. The individuals in each of the three groups are separated into three classes. The classes of the first group are those with only father treated, those with only mother treated, and those with both parents treated. In the second group the young are classi- fied as those from only father aleoholic, which means the father was descended from treated ancestors which may have been either treated males or females. In other words, this is not the record of a pure alcoholic male line, but merely the alcoholic effects, if any, that reach the recorded individual through an alcoholic father regardless of the origin of his alcoholism. The second column of this group shows the records of animals from aleoholic mothers. Here again the mother’s alcoholism may be due to treatment of any of her ancestors, male or female. It is not a purely female alcoholic line, but a maternal alcoholic line. The third column of the second group shows records of animals from parents both of which were alcoholic. In the entire second group the aleoholism of the parents is ancestral, not being due to direct treatment, while in the third group the aleoholism is either direct, ancestral, or both. The third group is, therefore, an arrangement of all the animals from alcoholic lines for a comparison of the influences of maternal and paternal alcoholism. In the first column of table 3 it is seen that when the father only is treated the results contrast decidedly with the control. MODIFICATION OF THE GERM-CELLS IN MAMMALS Al There is a high percentage of small litters and a low percentage of large litters, thus giving next to the lowest average litter con- tained in all the records, only 2.80 against 2.77 for the normal. The average litter weight is very low on account of the small average litter size. When this is corrected for the proportion of weight to number of individuals in the control litters these small litters from the treated fathers weigh more for their size than do the control, being over 6 grams heavier. This is not an actual advantage since the majority of young born in small litters of one and two are larger than those born in high litters of four or five. The percentage of mating failures is unusually high, 23.52 per cent against only 4.54 per cent in the control. All of these facts would seem to indicate that the treatment of the fathers had evidently lowered their productivity or fertility, causing them to fail to sire offspring in almost one-quarter of the matings and to beget unusually small litters in the other three-quarters of the cases. There must have also been a high ‘early prenatal mortality’ in view of the remarkably great per- centage of small litters and high percentage of mating failures. We must necessarily divide the mortality into prenatal and postnatal, and the prenatal again into ‘early prenatal,’ as in- dicated by the small average size litter and high number of mating failures, and ‘late prenatal’ based on the exact observa- tions of absorptions, abortions, and still births. Two-thirds of the offspring from treated fathers survived against over three-fourths from the control. The prenatal mor- tality is a larger proportion of the total than in the normal. The total mortality when corrected to the normal rate for the differ- ent-size litters in which the animals were born is 178 in terms of the control as 100. This is only slightly below the mortality rate of 189 for the entire non-inbred alcoholic group. When the mother alone was treated the records of the off- spring differ considerably from the above. The percentage of small litters is only slightly higher than the percentage of large litters, and the average-size litter, 2.78, is as large as the nor- mal. There are very few mating failures, in this regard again 172 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU almost a normal record. The productivity of these treated mothers is high and the size of the litters would indicate a very low ‘early prenatal mortality.’ Here, however, their good records stop. Although the litters contained as many individuals as the con- trol litters, their average weight was 26 grams below the nor- mal. The large litters from treated mothers actually weighed only as much as the very small litters from treated fathers; there- fore, the individual members of the litters from treated mothers were unusually small animals. The ‘late prenatal mortality’ was proportionately very high—three times the postnatal. Thus many of the young died in utero or were still-born, and those that were born alive were small specimens. ‘The total mor- tality was 51.08 per cent, corrected for the litter sizes, and ex- pressed in terms of the control as 100 it becomes 281—the highest mortality on record. We see from the table that treating the mother with alcohol does not appreciably affect her productivity, but greatly depreci- ates the quality of offspring to which she gives rise. While in the case of the alcoholic father the productivity is greatly re- duced, and although the quality of offspring which he begets does not compare favorably with the control, it is considerably superior to that from the treated mother. In the treated mother the alcohol may act not alone on the ova or germ cells, but on the developing embryo as well, while in the father it acts, of course, on the germ cells alone. Does the difference between the qualities of the offspring from these two cases represent the action of the treatment on the developing young in utero? Further, does the reduced productivity on the part of the treated male indicate that the spermatozo6n or male germ cells are more sensitive to the treatment than the egg? The remain- ing columns of this and the following table may throw some light on these questions. During the period of the experiments now under consideration practically no matings between treated males and females have been made, as the third column of this group shows. MODIFICATION OF THE GERM-CELLS IN MAMMALS 173 The next group in table 3 are animals derived from parents of alcoholic descent which had not themselves been treated. These are the same 408 animals recorded in the fourth column of table 2. The first class in this group are animals obtained from fathers of alcoholic ancestry and normal mothers; the second class are from mothers of alcoholic ancestry and normal fathers, and the third class are animals produced by two alcoholic parents. As mentioned above, the alcoholic father or mother may owe their condition to either male or female or to both male and female ancestors. These are not purely male or female alcoholic lines such as will be found in the next table. A comparison of these three columns with the normal records shows clearly the alcohol effects, though not so strongly ex- pressed as when the father or mother is directly treated. The father and mother columns of this group differ very little from one another, which is in marked contrast to the striking differ- ences when the fathers and mothers are directly treated, as seen in columns | and 2. In the present columns all of the modified conditions are due to an injury of the germ cells in the treated ancestral generations. This is equally as true of the aleoholic- mother column as of the alcoholic father. For example, the mortality records in the alcoholic father and mother columns are about the same, while there is a remarkable discrepancy between the mortality records of young from treated fathers and treated mothers in the first two columns. The extremely high mortality, largely late prenatal, among the offspring of directly treated fe- males is to some extent due to the direct action of the alcohol upon the early developing embryo in utero. If this action could be eliminated the treated father and mother columns of the first group might become as nearly similar as the alcoholic father and mother columns of the second group. The individuals in the latter two columns are on an average about the same dis- tance removed from the ancestral alcohol treatment, and, there- fore, the records would be little affected’ by a correction,on the basis of the generations treated. When both parents are from alcoholic ancestry the produc- tivity is considerably lowered as shown in the third column by 174 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU the high percentage of small litters and low percentage of large litters and consequently the very low average litter of 2.31. This is likely due to the male partner in the combination, as the preceding columns would suggest. The average litter weight, however, is high, so that the individual members of the litter are as heavy as the normal; this, again, may be due to the male influence as expressed in the high early prenatal mortality. The mortality records, though markedly inferior to the nor- mal, show an advantage over the two previous columns. There is probably a high ‘early prenatal mortality’ as indicated by the low average litter, but the ‘late prenatal mortality’ is lower than in any of the foregoing columns except that of the treated fathers, where again the Jitter was very small and the prob- able ‘early prenatal mortality’ high. This close association be- tween the small litters and the low late prenatal mortality makes it seem all the more probable that the litter size is associated with an ‘early prenatal mortality’ that occurs so near the begin- ning of development that it cannot be directly observed. On the other hand, this result could be interpreted as due to a lowered fertility. If this were brought about through an elimi- nation of the weaker germ cells we might except also the asso- ciated low late prenatal and postnatal mortality, and would have a condition in exact accord with Pearl’s interpretation of the results on fowls. We should be glad to accept such an ex- planation, but for the considerable amount of evidence in our records which points towards a high ‘early prenatal mortality’ rather more than infertility as the underlying cause of the small litters and low late mortality. It must also be remembered that the infertility among the fowls was found in the females as well as the males, while here it would be confined to the males only. The slight advantages which appear in favor of the records from both parents alcoholic as compared with records from alco- holic mothers or fathers are due largely to the distance from the treatment of the generations concerned. In the majority of cases the generations are more remote in the both-parent column than in either the father or mother column, and on the basis of MODIFICATION OF THE GERM-CELLS IN MAMMALS E75 the evidence shown in table 2 this may readily explain the apparent advantages. The last three columns show the results of the first two groups combined and in addition contain a few records from mixed eases that could not be properly included in any of the previous classes; for example, animals with one parent of alcoholic ancestry and the other parent directly treated, ete. Here again there is considerable contrast between the alcoholic- father and the aleoholic-mother columns, these differences being due to the influence on the totals of the F, records from the treated-father and treated-mother columns of the first group. The productivity when only the father is alcoholic is low, the litters being small and over 21 per cent of the matings result in failure. It may be inferred that there was a rather high ‘early prenatal mortality.’ The average litter weight, however, was about as good as normal. The late prenatal and postnatal mortality records are better than those from the alcoholic mothers. The average-size litters from the alcoholic mothers was rather large and the mating failures were much less frequent than from the alcoholic fathers, indicating a lower probable ‘early prenatal mortality.’ The average litter weight was lower than from alco- holic fathers, taking into account the size of the litters in the two classes. The total mortality from alcoholic mothers was high and the proportion of late prenatal to postnatal was ex- cessive. It is thus seen that a high prenatal mortality is fol- lowed by a low. postnatal death rate, and this is in accord with our assumption that a high ‘early prenatal mortality’ will be followed by not only a low postnatal, but also a low late pre- natal mortality. In other words, the more thorough the elimi- nation of defective embryos and fetuses the greater the prob- ability of survival for the selected few that remains to be born. The last column with both parents alcoholic has a mortality record as good as the alcoholic-father column and better than the aleoholic-mother, but this is only apparent and not real. The column contains only one individual from directly treated parents, and consequently the alcoholic treatment was applied 176 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU on the average to more remote generations than was the case in the two single alcoholic-parent columns. In spite of the generations concerned, there is a higher per cent of small litters and a lower per cent of large litters here than in any other class in the entire table. Consequently there is also the lowest average litter. In so extreme a case there was no doubt a high early prenatal mortality. The average litter weight is actually low, but allowing for the small-size litter the average birth weight of the individuals is about as much as the control, again indicating that an individual selection has occurred through an elimination of the weaker embryos during the early develop- mental stages. The extremely small-size litter and the high ‘early prenatal mortality’ may also in addition to the generations concerned ex- plain to some extent the relatively low total mortality and es- pecially the lower rate of late prenatal mortality as compared with postnatal. The questions involved in the present section may be still further analyzed by rearranging the data on the basis of only male ancestors treated or only female ancestors treated instead of only father alcoholic and only mother alcoholic. Table 4 pre- senting this arrangement will be reviewed in the following sec- ‘tion, after which several points of interest may be _ better discussed. 9, A COMPARISON OF LINES FROM ONLY MALE ANCESTORS ALCO- HOLIC WITH LINES FROM ONLY FEMALE ANCESTORS ALCOHOLIC AND WITH THOSE FROM BOTH MALE AND FEMALE ANCESTORS ALCOHOLIC The records tabulated on the basis of male or female ancestors treated supplement the arrangements in table 3, where the groups are classed for only father or mother alcoholic. In table 3 the alcoholic father may owe his alcoholism to the treatment of any of his ancestors, either male or female or both. The alcoholic effects, if any, are there due to the paternal ancestry. The same applies to the groups with only mother alcoholic. MODIFICATION OF THE GERM-CELLS IN MAMMALS 177 In table 4, on the other hand, the groups with only male ancestors treated owe their modified conditions, if such exist, entirely to the effects of the treatment on male animals, though the individual being considered may have inherited this alco- holic effect through its mother. Thus animals in the columns with only male ancestors treated were not necessarily derived | from alcoholic fathers, but may have been produced by alco- holic mothers which, however, owe their alcoholic condition to one or more treated male ancestors. The table permits a comparison of the action of the treatment on the male germ cells and the transmission of the effects with the action of the alcohol treatment on the female germ cells and the effects transmitted to the different generations. While the last table permitted a comparison of the animals derived from males of alcoholic stock with others derived from females of alcoholic stock. The two tables serve to analyze very completely the problem of the parts played by the sexes in the acquisition and transmission of the effects of the aleohol treatments. The three columns in the first group of table 4 are the same as those of the first group of table 3, being the records of F, animals derived from treated fathers, which have only one male ancestor treated according to the table 4 arrangement, and F, animals derived from treated mothers or from only the one female ancestor treated. This group was discussed in review- ing the third table. The points of chief interest in the present connection are the decidedly inferior conditions of the off- spring from the treated females as compared with those from the treated males, in so far as their measured mortality records and birth weights per litter are concerned. On the other hand, the records from treated males suffer as regards the ‘early prenatal mortality’ indicated by the small average-size litter and the high percentage of mating failures, while the records of the treated females in regard to these conditions are equally as good as those of the control animals. The next three columns of table 4 are highly important, since they contain the results of matings when, first, only male an- cestors are treated; second, when only female ancestors are THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 26, NO. 1 178 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU D\= igh= Ke= 8lis %T VAN (480g) 97°81 N Cce'ee) ALIS Axes AIL 2abc %) "11 @ 14 1) ne GB el) Und ia 4999 (oom) tr] ol | Sy ta ra atl AULA Le 156 oOolje & I peop |eFoL (HI: SE) (%LE ot) (% or) SUJUOWL ES UlYEIM Palg UNOG|1}S J ANPEWd.14 ‘poqdiosgy + q i © [Peet = I J i en eee Ml (Se aroma A icant: (%tPbL) (49% vl) (%b¥°897 (%99°SL) OG +e & Cl ve (%99°b5) (%te-€9) Asc yores STL 08 leccclellleeMe west Oo CL EPLE + |e Ol t9 +5 El Coy Say em SLY See Ge ae (%EC'¥9) 2) a fe © a {2G (Z8G'+9) (49899) Ont J brOnnl Ste EOC aa AV UD He R16 Side GG Ah U-ctl9l 9 inte GG ol a Fae Bf ik ce 14 © Saris 6 I (Sb IL) Acces LILA ZULLL ust ol Séewl € Siete ‘ (459) (“ib er) (%99°99) ‘oe asters iLs\8 209 toss Hooley veel] SUfUOU ¢ m% {> Cia lke @ I Uy sh 3) rn Chae eG ll \ 1QAO Poet] (4%79°9C) Yr cl ASSSS UBS AEE EL O 9 Sb bbe +i ae & ta” i (%99->9) Preul Ysss Ls corel 09 Gb Iv cl San toa tl TO 09 (%IL syerne} wees ee) s iw} aw. Lg i Li yenposd oy | 82 0Li pnpordasvamny BLE Arp army | OFT Layft|>Svsday stl (% eb tol) +1 sofa St CLI ypnpoad sony OST AAP amay’ LEG Amp smn] 799 T Arylyaray | cH eo AB|P | srary let Aa Ara | 9a°U Aa aay %rb bl %ChBS 1%4'SI Are |esecc XYLLIE |%CE'bl SESE | %LIb1 YRIT |VITEC Yar-eS “bell %sFot | or 2,04 . A ES A~ i, an ~ —— ————— a pao _a i ~_— “a BRS a O JE 18 85 GI} O 8z gol ov G | SI 9€-o0b +7 Bl 9 9 18 OG +) oO Ul & BI + | ol #0 9 Be 0) oy SM (te fay O tl $e gig Ae ton GN JE AA. Gi Cee eg [ll Sea el |) Sp ae ete ee At pecte weno Ge payeaty SAOL| POpCatEStof| papyea-tyS-io}|Popea.lf S10} |payeaty S10] |Po-eo.ly SIOF] Poet} SAAS (60 SG popyeot, | popesd.t} S90 UPR 9|-S2ouE §AUO]-sarue QAUC|Soque sg Y|sa2ue §A|UO] sarue gAuC|sjuo.ied yog}aupou 4(u].iaypyef uO 6G sor ( ; (ys-uf ¢dooxa ee ele 1s.f pup ire ouar ile) (SLOFEsoOUDS 1OUZO) Poped.|p (uoiperduags ysify spueied YUOPSEP | fra r1p fou (ng GLaISOP I!/0 YOrje |popee ee ee ie. Tee a rere SuGC Cone POPEOAL YPM sjeuUUe D{OYOI\ Ga LVaN_L SHOLSFONVY DIVWa4 & SIVW HLOd HO GSLVSYUL SYOLSSONY S1VWwad ‘OsIWaylL WOM SYOLSSONV SIVW XINO NSHM ANSOOUd SHL VO SLOSASS. AHL dO SISKIVNY NY IN SI WEIS MODIFICATION OF THE GERM-CELLS IN MAMMALS 179 treated, and third, when both male and female ancestors are treated with all first generation, F, offspring, excluded. The modified conditions shown by these records are due to an heredi- tary transmission of the defects and not in any case to the direct influence of the treatment on the developing animals. The fourth column from only male ancestors when compared with the normal stock in tables 1 and 2 shows a higher ‘early prenatal mortality’ based on the average litter size and high number of mating failures, a lower average litter weight, a higher late prenatal mortality, and a higher total mortality. The re- sults of these matings are, therefore, from any point of view worse than the results of normal matings. And they prove the hereditary transmission of the defects arising from the treat- ment of the male animals. The same can be said for the female column, the results shown here also being worse than from the normal matings. The ‘early prenatal mortality’ is higher, the average litter weight, indicating the total productivity is smaller, the late prenatal and total mortality are higher, while the mating failures are about the same as in the control records. Therefore, the treat- ment of female individuals also induces effects that are trans- mitted to later generations through the germ cells. When, however, the records of the fourth and fifth columns are compared, it is found that the treatment of male ancestors gives in every point considered more marked effects on the qualities of the descendants than the treatment of female an- cestors. Among the descendants of treated males there is a higher early and late prenatal mortality, a decidedly higher total mortality, and more mating failures than among those from treated female ancestors, while the first and second columns show the opposite to prevail so far as litter weight and mortality are concerned for first-generation, F,, animals from directly treated males and females. These inferior results, so far as late prenatal and total mortality are concerned on the part of the offspring from the directly treated female, may be interpreted as due to the direct influence of the treatment upon the young in utero. On the other hand, the improved records from the 180 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU treated-female line during later generations can probably be explained in part by the higher mortality of the offspring in the first generation, thus bringing about a greater elimination of the weaker individuals. In other words, these animals from only female ancestors treated have withstood a somewhat more severe selection during the first generation than have the off- spring from only treated male ancestors. As a final possibility it must be recognized that the superior records of the late generations descended from treated female ancestors as compared with the records of similar generations descended from treated males, may be due to a smaller influence of the alcohol treatment on the ova, or female germ cells, than on the spermatozoa. The sixth column from treated male and female ancestors shows, in comparison with the two preceding columns, the highest ‘early prenatal mortality’ based on the many small-size litters. There is also the lowest average litter weight. The late prenatal mortality, total mortality, and mating failures, while higher than for the treated-female line, are lower than in the treated-male line. The complete absence of matings between directly treated animals as seen in the third column of this table makes com- parisons and explanations of the results in this sixth column very difficult. The last three columns of the table show the combined re- sults from all generations. The column for both male and female ancestors treated shows the highest early prenatal mor- tality, the male treated line the highest late prenatal mortality, and the female treated line the highest postnatal mortality. In general it may be stated after reviewing this and the fore- going table that the treatment of males produces in their de- scendants a high early mortality, especially early prenatal. The treatment of females produces in their descendants a high later mortality, especially late prenatal and postnatal. The treatment of male and female ancestors produces in their descendants the highest early prenatal mortality, but the lowest late pre- natal and postnatal mortality. MODIFICATION OF THE GERM-CELLS IN MAMMALS 181 In table 4, male and female ancestors treated does not neces- sarily indicate that all records in the column were derived from matings between two alcoholic parents, since both males and females may have been treated among either the ancestors of the mother or the father, but not necessarily both. For this reason the sixth and ninth columns of table 3, in which both parents were in all cases from, alcoholic ancestry, show more decidedly that two alcoholic parents when mated together give the very highest early prenatal mortality, but a low late pre- natal and postnatal mortality. This last conclusion is extremely interesting in connection with Pearl’s results on fowls. Pearl found that when two alcoholic fowls were mated to- gether, the percentage of infertile eggs was higher than from any other combination, while the prenatal mortality, embryos dying in shell, and the postnatal mortality- were the lowest. This is exactly what the guinea-pig records show, provided our ‘early prenatal mortality’ (indicated by the small litter size, the frequent mating failures, and the observed mortality occurring in utero during all later stages of development) can be considered the same as many of Pearl’s ‘infertile eggs.’ Without intending any adverse criticism of the designation ‘infertile,’ we may again suggest the possibility that a certain proportion of these eggs had really: begun development, but had died in the early cleavage or gastrular stages, and yet on ex- amination, other than a minute microscopic study, they ap- peared as infertile or unfertilized eggs. If this were true, they could be classed in the early prenatal mortality records. Such an adjustment would serve to harmonize the fowl and guinea- pig records in another important respect. Pearl has attributed the good qualities of the offspring from his alcoholic parents to a germinal selection which has tended to cause all weak germ cells to be completely put out of commis- sion by the alcohol treatment and only the very best have sur- vived to produce embryos, and these therefore show a low per- centage of deaths in shell and a low postnatal mortality. A selection is also playing its réle in the case of the guinea-pigs, but here it isnot acting alone on the germ cells, but more evi- 182 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU dently on the developing individuals. The selection in our case is a continuous selection of individuals, eliminating, no doubt, certain of the least resistant germ cells, but continuing to act on the embryonic population to eliminate the most defective of these during very early developmental stages, and so on until the individuals born are a mixture of strong specimen and others only sufficiently strong to have reached birth and possibly to survive in a subnormal fashion for a shorter or longer time. This continuous, both germinal and individual, selection seems to us more to be expected than the abruptly broken germinal selec- tion advocated by Pearl, which completely eliminates all weak germs, and therefore no weak individuals begin development. We must admit that the data from Pearl’s double alcoholic matings considered alone strongly suggest only a germinal selec- tion, but the results from our double alcoholic matings, while leaning in the same direction, still show a greater late prenatal and postnatal mortality than do the control matings, and in addition present much evidence to suggest a very high early embryonic elimination. This same early embryonic elimination may be included among the high percentage of infertile eggs re- sulting from the matings of two alcoholic fowls, and in the case of the fowls it may be so much more severe that the later mortality records compare ‘favorably with the control. This again would lead us to an abrupt break after the high very early prenatal mortality and might be thought to vitiate our entire supposition, yet the guinea-pig records show almost all gradations up to the condition for the fowls. Our results show that in the alcoholic lines the higher the early prenatal mortality and consequently the smaller the aver- age-size litter, the lower the late prenatal and postnatal death rate, much as Pearl also finds for fowls. These findings will be still further discussed in connection with the sex ratio, table 6. MODIFICATION OF THE GERM-CELLS IN MAMMALS 183 10. TREATING MALES WITH ALCOHOL FOR ONE AND TWO GENERA- TIONS AGAINST TREATING FEMALES FOR ONE AND TWO GENERATIONS An experiment has now been in progress for some time in which straight male lines have been treated with alcohol for sev- eral generations in order to compare the results with those from the treatment of straight female lines for several generations. That is, the original males are treated, their sons are then di- rectetly treated, their grandsons, great-grandsons, and so on; these we consider the straight male lines. The treated females, their directly treated daughters, granddaughters, and so on constitute the straight female lines. There are now a few third- and fourth-generation individuals, though not a sufficient num- ber to tabulate. We shall thus for the present confine our attention to the records from the originally selected and treated males and females and the treated sons of these males and daughters of the females. The records are arranged in table 5. In considering the table it must be stated that the original animals in this experi- ment have been carefully selected large strong specimens that were particularly good breeders. Such a choice has been made on account of the severity of the treatment to which the de- scendants are to be subjected through a number of generations. Only the best animals are likely to produce descendants suffi- ciently strong to be treated with alcohol and to continue to re- produce for one generation after another. The fact that such a selection is possible does not refleet on the general population, since no population is so perfect that certain individuals are not better than others. This selection probably accounts for the presence in all the groups of table 5 of some offspring of unusually large size. The pedigrees or conditions of the animals in the different generations are expressed by the following symbols or formulae. A normal animal is represented by the letter N and one treated with alcohol by A. The symbol for the male is placed to the right of that for the female. Thus the first column NN are the normal control animals for comparison, the second column NA 184 . CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU (%sL I) Yi O Oo (Ses +) (%rS°9l) %q La LEE “ZEEE Bor | ° i | oF ve SS G Ss SLE=S%LY- I N(Zit'S9) AIV-I9 %05 oy @ 4 | © es eee (%EE-E1) (% CCT) oy fj | © + | a Gd Ze (%99°98) (2% LULL) mated with normal females, N. The next straight male generation treated and paired with normal NA females would be expressed by N “A, the offspring from such A a combination would have had their father, a grandfather, and a great-grandfather treated with alcohol and their mother, grandmothers, and great-grandmothers all normal, and so on for later generations. Animals of these higher pedigrees will be recorded in a future communication. In the table 5 only records from treated fathers are given in column 2 and from treated father and grandfather in column 3. The fourth column shows records from normal fathers and treated mothers, AN, and the fifth column from normal fathers mated with treated females which were derived from treated mothers, — N. The numbers in all of the columns are rather small, but in every case the records differ from the control. There is a re- markable similarity between the two treated-male groups and also between the two treated-female groups, but a striking contrast exists between the male records as a class and the female records. In the two male columns the average litter is very small and the mating failures high. The percentage of surviving young, though well under the control record, is equally above the fe- male records. The corrected total mortality in both columns is over 180 against 100 for the control. The proportion of late prenatal to postnatal mortality is slightly contrasted in the one treated male generation column, but more so in the two treated generations column. There are no defective animals in the NA column, but a small per cent of such are seen in the N a rou aN £ p- 186 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU The average litter size is high in both female columns and the mating failures lower than in the male groups. While the total mortality is extremely high, being in the two-generation treated column on the basis of litter size over twice as high as either treated-male column and about four times the control record. The proportion of late prenatal to postnatal mortality is in the first female column over three to one and in the last column over six to one. There were some defective animals in both female groups. The records of the females in this and the two preceding tables are out of accord with the records from fowls and do not fit an explanation based on a germinal selection or partial infertility. The total productivity is good and the late prenatal and the postnatal mortality are high. It is seen at once that the records from the treated-female generations are far worse than from the treated-male generations; in fact, so much worse that we are led to conclude that the alco- hol has not acted on exactly the same things in the two cases. The increased effect of the treatment in the double female column is much more evident than in the two male generation column. The results in the male columns are due only to an action of the treatment on the spermatozoa or male germ cells, while the re- sults in the female columns are also due to the effects of the treatment on the germ cells or ova, but more largely to the effects of the alcohol on the developing embryos within the uterus of the treated mother. Provided the effects of alcohol were equal on the sperm and ova of guinea-pigs, the difference between these two sets of records would then represent the action of the treatment on the developing embryo itself. Although the records in table 5 involve only small numbers, we are led to believe that they represent the true trend of the effects, since they harmonize so perfectly with the data of dif- ferent composition yet much more complete shown in tables 3 and 4. Here again, as in tables 3 and 4, the treated-male lines show the early prenatal mortality (based on the average litter size and frequent mating failures) to be unusually high while in the female line it is low. In the male lines the late prenatal and MODIFICATION OF THE GERM-CELLS IN MAMMALS 187 total mortality is low while in the female lines the late pre- natal mortality is extremely high and the total mortality very great. Finally, this table may be considered as supplying evidence of the increased effect of higher or longer alcoholic dosage. The double male records which have usually been derived from ani- mals that have had longer or more treatment during the two generations are somewhat inferior to the one generation male treated records, and this inferiority is very much more decided for the female groups in the case of the higher-dosed two-genera- tion records. 11. THE SEX-RATIO IN RELATION TO PATERNAL AND MATERNAL ALCOHOLISM AND TO THE TREATMENT OF MALE AND FEMALE ANCESTORS WITH ALCOHOL In the last group of table 3 it will be remembered that all of the non-inbred acoholic descendants were separated into three’ classes with only father alcoholic, only mother alcoholic, and both parents alcoholic. Again, in the last group of table 4, these 594 animals were rearranged into three classes, from only male ancestors treated, only female ancestors treated, and both male and female ancestors treated. The difference between these classifications are made clear in the discussion of tables 3 and 4. If we now record the number of males and females com- posing each of these six classes and express their sex-ratios on the basis of the number of males to every 100 females, a most peculiar result is obtained, and one for which it is very difficult to give a completely satisfactory explanation. The number of males and females and their mortality records in each of the six classes are shown in table 6. As a standard of comparison the 233 control animals are similarly recorded in this table. For further comparisons a total sex-ratio and the sex-ratios for animals born in different size litters are given be- low the table. The total sex-ratio calculated for about 1600 animals is 109.6; that is, 109.6 males to every 100 females. Many of these animals were from alcoholic lines, so that this sex-ratio may not be exactly normal. Yet a further perusal of the table STOCKARD AND GEORGE N. PAPANICOLAOU 188 CHARLES R. (8%t) 99-901 “ ” “yu «4 a (BLS) +L1Ol “ “ngewe i a“ (BLS) 98 v1 “a im wire fo 4 a SALE ee “ . (Fb1) BU-SH1 slemrmyreyyury fo «a %BIbC Y JoAryosrow jvpo €9°bO! OUCA xXoS JEAOL ee egos : | ae |) eae eee %oal Yoal Phsbse [LILES |Xbo'eb |Vly-ce |wexe | Lol |Zol'bz [Ibe |%e-Jb jyocge |%si-be | ool | Ere |%S 0d | 400) |ASloz | *A9'bI) Doan > Alb 2. i aN Gl | Oe Ee || Neen Oe bean pete cteps| Oa SUPE panel telat SsuluOW Uses ES g tL $9 ol sh 8c I+ ADAO POAT jot oe fi fos | feo | ral 619 el a |S Pl |S popeoit SOL | PopPotd, SiO} | PpopUv4l4 S.loL DYOUod(e VD joyuyoo|\e DWjoUocd|e So2UC & PUB Q | -SedUP + AjUDO | -SooUeg hug | spudted4ysod | sodqyou 4(UO | -reupef AjUO Ul] [EW.O) SYSHLOW UNV SYSHLV4 DMOHOONIVY WOYUS UO SHOLSSDNY S31VWas ONY AWW 40 LNEAWLVSY_L GHL YAL4IVY ANS9OUd SHL 40 SOILVY-KSS SHL UNS Sova si Vt MODIFICATION OF THE GERM-CELLS IN MAMMALS 189 will suggest a tendency on the part of the different alcoholic lines to level the sex-ratio to normal when they are combined as a grand total, yet we are not by any means comparing the sex- ratios from the alcoholic lines with an average alcoholic ratio. The sex-ratios of 194 animals born one in a litter was 113; of 578 born two in a litter, sex-ratio 114.8; of 579 born three in a litter, sex-ratio 101.7, and of 248 animals born in litters of four, the sex-ratio was 106.6. The first column of table 6 shows that of the 233 non-inbred control animals, 120 were males and 106 were females. The proportion of males to females is thus 113.2 to 100; that is, a sex-ratio of 113.2. The average-size litter in which these ani- mals were born is shown in parentheses in the sex-ratio space as 2.77. The-mortality record for the males was about the same as for the females, having only a very slight advantage. The third column of animals from only mother alcoholic are also in the majority of cases individuals from only female an- cestors treated, the sixth column, but not entirely so, as many of the mothers may have been alcoholic on account of a treated father or grandfather. In general, however, the third and sixth columns are rather the same in composition, the sixth being a purely female treated group while the third column is largely so but not entirely, and while not necessarily to be treated together they may be considered in connection with one another. A point of immediate notice is that the sex-ratios in both of these columns, 96.8 and 86.5, are very low. When only the father is aleoholic, second column, or only the male ancestors are treated, fifth column, the sex-ratios are higher, 101.7 and 109.1. While if both parents are alcoholic, fourth column, or male and female ancestors are treated, seventh col- umn, the sex-ratios are very high, 121.1 and 123.5. It must also be noticed that these differences in sex-ratios are more accentuated in the last three columns, giving the descendants from only female ancestors treated with alcohol the lowest sex- ratio in the entire table; those from only male ancestors treated a considerably higher ratio, and from both male and female ancestors treated the highest sex-ratio for all groups. A differ- 190 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU ence of 37 between the number of males to 100 females in ani- mals from treated-female ancestors as compared with those from both male and female ancestors treated is indeed very great. Are these differences in sex ratio a result of the direct influ- ence of alcoholism upon sex determination or sex differentiation? Or are they indirectly brought about by a difference between the early prenatal mortality rates of the two sexes in the sev- eral groups considered? Or are these merely chance differ- ences? There is no doubt that chance plays a large part in the make up of all sex-ratios, but to be consistent in six straight cases as the six groups show can scarcely be dismissed as a chance result. It is very peculiar that these different sex-ratios should coin- cide in a direct manner with differences in the early prenatal mortalities among the several groups of table 6, and thus suggests that the explanation for the sex-ratio differences may be in part at least along the line of the second of the above propositions. After considering this probability of differences between the early prenatal mortalities of males and females, we may then discuss the further possibility of the direct effects of the treat- ment on the sex-ratios. The groups having the lowest sex-ratios, female lines with ratios 96.8 and 86.5, also have, as shown in tables 3 and 4, the largest average litters, 2.69 and 2.66, or the lowest early pre- natal mortality. The lines having a somewhat higher sex-ratio, male lines with ratios 101.7 and 109.1, have correspondingly somewhat higher early prenatal mortalities, as indicated by the smaller average litters, 2.41 and 2.42; while the lines having the highest sex-ratios, double lines with ratios 121.1 and 123.5, have along with these the highest early prenatal mortalities as shown by the smallest average-size litters, 2.28 and 2.37. This is certainly a very suggestive parallelism. And if one now considers the fourth line of the table giving the total dead of each sex, in every column, with one exception, it will be seen that the female mortality is higher than the male. The ex- ception is in the column from both parents alcoholic; here the MODIFICATION OF THE GERM-CELLS IN MAMMALS 191 sex-ratio is very high and yet the late male mortality is higher than that for the females. The total mortality for the females is higher than that of the males, 31.71 per cent against 29.68 per cent. It is further shown below the table that small litters have a higher sex-ratio than large litters, the sex-ratios for litters of one and two young being respectively 113.1 and 114.8. While for litters of three and four young the sex ratios are 101.7 and 106.6. It has been pointed out before that the small litters are often due to an early prenatal mortality which has destroyed some of the original members, and since the sex-ratios of such litters are high the majority of embryos dying may have been females. We may see finally by a study of table 7 that female animals are generally smaller at birth than males in the same litter, and as their total higher mortality would indicate, they are probably also weaker. TAS yl THE BIRTH WEIGHTS OF MALE AND FEMALE MEMBERS OF MIXED LITTERS Number of Litters Total weight |Total weighf | Total excess |Average excess |Percent of excess Of males in|of females | weightof males|weightof males|weight of males grams ingrams joverfemales |Over females |Over females Litters of 105 736! giz Lifters of 13G 9513 Ea Litters of | 125 [8 22 A423 (3856) Table 7 only includes mixed litters; that is, those containing both male and female members. It shows that in 105 litters of two animals of opposite sex ‘the total birth weight of the 105 males was 7861 grams and of the 105 females only 7525 grams, or 336 grams less. The average excess weight of males over females in these litters of two was 3.2 grams, giving a percent- age of excess weight of 2.18 in favor of the males. One hundred and thirty-six litters of three, consisting of two males and one female, are recorded. The total weight of the 192 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU 272 males was 9513 grams; that for the 136 females was 4654 grams. If this total weight be doubled for comparison with the total weight of the double number of males, we have 9308 grams. The males again have a total advantage, amounting here to 205 grams. The average excess weight of the males is 0.75 gram, or a 1.08 per cent excess weight of males over females. It will be noted in this table that the average weight of the individuals is very low. This is due to the fact that a number of abortions in which the sex could be distinguished, as well as - premature still-births are included. These small specimens have brought the average in some cases almost below the birth weight which permits survival. It is also noticed that the 272 males in the second line only weigh about one-quarter more than the 105 males in the first line of the table, and this is due to the fact that there were many more abortions and early pre- mature births of litters consisting of three individuals than of two. While the males in the litters of one male and one female averaged almost’75 grams, the males in litters of two males and one female averaged less than 35 grams. | The third line of the table shows 125 litters of one male and two females. The 125 males weighed 4428 grams which may be doubled to give 8856 grams for comparison with the total weight of 8790 for the 250 females. There is a total advantage of 66 grams in favor of the males. The average excess of male weight is 0.26 gram, or 0.37 per cent over female weight. The last case of thirty-six litters, consisting of two males and two females each, gives a total excess of 85.5 grams to the males. The average excess weight of the males over females is 1.19 grams, and the per cent of excess of males over females is 1.87. It is thus seen that the males born in litters consisting of both sexes possess a superiority in body weight over the females in every combination. We do not attribute this constant excess in favor of the males to a sexual dimorphism in size. In a group of guinea-pigs both young and adult females are often larger in size than comparable males, and no constant size difference be- tween the two sexes is known. It seems more probable that MODIFICATION OF THE GERM-CELLS IN MAMMALS 193 this advantage in weight on the part of the males, the majority of which are of alcoholic ancestry, is in line with the lower mortality records of the males shown in the various columns of table 6. And this may further bear on the explanation of the high sex-ratios in those lines with high early mortalities or small average litters. There is, therefore, much evidence to indicate that among alcoholic guinea-pigs the females very probably suffer a much | higher early prenatal mortality than do the males, and it is shown that the female mortality is higher than'that of the males at all other periods, table 6. Before proceeding further with our theoretical explanation of the different sex-ratios in the several groups, which leads finally to a consideration of views expressed in a previous communica- tion, still another important relationship may be pointed out between early prenatal mortality and the sex-ratio, on one hand, and the late prenatal and postnatal mortality, combined in table 6 under ‘total dead,’ on the other. Stated concisely, the higher the sex-ratio and the early prenatal mortality, indicated by the small average litter, the lower will be the total late pre- natal and postnatal mortality, and vice versa. The columns with the highest sex-ratios, 123.58 and 121.17, and at the same time the highest early prenatal mortalities or the smallest aver- age litters, 2.37 and 2.28, show the lowest late mortalities, 25.55 and 27.12 per cent. In the opposite way the columns with the lowest sex-ratios, 96.8 and 86.51, and the lowest early prenatal mortalities, or the largest average litters, 2.69 and 2.66, have the highest later mortalities, 34.93 and 32.52 per cent. This is in line with what was brought out during the discussion of table 4 showing that the higher the early prenatal mortality, or the smaller the average litter, the lower will be the late prenatal and postnatal mortalities. There is one very evident objection to the foregoing explana- tion of the peculiar sex-ratios as being due to a differential sex mortality during the early prenatal periods. That is, among the normal stock the sex-ratio is rather high, although the early prenatal mortality is probably very low as indicated by the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 1 194 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU large average litter. With a large average litter the sex-ratio should be very low as in the female lines. We could only avoid this difficulty by assuming that the control lines are out of the consideration, since the other sex-ratios being discussed are all shown by modified alcoholic groups among which entirely dif- ferent conditions obtain from those existing in the control. Whereas there are reasons for such a position, it would seem preferable at present to admit that the case of the control is a real objection. And such an objection would serve to indicate that while a higher mortality on the part of the female embryos in the alcoholic groups might actually exist, yet it accounts only in part for the peculiar sex-ratios found. A recognition of the normal record also makes it difficult to account for the very low sex-ratios of the female lines. Here the early prenatal mor- tality was low on the basis of the average size litter, but if any early prenatal mortality did occur it could not have been partial to the female embryos, but must on the contrary have been confined almost totally to male embryos or else a sex-ratio could never fall 25 below the control. Is it possible that wherever a treated male is concerned, as in the male columns and the double columns of table 6, there is a high early prenatal mortality among the female embryos, and on the other hand where only a treated female is concerned there is a high early male mortality? It is difficult to believe so, and therefore differences between the early mortalities of the sexes can, on our present data, only partially explain the sex-ratios found in table 6. This leads to a final ex- planation which may seem highly theoretical, yet it does have a basis of fact. ; In an earlier communication (Stockard and Papanicolaou, 716), we presented some evidence which seemed to indicate a possi- bility that the action of the alcohol treatment not only differed in its effects upon the two sexes treated, but also acted differently on the two groups of spermatozoa in the male, the so-called male-producing and female-producing sperm. We suggested that the action of the treatment was more se- vere on the germ cells of the male than on those of the female; in other words, that the spermatozoa were more susceptible MODIFICATION OF THE GERM-CELLS IN MAMMALS 195 than the ova. The inferiority of the column from male ances- tors treated as compared with that from female ancestors treated in the second group of table 4 seems to substantiate such a position. The possibility exists, however, that the treatments of the male and female ancestors may not have been equally severe, since they have been treated in different fume tanks. This question is now being studied. At any rate, we believe it is proved that the germ cells of the female are as definitely in- jured and modified by the treatment as are the germ cells of the male. This is the point of importance in the present connection. The female offspring from treated fathers were found in the report cited to be inferior as a group to the male offspring as regards their powers of existence and structural perfection. The opposite was indicated among the offspring of treated mothers, the males being inferior to the females. Our explanation of these conditions was that the two classes of spermatozoa which differ structurally also differ in the degrees of injury suffered from the treatment. We are further testing these suppositions by selected matings and hope to report on them in the future. For further details regarding the supposed differences between the behavior of the two classes of spermatozoa, the reader is referred to.our 1916 paper. An explanation of the sex-ratios in table 6 may now be given along similar lines and the peculiarities found among these sex- ratios are exactly in accord with our previous theoretical consid- erations. If the male guinea-pig does possess, as has been claimed (Stevens, *11), heteromorphic spermatozoa, one class with a small Y chromosome, the male producing, and the other class with a larger X chromosome, the female producing, the follow- ing may be assumed: In the treated-male lines the female-pro- ducing spermatozoa are more decidedly affected, possibly on account of their larger quantity of chromatin, and therefore, in the competition to fertilize the eggs they are not so successful as the less injured male-producing sperms. Consequently, more male animals are produced than female. Or, if the female- producing sperm are not in any or all cases actually prevented from fertilizing eggs, nevertheless the individuals produced by 196 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU such a fertilization are inferior and more apt to die during early developmental stages, and thus a greater number of male em- bryos would survive and be born. When the alcoholic mother or early female treated lines were mated with untreated normal males, the sons were inferior to the daughters. Here again, taking into consideration the two structurally different classes of spermatozoa, the normal males paired with alcoholic females contribute a smaller amount of normal chromatin to the complex producing male offspring than to that giving rise to the female offspring. The records of the males are hence inferior to those of the females. And in the present connection such males might be expected to suffer a higher early prenatal mortality and so give rise to the very low sex-ratios shown by the columns from ‘only mother alcoholic’ and ‘only female ancestors treated.’ Such reasoning from the present data is admitted to be highly speculative; nevertheless, if the morphological differences which have been found to exist between the two classes of spermatozoa in a number of animal species have any significance, they must sooner or later be recognized as the underlying cause of such results as table 6 shows for sex-ratios in alcoholized guinea-pigs. These ideas also account for the fact that the sex-ratios of the normal animals is out of accord with the ratios of all the treated groups on the basis of the average litter size. This dis- cord was recognized as a possible objection to the purely dif- ferential sex mortality explanation previously discussed. In the present connection we may take the following position. The normal group has been subjected to no injurious action which has tended to modify the expression of the sex-ratio, while in the aleoholized groups there is evidence of a deviation from the normal, in one direction or the other, depending upon the combination concerned. And this deviation is imagined to be due to a lower fertilizing ability on the part of certain spermatozoa. There is another question to be considered in connection with the differences in response on the part of the two classes of sper- matozoa; that is, the possibility of certain eggs being more sub- MODIFICATION OF THE GERM-CELLS IN MAMMALS 197 ject to fertilization by either the X or Y type of spermatozoa. Even though the egg might be practically equally accessible to both types under normal conditions, a peculiarly affected egg might become much more readily fertilized by one class of sperm than the other, and almost all male offspring might. re- sult in one case and females in the second. One might feel that these are large suppositions on the basis of the minute differences between the two groups of sperm. But it may be replied that the differences are only minute from the standpoint of the minuteness of the structure considered. Corresponding differences between great things would necessarily: seem much more important, but with present powers of observation only very great differences between cellular structures are visible at all. There is evidence from a study of the control of sex-ratios in normal guinea-pigs to indicate that certain females have a very strong tendency to produce male offspring regardless of the male with which she is paired (Papanicolaou, ’15). Other fe- males have as decidedly marked tendency to produce female offspring. Such females may be said to have either a male or female tendency, while other females are in this regard indif- ferent, producing as many offspring of one sex as of the other. These tendencies may be explained in accord with the above dis- cussion as due to a high affinity for one type of sperm on the part of the ova of one female, while the ova of another female are ‘particularly susceptible to fertilization by the other class of sperm. The indifferent females are those with ova which are fertilized equally as well by one type of spermatozoa as the other. There are striking cases among the aseidians and other forms illustrating selective fertilization, and the above suggestions are by no means without foundation. Certain male guinea-pigs are also known to have a strong tendency to beget female offspring regardless of the females with which they are paired. Other males have a high male-producing tendency and still others are more or less indifferent in their sex-determining quality. This may be readily imagined to re- sult from a difference in the activity or fertilizing powers of the 198 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU two types of spermatozoa in certain male animals. And there is evidence to show, as cited in our previous paper, that the fertilizing power of the spermatozoa may be modified in such a way as to render them much less capable of success. If this is the case, we may be justified in assuming that one class of sperm may often, even under normal conditions, be at a disad- vantage as compared with the other. It is even more prob- able that under modified conditions the two morphologically different classes of spermatozoa will not be affected to equal degrees. In conclusion, then, it seems highly probable that the peculiar sex-ratios shown by the several groups of treated animals re- corded in table 6 are in part due to differential sex mortalities during early prenatal stages, on account of the close correlation between the sex-ratios and the average litter sizes. This differ- ence in early prenatal mortality between the sexes does not, however, completely satisfy the case. The sex-tendency of the animals considered and the possibility in the case of delicate treatment of affecting the two types of spermatozoa in different ways or degrees are certainly factors to be recognized in the production of the results obtained. Pearl found that for fowls treated with alcohol the relative proportions of the sexes produced were not significantly different from normal control series. Our results for the sex-ratio of the total alcoholic series agree with Pearl’s findings. The sex-ratio . of the 594 alcoholic animals considered in the present paper is 105.6, which, in view of the numbers involved, is not signifi- eantly different from the control series. Yet studying separately the several groups shown in table 6, we find strikingly wide dif- ferences in the sex-ratios and the arrangement of these differ- ences is decidedly consistent. From the standpoint of the above discussion it seems to us legitimate to consider the six groups individually, or at least as three classes, since there is a probability that different processes or conditions are affecting the results in the different cases. Several recent experiments on the modification of the sex-ratio would tend to strengthen such a probability. MODIFICATION OF THE GERM-CELLS IN MAMMALS 199 12. THE BIRTH WEIGHTS AND RATE OF GROWTH IN THE NORMAL AND THE ALCOHOLIC SERIES In the present section the birth weights and ability to grow of the animals born in the normal and the alcoholic series may be compared. Here again comparisons must be made between ani- mals born in litters of the same size. It may be expressed gen- erally, as was done above for the mortality rate, that the birth weight of an animal, either normal or alcoholic, varies inversely with the size of the litter in which it is born. The average daily increase in weight during the first month varies in the same way. So that when one month old the weight of a guinea-pig also as a rule varies inversely with the size of the litter in which it was born. ‘This condition holds up to three months, at which time the guinea-pig is mature. But the daily gain in weight during the second and third months after birth ceases to be greatest for the members of small litters. Yet the advantage in growth rate comes to the members of the large litters at so late a time that they are unable to make up their disadvantage sufficiently to equal ‘in size the members of small litters within three months. All of these statements apply equally to both the alcoholic and normal series, and thus the influence of the litter size in general is the same in both cases. The question then arises whether there is an actual difference in birth weights and growth rates between the two series. Table 8 contains the birth weights of 225 normal control and 531 ani- mals of the alcoholic series. This alcoholic group, as the fore- going tables show, not only includes F; animals, or offspring from directly treated parents, but also their descendants for several generations, F:, F;, and Fy. The animals of both series are arranged in table 8 according to the size litters in which they occur. A review of the table shows that the normal series is superior in the average birth weight of the individual and the average birth weight of the entire litter, as well as the average birth weight of the individual born in each of the five different-size litters. 200 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU MABEE Ml Bini’ | Giles zAND RAM ES OfsGROW Wit 70k INORINAL AND AECOHOEIC YOUNG 2 3 A \ 2 3 4 5 4023 1386) 1474! 5200 244 oe ime) 4G oq} 67 15 41 163 225 92 S at birth (108.5) (82-15) (Jo-23) (61.56) (62-13) (943.42) (82.54) (65.5!) (56-52) (49.20) (Average 71- 16) (Average 10-25) - 9.23% (ef the mean) Average productivity 197-12 poe Average productivity 170.0 | 2 Z} 4 S ! 2 3 2228 2902 142838 6750 2302 3046! 26398 Weight at the end of the 7 31 68 31 12 24 13) 133. 40 z first month (319-28) (240.54) (2lo.11) (182-43) (141.83) (247.63) (232-52) (143-48) (172-02) (164-0) (Average 228.64) (Average 213-94) — 6-63% Averagedailyincrease]| 6-44 5.28 464 = 4.02 4.30 inweight during the CAvevanets os) first Month Weight at ihe Anis! ase seen ee S$ end of the third i month 5 14264 44635 4175\ 13 3 33 62 33 12 =z! 105 102 32 2 (501.62) (433-0) (413.54) (330-36) (397.0) (460.12) (425.09) (404-32) (367-15) (354.0) (Average 425-11) (Average 4-04.13) - 5.06% ee as 3.05 3.20 BOG eeep ee-y) 2-J0 B20 3-5) 3-25 3.1L Inwet uyin ro) Z 2°93 and3td months (Average 3. 2b) (Average 3.16) The average birth weight of the individual in the normal series is 77.16 grams against 70.35 grams for the alcoholic, and the average litter weight is 27.12 grams heavier among the nor- mal animals. The average weight of the individual in a given size litter is shown in parentheses below the litter number; this is obtained by dividing the total weight in grams of all such litters by the total number of animals composing them. For example, in the alcoholic series there are 168 animals born in litters of two and their total birth weight was 13,867 grams, which gives an average weight of 82.54 grams per individual. The average weight of the individual is lower in the large litters than in the small ones in both series. The second line of the table shows in a similar way the total weight at the end of the first month of all individuals in the sev- eral-size litters and below this: the number of individuals con- cerned in each case. The quotient obtained by dividing the total weight by the number of animals is given in parentheses as the average weight of the individual in each litter at one month old. At this age the average weight of normal animals MODIFICATION OF THE GERM-CELLS IN MAMMALS 201 in litters of one was 318.28 grams against 297.68 grams for the aleoholic litters of one. The general average weight at one month for the normal series was 228.64 grams against a general average of 213.94 grams for the alcoholics. The average daily increase in weight during the first month is given in the third line of the table. It shows a mean daily increase for normal animals of 5.04 grams and for alcoholic animals only 4.78 grams. Members of small litters in both groups gained more rapidly than members of large litters. The weights at the end of the third month, when the animals are about mature, are given in the fourth line of the table. Nor- mal animals born one in a litter average over 500 grams, while comparable alcoholic animals weigh only 460.12 grams. The average normal animal at three months old weighs 425.11 grams against an average of 404.15 for the alcoholic animal. The last line shows that the average daily gain in weight during the second and third months was about as great for the alcoholic animals as for the normals. A much greater selection or elimination has taken place previous to this time among the alcoholic series than among the normal, as a reference to any of the mortality tables will show. All in all, table 8 would seem to indicate that in every case the normal offspring weigh more and grow more rapidly shortly after birth than do the young alcoholic specimens. The several points considered above and their general meaning may be much more clearly expressed in the diagram, figure 9. On the left side of the diagram are shown the records for the alcoholic series and the normal records are on the right. The shaded right-angle triangles represent the difference in average weight between the individuals in litters of one, two, three, four, and five at birth, at one month old, and at three months old from the two series. The altitudes of the right triangles measure the magnitude of the differences. Animals born one in a litter in the aleoholic and the normal series, as the bottom short triangle indicates, show a greater dif- ference in weight than those from any other size litter except that consisting of five individuals as the low long triangle repre- Normal lines Alcoholic lines AD (| UL S AQ Ul oy sap ULE sas UL Z 5o0o (8)t9 OM UL | Say | UL | ageroge Aayt| Ul 7 Apr) UL S Jot) “Ul +7 Aa | US 500 | | | | | | | { | | | | { ( Aoo . 1 | | | | ! | | CNolbe 1 32 5) 33-0 | ‘ (ce79e\O8S | 1 } (t9) bs ele los vi (sal) bo'st+ (Ze) TE" bO+e \ y (tE)SI LIS ! | «z) O-bSe I | | y | | | | A600 300 (B87 LEC i} 300 2o00 (eee-lbl ! (LE) Ev 200 {oo (siJelrg | ] | ! I ! | | | | | 100 (bs) 9s 19 |. if t (28) 88 ol 1 GEiss9 Cb) 1S°9G (S)O8 8+ ee Fig. 9 Diagram illustrating the differences in weight between normal and The weights are given at birth, at one month, and at three months old. Further explanations are to be alcoholic line animals born in litters of the same size. found in the text. 202 MODIFICATION OF THE GERM-CELLS IN MAMMALS 203 sents. There is little difference between the birth weights of normal and alcoholic animals born in litters of two, three or four. When one month old the middle group of triangles represent- ing by their position the weights in grams again show the largest differences between alcoholic and normal animals in litters of one, the short triangle, and litters of five, the long triangle. The normal animals in litters of one have passed the 300-gram line in weight, while the average alcoholic member of a litter of five weighs only 169 grams. Members of the two series in litters of two, three, or four do not show very great weight differences. The top triangle shows a very large difference in weight at three months between normal and alcoholic animals born one in a litter. The triangles for two and three in a litter animals are almost flat at three months, indicating very little difference be- tween such normal and alcoholic animals. Alcoholic members of litters of four are somewhat smaller in average than normal, while alcoholic from litters of five are far below the normal in weight as the long triangle shows at three months. We have here an example of the influence of the aleohol effect combined with the action of a normal condition, the condition being the size of the litter in which the animal is born. From a consideration of the diagram we may, therefore, conclude, first, that normal-stock animals born one in a litter are so strong as to run far ahead of the one in a litter alcoholic animals, although the latter at birth, at one month, and at three months are much heavier than all normal animals born in larger litters at similar periods. Consequently, the advantage of developing alone in the uterus is sufficient, so far as birth weight and rate of growth are concerned, to overcome the disadvantages resulting from alcoholic ancestry to such a degree that these individuals are better than control animals developing in larger litters. Yet in birth weight and growth rate these singly born alcoholic animals are further behind the singly born control than are the alco- holies from any other size litters behind the control from the same size litters. Thus, although being born alone tends to overshadow the alcohol effect, nevertheless the effect is still shown by comparison with control specimens born alone. 904 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU If we now recall the fact that aleoholic animals produce more small-size litters than do the control, and recognize that mem- bers of small litters in all cases weigh more, grow faster, and are more apt to survive than members of larger litters, it becomes evident that the production of a high percentage of small litters is a fortunate provision tending to preserve the alcoholic stock by counterbalancing to some degree the magnitude of the effects induced by the alcoholism. Second, animals born in litters of two or three have a tendency to weigh the same at birth and to grow at a similar rate dur- ing the first three months, whether they are from the normal or alecholic stock. In other words, being born in litters of this size gives no great advantage to the normal animals over the alco- holies, as does being born in litters of only one. Or stated re- versely, members of litters of two or three are not placed at a great disadvantage so far as birth weight and growth rate are concerned on account of their alcoholic ancestry, as is found below to be the case for the members of larger litters. In the third place, when animals are born in litters of four the alcoholic stock are at a disadvantage in birth weight when compared with the normal. The rate of growth of the alcoholic animals from litters of four is also slower than that of the com- parable control animals. Lastly, in the fourth place, alcoholic animals born five in a litter are very small and weak and only a few survive, yet these selected few fall far behind the normal animals from litters of five in their rate of growth. Thus at three months there is a greater difference in average weight between the alcoholic and control members of litters of five than between the members of any other size litters in the two series, except the animals born singly. The alcoholic animals as a group are at a disadvantage in birth weight and rate of growth, but when born in large litters of four or particularly five, this disadvantage is greatly exaggerated by the handicap which befalls the members of all large litters, the control as well as the alcoholic. MODIFICATION OF THE GERM-CELLS IN MAMMALS 205 13. THE RECORDS OF NORMAL MALES AND FEMALES PAIRED SUC- GESSIVELY WITH NORMAL AND ALCOHOLIC MATES: THE CRUCIAL DEMONSTRATION OF THE EFFECTS OF ALCOHOLISM ON THE OFFSPRING When the records of any group of experimental animals are compared with the records of a normal group, the possibility presents itself that some selection either conscious or uncon- scious may have played a part in forming the groups. Such a source of error is no doubt practically eliminated by many well- known methods of choosing control and experimental animals from a given population. We believe such a defect is entirely insignificant in the foregoing records which have involved many animals through several generations from the same stocks in the case of both the experimented and the control. It is, never- theless, satisfactory to consider the records of the same nor- mal animals paired successively with control animals and with animals of the alcoholic lines. ‘Table 9 presents all of the mating records of fourteen normal males and fifteen normal females that have been paired in this way. This table gives a most perfect control and shows most clearly the alcohol effects. BPA BEE TWX ANORMAL MALES AND FEMALES PAIRED SUCCESSIVELY Wa NOREAAIE AND ALCOHOLIC MATES Individual matings of 14 normal) Individual matings of (5 normal males, cach One mated succes-| females, cach One mated succe s- sively with sively with ie ere /Number of matings Total number of young /Negoative result Alcoholic females| Normal males |AlcOholic males A- ! (9.09% (5.84%) a (5.55%) Lived over ele 58 a BS months TOtal dead or 5 5 died within ee ee Ce SS) MOnsFhs (24.41%) (42.0%) (A0.0 % Defective L 206 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU The fourteen male animals are in no sense selected; they are all of the normal males in our series of animals between the num- bers 613 and 1909 which have been mated with both normal and alcoholic females. The record numbers of these males are 665, 666, 667, 669, 670, 676, 677, 679, 681, 682, 683, 854, 914, and 1052. The fourteen males, as the table shows, have been mated in all eighty times. The fifteen females recorded include also every normal female among the animals considered in this paper that has been paired with both normal and alcoholic males. The record numbers of the females are 645, 646, 650, 652, 657, 661, 662, 671, 674, 675, 703, 722, 760, 890, and 1043. These have been mated in all forty-nine times. There has been no selection or choice in mating these animals or in estimating the results, since it was only decided to arrange such a table after beginning the present study of the data. The first column of table 9 shows the results of thirty-six matings of the normal males with normal females. Two of the thirty-six matings failed to produce results, or 5.55 per cent, and the remaining thirty-four matings gave rise to eighty-six young. Sixty-five, or 75.59 per cent, of these lived to reach maturity, while 24.41 per cent died within three months. None of the eighty-six offspring showed any gross structural defects. When these same normal males were mated forty-four times with alcoholic females, the second column shows that four mat- ings failed, or 9.09 per cent, almost twice as many as the failures - with normal females. The forty successful matings produced one hundred offspring, only fifty-eight of which were capable of survival to maturity. Thus 42 per cent of the young animals died within three months against only 24.41 per cent of those from the normal mothers and same fathers. Six per cent of the young from the aleoholic mothers possessed noticeable structural defects. In every respect the matings of the fourteen normal males produced greatly superior results when paired with normal fe- males, as compared with their records by alcoholic females. The numbers are comparatively small, but the differences are large and the inferior records are consistently in the same column. MODIFICATION OF THE GERM-CELLS IN MAMMALS 207 The third and fourth columns contain similar records from the matings of the fifteen normal females with normal males and with alcoholic males. The twenty-six normal matings gave only one failure, while the twenty-three matings with alcoholic males failed to give results in five cases, or in 21.73 per cent of the trials. The alcoholic males always give a high percentage of mating failures even with normal females and, as this case shows, with females giving only a low per cent of failure by normal males. The normal matings produced fifty-nine young, fifty-one of which survived while only eight, or 13.55 per cent, died within three months. This is an unusually low mortality record and proves the ability of these females to produce strong viable young. None of the offspring from the normal matings were defective. The same females produced by alcoholic males fifty young, only thirty of which lived to maturity. Therefore, 40 per cent of them were non-viable, which is three times more than was the case with offspring from these females by normal fathers. Ten per cent of the fifty offspring were defective. ‘The contrast between the two groups of results from the same females is so great that the possibility of the difference being due to the smallness of the numbers involved would seem to be completely eliminated. The records in the entire table are perfectly con- sistent and very clear cut. It would seem only proper to interpret such results, along with the mass of evidence in the foregoing pages, as showing that alcoholic guinea-pigs, whether directly treated or descended from treated individuals, have had their ability to produce strong, viable offspring definitely and decidedly lowered. And it may be added in this connection that evidence from purely male treated lines as well as that given by later generations from the female treated and mixed lines, points directly to the fact that the germ cells have been affected. The effects of this modification are transmitted through several generations, only to be lessened by the elimination through death and sterility of the weakest individuals from the mating records and the con- stant introduction of more and more normal germ plasm into the line by matings with the normal stock. 208 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU 14. THE CONTRASTED QUALITIES IN THE CONTROL AND THE ALCOHOLIC SERIES The earlier reports on these experiments have given in the general text the various differences between the alcoholic and control lines; the case is made much clearer, however; if all the contrasted qualities be arranged together in summary fashion. In Pearl’s recent report on the influence of alcohol inhalation on the progeny of the domestic fowl, he has given a concise arrange- ment of the differences between the records of the experimented and control lines. The several qualities he has compared such as mortality records, fertility, abnormalities, etc., are the same as those considered in our previous papers. We have here con- structed a similar table to the one used by Pearl to show the qualities contrasted in the former sections of this paper. Definite numerical values have been presented for fourteen dif- ferent qualities studied in the two groups of animals. Several of these qualities are closely related, such as weights after dif- ferent periods of growth and the mortalities calculated at dif- ferent periods, yet these are stated separately since they were measured in this manner and help somewhat to give a clearer analysis of the entire problem. Table 10 shows the qualities measured. The first column of figures are the records from the control, the second column are the alcoholic records. In the last column a — sign indicates that the alcoholics are inferior to the control for the given quality; a zero, that the two groups are similar in the given respect, and a + sign would show that the alcoholics are superior to the control. It is seen at once that the alcoholic series suffers by comparison in every case except one, and in this case the two series are equal on account of an earlier unusually large difference. The alcoholic guinea-pigs are less productive, giving litters of smaller size than the normal, their matings more often result in failure to conceive; associated with these two facts there is a higher early prenatal mortality which is the only quality in- cluded in the table that cannot be numerically expressed for reasons brought out in previous pages. MODIFICATION OF THE GERM-CELLS IN MAMMALS 209 | TAD pie. QUALITIES CONTRASTED BETWEEN THE NORMAL AND ALCOHOLIC PROGENIES Qualities measured Alcoholic | 4}: que [IP Size oF initer; apie ZAT 2. Failure to conceive | 4.452%] 13.04% |S. Early prenatal death (sizeof litter, failuve,ete)} low high 14. Proportion late prenatal death | 51.92% 70.14% |5 Post-natal mortality 10.70%| 10.60% 6 Total mortality | 22.31% (100) 7. Abnormalities | O &. Oversize (+ 5008rs. at 3 mos.) | 5.57% 9 Undersize(— 500Srs. at 5S mas.) | 0.42% 10. Late generations alcoholic impreved, | 95 <10,-|Fi 42-40% mortality index | 22.5 Vole 2 I} f, Altered sex-ratios 1039.G0 12. Av. birth wt of fitter 13. Av. individual birth wf. l4 Av. wt. | month old 15 Av. wt. 3 months old PSGIEZ, HI TLO;O© UNG) Gees 223.64 1213.94 The alcoholics have a higher proportion of their total mortality occurring very early, so that there is a great elimination of weak embryos and fetuses; this lowers their later or postnatal mortality to about the normal record. In this case we have an elimination or selection of individuals or zygotes rather than a germinal selection. The total mortality record for the experimented group is far higher than for the control and a greater percentage of abnormal young are produced. The percentage of abnor- malities is lower than in our former records, as is also the total mortality rate. The improved mortality rate is partly due to better methods of breeding and caring for the animals. Yet the mortality record of the alcoholic group is very high, and when corrected for the normal rate on the basis of the size litters con- cerned it becomes 189 against the control as 100. Among the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 1 210 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU normal animals of the same general stock as the alcoholics, not one grossly deformed individual has been born in over 400 eases, and, as stated above, this is a remarkable record which argues strongly for the perfection of the stock. In considering the de- fective young, one must also keep in mind the fact that these are not worse, but, on the contrary, are better organized than individuals which die during early stages of development. At three months old, as No. 8 in the table indicates, fewer alcoholic than control animals were larger than usual or over size, though some were, while the next line shows that more alcoholic animals were small or under size, weighing less than 300 grams. The later generations of the alcoholic stock are improved by the continued elimination’ of weak and defective individuals which die or are unable to breed, and also by the introduction of more and more normal germ plasm from generation to genera- tion until’a mortality rate of 42.4 per cent for the F, generation becomes only 17.14 per cent for the F, generation. This is a clear demonstration of the aleohol effect and may also serve to show the action of increased germ dosage. ‘The earlier genera- tions being nearer the directly treated animals receive higher doses than do the later generations where in most cases the dose has been considerably diluted by a mixture of normal germ plasm. | The sex-ratio in the alcoholic group seem to have been modi- fied in ways which we have attempted to explain. The average weight of the alcoholic litter is less than the normal and the average individual birth weight of an alcoholic specimen is also less than for the normal. The average weight of the alcoholic individuals at one month old is below the nor- mal and the average weight at the age of three months, when guinea-pigs are about mature, is still below the weight of the ‘control animals. Therefore, in the fourteen measured points considered, the offspring of the alcoholic series are below the normal control in thirteen cases and apparently equal to the control in only one. The qualities are largely the same as those we have considered MODIFICATION OF THE GERM-CELLS IN MAMMALS 211 in former papers though analyzed in further detail. They are also very similar to those recorded by Pearl (’17) in his table 14. From a physiological standpoint it seems to us that these quali- ties are all closely associated and finally come down to the three related qualities: ability to develop normally, grow rapidly, and live to maturity. An animal possessing such qualities is usually termed a vigorous individual. At present it can only be stated that these properties are due to the vigor of the germ cells from which the individual arose. The qualities discussed might all involve a limited range of physiological factors so far as present knowledge permitsa separation of such factors and they only show on the part of the alcoholics a reduced capacity of development and growth. The same underlying cause may actually account for the abnormal sex-ratios, as has been pointed out in an earlier section. Leaving the environment out of account, the normal develop- ment, growth and length of life of a zygote varies with the perfection or vigor of the germ cells from which it originated. An experimental treatment may act upon the germ cells of an animal so as to modify them in some general way which lowers their ability to react normally in combination with germ cells from another individual. Thus zygotes are produced which tend to develop abnormally, grow slowly, or die during early stages of their existence, depending upon the degree of modifi- cation the treated germ cells have suffered. We are fully em- barrassed by the unsatisfactory nature of such statements, but have been unable to gather scientific facts that would permit any more definite estimate of the situation. All of our experiments on the modification of the germ cells have given results which express themselves in some such general fashion. Yet the germ plasm has been definitely modified and the subnormal condition is transmitted through a number of generations beyond the animals directly treated. This result is original on the complex materiai used, and is of primary impor- tance, although it may be disappointing in that it has not shown a modification in the mode of behavior of some particular char- acter known for its Mendelian inheritance. 212 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU The experimental modification of the inheritance of definite characters by a treatment of the germ cells is a future possibility. It must be recognized, however, that one is able to produce grotesque monsters by a treatment of eggs or spermatozoa, and yet all of the known characters which Mendelize in such an in- dividual may be expressed in a perfectly normal fashion. This may be due to the fact that comparatively few such characters are known. Aside from the future definite modifications of inheritance, it would seem from the present study that the ‘general qualities,’ for lack of a more suitable term, of an or- ganism may be affected, on account of an experimental modi- fication of the germ plasm from which it arose. The modifi- cation may have taken place several ancestral generations ago. This is really the inheritance of pathological conditions which were induced upon and transmitted by the ancestral germ plasm. Such a type of inheritance is no doubt important in its relation to the normal processes of development and inheritance. 15. GENERAL CONSIDERATIONS A discussion of the literature bearing on the influence of various chemical substances on the egg and spermatozoon has been given in former papers of this series, particularly Stockard (12 and 13). In all cases only the effects of the treatments on the zygotes immediately resulting from the modified spermatozoa or eggs have been studied. There has been no experimental investiga- tion of later generations arising from the affected specimens. And indeed, in almost all cases the developing individuals were lost during early embryonic stages as in the X-ray experiments of Bardeen and the radium studied of Oskar Hertwig which are the most satisfactory investigations on the direct injury of the sperm. These experiments really supplied no available material for an investigation of the inheritance or transmission of the induced defective conditions. Since the beginning of the present experiments other studies have been recorded which bear more directly on the results con- sidered in the foregoing pages. Of particular interest in connection with our supposed differential effects of the alcohol treatment on MODIFICATION OF THE GERM-CELLS IN MAMMALS ie thebehavior of the X and Y groups of spermatozoa is the ingenious double-mating experiment of Cole and Davis (14) with rab- bits. They found that when two male rabbits were mated with a single female, superfetation occurred in most cases, so that part of the resulting litter of young were sired by one male and part by the other. The males differed in their fertilizing abili- ties, so that one more often sired the majority of young of a given litter, and in the total number of competition matings he sired the greater number of young. This male with the fertiliz- ing advantage was then treated for a month or more with the fumes of alcohol by the inhalation method. As a result of this treatment his spermatozoa became affected in such a way that mated in competition with the same male he normally had beaten he now failed to sire any young. Yet when mated singly or alone with a female he still possessed the power to beget off- spring. This is a striking illustration of the debilitating effect ~ of a short alcohol treatment on the physiological behavior of these spermatozoa, thus lowering their fertilizing ability below that of other spermatozoa which were formerly less potent than they. When it is seen how definitely and readily alcohol treatments affect the behavior of thé spermatozoa, we are led to speculate as to whether the treatment might not affect the X and Y groups of sperm differently, and thus be partially responsible for a dis- tortion of the sex-ratios, should such occur. This responsibility may be due in the first place to a lowered fertilizing power on the part of one group of spermatozoa, thus giving rise to fewer individuals of one sex than of the other. Or, in the second place, even though both groups of spermatozoa should be equally capable of fertilizing the eggs, one group might be more affected as to its ability to produce viable zygotes in combination with normal ova, and thus an early differential sex mortality would occur causing a modification of the proportion of one sex to the other among the young born. We have elaborated somewhat on these possibilities in the section devoted to the sex-ratios of the alcoholic guinea-pigs. Cole and Davis originally devised their experiment as a cru- 214 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU cial control for the influence of alcohol treatment on the male germ cells. In mating two males to a single female any defective condition that might arise among the offspring from one of the males, as compared with those from the other, could not be attributed to differences in developmental environment or in the qualities of the ova, as might possibly be the case where different females are used. Cole and Bachhuber (’14) have employed the same method in a study of the effects of lead on the germ cells of the male rabbit and fowl. Their conclusion in regard to the rabbit is ‘‘that the offspring produced by male rabbits which have been poisoned by the ingestion of lead acetate into the alimentary tract have a lower vitality and are distinctly smaller in average size than normal offspring of unpoisoned males.’’ This conclusion is in exact accord with the conditions shown by our F, generation of guinea-pigs sired by alcoholized fathers. Cole and Bachhuber have not reported on the transmission of the effects to later generations. Their results with fowls ‘‘are interpreted as indicating that in fowls also poisoning of the male parent with lead results in offspring of a distinctly lower average vitality.” This again accords with the results on the offspring when male guinea-pigs are treated with alcohol. A later more extensive report concerning the influence of lead as a substance producing blastophthoric effects is given by Weller (15). This investigator has treated both male and female guinea-pigs with commercial white lead. The lead is adminis- tered by mouth in gelatin capsules, the same method as was em- ployed by Cole and Bachhuber (14). The effects from the lead poisoning on the guinea-pigs are very similar to those ob- tained by treating the rabbits and fowls. Weller has been careful not to overdose the animals and his precautions would make it seem probable that any effect from the treatment which might be shown by the offspring was actually due to the lead poisoning and not to impaired nutrition or other indirect causes. His conclusions are based on a total of ninety-three matings yielding 170 offspring. There were thirty-two control matings MODIFICATION OF THE GERM-CELLS IN MAMMALS 215 which produced only fifty-eight offspring. Whether or not every mating gave offspring is not definitely stated, but if so the aver- age-size litter was unusually small, being only 1.81. This would indicate either a stock of very low productivity or a high pro- portion of absorbed embryos and partial abortions, as a final result of which the litters would be small. In the foregoing tables where the numbers of matings and young are very much greater, not one group shows so small an average litter. From the thirty-four matings of lead-poisoned males with normal females, sixty-five offspring resulted, an average litter of 1.91, and from twenty-seven matings of normal males with lead females forty-seven young were born, an average litter of only 1.74. The fact that among the few individual litters recorded there were three cases of litters of four, and five cases of litters of three, makes it seem as though there may have been a high proportion of mating failures, giving rise to the small average litters ob- tained when the total number of young is divided by the total number of matings. The distribution and cause of these mat- ing failures, as is pointed out in the text above, may be of con- siderable importance. Weller has analyzed his results in some detail. He takes into account the influence of litter size on the birth weight and gives several individual mating records which illustrate the effects of a treated sire on the birth weight of the young from a normal dam. Weller has also taken into account the relationship between lead dosage and birth weight of the offspring without finding very consistent correlations. The relationship between germ dosage and the condition of the offspring in our records may be calculated for every individual born in the alcohol experiments, yet the result is uninstructive so far as at present studied. There are a great number of confusing factors involved in this seemingly simple proposition. Weller’s final conclusions from the study of lead poisoning closely accord with our previous statements regarding the influ- ence of alcohol on the same animals. He finds that chronic lead poisoning in guinea-pigs produces a definite blastophthoric effect. 216 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU This can best be demonstrated upon the male germ plasm, in which ease the blastophthoria manifests itself in some instances by sterility without loss of sexual activity, by a reduction of approximately 20 per cent in the average birth weight, by an increased number of deaths in the first week of life, and by a general retardation in development such that the offspring of a lead-poisoned male remains permanently under weight. These experiments with alcohol and lead on rabbits, fowls and guinea-pigs seem to their authors to modify the male germ cells in a definite manner. The offspring sired by treated fathers are inferior to those from control males. The transmission of the defects to subsequent generations has not been reported. In addition to the experiments on the direct treatment of the spermatozoa of lower forms, a few attempts have been made to treat the spermatozoa of higher animals directly with certain chemicals. Ivanov (13) has given a short note on the effects of immersing the spermatozoa of several mammals in solutions of aleohol. He finds that when fertilization is obtained after such treatments a normal development follows and normal offspring are produced. ‘To anyone who has studied the action of alcohol on the free swimming spermatozoa of lower vertebrates such re- sults are not surprising. The most probable explanation is that the spermatozo6n has been entirely protected from the action of the alcohol of the strengths used. When any action is ob- tained the usual effect on the spermatozo6n is to render it im- mobile. To obtain a fertilization the motionless sperm must be activated by the use of some alkaline substance, such as NaOH. Following this activation the spermatozoa may often give normal offspring after union with normal ova, thus indi- cating that their chemical nature has not been disturbed. It is most difficult to treat the spermatozo6n even of the very hardy fish, Fundulus heteroclitus, in such a manner as to injure it and afterwards obtain a fertilization. Dr. Wilson Gee (’16) experi- mented on the spermatozoa of fishes at Woods Hole for two seasons and found that the difference between an effective alco- hol dose and a fatal dose was so slight that it required the most delicate adjustment of solutions in order to injure the sperma- tozoa to such a degree that the development of eggs subsequently MODIFICATION OF THE GERM-CELLS IN MAMMALS Awe fertilized was rendered abnormal. Ivanov’s report is cer- tainly not sufficiently detailed to satisfy one that his results have any bearing on the problem of the modification of the germ cells by chemical treatment. There can be no doubt that if a spermatozodn is actually affected by a direct chemical treatment, the egg which it fer- tilizes will develop more or less abnormally. The radium and X-ray experiments of Bardeen and Hertwig, as well as fertiliza- tion by foreign spermatozoa give conclusive evidence on this point. The statistical research by Elderton and Pearson (’10) has frequently been quoted as if it shows that parental alcoholism was really to some degree beneficial to the human offspring. Their mathematical calculations were based on two series of statistics, the ‘‘Edinburgh Charity Organization Society Report and a manuscript account of the children in the special schools of Manchester provided us by Miss Mary Dendy.” ‘‘Sus- pected drinkers were included with drinkers,’ ‘‘the parents could be divided into two classes only, those who are temperate and those who are intemperate,’ and many other such state- ments make this biological data somewhat unsatisfactory to those interested in an experimental modification of the germ plasm. These authors, however, do not claim to find any effect, either good or bad, of alcoholism on the offspring, and finally state that On the whole the balance turns as often in favor of the alcoholic as of the non-alcoholic parentage. It is needless to say that we do not attribute this to the alcohol, but to certain physical and possibly mental characters which appear to be associated with the tendency to alcohol Such a conclusion on the part of the authors themselves would scarcely warrant anyone else in claiming that an effect of alco- holism on the parent,had given evidence of its existence in the quality of the children produced. A number of English physi- cians interested in alcoholism largely from a social and senti- mental standpoint opened a bitter attack on the memoir by Elderton and Pearson, not because it claimed a beneficial effect, \Ttalics are ours. 218 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU but merely because no harmful effect was shown. Such criti- cism is of little interest, yet one very serious point was cited against the data on which this study was based, and Pearson and Elderton (’10) in their reply failed to satisfy the objection. The children considered were in the neighborhood of nine years old at the time the statistics were collected and the fact that some parents were drinking at this time might not necessarily prove that they were drinking nine or ten years ago when the children were conceived. It is very evident that from our standpoint accurate data relating to this particular fact is most essential. This study really has no bearing in the literature on the chemi- eal modification of germ cells or the developing embryo, as Elderton and Pearson themselves state in the italicized portion of the quotation cited above. No one can confidently affirm that in their data alcoholics are being compared with normals or really whether any alcoholics or normals as such are actually being considered beyond the chance probability that some in- dividuals of both classes creep into the statistics to be included in the two groups arranged. Very recently Pearl (17) has published a most thorough analysis of the influences of parental alcoholism on the progeny of the domestic fowl. He states (p. 285): that a careful study of the present results makes it impossible to assert that the treatment of the parents has had no effect upon the progeny. The offspring of the alecoholists, as a class, are indubitably differentiated from the offspring of the non- -alcoholists. Such a statement agrees entirely with our results from the aleoholic guinea-pigs. In detail, however, Pearl finds that after treating fowls with aleohol the progeny produced are in some respects superior to'the control. This, he believes, is brought about by an elimination of all weaker germ cells through the action of aleohol which thus serves as a selective agent to 1m- prove the race. At first sight this would seem to be entirely contradictory to our results, since the guinea-pig progeny is decidedly the worse for the experimental treatment. Yet the treatment in both cases has affected the progeny through its MODIFICATION OF THE GERM-CELLS IN MAMMALS 219 action on the germ cells. This is the point of actual importance and the one of chief interest from the standpoint of these ex- periments. We are not here studying the alcohol problem from a social standpoint and it is immaterial whether the progeny be benefited or injured by the treatment of parental generations. Our interest lies in whether or not the germ cells are modified by the chemical treatment and whether the modification is of such a nature as to alter the qualities of the individuals which may compose the subsequent generations. Pearl, of course, fully agrees with such a position, and states (@ilGra, 2 258) : Our results seem to me to be supplementary to those of Stockard, and to throw an interesting light on the need for caution in reac: ing a correct interpretation of all experiments in which a mildly deleterious agent acts upon the organism. He also believes that his results are in no way contradictory to ours, but recognizes the fact that, although the same chemical substance may act upon the germ plasm of two different classes of animals, the visible response on the part of the animals need not necessarily be the same. In other words, one is not always within the realm of legitimate scientific speculation who assumes that since a given substance acts to induce a certain response on the part of one animal species that the same substance will call forth a like response on every other species. ‘‘ What is one man’s food is another man’s poison.”’ With this we fully agree; it is dangerous to draw universal deductions from experiments on any one or two classes of animals. Another possibility also recognized by Pearl presents itself in considering the opposite effects of the alcohol treatment on the progeny of guinea-pigs and fowls. Small doses of many sub- stances, one of which is alcohol, may form a physiological stand- point produce a stimulating effect, while larger doses produce decided depression. ‘There is a. possibility that the same may be true of the action of such substances on the germ cells. Pearl has discarded such an explanation after very fair consideration, and is possibly right in so doing. The experiences, however, with the guinea-pigs makes our opinion decidedly prejudiced in 220 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU favor of the possibility, that although a sufficiently large dose may have been used, yet it did not act solely to eliminate germ cells as such, but also caused the production of many zygotes which died during early developmental stages. The amount of dosage is very important. Treating female guinea-pigs with considerable doses of alcohol fumes only shortly before and during their pregnancies certainly does not injure the offspring to any noticeable degree. While the ‘same dose of treatment, if administered for a number of months or years, will render these mothers almost incapable of producing vigorous young, even when mated with normal males. Pearl (17, p. 281) finds regarding his 1915 results which were obtained after the treatments had been running for only a few months that considering the number of animals in the experi- mental series the individual differences are not in every case sufficiently large to be significant in comparison with their prob- able errors. The control in this case was also not what Pearl had, wished. He had originally chosen a carefully pedigreed control, taking as the one control male a half-brother of the three experimental males and using control females that were sisters of the treated hens as recorded in table 5, p. 158 (17). The only control male, No. 666, proved to be practically sterile and useless. This necessitated the use in paper No. III of an ordi- nary random sample control instead of the refined control originally planned in Part I of the series of papers, and nulli- fied the statement in the summary of Part I, p. 162, that ‘‘ Full brothers and sisters of treated are used as control.” For certain qualities, such as the fertility and hatching records of the eggs, the control was not in all cases the same cross as the experiment, which was invariably between Barred Plymouth Rock hens and Black Hamburg cocks. The hatching weight and rate of growth of the experimental chicks on account of want of con- trol data from the 1915 season were compared with chicks from a similar cross hatched and reared in 1913. Different keepers were in charge of rearing the chicks during the two different seasons. These unfortunate conditions, all of which are pointed out with conscientious fullness by Pearl, make it rather difficult MODIFICATION OF THE GERM-CELLS IN MAMMALS Pov \\ to fully estimate the actual significance of the differences between the experimental offspring and the control groups used. Fortunately, however, the data from the 1916 season is avail- able (Pearl, ’16b) for comparison with the 1915 results. The alcohol treatments were continued throughout the time so that the 1916 chicks are derived from more highly aleoholized parents. Should the alcohol continue to improve the race by ‘‘ completely putting out of commission all of the weaker germ cells,” the 1916 results should in all respects show a further improvement in the qualities that had been previously benefited. The percentage of infertile eggs given in the 1915 table may be reversed to per cent of zygotes formed and compared with this column in the 1916 table. The percentage of zygotes formed in the several combinations of alcoholic mating should be less than in 1915, and they are. When both parents were alcoholic in 1915, 40.8 per cent of the eggs formed zygotes, while in 1916 only 21.95 per cent produced zygotes; sire only alcoholic, 74.8 per cent zygotes in 1915 and only 53.52 per cent in 1916. This “ais in line with the lowered fertility and increased number of mating failures from the alcoholic guinea-pig records. The more decidedly alcoholic the guinea-pigs become, the smaller the litter size from double alcoholic and sire only alcoholic matings, and the greater the number of failures to conceive. With the guinea-pigs, however, this is not alone due to a destruction of weak germ cells by the treatment, but is cer- tainly in part due to an increased very early prenatal mortality for which much evidence is given in the body of the present paper. The smaller number of zygotes formed by the treated fowls is probably also due in some cases to death in very early stages, as blastulae or gastrulae, before the egg is laid; or in the. hen’s eggs these weakened zygotes may not be able to with- stand the developmental interruption following the laying of the egg. Embryos dying during such stages could not be iden- tified except by a most minute study. It seems to us in keeping with what is known of biological reactions in general and the guinea-pig histories in particular to take the following position. The aleohol treatment acts on the Se 222 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU germ cell populations of both fowls and guinea-pigs in such a manner that the weakest or least resistant ova and spermatozoa die from the effects of the treatment as germ cells without taking part in zygote formation. The somewhat more resistant ova and spermatozoa are greatly injured though still capable of forming zygotes. The zygotes, however, are so defective:as to be capable of only a short period of development and die during stages too early to be definitely detected by gross examinations of either the fowl’s egg or the mammalian mother. Still other em- bryos are capable of development to later stages and are actu- ally found dead, not only as the youngest embryos to be identi- fied, but from these early stages there occurs a continuous series of prenatal deaths up to the full-term still-births. Immedi- ately after birth the postnatal mortality is greatest and gradu- ally decreases until these specimens capable of reaching maturity may often enjoy a comparatively long life. At the present stage of the two experiments it would seem as though this elimination of defective germ cells and very early embryos was much more intense in the fowls than in the guinea- pigs as a group; so that the late prenatal and postnatal mor- tality among the fowl progeny was low and those specimens that hatched were the hardy survivors from this early vigorous process of germ cell and individual selection. The records from the double alcoholic and male treated lines among the guinea- pigs forms a second step. ‘The size of the litters and failures to conceive in these lines indicates a rather high degree of in- fertility or germ cell debility as well as early prenatal deaths, though this is not so extreme as among the fowls, and the late prenatal and postnatal mortality is higher. Finally the female treated guinea-pig lines produce large lit- ters and have few infertile matings, indicating a low germ cell and early prenatal mortality, and here the late prenatal and postnatal mortality is highest, not entirely on account of the action of the treatment on the developing individual in utero, since the same condition is found among other female gene- rations than the one directly treated. This presentation of the situation is somewhat similar to that é MODIFICATION OF THE GERM-CELLS IN MAMMALS 223 which Pearl (’17) has illustrated in his diagrams, figures 5 to 7, pages 290 and 291. The chief difference being that we would decrease the proportion of eliminated germ cells and increase the proportion of defective and non-viable zygotes, and thus emphasize the selection of individuals rather than of germ cells. A further consideration of Pearl’s 1916 résults as shown in table 1, p. 676 (16 b), may be used to argue in favor of our po- sition. The ‘prenatal mortality’ column of this table when compared with the ‘dried in shell’ column from 1915 records (table 1, p. 244, 717) should show lower percentages according to our interpretation of Pearl’s expectation for an improved stock from the alcoholic lines. Instead of this, in only one combina- tion is the prenatal mortality lower. In both parents alcoholic it has been lowered from 26.9 per cent to 11.11 per cent, and here the postnatal mortality as we would expect is increased. In the other cases dam only alcoholic, none of which were re- ported for 1915 on account of the useless control male, gives 80 per cent prenatal mortality sire only alcoholic increased to 47.08 per cent from 36.6 per cent; sire and one grandparent, 46.84 per cent; one or more grandparents, 46.02 per cent; all alcoholic ancestry, 45.95 per cent, which is a considerable increase over the 1915 records. The control of 1916 also shows a higher prenatal mortality than that of 1915, though it is not stated whether the same breed crosses are used in the two controls. The postnatal mortality of the 1916 control is, on the con- trary, lower than the postnatal mortality of the twenty-two ‘random sample matings’ of 1915. While the total mortality for all the alcoholic groups is about ‘the same, 17.6 and 16.5 per cent, for the two seasons, the indi- vidual combinations show wide variations. From both parents alcoholic the 1915 postnatal mortality was 10.6 per cent, while for 1916 it rose to 25 per cent, sire only alcoholic fell from 21.1 per cent, 1915 record, to 13.79 per cent, 1916 record. Sire and one grandparent alcoholic gave a postnatal mortality of 28.38 per cent, while the non-alcoholic postnatal mortality was 21.2 per cent. Considering the numbers involved, the records from the prog- 294 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU eny of the 1916 matings after longer alcohol treatment do not seem altogether improved as compared with the 1915 records. A comparison of individual lines in the tables frequently show disadvantages for the 1916 matings. This would seem as though some injured zygotes were present and all of the affected germ cells had not been’ completely eliminated by the treatment. The percentage of abnormal specimens among the 1916 alcoholics is about the same or slightly more than among the control, while Pearl had counted this point in favor of the alcoholics from his 1915 records. It would thus seem, as Pearl (17, 292) himself suggests, that “it might be supposed that with larger administration to the fowls (higher germ dosage) or more years of drinking behind them in the case of Elderton and Pearson’s workingmen, the conditions shown in figure 7 would gradually pass over into those shown in figure 5.” That is, that not only weak germ cells would be eliminated by the treatment,. but that also a considerable proportion of defective individuals would arise to be eliminated during various developmental stages or persist as degenerate specimens. From these conditions we believe that there is a really close agreement between the results on the fowls and the eulnea-pilgs. These suggestions are advanced only in a spirit of the most friendly criticism. We have worked long enough in accumulating and considering evidence bearing on. the various phases in- volved in this problem to highly appreciate the masterly manner in which Pearl has considered and analyzed his data; and we are thankful for many suggestions that have come to us through the contribution on parental alcoholism in the fowls. In the end our aims and objects are the same, to affect the germ plasm in so definite a manner as to be able to predict the quality and degree of the modifications subsequently expressed in the gen- erations to follow. MODIFICATION OF THE GERM-CELLS IN MAMMALS 225 LITERATURE CITED Aruirt, A. H., anp Weiis, H. G. 1917 The effect of alcohol on the repro- ductive tissues. Jour. Exp. Med. Vol. 26, p. 769. BarDEEN, C. R. 1907 Abnormal development of -toad ova fertilized by sper- matozoa exposed to the Roentgen Rays. Journ. Exp. Zodl., vol. 4, je dle Cots, L.J., anp Davis, C.L. 1914 The effect of alcohol on the male germ cells, studied by means of double matings. Science, N. 8., 39, p. 476. Cotz, L. J., anp Bacunuser, L. J. 1914 The effect of lead on the germ cells of the male rabbit and fowl as indicated by their progeny. Proc. Soc. Exp. Biol. and Med., 12, p. 24. Experton, E. M., anp Pearson, K. 1910 A first study of the influence of pa- rental alcoholism on the physique and ability of the offspring. Eu- genics Lab. Mem. X, Univ. of London, pp. 1-46. Ger, Witson 1916 Effects of acute alcoholism on the germ cells of Fundulus heteroclitus. Biol. Bull. 31, p. 379. Herrwic, O. 1913 Versuche an Tritoneiern tiber die Einwirkung bestrahlter Samenfiden auf die tierische Entwicklung. Arch. f. Mikr. Anat., 82, Abtslk Ivanoy, J. 1913 Expériences sur la fécondation des mammiféres avec le sperme mélangé d’alecool. Compt. rend. Soc. Biol. Paris, 74, p. 482. Meyer, A. W. 1917 Intra-uterine absorption of ova. Anat. Rec., vol. 12, pa. 293: Papanicotaou, G.N. 1915 Sex determination and sex control in guinea-pigs. Science, N S. 41, p. 401. Peart, R. 1916a On the effect of continued administration of certain poisons to the domestic fowl, with special reference to the progeny. Proce. Am. Phil. Soe., 55, p. 243. 1916b Some effects of the continued administration of alcohol to the domestic fowl, with special reference to the progeny. Proc. Nat. Acad. Sci. 2, p. 675. 1917 The Experimental modification of germ cells, Parts I, I, and III. Journ. Exp. Zoél., vol. 22, pp. 125-186 and pp. 241-310. Pearson, K., anp Experton, E. M. 1910 A second study of the influence of parental alcoholism on the physique and ability of the offspring. Being a reply to certain medical critics of the first memoir and an examination of the rebutting evidence cited by them. Eugenics Lab. Mem. XIII, Univ. of London. Srevens, N. M. 1911 Heterochromosomes in the guinea-pig. Biol. Bull. 21, Delo: Srockarp, C. R. 1910 The influence of alcohol and other anaesthetics on em- bryonic development. Am. Jour. Anat., vol. 10, p. 369. 1912 An experimental study of racial degeneration in mammals treated with alcohol. Arch. Internal Med., X. 1913 The effect on the offspring of intoxicating the male parent and the transmission of the defects to subsequent generations. Amer. Nat.. 47, p. 641. 226 CHARLES R. STOCKARD AND GEORGE N. PAPANICOLAOU Srocxarp, C. R. 1914 A study of further generations of mammals from an- cestors treated with alcohol. Proc. Soc. Exp. Biol. and Med. XI, p. 186. SrockarD, C. R., Aanp Craig, D. M. 1912 An experimental study of the influ- ence of alcoho] on the germ cells and the developing embryos of mam- mals. Arch. f. Entw’Mech. 35, p. 569. Srockarpb, C. R., anp Papantcotaou, G. 1916 A further analysis of the he- reditary transmission of degeneracy and deformities by the descend- ants of aleoholized mammals. Amer. Nat., 50, Part I, pp. 65-88, Part II, pp. 144-177. 1917 The existence of a typical oestrous cycle B in the guinea-pig with a study of its histological and physiological changes. Am. Jour. Anat., vol. 22, p. 225. Wetter, C. V. 1915 The blastophthoric effect of chronic lead poisoning. Journ. Med. Res., 28, p. 271. as AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 4 A DEMONSTRATION OF THE ORIGIN OF TWO PAIRS OF FEMALE IDENTICAL TWINS FROM TWO OVA OF HIGH STORAGE! METABOLISM OSCAR RIDDLE From the Station for Experimental Evolution, Cold Spring Harbor, L. I., N. Y. THIRTEEN TABLES Much has been written on identical twins; but, assuming that it is a fact that such twins arise from the first two blastomeres of a single ovum, not a single fact seems to be known concerning either of the first two questions which one is tempted to ask con- cerning the germs from which such twins arise. Probably the first of these questions is, If identical twins arise from two sep- arated blastomeres, why do the blastomeres separate in these special and particular cases? Second, What functional differ- ences characterize two such ova that produce male twins in one instance and female twins in the other? It is possible that the data presented here do not supply us with a fact concerning the reason for the separation of the two blastomeres in these occa- sional instances. A suggestion on this point is offered. But the present data do, beyond question, give us a fact concerning the functional status of two particular germs which produced two pairs of female identical twin ring-doves. In my earlier studies? on the eggs (yolks) of doves and pigeons it was learned that males arise from eggs (yolks) of lesser storage metabolism (small size, and higher metabolism), and females from eggs (yolks) of greater storage metabolism (large size, and 1 High storage metabolism is to be interpreted as low (oxidizing) metabolism. 2 See, a) Science, N.S., vol. 35, pp. 462-468, March 22, 1912; b) Carnegie Year Book, no. 12, p. 322, 1913; c) Bulletin of the American Academy of Medicine, vol. 15, no. 5, pp. 265-285, October, 1914; d) American Naturalist, vol. 50, pp. 885—410, July, 1916; e) Journal of the Washington Academy of Sciences, vol. 7, no. 11, June 4, 1917; f) Science, N.S., vol. 46, pp. 19-24, July 6, 1917. 227 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, No.2, JULY, 1918 : 228 OSCAR RIDDLE lower metabolism). I have now obtained two cases of female identical twins, and am able to know that the ova (yolks) which produced both of them were extraordinarily and abnormally large. We may first note that these two instances supply a strong confirmation of my earlier conclusion concerning the correlation of high-storage yolk values and femaleness. Next we may examine the data which demonstrate that the twin-producing yolks were of exceptionally large size. SIZE OF EGGS AND YOLKS IN RELATION TO THE TWO CASES OF TWINS The yolk of a dove’s egg cannot, of course, be directly weighed on a balance and then be incubated with the hope of producing young. But, as we shall see, the two procedures on the same egg are unnecessary to the demonstraiton that the twin-producing yolks were of extraordinary size. Again, the weights of two en- tire eggs (yolk + albumin + shell) from the same bird, and the same clutch, may be different and yet this difference not indicate which egg contains the larger yolk (ovum). But the amount which two such eggs may thus differ, without showing the direc- tion of difference in the yolks, is limited. And, for the species under consideration, as well as for several others, the limits of such difference are now approximately known. Accurate direct weighings of nearly 15,000 yolks of various pigeon species and hybrids have been made. Probably nearly 4,000 of these are egos of the species which produced these two instances of twins. Certain aspects of these data will be utilized here for the purpose just indicated. One of the twin-producing eggs was the seventy-sixth egg laid by that particular female. The other twin arose from the fiftieth egg of another female. In our main study the complete egg-laying history is kept of some hundreds of doves and pigeons. We are here able, therefore, to give the exact weight of every egg produced by each of these two twin-producing females prior to, as well as after, the appearance of their twin-producing eggs. For the blond ring female (No. A248) this record is given com- TWO PAIRS OF FEMALE TWINS 229 plete in table 1. The similar record of the hybrid* female (No. 60) is given in table 2. A glance at those two tables will show how conspicuously larger is each of the twin-producing eggs than is any other egg of the series to which it belongs; the tables give the necessary details for one hundred‘ eggs of one series and ninety-seven in the other. The extraordinarily large size of the eggs which produced the twins, in comparison with all of the other eggs produced by these particular parents (totals of 116 and 134 eggs), is itself sufficient to make it extremely probable that the yolks contained within them were of very large size. The certainty of their large yolk size becomes apparent in the light of the results of our accurate measurements on many thousands of eggs. It is, of course, im- possible to give all of these latter measurements here; this is also unnecessary since these will appear in connection with the com- plete account of our studies on sex in pigeons. It does seem necessary, however, to give here the particular segment of this evidence which is contained in the several following tables. Reference to tables 5 and 6 will show that the eggs which pro- duced the twins were, in one case, 24.9 per cent, and in the other, 43.1 per cent larger than the associated egg (of the same clutch). Such amounts of difference between the two eggs of the clutch are highly abnormal; except in cases of evident dwarfing of the smaller egg of the pair they practically do not exist. In the two pairs of twin-producing eggs it will be observed that the smaller eggs of the pairs are not dwarfed. A number of cases approximating to these extreme differences have, however, appeared in our records. An assistant has gone through the entire body of our breeding records and listed all pairs of eggs in which the members of the pair differ by 20 per cent or more (tables 7 to 12). In connection with these summaries it was thought advantageous 3 This female is hybrid between two very closely related ring-doves—Strep- topelia alba and St. risoria (2 alba, } risoria). 4 Sixteen additional eggs (to December 1) have since been laid by female A248; the largest egg weight among these is 9.40 grams. Thirty-seven eggs have since been added to the series containing ninety-seven eggs, and the largest egg weight among these is 9.11 grams. Most of the tables of this paper are sum- maries of data collected to April, 1917. 230 OSCAR RIDDLE to list all pairs of eggs (of whatever per cent of difference in weight) in which the yolks weights had been obtained and were found to differ by as much as 35 per cent. This latter amount is likewise abnormal for yolk weight differences; usually, this difference (in pure species) is only about 9 to 15 per cent. Since the egg weights of incubated eggs are thus included, the lists as given contain all of the strikingly abnormal or ill-matched pairs of eggs that have been encountered among the nearly 20,000 doves’ eggs that we have studied. These selected data are partially classi- fied in the tables (7 to 12) according to the kind of female which produced the eggs. Several of these pairs of disproportionately sized eggs have been incubated and the sex of the resulting offspring learned; these data are also fully given in the tables. Sex is certainly cor- related with the size, or storage metabolism, of the ova (yolks); and, in pure species, both of these are certainly correlated with the order of the egg in the clutch, as has also been pointed out in earlier publications already cited. In hybrids, however, and most markedly in the hybrids from the wider (generic) crosses, any regularity of the presence of smaller yolks in the first eggs of the clutch is lost. At the same time a high predominance of males from the first and of females from the second eggs of the clutch is also lost. Some instances of these conditions will be observed in the tables. But, as we have previously pointed out, there are conditions other than yolk size which also influence the sex that is to proceed from a particular yolk. A sperm from a different genus, or subfamily, may cause a male to arise from an ovum which, if fertilized by its own species, would have produced a female. Further, there is at hand considerable evidence in favor of the following interpretation: The sperms formed by hybrids, par- ticularly by generic, subfamily, and family hybrids, are of the most varying degrees of fertilizing power. That is to say, the sperms produced by a particular male vary thus, and these sperm differences are probably not devoid of power to influence both the degree of development and the sex of the offspring from the ovum with which the sperm unites. In still other words, differ- TWO PAIRS OF FEMALE TWINS 231 ent sperm from the same hybrid male may exercise opposite tendencies for the production of sex.» A probable instance (Q of 265) of this is partially described in the footnote to table 8. The females which produced the twins, and the eggs listed in tables 1 and 2, were mated to blond ring-dove (Streptopelia risoria) males. Indeed, these males are sire and son; and, further in one of the series the sire is mated to his daughter (A248). The hybrid female (alba-risoria No. 60) is thus mated with one of her parent species, and the two species which enter into her composition are closely related ones. In consequence both of these female tend—aside from special modifying conditions— to throw higher proportions of males from first eggs of the clutch and of females from the second eggs of the clutch (tables 3 and 4). The bisexual clutches (those producing the two sexes) are con- sidered in tables 5 and 6. Among these latter this proportion is 5:1, or 5:3 (?) for female A248,° and 12 : 7 for female 60. Reference to table 7 will show that all pairs of eggs obtained from pure blond rings which differed by as much as 20 per cent in weight had yolk weights which differed in the same sense as the egg weights; i.e., in this pure species, the yolk of the second egg was invariably (seven cases) larger than the yolk of its clutch mate, when the total weight of the second egg was 20 or more per cent larger than the egg weight of the clutch mate. In the case of the twin-producing egg belonging to this series the egg- weight difference was 43.1 per cent; this very wide difference fully guarantees the larger size of the yolk which it contained and which gave rise to the twins. The difference (43.1 per cent) is also far greater than that of any of the thirty-five pairs of yolks weight of alba-risoria hybrids given in table 8. That table shows that every egg-weight differ- ence of more than 14 per cent correctly indicated the direction of the difference between the pairs of yolks produced by these 5 The factual support of this unorthodoxy must, after consulting the papers cited under note 2, in part await the publication of, a) C. O. Whitman, Posthu- mous Works, vol. II (The Carnegie Institution, in press) and, b) our own forth- coming work. 6 Some evidence from inbred relatives of this bird possibly indicate a slight contamination of St. alba in this female. Za OSCAR RIDDLE hybrids. In only one case of the thirty-six clutches listed-— one in which the difference in egg weight was only 13.9 per cent— did the difference in egg weight fail to show which egg contained the larger yolk. This failure, moreover, concerns the first pair of eggs laid during the life of the bird, and this has long since been observed to be a clutch in which the usual order, both of sex delivery and of yolk size, is more often disturbed or reversed. It may be incidentally noted that the first clutch produced after prolonged reproductive rest is similarly disposed—even in pure species—to supply a large yolk to the first egg and a small one to the second egg, and to reverse the normal order of the resulting sexes in the same sense. : Table 8 will show that the other twin-producing egg, that of hybrid female No. 60, was an extraordinarily large egg when compared with the extraordinarily large eggs produced by any and all of our hybrids of similar or related kinds. Only three eges of the fifty-two pairs from all similar sources equaled it in size. The data of the table leave no doubt that this twin- producing egg contained a yolk of unusually large size. The effects of hybridity in the female on the regularity of delivery of larger eggs and yolks in second eggs of the clutch (as indicated by the segment of data contained in the several tables) may be summarized as follows: syste Eggs 2nd larger, 20; smaller, Pure’ St. Tisoria.c. eee ae nae eae Dad lereeneuenelion (table 7) d ye : Eggs, 2nd larger, 6; smaller, 0 Pure St. alba and T. orientalis. . ean Sud: dangers 42cm alleraaalh (table 9) Abort ¥ Eggs, 2nd larger, 48; smaller, 4 Risoria-alba hybrids............. ete Bad lager aan emilee ’ (table 8) ae a f{Eggs, 2nd larger, 14; smaller, Miscellaneous hybrids........... pee Sad lnbeen “1 ermetlen (table 10) : ; Eggs, 2nd larger, 2; smaller, Common pigeons’: =. career eae Sidi vlaeen tyes Abuse 12) Alba-orientalis hybrids (generic {Eggs, 2nd larger, 27; smaller, 3 (tanle’ti) hybrids) hea. ee wisi ee ok Yolks, 2nd larger, 26; smaller, 34 TWO PAIRS OF FEMALE TWINS 233 It is clear, therefore, that these data justify the preceding state- ment on the effects of hybridity upon the normal size relations of the two yolks (ova) of the pigeon’s clutch. They also empha- size the fact that the two twin-bearing eggs were from bird groups which assuredly throw a very high proportion of larger yolks in the second egg of the clutch. Both twin-producing yolks were, as already noted, the second of the clutch. The two series of breeding records which supplied these two cases of twins (tables 1 and 2) afford an opportunity partially to illustrate still another condition which affects the sex production, and to a certain extent the storage capacities, of the ova of pig- eons. In earlier papers we have referred to ‘crowded reproduc- tion’ merely as an aspect of ‘reproductive overwork.’ We can here particularize to the extent indicated in tables 1 to 4. The summaries at the bottom of tables 1 and 2, and the four divisions (columns 2 and 3) of tables 3 and 4, show that the sex ratio changes with the rate at which the eggs are produced. In the case of 9 A248 (table 1) those clutches which were separated from the preceding clutch by an interval of eight days or more yielded 269 : 1392 (sex unknown 9); those of seven-day inter- vals, 9o° : 239 (unknown 8); and those of six-day intervals, 4% :89 (unknown0). For thehybrid 260 (table 2), the corre- sponding figures are: 26c7 : 27? (unknown 8);50 :1192 (un- known 2); 2° :9@ (unknown 5). Under the most crowded reproduction (six and seven days) in these two series it is clear that there is an undoubted deficiency of males, even if all of the eggs of unknown sex value were classed as males. The fact that one pair of twins arose from a clutch with a six-day interval (table 1) and the other from a seven-day interval (table 2) is therefore significant for the purposes of the present paper. The actual time intervals involved predisposed, so to speak, these clutches to femininity, and both our published and unpublished data show conclusively that femininity is correlated with a high storage metabolism of the ova. Tables 3 and 4 supply still a different method of analysis of the relation of ‘crowded reproduction’ to sex. In column 2 of those tables the whole period of egg laying (shown in tables 1 and 2) 234 OSCAR RIDDLE is divided into its natural divisions; i.e., into the actual periods of work and rest of the female parent. It will be observed that both females threw highest proportions of male offspring from the longest clutch intervals and fewest from the shortest intervals. In the eight instances there is not an exception. Moreover, when each of these natural periods is subdivided into a first and last half and the clutch intervals and the sex ratios are calculated anew (columns 2 and 4, tables 3 and 4), the same fact is again demonstrated. When the undivided natural periods are put alongside the sex ratios, calculated as percentages, the figures speak for themselves: 2 A248 2 60 16.5da.= 26: 19 = 33.3% females 14.5da. = 129 :109 = 45.5% females 7.6da. = 160 :109 = 388.5% females 9.9da. = 10c' :102 = 50.0% females 7.4da. =179:1792 = 50.0% females 7.7da.= 60 :159 = 71.4% females 7.0da.= 47 :162 = 80.0% females 7.1da.= 47 :119 = 73.3% females Twin from six-day interval Twin from seven-day interval In this connection we may consider the question whether the parents continued to produce eggs immediately after the twin- producing eggs were laid and what was demonstrated as to the sex ratios in these eggs. Reference to table 2 will show that the hybrid female (60) laid only one other clutch of eggs before taking a rest (forty-four days). We are therefore unable to say what the sex ratio from this female would have been had she continued ‘crowded reproduction.’ The period of rest is, of itself, however, an evidence of the weakness (following overwork) which we have learned to associate with a high proportion of females. But even the next complete reproductive period following this period of rest yielded eleven females to four males, with two additional embryos too weak to hatch. In the case of the other twin- producing parent ( 9 A248, table 1) it will be observed that egg laying was continued at regular seven-day intervals for twelve clutches after the twin-bearing clutch. From these eggs four males and sixteen females were produced, and four eggs were too weak in developmental power to permit us to know their pros- pective sex. Here one series of eight consecutive eggs produced females; another unbroken series of five eggs produced females. TWO PAIRS OF FEMALE TWINS Pan Unquestionably, in the parents which supplied a test of the matter, the twin-bearing egg was immediately succeeded by the production of a high proportion of female-producing eggs. The question of variations in the relation of yolk weight to egg weight under the conditions of reproductive overwork and of ‘crowded reproduction’ is not answered by the data of the tables given here and lies outside the scope of this paper. For the present purpose, and by way of summary, it may be observed that both cases of twin-producing eggs occurred, a) in reproduc- tively overworked females; b) in periods of ‘continuous activity;’ ¢) in very short intervals—six and seven days—since the pre- ceding clutch, and, finally, that such crowded reproduction tends to produce an excess of females. DOUBLE-YOLKED EGGS IN DOVES AND PIGEONS There remains for consideration the possibility of the origin of the two cases of twins from ‘double-yolked’ eggs. Two sets of facts show that this did not occur. We may first note the con- clusive data obtained from the twins themselves, and from the eggs that produced them, and later produce the record of the few cases of double-yolked eggs that have appeared in our studies with doves and pigeons. The pertinent facts concerning the first pair of twins (from 260) are as follows: This egg was laid on March 7, and failing to hatch on March 22, was opened for examination. Two nearly full-term dead embryos were found; both birds were plainly smaller than normal birds ready to hatch, but both seemed practically completely formed and ready to hatch. A very considerable amount of yolk, however, remained unabsorbed, and both umbilici were plainly united at a nearly common point on the single yolk-sac. Both young were plainly females; in one of the two there was a distinct right ovary as well as the usual left ovary, but I was unable to make sure that a similar right ovary was also present in the other. The facts obtained on the second twin-bearing egg (from 2 A248) were as follows: When this egg was candled (held toward the light, as is done on the second to the fifth day for all eggs 236 OSCAR RIDDLE incubated) to test its fertility on the third day of incubation, two embryos were plainly seen. It was then noted that the two were close together; that in moving, turning, or shaking the egg, they invariably turned together, and that their position with reference to each other could not be altered. Clearly they were contained within the same ovum. It was thus known in advance that this was a twin-bearing single yolk. On the fourteenth day of incu- bation—just before the young were due to hatch—the egg was opened so as better to learn the conditions presented by the twins. Two young were found, one dead, the other alive. The dead young was practically a full-term embryo, perhaps slightly larger than the live one which seemed nearly ready to hatch. There remained here also a considerable amount of unabsorbed yolk; and, as in the previous case, the umbilici had a practically common point of union on the yolk-sac. Both birds were plainly females’ and both birds possessed right ovaries which were one-half as large as the left ovaries. Five ‘doubled-yolked’ eggs have appeared among the approxi- mately 20,000 doves’ eggs that have been examined. The size of these eggs compared with the other egg of the clutch, and with the size of the eggs of the immediately preceding and succeeding clutches, is given in table 13. Four of the five cases occurred among the eggs of hybrids. The one case of a female of pure species (Stigmatopelia senegalensis) was supplied by a female which otherwise showed the following reproductive abnormalities: Two clutches immediately preceding the double- yolked egg were clutches of one egg each; previous to these she had laid fifteen clutches, fourteen of which consisted of two eggs. The double-yolked egg was the last egg produced during the year (November 25), and the last in life for this bird, except that an ege was present in her oviduct when she died three and one-half months (March 5) later. The second of the double-yolks was produced by a female hybrid (alba x risoria) from her third egg in life. The two yolks of this egg were of most strikingly abnormal size—both together being * 17 The sex can usually be definitely learned in nine- to ten-day embryos of those species whose incubation period is from fourteen to fifteen days. TWO PAIRS OF FEMALE TWINS 237 little more than one-half the size of one normal yolk for birds of this kind. The immediately preceding eggs and the succeed- ing one were similarly much undersized. And, further, this female produced eggs at an abnormally slow rate during the entire year. The third double-yolked egg was produced by a generic hybrid (T. orientalis < St. alba), whose reproductive record seems to be fairly normal for her kind. She was mated to another female when the double-yolked egg was produced. | The fourth and fifth of these double-yolked eggs were also produced by generic hybrids. The fourth was from a Stig. senegalensis < St. alba hybrid. It was produced only five days after a preceding clutch—a thing of most unusual occurrence; in another case she produced two clutches, of one egg each, only four days apart. The fifth egg was from a complex generic hybrid (of three species, orientalis, risoria, alba), which produced eggs during only one season, and only three clutches after the abnormal one containing the double-yolk. From the above records it seems clear, therefore, that double- yolked eggs of doves are practically restricted in their production to hybrids from wider crosses or to birds showing striking repro- ductive abnormalities or to both of these. The history of the few cases of double-yolks that are known would indicate, then, that such eggs would not be expected to appear in the series in which the two cases of twins were found. ON THE CAUSE OF THE FORMATION OF IDENTICAL TWINS At the beginning of this paper it was stated that on the basis of the present data a suggestion could be offered as to the reason for the occasional separation of the blastomeres which leads to the production of identical twins. Possibly the data do not really provide such a suggestion; but, knowing that the two eggs that produced two pairs of identical twin ring-doves were from yolks § These eggs may of course be considered as of the same clutch; in this case the abnormality would consist in the time interval being a four-day period instead of the normal forty hours. 238 OSCAR RIDDLE of most extraordinary size, the writer can not but wonder if there existed a causal nexus between the extraordinary size, on one hand, and the unusual separation of the blastomeres on the other. Plainly the main question is, Why, or by what means, is ‘inde- pendent’ development instead of codrdinated, mutual, integrated development initiated in the two blastomeres? One means already known for obtaining this ‘independent development’ is that of physical separation of the blastomeres. Surely, the exact placement and position and inclination of the early blasto- meres (meroblastic eggs) are not wholly out of reference to the size and to the consequent polar configuration. And surely the type of cleavage, normal to normal blastula-formation, etc., is not out of reference to the normal size and shape of the ovum. A somewhat unusual disposition of the segmentation spheres at the animal pole—these being, at their outer borders, abnormally raised in extraordinarily large eggs and abnormally lowered in extraordinarily small ones—would thus seem to afford a possible clue to this relatively rare occurrence. in holoblastic eggs the egg-size might still be the conditioning factor, as in the case just noted of meroblastic eggs; for, although the blastomeres there are always in apposition, the centers of metabolic activity (nuclei, centrosomes, ete.) in abnormally large eggs would be separated to a degree unusual to the species, and thus conceivably afford a basis for the ‘independent action’ of the: first two segmentation spheres. Abnormally small ova, in divi- sion, would provide two cells with abnormally (for the species) large surface areas in proportion to their masses, and con- ceivably this may similarly result in the immediate assumption of ‘independent development’ in each blastomere. According to the view just sketched, identical twins should arise from the extremely large and the extremely small eggs of a species. Presumably such would be produced in approximately equal numbers. According to the theory of sex hitherto developed by the writer, males should develop from the smallest and females from the largest eggs. Apparently, the same size relations should hold for the sexes in twin-producing eggs. The two cases of identical twins described in this paper are two instances in sup- TWO PAIRS OF FEMALE TWINS 239 port of both of the above-mentioned views. A few cases of identi- cal male twins from extremely small ova of the pigeon would com- pletely establish the view as stated for twinning in pigeons. Our present data undoubtedly establish the relations of yolk size to sex and give much warrant for the prediction that if identical male twins ever arise from the ova of doves and pigeons they will arise from small ova. In conclusion, it may be emphasized that the correctness or incorrectness of the tentative hypothesis concerning the causes of twin-formation, as stated in the immediately preceding para- graphs, is unimportant to the main purpose of the present paper. The available data concerning the germs which gave rise to two pairs of female twins demonstrate that each pair arose from a single ovum and that each pair arose from an ovum of high storage metabolism. TABLES eee 919M 618 | 22/¢ | Id fo) yoy £9'2 | IT/21) TZ fo) L88V p92 | 22/9} TI fe) IFoo 96°8 | SI/¢ | ZO } youe Hy 9922) -8/ 20 Gx 22 || (FI) Pelt G¢°8 | S/S | cH re) Z6oM &1'8 | 8I/¢ | TO 5 POSH Gey L/S Tek } IS8V 96°9 | $1/¢ | TH Fs) * * ek Kk LK K 729M Go'6 | 9/¢ | ZN } youey 9¢°2 | 08/01] 2X ve) L81V 82°8 | 2/¢ | 2D Fs) “Iajuy 06°2| #/S | IN PSP TE) | BZ/Oll> LS ne) 99LV 10°24 | &/¢ | TD P yoy] 02°8 | 22/F | ZN } “1oyUy 188 | 61/01] 2M SILV 02°8 | 8%/F | 2a se) 62PM G68 | S8/F | TIN 2 qove Hy Go L | LT/0T| TA } 6SLV 082 | 90/F | Ta se) * Ok * KOK Sk K OK (1-9) “aquag 68°8:| S1/F |. Ca FOF 69'2 | 82/6 | ZA ne) (f) Aqua |sze°S | 8I/F | 2a 9PM SI°8'| 91/7 | TI ve) yoy G6°9 | 12/6 | TA re) ZIZLV = (06h 2 | OT/F | Ta ) SULT, $9'°01| 6/¥ | @M 56 S6FH ces | 81/6 | zn ye) you SOG6h'S | 8/F | 2a Fe) IZPM Sele Pa wT re) 98FA Cer alot) Gal. Ta: ne) ~ywy «ZIs'2 | 9/F | Ia d yous yy [G8 85\ eb/Fa\" ot Lg CSP GTS" | “S/68l-- cai se) 189V |20r'8 | 0&/E | ZO ve) yoy ST 28 08/25)" 18 ro) yoy yy Or 2a! “0/65 KL we) c99V -*|GZI'2 | 82/E | TO ne) ZLEM 69°8 | #2/E | ZI Lg seed O18 | §¢/8 | zg fo) ‘Iaju. = |08%'8 | 02/E | 2a oceM $6 24) SC/Eh) TI 2 qoywy Sonal 167s") ats Ossv |260°2 | 8I/E | Ta 2 LLEM Li Gs\ ST/e-| 6H ne) Le 9F°2 | 81/8 | SU } (Z) “Aquyy = 0482 (EN oSy OLEM 18°2Z | 81/€ | TH } 10SH o7-0F iS te } yey «G42°9 | 2/8] IV row fs otenairon rHpiaM| atva |nonaio| xas hee Cre IHpiaM| atva |Hozo1| xas Beal se |eHoram) auva |Hotat0| xas DIOSiL “WK aind azmb so hjwwau—grygy 6 fo psoovaay 1 ATava SYSII9}Sv JO SMOL JY,L, ‘(SyooM € 07 [) TeqUINU JO SV} G BATIDEL OF BUT} BIOJOG Pvap INq Suryoywy soyvorpur ,yo ey, *qSo1 9ATJONpOIder Jo sporsed yavw ‘(Sutuurdeq ay} WOIy payeqnour Ayeso[o a1v 9S9Y} VOUIS S859 PUODES IOJ UNIO} [[NJ = SAvp FT) Suryoywy jo yurod oy} of “aT ‘ue, T[Ny 0} pedoyaasp skvp CG] poyavur asoy} foArquia oy} Aq pourezye 98v oY} 97BOIpUT ,“AqUIO, UOTJVTAVIGGe ay} oJ1SOddo srzaquinu dILJ, ; O00T M 92631 P88 M 606M It8M SSM LO8SM ‘IOjUyT FOLM S62 (OL) “aqui O&M. OSL TOLM £0°6 uMouyUN x9g ZO 10 éd Id zO 10 oN IN oN TW oT iba | om IM or fo) © 241 Or OF Or " 0 8 6 58: Ob = "Bp 9 °S8S: 6 = “ep LS SST: 9% = Aude Skup +. g IO g sayoINIO youeyy 798 | 9/E | 2 © 987 youe yy OF. P/ Es 1s) re) 88ST } ZLEM tL 8 | Fe/e | Ca fo) udyorg re) 188M Iv'2 | 02/0 | Ta ro) Psa Fe) (¢) “aquany Ze8) | Fl/e) ca (#1) “aqui } yyy 69°2 | 21/2 | Ta Lg PICA ro) (8%) “aqui Thiel ala Gal ard @ (Z) “aquigy re) oye Ty OG se ict Seen rs) PSea 5 (FI) PETES ISS eave i eo ye) AG (ST) peTIDst OF LA 1e/ ae te) re) ‘IOjuy P i. "IOJUy 80°6 | SI/T | 24 SOL } S89 Se TEAL ol Le ye) Sa 6290 LPS Se iev } L8V } GZS So 24s E/ Pe Ly: ne) ISSV LI6I } yore yy 99°8 | 8z/ZI| ZVV| © PL8V } you] tL | 12/21] IVV| é Zo8V 62 UL 166/8 €9°2 | T&/Z SI'S | 86/Z 29°9 | 12/2 GIS | FI/2 PLL | Ste 60°8 | &/2 62 2 |. $/Z $L°2 | 92/9 F0°2 | ¥2/9 tP'S | 11/9 F0°2 | SI/9 98°8 | 6/9 66°9 | 2/9 62-8 1-1/9 LL°9 | 08/¢ If Fo Or ee EP | EG DL, OOO) Atl Gg NI 0)0) * * * ok o* IPOH €2°85| 8/61) AN ro) zI9v. |g90'8 | S/E}| Ta ne) yowy GL'8 ZI é * *« « |e «x |e «|e x | & x yorvey Lo° 2 II é Scr €0°S8 | 8/0T| ZA re) e24SV |0¢9'°8 | 22/2. | Ca o you yy 16°2| 92/01] TA ve) O6SV |089°8 | S2/Z | TH fo) 16ZV 80°6 cH re) cae eared Notre ely Dar) eerie ore (ST) “aqua jd TH é 029 crs | F1/0T| 20 o) 19SV._ |089°2 | 91/2 | ta 3} LLSA Gl | 20/00| 2a ne) es9V 088 | F1/z2 | Id é 126 00°6 ZO ne) * * « | x «x | & x | we x | & & |e 106 SPs jt) é COP C6 Zale eal cl ne) sosy 0808 | &/2| ZO v2) 6FSH 79'2 | 61/6 | TL a) 4eSV |00r'8 | $/2}| TO se) (FT) “Aquigy S16 oA } oe Re esken Gell cee Gk dl) eee || tek ‘AVP VIGIL, 8's LAL O6LH 29'L| ¥/6| ZS é 6scy |0r9'8 | e2/T | ro) : gcen 1p Z|, 2c/6-|| 318 fo) 916 £9°8 aC } $cc | Ke ee look ake les ORE! ee oe CSSV OST Sa til ck fo) $16 9¢'°8 TH 6 00ca 06°2 | 02/8 | ut Yo Ss9SV- (S162 | 6/1 | IV ne) 66rA GL LACS1/8 1) i) 886 89°8 AGI ro) * *« «|e «x |e & | & x | & x 9T6L 696 168 Id fo) SSea SP 2. | 2/8 | co — ———— 978 Be a Leal ye) TOPa = OZ. 8 | O8/ZT} ZA ve) 9&6 8Z'8 ZO re) eR a lhe oe | Te al| Pee ys emer SIfad |008°2 | 82/21] 10 ve (GT) “IqQuagy 09° 2 Te) 3} “quay 12°9 | 1/2 | 2d fo) POSH GL Quleod/ 24 tal 3} (8) “1quigg = «(08h S | ST/Zt| .L OFS 'T 61°8 od a 09F T PLL 1a “Iquagy reed | rd iballaera @) é S9CH (6222 | G/ZI| ZS ro) Pola 92°2 | 08/9 | TO } 66¢0. — 100181 €/al| 18 ve C19" T 6¢ V Z 4 v Q a 4 4 0 g z 4 4 & a eons 5 R ara 5 a5 5 ‘ 7h 7 Re 6 ie] (oe) . mw by b=} “Duos IS) §/T GID 48 8/L—09 6 fo psooog 6 HTAVL 242 , ; "4s0s dATJONpOIdol JO Spotiod YIVUI SYST10}Se JO MOI OYJ, *(SY9OM ¢E OF [) JOQUINU 10 5B} B BATOIOI OF UIT] BIOJOq PvOp nq SUTYO VY SoYVoIpUr Yo Vyy, ‘(SuruUIZeq oY} WOT] poyeqnout AJosojo 91B OS9Y} BOUTS “SIIO PUOIOS 1OJ UI10} [[NJ = sXvp FI) Suryoyey jo yutod ayy 09 ‘a'r ‘uI194 [[NJ 0} pedojeasp sup eT poyreur osoyy ‘oArquie oy} Aq poureyze o3v oY} OPVOIPUT ,“AQUID, UOTPVIADIQe O44 oytsoddo sioquinu 9qy, ; C Z 8 uMouyUN xog 56: 22 = Shep 9S SIT: OG = skep 2 641%: 9% = Wede skep + g JO g ‘soyoyNID c6ca SPT Ga ltko Uh PGE fo) SEH 06°24 | &2/9 | ON 6 O0Ea 06Z°8 | 61/TT| Tu e T9OPM TheS5|6/ Pale 6 Youve A ¥6°9 | 12/9 | IN ee OR Se ee oe eee oe: || Las RG oe €SPM 862] T/F| Ta ve) 698V €0'8 | #2/6| O fe) Gucl 888 | ST/9 | CIN 6 (FI) “rquaig O16 | S2/€ | CH 6 OL 06°2 | €1/9 | TIN o 99EV 06'S | F1/6 | 2d fe) TEP ST-Sesare.s|| Val 2 GSEV 18'L | 21/6 | Td é * * * & |x & |e & | oe oe | OF (G) “aquagy 16° 2 7/9 | @1 S0PM 8'9 | ¥/€ | 2a e SoSV G85 G/ 9a al fo) YyoueH 99°8 | 9/6] ZO Fo) “rojuy C6 elwee/2))) 1a yoy 90°8 | */6} TO é S8EM (Vee el rl WAGe rae) fe) Toca 19°21 €2/S>\" 1 6 e9ea 10°6 | 92/8 | oN fo) 668M Cor Zul eOL/S4|- tO 2 98a 1¥°8 | ¥¢/8 | IN fo) 189 09°2 | 2/1 | ca ie) Z88V 09°2 | ST/¢ | Te re Coed 216 | ST/8 | ZN o “TOJUT PLL | 2/1 | 14 OZE El 60°8 | &1/8 | TN 3} x * * «x |e & |e & |e & |e # GSLV O18 8/¢ al 2 (2%) “1quigy C02 Meals 6GLV €6°2 | 9/¢ II 6 (401) “1qQuay Sh'8 | 9/8] CT 899 Cyber Tall aL fe) qo ey Gog is /s |) Tel é OZLV 91°8 | 08/F | ZH re) LI61 eGLV 80°24 | 82/7 | TH é TIZV 62°8 | 8/L | OM 3} ke ee (ae | Ser aeealce: sae || 0G) rec rattay Cote a oh VAP 113 8 * * *& & | & & | ok ok | ke & |e # 899V OST'S | S/S |-ZD é Plt OGL s\eGGG| eux fe) yoveHE §=«f002'2 | eT/e | TO 2 CSCV 60°2 | 06/2 | SE Fo.."o 99h 6Z°8 | 0Z/ZT| TX © GIZV GES) (81/2) abe 243 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 2 TABLE 3 Further analysis of the data of table 2, 2 No. A248 (nearly or quite pure St. risoria) | NUMBER i) CLUTCH INTERVALS OF PRECEDING COLUMN WEIGHT OF FIRST AND PERIODS OF REPRODUCTIVE ACTIVITY DIVIDED INTO FIRST SECOND EGGS OF CLUTCH, AND REST AND SECOND HALVES; es AND NUMBER ©’: Q FROM NUMBERS, co’ AND Q ie EACH AND SEX NOT KNOWN Z TDKe ~ = e a 30 f Ratio Duration, kind and rate of reproductive | .& “ 36 8 2) = 3 period - D Seed g ie 8G “ gS LOS | aire re ia lagen 3 Hil eee iieeis eal eel al ae x 'S) e | Zia oi ZB tS Sa ite days 3/7/16—9,'23/16 Ist |7.14157 :59}(1) Continuous activity 7.2/8 :10 4 | 7.739)2nd |8.3430 :591](8) Average days between (22)| 7.6/9 :7 6 | 7.575|Ist 17.1446 :39/(8) clutches = 7.4 days (17c': Sara heal aoe 179) 2nd |8.00/4o7 :49](8) 9/24/16—10/30/16 Ist |7.39}/19 :19/(0) Interrupted activity 16.5/2 :1 1 ed GG Se Average days between (2) 2nd |7.93/1o" :09/() clutches = 16.5 days Qo: 19) 12/1/16—4/9/17 Ist |7.49|40 :39](0) Continuous activity 8.1 |7:4 4 | 8.06 |2nd |8.64/387 :19]| (8) Average days between (15) |7.2 |9 :61 2 | 8.34 jIst 17.67/50 :39)(0) clutches = 7.6 days (167: "cal hal nna nes |e 102)! 2nd |9.01)/407: 3911 (1) 4/16/17—8/3/17 Ist |8.10/20 :39)() Continuous activity 7.0/2 :8 2 | 8.52 /2nd|8.94/0 :59/(1) Average days between (12) | 7.0/2:8 2 | 8.26 jist |7.78|10 :49)(1) clutches = 7.0 days (4c: Frgt es ke 169) 2nd |8.73)10% :49)|(1) Total from Ist of clutch = 24° :229 (unknown, 4); 2nd = 16 o1:219?2 (unknown 13). ! The last egg of this period produced female twins. 2 There is an additional female twin. Reference to table 1 will show that there were 13 second eggs of the clutch, predominantly female-producing eggs, which were (in 12 cases) too weak in development to permit a test of sex. The sex of only 4 first eggs of the clutch is unknown; only 2 of the 4 were too weak to hatch. This undoubtedly makes both the apparent number (21) and the proportion of females from the second eggs of the clutch much too small. 244 TABLE 4 Further analysis of the data of table1, 9 No. 60 (hybrid?/8 St. alba, 1/8 St. risoria) | NUMBER | Duration PERIODS OF REPRODUCTIVE AC- TIVITY AND REST 4/24/15—9/24/15 Continuous activity CLUTCH INTERVALS OF PRECEDING COLUMN DI- VIDED INTO FIRST AND SECOND HALVES; NUM= BERS oO’ AND Q, AND OF SEX NOT KNOWN , kind and rate of repro- ductive period = = ha P Bos oO Soe + oe ago A, og BoA % o+y ae] aae¢d s 2 as 5 = 3 So % 6) fe wan days Average days between (16)| 7.7) 3:8 6 clutches = 7.7 days (6c 7159) 11/19/15—3/15/16 Continuous activity Average days between (10)! clutches = 9.9 days] 8.4} 5 : 6? 0 (100 :109)? 4/28/16—7/14/16 6.8) 3:4 1 Continuous activity Average days between (8) | 7.4, 1:7 2 clutches = 7.1 days (4¢ :119) 7/15/16—3/25/17 bpaelintita wilaty « Interrupted activity 14.5)12 : 10 3 Average days between (13) clutches (129 : 109) 14.5 days WEIGHT OF FIRST AND SEC- OND EGGS OF CLUTCH, aND 8 NUMBER oO’: 9 FROM EACH 9 Ratio - : : a b a = 5 ‘3 yi Bi Ze la |MoM Sue ties 8.372} Ist |8.06)/0n% :49] (2) 2nd {8.79/38 :32)|() 8.360) Ist |8.16)/10 : 42] (3) 2nd |8.60|)2c7 :32)] (38) Sing. |8.03} — :19)|() Ist -|8.03/47 :02/() 8.235] 2nd |8.44/1o7 :32| (0) Sing. |8.55} — :19/(1) Ist |8.22/301 :29)(0) 8.370| 2nd |8.52/3.7:4Q2/(0) 7.655) Ist 17.55/10? :32| 0) Onde N25 |2etecelaS 1) 7.502) Ist |7.39]1o :39/|(1) 2nd |7.61/0c :49}(1) Ist |7.78/7c7 :32](2) 7.820] 2nd |7.78150 : 79] (1) Sing. |8.23] —- :12/() known 6). Total from 1st of clutch = 17% :192 (unknown 8); 2nd = 16c7 : 252 (un- in rating the time between clutches. 245 1 Eleven clutches are present, but the first following a rest is left out of account 1 Female twins were produced at the close of this period. 246 SEX PER CENT OF DIFFERENCE CLUTCH WEIGHT 40 Qy 40 Qy Q, 40 Q, 40 Q, 40 | SEX 40 Q, Q, 40 40 G2| 8.28)+17.1 K2| 8.20)4-22.4))% L2 | 8.36/+-19.6 WEIGHT PER CENT OF DIFFERENCE QD _ ~] or i) C2 |8.28 |+10.4 G1 |8.48 G2 |9.00 |+ 6.1 M1 |8.09 M2 |9.12 |4+12.7 R1 |8.290 R2 |9.178/+10.7 S1 /8.100 r a ot 2 10 SEX 40 10 OSCAR RIDDLE TABLE 5 Individual clutches of 9 A248 which yielded the two sexes 58 Es 58 Be Ba & Ze Ze & ee m | 8¢ FI E oe Be) oe 2 an ra 5 5 w 5 fa 5 5 25 Bolte Ngee de |. Bical cel aera eae 7.63 Q| H1| 7.81 oa) M1)/8.22 7.79|\+ 2.11) of H2} 9.17)/+17.4|| 9) M2/8.70)+ 5.9 7.56 GUL 7s 9} Sl /8.01 8.48/+12.2|| ?o7| J2 | 8.51/+18.5))7| 82 |8.70/4+ 8.6 7.28 o| K1} 7.48 o'| W1\7.10 8.47/+16.3]| 2 9| K2|10.63)4+-43.1]| 9} W2/8.47/+19.3 TABLE 6 Individual clutches of 960 which yielded the two sexes 68 68 68 eA me a e ag x | GB « | o8 x O fe q Ft O fe q | O & o = & S I / gk 2 | P| ee gla] e | a? gle] 2] 2° 7.915 QiIl |7.93 rofl OU Leah’ /e fe 8.150/+ 2.9]/|7)12 |8.10)/4+ 2.1]/9/U2 |8.45)4+ 9.4 8.680 31 |7.60 o|X1 |8.29 8.650|— 0.3/|9|J2 |8.61/+13.3]| 9'X2 |7.26)—14.2 8.065 o|M1 |7.90 wC1 |7.69 10.080) +24 .9]| 9|M2 |8.38/+ 6.1]}9/C2 |7.79|+ 1.3 7.700 aIQI |7.42 A\E1 {8.15 8.180/+ 6.2|)9/Q2 |7.48)/+ 0.8]| 9/H2 |9.10)/+11.6 7.08 Q|T1 |7.64 8.16 |+15.2)|7/T2 |7.95|4+ 4.0 S2 |7.727|/— 4.8 Qy 40 61; HI 9° H2 Kl 9 K2 9 Ql Q2 Kl 9 9 K2 100 Bl B2 D1 D2 Fl F2 163 Al A2 9 D1 9 D2 118} @X1 9X2 * Eggs thus marked are the first eggs in the life of the birds that produced them. 6.96 | 8.55 6.77 8.29 6.62 8.13 7.43 10.63 8.270 9.970 7.500 9.265 8.410 10.445 5.145 6.550 7.47 9.04 7.005 8.635 TWO PAIRS +22 .8 +22.4 +22.8 +43 .1 +20.5 +227 +24.2 Zico +21.0 +23 .2 Hatched Hatched Hatched Hatched \Hatched Broken ' |Hatehed Twins 1.685 2.270 1.820 2.256 0.943* 1.170 Hatched Hatched Hatched 13-14 da. embr. TABLE 7 Clutches (pairs of eggs) of greatest size disparity laid by Streptopelia risoria +34.7? +48 .7 +23.9 +24.1 86 141 122 9 Bl oJ B2 JK1 WK2 9 M1 9 M2 981 S2 STI 9T2 Kl 9 E2 ?AF1 9 F2 Al A2 Ht; H2 Gl G2 1 The purity of the female parent is questionable. 2 The figures of this column, in this table and in those which follow, represent the per cent of difference between the weights of the two eggs or of the two yolks. the smaller egg or yolk is taken as 100 per cent. 7.290 8.870 7.300 8.780 6.980 8.410 for) 7.6 9.22 7.28 8.86 7.010 8.730 7.420 8.945 7.400 9.045 OF FEMALE TWINS 21.77 +22 .2 +25.0 +20.3 +20.5 +20.3 +21 .7 4+24.5 +20.5 +22.2 Hatched! Hatched | Hatched Hatched Hatched Hatched Hatched Broken Hatched Hatched Hatched! Hatched 10-12 da. embr. Hatched 1.550 2.052 1.600 2.035 1.648 2.132 247 +32. 4? +27 .2 +29.4 In all cases 248 73 179a 101 177 3l 125k PP 162 199 281a 288a 135 Fl 9 9F2 9 Kl K2 Bl B2 D1 D2 Cl C2 Ji J2 Al A2 Al A2 D1 D2 Nl N2 D1 D2 Al A2 Bl B2 Cl 9 C2 Al 9 A2 8.065 10.080 +24.9 —23.7 +22.4 23.3 +25.3 +28.7 +20.7 +13.9 —15.2 +20.3 +14.9 +-10.9 +21.5 +20.9 +21.2 OSCAR RIDDLE Hatched Twins Hatched Infertile 1.806 2.038 1.300 1.700 1.445 1.950 1.405 2.080 1.100* 1.370 1.720* 2.080 1.827 1.330 1.170 = .400 .910 — 1.601 2.190 1.480 2.050 Hatched Hatched 0.870* Hatched TABLE 8 Clutches of eggs of greatest size disparity laid by hybrids of St. alba and St. risoria +12.8 +30.8 +34.9 +48 .0 +24.5 +20.9 —ol.o +51.5 +36.5 +36.8 +38.5 173 311 198 172 127¢ 1 N2 OD Cl C2 El K2 Cl C2 Cl C2 Al A2 Fl F2 J1 J2 Bl B2 Cl C2 D1 D2 Al A2 Bl B2 Cl C2 Bl B2 Fl F2 8.800 7.820 SAE +20.2 +19.3 eld ae +24.4 +21 .2 +22 .4 +22.4 +24.7 +14.8 +20.3 +21.5 +21.3 2.010 1.485 Infertile Infertile 1.625 2.200 0.990 1.435 0.930* 1.960 2.932 1.240 1.725 1.818 1.903 1.780 2.350 —35.3 +35.5 +44.9 +45.2 +49.6 +091 = el +32.0 +36.4 +22.6 +23 .3 +44.5 153 Bl B2 188 El K2 153a Al A2 291 Fl F2 6lb| Fi 9 K2 9X1 SX2 105 | ??Bl1 oB2 2o9'C1 C2 M1 M2 Rl R2 137 JB1 o'B2 TWO PAIRS OF FEMALE TWINS TABLE 8—Continued (aon 1.672 9.015/+22.8} 1.818 7.465 1.825 9.050/+21.2) 2.055 6.850 0.750* 8.190)+19.5} 1.800 7.140 1.550 8.470/+18.6} 2.100 8.200 Hatched 10.080) + 22.9) Hatched 7.580 Hatched 9.970|+31.5) Hatched 7.010 Hatched 8.590) +22 5 Hatched Wa2A2 22 da. (!) live embr. 8.800)-+21.5/14 da. dead embr. 7 YAS 1.625 9.340/+23.5) 2.039 7.630 1.575 9.690|+26.9} 2.125 7.140 Hatched 8.700/+21.8)Hatched +140.0 +35.5 +25.5 +34.9 295a 156 159 146 265 9 P2 2Gl 9 G2 Ll JL2 9?M1 oM2 JE1 9 E2 Bl B2 2 Ql J Q2 8.325 9.890 8.36 10.31 +20.0 +24.1 +24.4 +24.7 +23.3 249 2.890 |+46.6 2.321 |+32.6 SK Ce 1.975 |+25.0 Infertile Hatched Hatched Hatched Broken Hatched Hatched Hatched Hatched Hatched 1.520 1.810 |+19.0 Hatched! Hatched * Eggs thus marked are the very first eggs in the life of the birds that produced them. 1 These eggs were the first after a long rest that followed a long peroid of reproductive overwork. It is also important to observe that they were fertilized by a generic hybrid (Zenaidura-Zenaida) male, and that these genera belong to a different subfamily from the female that produced these eggs. sperms exercise opposite influence on the sex development. It is of some interest to add that the male that developed from Q2 was killed when quite healthy; two testes were found, but, contrary to the rule for normal males, the left testis was larger than the right, i.e., the size relations of the glands were those of a female. (Riddle, Anat. Rec., vol. 14, 1918, pp. 283-334.) There is evidence that in this type of cross different 250 OSCAR RIDDLE TABLE 9 Clutches of greatest size disparity laid by St. alba and by Turtur orientalis 23| 9 H1) 7.210 Hatched 136) Ki {7.14 Infertile ?H2| 8.860)-+22.9) Hatched ?9K2 |8.85 |+23.9|Hatched 139} L1 | 8.320 Lost eal ie/ atsn(0) Hatched L2 |10.042|/+20.7) 3 da. ?HJ2 |9.030|+20.2| Hatched embr. 155} R1 | 9.10 1.485 97) Jl |7.210 1.276 R2 {11.21 |+23.2) 2.486 |+67.4 J2 /8.920)+23.7| 1.503 |+17.8 TABLE 12 Size disparity of eggs of common pigeons 291a | B1 | 15.00 3.080 124 | Al |17.365 3.928 B2 | 15.38)+ 2.5)1.760'|—75.0(?) A2 |14.065) —23 4) 3.220)—21. 3 Ft. | Bl | 16.18 3.190 305 | Al |10.50 0.995 B2 | 13.62)/—18.8)2.350 |—35.7 A2 |11.65 |+10.9| 1.600|+60.8 1 Tt is quite probable that this figure was copied wrong when the weighing was made, and should be 2.760. TWO PAIRS OF FEMALE TWINS ; 25 TABLE 10 Clutches of greatest size disparity laid by miscellaneous hybrids (a few from pure species) 140b| Al | 6.618 1.543* 97 | K1 | 7.98 Hatched A2 | 8.357/+ 26.2) 1.940 |+25.7 oK2 | 9.84 |+23.3) Hatched 127a) KI | 5.39 1.095 196 |“E1 | 7.93 Hatched K2 | 6.72 |+ 24.7) 1.160 |+ 5.9 Q@K2 |10.03 |+26.5| Hatched 286 Al | 6.34 1.170* 279) Ady s524i 1.080* A2 | 8.35 |+ 31.7| 1.850 |+58.1 A2 | 6.73 |+24.4; 1.540 |+42.6 17a SDs |A7A28 Hatched 286 | Al | 6.34 1.170 D2 {10.31 |+ 41.6) 2 da. embr. A2 | 8.85 |+31.7) 1.850 |+58.1 281 Bl 0.840 174 | Jl | 8.34 1.710 B2 0.450 |—86.7 I2 | 5.28 |—57.9| 1.738 |+ 1.6! Gl 0.690 281 JI oelO 1.075 G2 0.450 |—53.3 J2 | 5.23 |+ 2.5) 1.480 |+37.7 L1 |3.4802 0.550 273 |9Al | 5.94 Hatched L2 {4.780 |+ 37.3) 0.965 |+75.5 A2 | 7.25 |+22.0| 2-3 da. devel. N1 | 4.20 0.785 SR | C2700) 1.205 N2 | 1.96 |—114.3} 0.475 |—65.3 P2 | 6.920/+ 3.3) 1.783 |+47.9 181 Ci |) 3'.985 0.628 El | 6.800 1.420 C2 | 3.380/— 17.9} 0.448 |—40.2 E2 | 5.650)—20.3) 1.043 |—36.1 54 D1 | 9.165 1.269 114a} Lil | 7.940 1.530 D2 |10.120/+ 10.4; 1.713 |+35.0 L2 | 9.690)+22.0} 1.640 |+ 7.2 * Eggs thus marked are the very first eggs in the life of the birds that produced them. 1 This second yolk absorbed water from the albumen during fifty-two hours, the first yolk for only five hours; if this correction were made the second yolk would be shown to be smaller than the first. 2 Many eggs with imperfect shells or without shells in this series. D2 ; OSCAR RIDDLE TABLE Il Clutches of greatest size disparity laid by hybrids of St. alba and T. orientalis 3) El | 9.610 2.244 119 | Bl |10.345 2.840 E2 | 8.100|—18.6/1.485 |— 51.1 B2 | 8.005|—29.2 1.620 —75.3 98| L1 | 8.600 1.050! 100 | Li {10.050 2.155 L2 | 6.970) —23.4|1.560 |+48.6 (?) L2 | 8.360|—20.2 1.870 —15.2 49| H1| 9.535 1.992 109 | Jl | 8.575 1.430 H2| 8.270|—15.3]1.370 |— 45.4 || J2 |10.200|/+18.9 2.220 +55 .2 E1 | 8.620 1.534 B1 | 9.310 2.080 E2 | 9.800)+13.7|2.220 |+ 44.7 B2 |11.500)/+23.5 3.020 +45.2 103} H1| 8.915 2.057 S9n Del 227 1.190 H2| 6.730)—32.4/0.975 |—110.97 D2 | 8.27 |+13.7 1.650 +38.6 258] G1 | 7.730 1.450 El | 8.76 1.830 G2 | 9.590|+24.0/2.440 |+ 68.3 E2 | 6.66 |—31.5 1.100 — 66.3 33] T1 | 7.260 1.615 65 | Al | 7.245 1.498 T2 | 9.120/+25.5/2.110 |+ 30.6 A2 | 8.650/+19.4 2.033 +35.7 49 L1 | 9.700 2.088 Dig 4505 1.250 L2 | 7.360) —31.8}1.480 |— 41.1 D2 | 8.840|/+17.8 1.960 +56.8 106| G1 | 8.75 [1.510 T1 | 8.870 2.070 G2 | 9.94 |+13.6)2.162 |+ 43.2 12 | 7.210)—23.1 1.230 —68.3 Cl | 8.615 1.466 K1 | 7.670 1.380 C2 | 9.470|4+ 9.9/2.080 |+ 41.9 K2 | 9.250/+-20.6 2.001 +45 .0 115) D1 | 8.370 1.905 Fl | 7.585 1.355 D2| 6.017|—39.110.915 |—108.2 F2 | 9.035/+-19.1 1.906 +40.7 30| B1 | 8.455 1.602 56a} J1 | 8.670 1.890 B2 | 9.800) +15.9'2.2°0 |+ 37.3 J2 | 6.505|—33.2 1.220 —54.9 Th 37.285 1.345 29) Gl | 9.520 2.100 I2 | 9.605)+-23.8) 2.115 |4+ 57.2 ' G2 | 8.070|—17.9 1.520 —38.2 16) C1 |10.030 2.530 W1 | 8.100 1.483 C2 | 8.600|—16.6/1.810 |— 39.8 W2 | 9.340)+15.3 2.015 +35.9 11Tt is probable that this weight was 2.050 instead of 1.050 as recorded. The usual ‘underscoring’ of yolks of abnormal size for the bird or species was here omitted at the time of weighing. TWO PAIRS OF FEMALE TWINS 253 TABLE 11—Continued 14| El | 8.120 1.295 110 | Hi | 8.360 1.525 E2 | 8.350/+ 2.8/1.845 |+ 42.5 H2 | 9.255|+10.7 2.315 51.8 112) C1 | 6.970 1.150 F1 | 9.320 1.925 C2 | 8.870/+-27.2|1.970 |+ 71.3 F2 | 6.990|—33.3 1.630 8! Lil | 8.367 2.120 206 |9Z1 | 6.04 Incubated L2 | 6.795|—23.1/1-700 | 24.7 6Z2 | 7.31 |+21.0) Incubated B1 | 8.062 1.940 264 | Al | 9.110 1.520 B2 | 6.760|/—19.2/1.325 |— 46.4 A2 | 9.850/+ 8.1 2.085 LEE 152| Al |11.220 3.035 220) | Gt neue 1.290 A2 | 8.150|—37.6|1.610 |— 88.5 G2! 5.7 alo 0.850 |— 51.8 C1 | 8.340 1.855 Hi | 7.34 1.150 G2 | 7.050|—18.3|1.147 |— 61.7 H2 | 5.76 |—27.4 0930. |=223..7 H1| 8.180 1.925 K1 | 5.57 0.774 H2| 7.095}—15.3|1.240 |— 55.2 K2 6579 1.268 -|4= 63:8 E1 | 8.950 2.130 D1 | 5.27 0.645 E2 | 5.030|—77.910.495 |—330.3 D2 |. 6.82 |--29.3 1.320 |+104.6 F1 | 7.270 1.400 2850 (ed 9reso 2.335 F2 | 9.180|+26.2/2.280 |+ 62.8 L2 | 8.600/—14.3 1.620" »|-43.43.3 H1 | 7.940 1.870 H1 | 7.440 1.235 H2| 7.240|— 9.6/1.370 |— 36.5 H2 | 9.180|+23.4 2.160 |+ 74.9 Ii | 7.990 1.820 D1 |10.098 2.502 12 |10.010/ +25 .3/2.610 |+ 43.4 D2 | 8.677|—16.3 158354 3 |2236-3 9| F1 | 9.410 2.025 18 | Al | 8.620 1.625 F2 | 7.500|—25.4/1.430 |— 41.6 A2 | 6.300|—36.8 0.620 |—162.1 G1 | 7.625 1.267 B1 | 9.075 1.840 G2| 9.155|+20.1/2.315 |+ 82.7 B2 | 7.340|—23.6 1.580 |— 16.5 K1| 9.750 1.873 104 | G1 | 9.560 1.968 K2| 7.130|—36.7/0.910 |—105.8 G2") asi7| Soa 16650 |—="18.2 Y1| 9.320 2.213 158 | C1 | 9.590 1.920 Y2| 7.440|—25.3/1.443 |— 53.4 C2 | 7.680|—24.9 1.350) . |p 409 6| Y1| 6.640 1.573 1S eC17 $5680 2.080 Y2| 8.200/+23.5/1.760 |+ 11.9 C2 | 7.550/—14.9 LjA72.~ | eARe Ss 49| B1 | 8.290 1.770 46| I1 | 6.940 1.205 B2 | 7.440|—11.4|1.310 |— 35.1 I2 | 5.570|—24.6 0.782 2 5a i OSCAR RIDDLE 254 “‘poys1om Ajozeiedes jou a19M Sy[OA OM} OL g “MOT 00} A[GBAIPISUOD A.10JOIOYY ST UIATS SB IYSTIOM SFI ‘]]OYS B JNOYIIM SBA B99 SIU, + cor T = 009°8 = “IM ZI O16S = 079 I 8 26 898 — AW CE COR ae CS Gos = AMG ¢6¢'I = 080°8 = “3M ITI 2099 3G OOTT = 7M TH 08S: IS = 00S = UM. TD OI6I ‘(Bl1osta — BqyTe X Bqye — “yusTIO “IQAY) GSTS ; L0L = 008 T-8 08 = 626 2 = M CGO S821 = 026'9 = “IM Sd oor’ T eee Lee he Se 0ZF° 2 = “3M ON G06:1.=.002 9 = IMald a Eos Set? LG at ea ae 0199 = “3M IN S16 ‘(BIe X Ssue[BBoues “1qoy) STVS.L Sd OSG = OLT 6 = IM Cf a O€ + 09'S GE + 02601 = 9M GI OCG ~ ~SE9 8 = IM CH OST 'e = 0096 = IM IL a OSTeGs-=— EAS SE O8GS Irene aes 016°8 = “3M TH GI6I ‘(BqT® X StTBpuaTIO “1qAY) 0186 Iq ul syxjod oy} pozeaedes AjoJ9[duI0D oUBAGUIOUT OUTI[AIIA 94} ANS JON a ie ard 6.08 — ST9°o = YM Cel cer 1 = 092° 2 = tM-GO £&8'0 cee OVO Or i es 002°S = “IM 6V 066-0°= 0c9 9 = IM 1D ogra ae a Gc9 0 os SOST A TAY EV. GIGI ‘(Bltostr X VqTe “Iq4Y) FETS (1194s +) a a gene pa Rat a eee OZI'T = 1¢80°F = "IM V OLLI 961 +9108 = IMS 00¢ T 0029 = IM E16T ‘(stsuayesoues “319g aand ‘qoid) 06 sayajnjo Bururolpp YUN pup yoIn70 ay} fo sbba wayjo fo syjoh puv sbba fo syybram yun paswvdwos ,payjoh-ayqnop, aq 0) unouy sb66a aay fo sjy bay el AIAVL AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 4 THE EFFECTS OF THE DUCTLESS GLANDS ON THE DEVELOPMENT OF THE FLESH FLIES B. W. KUNKEL From the Department of Biology, Lafayette College, Easton, Pa. The effects on embryonic development of the ductless glands have not as yet been very widely studied, and our knowledge of their action is extremely meager. As practically all the experi- ments on the developing embryo have been made on vertebrates, it seemed desirable to determine if possible how some of the invertebrates react to these glands in which no organ comparable to a ductless gland has, so far as I know, been demonstrated. In addition to the work of Gudernatsch on frogs and rats, recently Northrop has experimented on Drosophila and found that although the larvae of this fly do not develop normally on a sterile paste made by grinding fresh thyroid and thymus with glass, the larvae develop quite normally when bacteria are present in the cultures. Aside from these studies just mentioned, the only other animals which have been used have been Paramecium by Nowikoff, Shumway, and Budington and Harvey who also used Stylonichia. In view of the suppression of growth and acceleration of differ- entiation which has been observed in the tadpole by Gudernatsch, it seemed especially desirable to test the effects of thyroid gland on an animal in which the processes of growth and differentiation are more or less completely separated, as is the case in the insects which have a complete metamorphosis. ‘The flesh flies seemed to lend themselves to this purpose most perfectly because of their abundance and the ease of rearing them as well as because the larvae consume the ductless glands with avidity and thrive on them. The flesh flies which I found visiting exposed meat during the summer months are: Lucilia caesar, L. sericata, Calliphora erythrocephala, and a species of Sarcophaga, probably sarracena. 255 256 BEVERLY WAUGH KUNKEL Lucilia, as a rule, lays its eggs on meat before decay has set in, within a few minutes after the animal has been killed, while Sarcophaga is attracted to meat which has advanced somewhat — in the process of decay. Most of the observations made in this investigation were on Lucilia which is the most abundant genus of flesh flies in this region during the period of these observations. The eggs of Lucilia are laid in masses of about one hundred. In a large number of observations I found that the flies prefer crevices or cavities in which to deposit their eggs, at least these situations are the first to receive eggs, later the more exposed portions become ‘blown’ as well. As flies of different species sometimes lay their eggs in close contact with each other so that the limits of the two or more masses cannot be distinguished, care was taken to use only isolated egg masses of comparatively small size which were laid soon after the meat was exposed to the flies. The eggs of Lucilia hatch in about twenty-four hours and the larvae at once begin to liquefy the food material. They work beneath the surface, excavating cavities into which they sink. Finally the whole mass of food may be reduced to'a'thick, gray broth through which the larvae move freely. The larvae are very strongly negatively phototactic and are apparently very sensitive to variations in temperature and in the composition of the atmosphere. If for any reason, such as improper ventilation, the conditions in the culture glass in which the larvae are feeding become unfavorable, within less than a minute there is a whole- sale movement on their part to escape. L. caesar and sericata are very closely related species which are distinguished by the arrangement of the spines on the thorax. I am indebted to Mr. V. A. E. Daecke of the Pennsylvania State Agricultural Department at Harrisburg for the identification of the flesh flies. Lucilia may be distinguished by the bright metallic luster of the body, which is green. Calliphora has a black thorax and an abdomen which is a deep metallic blue. The species of Sarcophaga are grayish, and can usually be dis- tinguished by the checkerboard pattern of gray and black squares on the abdomen. EFFECTS OF DUCTLESS GLANDS ON FLIES 257 The method of procedure in the experiments made was very simple. The food material with the eggs on it was placed in ordinary glass tumblers having a capacity of about 200 ec. which were covered with voile fastened with a rubber band. By examining the glasses several times a day any eggs deposited on the voile may be removed before they hatch and contaminate the culture. In later experiments a canopy of mosquito netting was stretched over the glasses to prevent the near approach of flies. When it was desired to obtain the pupae, it is more convenient to place the food in a Petri dish and place this upon a layer of sand 1 or 2 cm. deep in the bottom of a large culture glass 200 cm. in diameter. This large dish was covered with a glass plate elevated at one side by a few folds of paper for ventilation, and the whole was covered with a mosquito net and kept shielded from direct sun light. The culture dishes were always kept side by side under as nearly identical conditions of illumination and temperature as possible and in order to reduce any accidental differences, the positions of the dishes were reversed from time to time. As far as possible, in each observation a single egg mass wasemployed, one portion was used for the experimental feeding and the other for the control. It is somewhat difficult to handle the eggs of the flesh flies with- out injury because of their stickiness. In consequence it fre- quently occurred that the division of the egg mass was far from equal. The glands used for food were obtained fresh from the slaughter house and for the most part were obtained from calves, although some were obtained from steers and sheep. The control food was a mass of muscle approximately equal in bulk to the thyroid or thymus used. Anadvantage in using these relatively small:masses of food is that the heating due to active putrefaction is less. The effect of the ductless glands on growth and differentiation was measured by the length of the larvae, the length of the pupae, and the duration of the larval and pupal stages. In order to measure the larvae, they were killed by immersion in boiling water or 95 per cent alcohol. Both these agents seemed to cause very little contraction or distortion. It was 258 BEVERLY WAUGH KUNKEL found difficult to measure the length of the larvae very accu- rately because of their consistency and the protrusion of the mouth parts. The length was obtained with dividers and a milli- meter scale. It was soon found, however, that sufficient accu- racy was not possible by this means, and so in later experiments the length of the pupae was used as an index of the effect of special feeding on the growth of the larvae. By means of micrometer calipers it was possible to measure the pupae to .001 inch. A number of observations was made to determine the effect ofjthyroid feeding on the length of the larvae. From these, the following have been selected because of the completeness of the data regarding the flies. In all the observations on the length of the larvae there was an abundance of food present so that effects are not due to insufficient food. Unfortunately, I did not identify the adults, but judging from the characters of the larvae, there is no doubt but that we are dealing with Lucilia. Measurements were made of larvae varying in age from two to eight days and in all these cases except in the youngest larvae of two days’ age, the muscle-fed larvae were slightly longer than the thyroid-fed. Except in the eight-day larvae, the difference in length is rather insignificant. Great care was taken in the meas- urement of the larvae not to estimate the thyroid larvae low and the muscle larvae high. The results are shown in the following table: TABLE 1 Showing the effect of thyroid feeding on the average length of larvae of different ages Agee ofdarvae inidays,. Tin week iee. aoeek 2 3 6 7 8 Ree mass mumber: (A sc tsbaeeerte ede ck aie 15 17 18 5 4 Average length in mm. thyroid-fed............| 18 12.4 | 15.6 | 10.9 | 13.4 Average length in mm. muscle-fed............ 12.6 | 12.6) 15.9 | 1174 | 14.9 Table 2 indicates the distribution of the larvae in egg mass No. 5. In this experiment, after seven days of feeding twenty larvae from the thyroid and twenty from the muscle, chosen at random, were measured after killing with boiling water. EFFECTS OF DUCTLESS GLANDS ON FLIES 259 TABLE 2 Showing the effect on the length of larvae of feeding thyroid exclusively. The first line indicates the distribution of the thyroid-fed larvae according to length, the second line indicates the distribution of the controls which were fed on muscle Gheruide er er et 1 5 0 9 3 1 Mruscletas cc. 8 ase 0 0 1 3 6 5 5 Length in mm......./9.5-9.9/10—-10.4/10.5-10.9/11-11.4/11 .5-11.9}12-12.4)12.5-12.9 In order to compare the effects of thyroid and thymus feeding on the length of the larvae, reference may be made to egg mass No. 20. This was a very large egg mass numbering nearly 200. Portions of the eggs were placed on thyroid and thymus of a calf, respectively. At the end of four days the larvae were killed with boiling water and measured. The average length of 107 thyroid larvae was 11 mm. and that of 71 thymus larvae was 14 mm. (table 3). } TABLE 3 Showing the effect on the length of larvae of feeding thyroid and thymus, respectively. The jirst line indicates the distribution of the thyroid-fed larvae according to length, the second line indicates that of the thymus-fed larvae Thyraid....| 6 | 34| 29 6 gen fot lente 3 | 0 0 Thymus... 2/0 0 0 4 7, 15 12 15 6 2 Length in MMi cas 8-8.9|9-9.9/10-10.9}11-11.9 12-12..9|13-13.9|14-14..9)15-15.9| 16-16 .9]17-17.9 The effect of thyroid feeding on the duration of larval life could not be determined with great accuracy, as pupation is retarded if the larvae are prevented from burying themselves or if they are disturbed after burial. The length of time during which the larvae feed is not affected markedly by the ductless glands selected for the experiment. This is rather interesting in view of the fact that the pupation of mature larvae of Lucilia caesar may be retarded in several ways; such as by cutting off the supply of oxygen, as Dewitz (’01) has shown. I have noticed also.in my own experiments that if the substratum is too dry the pupation of mature larvae may be delayed from f en hours, which is normal, to four or five days. Nearly all my evidence indicates that there is some hastening of maturity effected by feeding thyroid gland to larvae of the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26,NO.2 260 BEVERLY WAUGH KUNKEL flesh flies, although the acceleration is not great. In one excep- tional case, that of egg mass No. 28 of L. caesar there was a slight retardation in pupating effected by thyroid feeding. My record in this experiment shows that on the sixth day after deposition of eggs, the muscle larvae were all buried while the thyroid larvae were in proeess of burial. The following day 8 per cent of the thyroid larvae had pupated and 16 per cent of the muscle larvae. On the other hand, eighteen days after the deposition of the eggs, the thyroid-fed had outstripped the muscle-fed, for 57 per cent of the former had emerged and only 18 per cent of the latter. Egg mass No. 16 of L. caesar, in which the thyroid-fed larvae were markedly smaller than the muscle-fed affords evidence that thyroid feeding accelerates development. ‘Ten days after depo- sition of eggs, the thyroid culture contained 94 pupae and 1 larva, while the muscle culture contained only 43 pupae and 11 larvae. That is, 99 per cent of the larvae had pupated in the thyroid and only 80 per cent in the muscle culture. Egg mass No. 9 also indicates that the pupal period may be shortened slightly by thyroid feeding. In both experimental and control cultures pupation was first noted on the tenth day following the deposition of the eggs. On the fourteenth day, five out of fourteen pupae had opened while none of the muscle pupae had matured. Within eighteen days after the deposition of the eggs all ten flies from the muscle-fee larvae had emerged, but the rest of the thyroid-fed were dead. Egg mass No. 16 also affords some evidence that the pupation period may be slightly shortened by thyroid feeding. Fourteen days after deposition of eggs, 43 out of 94, 46 per cent, of the thyroid pupae had emerged while in the same time only 2 out of 50, 4 per cent, of the muscle pupae had emerged. ‘Two days later there were 36 muscle flies emerged and 88 thyroid flies, or 72 pert cent of the muscle-fed flies had emerged in the same time that 93 per cent of the thyroid-fed had. Egg mass No. 22 of L. sericata also shows that there is a slight hastening of pupation effected by thyroid feeding. Six days after egg deposition there were found in the muscle culture ©) pupae and 57 larvae which had buried themselves, while in the EFFECTS OF DUCTLESS GLANDS ON FLIES 261 thyroid culture there were 46 pupae and 29 larvae. That is, less than 9 per cent of the muscle larvae had pupated in the time in which over 60 per cent of the thyroid-fed had pupated. Eleven days after deposition there were 54 pupae in the muscle culture and 66 in the thyroid. The experiments to determine the effects of thyroid feeding on the size of the pupae were on the whole rather disconcerting, for although accurate measurements were possible there were less constant and definite differences from the controls than in the experiments with the larvae. In two experiments with L. sericata the pupae derived from muscle-fed larvae were markedly longer than those derived from thyroid-fed. A similar difference was noted also in one of the experiments with L. caesar. In two experiments to compare the effects of thyroid and thymus feeding, there was a very slight superiority in the length of the thyroid-fed pupae. In two experiments, on the other hand, in which egg masses were placed on muscle and thymus, respectively, there was a superiority in length among the thymus-fed. It is evident that the results are not striking and that the effect of thymus and thyroid probably are to a slight extent in accord with the findings of others who have worked with tadpoles. Egg mass No. 13 of L. sericata is typical of a number of experi- ments. Larvae were fed exclusively on calf thyroid and muscle. After ten days all the larvae in both cultures had pupated. The distribution of the pupae is indicated in the following table (table 4) : TABLE 4 Showing the effect on the length of pupae of Lucilia sericata of feeding the larvae exclusively on thyroid. The first line indicates the distribution of the thyroid-fed pupae according to length, the second line that of the muscle-fed Thyroid...) 2 4 2 3 7 5 10 4 3 Muscle.... 0 0 4 8 5 4 6 5 3 Length in .001’s inchs: i) 270-274 275-279, 280-284) 285-289) 290-294 295-299 300-304 305-309 310-314 The average length of the thyroid pupae was .295 inch; that of the muscle pupae was .296. This difference, of course, is too small to be significant. 262 BEVERLY WAUGH KUNKEL The pupae were then removed from the sand and placed in Petri dishes with loose-fitting covers. After seven days all the living pupae had emerged. There seemed to be greater mortality among the thyroid fed flies than among the muscle-fed, for out of 35 muscle pupae there emerged 33 adults, while from the 40 thyroid pupae only 23 emerged. In egg mass No. 24 of L. caesar, the thyroid-fed larvae produced pupae having an average length of .315 inch, and the muscle-fed larvae produced pupae having a length of .314 inch. In egg mass No. 26 of L. caesar thyroid feeding was followed by a slightly decreased length of pupae in comparison with those fed on muscle. The respective average lengths were .293 and .297 inch (table 5). TABLE 5 Showing the effect on the length of the pupae of Lucilia caesar of feeding the larvae on thyroid. The first line indicates the distribution of the thyroid-fed pupae according to length, the second line indicates that of the muscle-fed aN POI peeks cee ere ae: 1 5 6 ih 4 1 1 IMiisclesJe)s) ete seat. e 0 1 3 2 3 4 0 Length in .001’s inch..... 280-284| 285-289) 290-294) 295-299) 300-304|305-309|310-314 TID AAD SS, AMEE i ae lt RRR Pa Pega pC Ml a lite od J Wt e i a Egg mass No. 16 of L. caesar shows by far the greatest differ- ence in length between the thyroid- and muscle-fed pupae. The distribution of the pupae is shown in table 6. The average length of the muscle-fed pupae was .350 inch, and that of the thyroid-fed was .308 inch. TABLE 6 Showing the effect on the length of the pupae of Lucilia caesar of feeding the larvae on thyroid. The first line indicates the distribution according to length of the thyroid-fed pupae, the second indicates that of the muscle-fed Thyroid...,.-..-'- SON TON salon ele) 15, ) etd | A eS Ae On OF | Ora) ele rnOn mel Muscle... ::... 0; O| G OF O01) 0 EP Dpeteege | Qn ga teas eT ON ai a Length in 270—|275—|280-|285-|290-|295—|300-|305-|310-|315-|320-|325-|330-|335—|340—|345-/350-|355—/360- 0.061’s inch. | |274 |279 |284 |289 |294 |299 ie 309 |314 |319 |324 |329 |334 |339 |344 |349 [354 359 364 EFFECTS OF DUCTLESS GLANDS ON FLIES 263 In the case of egg mass No. 22 of L. sericata the food material was raised to the boiling-point before the eggs were placed on it. The average length of the thyroid pupae was .291 inch; that of the muscle pupae, .303 inch. The distribution of the pupae is Shown in table 7. TABLE 7 Showing the effect on the length of the pupae of Lucilia sericata of feeding the larvae on thyroid. The first line indicates the distribution according to length of the thyroid-fed pupae, the second indicates that of the muscle-fed Ui 2010 ER eee le Serovar Losi aaa ale dal oR. Ol. ININIS ClO a2 eke 3 n= OF AOR OncOn) 4) 10) Sorrel Gr Siete eet) ob Length in .001’s_|265-|270—275-|280-|285— 290—|295-|300-|305-|310-/315-|320-|325- WAGE ess oye a |269 274 |279 |284 |289 |294 |299 |304 |309 |314 |319 |324 |329 In a number of experiment the effects of thyroid and thymus feeding were compared. Egg masses No. 24 and No. 26 show that thymus probably has an effect upon the size of the pupae. In one lot in which there were 25 muscle pupae and 52 thymus pupae the average respective lengths were .314 and .322 inch (table 8). TABLE 8 Showing the effect on the length of the pupae of Lucilia caesar of feeding the larvae on thymus. The first line indicates the distribution according to length of the thymus- fed pupae, the second line indicates that of the muscle-fed Db yas. sae eet ek eee OPO FL ANS D OOM Gx eG | csl 6 MPa le ol es en eee AST LY AS ORES. Sere ie Feb OUD Length in .001’s inch........ 290-—|295-|300-|305- 3 10-|315-|320-|325-330-—335-|340— 294 |299 |304 [309 |314 |319 |324 |329 |334 /339 |344 In another series, however, in which there were 15 muscle pupae and 13 thymus pupae, the respective lengths were .297 and .298 inch (table 9). TABLE 9 Showing the effect on the length of the pupae of Lucilia caesar of feeding the larvae on thymus The first line indicates the distribution of the thymus-fed pupae accord- ing to length, the second line indicates that of the muscle-fed pupae AAW UIST ce een ore 2 1 + 3 3 1 1 AVITISC] Ge Rees tet ton cian ans 0 1 3 D, 3 4 0 Length in .001’s inch.... 280-284 | 285-289 290-294 295-299) 300-304 305-309 310-314 264 BEVERLY WAUGH KUNKEL In summing up the results of these experiments, it may be said that they agree essentially with those already determined for vertebrates, but are far less striking. Feeding larvae of Lucilia exclusively upon mammalian thyroid tends to retard slightly the growth of the larvae and consequently to reduce the size of the resulting pupae, while thymus tends to increase their size. Thyroid feeding tends to hasten the onset of pupation and to shorten the period of pupation. BIBLIOGRAPHY BupinerTon, R. A., AND Harvey, H. F. 1915 Division rate in ciliate Protozoa as influenced by thyroid constituents. Biol. Bull., 28, 304-314. DewitT7, J. 1901 Verhinderung der Verpuppung bei Insektenlarven. Roux’s Archiv, Bd. 11, 690. GupDERNATSCH, J. F. 1912 Feeding experiments on tadpoles. I. The influence of specific organs given as food on growth and differentiation. Roux’s Archiy., Bd. 35, 457-483. 1914 Feeding experiments on tadpoles. II. A further contribution to the knowledge of organs with an internal secretion. Am. Journ. Anat., o.. 14, 481-478. 1915 Feeding experiments on rats. III. A further contribution to the knowledge of organs with an internal secretion. Am. Jour. Physiol., vol. 36, 370-379. Norturop, J. H. 1917 The réle of yeast in the nutrition of an insect (Droso- phila). Jour. Biol. Chem., vol. 30, 181-187. Nowrkorr, M. 1908 Die Wirkung des Schilddriisenextrakts und einiger an- deren organischen Stoffe auf Ciliaten. Arch. f. Protist., Bd. 11. Suumway, W. 1914 Effect of thyroid on division rate of Paramecium. Jour. Expt. Zool., vol. 17, 297-311. 1917 The effects of thyroid diet upon Paramecium. Jour. Expt. Zool., vol. 22, 529-563. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 1 IS THE THEORY OF AXIAL GRADIENT IN THE REGENERATION OF TUBULARIA SUPPORTED BY FACTS? MARIO GARCIA-BANUS From the Laboratories of the Rockefeller Institute for Medical Research In his book ‘‘ Individuality in Organisms,” C. M. Child assumes the existence of ‘metabolic gradients’ in a great number of species of animals and plants, and on this assumption he builds a theory of individuality. In the case of the hydroid Tubularia, which he uses extensively to prove his theory, the metabolic gradient would lie in the axis of the animal. The apical end of the metabolic gradient of the major axis is the apical end of the hydranth and from there the rate (metabolic rate) decreases basally through the hydranth. In the stem the metabolic rate is lower than in the hydranth and there is a slight decrease in the basal direction, but at the growing tip of the stolon there is a short gradient in the opposite direction.1 Child has made no measurements of the rate of metabolism of different regions of the stem of tubularia. What he really means is that an excised piece of the stem of Tubularia regenerates a new hydranth at the oral end the more rapidly, the nearer this end lies to the original apex of the stem. Such differences he assumes to be due to alleged differences in what he calls ‘meta- bolic rates.’ We are therefore only concerned with the question whether the regional differences in the rate of regeneration which Child assumes in Tubularia really exist. If a piece is cut from a stem of Tubularia a new hydranth will regenerate at.each end, and, according to the old experiments of Loeb, the oral end of the piece will, as a rule, regenerate the 1 Child, C. M., Individuality in organisms. Chicago, 1915. 265 266 MARIO GARCIA-BANUS hydranth sooner than the aboralend. Child in his experiments? cut the stems of Tubularia in two or three pieces of equal length, and states that in such cases the oral end of the most apical piece will regenerate a hydranth sooner than the oral end of the most basal piece. This is the actual basis for his theory of ‘axial gradient.’ The differences found by Child between the times of emergence of the oral hydranths of the two pieces are, however, so slight that the suspicion arises that they may be merely within the limits of error and individual variation. As a matter of fact, only in one single case of those that he presents, is there a marked difference between the time of the regeneration of the oral hydranths of the apical and basal pieces. And in this case the average is based on the observation of only eight individual stems. The apical pieces give an average time of emergence of the oral hydranths equal to 27 hours, while for the basal pieces the average is 36.5 hours. In all the other cases which he gives, the differences are far from being so marked. Thus in one case the averages for the hydranth formation of eight stems were 38 hours for the apical pieces and 39 hours for the basal pieces, while in a third case the averages were 42 for the apical pieces and the same number 42 for the basal ones! In the case of the stem cut into three pieces of equal length the averages for the oral hydranth formation are for one of the series of twenty stems, 42.5 for the most apical piece and 44 for the most basal one, and in another series of ten stems the averages are 117 for the apical piece, 118 for the middle piece, and 120 for the most basal one. The differences in all these experiments, except the first, are so small that they may well lie within the limits of variation and of error of such experiments. In order to be sure we repeated Child’s experiments to try to ascertain the foundation for his conclusions. 2 Child, C. M., An analysis of form-regulation in Tubularia. IV. Regional and polar differences in the time of the hydranth formation as a special case of regulation in a complex system. Arch. f. Entwicklungsmech., 1907, 24, 1. THEORY OF AXIAL GRADIENTS 267 TABLE 1 Time in hours required for emergence of the oral hydrants of two pieces of equal length NUMBER OF INDIVIDUAL APICAL PIECE BASAL PIECE 1 75 71 2 81 73 3 71 77 4 71 71 5 7 77 6 71 71 7 95 71 8 95 79 9 81 71 10 71 71 11 71 73 12 71 — 13 79 73 14 83 95 15 71 73 16 95 71 17 77 95 18 75 79 19 71 83 20 71 73 VIET AR CRO Sorc 5. hy i. 1353s 77.6 hours 76.1 hours The experiments were carried out at New York (with material from Long Island Sound and from the Lower Bay) and at Woods Hole, and the results were similar in both places. The material and conditions, especially at Woods Hole, were so excellent that we were able to keep the regenerating stems alive in finger bowls for more than fifteen days. In each case we recorded the time of emergence of the hydranth outside the perisare, and observations were made at least every two hours. In the first group of experiments (table 1), the stems were cut into two pieces of equal length; several series of this type, with twenty stems each were made with similar results, namely, that there 1s no marked difference between the times of appearance of the oral hydranths of the two pieces. Very often the oral hy- dranth of the basal piece is the first to appear. As an example of 268 MARIO GARCIA-BANUS this may serve the experiment recorded in table 1. The size of both pieces was 20 mm., so that they were long enough to show a marked difference according to Child’s opinion. Of the twenty stems included in this experiment, in eight of them the oral hydranth of the apical piece appeared first, but there were seven cases in which the oral hydranth of the basal piece was the first to appear and four cases in which they both appeared at the same time. The averages of the time of appearance are 77.6 hours for the oral hydranths of the apical pieces and 76.1 hours for the oral hydranths of the basal pieces. The difference is even slightly in favor of the oral hydranth of the basal piece, but it is so small that it is within the limits of error or of individual variation. In a second group of experiments the stems were cut into three pieces of equal length. The pieces in this case were each 10 mm. long and the results are similar to those of the other series. The experiment in table 2 may serve as an example. There are in this case ten stems in which the oral hydranth of the third and most basal pieces emerges before the corresponding hydranths of the most apical pieces, while there are only five stems in which the oral hydranth of the most apical piece is the first to appear and there are two cases in which both appear at the same time. If we compare the apical piece with the middle piece, the results are the same; in seven cases out of the twenty, the oral hydranth of the middle piece is the first to appear, while there are only two cases in which the contrary happens, and six cases in which both appear at the same time. The averages in this experiment are: 78.6 hours for the oral hydranths of the apical piece, 77.8 for those of the middle piece, and 74.2 for those of the basal pieces. This shows again that the slight difference in the time of appearance turns out to be in the opposite sense of what we ought to expect from Child’s theories. We have made several series of this type with similar results, and in some cases the differences in favor of the basal pieces were even greater than in the case just shown. For instance, in one of these experiments in which twenty stems were cut into three pieces of 10 mm. each, the averages for the time of emergence of THEORY OF AXIAL GRADIENTS 269 TABLE 2 Time in hours required for the emergence of the oral hydranths of three pieces of equal length, from the same stem of Tubularia Length of the pieces = 10 mm. NUMBER OF INDIVIDUAL APICAL PIECE MIDDLE PIECE BASAL PIECE 1 91 69 67 2 77 71 69 3) 71 77 “133 4 69 91 ral 5 67 69 67 6 69 66 67 7 71 71 73 8 73 73 71 9 ~- _ 79 10 79 69 73 11 75 75 77 12 79 73 73 1183 73 73 75 14 — 91 91 15 91 91 79 16 — 95 71 17 — 91 69 18 91 73 73 19 91 91 75 20 91 69 91 ASVETARESE Metis ative 78.6 hours 77.8 hours 74.2 hours the oral hydranths were in hours: for the apical piece 92, for the middle piece 84, and for the basal piece 76. We have chosen the experiment, the results of which are given in table 2, as an example, because in it the individual differences are not very great, so that the data are more constant and the averages more reliable. It is therefore a mere matter of chance whether the oral hydranths of the apical or basal pieces emerge first, as long as both pieces have equal length. Comparing pieces of different size from the same stem, Child found that the shorter pieces will form hydranths a little later than the longer ones, but he states that since the more basal pieces will form the oral hydranths later than the more apical, the pieces must be cut in such a way that the shorter piece is always the more apical one. Thus he says the two factors of size and level are opposed to each other instead of acting in the same 270 MARIO GARCIA-BANUS TABLE 3 Time in hours required for emergence of the oral hydranths in pieces of different length Ratio of the length of the pieces = 1 : 2. Series A Apical piece = 10mm. Basal piece = 20 mm. Series B_ Apical piece = 20 mm. Basal piece = 10 mm. SERIES A SERIES B Number of Apical piece Basal piece Number of Apical piece Basal piece individual 10 mm. 20 mm. individual 20 mm. 10 mm. 1 67 67 1 67 67 2 53 53 2 67 67 3 55 55 3 53 69 4 55 67 4 55 69 5 55 55 5 67 97 6 53 55 6 — 67 7 67 67 7 79 97 8 — 67 8 91 67 9 67 67 9 91 al 10 73 67 10 67 67 11 75 67 11 59 71 12 67 55 12 91 91 13 55 67 13 59 67 14 71 67 14 67 67 15 69 67 15 67 67 16 91 67 16 67 67 17 95 73 Ve 67 67 18 53 66 18 77 67 19 67 67 19 67 67 20 55 67 20 53 67 Averages....... 65.4 64.2 69.0 70.3 Witterenceseuee. coe oe eee 1.2 hours 1.3 hours sense and adding their effects, as would be the case if the shorter piece were basal to the larger one. We have found the existence of this factor of the size of the pieces to act constantly but irregularly, and we have tested the other factor of level comparing a series of stems cut in such a way that the shorter piece is the more apical with another series of stems, under the same conditions and cut into the same proportions, but in which the shorter piece is the more basal one. The results are given in tables 3 to 5. In the experiment recorded in table 3 the ratio of the lengths of the pieces was | : 2. THEORY OF AXIAL GRADIENTS Dit TABLE 4 Time in hours required for emergence of the oral hydranths in pieces of different length Ratio of the length of the pieces = 1: 3. Series A Apical piece = 10mm. Basal piece = 30 mm. Series B_ Apical piece = 30mm. Basal piece = 10 mm. SERIES A SERIES B Number of Apical piece Basal piece Number of Apical piece Basal piece individual 10 mm. 30 mm. individual 30 mm. 10 mm. 1 99 73 1 79 79 2 69 69 2 67 71 3 69 67 3 al 73 4 67 67 4 77 77 5 69 69 5 71 69 6 67 67 6 71 69 Gi 67 67 a 79 91 8 67 67 8 91 91 9 69 69 9 91 91 10 73 71 10 67 lid, 11 69 69 11 69 73 12 67 67 12 tide 77 13 91 Wiad 13 67 71 14 71 a 14 67 71 15 tk 71 15 di 77 16 69 69 16 69 al ilzi 77 67 17 73 73 18 67 67 18 71 91 19 67 67 19 69 71 20 69 71 20 Cpl 79 Averages....... 72.0 68.8 73.9 Tied BReRe NCES oii s.gee! sme eae 3.2 hours 3.2 hours We took two series of twenty stems each; in series A the shorter piece of 10 mm. was the apical one and the longer of 20 mm. was the basal one; while in series B, on the contrary, the shorter _ piece was the basal one and the longer the apical one. The averages are in series A: for the apical piece (10 mm.) 65.4 hours, and for the basal piece (20 mm.) 64.2 hours; in series B: for the apical piece (20 mm.) 69.0, and for the basal piece (10 mm.) 70.3 hours. The oral hydranth of the larger piece emerges about an hour earlier than that of the smaller pieces, a difference that bo ~J i) MARIO GARCIA-BANUS TABLE 5 Time in hours required for the emergence of the oral hydranths in pieces of different length . Ratio of the length of the pieces = 1 : 4. Series A Apical piece = 10 mm. Basal piece = 40 mm. Series B Apical piece = 40 mm. Basal piece = 10 mm. SERIES A SERIES B Number of Apical piece Basal piece Number of Apical piece Basal piece individual 10 mm. 40 mm. iudividual 40 mm. 10mm. 1 72 72 1 70 94 2 70 70 2 76 82 3 94 74 3 74 82 4 72 hie, 4 74 70 5 74 70 5 82 78 6 94 94 6 94 74 a 74 78 7 74 82 8 98 74 8 94 74 9 118 70 9 72 74 10 70 70 10 — — 11 82 96 11 72 82 12 64 54 12 = — 13 98 94 13 70 70 14 82 106 14 70 74 15 70 82 15 72 82 16 94 96 16 70 94 il? — 74 Ne 70 78 18 102 72 18 76 = 19 94 94 19 72 82 20 70 70 20 70 94 Averages....... 83.8 79.1 75.1 80.0 Die nENCES!. “Acne ot ous ke one: 4.7 hours 4.9 hours is very small indeed but the same in both series. There was no influence of the level factor. In the experiment recorded in table 4, the ratio of the lengths of the pieces was 1:3. In series A the shorter piece of 10 mm. is the apical one and the longer piece of 30 mm. is the basal one, while in series B the reverse occurs. The averages are in series A: for the short apical piece 72.0 hours, and for the longer basal piece 68.8 hours; in series B: for the longer apical piece 73.9 and for the shorter basal one 77.1. The differences between the THEORY OF AXIAL GRADIENTS 273 time of regeneration of the shorter and the longer pieces are 3.2 hours in favor of the larger piece in both series. In table 5 the ratio of the sizes of the pieces was 1 : 4 and the averages are in series A: for the shorter apical pieces 83.8 hours, and for the longer basal one 79.1 hours, with a difference of 4.7, while in series B the averages are: for the longer apical piece 75.1 hours, and for the shorter basal piece 80.0, with a difference of 4.9. In any of these cases there was no evidence of the existence of level or regional differences on the stems of Tubularia. CONCLUSIONS Child has based his theory of the ‘metabolic gradients’ on the assertion that if a stem of Tubularia is cut into two pieces the oral end of the apical piece will regenerate a hydranth more rap- idly than the oral end of the basal piece. The differences in the time of regeneration observed by him were so small that they seemed to lie within the limits of error of such experiments. The writer repeated Child’s experiments and found this sus- picion justified. The rate of regeneration of the oral hydranth of an apical piece is on the average identical with the rate of regeneration of the oral hydranth of the basal piece. If in one series the average is in favor of the apical piece, in another the reverse may be found. There is no evidence of the existence of level or regional differ- ences of the rate of regeneration in the stem of Tubularia, and as a consequence there is no basis for the theory of ‘axial gradient’ in this species. I have to acknowledge my thanks to Dr. Jacques Loeb, who has suggested and directed the present work. Bais RARER OAR, iH6% a a 2) shirt penny isa) He Hie aihondacli Sala Debden Liste ahead heii er .

. 312 1. The nature of the contraction of melanophore pigment.......... 312 2. Résumé of the reactions of the melanophores of the horned toad. 314 3. Comparative physiology of the melanophores of vertebrates..... 316 a. The influence of hormones upon melanophores............. 316 b. The coérdination of melanophores and smooth muscle...... 317 ce. The physiological basis of the emotions:................ rae OLY VATED PS UTED TE Vig Ae tA Sire SR ale OR es en tne Bch AE UR aah Ue St me oa 319 VilliieS bitenagurer cited). 44 209-6 oc Seco at Re ares SEL OR a ee eB TEMP AGES ee ts. Sts sais Bite ale ESS RTECS tn atic tei rele) SRR EM Not gue 323 I. INTRODUCTION The results of an investigation of the reactions, codrdination, and function of the melanophores of the horned toad Phyrnosoma cornutum Harlan are described in the following paper. The facts which are presented possess, in addition to their intrinsic interest, a bearing upon several aspects of comparative physiology. Earlier studies upon the melanophores of lizards indicated that in the chameleon (Briicke, ’52; Keller, 95) the color changes are produced under the influence of nerves which cause the melano- phore pigment to contract, while in Anolis (Carlton, ’03) the nerve impulse appears to cause the opposite effect, an expansion of the melanophore pigment. The present investigation was initiated in the hope of gaining more light upon the relation of the nervous system to the melanophores. It very soon became evident that the melanophores are not only under the control of nerves, but that certain of their reactions result from the direct effect of stimuli upon them, while other reactions indicate clearly that some codrdinating mechanism other than the nervous system is involved. The discovery that these reactions are due to a hormone (Redfield, ’16) and the identification of the hormone as adrenin form the most novel result of the investigation. A few observations have already been made upon the color changes of this lizard by de Grijs (’99) and upon the closely related species Phrynosoma orbiculare, by Wiedersheim (Hoffmann, 90, p. 1353), and Phrynosoma modestum, by Weese (’17). Parker (’06) has made a more intensive study of the reactions of the melanophores of Phrynosoma blainvillei to illumination MELANOPHORES OF THE HORNED TOAD ZA and temperature. The great mass of data upon the chromato- phores of other lizards has been reviewed recently by von Ryn- berck (06) and by Fuchs (’14). Horned toads have been obtained from collectors in various parts of Texas and Oklahoma. They may be had in large num- bers between April and September. In the laboratory they were kept in a large sunlit cage, where they thrived on a diet of meal worms. In summer the animals were kept in cages out of doors and fed on various insects in season. Horned toads do not feed if kept in the laboratory through the winter. They become greatly emaciated and usually die before spring If these lizards are placed in a dark, cool cellar, they go into hiberna- tion and remain in very good condition all winter. Experiments have been carried on in the Zoélogical laboratory of Harvard College at Cambridge, Massachusetts, in the physi- ological laboratory of the Harvard Medical School in Boston, and in the laboratory of the United States Bureau of Fisheries at Woods Hole. I wish to acknowledge my indebtedness to Dr. G. H. Parker for his kindness in directing this investigation and to Dr. Walter B. Cannon for valuable advice on certain phases of the work. Il. THE MELANOPHORE REACTIONS OF THE HORNED TOAD 1. Description of color changes Under certain conditions the ground color of the upper surface of the horned toad is fuscous.?. Across the back run three irregu- lar bands of fuscous-black, bordered posteriorly with a bright amber-yellow line, which stands out prominently in contrast to the dark ground-color. Similar fuscous-black bands, not bor- dered with yellow, extend across the legs. The flattened scales which extend in a row along the sides of the body are black at the base, white at the tip. Under other conditions, the ground- color becomes drab, drab-gray, or cinnamon-buff, frequently flecked with black. The amber-yellow lines across the back no * The color nomenclature of Ridgway (’12) is used throughout this description. 278 ALFRED C. REDFIELD longer stand out in contrast with the ground-color, but the fus- cous-black bands now become very prominent both on the back and legs. The lateral scales become white throughout, with the exception of two on each side which are located at the ends of the fuscous-black cross bands of the back. These scales rarely be- come entirely free from pigment. Plate 1 illustrates the extremes of these color changes.* The color changes of lizards have been shown by Briicke (52), Keller (95), and Carlton (’03) to be due to the migration of pig- ment contained in cells known as chromatophores, which are situated in the dermal layer of the skin. An examination of sections of the skin of the horned toad reveals the presence of chromatophores, filled with black pigment, such as Keller (95) has called melanophores. When the melanophore pigment expands the ground-color becomes fuscous; when it contracts this portion of the skin becomes drab. The fuscous-black bars across the back and legs contain such an abundance of non- motile pigment that their color does not change. In addition to the melanophores there exist in the skin pigment cells containing a yellow pigment. Whether this pigment is motile has not been determined, certainly it causes no prominent color changes. 2. Daily rhythm of color changes At night the melanophore pigment of the horned toad is con- tracted, giving the lizards a pale appearance. In the early morning this pigment expands and the skin becomes uniformly dark. During the heat of midday the melanophore pigment contracts again, but as evening approaches there is a second ex- pansion, followed finally by a contraction once more as night sets in. Thus the pigment is contracted and the animals pale at night and at midday; the pigment is expanded and the animals dark in the morning and afternoon. 3 Photographs fail to bring out fully the contrast between the dark and the pale condition of the skin. This is because the skin is tinged with yellow, a non-actinic color. Orthochromatic plates and a color screen have been used, but without complete success, in an attempt to overcome this difficulty. MELANOPHORES OF THE HORNED TOAD 279 3. Adaptive color changes Horned toads show striking color changes which depend upon the color of the environment in which they live. If they are placed in a pen which is lined with black cinders, the ground- color becomes so dark that the transverse bands are almost indis- tinguishable. When such animals are transferred to a pen lined with white sand, the ground-color becomes very pale in the course of a few days. The contrast between these pale animals and horned toads which have been left in the cinder-lined cage as controls is illustrated in plate 1. This change in color becomes noticeable within one day after the lizard has been placed in its new environment, and reaches a maximum within one week. Upon the adaptive state of the chromatophores is superimposed the daily rhythm of color change. Thus horned toads which are adapted to a dark background become paler at night and at mid- day, but never attain the extreme paleness of specimens adapted to a light background. Similarly, horned toads which are adapted to a light background become darker in the morning and afternoon, especially on cool days, but never attain the dark color of specimens which are adapted to a dark environment. 4. Color changes due to nervous excitement Whenever horned toads are thrown into a state of nervous excitement the melanophore pigment throughout the entire surface of the body is contracted and the animal assumes the maximal pale condition. This reaction is observed whenever one of these lizards is subjected to any treatment which causes it to struggle in an effort to free itself. Simply holding a horned toad upon its back, attempting to pry open its mouth, or pinching its tail may suffice to cause this reaction. A male horned toad, attempting to copulate with a female, has been observed to become very pale at a time when the pigment of the other ani- mals in the pen was expanded fully. The contraction of the pigment which accompanies nervous excitement occurs more rapidly than the two other types of reac- tion which have been described. The complete change in color 280 ALFRED C. REDFIELD may occur in as short a time as three minutes and never requires longer than ten minutes. This reaction is also a dominant one, occurring irrespective of the condition which the cycle of day and night and the color of the environment has imposed upon the melanophores. III. EXTERNAL STIMULI INVOLVED IN MELANOPHORE REACTIONS 1. Illumination and temperature The series of color changes which accompanies the cycle of day and night is dependent upon the accompanying changes in light and temperature. The rhythm of the color changes may be broken by a disturbance of the normal rhythm of the changes in light and temperature. Thus on cool days the midday con- traction of the melanophore pigment does not occur; on cloudy days or when the animals are buried in the sand the skin may remain in the pale condition throughout the entire day. If a horned toad is exposed to bright daylight or to sunlight which is not too warm, after a short time an expansion of the melanophore pigment occurs. If the animal is placed in the dark, the melanophore pigment is contracted. These changes may be followed most clearly by observing the appearance and disappearance of the dark areas upon the bases of the lateral scales. The rapidity with which these changes occur in either direction varies from ten minutes to one-half hour, depending upon the intensity of the illumination, the temperature, and the idiosynerasy of the individual. In order to determine whether the effect of light upon the melanophores is due to the heat developed by its action, the following experiment was performed. An apparatus was devised in which the temperature could be kept quite constant and to which light could be admitted at will. This apparatus consisted of a glass aquarium 30 cm. in diameter and of equal height. Into this was inserted a second glass vessel 15 em. in diameter. Between the walls of these two jars ice or water of any desired temperature could be introduced. By this means the air in the MELANOPHORES OF THE HORNED TOAD 281 inner jar could be kept within a degree or two of a given tem- perature. In the inner jar was placed a layer of sand, which served to keep it floating upright and gave a natural support for the animal. This jar was kept covered to maintain the temperature and humidity as constant as possible. When desired light was excluded from this apparatus by inverting over it a large paste- board box. November 26, 1913. Two horned toads, the melanophore pigment of which was expanded, were placed in the apparatus at 10.45 A.M. They were illuminated by the light from the overcast sky. Warm water was placed in the outer jar. After one hour (11.45 A.M.) the temperature of the air in the inner jar had risen to 40°C. and the melanophore pigment of the animals had contracted. One-half hour later (12.15 P.M.), the temperature of the air in the inner jar had fallen to 26°C. and the melanophore pigment had ez- panded slightly. Ice was added to the water. Nearly two hours later (2.00 P.M.), the temperature was found to be 16°C. and the pigment was fully expanded. The apparatus was then covered with a box to exclude the light. After one-half hour (2.30 P.M.) the melanophore pigment was still expanded. Temperature 17°C. The ice-water was replaced with warmer water. One hour later (8.40 P.M.) the temperature had risen to 25°C. and the pigment had contracted again. It remained in this condition as the temperature was raised to 36°C. at 4.20 P.M. Repeated experiments confirm the conclusions which may be drawn from these data. It makes no difference in which order the changes of temperature and illumination are arranged. The following summary indicates the condition of the melano- phore pigment under various conditions of temperature and illumination: 1. High temperatures produce a contraction of the melano- phore pigment, irrespective of illumination. 2. Low temperatures produce an expansion of the melanophore pigment, irrespective of illumination. 3. At intermediate temperatures (30° to 20°C.) the state of the melanophores is conditioned by the illumination, light causing an expansion and darkness a contraction of the pigment. 282 ALFRED C. REDFIELD Light and heat thus act in opposite ways and their effects need not be confused. The expansion of the melanophore pigment in the light cannot be due to heat, as this stimulus causes a con- traction of the pigment. In certain individuals light dominates over temperature in determining the state of the melanophores through a larger range of temperatures than that indicated above. Some horned toads maintain a contracted condition of the pigment in the dark at temperatures as low as 5°C. The pigment of others may expand upon illumination when confined at temperatures as high as on Gs These conclusions are in very good agreement with those which Parker (06) drew from experiments upon Phrynosoma blainvillei. By their means may be explained the rhythmic changes in the color of horned toads which are correlated with the cycle of day and night. The expansion of the pigment in the morning is due to the stimulation of light. The heat of midday causes a con- traction in spite of the light, but as the air cools in the afternoon the light effect again dominates and the pigment expands. When the light fails, at night, the pigment becomes contracted as a result. 2. The color of the substratum It has been pointed out on a preceding page that the color of the environment has a marked effect upon the condition of the melanophores. A dark substratum produces an expansion of the pigment in horned toads which live upon it; a light colored sub- stratum has the reverse effect. The color of the substratum must consequently be considered the stimulus which initiates the adaptive changes in the color of this animal. This stimulus must not be confused with the effect of illumina- tion upon the melanophore. The reaction which it produces is quite a different phenomenon, as the following considerations in- dicate. The results of the two forms of stimuli are diametrically opposed; light causes an expansion, a light background a contrac- tion of the pigment. The photo-receptors concerned with the two sources of stimulation are quite distinct, as will be shown in another place. MELANOPHORES OF THE HORNED TOAD 283 3. Mechanical stimuli Mechanical stimulation, such as tapping or pressing the skin, does not produce any effect upon the melanophores. Mechan- ical stimuli of a more violent nature cause profound changes in the condition of the melanophores. Such stimuli, however, do not act directly upon the pigment cells to which they are applied, but affect the pigment of the entire body by producing a state of nervous excitement. Such stimuli are better classed with noxious stimuli, to be considered later. 4. Faradic stimuli The application of a faradic stimulus to the surface of the skin of the horned toad fails to produce a change in the state of the melanophores. Apparently the dry horny layer of the skin forms a very poor conductor for the electric current. If the electrodes are applied to the under surface of the skin, the melanophore pigment of the region becomes completely contracted in the course of five minutes. Faradic stimuli, if sufficiently violent, may produce effects upon the melanophores of the entire body. Such stimuli are conveniently classed as noxious stimuli. 5. Noxious stimuli In this category may be placed a variety of stimuli which have a common effect upon the state of the nervous system of the horned toad. Noxious stimuli are those which may be considered harmful, painful, or otherwise unpleasant and produce in the horned toad a state of nervous excitement. Whenever such a nervous condition is established in this animal there ensues a complete contraction of the melanophore pigment. A descrip- tion of some of the conditions which produce nervous excitement has already been given on page 279. For experimental purposes the most convenient noxious stimulus is the electric current. By placing platinum-point electrodes in contact with the mucous surface of the mouth or cloaca of the animal and passing a faradic current between these points, a complete contraction of the pig- 284 ALFRED C. REDFIELD ment may be produced. The current used for this purpose was just strong enough to cause an unpleasant sensation when applied to the human tongue. Summary The responses of the melanophores of the horned toad to exter- nal stimuli are the following: 1. Light produces an expansion; its absence a contraction of the melanophore pigment. 2. High temperatures produce a contraction; low temperatures an expansion of the melanophore pigment. 3. Light and heat interact in such a way that the heat effect dominates at extremes of temperature, the light effect dominates at mean temperatures. 4. Light coming from a dark substratum produces an expan- sion of the pigment; light coming from a light substratum pro- duces a contraction of the pigment. 5. Mild mechanical stimuli do not affect the melanophores. 6. Mild faradiec stimuli cause a contraction of the melanophore pigment. 7. Noxious stimuli, such as violent mechanical or faradic stimuli, produce a contraction of the melanophore pigment. IV. RECEPTORS INVOLVED IN MELANOPHORE REACTIONS 1. Direct action of stimuli wpon melanophores It has been pointed out by Spaeth (’13) that the direct action of stimuli cannot be studied satisfactorily by experiments upon living animals, owing to the complications introduced by the presence of the nervous system, circulation, etc. He recommends the study of bits of tissue separated from the body of the animal. Unfortunately, the horned toad does not lend itself easily to this method of investigation. Pieces of skin placed in Ringer’s solution show no melanophore reactions, except that the pigment is contracted very soon. This contraction may be attributed to anemia (p. 292). MELANOPHORES OF THE HORNED TOAD 285 By means of carefully controlled experiments, however, it is possible to gain some idea of the significance of the direct action of stimuli upon the melanophores without removing them from the organism. If these pigment cells react in only those places on which stimuli act, a direct response of the melanophore may be suspected; if the pigment throughout the entire body is affected by a local stimulus, it may be concluded that the reaction is due to the intermediation of nerves, or hormones. The fol- lowing experiments are instructive from this point of view: November 1, 1913. Three horned toads, the pigment of which was contracted, were shielded locally from light by placing a small piece of modeling clay over two or three lateral scales. After an hour of exposure to sunlight the melanophore pigment of all the lateral scales of all the animals had expanded, with the exception of the shielded scales, the pigment of which remained contracted. Two horned toads, the melanophore pigment of which was ex- panded, were shielded in a similar fashion and exposed to sunlight. The melanophore pigment of all the scales remained expanded excepting those shaded by the clay, which contracted. From this experiment, which has been amply confirmed, it is evident that if a small part of the surface of a horned toad be shielded from light, the melanophore pigment of this part will remain contracted while that of the rest of the body expands, and if the shaded melanophore pigment be expanded, it will contract without reference to the illumination of the rest of the body. It might be objected that this effect is due to the tactile stimu- lation of the clay or to the exclusion of air from the surface of the skin. That this objection is invalid is indicated by the fact that if the clay is applied so as not to touch the lateral scales, but only the surface of the skin on both sides of it, the reaction follows in the same manner. Moreover, if the scales are painted with cel- loidin, which should provide a similar tactile stimulation and exclude the air from the skin, but which does not keep out the light, the pigment is able to expand as readily as on those scales which are not covered. The converse experiment consists in illuminating a few scales of the body when the remaining parts are covered. This experi- 286 ALFRED C. REDFIELD ment has been performed in various ways and has yielded variable results. In a few cases the outcome has been in good agreement with the expectation raised by the experiment just described. August 24, 1915. A horned toad, having the melanophore pigment expanded, was tied to a board and covered with a piece of black woolen cloth. Through a hole in this cloth a portion of the right half of the back was exposed to the hght. 2.45 P.M. Placed in sunlight. 3.00 P.M. Exposed part of back unchanged. Melanophore pigment of unexposed part of skin is contracted. 3.15 P.M. Body completely covered with black woolen cloth. 3.30 P.M. Melanophore pigment of entire skin contracted. August 24, 1915. A horned toad, having its melanophore pigment completely contracted, was tied to a board and covered with black woolen cloth. Through a hole in the cloth 1 sq. em. of the anterior part of the left side of back was exposed to the light. 2.40 P.M. Placed in sunlight. 3.00 P.M. Melanophore pigment of exposed part of skin has expanded; unexposed skin is still pale. From these experiments it is again evident that the ilumina- tion of a restricted area of the skin will cause an expansion of the pigment of that area, without regard to the illumination of the rest of the body or to the reactions of melanophores of the other parts of the skin. Similar experiments have been performed in which heat was used instead of light as a local stimulus. The temperature of a few lateral scales could be modified by the following method. Against the side of the body of a horned toad a nozzle was applied through which ran a current of water of the desired temperature. The nozzle was made of a piece of glass tube 2 cm. in diameter. In one end was inserted a rubber stopper carrying an inlet and outlet tube and a thermometer. Over the other end a piece of sheet rubber was tied loosely. When pressed against the skin of a horned toad, the rubber end of the nozzle wrapped itself snugly about the scales and soon imparted to them a temperature which must have approximated that of the water within the nozzle. December 15, 1913. Water at a temperature of approximately 45°C. was passed through the nozzle applied to the right side of a horned toad the melanophore pigment of which was expanded. The experiment was performed in the dark so that the exclusion of light from the skin by the - MELANOPHORES OF THE HORNED TOAD 287 nozzle could not form a disturbing factor. After three minutes the right side was much paler than the left; after six minutes the melanophore pigment of the right side was completely contracted. The left side remained unchanged until the nozzle was applied to it, when it, too, became pale. This experiment, repeated on a number of animals, indicated that if a portion of the skin of a horned toad is heated while in the dark, the expanded melanophore pigment of that portion will contract much more rapidly than that of the other parts of the body. That this result is not produced by the contact of the apparatus, but is a true temperature effect, is shown by the fact that the application of the nozzle to the skin of a horned toad does not produce a more rapid contraction of the melanophore pigment of that part of the skin if the water within the nozzle is at room temperature. The converse experiment, the application of a low temperature to a small part of the skin, is illustrated in the following protocols: December 18, 1918. A nozzle through which ice water was running was applied to the side of a horned toad the melanophore pigment of which was contracted. Throughout thirty minutes of this treatment, during which the animal was kept in the dark, no change occurred in the condition of the pigment cells. In the same way a low temperature (10.5°C.) was applied to the side of a horned toad of which the melanophore pigment was expanded. The animal was in the dark. After six minutes the melanophore pig- ment of the animal had contracted, as might be expected in the dark. The melanophore pigment of the chilled portion of the skin did not con- tract in twice that time. From these experiments it appears that the local application of low temperatures to the skin of the horned toad cannot call forth an expansion of the melanophore pigment,‘ but that an expansion previously established may be maintained in the region of low temperature, though the rest of the skin becomes pale. 4 The action of a low temperature when applied locally to the melanophores is discordant with the results of applying light and heat and excluding light locally. It may be pointed out that the facts are no more readily explained by the assump- tion that nerves are concerned in the reaction, so that the point does not argue for or against the ability of the melanophores to respond directly to the stimula- tion of light and temperature. , 288 ALFRED C. REDFIELD To summarize the effect of local photic and thermal stimuli upon the melanophores of the horned toad, the following points should be brought out: 1. The local exclusion of light and the local application of heat, both of which produce a contraction of the melanophore pigment in the parts so treated, do not affect the melanophores in other parts of the body. 2. The local application of light, which produces an expansion of the exposed melanophore pigment, does not affect the melano- phores of other parts of the body. 3. The local application of a low temperature fails to produce any expansion of melanophore pigment. An expansion previously established may be maintained locally by this stimulus. Fig. A For explanation see text (p. 288). These experiments do not prove that the melanophores are stimulated directly by light and heat, because the same results might be obtained by a different mechanism. If the melano- phores were under the control of nerves so arranged that the stimulation of a point on the skin sets up impulses which passed over a reflex are terminating in the pigment cells of the same point and no other, then a local stimulus would also produce a local response. An experiment designed to test the presence of such a system of reflex arcs consisted in making cuts through the body wall of a number of horned toads in such positions that had all the cuts been made upon a single animal a part of the skin would have been completely severed from the rest of the body. ‘Thus one animal received a eut in a position indicated by the line a-—a, in figure A; a second, third and fourth animal in positions indicated MELANOPHORES OF THE HORNED TOAD 289 by the lines b-b, c—c, and d—d respectively. Whatever the course of any nerves connected with the melanophores within the region inclosed by these four lines, one of the cuts must be expected to interfere with the passage of nervous impulses to and from the region. The animals upon which these operations were performed recovered rapidly. Four days after the operation, when the melanophore pigment was fully expanded, they were placed in the dark. Within two hours the melanophore pigment of all parts of the skin in every one of the four cases had contracted. Since this contraction cannot be attributed to a nervous reflex, the con- clusion must be drawn that light acts directly upon the melano- phores of the horned toad.* In a similar way it was found that a high temperature would cause a contraction of the melanophore pigment of all parts of the skin of these animals. It is clear that a local response to local stimuli cannot be explained by postulating the action of hormones or of a diffuse nervous mechanism, such as anerve net. It has been shown that these responses cannot be attributed to a nicely arranged system of reflex ares. The conclusion seems unavoidable, therefore, that the action of light and temperature upon the melanophores of the horned toad is a local one, which does not involve the coérdinat- ing systems of the body. 2. Receptors of noxious stimuli Noxious stimuli can bring about melanophore reactions when- ever they affect sensory nerves with sufficient intensity. It has 5 The exact point upon which these stimuli act is still a subject for speculation. The simplest conception is that light and cold stimulate the melanophores directly and so produce their characteristic effects. If this be true, the melanophores may then be considered to serve as receptors for these stimuli. This conception, however, may be naive. Light and heat may act not directly upon the melano- phores, but upon some closely associated tissue and still bring about the observed results. Thus it will be shown (page 306) that the melanophores are acted upon directly by nervous impulses, which maintain a contraction of their pigment. Assume that cold acts upon the motor nerve endings which connect with the melanophores so as to block the passage of nerve impulses to the pigment cells, and the expansion of the melanophores by cold may be explained as an inhibition of the nervous control, rather than a direct positive effect of the stimulus upon the pigment cells. The present investigation has offered no grounds for choice between the two possible mechanisms. 290 ALFRED C. REDFIELD been found that an induced current of electricity will stimulate the terminations of sensory nerves which are found in the month and cloaca and thus bring about a contraction of the melano- phores. Stimulation of receptors in the ear are probably re- sponsible for setting up the state of nervous excitation which produces a contraction of the melanophores when the horned toad is held upon its back. One cannot say exactly what nerves are stimulated when an attempt is made to pry open the mouth of a horned toad, but this procedure also sets up a nervous dis- turbance which produces a contraction of the melanophore pig- ment. It seems clear that there are no specific receptors involved in the reaction; any receptor, the stimulation of which produces the necessary state of nervous excitement, is capable of initiating the reaction by which the melanophore pigment is contracted. 3. Receptors involved in adaptive reactions The reactions of the melanophores to the color of the substra- tum upon which the horned toads are kept depend upon stimuli received through the eyes. If horned toads which have been kept upon a bed of dark cinders are transferred to a bed of white sand, they become noticeably paler after one day and reach the maximum state of contraction of the melanophore pigment within five days. If such animals are blindfolded by fastening bits of woolen cloth over their eyes with celloidin and adhesive tape, they will retain the dark color, although they are kept upon a light colored background for several weeks. The failure of the melanophore pigment to contract is not due to any influence of the bandage, for animals which have had only one eye bandaged become pale as rapidly as specimens which have not been bandaged at all. Plate 2 shows a number of animals which illustrate such an experiment. In figure 3, a is a horned toad which has been kept upon cinders, 6 an animal which has been kept upon sand without blindfolding. The contrast in the pigmentation occurring under the two conditions is obvious. Figures 4 and 5 show four horned toads all of which have been upon sand for ten days: d and f, having had the use of their eyes, a MELANOPHORES OF THE HORNED TOAD 291 have become pale; c and e, having been blindfolded, remain as dark as the animals which have been kept upon cinders. Figure 6 shows a control experiment, in which both animals have been upon white sand for ten days; g was blindfolded in one eye, yet it has become as pale as h, which was not blindfolded. This experiment leaves no doubt that the eyes are the receptors involved in the adaptive reaction of the melanophores to the color of the environment. Summary 1. The melanophores or some closely associated tissues are receptors of photic and thermal stimuli. 2. There are no specific receptors for noxious stimuli. 3. The eyes are receptors for stimuli which cause adaptive reactions of the melanophores. V. COORDINATING MECHANISMS OF MELANOPHORE REACTIONS In the foregoing pages there have been considered the stimuli which initiate reactions of the melanophores of the horned toad and the receptors upon which these stimuli act. There remain to be considered the mechanisms involved in transmitting the effect of a stimulus to the pigment cells. By what means are the reac- tions of the melanophores coérdinated, so that a localized stimulus can cause a change in the color of the entire skin? Obviously this question need not be answered for those reactions in which the melanophores or some closely associated tissues are stimu- lated directly, as by light and heat. Noxious stimuli and stimuli arising from the color of the environment, however, may act on a restricted sensory area and yet cause a reaction of pigment cells over a large surface of the body. How is this integration accomplished? Two systems of the body are concerned in coordinating its parts. These are the circulation and the nervous system. The relation of these systems to the reaction of the melanophores of the horned toad will now be considered. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, No. 2 292 ALFRED C. REDFIELD 1. Coérdinative action of the circulation a. Influence of the respiratory function of the circulation upon melanophores. The circulation of the blood is obviously impor- tant in maintaining the metabolic activity of the melanophores. Stoppage of the circulation might be expected to produce effects upon the state of the melanophores, either on account of failure in the oxygen supply which must result or because of the accumu- lation of carbon dioxide or other katabolic products in the tissues. This expectation is readily realized. If a ligature is tied tightly about the leg of a horned toad, the melanophore pigment of which is expanded, no immediate result follows. In the course of two hours the skin of the ligatured leg becomes very much paler than that of the rest of the body (fig. 7). Evidently the stoppage of circulation has caused a con- traction of the pigment. Upon removing the ligature the circu- lation is restored, and after an hour and a half the melanophore pigment becomes expanded once more so that the color of the leg is indistinguishable from the other parts of the skin. If small bits of skin containing expanded melanophore pigment are cut from a horned toad and placed in a normal salt solution, this pigment becomes contracted very rapidly, usually within five or ten minutes. The skin of the horned toad also becomes pale within a short time after the death of the animal. The contraction of the melanophore pigment in these cases seems to be due to the interruption of the circulation. Whether the reaction is due to a deficiency of oxygen or to an accumulation of carbon dioxide or other waste materials in the tissue has not been determined. The effect of anemia upon the melanophores probably is not a thing which comes into play in producing any of the usual melanophore reactions. Consequently, the other mode of action of the circulation upon these cells will prove of greater interest. b. The coérdination of melanophores by hormones. The circu- lation of the blood is known to influence the tissues to which it flows by carrying to them chemical substances, such as hor- mones, which excite the tissues to various forms of activity. In MELANOPHORES OF THE HORNED TOAD 293 attempting to discover the mechanism through which a noxious stimulus, locally applied, is able to produce a general contraction of pigment of the entire body, it became apparent that the melano- phores are codrdinated by means of a hormone (Redfield, ’16). When the nerves supplying a region of the skin are cut, and the surface of the mouth or cloaca is stimulated with a weak faradic current, the melanophore pigment of the entire surface of the skin contracts. There is no difference between the behavior of the melanophores in the part which has been denervated and the rest of the skin. For example, the sciatic nerve may be cut close to the junction of one hind leg with the trunk. Upon applying faradic stimulation to the mouth, both hind legs become equally pale as the melanophore pigment of the entire skin contracts. Again, the series of horned toads operated upon in the manner described on page 288 may be subjected to a noxious stimulation. In none of these animals does the operation, which in some cases at least must have cut the nerves to the portion of skin under observation, interfere with the contraction of the pigment cells upon this region of the body. It is clear that the reaction may be completed without the aid of nerves which connect directly with the melanophores. A hormone, set free through the action of the noxious stimulus, is the natural alternative to account for this result. To test the hypothesis that the pigment is contracted by a hormone, carried to it by the circulation, the following experiment was tried: August 80,1915. Tied a ligature firmly about the right hind leg of a very dark horned toad. Movements of the lower part of this leg indi- cated that pressure of the ligature had not blocked the nervous impulses to the leg. 12.20 Started stimulating mouth with weak faradic current. 12.25 Color of skin of right leg unchanged. Skin of left leg has become much paler, in striking contrast to the right leg (fig. 8). Mel- anophore pigment of the other parts of the body has contracted. Stopped stimulation. 12.27 Removed ligature. 12.36. Right leg has become as pale as the left. 294 ALFRED C. REDFIELD From this experiment, which has been repeatedly confirmed, it is apparent that by blocking the circulation to one leg of the horned toad without interfering with the innervation of the leg noxious stimuli are prevented from influencing the melanophores of that leg. This result is to be expected if the melanophore pig- ment is contracted by a hormone liberated in the circulation by noxious stimuli. The fact that the melanophore pigment of the isolated leg becomes contracted upon the removal of the ligature (fig. 9), sev- eral minutes after the cessation of stimulation, confirms the hy- pothesis further. If noxious stimuli cause the production of a hormone in the blood which contracts the melanophore pigment, one might expect that blood drawn from a horned toad which had been thrown into a state of nervous excitement by this means would cause the contraction of the pigment of an unexcited animal when injected into its body. How far this expectation is realized is indicated by the following account of such an experiment: July 28,1915. 2.00 P.M. Needle of hypodermic syringe inserted into the lymph space under the skin of the right side of a horned toad the melanophore pigment of which was expanded fully. 2.52 P.M. Through this needle was injected 0.5 cc. of defibrinated blood drawn from the neck of a second horned toad, which had been made very pale by stimulating the mouth with a weak faradic current for ten minutes. 3.04 P.M. The skin of the first horned toad has become clearly pale along the right side of the back (fig. 10). The melanophore pig- ment of some lateral scales near the point of insertion of the needle has contracted. 3.18 P.M. No change. 3.28 P.M. Pale patch is becoming darker. 3.50 P.M. Pale patch is almost indistinguishable from the rest of skin. 7.00 P.M. Pale patch has entirely disappeared. This experiment indicates that the blood of an excited horned toad will cause a contraction of the melanophore pigment of an unexcited animal. The size and intensity of the pale area varied considerably in different experiments. In no case did it extend beyond the side of the body into which the blood was injected. MELANOPHORES OF THE HORNED TOAD 295 The foregoing experiment must be carefully controlled, for it is quite conceivable that blood from any horned toad might have an effect upon the melanophores of another individual, irrespective of the presence of hormones in it. That this sup- position is untenable is indicated by the following experiment: August 4, 1916. 12.30 P.M. Hypodermic needle inserted into the lymph space under the skin of the left side of a horned toad the pigment of which is expanded. 1.42 P.M. Injected through the needle 0.2 cc. of defibrinated blood drawn by quickly decapitating a second horned toad. This decapitated animal had had a portion of the thoracic spinal cord destroyed, an opera- tion which prevents any noxious stimulation from producing a con- traction of the melanophore pigment (see p. 300). In this way the presence of the suspected hormone is avoided. The first animal still had its pigment expanded. 1.53 to 3.00 P.M. No change in the color of the first horned toad. Since the exclusion of blood from a part of a horned toad during excitement causes the melanophore pigment of that part to remain expanded, and since the injection of blood from an excited horned toad into an unexcited individual causes the melanophore pigment of the latter to contract, it may.be concluded that the melano- phores of the herned toad are influenced by a hormone, present in the blood during nervous excitement, which causes the melano- phore pigment to contract. c. The nature of the hormone involved 1n melanophore reactions. What is the chemical substance which causes a contraction of the melanophore pigment of the horned toad and where is it produced? The conception of a hormone regulating the activity of melano- phores is not altogether novel. Fuchs (’14, pp. 1546 to 1547, 1651 to 1652) has attempted to explain the behavior of pigment cells in amphibian larvae and reptiles by assuming that sub- stances, perhaps internal secretions, which contract the pigment, are produced in the body under the regulation of the pineal organ. Laurens (’16) has recently shown this hypothesis to be inap- plicable to the phenomena observed by him in Ambylstoma punctatum. ' The pineal body of the horned toad (Phrynosoma douglassii and Phrynosoma coronata) has been described by Ritter (91), 296 ALFRED C. REDFIELD who finds it to be a highly developed organ, in part closely resembling an eye. The structure of the epiphysis, with its folded epithelial walls containing great quantities of blood cor- puscles, does not forbid the interpretation that it may function as a gland of internal secretion. It may be readily shown, how- ever, that it is not the organ concerned, primarily at least, with the production of the melanophore hormone. The skull of a horned toad may be trephined and the entire brain anterior to the cerebellum, and including the pineal organ, may be removed. The animals recover from this operation and their pigment be- comes fully expanded. Stimulation of the cloaca with a weak induced current is still able to produce a contraction of the melanophores. A more favorable clue to the identity of the melanophore hor- mone is found in the work of Cannon and his collaborators (Cannon, 715). Cannon and de la Paz (’11) have demonstrated that during states of emotional excitement there is an increase in the concentration of adrenin in blood drawn from the adrenal vein of the cat. A similar increase is produced by stimulation of sensory nerves and by asphyxia (Cannon and Hoskins, 711), conditions which will be shown to produce a contraction of the pigment of the horned toad. It is well known that adrenin occupies an especially significant place in the physiology of smooth muscle. Langley (’01), Elliott (’05), and others have shown that the action of adrenin upon this tissue in a great number of cases is identical with the effect of stimulating the sympathetic fibers innervating the smooth muscles of the body. Spaeth (16a) has concluded that melanophores ‘‘are to be considered function- ally modified smooth muscle cells.”” He has accumulated evi- dence which shows at least that there is a very close analogy between the two types of cell. ‘In view of these facts, it would not be surprising to find that adrenin is the hormone concerned in coérdinating the melano- phores of the horned toad. It has been pointed out by Coronae Moroni (’98), and by Lieben (06) that adrenin produces a contraction of the melanophore pig- MELANOPHORES OF THE HORNED TOAD 297 ment of the frog. Stockard (’15) and Spaeth (16a) report the same to be true of the melanophores of Fundulus. If a dilute solution of adrenin‘ in isotonic sodium chloride soiu- tion is injected under the skin of a horned toad, a complete con- traction of the pigment ensues; the dark bands across the back and legs stand out boldly against the ground-color, which becomes bright cinnamon-buff (fig. 12). This condition is produced by injections of 0.2 cc. of a solution of adrenin diluted to one part in 100,000 and occasionally by a solution of one part in 1,000,000. Weaker solutions, 1 : 10,000,000 and occasionally 1 : 100,000,000, produce a contraction of the melanophore pigment over a more circumscribed area, centering about the point at which the solu- tion is introduced (fig. 11). The latter effect is especially inter- esting because it is a close duplication of the condition produced by injecting blood from an excited horned toad into an unexcited individual. A comparison of figures 10 and 11 will make this point obvious. It may be concluded from these facts that adrenin produces an effect upon the melanophores of the horned toad identical with the action of the hormone which appears in the circulation of these lizards during rervous excitement.? The adrenal glands of the horned toad consists of two elon- gated, yellow bodies situated in the membranes which support 6 « Adrenalin-chloride’ of a concentration of 1 : 1000 in physiofogical salt solu- tion prepared by Parke-Davis and Company has been employed as a starting- point in making up the solutions used throughout these experiments. 7 Gudernatsch (’14) has reported that tadpoles of Rana temporaria ‘‘fed on (horse) adrenal cortex become much lighter than those fed on adrenal medulla or any other food.”’ This observation suggests that some secretion of the cortical part of the adrenal gland may be the melanophore hormone, instead of adrenin itself. This hypothesis was investigated in the following way. Extracts of the cortex and of the medulla of fresh beef adrenals were made. It was found that both extracts produced a contraction of the melanophores of Fundulus heteroclitus and of the horned toad. The extract of the medulla was effective at much greater dilutions than that of the cortex. Upon examining these extracts chemically it was found that both contained adrenin, but the extract of the medulla contained a much greater quantity than did the cortical extract. It seems justifiable, therefore, to conclude that adrenin is the substance concerned in the contraction of the melanophores, and not some constituent of the cortical part of the adrenal gland. 298 ALFRED C. REDFIELD the gonads. When they are fixed in a mixture of formol and Miiller’s fluid and sectioned, it is found that the peripheral portion is composed of alveolar masses of tissue which take on the yellow stain characteristic of the medullary substance of the mammalian adrenal gland. Extracts made by grinding a pair of these glands in 2 ec. of physiological salt solution cause an inhibi- tion of the tonic contractions of the intestine of the white rat (figure C). This reaction is characteristic of adrenin (figure B) Extracts made in a similar way from other tissues of the horned Fig. B Record of the effect of adrenin upon the rhythmic contractions of the intestine of the rat. At X adrenalin chloride 1 : 100,000 in Ringer’s solution was introduced. At Y Ringer’s solution replaced the adrenalin solution. Fig. C Record of the effect of an extract of the adrenal glands of the horned toad (introduced at X) upon the rhythmic contractions of the intestine of the rat. At Y Ringer’s solution replaced the adrenal extract. Fig. D Record of the effect of an extract of the epididymis of the horned toad (introduced at X) upon the rhythmic contractions of the intestine of the rat. At Y Ringer’s solution replaced the extract. toad, such as the liver (figure /#), skeletal muscle (figure F), and testis (figure G), do not produce this reaction. Extracts pre- pared from the epididymis do cause an inhibition of intestinal muscle (figure D), but this action is readily understood when it is considered that the adrenal glands lie in immediate juxtaposition MELANOPHORES OF THE HORNED TOAD 299 to the epididymis and that a small portion of adrenal tissue may readily have been incorporated in the extract. When extracts made from the adrenal glands of the horned toad are injected under the skin of living specimens of this lizard, a contraction of the melanophore pigment is produced. Extracts of liver, of skeletal muscle, and of testis do not cause any con- traction of the pigment. Extracts of the epididymis produce a contraction of the melanophore pigment of horned toads into which they are injected, but, as pointed out, these extracts have been shown to contain adrenin. Fig. E Record of the effect.of an extract of the liver of the horned toad (intro- duced at X) upon the rhythmic contractions of the intestine of the rat. At Y¥ Ringer’s solution replaced the extract. Fig. F Record of the effect of an extract of the skeletal muscle of the horned toad (introduced at X) upon the rhythmic contractions of the intestine of the rat. At Y Ringer’s solution replaced the extract. Fig. G Record of the effect of an extract of the testis of the horned toad (introduced at X) upon the rhythmic contractions of the intestine of the rat. At Y Ringer’s solution replaced the extract. From the foregoing it appears that the adrenal glands of the horned toad contain a substance, adrenin, which is capable of producing a contraction of the melanophore pigment of this animal. The question is thus raised: can the adrenal glands, 300 ALFRED C. REDFIELD during functional activity, produce enough of this hormone to cause a contraction of the melanophore pigment in the skin? A difficulty arises in attempting to discover the answer to this question. In order to stimulate the adrenal glands, the body cavity must be opened. If the animal is etherized before per- forming this operation, the melanophore pigment contracts and renders the continuance of the experiment futile. If eitheriza- tion were omitted, the operation of itself would constitute a noxious stimulus and produce a contraction of the pigment. Fortunately, it has been discovered that the destruction of the spinal cord between the eighth and thirteenth vertebra prevents the contraction of the melanophore pigment which normally follows noxious stimulation. After operating on horned toads in this way, it is possible to open the body cavity without causing the melanophore pigment to contract. The use of anaesthetics is rendered unnecessary because the destruction of the cord pre- vents the passage of nervous impulses from the posterior portion of the body to the brain. If the adrenal glands of such a preparation be stimulated directly by a weak faradic current, a complete contraction of the melanophore pigment will occur in a few minutes. That this result is not due to a nervous reflex is shown by the fact that if a ligature is tied about one of the hind legs before the stimulation is commenced, the leg will remain in its original dark condition after the rest of the body has become pale (fig. 138). Upon removing the ligature, the leg will become as pale as its mate, though stimu- lation has ceased several minutes before. If a ligature is tied about the blood-vessels which supply the adrenal gland, no change in the condition of the pigment cells can be detected upon stimulating the adrenal gland until after this hgature is removed. From these experiments it becomes quite clear that the adrenal glands of the horned toad can produce a hormone when they are brought into functional activity which contracts the melanophore pigment. The coloration produced in this way resembles in every respect that occurring during nervous excitement. : MELANOPHORES OF THE HORNED TOAD 301 If adrenin is the hormone responsible for the contraction of the melanophore pigment following noxious stimulation, it should be possible to recognize its presence in the circulation at such a time by its characteristic effects upon other bodily conditions. The action of adrenin upon the heart, blood pressure, alimentary mas- culature, and the iris do not form good criteria of the presence of adrenin in the circulation during emotional states, because in such a small animal as the horned it is very difficult to be sure the effects are not produced through the sympathetic nervous system. The presence of adrenin in the circulation, however, causes a change in its sugar content. It has been shown by Paton (’03), Bierry et Gatin-Gruzewska (’05), and others that glucose appears in abnormally large quantities in the blood stream after the injection of adrenin into it. Blum (’01) has discovered that adrenin injections will produce glycosuria, another manifestation of the same action of this hormone. Cannon, Shohl and Wright (11) found that when cats are excited for even so short a time as one-half hour they exhibit glycosuria. This emotional gly- cosuria is produced by the secretion of adrenin which accompanies emotional excitement. If adrenin occurs in abnormal quantities in the circulation of the horned toad during states of excitement, its presence should be indicated by an increase in the concen- tration of blood sugar. The sugar content of the blood of fifteen horned toads was determined by the method of Myers and Bailey (’16). The first series, representing the normal sugar content of the blood in five individuals, was made by drawing blood from animals which were unexcited and in which the melanophore pigment was fully ex- panded. The second series represents the sugar content of the blood of five horned toads which were thrown into a state of nervous excitement by stimulating the mouth with a weak induc- tion current for thirty minutes and then waiting thirty minutes before drawing the samples of blood. The third series represents the sugar content of the blood of five horned toads into which 0.5 ec. of a1 : 1000 solution of adrenin had been injected one hour before. Table 1 shows the result of this experiment. 302 ALFRED C. REDFIELD TABLE 1 Per cent of sugar in blood of horned toads NORMAL EXCITED INJECTED WITH ADRENIN 0.09 0.13 0.10 0.12 0.16 0.21 0.14 0.17 0.23 0.15 0.18 0.34 0.15 0.20 0.44 Mean 0.13 + 0.01 Mean 0.17 = 0.01 Mean 0.26 + 0.04 It is clear that the sugar content of the blood in those animals which had undergone nervous excitement is higher than in the normal animals, but not as high as in animals into which adrenin has been injected. The conclusion is warranted that the horned toad exhibits an emotional hyperglycemia, which indicates that adrenin is secreted into the blood stream during nervous excitement. Although a number of very sensitive tests for adrenin are known, no one, according to Stewart (12), has succeeded in demonstrating the presence of adrenin in the general circulation of mammals. In the hope that the adrenal secretion of the horned toad might be greater than that of mammals, an attempt has been made to detect adrenin in the blood of these animals during states of nervous excitement. The method devised by Cannon and de la Paz (’11) was finally selected for this purpose. This test depends upon the inhibition of the rhythmic contrac- tions of the longitudinal muscles of the intestine by minute quantities of adrenin. The method was successfully modified so that adrenin could be detected at dilutions of one part in ten million, and in quantities of blood so small as 0.7 cc. Although a number of positive tests for adrenin were obtained from blood from excited horned toads, the majority of the results were so variable that no weight can be placed upon the experiments. The difficulty which must attend the detection of adrenin in the circulation prevents this negative evidence from having any effect upon the hypothesis developed in this paper. If adrenin is produced in the circulation of the horned toad dur- ing nervous excitement and is responsible for the contraction of MELANOPHORES OF THE HORNED TOAD 303 the melanophore pigment at this time, one might expect to find the pigment contracted under other conditions which are known to activate the adrenal glands. Cannon and Hoskins (711) found that during asphyxiation the adrenal glands of the cat are activated. Itami (’12) suggests that the effect of carbon dioxide upon the vascular system is probably due in part to an increased secretion of adrenin. If horned toads are placed in a glass vessel through which carbon dioxide is passed, the skin becomes very pale in ten minutes. At this time the animals have become lethargic, but are still able to recover from the treatment. Etherization decreases the residual adrenin in the cat’s adrenal glands about 50 per cent, according to Elliott (12). The melano- phores of horned toads almost invariably exhibit a contraction of their pigment when the animals are etherized. Morphia has been shown by Elliott (12) to cause exhaustion of the adrenal glands. If 0.1 ec. of a 10 per cent solution of morphia oleate is injected under the skin of a horned toad, the pigment be- comes contracted completely within a few minutes. Nicotine has an effect upon the adrenal glands similar to that of morphia (Cannon, Aub and Binger, ’16). Injection of nicotine also produces a contraction of the melanophore pigment of the horned toad. This effect is due in part to the direct action of this drug upon the pigment cell, for if nicotine is injected into a leg about which a ligature is firmly tied, this leg alone will become pale. It is quite possible that the effects of asphyxiation, etherization, and morphia are also due to a direct action upon the pigment cells. At least it can be said that there is no contradiction between the observed behavior of the pigment cells and the behavior to be expected if they are affected by the secretion of the adrenal glands. The melanophore pigment of the horned toad is contracted during nervous excitement, asphyxia, ether anaesthesia, and poisoning by morphia and nicotine—conditions which are known to produce activity of the adrenal glands in mammals. 304 ALFRED C. REDFIELD If it could be shown that changes occurred in the histological condition of the adrenal glands of the horned toad or that these bodies become exhausted as the result of any stimulus which will cause a contraction of the pigment, additional evidence that the melanophores are coérdinated by the secretion of adrenin would be afforded. It has not been possible to produce any recognizable change in the condition of the glands either by prolonged noxious stimulation or by stimulating the glands directly with an electric current. The method employed, fixing with a mixture of formol and Miiller’s fluid, evidently was not suitable for detecting any changes which may occur in the adrenal glands of these lizards. Elliott (’12) was able to demonstrate only a very slight decrease in the residual adrenin after faradizing one of the splanchnic nerves of the cat. Although the preceding experiment failed to demonstrate any exhaustion of the adrenal glands, an observation has been made which is significant in this connection. When shipments of horned toads first arrive from Texas the skin of the animals is always very dark in color. If these lizards are subjected to a noxious stimulus, the melanophore pigment becomes contracted only slightly and much more slowly than normally. Some individuals fail to exhibit the reaction at all. Not until after a week does the reaction become normal. When it is remembered that these animals have been crowded together in a small box, and shaken about in an express car for at least five days, it 1s not surprising that any gland which is brought into activity by nervous excitement should become exhausted. The failure of the reaction in these animals may be explained readily in this way. The bearing of this observation upon the relation of the adrenal glands to the reaction of the melanophores to noxious stimuli will become evident when it is pointed out that Elliott (12) has found that the adrenin content of the adrenal glands of cats which have been recently brought into the laboratory is below that of animals which have become accustomed to their sur- roundings. If the contraction of the melanophore pigment of the horned toad which follows noxious stimulation is due to a secretion of the MELANOPHORES OF THE HORNED TOAD 305 adrenal glands, removal of these bodies might be expected to check the reaction. It is not difficult to open the body cavity of a horned toad, tie ligatures about the membranes supporting the adrenal glands, and cut out these bodies. It is necessary to take with them the gonads and a portion of the genital ducts. Although the operation involves tying off the postcava, other veins are able to compensate for the loss of this vessel, and the animals live for a week or more. It was stated in a preliminary communication (Redfield, ’16) that this operation does not check the contraction of the melanophores by noxious stimuli. This observation was based upon experiments upon only a few animals. A larger series has since been operated upon and two horned toads found which fulfilled the requirements of the hy- pothesis. Although these animals became very pale in less than four minutes when the mouth was stimulated by a weak faradic current of electricity, after the removal of the adrenal glands no contraction of the melanophore pigment resulted, although the mouth was stimulated in the same way for twenty minutes. ‘The conclusion is unavoidable that the hormone responsible for the original con- traction of the melanophore pigment, as the result of noxious stimuli, is produced by the adrenal glands. Turning to the larger number of animals in which removal of the adrenal glands proved ineffective in checking the reaction, it must be pointed out that in these horned toads the contraction of the melanophore pigment was somewhat slower and the skin never became as pale as before adrenalectomy. On page 308 an experiment will be described which proves conclusively that this incomplete reaction is due to the action of the nervous system upon the melanophores and shows that when the nervous system has been destroyed removal of the adrenal glands is necessary before the reaction can be blocked. In the foregoing pages evidence has been presented indicating that: 1) adrenin contracts the melanophore pigment of the horned toad; 2) the adrenal glands contain this substance; 3) stimulation o° these glands causes the melanophore pigment to contract; 4) following noxious stimulation, adrenin occurs in the circulation and the melanophore pigment contracts; 5) the 306 ALFRED C. REDFIELD melanophore pigment is contracted at other times when he presence of adrenin in the circulation is to be expected, and 6) the removal of the adrenal gland hinders or prevents the contrac- tion of the melanophore pigment by noxious stimuli. In view of these facts, it must be concluded that the hormone which causes the melanophore pigment of the horned toad to contract during states of nervous excitement is adrenin, the active principle of the adrenal glands. 2. Coérdinative action of the nervous system a. The nervous control of melanophores. Any experiments designed to demonstrate the innervation of pigment cells are invalid unless care is taken to eliminate the action of hormones upon the part of the body under examination. The production of adrenin in response to noxious stimuli presents a real difficulty in such an investigation because the operative procedure in- volved in cutting and stimulating nerves is quite sufficient to produce a contraction of all the melanophore pigment through the action of this hormone. This difficulty has been overcome by destroying the anterior part of the spinal cord of the horned toad to be experimented upon. When this is done the animals recover completely, but no longer respond to noxious stimuli by contracting the melanophore pigment. Apparently the operation destroys the ‘center’ or the tracts of efferent nerve fibers which activate the adrenal glands. The operation has the additional advantage of allowing one subsequently to operate on the posterior part of the body without the use of anaesthetics. The following experiment shows the effect of stimulating the spinal and sciatic nerves of horned toads prepared in this manner: August. 11, 1915. Etherized two horned toads and removed the spinal cord between approximately the eighth and the thirteenth vertebrae. August 12, 1915. Skin of right side of back of one of these animals was cut open and a spinal nerve (the twelfth?) dissected out and stim- ulated with a weak faradic current for ten minutes. Produced abso- lutely no change in the melanophores of any part of the skin. Stimulated the sciatic nerve of the left hind leg of the other animal for five minutes with a weak induction current. Produced a very MELANOPHORES OF THE HORNED TOAD 307 clear contraction of the melanophore pigment of the left leg, which brought out in sharp contrast the dark and light bands (fig. 14). These experiments have been repeated many times. In no case has stimulation of the spinal nerves produced any contrac- tion of the melanophore pigment. ‘The results of stimulating the sciatic nerve are extremely variable; often no contraction of the pigment is produced. In many individuals, however, stimulation of fibers which occur in the sciatic nerve produced a contraction of the melanophore pigment. Many attempts have been made to isolate various portions of the skin from the central nervous system by cutting through the nerves supplying these parts. The sciatic nerve has been cut in the thigh, groups of spinal nerves have been severed close to the spinal cord, and cuts have been made through the body wall so as to isolate completely portions of the skin in the manner described on page 288. In no case have these operations altered the reactions of the melanophores of the isolated regions in any way. The responses of the melanophores to direct stimulation and to hormones evidently suffice to bring about all ordinary melanophore reactions without the aid of nerves which connect with these cells directly. One operation of this type has yielded a pertinent observation. When the spinal cord of a horned toad is transected at about the level of the thirteenth vertebra, the melanophore reactions are ordinarily unaffected in every way, except that a noxious stimulus applied to the posterior part of the body no longer produces a contraction of the melanophore pigment. Under usual laboratory conditions, the skin of such animals remains in a uniform dark condition. Three cases have been found to which the foregoing statement does not apply. On the day following the operation these horned toads had recovered fully, but the condition of the melanophores was exceptional. On the anterior half of the back the melanophore pigment was completely contracted, giving this portion of the skin the maximum pale condition. Posterior to the point at which the cord had been sectioned, the melanophore pigment remained fully expanded. The skin of this portion was quite dark (fig. 15). THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 2 308 ALFRED C. REDFIELD Why the anterior portion of these animals became pale is not obvious. The effect cannot be due to a secretion of adrenin or any other hormone, nor to the direct action of any environmental factors, since these would affect all parts of the skin uniformly. The coincidence of the point of operation upon the spinal cord with the line of demarkation between the dark and pale areas of the skin point to a causal connection between the two. The conclusion is thus suggested that the posterior part of the body is - isolated by transection of the spinal cord from some nervous disturbance which is acting upon the melanophores of the anterior part of the body through nerves connecting directly with these cells. These observations, although exceptional, lend strong support to the conclusion drawn from the contraction of the melanophore pigment of the leg by stimulation of the sciatic nerve; the melanophores are acted upon directly by nerves, and their pigment is contracted by nervous impulses. That the preceding interpretation is correct is shown clearly by an experiment which demonstrates, in addition, the relation existing between the adrenal glands, the nervous system and the melanophores. It has been pointed out that cutting the nerves distributed to a part of the skin does not interfere with the con- traction of the melanophore pigment which follows noxious stimulation. The contraction which occurs at this time has been attributed to the action of the secretion of the adrenal glands upon the melanophores. If the adrenal glands are removed from horned toads which have had the nerves to a portion of the skin transected, the isolated region is no longer affected by noxious stimuli, although the melanophore pigment of the remainder of the skin is made to contract. By this procedure the effect de- scribed in the preceding experiment may be exactly duplicated (fig. 16). August 9, 1916. Stimulated the mouth of a horned toad for five minutes with a weak faradic current. A complete contraction of the melanophore pigment resulted. Etherized the animal and transected spinal cord at thirteenth vertebra. August 11, 1916. Melanophore pigment was expanded. Stimu- lated mouth as before for ten minutes. ] i Sree omoedio- %) © OD © oO MOMSOKROMHHOKRKRDKRHDHDHDRADKHDHDOKRON (o ) INAHAN WOMAN on) 6F8'E €8P £02 92 002 GZ €8T GZ PES 1G LIZ 1X6 GG 1G SLT &@ Z61 &@ PRG 1G GOT IG c61 6G SsI IZ OST GG PST 0G OLT IG 9ST 0G reveal LI aa LT 661 9T GOT oT O&T 8T LL al 9% g 1G ¥ Z I e"L | OPV ET a2 1 2 | 6292 9FE LOL@ | BBE 6 Wo SEP TZ 99 06 69 | FOr ST GL | 99T £6 ays TOL 66 929 Out IZ 08 =| 221 GG €L GLY 16 PL AeGeL 61 Vee le LOE GG 08 | 892 $6 9-0) Shr LT 06 | 8Iz ¥G 18 98 cOT G8 | Z8T GG G8 | 96 VL 969 ¥6 QL | O&T LT S50" COR &@ VL ST9 &8 82 | SII ei Sada Gua GG 62 69 08 § 2 | OI ST 9°8 | POT 61 68 -| SB cOl LiL -| 821 & C55 sesh GZ 4 16h OL Gro she c8 OI 9) 4) Tek 8ST 08 §29 8 0'8 | 9ET ZI G8 SSGZ0 1Z 6rZ cr 18 cL, Past ST 18 | 2st 1Z 9°2 €o9 G8 Sz | 291 0Z VL | 9ST 1G LL 16¢ 92 & 8 | OFT LI Oc8 9 5GSir 61 9°2 €19 08 SoZ) GET ST 08 | F9I 0G 9°L S8¢ LL 08 | ZéT LT 08 | I9T 02 8°9 SGP G9 9 =|, 92 al ya S| eal 1 a 4 OCP Lg Onze 208 8 9794/2001 ST TL | 007 9S Ze9) IL9 6 GL | 601 ST G9 | SPE g¢ LS ANSe9 IT ty eee iss ia 89 | OTP. 19 (che ieee fate! Or VL UTE ST 09 | && 6€ SN Si 9 G9) Aiea 8 0's 601 GG G¥ —F1¢ . VG EAG G TsO" S08: €1 97 |-¥I 6 9 — | 6D 6 GL 66 Vv 06 | 6 I 09: | 9 I Sune 40] -tpur jo | SMALE lungs yo] TENPL | SOT |'sunog yo| STERPLA | SH aaa eA SeaeloNe ied eo annals Seanrane cates eases reqerni Vi [ROL OSBIOAV OSBIBAV | AUVWNWOS pearqury[ey) 810}91] YANO 1a}}1[ 10d sunod jo Jequinu BBBIOAV gjenpia | s10}}I] a Ran IO 40 roquinyy |requinyy 9°9 | SZT 9% (5/4, = Weiss! S14 Vee |p 98it GS Sey |). rahlte KG VL | 661 1X6 keg -teto31! KG & 9 | ZI €% SED ESOL €% Lee RONG KG S59 NESE IZ VL TZT &%@ 9°9 | 6&T 1G G2 | OOF 6G Gt Neil 0G 38°9 | HI IZ ¢°9 II 0G Gage eSOT LT Lz al LT toe) IOL OT 69 | ¥6 CT C29) SAE 8T Vr | 8S €I og | SE L L°9 | 06 1 Olea: I 19}41] od gjenpta | 819931] Preece cxeuts roa lees agtaoay |@AUON |zequinn (perqur-ypey) $1049] PAY L, (peiqut) $10}41] puoveg salugs UALLIT sypl pasqur fo (perqut) 810991] 821 on = re mA N OO HID Or CO NOILY -HaNGD saisas p ay) fo suownsauab aay-hjyuany ysl ayy fo yova ur poonpoud s19q3y ay) fo azis abpsaan ayz pup soqunu ay} Burmoys Tt HTaV.L 339 AND VIGOR OF INBREEDING ON FERTILITY EFFECTS vL | gee‘et| 9co'T | 22 | Gee's | 20e | 82 | 922'e | GIF | BL | 909 | ZOW | 69 | GIZE | SOF | Se-T TZ G69 16 89 | 9ET 0Z Tez eO2T GZ i724 || t30! 96 OL | F8T 9% 6 OL G89 16 L°9 | VIL LT LO SLE 9% OL | 681 KG 92 206 1K6 ¥% 18 L¥9 62 FS | 921 CT GL CFT 0G 88 | S6I (a6 a8 18] GG &% 8 L PEL G6 SeZee|68L 61 88 11Z ¥G 0'8 | O12 9% LO BLE 9% GG SLE, C19 c8 OL GIL 9T 92 SéT ST 82 68T ¥G Sice | 92E ¥G IZ Teed 86¢ ¥8 08 | 88 II Fan ese IG $9) | 625 9G LG) VA 92 0Z v8 869 €8 PL 1-96 €I 8°8 | S6I GG G6 1GG FZ LL | 980 ¥G 61 Lek 899 98 PL | 6F1 0G _ G8 IZT 0G 82 | O8T &% (74 |) tei €% SI PL FI9 o8 9°9 rea a jail a8 | &ZI IG Lite OZ (a6 eZ esr (26 LT PL $99 68 6°9 | C&L 61 9°8 | 061 GG 92 | PST FG CEOs S8Sir FG 91 GL G8G LL 68 | OFT LT OFS) | -SZk 06 6°9 | S&I 0G cone Tél 06. cI Tez OLS O8 (Sv GET ST 0's I9T 0G ¥9 GT IG L°9 | Chl 1G FI Joede, 92S vL 08 | FOL €I 6°72 OST 61 08 | 891 IG S25 ari IZ €1 GL 9€¢ IZ 9°¢ | 89 al Tash erst 61 Lease eGZe 0G OE9h GEL 0G al TZ LCS vl 68 | 6F1 ST 7s |ltetell SI L OF Sor 61 GaSe GEE 61 IT OL OFF F9 9°9 | £6 ial! 99 | ZOL 9T ¢'8 -| StI LT 19 | FOr LT OL Gia CST L9 67 | 67 Or PL | 2Cr LT (0) ("IP tesa 0G 69 | 9er 0G 6 - 92 CGE OF ORZ IG € GL | 6 §I 9°8 | 6eI cT 972) | -80E oT 8 9°9 99€ Gg Siczy les OL 6°9 | POT ST OrZ- | SOL ST See sou SI Ht LL (OF GG (6 (eM 175 8 6/4 |) C40 FI S24. | SiD CT OL | 901 cI 9 GL 9€¢ TZ Sie e860 €1 Sasa (OST SI ee eSSih 0G COORReGGL 0G ¢ Gal 606 66 09 | 8I € Ors | 9S Li 98 | 69 8 09 | 99 IT ¥ OS TZ tI Chan |n0 G GIS 596 g eS ieos i € 18 L¢ L 06 | 6 I CRSi se |e4h G ¢6 | GF G 09 | GL G G 0¢ ¢ I ORGS ES I it 12}}1[ 1ed| sjenpra 194}, 13d 19}}1, 10d Joqyy 10d 10941] ea Banod jo) put jo | Toquanis [MES 20| “our yo | yo". [AERC 20| puro | yo. PUTO 29] puro | yo ALCOA JO) pur jo| 30 e¥eroay | [BIO], 1®30.L aBBIoAY IoquUINN |Jequinyy asusoay Joaquin |requin yy asRioay IoquUINN |1oquUIN yy aBBIOAW JequinN |requin yy AUVARAOS (perqur-jyeq) 810991] YWNoy parqury{ey) 819991] PANT L (peiqut) 810441] puovag SaIuNgds YUALLIT (perqut) 819341] ISI S]DL palquir NOILV “HANGS fo sarsas g ay} fo suornsauab aary-Ajyuany psa ayj-fo yaa ur paonposd 8.1ajq4 9Y} fo az2s abviaan ay} PUD saqunu ay} Buinoys o WTAVL SING HELEN DEAN 340 GL ZSP‘GZ| SOPs $l | POLP €S9 SL 6¢9'9 L198 62 | SSPr'Z CcP6 69 | $49'9 €96 | Sc-T CL ges‘T | L8T 6°9 | OFZ cé Coe eae SP 92 968 GS 8'9 LSE (6 Gc Say 98ET | O6T 19 | SG 8€ Gee (Gas SP GL | 68€ 6G GL 16€ GS VG Ley, 61ST | OLT 6°L | 896 vE €°2 | 90E (Gi OS e828 ye Ay. ea, Pe /Aghs LV &@ (a) ZOS‘T | O61 O02 | 6S¢ 9€ 6°8 | 6GF SP €8 | Vr €¢ OFZ 1245 g¢ GG owe IFT | PST 62 666 SE 68 T9E TP 08 | 90F Tg Sele S26 1g IZ (G4 760 1 | SAT Lake ABS 8G O24 | 908 TV 9°L | €0P &S 6°9 | L9E €¢ 0Z 62 SIe'T | 99T 9°2 | PIG 8G £83 LOE VV G8 | 666 LV a | atts LV 61 SL L08°'T | 99T WwL, 6G cE COSinlaGes: 6€ 68 LLE OF is 4) |) Slee OF 81 62 6SPh'T | P81 GL 166 OF VS | 98¢ OF VS | PIP 67 GL | 89€ 6P LT Cee; GST‘T | 691 09 | F6I 66 9L eit OF Li, OPS cP 19 10€ cP 9T Sid Goat | 191 T8 | 926 VE VS | OVE IP rer I€€ €P (A | ANS eV CT Ci GIZ‘T | T9T GL | 89G 9€ VS | SPE Th LL, GGG GP 19 18G GV al Me P 622 T | 6ST 62 196 €€ 9°L | 90€ OF 0S | 8PE €P (S Ae | Ale €P €T Oe VEL. \SLV1 GL | 806 6G L8 | 90€ 8é G8 | 66E OF TL |s¥8% OV al VL OFI'T | FST 62 | 86 9€ 0'8 | GOE 8€ VL | 866 OF v9 996 OF II Ge; FEO LT | Trl PL, 12086 165 VL | 896 9€ 18 LOE LE €°9 CEs LE OL” OL O16 661 (ee = | eal GG GL ) 8¥G €&€ €8 | 90€ EG 69 T&% LE 6 CL CLL €01 O02 LL 1 6°9 | P6T 8G G3 TLz GE Gud OSG (Gs 8 6°9 992 JEU Ge BOG 61 (Tez? €1Z 0€ Gye 2 \neG I€ 8°¢ | O8T its le 6°9 GPL LOT GL LET 61 19 | 881 8% €'L | 066 O€ 9°9 006 0€ 9 GL 6G6 GET 9°9 | ST €&% (a) 193 €€ 9°L | 88z Sé 9°9 | 0SZ 8é g ¢'9 GEV 89 TL | 09 6 GZ -| S0r CT Guz 12901 0% LG EGE VG v O'S OST 9€ GV IZ ¢ eG) SOS. ZL GG | 6 Or T¢ iW, iat € 8'9 LET 0G LAG IESG V (Ge Woks g Oey 89K 9 v9 | G ¢ G 8°9 VE ¢ 06 | 6 T 09 |9 I! OL yy T 09 | 8I 6 I song sod} sTeNPEs lesomnat solgatk go) stemprs | sion mde gol swmpes | ston lade go] siempre | exon Ggnas yo] SUMPES | SoH ae, Sse ar Baie ae eanea N See anny cone N oat ae Pe an rein N SRueAty geri ranean N a: 0 mtrcestcts) AUVWWOS (perqur-j[By) $10}}1] Yo y (perqur-jyBy) 10991] Pry L (peiqut) 810441] puodveg (perquy) 810991] 98I1Y SAIHaS YALLIT g 919QD) UL puy J 91QD} Ur DIDp ay} fO WOYDULQuod D :(g{‘V) sales pauqui on ay) fo suoynsoudb saryf-hzuanj jsf oy) fo yova ur poonpowd s.19qqr) oY} fo azis abpuean pun saqunu oy} Buimoyy € WIAV.L 341 VIGOR OF INBREEDING ON FERTILITY AND EFFECTS Gr 12OLL ell cout 692) e6c9'S 4) 10S Pe | CCGG le Sey. 6'2 | 6F8‘°E | &8F 6°9 | ¢cs’e | Ssh [Se-1 VL 610% | FLz One 98€ GG GL FOS 19 / ey 98¢ 92 eee, €F¢ 92 GC-ES Lee G06‘Z | PLE Ls Ly SPs TZ 0'8 692 C6 68 €S8 POL ee (Ga FOL |Zc-61 thee 8F9°% | 98 GL | 98P c9 0'8 TL9 €8 SoSk | 062 F6 €'L | 669 76 = |SI-4T Lil; GOS‘ | BSE SL 89S &Z 0'8 799 1s G8 | c69 t8 O02 | 88¢ 78 = =|PI-IT CL O&8‘T | 2&2 ess O€& OF Wee 167 99 tel OsS¢ OL G9 6SP OL OI-2 69 ZIZ‘T | $61 6S Iz 9€ G9 TOE OF L°9 69€ GG 8°¢ T&€ LS 9-T 194p1[ Jed} sjenpra s19}71T |104}1] Jed 10}}1] 10d Ie4}IT 10d 10}}1, 10d ae MUNA 70] APELIO| JO, (BENOK 19 ome o|. go ERE erouryo | go, a | tomo a ye clases eel eetaeio |e aa aReIoAy [210 ae ay see Pec eatNe OTe Barone Ze QU: SQREEN: OSBIOAV AOE ORANs ZOCOXTEN: OSBIOAY TO CLOATEN “ECL EEEEN, SNOILV -HaNaAD AUVWWOS perqui-s[eq) $10}41] YANO (porqur-j[eq) $10}41] PAT, (perqur) 810}}1[ puovsg SHINaS UALLIT (pearqur) 810941] 481 SIDL paiqur fo sarios Y oy) ur poonpoid $1997) 9y7 fo az1s obpwvan pun soqunu ay) sdnowb uoyn.iouab snorwa sof Buinoyy * ATAVL 342 HELEN DEAN KING An examination of table 4 shows that all of the litters produced in the first generation group of the A series were smaller, on the average, than corresponding litters in the later generation groups. The relatively low fertility of the animals in the early generations was not due to inbreeding, but to the fact that these individuals suffered from malnutrition. As soon as the nutritive conditions were improved there was at once an increase in the number and in the size of the litters produced, as the data for the fifth and for the sixth inbred generations show (table 1). As indicated in the last column of table 4, the groups compris- ing the tenth to the twenty-fifth generations of the A series showed, as a whole, comparatively little variation in the average size of the litters. The maximum average size (7.8) came in the group including the fifteenth to the eighteenth generations. This maximum was, however, only 0.1 greater than the average litter size for the preceding and for the following group, and therefore it can have little, if any, significance. Litter data for various generation groups in the B series of in- breds are shown in table 5. . As the average size of the litters produced in the first gener- ation group was greater than that in the second group (table 5), it might appear that the fertility of the breeding females in the B series was not lessened by malnutrition. In the beginning of these experiments many more females of the B series than of the A series were completely sterile, but the females of the B series that did breed were the more productive. Malnutrition, in this instance, was a selective agent that helped to eliminate the tend- ency to sterility in the B series by preventing the breeding of any except the most fertile females. In the B series the maximum average size of the litters was found in the group comprising the nineteenth to the twenty- second generations, but, as was the case in the A series, this max- imum was not great enough to be considered significant. Litter data given in table 4 and in table 5 have been combined in table 6. The data for each of the two inbred series, as well as that given in table 6, shows that in all generation groups the first litter cast 343 EFFECTS OF INBREEDING ON FERTILITY AND VIGOR pL | gee‘st| 9S9'T | aL | sez'% | 20Oe | SL | Ole | GIF | BL | 9098 | ZO | 69 | GI's | 89F [Se-T VL 70'S | E22 GL | OLE Gg V8 | 667 We Leh N\A GL 92 | GZS GL |SG-82 QL CF9'S | THE Sel Ser 6¢ 68 | 102 G8 08 | 662 OOT Te Ze | OEZ OOL (22-61 GL 825% | PSE & 2 | v&sS €2 tee OPA 0) 68 GL | G9 68 Geos eSLo 68 |8I-ST inv 606° | 662 V2, |08SP 19 624 | 809 92 % 4 \-909 18 £9 \\"Z¥S Ig |PI-IT Tey CE9‘T | BES G9 | The LE OL | C&P 19 Sie GOs: L9 69 | ZIP L9 =«|01-L Sed 8161 | PLT 82 | 96T GS LEN ARS &F 82 | 068 OS SHOE ESaG Gh a9 sk aunos yo| pur jo /SSemilFo runs yo| SenPes | ston rungs yo] SMNPES | SMH runes yo| SeMPEs | Som lounge yo] SenPES | SNH ae ae [830], amnaae requiny |zequinn esas jequiny |zequiny eee joquiny |roquinny Bee jaquny |requinyy SNOILY -4INAD pamqur-sey s199}1] WJM | (perqur-yey) sx0yy pry, | — (perqur) 810491] puodag (perqur) 810491] ISL AUVWWOS SaIHas YALLIT U sjoi posqui fo saiuas g ay] ur paonpoid s4aj}1) ay) fo azis abpsvan pun waqunu ay? sdnowb uornsauab snoruwpa sof Buinoyy $ HIAVL GL CSP‘GZ| 80'S @L | P9L'F §S9 82 | 699'9 LS8 62 | Sor, CF6 69 | F2g°9 696 PL | &h0'r | Lye 2 cou LOT Sry e003 | ser eee Oda on Poe Gi iey ie Ten ey de ogs'¢ | STZ Ory, €86 0&1 18 | Sort OST T'S | g9'T POS Tee espe FOG “3 Les One Gs O29 F 2 |000T a) Selle 8.3 | S260 al Gale al Or Senlek op tl cella eel ee O tania meas q G2 | IZ | 129 D2. ISCO a FET 3) 0 8.) 2eC or |) SAS a 6 Ze ol meo Tale SON 8 Ube eee ial ale ol as GL C8P's | FSP 6°9 TLS €8 GL | &6 LCL Lise Clee LET V9 | 928 LET ng 8°9 06% | 89 6°9 LOF 19 O02 | SE9 68 GL 6S2 SOT 19 | 689 SIT = : g sae al toa te | ya gee 28] srompra | sroxgt [OMT syenpra | ssoxy [OM syonpra | sxoxy [OME syenpra | sxoy91 goquuinu Jroqunyy jaoqtunyy | zaquinu | tPUTjJo | jo zequint | 1PUT JO jo goquinu | “tPUt Jo Jo gequinu | PUrjo |) jo Z aseBloay | [B}O], [®10]., | osBvi0Ay ENG |2esa Nt OSBIOAV PEISMEELNINT | EXC SUE INT oavIDAV TOCRETN, 4 | SPSUUN OSBIOAY ROS OTN TeGUTTN, ea Ed peiqui-jjey) s10441] YANO (perqui-j[By) s104}1; Paryy, (periqut) s819}}1] puoseg (perqut) 810341] 4SsIt7 an AUVWWOS = g 919] Ur puD F 97qQD] Ur DID BY} [oO UOYWULQuUOD D SG1UaS YaALLIT SNOILVY -HONGSD i(Q‘V) salsas pasqui ony ay) ut paonpoud 819j}1) 9y) fo az1s abpwaan pun soqunu 9y) sdnosb uoynsauab snorwa of Buinoysy 344 9 WIaVL EFFECTS OF INBREEDING ON.FERTILITY AND VIGOR 3409 was the smallest in the litter series, as a rule; the second litter was the largest; the third and fourth litters were intermediate in size between the first and the second; a similar relation in the size of litters has been found, also, in two groups of stock Albinos. In both inbred series the first two litters cast in each generation were the offspring of brother and sister matings; the third and fourth litters were produced by the mating of an inbred female with an unrelated stock male. In the first of the stock series noted (King and Stotsenburg, 715, table 7) all of the litters were produced by the pairing of unrelated stock animals; in the second stock series, for which data are given in table 7 of the present paper, all of the litters obtained were the offspring of brother TABLE 7 Showing the average size of each of the first four litters produced by a series of stock Albino females AVERAGE NUMBER OF LITTER SERIES NUMBER OF LITTERS NUMBER OF INDIVIDUALS YOUNG PER LITTER 1 116 717 6.2 2 116 843 m3 3 103 671 6.5 4 89 587 6.6 494 2818 6.7 and sister matings. Since in all three groups the average size of the litters in the litter series varied in a similar way, it is evi- dent the litter size does not depend at all on the relatedness or the unrelatedness of the parents, but chiefly on the age of the female. Young females tend to be somewhat less prolific than older ones, as Crampe (’83) noted. The litters reach their maximum size when the females are about five months old, but the number of young does not decrease appreciably in the various litters cast until the females have passed the height of their repro- ductive power at about seven months of age (King, ’16b). As shown in the first paper of this series (King, 718), the rats in the seventh to the ninth inbred generations were considerably heavier, at any given age, than the individuals belonging to sub- sequent generations. The cause for this unusually vigorous 346 HELEN DEAN KING growth was attributed to a stimulation of the growth processes produced by adequate nutrition following a period of semi- starvation. During this period the productiveness of the females was increased considerably, since the average size of the litters in the group comprising the seventh to the tenth generations was 0.4 greater than the average for the previous generation group (table 6). The period of maximum fertility in the inbred series did not, however, coincide with the period of maximum growth in body weight, but came at a much later time (fifteenth to the twenty-second generations), when the litters contained 7.7 young, on the average. The fact that the average size of the litters in the last three generations of the inbred series was slightly lower than the maximum can be attributed to a change in diet made necessary by the economic conditions of the present time. This diet does not seem to be quite as favorable to growth and fertility as was the more varied diet used until the beginning of last year. The graph in figure 1, showing the average size of the litters produced in the various generations of the inbred strain, was con- structed from the data in the last column of table 3. Starting at the point of 6.8, the graph in figure 1 drops at the third generation to 5.0, the lowest point in its course. From this point it rises slowly, and after the fifth generation tends to be a’ fairly horizontal line, since it never falls below 6.9 nor does it rise above 7.9. At more or less regular intervals the graph drops slightly below the normal level. The most pronounced depres- sion is at the point of the third generation; a second drop comes at the ninth generation; other depressions of about the same depth are found at the point of the sixteenth, the twentieth, and the twenty-fourth generations. As the last three depressions in the graph occur at intervals of four generations, it is evident that they were not due to a chance variation in the data, but that they must express periodic changes in the reproductive cycle of the females that tended to reduce the number of young born. In whatever way this reduction was effected, whether by a lessening of fecun- dity or by limiting the number of embryos that were capable of normal development, the cause for it, I believe, lay in the seasonal changes in temperature which always have a marked effect on the EFFECTS OF INBREEDING ON FERTILITY AND VIGOR 347 physical condition of the animals. During the summer months rats suffer severely from excessive humidity and from high tem- perature, since their mechanism for heat regulation, under these conditions, is inadequate. At this season their sexual activity is at its lowest point, and the litters that are produced tend to be relatively small. Severe cold checks reproduction, but litters born under these conditions are usually of normal size and the ppt yee Average Size of Litter SESGS 0000000055055 BREDSEaES : GENERATIONS H EEE CEE EEE EEE EEE EEE FF 4 22 DODO S PSR LOSSeSeBeoseeaeRR0neea Lee27 5. 37 ee C8) 9) 10 <0) 12 1S 14°16 “t6,UI7'- 18 “19° 20° 21 22 23 124 °25 Fig. 1 Graph showing the average size of the litters produced in the various generations of the inbred strain (data in table 3). individuals strong and vigorous. At about every fourth genera- tion the majority of litters produced in the inbred strain were born at the most unfavorable season of the year, the summer and early fall. In this generation, as the records show, the litters were smaller, as a rule, than the litters in the preceding and in the following generations. This decrease in size was sufficient to account for all of the depressions in the graph in figure 1, except the first one, which was doubtless due to the fact that the maxi- 348 HELEN DEAN KING mum effect of malnutrition in lowering the fertility of the females was reached at the third generations. Cyclic changes in productiveness were noted by Castle et al. (06) in an inbred strain of Drosophila, in which, for three suc- cessive years, there was a gradual rise in fertility followed by an abrupt decline. ‘These changes in productiveness were likewise ascribed to the variations in temperature at different seasons of the year. The data given in table 1 to table 6 and the graph in figure 1 show clearly that, despite all theories to the contrary, it is possible to maintain a high degree of fertility in a mammal for at least twenty-five generations of the closest possible form of inbreeding, by a careful selection of breeding stock and by keeping the ani- mals under environmental conditions that are favorable for their growth and reproduction. While, in general, the size of the litter varies according to the age of the mother, individual females differ greatly regarding the number of offspring that they produce in any litter of the litter series. Sisters from the same litter, mated to the same male, will show marked variations in their fertility at the same age. One female may never have a litter that contains more than five young; the other may always throw litters in which there are nine or more young. Some females, regardless of their age, tend to east the same number of offspring in every litter. One female so noted had ten young in each of her four litters. Marked indi- vidual differences in fertility are also found among female guinea-pigs, according to Minot (92). The average number of young in a litter of albino rats is 6.3, according to the data for 394 litters collected by Crampe (84); Cuénot (99) found an average of 8.5 young in the 30 litters that he examined. Records for 1089 litters of stock Albinos born in the Wistar Institute animal colony during the years 1911 to 1914 give 7.0 as the average number of young per litter (King and Stotsenburg, 715). When this last series of data was collected it was not realized that litter size in the rat depends to such a marked degree upon the age of the mother, and that in this species the maximum fertility comes at a relatively early age, as it does EFFECTS OF INBREEDING ON FERTILITY AND VIGOR 349 in the human race (Powys, ’05) and also in poultry (Pearl, 717). Most of the litters recorded were cast by young females that had not reached the height of their reproductive power; such litters tend to be larger than those cast after this time (King, ’16b). Data for litters cast by females of unknown age, however exten- sive they may be, cannot, therefore, properly be used to furnish the norm for litter size in the albino rat. In order to obtain standards for litter size with which the data in the inbred strain might justly be compared, the complete breeding history of a considerable number of stock Albino females was recorded during the past three years. Data for the first four litters produced by 116 females belonging to this group are given in table 7. All of the stock rats from which these litters were obtained were reared under the same environmental conditions as the inbred strain. In table 7, as has already been noted, the litters of the series bear the same size relation to each other as that found in the litter series of the inbred rats. The first litter was the smallest, aver- aging 6.2 young; the second litter, with an average of 7.3 mem- bers, was the largest of the series; while the third and fourth litters were somewhat smaller than the second. The entire series of 424 litters gave an average of 6.7 young per litter. ‘This average is 0.3 less than that in the random collection of stock litters previously recorded (7.0), and 0.4 more than the norm as given by Crampe (6.3), so it is probably a fair standard for litter size in any similar series of Albino litters. There is no reason to believe that the stock females from which the litters recorded in table 7 were obtained were, as a group, inferior in reproductive power to other stock females, and presumably their fertility at any given age is fairly representative of that in the general run of stock Albinos. Each litter of the stock series, shown in table 7, contained a smaller average number of young than the corresponding litter in either of the two inbred series when the data were arranged according to generation groups (tables 4 and 5), and, omitting the records for the first five generations where the fertility was les- sened by malnutrition, there was not a single generation in either 350 HELEN DEAN KING of the inbred series where the average size of the first four litters was as low as that in the stock series (tables 1 and 2). In the inbred series as a whole, the average size of the litters was 0.8 greater than that in the stock series. Even if the previous finding of 7.0 be taken as the norm for litter size in the rat, the difference between the average size of the litters in the inbred strain and the norm chosen is 0.5. This difference is great enough to preclude the possibility that it was due to chance, and it cannot be attrib- uted to the differential action of environment, since stock and inbred rats were constantly under the same environmental con- ditions. According to these findings, fertility in the inbred strain of Albinos, in as far as it may be judged by the size of the first four litters cast by a large number of females, was greater than the fertility in stock Albinos that were not inbred. B. Frequencies of litter size According to the several series of observations that have been recorded, there is a wide range in the size of the litters cast by Albino females. Kolazy (’71) reports litters containing from five to seventeen young, although Crampe (’84) states that he never found even fourteen young in a litter of albino rats. Litter size varied from four to twelve in the series of Albinos studied by Kirkham and Burr (15); while in the litters recorded by King and Stotsenburg (15) the range in size was from two to fourteen. Data for litter frequencies in the two series of inbred rats are shown in table 8. TABLE 8 Showing the frequencies of litter size in the two series of inbred rats SIZE OF LITTER 2 3 4 5 6 7 8 9 10 11 12 13 14 15 LO ey 1 A | 1 | 40) 59) 103] 193] 182} 288} 268] 258) 181) 99) 45 | 27} 5] 2{ O/| 1 0 | 35} 72| 102} 168) 206] 263) 260} 208) 152) 110) 49 | 22; 6| 38] 0O| O 1 | 75] 131] 205} 361} 388} 551] 528] 466| 333) 209} 94 | 49/} 11] 5] 0| 1 EFFECTS OF INBREEDING ON FERTILITY AND VIGOR dol In the A series of inbreds the range in litter size was from one to seventeen. The litter of one was cast by a female of the nine- teenth generation that was suffering from pneumonia and had to be killed three days after parturition. This is undoubtedly a case where the physical condition of the female prevented the normal development of all of the embryos except one; the other embryos probably became atrophic and were absorbed. The litter containing seventeen members occurred in the fifteenth generation. All of the individuals were born alive, but they were all very small, weighing not more than three grams each: the average weight of the albino rat at birth is about four grams (King, 715b). In the B series, as table 8 shows, the range of variation in litter size was not as great as that in the A series: no litters smaller than two or larger than fifteen were obtained. In both series litters containing seven young were the most frequent, while those with eight young were only slightly less in number. Figure 2 shows graphs for litter frequencies in the two series of inbreds that were constructed from the data given in table 8. In figure 2 each graph rises quickly to the modal point at seven, falls slowly at first and then rapidly. The drop in graph A at the point of 6 has apparently no significance, since there is no similar drop in the B graph. Each graph is a simple frequency curve with one modal point, and is exactly the sort of graph that one would expect to obtain from the data for litter frequencies in a large series of animals belonging to a pure race. C. Puberty Under normal conditions puberty tends to appear at approxi- mately the same age in the different individuals of a given race, but the time of its appearance is seemingly more dependent on the growth changes incident to age than it is on age itself. In the albino rat both males and females attain sexual maturity when they are about two months of age (Donaldson, 15), but the environmental and nutritive factors that hasten or retard growth have considerable influence on the reproductive activity of the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO.2 — KING HELEN DEAN 302 an) “<8 ~~ "Sa ao {TT tT pe SERERERERESS | SERS Graphs showing the frequencies of litter size in the two series of inbred rats (data in table 8). Fig. 2 EFFECTS OF INBREEDING ON FERTILITY AND VIGOR 353 individuals. If young rats are fed exclusively on a meat diet, puberty is considerably delayed (Watson, ’06); the same effect is produced by underfeeding (Osborne, Mendel and Ferry, ’17). According to my observations, the time of year in which the ani- mals are born affects their subsequent growth and also the time of their maturing. Rats born in the winter and early spring grow rapidly, and usually breed at about three months of age; those born in the summer and autumn grow more slowly and compara- tively few of the females cast litters before they are four months old, many not breeding until spring, which is the season of the most pronounced sexual activity for the rat. Convincing evi- dence that age alone does not determine the beginning or the end of the reproductive life of the rat is given by Osborne and Mendel (15, 717), who found that Albino females, stunted at an early age by underfeeding, were completely sterile until they were properly nourished, when they grew rapidly, attained a normal size, and were able to breed long after the age at which the meno- pause usually appears. According to Darwin (’75) and others, favorable environment tends to delay sexual maturity, though not necessarily to decrease fertility. Since these inbred rats were reared, for the most part, under environmental conditions that seemed well adapted to their needs, and since they lacked the stimulus to reproductive vigor which is said to come from outcrossing, it might be expected that they would tend to mature much later than stock Albinos which were not inbred. Table 9 shows the approximate age at which the breeding females belonging to various generation groups of the two inbred series cast their first litter. The records for the first generation group, given in table 9, confirm Osborne and Mendel’s findings that underfeeding tends to retard sexual maturity, since they show that about one-half of the breeding females in this group did not cast their first litter until they were four months old. In subsequent generations, when the animals were adequately nourished, they began breed- ing at a much earlier age. Under the conditions of this experi- ment, inbreeding seemingly hastened the onset of puberty, for 354 HELEN DEAN KING in both series, as the inbreeding advanced, there was a marked tendency for relatively more of the females to breed at the earliest possible age. About 30 per cent of the breeding females in the eighteenth to the twenty-fourth generation group of each series threw their first litter at or before the age of ninety days; only a small proportion of them failed to breed before reaching the age of four months. As a whole, the females of the A series of inbreds tended to mature slightly earlier than the females of the B series, but the TABLE 9 Showing the approximate age at which breeding females in various generation groups of the two inbred series (A and B) cast their first litters £2 SERIES A SERIES B SUMMARY (A,B) Bee|%, gs i8ee |88 |=, |22 (285 |28 | 2,22 [es | 8 Bae| we |ee jgee [es | wa | HE [EEE | Es | se] ES [28 | es AO | 2E [Hoje 8 0 | Se, | GE | So, (S80 | SH, | ok | as. (Ass | fe, aaa gk 2 bbl S Si | ge 2 hp — oo ga 2g S % ep BORER AL | Er a 1S senea) ce 0 ee i ese eee) eh | Te aa) nel e Bis ee shal 3 | S8eclesy | 85o| 3 | Sao [8st 83 oe = | 830 [See oye ae Taw Salo S So S>| as 8m>| $a 8 ral @ Sal Wiles 8 Bes | 22 | ks|,ko8| oko | 22 | 583 |e 8| -£2| 25 | s£S |eko9) £5 ) a m4 Ay Ay a Ay ay Ay Ss AY ay a 1 5] 95811220) $6.21" Sis) 56 | 14.3 | 4 | 44.6 | 114 | 18) Beh ease 69] 70! 21.4) 6876 | 10:0) 674) 7:4+) 7427 1917.9.) 1387 19146 | 71/6 tae 10-13), $84. |:25.0).71-4)), 3264) S136 Fa) 92.3) | 465) 19s B72 Beas 14-17] °G4 | 9897) 62.8.) 895 10°89 | 25.8: 6220) ) 11.3") 183-10 2723) | Goeo aes 18-21! 104 | 36.5) 62.6 | 0.9 | 100 | 31.0 | 64.0| 5.0! 204 | 33.8} 63.3 | 2.9 22-241 76 | 31.5} 60.6 | 7.9 | 75 | 20.0 | 69.4 | 10.6 | 151 | 25.8 | 65.0] 9.2 1-24| 486 | 27.1] 61.6 | 11.3 | 468 | 19.8 | 65.3 | 14.9 | 954 | 23.5 | 63.4 | 13.1 difference between the two series was not great, and correspond- ing records were in nearly as close agreement as were those for litter size. The youngest breeding female in the inbred strain was a member of the A series of inbreds, and she was eighty days old when she cast her first litter of five young. As the gestation period in the albino rat is about twenty-two days (Donaldson, 15), this female must have conceived when she was two months old. Kirkham and Burr (15) state that one of their Albino females gave birth to a litter when she was only seventy-seven EFFECTS OF INBREEDING ON FERTILITY AND VIGOR 3955 days old; while Lantz (10) reports a case in which an albino rat was said to have produced a litter at the age of fifty-six days. This last case is certainly a remarkable one, and its parallel has not been found among the 50,000 rats bred in our colony. The last section of table 9 shows that, after the tenth genera- tion, there was no marked change in the proportion of females that bred at three and at four months of age, respectively. Nearly 24 per cent of the total number of females used for breed- ing cast their first litter by the time they were three months old; over 60 per cent of them bred for the first time when they were between ninety and one hundred and twenty days of age; while about 13 per cent did not breed until after they were four months old. The latter group was made up, for the most part, of females that were born in the summer or autumn. Although Diising (’84) states that inbred animals tend to mature very early, I do not think that inbreeding alone was responsible for the fact that relatively more of the females in the later than in the earlier generations of these inbred rats bred at three months of age. In these experiments, when two or more females of a litter were reared as possible breeding stock, the first _ female that became pregnant was the one taken to continue the line, provided she fulfilled all requirements as to size and vigor. Thus the manner in which breeding females were selected pre- served those individuals that tended to breed at an early age, and this tendency. to early maturity, if heritable, must have been retained in the stock and intensified to some extent through continued brother and sister matings. Inbreeding, aided by selection, would thus seem to be the factor involved in producing a strain of rats in which the females attained sexual maturity at a relatively early age. D. Sterility Sterility occurs normally in the Albino, as in other strains of rats, and therefore it might be expected to appear at times in any strain, regardless of whether the animals were inbred or outbred. Crampe (’84) states that of 221 Albino females which he selected for breeding forty-six, or 20.8 per cent, were sterile. 356 HELEN DEAN KING Out of 124 stock Albino females reared in our own colony during the past three years and intended for breeding purposes thirty- two, or 28.8 per cent, were completely sterile, while about 10 per cent of those that did breed cast only one or two litters. Un- fortunately, no records have been kept that give information regarding the exact proportion of sterile females in the first six generations of the inbred series. The number was relatively very large, and must have included at least one-half of the total number of females that lived to be six months old. Sterility in these females was, for the most part, the result of poor nutrition, and it disappeared as soon as the nutritive conditions were im- proved. Loeb (’17) has shown that in the guinea-pig ‘under- feeding prevents maturation of the follicles and thus causes sterility which lasts as long as the effect of the underfeeding is present in the ovary.’”’ In the guinea-pig, as in the rat, adequate nutrition reéstablishes normal conditions in the ovary and sterility almost entirely disappears. In Drosophila, according to Castle et al. (06), low productive- ness (sterility) is directly transmitted by inheritance and is amenable in selection. In the rat, sterility seems to depend not entirely on genetic factors, but to a marked extent upon condi- tions, such as malnutrition and disease, that act unfavorably upon reproduction. In the present experiments, by selecting for breed- ing only the most vigorous individuals (which it seems were also the most fertile), sterility in as far as it may depend on genetic factors would seem to have been practically eliminated from the strain, and it has not reappeared even after twenty-eight genera- tions of brother and sister matings. . Of the 954 inbred females that were used for breeding during the course of these experiments, 653, or 68.5 per cent, cast four litters each, and many of them, kept for body-weight records, produced several other litters which were not recorded. Of the females that did not cast the required four litters, the great majority died from pneumonia, or were killed because they showed unmis- takable evidence of illness. is ahweaiott oitd ta 16} ponte K Oh ian Coe be ysot ed hy Agr Oi gins bea A Re amr a al nie, {yenairels Melis bets 3 res RRR: NS) HSA fia a iy Hg +a tle rsp hig Pe in a Hen aes ray tom LAS wits Cat ie Ee ut alae is ae (UAE RGAE he deni Cae? GA Te Mora ttt haat! yf AN rend Pe A fe bn 4) ast na 1h hia) ae th. ae Al,” oo) es FRO Ns Tek mse PTE te ae oe eerie WAP hz brag gs PE ea ae As ES tee bi lo Of, rae ‘ At d ¢ ie K AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE HYBRIDS BETWEEN FUNDULUS AND. MACKEREL A STUDY OF PATERNAL HEREDITY IN HETEROGENIC HYBRIDS H. H. NEWMAN Hull Zoological Laboratory, University of Chicago FOUR FIGURES I. INTRODUCTION The present status of heterogenic hybridology in animals Experiments in hererogenic hybridization of animals have been performed largely upon two groups, Echinoids and Teleosts. A much larger amount of attention has been paid to the former than to the latter group, although it now seems certain that the Teleosts in a number of important respects are more favorable than the Echinoids for this kind of work. The list of investi- gators who have contributed to the literature of hybridization among the Echinoids is imposing, including as it does such names as Boveri, Seeliger, Morgan, Vernon, Driesch, Delage, Stein- bruck, Loeb, Doneaster, Peter, Fischel, Herbst, Godlewski, Kupelwieser, Hogedoorn, King, Moore, Tennant, McBride, Fuchs, Debaisieux, Shearer, and DeMorgan. Only a few, largely Americans, have studied hybridization in Teleosts and the work has been comparatively superficial. Only the following seven or eight investigators have contributed to this field: Appellof, Bancroft, Hertwig, G. and P. Loeb, Moenkhaus, Mor- risand Newman. It is my conviction that had the same amount of attention been applied to Teleost hybridization as has been given to that of Echinoids, we would have a vastly better under- standing of the truth about heterogenic hybridization phenomena that we now have. 391 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, No 3 AuGustT, 1918 392 H. H. NEWMAN The hybrid situation in Echinords A close comparison of the results of hybridization in Echinoids with that of Teleosts has served to place the whole matter in a new light and to emphasize the need of further work on Teleosts. It seems necessary in making such a comparison to show graph- ically the interrelations of the species that have been crossed. The following table gives in italics the genera of Echinoids used in all of the various experiments, together with an outline of their classification (table 1). TABLE 1 GENUS FAMILY SUBORDER ORDER SUBCLASS Arbacia Arbaciidae Arbacina Echinus Hipponoe Triplechinidae Toxopneustes wohinina ( Diade- Regularia moida Strongylocentrotus \ : J Salaeresnais Strongylocentroidae Echinocardium Spatangidae Sternata | Atelosto- | Irregu- mata laria A goodly number of investigators have crossed Strongylo- centrotus and Sphaerechinus, using a number of species of each genus. Such crosses are of merely intergeneric width and are hardly to be classed as heterogenic. Interfamily crosses have been carried out between Hchinus of the family Triplechinidae and Sphaerechinus of the family of Strongylocentrotidae, and there is little difference of opinion as to the evidences of paternal heredity. Inter-sub-order crosses between Arbacia of the suborder Arbacina and members of the suborder Echinina have given less definite results as to the hereditary effect of the foreign sperm, Driesch stating that the larvae were pure maternal and Fischel that the paternal influence is shown in form, size, pigment, and skeleton. Neither author was able to rear the larvae far enough to study very definite characters. : HYBRIDS BETWEEN FUNDULUS AND MACKEREL 393 Only one cross of greater than suborder width has been made within the bounds of the class Echinoidea and that is of sub- class width. McBride and Fuchs have independently crossed a representative of the subclass Regularia with one of the sub- class Irregularia, using Echinus and Echinocardium. The percentage of fertilizations is so low and the larvae are so un- healthy that no definite conclusions were reached by either author. By means of chemical aids to insemination, ‘crosses’ of inter- class width have been made between Echinoidea eggs and the sperm of Astereridea, Holothuroidea, Ophiuroidea and Crinoidea. Though extremely few larvae resulted—and these were decidedly abnormal—they were described as pure material. By the same chemical means Echinoid eggs have been inseminated with the sperm of Mollusca and Annelida, and occasional stunted larvae have resulted that also are said to be pure materinal. These interphylum ‘crosses’ mark the extreme limit of heterogeneity in hybridization, if indeed they are hybrid phenomena at all. One step further and we would have crosses of subkingdom value (Protozoa x Metazoa) or even of kingdom value (plants x animals), which would amount to a reductio ad absurdum. It must be borne in mind that, in Echinoids, ‘crosses’ of greater than suborder width are impossible in nature and can be made only by the use of chemical agents that serve to break down the normal incompatibility of the heterogenous germinal materials, and may well play a réle analogous to that of parthenogenetic agents. ‘Crosses’ of greater than suborder width are described as being purely maternal, although the criteria for this judgment are ex- tremely doubtful since the embryo and larvae reared are highly pathological. No paternal heredity is to be expected in view of the fact that the foreign chromatin forced into the egg unnaturally, utterly fails to coéperate in cleavage or in more general cell metabolism, but remains an inert mass till gradually eliminated by absorption. An exception appears to exist in the case of the cross between Echinus and Antedon, in which Godlewski claims a normal behavior of the paternal chromosomes, except that 394 H. H. NEWMAN the latter come to be indistinguishable from Echinus chromo- somes; yet the larvae are pure maternal in so far as they are any- thing definite. In these wide ‘crosses’ it appears reasonable to conclude that the foreign sperm at least assists in the initiation of development, duplicating in a sense the rdle of partheno- genetic agents. According to Loeb: ‘‘The egg behaves exactiy as we should expect from the fact that the spermatozoén removes only certain obstacles for the development of the egg, but does not cause its development by carrying any activating enzyme.” It may be concluded, then, that in Echinoidea crosses of generic and family width exhibit paternal heredity, that when Echinoids are ‘crossed’ with other classes or other phyla that heredity is not concerned, but merely initiation of development. The really interesting and critical phases of Echinoid hybridiza- tion lie between these extremes: in crosses of subordinal, ordinal, and subclass width, and there are, unfortunately, very few data on this point. A cross of subordinal width was studied by Driesch and by Herbst. Their results and conclusions are diametrically op- posed; the former concluding that the hybrid larvae were purely maternal and the latter that the paternal influence was seen in several characters. More work is needed upon crosses of this width. A cross of subclass width was made by McBride and repeated by Fuchs with results quite different in the two cases, and it seems certain from what these authors say that the larvae are so few and so abnormal that little conclusive evidence as to heredity is available. In conclusion it may be said that the results of hybridization ex- periments of greater than interfamily width among the Echinoids have been entirely inconclusive, and it would appear necessary to transfer our attention to Teleost material if we wish for any approach to a definite answer as to the question whether heterogenic hybrids show paternal heredity or are merely parthenogenetic in character. HYBRIDS BETWEEN FUNDULUS AND MACKEREL 395 The hybrid situation in Teleosts as compared with that in Echinoids Taxonomically the group Teleostei is of lower value than the group Echinoidea, the former being merely an order of the sub- class Teleostomi of the class Pisces, while the latter is a class on a par with the class Pisces. A group of Echinoidea correlative with the order Teleostei is the order Diademoida of the sub- class Regularia. A strict comparison of hybrid phenomena would therefore be limited within the confines of these two orders, Diademoida and Teleostei. As a matter of fact, only one Echinoid cross has been made that exceeds the limits of the Diademoida. No one has as yet succeeded in crossing a Teleost with any of the Ganoid orders of Teleostomi nor with any other subclass of Pisces, much less with members of other classes of vertebrates or with invertebrate phyla. By the use of appropriate chemicals it might conceivably be possible to inseminate Teleost eggs with other than Teleost sperm, but my impression is that the micro- pyle would be too much of an obstacle for a type of sperm very different from that of a Teleost. Even if a foreign spermatozoén could be forced past the micropyle barrier, it would doubtless be unable to initiate development for the same reasons that the several sorts of parthenogenetic agents, that have been extensively tried upon Teleost eggs, have failed to bring about any develop- mental result; for Teleost eggs, unlike Echinoid eggs, are extreme- ly refractory to parthenogenetic agents and would therefore probably fail to respond to foreign sperm. Hybridization ex- periments among Teleosts are, therefore, confined entirely within the order and at most will be of subordinal width. A very large number of crosses of subordinal, of family, and of generic width have been made. A companion of the following table (table 2), which represents the classification of the genera of Teleosts, with table 1 will be of interest. A survey of this table shows that nearly half of the genera used are members of the large suborder Acanthopterygii, just as most of the Echinoid genera used in crosses belong to the suborder Echinina. The four suborders in the table have been crossed in a great many ways and with varying degrees of success. At 396 H. H. NEWMAN least as wide a cross as any possible within the bounds of the genera given in this table is that between the first and the last genera, Fundulus and Scomber. In general it may be said that any Teleost will cross with any other with the exception of vivi- parous species, and those with peculiar breeding or brooding habits, such as the pipe-fishes, where secondary obstacles to cross insemination present themselves. It is also worthy of note that no artificial aid to insemination is necessary In any Cross. - In the ease of crosses between closely allied species, as, for instance the various species of Fundulus (F. heteroclitus, F. ma- jalis, F. diaphanus), in which paternal heredity is as obvious in many ways as is maternal, in that paternal characters are often TABLE 2 Genera of Teleoster used in hybrid experiments GENUS FAMILY SUBORDER Fundulus ; , E 3 H Ciamaton } Cyprinodontidae aplomi G eus | : ; BBE OBUELE Gasterosteidae Cateasteomi Apeltes Menidia Atherinidae | Peuseseeid Pronotus Stromatidae { ity Morone Serraninae | \ Stenotomus Sparidae | & ) c t ) y 8 Tautogolabrus Tee ee Acanthopterygli Tautoga Scomber Scombridae | dominant over maternal. The same is clearly true for crosses between two genera of the same family as when the genus Gas- terosteus is crossed with Apeltes, both genera belonging to the family Gasterosteidae. In this case a large percentage of healthy hybrid larvae hatch and they are by no means pure maternal. Crosses of interfamily width such as that between representatives of the different families of Acanthopterygii give similar results, in so far as success in development and strength of paternal heredity are concerned, as do crosses of suborder width and need not be considered at length. As a very typical case of heterogenic hybridization of suborder width among the Teleosts I have chosen that between Fundulus heteroclitus of the suborder HYBRIDS BETWEEN FUNDULUS AND MACKEREL 397 Haplomi, and Scomber scombrus of the suborder Acanthop- terygii. This cross between mud-minnow and mackerel is an excellent one for our purposes because it shows in the clearest way what many other crosses show, but show in a less obvious way. One reason for making an intensive study of one particular heterogenic cross was because | have felt that certain incorrect conclusions as to the nature of heterogenic Teleost hybrids had been published by well-known writers. Moenkhaus holds that the development of these hybrids is pure maternal and that the cleavage rate of the hybrid egg is unaltered by the foreign sperm even if the cleavage rate of the paternal species is much slower or much more rapid than that of the maternal species. I have shown that this is incorrect by the use of more exact meth- ods of comparison. There is an acceleration of cleavage and development, accompanied by a heightened vigor when two very closely allied species are crossed. In any crosses except those of very closely allied species the rate of cleavage and subsequent development is retarded and more or less abnormal (subnormal) larvae result. This does not depend directly on the distance of the cross, for sometimes crosses of suborder width show less retardation and a higher viability than crosses of generic width. Loeb, on the basis of certain rather unfortunately chosen crosses, supports Moenkhaus in his statement that cleavage rate and that of subsequent development is pure maternal and goes farther in claiming that heterogenic hybrids are pure maternal in their hereditary characters. He thinks the hybrid larvae, which in all of his experiments appeared to be quite unhealthy, are merely pure maternal or parthenogenetic larvae that are poisoned by the presence of materials brought in by the sperm. The argument in favor of the pure maternal character of devel- opment, which implies parthenogenesis, is that ‘‘if the develop- ment of the egg were caused by an enzyme carried into the egg by a spermatozoén, developing eggs should be accelerated by a spermatozo6n of a species developing at a faster rate.’’ This conclusion does not appear to me to follow at all, for we must not forget the high specificity of enzymes. An enzyme that is ca- pable of setting up a high rate of activity in one species of proto- 398 H. H. NEWMAN plasm would not be expected to activate so readily another species of protoplasm. It might conceivably retard develop- ment beyond the normal rate even in a slowly developing egg and still play the typical réle of sperm materials in heredity. This is, I believe, exactly what happens in all heterogenic hybrids, for they all exhibit more or less pronounced evidences of early and long-continued retardation. That the sperm actually do cooperate in development even in suborder crosses is shown by the unmistakable cases of paternal heredity. It is inconceivable that a foreign sperm could function in heredity without effective functioning in development, and any hybrid in which paternal heredity is demonstrated is neither a parthenogenetic individual nor pure maternal. It is just exactly this point that the present experiments demonstrate, to my mind at least, beyond controversy. I have selected the Fundulus x Mackerel cross out of nearly one hun- dred crosses that I have personally made among the Teleosts because it is especially favorable. In no sense, however, is it exceptional or peculiar in character, merely a little clearer and more diagrammatic than others that might have been chosen. Any one of half a dozen other crosses would have done nearly as well. II. EXPERIMENTAL Differences between adults of Fundulus heteroclitus and Scomber scombrus Fundulus and the mackerel are sharply contrasted in all of their adult characters as might be expected in representatives of differ- ent suborders. They differ radically in habitat, ecological relations, breeding habits, eggs, and larvae. Fundulus is a minnow with shore-feeding habits, is found in both salt and brack- ish water, is tolerant of foulness in water and to low oxygen and high CO, concentration. The large eggs (25 mm. in diameter) are laid during a clasping act on the part of the male which insures fertilization of the eggs with a minimum expenditure of milt. The eggs sink to the bottom and adhere to stones and seaweeds at or near the bottom by means of the sticky egg HYBRIDS BETWEEN FUNDULUS AND MACKEREL 399 envelope, which probably also protects the eggs from injury of various kinds. In correlation with the fact that eggs are ferti- lized immediately on their emission from the oviduct, it is note- worthy that the spermatozoa live only a few minutes, at most five, in seawater. Consequently, any females that have been isolated for any appreciable length of time could not possibly carry sperm upon their bodies. Therefore any criticism of these results, based on the assumption that occasional heterogenic hybrids that go through successfully to hatching might be due to chance fertilizations by sperm of the same species, have no foundation in fact. All that one has to do to obviate the possi- bility of any such contingency is to isolate the Fundulus females and handle no Fundulus males during the hybrid experiments. This was done in every case. The mackerel, in contrast with Fundulus, is a rather large fish, frequenting the off-shore waters except when they come in closer to breed. They are intolerant of water impurities, die soon in aquaria and show themselves extremely sensitive to lack of oxygen or high CO, concentration in the seawater. The eggs are about 1 mm. in diameter, typically pelagic in character, but differ from most pelagic eggs in having a distinct pinkish cast. The milt is extremely abundant and is evidently shed in clouds in order to ensure fertilization of the scattered eggs. The spermatozoa live in aquarium seawater for about twenty minutes and probably live even longer in natural seawater. Differences between embryos and larvae of the Fundulus and mackerel The egg of the Fundulus develops much more slowly than does that of the mackerel, taking about two weeks to hatch as compared with about two and a half days for the latter. Only in late larval stages are there significant differences of bod- ily structure and proportions. These do not need description here as none of the conclusions here brought out have to do with such general characters. Our attention must be focused upon one type of character, the peculiarities of the yolk and body chromatophores. 400 H. H. NEWMAN Fundulus has two well-defined types of chromatophore: a type of melanophore (black chromatophore) characterized by large squarish body and few short branches, showing also a tendency to fuse into syncytia, and a red or red-brown chromatophore which in its definitive or fully expanded condition is very intri- cately branched. The black type, if unhealthy, may remain small, and give out a few slender branches, but is never a very intricately branched cell in pure Fundulus embryos. The red chromatophores also may be relatively simple or unbranched if the embryo is pathological or retarded. The mackerel larvae also have two types of chromatophore equally characteristic of the species and quite distinct from those of Fundulus. One type is a melanophore which is characterized by a small core or body and very slender anastomosing branches. They never exhibit the tendency to fuse into syncytia. The second type is an olive-green chromatophore that occurs in the larvae both on body and yolk sac. Usually there are two large green cells just back of the eyes, two more a short distance behind the otic vesicles, and two or three adjoining the Kupfer’s vesicle on the yolk sac. There are no red cells in mackerel nor any green cells in Fundulus. So in these two opposed char- acters there is a sharp contrast, and the finding of a red chromat- ophore in a mackerel egg-hybrid or a green chromatophore in a Fundulus egg-hybrid could not be interpreted as within the range of variability of the maternal species or as a pathological occurrence. A fair-minded critic will admit that the only rea- sonable explanation is that the factors for the foreign type of cell have been introduced by the foreign sperm. Methods and general results of the hybridization experiments During the latter half of June, when both Fundulus and mack- erel are at the height of the breeding season, is the best time to make the crosses. Results of experiments made later in the season, although to all appearances development is more fre- quently normal, may be quite different. HYBRIDS BETWEEN FUNDULUS AND MACKEREL 401 The mackerel egg gives very poor results when inseminated with foreign sperm of any of the numerous species tried. A very large percentage of eggs develop, but many cleave quite irregu- larly. A somewhat smaller percentage proceed normally or nearly so up to gastrulation and then die. I have never succeeded in raising mackerel eggs inseminated with the sperm of any foreign species up to the point of embryonic differentiation, and hence no chromatophores are formed. ‘This study of heredity is there- fore confined to one of the two reciprocal crosses: that produced from the Fundulus egg inseminated with mackerel sperm. The striking difference between the developmental success of the reciprocal crosses is not at all an uncommon phenomenon. I described in a former paper numerous other cases of the same sort (Newman, °15). The explanation seems to be that the mackerel egg is extremely sensitive to foreign sperm material and fails to tolerate it beyond the cleavage period. A detailed account of a single typical experiment in hetero- genic hybridization will serve to bring out the essential facts. A typical hybridization experiment between Fundulus 2 xX mackerel ff The following account gives the details of five experiments performed on June 12 at the Woods Hole fish traps. An un- usually fine lot of female Fundulus were segregated from males and taken in a large aquarium in order to have the best possible conditions. A large female was chosen, the eggs stripped into a bacteria dish and inseminated with milt from a vigorous male mackerel just taken from the traps. After about ten minutes the excess of milt was washed out carefully to avoid contamina- tion of the water. Four other exactly similar experiments were performed using other males and females. This one experi- ment here described gave a somewhat wider range of types than the others, but all gave substantially the same results as far as fundamental matters are concerned. After about five hours the eggs in all experiments were examined and all unfertilized eggs removed. In the five experiments the following percentages of 402 H. H. NEWMAN eggs developed at least as far as the early cleavage stages: 82, 91, 77, 85, 72. There are always a few immature and possibly some overripe eggs and one gets about these same percentages in a pure-bred lots of Fundulus eggs, so that we may conclude that the percentage of fertility of Fundulus eggs with mackerel sperm is about normal. Although the eggs in these experiments were well cared for and casually examined from time to time, no detailed study was made until nearly one week after fertilization, when the chromat- tophores had made their appearance in abundance. On the seventh day, however, a complete census of the hybrid embryos was made on the basis of their relative success in development and the heredity of maternal and paternal types of chromatophores. It was possible roughly to divide the entire lot into eight classes, beginning with those that showed the most nearly normal devel- opment and ending with those that had died since development began. A full account of the conditions seen on the seventh day is herewith given and is to be compared with another census taken after nearly three weeks. Class A. A small group of three individuals apparently normal in every way. The heart-beat is strong, the circulation abun- dant and vigorous. Although a little belated as to stage of devel- opment as compared with pure-bred F. heteroclitus embryos of equivalent age, they are exactly like the latter even in the chromatophores which are pure maternal. It may be said that from one to five such embryos appeared in all five experiments and cannot be due to accidental insemination with F. heteroclitus sperm. Class B. Five individuals are slightly retarded as compared with class A, due evidently to their failure to establish a circula- tion. The heart in each is large and pulsating vigorously, and in two individuals contains some blood which moves back and forth in the heart chambers as the pulsation drives it. The chromatophores are all of the maternal type except that the black chromatophores are somewhat more branching than one expects in pure-bred Fundulus embryos. HYBRIDS BETWEEN FUNDULUS AND MACKEREL 403 Class C. In asense this class is ahead of class B, but in another ranks below it. Two individuals of a decided paleness and more slender of body than those of class B have established a circula- tion. The heart, however, is of small proportions and the red blood corpuscles are few in number. The blood stream is quite sluggish as compared with that in class A. The feature of most significance in these embryos that marks them off sharply from those of earlier classes is that they both exhibit paternal heredity, in that green chromatophores are found scatteringly on the yolk, but none on the body. The black and red chromatophores of the maternal species greatly predominate. Class D. Eight individuals showing chiefly subnormal condi- tions in the head region. The eyes are small or asymmetrical in size or position; in one the eyes are very close together ap- proximating a eyclopic condition. ‘The hearts are typical string- hearts pulsating vigorously. Both Fundulus and mackerel types of black chromatophores are present in abundance, but the maternal type predominates slightly. Red Fundulus chromatophores are more abundant than green mackerel chromat- ophores, but the latter are present in abundance. Class E. < Scomber scombrusc toward the middle of June. If the cross is made two weeks later it will be difficult to find the green chromatophores. The correlation between success in development and the strength of paternal inheritance It has been pointed out that in this cross, as well as in a con- siderable number of other heterOgenic crosses, that the most HYBRIDS BETWEEN FUNDULUS AND MACKEREL 409 nearly normal hybrids, those that reach an advanced stage of development and occasionally hatch, show the least evidences of paternal heredity—sometimes being pure maternal in their chromatophores and in body form. If these larvae are absolutely pure maternal, and I see no reason for doubting it at present, the explanation of this hybrid phenomenon offers a very pretty prob- lem in genetics. Loeb, as we have seen, classes all types of heterogenic Teleost hybrids with his extreme interclass echinoderm ‘crosses,’ and con- cludes that they are to be explained as products of a sort of parthenogenesis, the foreign sperm playing merely a development- initiating réle and not functioning in later development and heredity, except negatively as a retarding agent. The present studies show that this view is quite unacceptable on three counts: 1. The experiments just recounted prove that the sperm functions in development and heredity. The same is no less true for other crosses. 2. These nearly normal pure maternal hybrids are merely part of a graded series of types, the next grade showing a slight degree of abnormality and a less positive pure maternal character, and the next grade showing less normal development and positive evidence of paternal heredity. Other grades show progressively stronger paternal heredity and progressively less normal develop- ment. The question would be pertinent, then: if the most suc- cessful and apparently pure maternal hybrids are parthenogenetic, what is the explanation of the rest of the series? There are no sharp lines of demarkation between the normal and abnormal nor between the pure maternal and the mixed type of hybrid. If one results through the real codperation of the sperm cell in development, so do the others. 3. The traditional factual basis for the belief in the partheno- genetic nature of heterogenic hybrids is derived from the cyto- logical studies of cross-fertilized eggs of echinoderms. In those eases the usual situation in very wide crosses is one in which the sperm head enters the egg but takes no part in cleavage, remain- ing an inert lump in the egg cytoplasm. In Teleost hybrids the 410 H. H. NEWMAN situation is entirely different. All cytological studies of Teleost hybrids of suborder width show that the paternal chromosomes retain their specific size, shape, and number, and are distributed to the blastomeres nearly or quite normally during the cleavage period. That occasional irregularities in chromosomal distribu- tion oceur is seen in this and other types of hybrid in which there occur numerous eggs in which pronouncedly irregular cleavage occurs. Such embryos die in early stages of cleavage or before embryonic differentiation begins. It seems certain, therfore, that in all of those embryos that reach an advanced embryonic stage the chromosome distribution (both maternal and paternal) has been at least fairly normal. Moenkhaus found that the paternal chromosomes were still normal in their distribution in advanced embryonic stages in the Fundulus x menidia cross, a cross of the same width as that between Fundulus and the mackerel. If, therefore, the paternal heredity materials co6perate during the developmental process with the maternal, it seems logical to expect them to function in heredity. It is hardly conceivable that the chromosomes, which are living en- tities, should be growing, dividing, and being distributed from cell to cell in cleavage and later development, unless they are functioning, and if they are functioning, how else can they function than in the expression of the changes we speak of as differentiation and heredity? In view of these considerations, we must seek some other explanation of the pure maternal appearance of the most suc- cessful Teleost hybrids. It is my belief that the occasional occurrence of a hatching larva in such a cross as that between Fundulus and mackerel is to be interpreted as a case of early and complete recovery from an initial mhibiting agent. Though the foreign sperm fertilizes the egg and codperates with it in development and heredity, it must also introduce some materials that are inimical to normal development. This would seem only likely from what we know of the high degree of specificity of protoplasmic materials. In a few cases the egg materials are able either to transform, absorb, or in some other way to neutral- ize these foreign elements so that they have little or no harmful HYBRIDS BETWEEN FUNDULUS AND MACKEREL 411 effect upon the development. At the same time the specificity of the paternal chromatin seems to be neutralized and it is unable to cope with the unaltered maternal hereditary factors in the determination of chromatophore and other characters. Other hybrid eggs react a little less promptly or less completely to the foreign materials and are in consequence affected by them more or less strongly. Any agent inimical to development seems to express itself primarily as a retarding agent. As Child has shown, an inhibiting or retarding agent has the most pronounced effects upon the structures that have the highest rate of metabo- lism, usually the head, heart, and especially the eyes. Conse- quently, when the foreign materials are most active, as shown by the presence of many of the paternal type of chromatophores, the effects of retardation in development are most strikingly shown in pronounced abnormalities or early death and disinte- gration of embryos. In the average case, however, the inhibiting effect, though long continued, is not unduly severe, and there result large numbers of embryos with abnormal head parts of all grades of severity and fairly well-developed posterior structures. This type of result would be classed as the effects of differential inhibition. There occur, however, not infrequently embryos in which the head parts are better developed than are the posterior parts, or, in extreme cases, only isolated eyes or hearts occur on an other- wise undifferentiated blastoderm. These are, I believe, to be interpreted as the result of differential inhibition followed by differential recovery. According to Child’s theories, the apical or head parts, although most susceptible to retarding conditions, also have the greatest capacity for recovery when the inhibiting agent is removed or reduced in intensity. Evidently some embryos recover from the inimical effects of the foreign sperm, either by becoming acclimated or because the foreign element has grown weak, or perhaps been absorbed, and the anterior struc- tures take up the process of differentiation and develop principally eyes and adjacent head parts, the rest of the body remaining in the original retarded condition incapable of recovery. Thus we can account for trunkless heads and even isolated eyes and 412 H. H. NEWMAN hearts. Toa certain extent the occurrence of the paternal types of chromatophores bears out this interpretation, for in these recovery cases there are few paternal chromatophores and some- times none. When the hindering effect is lost, the hereditary effect is correspondingly lost or diminished. Although the paral- lelism between the occurrence in these hybrids of degree of ab- normality and abundance of green (paternal type) chromato- phores is not exact, it is certain that lack of green chromatophores is accompanied by nearly normal development and that whenever the green chromatophores oecur in abundance development is decidedly abnormal. The need of better criteria of hybridization What is a hybrid and by what criteria may a true hybrid be distinguished from a pseudohybrid? These questions have forced themselves upon my attention during the present investi- gations and demand an answer. Is it scientifically a valid pro- cedure to include as hybrids those egg and sperm unions between representatives of different phyla, such as those between echino- derms and molluscs, and those between different classes, such as Echinoidea and Asteroidea? Itis admitted by those who have studied this type of union that the sperm chromatin acts as an inert body within the egg and in no way, except as it hinders development, plays a réle in development and heredity. More- over, there is no initial attraction or reaction between egg and sperm in these cases, the union being brought about by the aid of chemical agents that serve to break down specific incom- patibility between these foreign materials. I would therefore suggest that these and all similar forced unions between diverse germinal products be called what they are, merely mechanical mixtures of foreign protoplasms. I would further suggest as a definition of hybridization, a natural union between diverse germ cells that involves a fertilization reaction between egg and sperm and in which the sperm chromatin shows evidences of codperation in cleavage and subsequent development. The acceptance of these distinctions would appear t6 clear the atmosphere of mis- HYBRIDS BETWEEN FUNDULUS AND MACKEREL 413 understandings. We shall all then be able to admit that these ‘mechanical mixtures,’ which some authors have called ‘hetero- geneous’ or heterogenic hybrids, are pure maternal and may be parthenogenetic as to the initiation of development. On the other hand, we must expect to find in true hybrids (i.e., all those in which the male germ cell shows morphological evi- dence of playing a rdle in development) an accompanying physio- logical effect upon development and heredity. Personally I feel that a morphological picture must represent an equivalent result of functioning and must presage further functioning. It may well be that the egg develops more normally when the activities of the foreign sperm are suppressed, as is evidenced by nearly normal and pure maternal larvae among Teleost hybrids, and that marked activity on the part of the sperm material ereatly retards and inhibits normal development; but this is only what we should expect in true hybrid combinations of consid- erable heterogeneity. In crosses between closely related species, however, the effect of only slightly diverse materials appears to act as a stimulus, and more rapid development and more vigor- ous offspring frequently result. It is interesting to know that there is every degree of gradation between the two extreme results in true hybridizations. In some extreme heterogenic crosses the functioning of the sperm is so vigorous as to retard normal development during cleavage so seriously that differenti- ation is wholly inhibited. At the other extreme are those cases in which even during cleavage the processes of development are accelerated and supernormal individuals result. All of these results are rightly to be included within the category of hybrid- ization phenomena. SUMMARY 1. A survey of the literature on Echinoid hybridization is attempted in order to gain an accurate idea of the relative avail- ability of Echinoids and Teleosts for the study of hybridization phenomena. It appears that by chemical means inseminations of Echinoid eggs with sperm of other orders, classes, and even phyla may be made, but these involve no real fertilization reac- 414 H. H. NEWMAN tions and are considered as mechanical mixtures. Real hybrid phenomena appear to be restricted to species within the order Diademoida of the class Echinoidea. 2. In fish hybridization the same situation appears. The hybrids known for fish are restricted within the confines of the order Teleostei. 3. Subordinal crosses in both Echinoids and Teleosts have been made, but the data for crosses of this width among Echinoids is contradictory and very unsatisfactory; hence the especial value of an intensive study of Teleost hybrids of this width of cross. Practically all Teleosts may be crossed without artificial chemical aids. 4. As an example of a heterogenic Teleost hybridization of suborder width, that between Fundulus heteroclitus and Scom- ber scombrus is chosen after an extensive survey of a wide range of crosses, because it is especially well adapted to show the facts about heterogenic hybridization in a clean-cut and unequivocal way. Other crosses nearly as good might have been chosen. 5. Although the adult characters of the two species are in striking contrast and the larvae at all stages are different in all details, the only reliable differentiating characters of the younger larval stages are the chromatophores. Fundulus has red chromat- ophores and the mackerel green ones; Fundulus has solid black chromatophores and the mackerel delicately branched black ones; these differentiating characters serve to indicate the degree to which the two parental elements function in development and heredity. 6. Cross fertilizations between the two species result differently according to which is used for the egg species. When Fundulus eggs are used, all grades of success in development up to hatched larvae are obtained. When mackerel eggs are used, all embryos die before or during. gastrulation, before any specific characters are differentiated. The study of heredity is therefore confined to one of the two reciprocal crosses, that between Fundulus ? and Scomber <&. 7. If the experiment is performed near the middle of June, when the breeding season of both species is at its height, one gets HYBRIDS BETWEEN FUNDULUS AND MACKEREL 415 results nearly uniformly as indicated in the typical experiment given in detail. Later in the season the results differ materially. In this paper our attention is confined to the results obtained in a number of crosses made about June 12 to 15. 8. The general conclusion derived from a study of the data are: in proportion as the paternal element vigorously exercises its functions, in like proportion is development retarded and the various types of monster appear. Likewise the paternal types of chromatophore tend to be more numerous the more pronounced the general abnormality. The most successful embryos are, in so far as the chromatophores are concerned, pure maternal, a fact that leads to the belief that these practically normal indi- viduals are not the product of parthenogenesis, but represent cases of complete recovery from an influence that in most cases retards development. It seems that the disharmonious elements of the paternal contribution are either completely neutralized or so modified as not to oppose the normal activities of the maternal materials. 9. The great majority of hybrids are subnormal, especially in their apical structures (eyes, hearts, ete.), and at the same time show obvious paternal heredity. The persistence of pater- nal types of chromatophores indicates that the paternal contribu- tion is still active and that complete recovery is therefore impos- sible. The embryos are largely to be classed as the results of permanent but nonlethal inhibiting agents, the effect of which is always, according to Child’s ‘axial gradient’ hypothesis, to pro- duce embryos with more abnormalities in the apical than in the basal parts. Hence we explain the wide range of ophthalmic and cardiac abnormalities in these hybrid groups as the result of differential inhibition. 10. A somewhat smaller group of hybrid embryos show large apical parts and reduced basal parts. Occasionally isolated apical parts (eyes and hearts) occur in an otherwise undifferenti- ated egg. Such embryos are usually pure maternal as to chro- matophores or at least in the region of differentiation. Such anomalies are to be interpreted as differential recovery products, but differ from the complete recovery cases in that recovery 416 H. H. NEWMAN occurs only after a prolonged inhibition. When too long in- hibited the whole organism cannot completely recover, but only the apical structures, whose capacity for recovery has been shown by Child to be greater than in basal structures. Thus we find heads differentiated without bodies, or with at best rudimentary bodies, and occasionally isolated eyes and hearts. 11. These experiments may be taken to indicate that in hetero- genic crosses no harmonious structural differentiation can result, without the neutralization or elimination of the disharmonious paternal materials. If the latter function actively, so as to ex- press this functioning in heredity, there can result only retarded and subnormal embryos and larvae, and the vast majority of heterogenic hybrids are of this type. HYBRIDS BETWEEN FUNDULUS AND MACKEREL 417 BIBLIOGRAPHY Appetuor, A. 1894 Uber einige Resultate der Kreuzungsbefruchtung bei Knochenfischen. Bergens Museums Aarbog, no. 1, pp. 1-19. Bancrort, F. W. 1912 Heredity of pigmentation in Fundulus hybrids. Jour. Exp. Zool., vol. 12, no. 2, pp. 153-178. Cuitp, C. M. 1915 Individuality in organisms. The University of Chicago Press. 1916 Experimental control and modification of larval development in the sea urchin in relation to axial gradients. Jour. Morph., vol. Xx¢iil, no. 1. Hertwic, G. unp P. 1914 Kreuzungsversuche an Knochenfischen. Arch. f. mikrosk. Anat., Bd. 84, Abt. 2, pp. 49-88. Lors, J. 1912 Heredity in heterogeneous hybrids. Jour. Morph., vol. 23, no. 1, pp. 1-16. MoenxkuHats, W.J. 1894 The development of hybrids between Fundulus hetero- clitus and Menidia notata, with especial reference to the behavior of maternal and paternal chromatin. Am. Jour. Anat., vol. 3, pp. 29-65, 196 Seal Cross fertilization among fishes. Proc. Indiana Acad. Sciences, pp. 353-393. /9/0 Morris, M. 1914 The behavior of chromatin in Teleost hybrids. Jour. Exp. Zool., vol. 16, no. 4. Newman, H.H. 1907 Spawning and behavior and sexual dimorphism in Fundu- lus heteroclitus and allied fishes. Biol. Bull., vol. 12, pp. 314-349. 1908 The process of heredity as exhibited by the development of Fundulus hybrids. Jour. Exp. Zodél., vol. 5, no. 4, pp. 503-511. 1910 Further studies of the process of heredity in Fundulus hybrids. Jour. Exp. Zo6l., vol. 8, pp. 1438-162. 1911 Reply to M. Godlewski’s Bemerkungen zu der Arbeit von H. H. Newman, ‘Further studies of the process of heredity in Fundulus hybrids.’”’ Arch. f. Entw.-mech., Bd. 32, pp. 472-476. 1914 Modes of inheritance in Teleost hybrids. Jour. Exp. Zodl., vol. 16, no. 4, pp. 447-499. 1915 Development and heredity in heterogenic Teleost hybrids. Jour. Exp. Zodél., vol. 18, no. 4, pp. 511-576. 1917 On the production of monsters by hybridization. Biol. Bull., vol, 32, no. 5, pp. 306-321. 1Qoy . PLATE 1 EXPLANATION OF FIGURES 1. A hybrid embryo (Fundulus 2 Scomber 0’) nearly three weeks old. It is nearly normal in structure and practically pure maternal in the chromatophore complex. Note the total absence of the green chromatophores of the paternal parent species, as contrasted with their presence in figures 2, 3, and 4. The egg is 2.5 mm. in diameter. 2 Asubnormal hybrid embryo (Fundulus? Scombero’), showing the anterior end of the body and the rather poorly differentiated eyes. On the large yolk mass note both the square black chromatophores of Fundulus and the branching black chromatophores of Scomber. Note also that the red chromatophores of Fundulus and the green ones of Scomber are about equally numerous. The drawing was made about two weeks after fertilization. (Drawn from life by Kenji Toda.) 418 HYBRIDS BETWEEN FUNDULUS AND MACKEREL PLATE 1 H. H. NEWMAN PLATE 2 EXPLANATION OF FIGURES 3 Asubnormal hybrid embryo (Fundulus 2 Seombero”), two weeks old, show- ing abnormal eyes and a rather pronounced ascendency of the paternal types of. chromatophores. Only afew of the red chromatophores of Fundulus occur, while the green chromatophores of the Mackerel are numerous. The whole embryo has a greenish cast that one never sees in a pure-bred Fundulus embryo. 4 IN echo 0 * * |v 9T LF 18d} ¥ TT 8°9 | 8-0 °9)¢8 &l + 699) ¥ an 9°L4 0 II-0'9 |6F IN $200°0 x he ||-08G OT + 69T| ¥ ty. 22S5) GF |8°S-0°5 Si L + T&|* | ASV | 89/9809 ($8 IN ¢z00'0 * x | GP 9°0 + 6F | ® Vy 998/82 8 8-029 OF 6L \* | G®V | FS) 8808 i IN $2000 erqo,ydoydeT * » | 98 SUP 6 FL LE | ¥ V 998 | 9% (0'S-E'S/Ol S14 EF FG 8h ¥ | “ASV | 9°8 0 TT09 ([6T |N 1000000 * x | 86 OT + 8hT| ¥ &T 9° /$°8-0'7/8 0% + LPS ¥ L 69 |)02¢'¢€ {8 |N 10000°0 * a> [neh b FZ | ® V 998 | 2°¢ (0°8-0 F6I or + eot|¥ | ASV |09]08-4'E [ZT IN 10000°0 * Brat ets Saat PERem Vv O2Se) Var (9-9-0708 05 8+ ecti*¥ | ‘ASV |69|/0609 (cr |N 10000/0 sajnum ‘OQ ‘aap UU WU | saynurue “2 ‘bap “UU “UU sni0sdy q uorstard | V woIstarq n n Ps) DB w 4 > 2 n mn 8 > 2 4 n oo Ho rao A aq re] > il g = aq 8 2 i g Op Bom | BR ic RS ae rs a I 4 < cas ry a PI = Ag moist er te: e.g a a 5 | i a 4 E ag 5 A 4 3 i EE ie By 8 Shs q a @ 5 4 Igg q w a SI 6 a5 | § (|g84 2 Bol 6 5 3 Bo] 8 S zg : a1 84 5 ge y B® 8 ws 1j0qnjaU fo a}D4 1aM0) AY} fippDIYWa109Y4) BADY P]NOYs JY] S]DULUD UO s7z]NSat s72GQ2YxA 27qQD} ay? fo P UoIlsuany “viqajydojdaT fo ws20 -qnjau fo aD ay} sainspaut fijj094~p “N F800'O 42Y72YN 07 SD OS]P “70 Saunspaw fiq}0a4YpUr *N 100000 0 puv sydwfiu snioad yz fo ajp4 ayoqnjaw ay) saunspaw fiqj2a4~p apruvfio wnisspjod fo suoYNzOS JDUW.LOU TQO00'O 49Y72YM 0} Sv s}8a} fo synsau ay) Burnoyg T WIaVL LIGHT REACTIONS—-METABOLISM—MAY-FLY NYMPHS 429 DOES RESISTANCE TO POTASSIUM CYANIDE MEASURE THE RATE OF METABOLISM IN MAY-FLY NYMPHS? In this work the cyanide resistance method of Child (’13) was used without modification except in the strength of the cyanide solutions employed. In this method with a relatively high concentration the animals with a higher rate of metabolism (cf. Geppert, 99, Hyman, ’16) die before those with a lower rate. In much weaker solutions acclimatization occurs and the effect is reversed. The resistance of the nymphs was tested in 500-ce. Erlenmeyer flasks. Control experiments showed that the more sensitive Epeorus nymphs could live from five to fifteen days in this amount of unchanged tap water, provided the temperature was approximately constant. Some difficulties were encountered in determining the death point. The nymphs were usually observed every thirty minutes in later experiments every twenty minutes), and those that were apparently motionless were removed with a large pipette and care- fully inspected under a lens. If no motion was apparent they were stimulated with a needle. If they failed to show any mo- tion under these conditions they were considered dead. It is obvious that this treatment would stimulate live nymphs and more frequent inspection would increase rather than decrease the experimental error. Epeorus proved to be much less resistant to the cyanide than Leptophlebia. With the former (table 1) a solution 0.00001 normal gave a measurement of metabolism by the ‘direct method; with the latter the same measurement was obtained by a strength of 0.0025 normal. With Epeorus 0.000001 normal was found to measure metabolism by the acclimatization method. With this dilution so great care was necessary in order to maintain the solution at even its approximate strength that no serious tests were run. A new solution of 0.1 normal potassium cyanide was made up weekly, and from this dilutions were made to the desired strength. With extreme dilutions and with certain preliminary experiments which ran several days the solutions were made up fresh twice daily. 430 W. C. ALLEE AND E. R. STEIN, JR. The inquiry as to whether resistance to potassium cyanide in May-fly nymphs is affected by the rate of metabolism of the ani- mal was prosecuted along the following lines: 1) What is the relation between the resistance of large (old) and small (young) nymphs? 2) What is the effect of stimulation upon the resistance to the cyanide? 3) What is the effect of differences in temperature? The results of these experiments are summarized in table 1. Nymphs of Epeorus were less resistant to 0.00001 normal potas- sium cyanide when they were small, or stimulated, or at increased temperature. All of these conditions cause a higher rate of metabolism (Child, ’13; Allee, 714). On the other hand, in a 0.000001 normal solution the smaller (younger) nymphs were more resistant than the larger ones, which is what the theory demands if a solution of this strength to measure indirectly the rate of metabolic processes. Leptophlebia in 0.0025 normal solution was less resistant when young, or stimulated, or when the temperature was increased so that this strength of cyanide directly measures the metabolic rate of these nymphs. A solution 0.00025 did not indirectly measure the rate of metabolic processes of these nymphs and the experiments were not continued long enough to find a solu- tion strength that would do so. In all the above instances in which the evidence is that cyanide resistance does measure the rate of metabolism the differences in the survival times exceeds twice the probable error and we may safely hold them to be statistically significant. RELATION BETWEEN THE SIGN OF LIGHT REACTION AND RESIST- ANCE TO THE CYANIDE 1. Epeorus The average survival time of fifty-one positive Epeorus nymphs (table 2) which had been taken directly from tap water was 108 + 5 minutes. These nymphs averaged 5.4 mm. long. A reversal of the phototactic reaction of forty-seven other posi- LIGHT REACTIONS—METABOLISM—-MAY-FLY NYMPHS 431 tive Epeorus nymphs was caused by treatment with alcohol (1 or 2 per cent) and decreased temperature (2° to 8°C.). These reversed nymphs gave a survival time of 131 + 6 minutes in the same strength of cyanide. Their size average was 6.5 mm. The difference in survival time is only 2.1 times the probable error, and since the difference may have been affected by the difference in size, too much emphasis cannot be laid on these results. The survival time of eighteen nymphs of the same species that failed to reverse in the alcohol-reduced temperature treat- ment was 88 + 3 minutes. Their average length was 6.1 mm. The nymphs that were reversed by this treatment lived forty- three minutes longer in the cyanide than those that remained TABLE 2 Showing the relation between the sign of the phototactic reaction of Epeorus nymphs and their resistance to potassium cyanide AVERAGE SURVIVAL TIME AVERAGE SURVIVAL TIME NUMBER TESTED NUMBER TESTED AVERAGE LENGTH AVERAGE TREATMENT BEFORE KILLING LENGTH Positive nymphs Negative nymphs mm. minutes mm. minutes 51 5.4 108 + 5 Tap water 18 6.1 88 + 3 Alcohol 47 6.5 |131+ 6 Lowered temperature 11 6.1 |160+16 positive. Since this is 4.8 times the probable error, it must be significant. The majority of the Epeorus nymphs treated with alcohol and decreased temperature reversed their reaction to light and had a slightly (perhaps questionably) lower rate of metabolism than the control animals and a decidedly lower rate than those not re- versed by the treatment. The nymphs that remained positive, although given the alcohol-low temperature treatment, appar- ently had been stimulated by the alcohol, for they showed a higher rate of metabolism than the control animals. This can be explained by assuming that the alcohol acted as usual, first stimulating and later depressing. If this be true, the fact that aleohol plus reduced temperature was more effective than the 432 W. C. ALLEE AND E. R. STEIN, JR. latter alone becomes significant if the metabolic differences prove to be causal rather than incidental or resultant, for the difference between a stimulated period followed by depression may well be greater than mere depression. Sometimes quantitative reversals were obtained by reducing the temperature. Eleven nymphs that had been thus reversed (average length 6.1 mm.) were killed in cyanide and gave an average resistance of 160 + 16 minutes. This is fifty-two min- utes longer than that given by the control nymphs which is 2.5 times the probable error and is probably significant. Taken altogether, the evidence here presented indicates that reversed Epeorus nymphs have a lower rate of metabolic activity than do positive animals. One other observation confirms this idea. Epeorus nymphs collected in October were kept in a large aquarium that also contained some fresh-water mussels. In December and January the nymphs were found to be dying in large numbers. When tested all were negative to light, although when first collected they had given almost quantitatively posi- tive reactions. Obviously the metabolic process of the nymphs was strongly retarded and this is correlated with their reversal to light. 2. Leptophlebia Leptophlebia nymphs were usually negative in their reaction to light giving a ratio of twelve negative to five positive animals. The average survival time of sixty-one untreated nymphs (average length 7.9 mm.) that gave the usual negative light reaction was 131 minutes. Forty-two positive nymphs (aver- age length 7.8 mm.) under conditions similar in every way re- sisted the same strength of cyanide 130 minutes. Thus there was no difference in the metabolic condition of these two groups that could be measured by the cyanide resistance method. Hydrochloric acid was very effective in causing reversals in Leptophlebia. The survival time of thirty-six nymphs so reversed (average length 7.7 mm.) was 127 minutes. Twenty nymphs similarly treated that remained negative (average length 7.8 mm.) gave a mean survival time of 120 minutes. This LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 433 difference is of course negligible as is also the difference between these acid-treated animals and the control. Aleohol was also effective in making these negative nymphs positive. Fourteen nymphs that were made positive gave a resistance of 91 + 7 minutes to the cyanide as compared with 130 + 4 minutes for the sixty-one control animals. This is 3.5 times the probable error. The eleven nymphs that were treated with aleohol and remained negative showed some stimulation TABLE 3 Showing the sign of phototactic reaction of Leptophlebia nymphs and their resistance to cyanide NUMBER | AVERAGE SURVIVAL maratuest suronn xiiiwa | YOUBET | AVSEAGE | sunvivat Positive nymphs Negative nymphs mm. minutes mm. minutes 36 Gad 127 HCl 20 7.8 120 14 7.8 9147 Alcohol 2 per cent 11 7.7 |109+9 9 7.8 119 H2SO, 8 7.8 115 Acetic acid 2 7.0 136 a 8.5 127 NaOH 8 7.8 146 8 6.9 132 KOH 2 7.0 130 9 6.8 133 KCN 2 7.4 103 3 8.3 115 KCl 2 8.0 103 6 8.7 100 NaCl 4 8.7 105 MgCl 2 8.0 129 2 7.8 129 CaCl, 3 7.8 123 6 7.8 161 Chloretone 8 8.1 165 108 7.8 123 Totals and averages 64 7.8 129 42 7.9 | 13144] Tap water control 61 7.8 |130+4 when tested with cyanide, but not so much as the positive nymphs. Considering the numbers tested and the small deviation from the resistance shown by the control animals, none of the results obtained with other reagents and listed in table 3 are significant with the exception of those with chloretone. This drug was not very efficient in causing reversals, but did cause decided de- pression, as measured by the cyanide method, and caused re- versals in about 30 per cent of the nymphs treated. 434 W. C. ALLEE AND E. R. STEIN, JR. The average effect of all these reagents upon the resistance to cyanide, if such is worth anything, shows no marked difference between positive and negative treated animals nor between the treated animals and the control. From these experiments with Leptophlebia we have the inter- esting results that these nymphs were reversed without affecting their resistance to potassium cyanide (HCI and averaged results) ; with accompanying stimulation (alcohol) and with accompanying depression (chloretone). The quantitative reversal of positive Epeorus nymphs and negative Leptophlebia by 1 per cent alcohol in the same ex- perimental dish at the same time was repeatedly demonstrated. Thus conditions absolutely identical caused opposite reversals in the two species. At first sight this would appear to mean that both positive and negative animals were reversed for the same reason. This is not necessarily true. In the tests to find the strength of potassium cyanide that would directly measure the metabolic rate of the two species it was found that Leptophlebia- was only one-fourth as sensitive as Epeorus. Since the Lepto- phlebia are much more resistant, a strength of alcohol that only stimulated them may have clearly depressed Epeorus. That such is the true explanation is indicated by the effect of alcohol on the resistance of the two species of nymphs to cyanide. The Epeorus that had been made negative were found to be depressed, while the Leptophlebia that were made positive were clearly stimulated. II. REVERSALS OF PHOTOTAXIS AND CARBON DIOXIDE PRODUCTION? ® METHODS The experiments upon which the second part of this report is based were carried on at Lake Forest College upon a May-fly 2 This section is based on experiments by the senior author, now being continued, which were started in the spring of 1916. They were made possible by money grants from the Elizabeth Thompson Fund and from the Bache Fund of the National Academy. 8 The nymph whose reactions are described in the sccond part of this paper is Heptagenia pulchella Walsh. We are indebted to Professor J. G. Needham for this identification. LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 4395 nymph belonging to the Heptageninae. These nymphs are quite common in Pettibone Creek (Shelford, ’13, maps) where they are usually found in the stones between rifles. Pettibone Creek is a brook about the same size as the Berkshire streams mentioned in the preceding part, but with much less rapid current. The nymphs were kept in the laboratory for long intervals during the winter in well aerated running-water aquaria. Ani- mals to be experimented on were transferred to room-tempera- ture aquaria aerated by means of an air-pressure pump operated by water pressure. The experiments were conducted as at Williamstown save that a daylight, concentrated filament, 100-watt Mazda (C2 of the General Electrical Company) was used as a source light. The experimental box was placed in a darkened room so that only light from the source lamp could enter during the experiment. In most of this work an assistant plotted the reactions of individual nymphs to light and manipulated their change to experimental solutions. Careful controls were run. When the nymphs were to be changed to a new experimental solution, the controls were changed in the same manner to fresh tap water. The assistant also prepared the pairs of nymphs for their carbon dioxide test in such a way that the experimenter had no idea which was the experimental and which the control nymph. The carbon dioxide production was determined in Tashiro’s biometer (Tashiro, 713, ’17) as follows: The assistant placed two nymphs, whose rate of carbon dioxide production was to be com- pared, momentarily on filter paper and then transferred each to ‘a shallow glass cell. The nymphs used were of the same size and were selected so that the experimental factor was the only known cause for varia- tion in their rate of carbon dioxide production. The containers were marked for future identification. These were handed to the experimenter and immediately placed in the apparatus. Within five minutes from their removal from the experimental dishes one could get an indication of their relative rate of carbon dioxide production. Under optimum conditions the entire process of determination could be repeated at the rate of three per hour. ‘This was much in excess of the usual rate. s 436 W. C. ALLEE AND E. R. STEIN, JR. TABLE 4 9/1/16. 1:20 p.m. Put two lots of eight May-fly nymphs each in aquarium water in dark box. Light: 60-watt Mazda, 50 cm. distant. The nymphs had been in the laboratory two days. Temperature aquarium 21.5; of room 22. Lot 1 LoT 2 a | ale | (Sv) bo i — wo —_ bo on or is 100) —_ rr Oo — a W bo Se i ha —y | oN NO et — — UTIAMOWUMUWHNMIWAA an a»n ® @ NONNNNNNNNNNN HK KE (=) oO — | Os See dD NR RK ee DO NINN MDWMWANNNOWOD 218 Le &* * Positive one has been positive throughout put in biometer 2:20 with a negative one that has been negative all the time. Washed with CO, free air 5 minutes. Positive nymph in left chamber, negative in right. Bubble 2: 25. First ppt. Rt. 2: 28. Much more left 2: 38. Took out biometer, 2: 40. Both active; put back in dark box in separate dishes. 3:00 Both reacting as before testing in biometer. 3:20 Do. ono. Dos 4:10 Do. 4:15 Replaced in biometer as before. 4:25 Bubble. 4:27 First left, positive. 4:40 More on left as before. 4:45 Out of apparatus. Reintroduced at negative end of separate dishes. 5:00 Nymph that had been positive throughout, still positive; other nega- tive as before. 6:00 Both negative. 7:40 Nymph negative throughout now positive, other negative. Light on all night. 8:00 a.m. Both negative. LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 437 Some idea of the nature of experimentation together with its effect on the nymphs may be gained from table 4. This table is a slightly expanded copy of a portion of the laboratory record for the day. It shows that during almost four hours of observation, in which time three biometer tests were made, the nymphs maintained their original light reaction and that each test showed the positive nymph had the higher rate of carbon dioxide production. Whatever the faults of the method used, it at least has the merit of being always comparative. As was to be expected, large nymphs gave off carbon dioxide more rapidly than small ones and active nymphs more rapidly than inactive ones. These nymphs are strongly thigmotactic, and this usually caused them to remain quietly in their container. Occasionally one would move. Such determinations were of course thrown out. PHOTOTAXIS AND CARBON DIOXIDE PRODUCTION IN UNTREATED NYMPHS These nymphs, like Leptophlebia, are usually negative to light. This normal reaction to light is graphically shown in chart 1 together with the preliminary and control records shown in charts 2, 3 and 4. About 20 per cent of the untreated nymphs were positive to light. When these were tested they were found to have a higher rate of carbon dioxide production than the negative nymphs. Lad oe Their N ratio, based on 332 non-selected control readings was 5.5 This is indicated by the graphs in columns | and 2 of chart 1 and the details are given in table 5. Exposure to light for considerable time sometimes caused negative nymphs to reverse their light reaction and become positive. Such animals were found to be more stimulated, as determined by the rate of carbon dioxide production, than simi- lar nymphs that had not been exposed to light (table 6). This is just the opposite to the result obtained by MacCurdy with nega- tive starfish which he found gave off less carbon dioxide when exposed to strong light. PC OO lasrotocreebreere tains = S i} = 6 © @ ~ o o - e a a 8 S 8 x 2 8 8 8 3 5 : MILI UI CCC PL PB OL COL OL COL CL LLY LC TO 1 2 + 5/22/17_| 4. 5/22/17 — i « - % & =] Ss © co “ UT PLL LL LL OO = Joo eee B p 3 = 5 s Es TO CY PO 8 8 More COs 8 8 8 & Bs 2 nobel tiled tiebereretiree tiers 3 Chart + 5/19/17 Water oun “1th ‘+ “uta oO p97} wo Water Pd44two "uTW ST "50' omitted. 438 1 4 5/19/17 — ' Poslisbinboebeee bebe lartuidicriebirin botrloprelieebiebee bebe ebelchabederebeboe tebe tere i cr ~ o a - © COUTURE PLLOCOCI PLO. LOLOL SLC LOLOL LLL LL PB Ye OL = B 3 8 8 LY} 8 a) 2 8 ULC LCOS LA UOC LACIE eS CC CY MI COg test before exposure bibeelauiny B Harlot adobe beer bcbebeern lvvevelevereloeroeberertins 8 ‘ £ w 8 8 3 Ps 3 ee a i) 3 5 ro} = © a a oa = a “ a ra POO UOC OCU GUC UU AGU IU UO )OUCG OOo UU cn inc cc cc i Cc PC A LIGHT REACTIONS—METABOLISM—-MAY-FLY NYMPHS 409 TABLE 5 Showing the comparative rate of carbon dioxide production of positive and negative - nymphs under control conditions (p. 625) SIZE DATE opi 14: SI PE A MORE CO2 + = mm. mm, 9/ 5/16 8.0 8.0 | Positive 9/ 1/16 6.0 6.5 | Positive 6.0 6.5 | Positive 9/ 2/16 5 6.0 | Positive Hoe 5.5 | Positive 10/26/16 7.0 7.0 | Positive 10/27/16 Use 8.0 | Positive 10/28/17 eal 7.0 | Positive 7.0 7.0 | Negative. More active before test 8.0 8.0 | Negative. More active before test 11/ 2/16 6.5 7.0 | Positive 3/12/17 115 12.5 | Positive eS 12.5 | Positive 3/13/17 15 11.0 | Positive 11.0 11.0 | Positive MO 11.0 | Positive 11.0 11.0 | Positive 11.0 11.0 | Positive. Not marked. Demonstrated 3/15/17 10.0 10.0 | Positive 3/16/17 10.0 10.0 | Positive 3/20/17 11.0 11.0 | Positive 3/21/17 11.0 11.0 | Positive 3/22/17 10.0 10.0 | Positive Number tested 23 pairs. Positive more, 21. Negative more, 2. Chart 1 Showing graphically the reaction of six untreated nymphs to light from a 100-watt daylight Mazda (C2) placed 50 em. from the experimental dishes. The charting was done by an assistant from whose records these graphs were copied. The scales show time in minutes. The left-hand side represents the positive end of the dish, i.e., the end toward the light. Where the line is approx- imately straight vertically the nymph was resting quietly. Curves and kinks show movement which did not markedly change the position in the dish in respect to light. Column 1 gives the reactions of a nymph that was predominantly positive to light, which after twenty-three minutes’ exposure gave more carbon dioxide in the biometer than the negative nymph whose reactions are recorded in column 2. Column 3 shows a nymph made positive by long exposure to light. In this case the reversal came suddenly. Column 4 gives the same result ob- tained by a different method. Perhaps column 3 represents a ‘tropic’ and column 4 a ‘trial’ reaction. Columns 5 and 6 show the reactions of two nymphs that were tested in the biometer before exposure to light. The animal with the higher rate of carbon dioxide production became positive. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 26, NO. 3 440 W. C. ALLEE AND E. R. STEIN, JR. TABLE 6 Showing the effect of long exposure to light upon carbon dioxide production SIZE TEMPERATURE MINUTES MORE CARBON EXPOSED DIOXIDE Exposed Unexposed Exposed Unexposed mm. mm, deg C. deg C. 13 13 21 19 120 Exposed 13 14 21 19 150 Exposed 10 10 23 21 204 Exposed 10 10 23 21 234 Exposed 10 10 15 13 52 Exposed EXPERIMENTS WITH HYDROCHLORIC ACID 1. Effect on negative nymphs As with the other species studied, hydrochloric acid was one of the most effective reagents in causing reversals. In one set of carefully controlled, fully plotted experiments there were a total of 220 control or preliminary readings lasting from fifteen to ninety-three minutes. Of the nymphs thus tested, thirty-seven were or became positive without other treatment than exposure to light. This is 17 per cent of the number tested. In this same series of experiments 125 nymphs that had been consistently negative through a preliminary testing period of at least fifteen minutes, were treated with N/25 hydrochloric acid. Under this treatment, seventy-five nymphs, 60 per cent, became positive. Of the fifty nymphs that did not reverse, twenty-five were kept under observation for less than twenty-five minutes and only two were kept until they died. If it had been the pur- pose of the experiments to ascertain how many nymphs could be reversed by the treatment, doubtless about 90 per cent would have become positive before death resulted. The typical effect of the acid upon the light reactions of these nymphs is shown in chart 2. The graphs show that the nymphs may reverse soon after being placed in the acid or the reversal may come only after long exposure. Forty-eight per cent of the reversals came within fifteen minutes, but reversals occurred after seventy minutes’ treatment, Death frequently followed close upon the later type of reversals. LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 441 The effect of this treatment upon the carbon dioxide produc- tion is shown in table 7, which lists all the determinations, and in table 8, which partially analyzes the results listed in the pre- ceding table. The biometer tests show that the nymphs were stimulated when first put into the acid and that this period of stimulation lasted approximately fifteen minutes. The time limits varied with different individuals. After this period of stimulation the nymphs were depressed. This gives two periods in the carbon dioxide production corresponding to the two periods in the re- versals by this strength of the acid. The tables show that all nymphs were not tested immediately after reversal, but of those whose carbon dioxide production was found within two minutes after reversal and which had been stimulated by the acid the average time of treatment was 13.2 minutes. Thirteen of the twenty-five nymphs so tested had reversed within ten minutes after being placed in the hydrochloric acid. On the other hand, the average time of treatment of the animals similarly reversed and measured, but giving more car- bon dioxide in the control than in the acid, was twenty-seven minutes, and seven of the seventeen nymphs reversed after twenty-seven to thirty-seven minutes’ treatment. From these experiments it appears that either a marked increase or a decrease may accompany phototactic reversals of these nymphs when treated with hydrochloric acid. 2. Effect on positive nymphs About 20 per cent of the Heptageninae tested were positive to light when first tested or became positive under the influence of exposure to light for a short time. This positive reaction is much less stable than the usual negative reaction. Of fifteen carefully plotted tests lasting from 15 to 114 minutes, seven, or 47 per cent, showed a change to the usual negative reaction without treatment. In nineteen tests lasting from 1 to 68 minutes seventeen of the nymphs, 89 per cent, were made negative by N/25 hydrochloric acid. Most of these became negative within the first five min- utes of treatment and, as was to be expected from the preceding ry - o = Ss TY CY 7 ~ S B Vevserdervevdeveuaborrsebettritoecttrettdidboatieloibtidbeleo ddd 8 8 i} | TORUS ROLL OL OO 8 £ 8 8 8 prunnyeeny TAT Piru 1 + 6/15/17 Water | Changed to u/25 HCl eo CO2 N/25 HCl 5/26/17 2 6/15/17 — Water Changed to water 5/26/17 Vvrrboedembeteberadobie bbb Water TRULORU TDR ERER TORE RRR NEN nn An 8 Chart 2 3 4 + 5/18/17 -+5/18/17 — - e ~ o ; MEE CY RR CY RS = & SOOO COO (COC Cn (OO RON OC cc Is ‘SJ CUIPEL PEPPER, PEEP EPP UEE EEE EP EP TPT LPP ee ep Peep eee yee . F Water Changed to wate NN r Water ~ cy co o ~ © Changed to w/25 HCl trrertetletieant earn iumtcry lnenatrlarevall cninaaa (areata ei MRC CST MMRRT Accrual attain Gawciaivencketer ania emergent nnn aR R 8 x & TENEOYODAAENRNSAEUG (HORUS VCOSTD RES UV ESM AAS SASL TNS USAST SUTRA YUASA USTS SUA ARING(SUDTOTRTA AIST TSTSUSTENSRSUS UATE] USUETOTORAYORUSUGASS ETAT TSAI MSTA TCU T DELS Pe LPL oO 8 briploptrtboedieeebeeiin a More COp 5 + s/es/i7 Acid o o ~~ o a - we i) - UU OIL ULI PUI WL PB BO 5 = 9 1/2 minutes omitted g g LS = 3 is & 3° S & S a pemen = SOU UT COUT URC TULL PRC CCL LULL PLLA OC I PCO I CLO OC CPL BO A DO BO 8 8 19‘omitted Less COs surtvileteisticveetoverelieitireeliertitebrectodibebtt be bee More C02 LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 443 experiments, they were found to be stimulated by the treatment (table 9). Two instances of the reversal of positive animals by acid treat- ment are given in the first two columns of chart 4, p. 450. In both instances shown the reversals are typical in that they are very marked and occur almost directly upon the start of acid treatment. In the second instance shown the experimental dish was twice turned end for end and each time the nymph moved directly negative. Here we have the same reagent making negative animals posi- tive by either stimulating or depressing them and making positive animals of the same species negative by stimulating them. 3. Negative vs. positive nymphs; both treated with HCl The treatment with hydrochloric acid did not cause all the nymphs to reverse their light reactions. When animals that had been made positive were compared with those still negative, it was found (table 10) that the former had a higher rate of carbon dioxide production when the time of exposure had been approxi- mately the same for both nymphs compared. Almost all the nymphs tested came from the preliminary period of stimulation. Theoretically, nymphs long exposed to acid and positive would give less carbon dioxide than those more recently placed in the acid and still negative. This possibility was not tested, since the results can easily be interpolated from the tests given in pre- vious tables. Chart 2 Showing graphically the light reactions of four nymphs treated with N/25 HCl and their carbon dioxide production as compared with that of their control nymphs. Column 1 shows two reversals with the acid treatment, the first of which occurred within three minutes. The reversals shown in column 1 and that in column 4 were accompanied by stimulation. On the other hand, the reversal shown in column 5, which came after twenty-three minutes’ exposure to the acid and which was allowed to remain positive for thirty-five minutes before testing, showed depression. The preliminary test in columns 5 and 6 is not charted, but was essentially like the first eight minutes shown. It will be noted that in this test the control became positive due to exposure to light and was therefore more stimulated than the ordinary negative contrel. Other tests show depression by treatment with acid for this length of time when com- pared with negative nymphs. 444 W. C. ALLEE AND E. R. STEIN, JR. TABLE 7 Showing the effect of N/25 hydrochloric acid upon carbon dioxide production, All the nymphs here listed were decidedly negative in their light reaction before being placed in the acid SIZE SIGN OP LIGHT REACTION Lj) Setenwiets : MORE CO: NACL iN Read hilar + _ HCl Control AND TEST 13 13 + Acid 2 il 13 13 + oe Acid 2 1 11 11 + _ Acid 2 1 13 13 + = Acid 3} i 13 13 -e — Acid 5 2, 13 13 -- _ Acid 5 D, 12 12 a - Water 5 2 11 11 ae — Water 6 1 10 10 + om Acid 6 1 10 10 + = Acid 6 1 11 11 a — Acid 6 1 12 1? + a Acid 7 -] 10 10 _ Acid a 1 11 iil oo — Acid 8 1 10 10 - Acid 10 1 10 10 ae — Water 10 ii 11 iil — — Acid 1) 1 11 11 + ~ Acid 12 3 12 12 Acid 1 1 13 13 + -- Water 12 1 13 13 + = Acid 13 2 13 13 _ a Acid 14 2 11 ital an _ Water 14 4 15 15 ao — Water 15 15 13 13 - Acid 16 1 14 14 4- os Water il7/ 1 13 13 a a Acid 17 1 13 13 aad = Water — 17 1 13 13 + — Water 18 1 13 13 - — Acid 18 1 12 194 + — Water 19 1 11 11 of. — Water 20 12 ilal ili! -- — Acid alt 0 12 i aa + Water 22 il 14 14 — Acid 73) 1 11 11 — — Acid 433 0 14 14 ao _ Water 28 1 13 13 “fe — Water 28 5 10 10 + Water 28 di LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 445 TABLE 7—Concluded MINUTES SIZE SIGN OF LIGHT REACTION pene Co. MINUTES IN BETWEEN of - HCl Control AND TEST 10 10 _— -- Water 29 0 11 11 + — Acid 30 5 11 11 +- _ Water 31 6 13 13 -b = Water 33 1 10 10 _ -— Water 33 0 12 12 +: — Water 33 1 12 12 + _ Water 35 1 12 12 4- = Water 37 1 15 13 + -- Water 37 1 13 13 + _ Water 38 38 13 13 + — Water 41 1 ite 13 -b — Water 41 14 11 11 + _— Water 42 33 13 i183 + - Water 44 1 11 11 _ _ Acid 44 0 10 10 — - Acid 48 0 13 13 + = Water 51 14 ie 13 + - Water 58 3 13 13 + — Water 59 a 11 11 + _ Water 67 0 13 13 + + Water 70 30 11 10 + a Water 78 10 TABLE 8 Showing relative carbon dioxide production of experimental and control nymphs analyzed on the basis of length of exposure to hydrochloric acid. This table is based on table 7 MORE CARBON DIOXIDE MINUTES IN HCl HCl | Water Part I. Showing all tests listed in table 7 1- 5 6 1 6-10 8 2 11-15 5 3 16-20 2 5 21-25 3 1 26-30 1 4 31-40 0 7 41-50 2 4 51-60 0 3 61-70 0 2 71-80 0 1 TABLE 8—Concluded MORE CARBON DIOXIDE MINUTES IN HCl HCl | Water Part Il. Showing results of the biometer tests made within two minutes after reversal I- 5 6 1 6-10 i 1 11-15 5 1 16-20 PE 3 21-25 3 1 26-30 1 2 31-40 0 5 41-50 1 2 67 0 1 TABLE 9 Showing the effect on carbon dioxide production of treating positive nymphs with N/25 hydrochloric acid until they became negative SIZE SIGN OF LIGHT REACTION REOSN II MOI ila ae er i ae Ei ei te a ea HCl Control HCl Control AND TEST 13 13 _ _ Water 2 2 12 12 - ~ Acid 3 3 13 13 _ _ Acid 3 2 13 13 _— — Acid 3 3 10 10 — + Acid 3 2 12 12 — = Acid 4 1 13 13 _ _ Acid 4 4 14 14 = —- Acid 9 i TABLE 10 Showing the relative carbon dioxide production of negative nymphs made posilive by treatment with N/25 hydrochloric acid, compared with those similarly treated but not reversed 1 OUEST Ob TON SER « wvccnia whawaaimeer 0! oy Ale 0 ee eae MINUTES IN HCl MINUTES weknGo, OMe Ee = a ak . - TEST 13 13 + 2 2 1 14 14 oe 4 4 2 12 12 + 3 3 1 14 14 + 10 10 1 12 i, 4. 13 13 13 14 13 + 14 14 3 12 2 + 15 15 3 12 12 ok 15 15 6 11 11 > 25 25 2 11 11 Both same 32 29 3 11 11 — tH 21 2 13 13 + 15 115% 2 446 LIGHT REACTIONS—-METABOLISM—MAY-FLY NYMPHS 447 EXPERIMENTS WITH POTASSIUM CYANIDE In the experiments with the negative Leptophlebia potassium cyanide had proved an effective reagent in causing reversals. With these Heptageninae at a concentration of N/500 it was almost as effective in causing reversals as hydrochloric acid. Of thirty tests run, 57 per cent were reversed by the cyanide while the controls showed no reversals at all. Biometer tests proved that the cyanide clearly depressed the nymphs. The details are listed in table 11 and specimen results are shown in chart 3. In all cases the nymphs were washed in water after treating with KCN in order to keep the potassium from in- terfering with the CO, determination. In addition to the carbon dioxide tests, one nymph died while moving positive, another just after reaching the positive end. In all ten nymphs died in the experimental dishes as a result of the cyanide treatment, of these six had reversed their reaction to light shortly before dying. TABLE 11 Showing the effect of N/500 potassium cyanide upon nymphs negative to light and upon their rate of carbon dioxide production SIZE SIGN OF LIGHT REACTION SEN OME SIS nop Co Moore rd STE KCN Water KCN Water AND TEST 13 13 + Water 3 2 1) 12 - = Water 4.5 1 14 14 + Water 6 3 13 13 aa _ Water i 2 12 12 + — Water 8 2 13 13 ++ _ Water 6 1 118° 13 + - Water 9 3 13 13 + — Water 14 4 13 13 t- — Water 16 5 12 4 + — Water 19 2 12 12 + — Water 41 1 Poorer eee eee eee beeebee et e o - ts n ~ “ 8 8 y g as 2 8 B 8 3 > ) i] a a = & B & Ss © CL LCCC CO PL UU WLLL) LL LU Le aL LO LoL A LO LOL LP Ce Chart 3 3 + 6/10/17— Water iL 2 +6/10/17 +6/10/17 — Water Water & 5 i] Ss © co ~ a oe - wo ns “ — = a Changed Changed Changed to KCN to KCN to KCN & 2 8 3S & S a 8 \ Hitec bee tbe beberle bebe beccebeer tera S 8 (4 £ SO RUS ULC UU LL Ln PC LC DC URL PULL UL LLL LL LL 8 8 Deal 15! Les later Dead BCee 448 4 6/10/17 — Water Changed to water More C02 o - « w _ a ribtirbobiliseldceebiecbeeeeetoee bere & Ss S a = & & i=} S o of TO . > 3 8 8 8 a S 8 8 8 milter bei bie pooettaniebrreerloreeetes TO é 5 + 6/10/17 KCN Less COo 6/11/17 KCN 6/11/17 KCN 6/12/17 KCN Less CO than costro2 6 6/12/17 — KCK Less CO5 6/10/17 KCH Died withi 5 minutes 6/10/17 KCN ritirrsbecibreteebcledeetereerbetieadicers retplovperteseertiereboreteereedia 8 pipslrreliviebissebecrbiselicee eet Pa tesreetiprertaaagada TOO Cc YC ee = & ] a & = i] i=] = 5 . © ~ a a - e n - eT Ty eT '8 8 g 8 8 ae 2: 8 B re 3 LIGHT _REACTIONS—METABOLISM—MAY-FLY NYMPHS .449 EXPERIMENTS WITH ALCOHOL AND STRYCHNINE Ethyl] alcohol had less marked effect in causing reversals among these Heptageninae than with the species used in the first part of the work. In twenty-five well-studied cases, only five nymphs were by treatment with 2 per cent solution reversed. Of the fifty-two accompanying preliminary or control readings, one nymph became positive. When alcoholic nymphs were tested in the biometer (table 14), it was found that alcohol acts here according to its usual effect as a narcotic, first stimulating and later depressing. Some typical results obtained with the alcohol treatment are given in chart 4. This shows two tracings of the reactions of nymphs that were not reversed by the treatment and of one that was, together with a type of control reaction which occurred frequently. Some preliminary experiments with strychnine showed a marked increase in the activity of the nymphs without a reversal of their light response. The animals usually remained at. the negative end of the dish in almost continual motion. In the biometer, they usually showed a markedly increased carbon dioxide production. I am not able to state why these nymphs did not reverse their light reactions, but offer these results with alcohol and strychnine as evidence that all stimulation and depression do not cause reversals in phototaxis. Chart 3 Showing the light reactions of ten nymphs treated with potassium cyanide together with three preliminary test periods and one complete control. In the cases where the preliminary and control graphs are not given it is under- stood that they are in no essential respect different from those shown. In order to economize space the time spent in changing from water to cyanide is not fully shown as it was in chart 2. Where such a change is indicated it must be understood that two or three minutes elapsed. In all cases given in columns 5 and 6 the control, with which the carbon dioxide production of the treated nymphs was compared, is not given in the chart. For text reference see p. 535. Chart 4 al 2 3 4 S/POValre SI RG/AGY yee 4/28/17 4/27/17 Water Water Water Water eo w a o oo ~ © s SUL UNUSUAL PCCD ATU PRATAP SSE PULAU ERAT TET Ss TSPpEDDNRURETN REMEDIES PTS VES MST ASARUE eS CR ATS OGRA MRE US ERTECSTN CALC brobritiebe eee pouidieebreb eed 5B i S & a Changed to Changed u/25 HCL to water & & 5 & Changed More C02 Less CO2 to Alc .2% os 3 Water Water B'omitted 18 ‘omitted 6/22/17 6/12/17 8 CTT CUCL PLLOLOC I) LOLOL LOC LOL LLL WL HL TO PPL 8 | Changea to Alc .2% 8 8B & B 8 UF LOLLIPOP Hirt TT LCCC PL is : rs) 8 g 8 8 HC] N/25 8 g Reversed Ss 8 Reversed age) ests ! E ° 4 i) Q o = hore CO - than control/|/ than control Hobbit bebe be be be prrelovreltrbimel deed poilontretietecrbeiboe 450 7 5 6 4/20/17 4/20/17 Alcohol Water 15' omitted 15" omitted 6 oe a Ss COOTER POL PL PO Fololobrbe carted 3 ry = 3 a s 5 - Hlassertavstetarerrdscrerbereredivieg | METRE TY ET EY 8 ‘S 8 xf & 8 Te wi Less CO More C02 ~ 30 & & = Ss 2 @ =) o ow - e “s ~ = 8 8 8 2 8 B 8 8 s 5 5 S SOOO OT OD GUO CUCU GUO) CGO UC COUN CUO Occ UOC OC PP BB 8 Hotbot bia LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 401 TABLE 12 Showing the effect of 2 per cent ethyl alcohol upon the phototactic reaction of nymphs negative to light and upon their carbon dioxide production SIZE SIGN OF LIGHT REACTION MORE CO? PRE ta Alcohol Water Alcohol Water 10.5 10.5 + + Alcohol a 8.5 8.5 ae St Alcohol 8 12.0 12.0 — _ Alcohol 9 9.0 9.0 - -- Alcohol 9 11.0 11.0 _ — Water gl 9.0 9.0 - = Water g} 8.0 8.0 — _ Alcohol 10 1S SS + + Alcohol 10 10.0 10.0 — + Water 10! 10.5 10.5 - - Water 10! 11.0 11.0 + + Alcohol 11 9.0 9.0 - = Alcohol 12 11.0 11.0 — a Alcohol 12 11.0 11.0 _ ~ Alcohol 13 9.0 9.0 — - Alcohol 16 10.5 10.5 a _ Alcohol 28 10.0 10.0 = — Water 44 1015 Mya - - Water 50 8.5 8.5 a -E Water 50 10.5 10.5 — — Alcohol 52 11.0 AO) _ — Water 55 10.0 10.0 — — Water 83 9.0 9.0 = — Water 180 1 These control nymphs were more active than the experimental ones before the biometer test was made. Chart 4 Showing the reactions of two positive nymphs to light after treat- ment with N/25 hydrochloric acid together with their controls and the result of treating three nymphs with 2 per cent ethyl alcohol with one control tracing. The control shown in column 6 remained quietly at the negative end during almost all of its exposure and yet gave more carbon dioxide than the more active alcohol treated nymph whose response is shown in column 5. The nymph shown at the bottom of column 1 became negative immediately upon treatment with the acid. The experimental dish was then twice turned end for end each time the nymph moved directly away from the light. 452 W. C. ALLEE AND E. R. STEIN, JR. DISCUSSION Since this is an inquiry into the problem of a possible relation- ship between the sign of the phototactic reaction and the general rate of metabolism of May-fly nymphs, it is pertinent to question the effectiveness of our means of measuring the rate of the meta- bolic processes. The cyanide-resistance method is best checked by comparing results obtained with it with those given by Ta- shiro’s method of determining carbon dioxide production. This has been done for the isopod Asellus communis (Allee and Tashiro, 714). There it was found that the two methods gave the same results in direct tests and that two isopods subjected to daily variations of oxygen tension for ten successive days with daily quantitative estimations of carbon dioxide production by Dr. Tashiro behaved according to expectation based upon previous experience with the cyanide method. Regarding the work with Tashiro’s biometer in measuring the carbon dioxide production, I consider this the best, though by no means the fastest, method of making such comparative car- bon dioxide tests as those recorded in this paper. It is less complicated than the electrolytic determination of the hydrogen ion concentration and more accurate than the colorimetric meth- ods. The instrument is somewhat complicated in appearance, but it is in reality as simple in manipulation as a modern micro- scope equipped with oil-immersion lens, mechanical stage, and camera lucida. I have repeatedly demonstrated end points to coworkers at Woods Hole and even to college freshmen. My only change from Tashiro’s technique (’17, p. 109) consists in the use of a low-power binocular in reading end points. The sources of error in the method as applied to May-fly nymphs are: 1. May-fly nymphs are water-dwelling animals and were tested in as nearly a dry atmosphere as possible. 2. Five minutes or more intervened between the time the nymphs were taken from the water and the first indication of the relative rate of carbon dioxide production. During this interval the nymphs must be picked up, partially dried, and placed in glass containers. LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS 453 Regarding these points it must be borne in mind that the readings here given are all comparative ones in which the control and experimental nymphs were treated exactly alike. Since one must needs be taken from the water before the other, this was varied in the different experiments. The nymphs often lived twenty-four hours in the apparatus and at times they lived as long as forty-eight hours in the slight amount of moisture present. 3. Difference in carbon dioxide production may be due to difference in movement. This source of error was eliminated by the simple method of throwing out all tests where movement occurred. 4. Unconscious personal preference for one of the nymphs producing more carbon dioxide. This was eliminated by the experimenter’s ignorance of which was experiment and which was control. The relationship between the general metabolic processes of animals and their reaction to light may conceivably be one of the following: 1. Conditions that depress positive animals may make them negative (Mast, Bohn, Drzewina) and conditions that stimulate negative animals may make them positive (Holmes, Carpenter, Bohn, Jackson). 2. The above relationship may be reversed (Phipps in part). 3. Conditions that stimulate animals may cause reversals of their normal reaction or vice versa. 4. Conditions that markedly change the metabolic rate may cause reversal of either positive or negative animals. 5. Changes of metabolism accompanying changes in light reac- tions may be incidental or resultant rather than causal. 6. The relation between metabolic processes and the reaction to light may vary in different species so that no general law can be worked out. 7. There may be no relationship between the rate of metabol- ism and the phototactie reaction. With the positive Epeorus nymphs only depressing agents were used and these caused reversals. With both the negative 454 W. C. ALLEE AND E. R. STEIN, JR. species both stimulation and depression resulted in reversal and in positive members of the Lake Forest Heptageninae stim- ulation of positive nymphs caused reversal. In Leptophlebia thirty-six tests with hydrochloric acid shewed no effect on the average resistance to potassium cyanide. The biometer tests with the Heptageninae give the explanation. When first sub- jected to the acid the nymphs are stimulated; later they are de- pressed. If taken during the first period the resistance to cya- nide would offset that of the later period and so yield an average about the same as the control. An examination of the carbon dioxide production records shows that about as many nymphs were stimulated as were depressed by the treatment, so that a general average not considering the time factor would show re- versals with no relation to the rate of carbon dioxide production. In the May-fly nymphs studied the results obtained demon- strate that there is a relationship between the rate of metabolic processes and the sign of the phototactic reaction. It is also clear that reversals are accompanied by either a marked stimula- tion or marked depression. The experiments indicated but do not prove that the stimulation or depression is causal. From one point of view it makes little difference whether the metabolic changes are causal or only symptomatic; the fact that they are correlated at all is important. If the change in metabolic conditions is causal, the fact is evi- dent that all changes do not cause reversals. This was shown particularly by the action of the ethyl alcohol upon the Like Forest nymphs. This stimulated and later epressed the nymphs with or without an accompanying reverscl. On the assumption that metabolic change is causal the non-action of alcohol in 80 per cent of the cases might be explained by supposing that it did not cause a change quantitatively large enough. This suggestion is supported by the observation with the other species that aleohol accompanied by decrease in tem- perature ws more effective in producing reversals than alcohol alone and by the fact that in general these species were more susceptible. LIGHT REACTIONS—-METABOLISM—MAY-FLY NYMPHS 455 The idea that a certain quantitative change must occur before reversal in light reaction takes place is also supported by carbon dioxide determinations made before the nymphs were exposed to light. In some of these, as, for example, the results shown in columns 5 and 6 of chart 1, the nymph with the higher speed of earbon dioxide production became positive upon exposure to light while the other was decidedly negative. It more frequently happened that both nymphs so treated were negative in spite of the fact that one had a higher rate of metabolism than the other. Evidently the biometer is more sensitive to changes in carbon dioxide production than is the !ight reaction. Strychnine, again, did not cause a high degree of reversals, but it did strongly stimulate the nymphs. There is no question but that this stimulation is as strong as that caused by hydro- chloric acid which caused a high percentage of reversals. This difference in result can only be explained by the assumption that while light reversals are accompanied by changes in metabolism that are probably causal, all such changes do not cause reversals in reaction to light. The untreated positive nymphs were found to give off more carbon diox de than untreated negative ones. These negative animals often moved back and forth in the dishes, spending the major part of the time at the negative end. Some nymphs do this more than others. This brings them to the positive end more frequently and gives them more opportunities to come to rest there. A reversal apparently on this plan is shown in column 4, chart 1. It appears that this is one mechanism for reversal to light and here the relation between metabolic rate and the reversal is obvious. The higher the metabolic rate, the greater the tendency to move back and forth; the more random move- ments, the greater the chance of becoming acclimated to the positive end and of reversing the normal reaction. All eversals even under the influence of light alone were not of this type, witness column 3, chart 1. This reversal, like most of those experimentally obtained by the use of chemicals, was not preceded by random movements, but took place as though the animal had suddenly discovered the attractiveness of the positive end and must needs go there even though it died in the attempt THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, No. 3 456 W. C. ALLEE AND E. R. STEIN, JR. (as in cyanide reversals). This excursion was often the first one the nymph had made to the positive end and frequently the nymphs remained there until they died. The reason for ‘he rela- tion of metabolic rate with this type of reversal is not evident, especially when the reversal followed a very strong depression. An explanation of reversals sometimes advanced (Ewald, 13) is that animals change in their sensitivity to light and hence reverse their reactions. On this basis the change in sensitivity is causal; the light reversal and change in carbon dioxide produc- tion, resultants. In May fly nymphs these supposed resultants are correlated in such a way that either stimulation or depression may accompany reversals from negative to positive light reac- tions. This means that either increasing or decreasing sensi- tivity to light will make negative May-fly nymphs positive, or that an increase (or decrease) in sensitivity wil cause now de- pression and now stimulation. Either of these necessary assump- tions expand the sensitivity hypothesis beyond the limits justified by known facts. CONCLUSIONS The light reactions of the positive May-fly nymph, Epeorus, was reversed by treatment with alcohol, lowered temperature, calcium chloride, and other reagents. Nymphs so reversed had a lower rate of metabolism, as measured by resistance to potas- sium cyanide, than normal untreated nymphs. The negative nymph, Leptophlebia, was similarly reversed with accompanying stimulation or depression as measured by resistance to the cyanide. A negative nymph belonging to the Heptageninae was reversed in its light reactions with accompanying increase or decrease in carbon dioxide production as measured by Tashiro’s biometer. The experiments conclusively demonstrate that the photo- tactic reaction of these nymphs is correlated with the metabolic condition and indicate, but do not prove, that certain changes in metabolism cause the reversals in reaction to light. All nymphs that reversed their light reactions we e e ther stimulated or depressed, but stimulation or depression did not necessarily involve phototactic reversal. LIGHT REACTIONS—METABOLISM—MAY-FLY NYMPHS. 457 LITERATURE CITED AuLeE, W. C. 1914 Certain relations between rheotaxis and resistance to po- tassium cyanide in Isopoda. Jour. Exp. Zoél., 16, pp. 397-412. ALLEE, W. C., AND TasHtRO, SuHrro 1914 Some relations between rheotaxis and the rate of carbon dioxide production in Isopods. Jour. An. Beh., 4, pp. 202-214. Boun, Georces 1912 Les variations de la sensibilité en relation avec les variations de l’Etat Chimique Interne. C. R. Acad. Sci. Paris, 154, pp. 388-391. CARPENTER, F. W. 1905 Reactions of the pomace fly Drosophila ampelophila Loew to light gravity and mechanical stimulation. Am. Nat., 39, pp. 157-171. Cuitp, C.M. 1913 Studies on the cynamics of morphogenesis and inheritance in experimental reproduction. V. The relation between resistance to depressing agents and rate of metabolism in Planaria dorotocephala and its value as a method of investigation. Jour. Exp. Zodl., 14, pp. 153-206. 1913 a. Certain dynamic factors in experimental reproduction and their significance for problems of reproduction and development. Arch. Entw. Mech., 35, S. 598-641. — Drzewina, ANNA 1911 Sur la résistance des crustacés au cyanure et les effets sensibilisateurs de cette substance. C. R. Soc. Biol., 70, pp. 855-857. Dr7EWINA, ANNA, ET BoHn, GEorGES 1912 Modifications des réactions des animaux sous l’influence du cyanure de potassium. Ibid., 70, pp. 855-857. Ewaup, Woutreana F. On artificial modification of light reactions and the influence of electrolytes on phototaxis. Jour. Exp. Zodél., 13, pp. 591- 612. Geprert, J. 1899 Uber das Wesen der Blausiure-vergiftung. Zeitschr. klin. Med., 15, S. 208-307. Howimes, 8. H. 1905 The reactions of Ranatra to light. Jour. Comp. Neur., 15, pp. 98-112. Hyman, L.H. 1916 Onthe action of certain substances on oxygen consumption. I. The action of potassium cyanide. Am. Jour. Physiol., 40, pp. 238- 248. Jackson, H. H. T. 1910 Control of phototactic reactions in Hyalella by chem- icals. Jour. Comp. Neur. and Psych., 20, pp. 259-263. Logs, J. 1904 The control of the heliotropic reactions in fresh-water crus- taceans by chemicals. Univ. of Calif. Publ. in Physiol., 2, pp. 1-3. Logs, J. 1912 Comparative physiology of the brain and comparative psy- chology. MacCurpy, Hansrorp 1913 Some effects of sunlight on the starfish. Sci., N.S., 38, pp. 98-100. Mast, 8. O. 1911 Light and the behavior of organisms. Wiley, 410 pp. Puipps, C. F. 1915 An experimental study of the behavior of Amphipods with respect to light intensity, direction of rays and metabolism. Biol. Bull., 28, pp. 210-222. 458 W. C. ALLEE AND E. R. STEIN, JR. SuHetrorp, V. E. 1913 Animal communities in temperate America. Chicago, 362 pp. TasHIRO, SHtRO 1913 A new method and apparatus for the estimation of exceedingly minute quantities of carbon dioxide. Amer. Jour. Physiol., 32, pp. 137-145. 1917 A chemical sign of life. Univ. of Chicago Press, 142 pp. WopseEeDALeEK, J. E. 1911 Phototactic reactions and their reversal in May- fly nymphs. Biol. Bull., 21, pp. 265-271. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 24 STUDIES ON THE PHYSIOLOGICAL SIGNIFICANCE OF CERTAIN PRECIPITATES FROM THE EGG SECRETIONS OF ARBACIA AND ASTERIAS ALVALYN E. WOODWARD Simmons College, Boston, Massachusetts TWO CHARTS AND THREE FIGURES CONTENTS Dei ntrod uGtlOne hc iot. Kegs werner tek ps eiieek « eto te AAR ps ee Ws. eet wile eee | ids oases 484 III. Concerning activators—particularly lipolysin—and theories of activa- GL ORPR PE ayy SEER ers. 2.2 oh aie se eeste wand «ia oeeed ed od Mn eee aes 484 Aaa AT GAOL A ROLE MANN OME) 92 iota; eogis Seoetshs ate atiase fee Soon seit ees 485 Be ONCE AONMO NE BDOLIT Gro 5 oso Poise oN are vc aestcids doo. chd sroneliphom ated aie ae 494 DOV) S Ue yee Perret oe tac hae ei Pe Ske oe. eaters alts Sere 495 MAU ere HES) Tet 210» oS ah) 5 ta OD Poa Ree Pee RRR OEE ae? ee a 497 I. INTRODUCTION A new phase in the physiology of fertilization practically began in 1912 when F. R. Lillie discovered the effect of egg extracts and secretions on sperm. He found that sea-water in which Arbacia eggs had been standing caused sperm of the same species to be activated, directed, reversibly agglutinated, and finally paralyzed. If a drop of this egg-sea-water is introduced, by means of a 459 460 ALVALYN E. WOODWARD capillary pipette, into a suspension of sperm, the latter become more active and gather around the drop as if chemically attracted. In addition, they form small clusters, ‘agglutinations,’ which last for several seconds. After this the sperm separate. The period of agglutination depends upon the activity and density of the sperm suspension and the strength of the egg-water. The effective material in the water Lillie (712) called an iso-agglutinin. The eggs of Nereis were found to produce a similar substance, which seemed to appear at the moment of fertilization. In this case, however, the active substance is destroyed by boiling, while Arbacia agglutinin is not entirely destroyed when boiled for seventy minutes. Lillie also found that an extract of Arbacia eggs agglutinated Nereis sperm. Because of differences in la- bility, he concluded that the Arbacia egg produced both iso- and hetero-agglutinins. He also discovered that if sperm be added in quantity to egg-water, the agglutinin is ‘fixed,’ and no longer affects sperm subsequently added. For purposes of quantitative comparison, Lillie prepared a ‘standard solution’ of egg secretion by allowing one volume of ‘dry’ ripe eggs to stand in two volumes of sea-water for ten minutes. During this time the eggs were occasionally agitated. They were then precipitated with the centrifuge, and the super- natant fluid, carefully decanted, was subsequently used (Glaser, ’14.¢, p. 388). Since this secretion had such peculiar effects on spermatozoa, and since eggs from which it had been removed by repeated washings could not be fertilized Lillie has since called it ‘fertilizin’ (’13 b). A substance to which the cell membrane is not permeable was also found in the egg. This material could be obtained only by laking with distilled water or crushing. Since this material was able to neutralize the agglutinative power of fertilizin, Lillie ealed it ‘anti-fertilizin.’ Antifertilizin does not prevent the chemotactic effect of egg secretion. In addition, Lillie likewise discovered that Arbacia blocd in- hibits fertilization in the same species, but does not affect the agglutinative power of fertilizin. Moreover, the inhibitory effect of blood can be counteracted by an excess of fertilizin. EGG SECRETIONS OF ARBACIA AND ASTERIAS 461 From these facts, Lillie deduced his ‘side chain’ theory of fertilization, in which he applied Ehrlich’s conceptions to the subject at hand. He postulates that spermatozoa are chem- ically unable to unite directly with the egg, but that a third body, or amboceptor—in this case, fertilizin—is necessary to join them. Fertilizin, he believes, has two side chains, an ovophile group which can unite with a side chain of the egg, the egg receptor, and a spermophile group, which can unite with a sperm receptor. Fertilization may be blocked by 1) absence of the amboceptor; 2) occupation of either side chain by a foreign body; 3) occupation by similar means of either sperm or egg receptor. Further, “The sperm activates the fertilizing substance (fertilizin) already present in the egg. The egg is self-fertilizing (Lillie, 14, p. 587).”’ Il. CONCERNING THE FACTUAL BASIS OF THE FERTILIZIN THEORY The fertilizin theory is founded upon three essential elements: an amboceptor, an egg receptor, and a sperm receptor. The presence of the amboceptor, Lillie believes, is indicated by a. sharply defined reaction-agglutination of the sperm. This re- action is considered (’15) only as a visible, but not essential, symptom of a fundamental physicochemical change in the sper-. matozoa themselves. It is possible that in many species the egg secretion produces a significant change in the sperm without agglutinating them. The presence of sperm receptors, according to the same author, is indicated, among other things, by the Godlewski phenomenon, in which the inhibitory effect of foreign sperm is explained as due to the occupancy of the normal sperm receptors. The presence of an egg receptor he has been unable. to demonstrate directly. It is, however, indicated by other in- hibitors which effectively block the fertilization reaction. A. Concerning egg secretions in general The observations of Lillie with regard to the egg secretions of Nereis and Arbacia have been extended by several workers upon. the same and other forms. 462 ALVALYN E. WOODWARD Just (15 a) found that it was difficult or impossible to initiate development in Nereis eggs which had been washed. The eggs of Platynereis, also, are very sensitive to an excess of sea-water (15 b). From these facts he concluded that a substance formed by the egg and necessary for development had been removed. Fuchs found that the fertilizing power of the sperm of Ciona, Arbacia pustulosa, Ascidia mentaul, and Strongylocentrotus lividus could be increased by treatment with egg secretions of the same species. Moreover, the fertility of Ciona sperm could be enhanced by treatment with the egg secretion of Phallusia, Arba- cia, or Stronglyocentrotus. Likewise, Strongylocentrotus sperm are made more effective by the secretions of Ciona, Echinus, Sphaerechinus, and Asterias eggs. The egg extract, obtained by crushing the eggs, had the same effect as the secretion, except in the case of Asterias. In this instance, the secretion stimulates, while the extract poisons, Strongylocentrotus sperm. The method followed by Fuchs prevented the observation of sperm agglutination. He concluded that the secretions affected sperm only. Asterias eggs, Glaser found (’14 b), form a secretion similar to that of Arbacia, except in color and a few minor points. More- over, Asterias sperm are directed, activated, agglutinated, and paralyzed as well by Arbacia secretion as by that from their own species. Arbacia sperm react similarly to Asterias secretion. The writer has been able to confirm these observations and to demonstrate the reactions to colleagues. While mature Arbacia eggs in a healthy condition are easy to obtain and always secrete a sperm agglutinin, the conditions -differ in the case of Asterias. It was difficu!t, during the summers of 1915 and 1916, to obtain Asterias which contained many ferti- lizable eggs. In a good batch 10 per cent would produce larvae. Therefore, it was difficult to obtain the agglutinin in any usable amount. In June, 1914, however, Asterias at Woods Hole were in the best condition I have ever known. It was almost impos- sible to keep them in the laboratory three or four hours without their shedding eggs or sperm. Mr. Gray reported that the live- ears in which they were kept, though very open, were always EGG SECRETIONS OF ARBACIA AND ASTERIAS 463 slimy with eggs. At that time, when 98 per cent to 100 per cent of the eggs would develop normally, most of the observations on the sperm agglutinin were made. All of them, however, were repeated and confirmed in 1915, 1916, or 1917. As observed by Glaser, if a drop of sea-water, which had been standing over mature Asterias eggs, was injected under a cover- glass into a drop of sperm suspension, the sperm would gather into small, irregularly angular clusters of six to eighteen and remain agglutinated for a number of seconds which varied with the strength of the solution. When they separated, they were much more active than before. The general process of agglutina- tion and activation is the same in Asterias as in Arbacia, where it has been observed by many. The chief difference is in the size of the clusters, which are smaller in Asterias. As is known, Asterias eggs are obtained with the germinal vesicle still intact. Maturation changes begin almost immedi- ately and are complete in forty-five to sixty minutes. Directly after obtaining them, some of the eggs of a large female were put into a centrifuge tube with two volumes of water and allowed to stand ten minutes to obtain fertilizin in the standard way. After centrifuging, the supernatant fluid was tested for its agglu- tinative power. When diluted 1 : 200, Asterias sperm remained agglutinated six seconds, a unit reaction. An hour after shed- ding, the rest of the eggs from the same female, which had been kept in a large amount of water, were washed twice with fresh sea-water, and then the secretion obtained as before. This, when diluted 1: 200, caused a fresh sperm suspension to remain agglutinated over four minutes. In other words, it was over sixty times as strong as the other. Experiments of Lillie and Just indicate that in Arbacia and Nereis, as well as in Asterias, fertilizin is formed by the maturing or mature, but not by imma- ture eggs. I believe that Lillie’s failure (quoted by Loeb, 16, p. 83) to verify the observations with Asterias secretion was due to abnormal material. Loeb (’15) has observed activation of sperm of the echinoderms at Pacific Grove by eggs of the same and related species. He also noted that the sperm of Strongylocentrotus purpuratus are 464 ALVALYN E. WOODWARD agglutinated by the egg secretion of the same species, while sperm of 8. franciscanus are agglutinated not only by secretion from eggs of the same species, but also by that of S. purpuratus. Miss Margaret V. Cobb (unpublished) discovered that the eggs and egg-water of Cumingia produce positive chemotactic response on the part of the sperm. These results are summarized in table 1 and warrant the assertion that the eggs of at least four phyla of marine animals secrete into the sea-water substances significant either in their effects on sperm or in fertilization. B. Concerning the secretions of Asterias and Arbacia in particular While these facts indicate that the egg secretion may play an important role in fertilization, they do not make clear either the nature of the secretion itself or the manner in which it oper- ates. It may be one homogeneous substance or a mixture of two or more specific substances. Conceivably, there are three ways in which the problem can be attacked: 1) by physiological analysis of the secretion; 2) by general chemical analysis; 3) by the removal of specific ele- ments from the secretion and the further analysis of their indi- vidual properties. 1. Physiological analysis. a. The secretion is necessary for fer- tilization in some species. As stated above, Lillie gave the name ‘fertilizin’ to the egg secretion because he considered its presence absolutely necessary for the fertilization of Arbacia eggs. Just found it equally indispensable for the fertilization of Nereis and Platynereis. Loeb (’15) criticised their conclusion on the ground that ine washed eggs had stood so long after shedding that they were dead, but the following experiment disposes of this objection. The eggs from ripe starfish were divided into two lots. One lot, A, was put into a finger-bowl of sea-water as control. The rest were placed into a large glass tube, closed at the bottom with chamois skin. The upper end was similarly closed except for an opening into which fitted closely ‘a smaller tube through which (prea SEO ‘IOSB[4)) | (paear peyeu -pooM -Iqn[ S08 ‘I98B]5)) ‘peqes *poyeu -01558 -Ipn[ 538 TIS9qIOJ peyearoy ‘peyearjoy SBLIOJSY (syong) OT 1OF o1OUL B[N}UeUr aul0deg BIiplwsy (sqon,y) (syon,q) OEP IEF Soy 910uL 910UL Bsopnjsnd eulodeg euloveg, BlIoBqry (prea (pasa -PpooM -poo\y ‘IO8v[D) ‘IOSBlD) pezsy ‘Our D) -Bied :! pezA[eied peyeu ‘peyeu -Iyn[s3e -1}n[ 398 (@NTTYT) : ‘pozed ‘pozes | pezeu -01938 01338 -njyound “1yNa3V PoyBArjoy, “PeyBalpoyV BIoBqry USE AN NN or Etaeen sdoyesour | vyequiy Ouro sIpeurysoy voovryoo | mseqaoy | epnjzueur |esopnysnd | eyepnyound Rotmeepee i aseony Sracana bee sor0uAye[g | sielaNy Boerne AEN) -ur wuory | PUM°ISV | serroysy | seueysy | erprosy | vloeqry | loequy fresh sperm, (controlaayen. . « toa) Pr > q ee ee q ee] jee] ee] an =F + ale 3 35 F a =F minutes 2.5 2 ZA WG) PAL PAL XO ales es} On 8 | 13 4 6 550 1 aa ale OO 22a eZee aramid an inltes 6 8 3 5 ae 7 HA TSO 19) 20s eee 26 || 28 GES 2 5 10.0 3 3 Gale 2A 2 ee 20) 26m es ag AD ets 3 6 Ze ia adel 3 9 | 14 | 78 | 20 | 16 | 29 | 16 || 26 3) 4 15.0 ZOUPOOM ede || 2S | wlan 2d Oe eele) 2 5 3 a 20.0 21 | 24 9 Ne UI bE | AN hes) 2 8 4 5 25.0 15 | 20 Soe: 7 9 does not improve eggs that are already fairly normal, like those in the last column of table 17. In this connection it is interesting to recall the experiment of EK. P. Lyon and L. F. Shackell (10). They found that iodine, as indicated by the starch reaction, disappeared more rapidly from sea-water in the presence of unfertilized eggs than of eggs which had been fertilized. From this they concluded that the unfertilized eggs are more permeable to iodine than the fertilized. In the light of my own experiments, it seems more probable that the unfertilized eggs contained more unsaturated fatty acid with which the iodine might combine. EGG SECRETIONS OF ARBACIA AND ASTERIAS 493 On the other hand, if the eggs be allowed to settle in test-tubes and only the supernatant fluid be tested, it will be found that the liquid above the unfertilized eggs combines less readily with io- dine than does that above the fertilized eggs. The first inference from this was that the fertilized eggs had excreted unsaturated fatty acid into the sea-water. If, however, we add to the water poured off from unfertilized eggs an amount of sperm suspension equal to that present in the other tube, we find that iodine is ab- sorbed in equal amounts from each. Therefore, the results were due to the presence of sperm in the water above fertilized eggs, rather than to unsaturated acid excreted at the time of fertilization. TABLE 17 The effect of iodine on resistant or overripe Arbacia eggs PER CENT PehOLCCBEg Pat SHEL COMbROl. ta tatty icc Dee a 24 29 81 2. 10 cc. eggs +1 drop saturated iodine 10 minutes + STACI Petes ava MWh las oS dee SMA Nna Me hE SU ce as cud etl 28 39 40 3. 10 ce. eggs + 2 drops + saturated iodine 10 minutes + SHOEI 6 Senet CVE EPR TAREE ETN hs Ok pS Se 36 44 62 4. 10 cc. eggs + 1 drop } saturated iodine 10 minutes + OCT ace eo tas OE SEIS MO EEA toh: RAP Ee Eee ae es 50 48 61 5. 10 ce. eggs + 1 drop + saturated iodine 10 minutes +- SOCTUA MER Near mnt shins is aan. Yama oA I. lu oR tani | 3l Neither Asterias nor Arbacia lipolysin seems to combine with iodine. The chemical parthenogenetic agents, however, react with the fatty inhibitor, and, as we saw, the egg secretion, one of the three physiological activators, also combines with it to produce a compound soluble in water. According to this view, then, the resting egg contains enzymes which control metabolism, unsaturated fatty acid which inhibits enzyme action, and lipolysin, which reacts with the unsaturated fatty acid to make it innocuous. We can conceive that the rela- tive amounts of these substances may change from time to time in the same egg and may differ in different eggs. If there is more than sufficient fatty acid present to combine with all the lipolysin, the surplus will inhibit the action of the enzymes and the egg 494 ALVALYN E. WOODWARD will lie dormant. As the egg ‘ripens,’ however, it secretes more and more lipolysin. If this accumulates within the egg to such an extent that the inhibitor is relatively greatly reduced, the metabolic enzymes become extremely active and development or cytolysis results. Perhaps natural parthenogenesis, such as occurs in Daphnia and aphids is the result of such conditions. The membrane of starfish and sea-urchin eggs is easily perme- able to lipolysin, so that it does not normally accumulate within the egg. If, however, the eggs are crowded in a small amount of sea-water or, which amounts to the same thing, are placed in a bath of lipolysin, the loss of this substance from the egg is pre- vented. It therefore reacts wth the inhibitor, and the egg de- velops—a case of autoparthenogenesis. Any other method which disturbs the ratio of activating enzymes to fatty acid so that the proportion or effectiveness of the enzymes is increased must lead to development or cytolysis. As already stated, many chemicals have this effect, and produce parthenogenesis. B. Activation by sperm Can the action of the spermatozo6n be explained along similar lines? Its effect probably varies with the group of animals. Ostwald (’07) measured the amount of peroxidase and catalase in 1) extracts of Amphibian eggs; 2) extracts of sperm; 3) a mixture of the two extracts. He found that the mixture contained con- siderably more enzyme than the sum of (1) and (2). Hence we may conclude that in Amphibia the sperm carries into the egg something which changes pro-enzymes into enzymes. It in- creases the activator. On the other hand, Amberg and Winter- nitz (11) were unable to find any more oxidase or catalase in echinoderm eggs laked after fertilization than in those laked be- fore. Moreover, Warburg (’08) and Loeb and Wasteneys (’13) found the same rate of oxidation in parthenogenetic as in fertilized sea-urchin eggs. S nce, in this group, fertilization does not in- crease the amount of oxidase in the egg, the spermatozo6én may act by reducing the amount or effect of the inhibitor. EGG SECRETIONS OF ARBACIA AND ASTERIAS 495 In conclusion, I wish to express my gratitude to Dr. O. C. Glaser, who suggested this problem and directed the work; to Dr. F. R. Lillie for the privileges of the Marine Biological Lab- oratory, Woods Hole, and to the several colleagues who kindly allowed me to demonstrate to them various reactions. IV. SUMMARY In confirmation of the work of Lillie and Glaser, it was found that Asterias and Arbacia eggs secrete into the supernatant sea- water a substance which causes the sperm of the same species to be activated, aggregated, reversibly agglutinated, and paralyzed. The secretion is also a parthenogenetic agent. Further study of its physiological properties brought out: 1. That its presence is necessary for the fertilization of the egg, since, a. Immature eggs of Asterias, which cannot be fertilized, pro- duce a secretion with less than one-sixtieth the agglutinating power of that produced by the same eggs when mature. b. Eggs from which the secretion has been washed do not de- velop when inseminated. If, however, secretion be added before insemination, they develop. c. Arbacia eggs which are ‘resistant’ to fertilization late in the season, also produce little secretion. They fertilize normally if secretion is added. 2. That the secretion has a dual nature, as shown by the fol- lowing facts: a. It reacts with both the sperm and the egg. b. Boiling destroys its value as a parthenogenetic agent, but not as an agglutinin. c. Perivisceral fluid of the same species inhibits autopartheno-- genesis, but not agglutination. A qualitative study of the chemical properties of the secre- tion confirmed Glaser’s observations, and indicated: 1. That it does not dialyse through a collodion sac, and so is probably colloidal. 2. That it contains carbon and nitrogen, but gives no clear response to protein tests. It gives a faint yellow color, in the 496 ALVALYN E. WOODWARD xanthoproteic test, however, which indicates the presence of tyrosine, phenylalanine, or tryptophane. 3. That two substances can be precipitated from the same secretion: a. By saturation with (NH4,).SQ., a sperm agglutinin is thrown down. b. By a method of Robertson and others, using BaCl, and acetone, a parthenogenetic agent is obtained. In an attempt to learn whether or not enzymes are present, it was found that the agglutinin resembles an enzyme in the effect upon it of x-radiation. It also follows the law of Schitz and Borissow, i.e., rate of action = KC ferment. The secre- tion does not give a positive reaction to tests for oxidase, cata- lase, or a proteolytic enzyme. Since the, parthenogenetic agent dissolves a fat obtained from the eggs, it may contain a lipase. Hence it was called, provisionally, a lipolysin. A study of methods of inhibiting the initiation of cleavage brought out that both autoparthenogenesis and sperm fertiliza- tion are inhibited by: 1. Asterias and Arbacia serum, which may, like mammalian sera contain antiferments specifically protecting the cells of the organism itself. . 2. ‘Purple x’ and ‘Salmon x,’ substances obtained by boiling the testicular or ovarian tissues of Arbacia and Asterias. 3. Antifertilizin, obtained from eggs which have been freed from fertilizin. This inhibitor may be identical with 4. An unsaturated fatty acid obtained by extracting the eggs with ether. Parthenogenetic agents fall into three classes: 1) The fat solvents, including lipolysin, ether, chloroform, butyric acid, etc. 2) The halogen compounds—iodine and various salts. 3) Physi- cal methods, which may change the physical or quantitative relations of substances within the egg. Not only do these agents produce development in normal, uninseminated eggs, but they change resistant Asterias and Arbacia eggs so that they may be fertilized. EGG SECRETIONS OF ARBACIA AND ASTERIAS 497 A consideration of the above facts makes it seem probable that the factors tending to produce development in the resting egg are of the nature of enzymes. The action of these, Jobling found, may be inhibited by unsaturated fatty acids. The egg remains in the resting stage so long as the action of these enzymes is inhibited by the unsaturated fatty acid. The egg itself, when mature and in a suitable medium, produces a lipolysin which binds this inhibitor. The efficiency of the inhibitor may also be lowered by physical and chemical means. In some groups, the spermatozoon appears to bind the inhibitor, in others, to increase the activity of the enzymes. V. LITERATURE CITED AMBERG, S., AND M.C. WinTERNITz 1911 The catalase of sea-urchin eggs before and after fertilization, with especial reference to the relation of cata- lase to oxidation in general. Jour. Biol. Chem., vol. 10, pp. 295-302. Evuter, H. 1912 General chemistry of the enzymes. J. Wiley & Sons, New York. Fucus, H. M. 1914 Studies in the physiology of fertilization. II. The action of egg secretions on the fertilizing power of sperm. Archiv f. Entw’- mech. d. Organismen, Bd. 40, pp. 205-252. Guaser, O. C. 1913 On inducing development in the sea-urchin (Arbacia punctulata), together with considerations on the initiatory effect of fertilization. Science, N.S., vol. 38, pp. 446-450. 1914a The change in volume of Arbacia and Asterias eggs at fertiliza- tion. Biol. Buil.,:vol. 26, pp. 84-91. 1914b A qualitative analysis of the egg-secretions and extracts of Arbacia and Asterias. Biol. Bull., vol. 26, pp. 367-386. 1914e On auto-parthenogenesis in Arbacia and Asterias. Biol. Bull., vol. 26, pp. 387-409. GoptewskI, E. 1911 Studien iiber die Entwicklungserregung. II. Antagon- ismus der Einwirkung des Spermas von verscheidenen Tierklassen. Archiv f. Entw’mech. der Organizmen, Bd. 33, pp. 233-254. JOBLING, J. W., AND W. PererRsEN. 1914a Soaps as ferment-inhibiting agents. Jour. Exp. Med., vol. 19, pp. 239-250. 1914b Nature of serum anti-trypsin. Jour. Exp. Med., vol. 19, pp. 459-478. Just, E. E. 1915 a Initiation of development in Nereis. Biol. Bull., vol. 28, pp: 1-17. 1915 b An experimental analysis of fertilization in Platynereis mega- lops. Biol. Bull., vol. 28, pp. 93-114. 498 ALVALYN E. WOODWARD Lituiz, F. R. 1912 The production of sperm iso-agglutinins by ova. Science, N.S., vol. 36, pp. 527-530. 1913 a Studies of fertilization. V. The behavior of the spermatozoa of Nereis and Arbacia with special reference to egg-extractives. Jour. Exp. Zool., vol. 14, pp. 515-574. 1913 b The mechanism of fertilization. Science, vol. 38, pp. 524-528. 1914 Studies of fertilization. VI. The mechanism of fertilization in Arbacia. Jour. Exp. Zool., vol. 16, pp. 523-590. 1915 Sperm agglutination and fertilization. Biol. Bulk, vol. 28, pp. 18-33. Linurr, R. 8. 1911 The physiology of cell division. Am. Jour. Physiol., vol. 27, pp. 289-357. 1912 Certain means by which starfish eggs naturally resistant to fertilization may be rendered normal, and the physiological conditions of this action. Biol. Bull., vol. 22, pp. 328-346. Logs, J. 1910 The rdle of alkali in the development of the sea-urchin. Proce. Soc. Exper. Biol. and Med., vol. 7, pp. 119-120. 1912 The mechanistic conception of life. Univ. of Chicago Press. 1913 Artificial parthenogenesis and fertilization. Univ. of Chicago Press. 1914 Cluster formation of spermatozoa caused by specific substances from eggs. Jour. Exp. Zool., vol. 17, pp. 123-140. 1915 On the nature of the conditions which determine or prevent the entrance of the spermatozoon into the egg. Am. Nat., vol. 49, pp. 257- 285. 1916 The organism as a whole. G. Putnam’s Sons. New York. Logs, J.. anp M. M. CuamperuaIn 1915 An attempt at a physico-chemical explanation of certain groups of fluctuating variation. Jour. Exp. Zool., vol. 19, pp. 559-568. Logs, J., anp H. Wasrenrys 1913 The influence of hypertonic solution upon the rate of oxidations in fertilized and unfertilized eggs. Jour. Biol. Chem., vol. 14, pp. 469-480. LoEvenuHART, A. 8. 1905 Further observations of the catalytic decomposition of hydrogen peroxide. Am. Jour. Physiol. vol. 13, pp. 171-185. Lyon, E. P., anp L. F. SHacketi, 1910 On the increased permeability of fer- tilized sea-urchin eggs, following fertilization. Science, vol. 32, pp. 249-251. MacMunn, C. A. 1885 On the chromatology of the blood of some inverte- brates. Quart. Jour. Mic. Sci., vol. 25, pp. 469-490. McCienpon, J. F. 1910 Further proofs of the increase in permeability of the sea-urchin’s egg to electrolytes at the beginning of development. Science, vol. 32, pp. 317-318. Matuews, A. P. 1913 An important chemical difference between the eggs of the sea-urchin and those of the starfish. Jour. Biol. Chem., vol. 14, pp. 465-467. OstwaLp, W. 1907 Uber das Verkommen von oxydative Fermente in den reifen Geschlechtszellen von Amphibien. Biochem. Zeitschr., Bd. 6, pp. 409-472. EGG SECRETIONS OF ARBACIA AND ASTERIAS 499 Ricwarps, A. 1914 The effects of x-rays on the action of certain enzymes. Am. Jour. Physiol., vol. 35, pp. 224-228. Ricwarps, A., AND A. E. Woopwarp 1915 Note on the effect of x-radiation on fertilizin. Biol. Bull., vol. 28, pp. 140-148. Rosertson, T. B. 1912 On the isolation of oécytase, the fertilizing and cyto- lysing substance in mammalian blood sera. Jour. Biol. Chem., vol. 11, pp. 339-346. Warsura, O. 1908 Beobachtungen tiber die Oxydationsprozesse im Seeigelel. Zeitschr. Physiol. Chemie, Bd. 57, pp. 1-16. Witson, E. B. 1901 Experimental studies in cytology. II. A cytological study of artificial parthenogenesis in sea-urchin eggs. Archiv f. Entw’mech. d. Organismen, Bd. 12, pp. 529-596. Woopwarp, A. E. 1915 Note on the nature and source of ‘purple x.’ Biol. Bull., vol. 29, pp. 185-137. Woopwarp, A. E., anp F. S. Hagur 1917 Iodine asa parthenogenetic agent. Biol. Bull., vol. 33, pp. 355-360. PLATE 1 EXPLANATION OF FIGURES Photomicrographs of inseminated Asterias eggs: 1 and 2 Controls. Note that the spermatozoa are scattered. This shows especially well in the right hand part of 1. 3 Eggs from the same female treated for one hour with 0.3 per cent ether in sea-water, washed, and then inseminated, with the same sperm suspension as above. Note the spermatozoa in halos around both mature and immature eggs. 500 EGG SECRETIONS OF ARBACIA AND’ ASTERIAS PLATE 1 ALVALYN E, WOODWARD AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JULY 19 EFFECTS OF CHEMICALS ON REVERSION IN ORIENTA- TION TO LIGHT IN THE COLONIAL FORM, SPON- DYLOMORUM QUATERNARIUM S. O. MAST Zoological Laboratory of the Johns Hopkins University CONTENTS TRGT OUI CHLOM VG as Peo cctre: LS PRE Pe Teen She nits vce, cle se PEN AER eh costs loten crete 503 WraveniaeancdrmeynOdsinn nc aman hens che ee aie oe ae on anna 506 Effects of chemicals—General statement................. ccc cece ee ceceeees 507 1 EY STCTOUTLOVERE A 0 CI Reef eo are A ee rn av eS, ee ey ene meets FF 508 Te GUO feel a IS seem rte rey ie Ae INTE EA team gk ant ch ty Oubmied Se ac os Rea 512 ecto chemiealeconcentrationa veers. stake aie ee Meee ee 514 Effect of time-rate of change in chemical concentration................... 515 Mitechor nhysiahegicalstaten atest slate ists. .akioet Seba ss 516 Bite ctzotmanestiNeci Cs teaee Wh hehe. min, Sa Skene Mr Ree ae Ra 516 Hitectoltempenrssure;and dllmuminatlonesss 5) s4. 24 4sogee oa ae ccm 517 DISCUSSIONS ARN ee ieee eR AT Rs hae SAN NEA A a Rs 518 ROTA TiacIn Viaded LE Sec A Ot Rn en Md 2 0 Cee Ce at a mm Le Bee ARR ce 8 Bd 519 Ronn pMmONeabes sh cries talus aha con Abit Cone. A ody, creep tn Beha AR pa ua 520 INTRODUCTION The problems in behavior may be divided into two groups, one containing those which concern the nature and the cause of responses, the other those which concern the cause of changes in responses. The problems in the former group deal extensively with the relations between the responses and the environment or certain processes within the organism; those in the latter deal primarily with habit formation or learning, regulation or adapta- tion and evolution, i.e., with some of the most fundamental char- acteristics of living matter. The investigations which have thus far been made in behavior belong largely to the former group. Many observations have, however, also been made on changes in the responses. In fact, from a qualitative point of view, the whole field has been surveyed 503 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 3 504 Ss. O. MAST to such an extent that, while doubtless many more facts of a qual- itative nature will be discovered, this is not likely to result in the formulation of principles of much importance. Advance in behavior undoubtedly depends largely upon intensive quantita- tive research. This paper deals with the beginning of a series of quantitative investigations concerning changes in the nature of responses. Reversion in the sense of orientation was selected for this study because it appears to be among the simplest of the changes in responses, and if it occurs at all it is always fairly complete and fairly well defined, so that it serves well for work requiring pre- cise measurements. The selection of an organism that can be cultivated in the laboratory is also of importance, for in organisms that thrive in the laboratory normal behavior can be much more readily ascertained than it can in those which do not. Thus difficulties such as Esterly (’17), for example, encountered in work on the marine copepods are readily avoided. At present we are interested primarily only in reversal in orientation in light. The literature on this subject has recently been fairly thoroughly reviewed. (Mast, ’11, pp. 265-287; Holmes, 716, pp. 98-119; Washburn, *17, pp. 200-208). Refer- ence to these reviews and to a few investigations not mentioned in them leads to the following conclusions: 1. Organisms are usually positive in weak and negative in strong illuminations. Holmes (’01,-’05), however, maintains that the opposite holds for Orchestia and Ranatra, and there are many organisms in which reversion cannot be induced by light. 2. Reversal in the sense of orientation in light is probably usually dependent upon the amount of light energy received. This has been fairly clearly demonstrated by Arisz (’15) for plants and Mast (’07, pp. 154-162) for Volvox. Under certain condi- tions, however, reversion may depend upon the time-rate of change in light intensity. Gamble and Keeble (’03, p. 397) maintain that sudden increase in illumination causes positive Convoluta to become temporarily negative. Ostwald (07) con- tends that the same reversion is produced in Daphnia by either a sudden increase or a sudden decrease in illumination. Ewald REVERSION IN ORIENTATION TO LIGHT 5905 (14), on the other hand, holds that a sudden decrease makes Daphnia positive and a sudden increase negative. 3. Increase in temperature generally makes organisms positive to light and decrease usually makes them negative. But Loeb (05, p. 276) maintains the opposite holds for Polygordius larvae and Holmes (’05) comes to the same conclusion in reference to Ranatra; moreover, in some organisms temperature appears to have no effect on reversion. 4. Nearly all of the observations on the effect of chemicals on reversion in orientation have been made on crustacea (larvae and adults) and on Arenicola and Balanus larvae. None of the unicellular forms have been tested in respect to this and only one colonial form, Volvox, has been studied, and in this one only the effect of acids. These observations lead to the following conclusions: Acids tend to make all of the forms studied positive with the exception of Arenicola, but in many instances some of the salts, alkalis, and narcotics have the same effect as the acids. It is thus evident that chemicals which are fundamentally different in properties may have the same effect on reversion. Conse- quently reversion in the crustacea cannot be specifically related to the chemical constitution of the environment. It is therefore probably related to the physiological state of the organisms as a whole, but as to the nature of the physiological states involved in reversion we are as yet in total darkness. The interesting experiments of Allee (18, p. 95) show that in the May-fly nymphs it is not specifically associated with stimulation or depression. Allee found that hydrochloric acid and ethyl alcohol cause re- version from negative to positive, but that in some cases the nymphs were stimulated and in others depressed as indicated by the rate of production of carbon dioxid. Thus it is evident that in these relatively complex forms the problem is greatly involved. In the simpler forms it is not unreasonable to expect a more direct relation between the environmental factors and reversion. In a few of the investigations on the effect of chemicals on re- version solutions of known concentration were used, but in most of them the chemicals to be tested were merely added in indefinite 506 Ss. O. MAST amounts to the solutions in which the organisms live. Under such conditions it is, of course, impossible to say precisely what the chemical constitution is and how it is affected by the addition of chemicals. For example, most fresh waters are distinctly alkaline. When acid is added to such a solution it is evident that it can affect organisms as acid only after. the alkalis are neutralized. Addition of acids may, therefore, merely subject the organisms to an increase in salts and a decrease in alkalinity. The importance of this will become evident as we proceed in our discussion. MATERIAL AND METHODS Spondylomorum is a colonial organism consisting of sixteen zooids. It is ellipsoidal in form, about .05 mm. long and .035 mm. wide. Each zooid contains among other structures a prominent eye-spot and two flagella, considerably longer than the colony. The colonies are fairly active. They are definitely postero-ante- riorly differentiated and always swim with the anterior end or surface ahead, rotating continuously on the longitudinal axis. They respond definitely to light and orient fairly precisely, being positive under certain conditions and negative under others. Spondylomorum is not very common. It is usually found in stagnant pools rich in decaying organic matter. In October, 1912, it was found in great abundance by my colleague, Prof. E. A. Andrews, in a small puddle near a well frequented by ducks and chickens. I am greatly indebted to him for supplying me with numerous collections of this material. In the laboratory I succeeded for several months in raising the colonies in hay infu- sion, and in some cultures they became very abundant, but in January they all died out. I had intended to continue the work on these forms and delayed publication, hoping to obtain more laboratory cultures, but in this I have been unsuccessful. Two methods were used in ascertaining the effect of different chemicals on the sense of orientation in light: 1) Numerous colonies were put into about 1 ce. of solution from the culture jar or pure distilled water in a square watch-glass. This was then placed at a given distance from a window and left until the col- REVERSION IN ORIENTATION TO LIGHT 507 onies had collected either at the window side or at the room side of the dish, depending upon whether they were positive or nega- tive; then traces of the chemical to be tested were successively added, the solution being thoroughly stirred after each addition, until the sense of orientation changed or until there was no longer any orientation. The solutions were then tested for alkali with neutral red, and for acid with salts of neutral red prepared by treating neutral red with ammonium hydrate and washing the crystals produced in pure distilled water. 2) Ten watch- glasses containing a given amount of solution from the culture jar without any colonies were placed in the same illumination in front of a window. Chemicals to be tested were then added to the watch-glasses in such amounts as to make a series of solutions differing in concentration. A drop of solution from the culture jar containing colonies was then added to the solution in each watch-glass and the effect on the behavior of the colonies noted. In all cases the solutions were tested with neutral red. In some experiments distilled water was used in place of the solution from the culture jar and the colonies were washed in distilled water before they were used. In each experiment the temperature and illumination were prac- tically constant throughout. EFFECT OF CHEMICALS, GENERAL STATEMENT The results obtained in reference to the general effect of differ- ent chemicals when added to culture solutions may be summarized as follows: All the acids tested (carbonic, hydrochloric, nitric, sulphuric, formic, boric, chromic, tannic, tartaric, and oxalic), chloroform, ether, and chloral hydrate cause negative specimens to become strongly positive. They have no effect on positive specimens except perhaps to make them more strongly positive. There is also some evidence indicating that ethyl alcohol, ammonium chlorid and pure water induce reversion from negative to positive orientation, but if these substances actually have any effect it certainly is far less pronounced than that produced by any of the substances mentioned in the preceding paragraph. 508 Ss. O. MAST Formalin, sugar, oxygen, hydrogen peroxid, all of the alkalis tested (NaOH, KOH, NH,OH) and all of the salts tested except ammonium chlerid (MgSO,, NaCl, CaCl, KNO;) have no appre- ciable effect on the sense of orientation. These results, as presented, throw but little light on the ques- tion as to the factors involved in reversion. Let us, therefore, consider in detail the effects of the substances mentioned. EFFECT OF ACIDS If a little acid is added to a solution containing negative colonies of Spondylomorum, they usually become strongly positive almost immediately, but they soon become negative again. If now more acid is added, they become positive again, but in a few moments they are again negative. Thus they con- tinue to become positive after every addition of acid and then negative again until they are killed. These features in the response of Spondylomorum are clearly brought out in the fol- lowing details regarding two experiments: 1. On November 27, at 2.40 p.m., numerous colonies were put into a watch-glass containing about 1 cc. of solution from the culture jar and exposed in diffuse daylight. The colonies imme- diately collected at the side of the watch-glass away from the source of light. They were strongly negative. At 2.51 p.m., a trace of 10 per cent sulfuric acid was added, the solution being thoroughly stirred at the same time. At 2.53 p.m., the colonies were clearly slightly positive; but at 2.55 p.m., they were begin- ning to become negative, and at 3.00 p.m., they were strongly negative again. A trace of acid was now added, after which the colonies immediately became fairly strongly positive; but at 3.03 p.m., they were again negative. More acid was added at this time, and the colonies again immediately became distinctly positive. At 3.06 p.m., they were again negative. Acid was again added, and reversion followed immediately with a return to negative orientation at 3.10 p.m., when the process was again repeated with the same results. A trace more acid was now added. The colonies became inactive, but in a few moments they became active again and swam definitely toward the light, REVERSION IN ORIENTATION TO LIGHT 509 but three minutes later they were again negative. A minute trace of acid was again added, the colonies became inactive and all died in a few moments. 2. On December 6 sodium hydrate was added step by step to a solution containing numerous Spondylomorum colonies, until enough had been added to make the solution n/62 NaOH on the basis of pure water. It gave a faint alkaline test with litmus- paper. In the beginning of the experiment some of the colonies were positive, others were negative. At the close all were neu- tral, but no reversion in orientation had occurred. The following day the colonies in the alkaline solution were in excellent condi- tion, and five days later, December 12, they were still apparently normal and strongly negative. At this timen/100 HCI was slowly added step by step. The solution was tested with neutral red from time to time and the orientation was noted. The results obtained are presented in tabular form in table 1. The results obtained in these two experiments are essentially like those obtained in all of the numerous other similar tests made. They show clearly that the addition of acids causes nega- tive Spondylomorum to become positive, but they also seem to _show that the reversion in orientation produced by the addition of acids is not due to the direct effect of the acids on the organisms, for the colonies in most cases became positive before the alkalinity of the culture solution was neutralized. Moreover, the results obtained with pure distilled water support this contention. In distilled water of high! purity Spondylomorum lives for days and responds normally. In such solution the addition of a very small amount of acid proves fatal. For example, in one experiment, specimens put into n/15,000 HCl lived only a few moments. In n/6666 HCl they lived for some time, and in n/20,000 HCl they were still alive and apparently normal after two days. Whether or not addition of acid to pure distilled water causes reversion in orientation was not definitely ascertained, 1 The water used in these experiments was very generously supplied by the late Prof. H.C. Jones. It was redistilled from two Jena glass flasks in series, one containing chromic acid, the other barium hydrate, and condensed in a block-tin condenser. 510 S. O. MAST TABLE 1 The effect on the sense of orientation in light of adding acid to culture fluid TIME 10.55 10.56 10.57 10.58 10.59 11.00 11.03 11.04 11.07 11.09 2 bag 1a Aye 11.38 11.42 11.45 2.00 2.30 DY BY 2.42 2.55 3.00 December 13 a.m. 1.00 p.m. 1.30 2.00 2.10 24 hours later HCl concENTRA- TION ON BASIS OF PURE WATER n/1000 n/500 n/333 n/250 n/200 n/166 n/166 n/143 n/143 n/125 n/125 n/111 n/111 n/100 n/90 n/83 n/83 n/71 n/71 n/62 n/62 n/55 n/55 Acid grad- ually add- ed No acid add- ed Trace more acid added No acid add- ed No acid add- ed Trace more acid added SENSE OF ORIENTATION Negative Negative Negative Negative Negative Clearly but only slight- ly positive Negative Definitely positive Clearly negative More clearly positive Definitely negative Definitely positive Slightly negative Positive Positive Positive Fairly strongly negative Definitely positive Definitely negative Definitely positive Definitely negative Definitely positive Definitely negative Definitely positive Definitely negative Definitely positive Clearly negative Clearly negative All dead ALKALINITY AS INDICATED BY NEUTRAL RED Strongly alkaline Strongly alkaline Strongly alkaline Strongly alkaline Strongly alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Definitely alkaline Slightly alkaline Neutral? Definitely alkaline Very slightly acid Very slightly alkaline or neutral REVERSION IN ORIENTATION TO LIGHT Stik owing largely to the disappearance of the organisms in my cul- tures before sufficient tests were made. Ten distinct tests were made on two different days. In all of these tests the colonies were thoroughly washed in pure distilled water so as to remove all traces of the culture solution before the acid was added. The colonies were, under the conditions of the experiments, definitely negative in the pure water. In seven there was no indication of reversion in orientation. In one of the remaining tests some of the colonies probably became positive and in the other two a large majority of them clearly became positive. In these two tests the colonies had been in pure water overnight. They were tested in fresh pure water, but I am not certain as to whether or not they were washed in transferring them from the water in which they had been during the night to the fresh water. If pure water containing colonies is left in a watch-glass for some hours, it ordinarily becomes slightly alkaline. It is therefore possible that the solution in which these colonies were tested was slightly alkaline, and it may be that this is the reason why rever- sion was obtained in these tests and not in the others. It is consequently fairly certain that the reversion from negative to positive orientation in Spondylomorum is not due to the action on the organisms of acid or free hydrogen ions, as is maintained by some investigators. If this is true, reversion must be associated either with certain chemical compounds produced by the action of the acids, as, for example, salts, or with the concentration of the hydroxy] ions. There is no evidence at hand in favor of the former supposition, as will be demonstrated presently, but there is a certain amount of evidence in favor of the idea that reversion is associated with the concentration of hydroxyl ions. This evidence we shall now present. Under certain conditions, changes in the alkalinity of the culture solution produced marked effects on the sense of orientation; under others it did not. We shall discuss the latter first. 52 Ss. O. MAST EFFECT OF ALKALIS Experiments on the effect of changing the alkalinity were made at eight different times by adding sodium hydrate directly to the culture fluid containing positive colonies. Reversion was not obtained in any of these experiments, with the possible exception of those in which the colonies had been made positive by the addi- tion of acid, and in none of these were the results clear-cut and definite. Dilution of the culture fluid with pure distilled water also failed to produce definite results. The effect of such dilution in various degrees was tested in experiments made at nine different times. In four of these series of experiments the addition of pure water had no observable effect on the sense of orientation. In five the colonies became slightly positive when the pure water was added. To begin with, the colonies used in these five experiments were not strongly negative, and they soon became negative again after the water had made them slightly positive, and now the addition of more pure water did not make them positive again; on the contrary, it appeared to make them more strongly negative. It is consequently evident that reduction of alkalinity due to the addition of pure water was far less marked in its effect on rever- sion in orientation than reduction produced by the addition of acids. The results obtained by mixing culture solutions differing in alkalinity, as described below, were, however, quite as marked as those obtained by adding acids. A total of nine series of tests were made in studying the effect of such mixtures on the sense of orientation. In some the alka- linity was increased, in others it was decreased. ‘The concentra- tion of the alkaline in the solution used was changed either by slow evaporation or by adding sodium hydrate. The results obtained were definite in every test. They may be summarized as follows: 1. Negative colonies in a culture solution concentrated by evaporation become definitely positive if fresh culture solution is added. 2. Positive colonies in fresh culture solution become definitely negative if culture solution concentrated by evaporation is added. REVERSION IN ORIENTATION TO LIGHT als 3. Negative colonies in culture solution which has been made more alkaline by the addition of sodium hydrate, become defi- nitely positive if fresh culture solution is added. 4. Positive colonies in fresh culture solution become definitely negative if culture solution containing a little sodium hydrate is added. Reversion is, however, not permanent under any of these conditions. The results obtained in one of the experiments described in detail below are typical. Culture jars containing Spondylomorum were frequently left uncovered in the laboratory. In some of these, owing to evap- oration, the culture solution became considerably concentrated, and tests with neutral red showed that they were definitely more strongly alkaline than were the solutions in the covered jars in which evaporation had been prevented. On Dec. 18 numerous colonies were taken from one of these exposed cultures with some of the relatively strongly alkaline fluid and exposed in a given illumination. They were found to be strongly negative. After having been exposed for five minutes in this solution they were carefully transferred, without changing the illumination, to an equal amount of solution taken from a covered jar, i.e., to a solu- tion not so strongly alkaline as indicated by the neutral-red test. In this solution they were first momentarily negative, then they became strongly positive, and remained so for six minutes, when some of them began to swim away from the light. Three minutes later, practically all of them were strongly negative. These results show clearly that a decrease in the concentration of the salts and alkalis in a culture solution causes negative col- onies of Spondylomorum to become positive, that is, it has the same effect as addition of acids. Now addition of acids produces an increase in the salt contents and a decrease in alkalinity. It is consequently evident that if the effect of acid is due to its effect on compounds in the solution it is not due to the production of salts, but to the neutralization of alkalis. We thus have a considerable amount of evidence favoring the idea that reversion in the sense of orientation produced by the addition of acids is due to a change in the concentration of hydroxyl ions. But whatever the action of these ions may be in the process of rever- 514 Ss. O. MAST sion, it is certain that reversion is not specifically associated with their concentration, for organisms were repeatedly observed at different times to be negative as well as positive in practically all concentrations in which they oriented at all. Thus while the organisms tend to be positive in relatively weak and negative in relatively strong alkaline solutions, in four different experiments they were actually observed to be negative in solutions which were clearly slightly acid. Moreover, as may be observed by referring to the detailed description given above, of the effect of acids on the sense of orientation, the colonies change from positive to negative orientation without any appreciable change in the chemical constitution of the environment. This clearly shows that the sense of orientation is not specifically dependent upon the hydroxyl ion-concentration. Reversion is, however, not wholly independent of concentration. EFFECT OF CHEMICAL CONCENTRATION If acid, as previously stated, is added to a solution containing negative colonies they become positive, but after a certain time they become negative again without any further change in the solution. The time required for the reversion from positive to negative orientation depends upon the amount of acid added. For example, in one experiment, with sufficient hydrochloric acid added to make the culture fluid n/10,000 HCl, on the basis of pure water, it required approximately 3 minutes; with sufficient added to make it n/500 HCl, it required 3 minutes; with sufficient to make it n/333 HCl, it required 5 minutes; with sufficient to make it n/250 HCl, it required 6 minutes; with sufficient to make it n/200 HCl, it required a little over 8 minutes, and with sufficient to make it n/167 HCl, it required 84 minutes. In another ex- periment with sufficient hydrochloric acid added to make the culture solution n/1000 HCl, thé colonies did not become posi- tive at all; with sufficient added to make the solution n/200 HCl, they become definitely positive, and it required 6 minutes to be- come negative; with enough acid added to make n/143 HCl, they remained strongly positive for nearly one hour and then became neutral; whether or not they became negative later was not REVERSION IN ORIENTATION TO LIGHT a15 ascertained. The solution gave a definite acid reaction with neutral red. The following day all of the colonies were dead and the solution was still distinctly acid. The fact that the time required for positive individuals to become negative is greatly extended when the proper amount of acid is added, seems to indicate that under certain conditions the colonies may be permanently positive. I was, however, unable to obtaiti such conditions. While it has thus been clearly demonstrated that concentration is a factor in reversion, the following experimental results demon- strate equally clearly that the time rate of change in the concen- tration of the effective elements in the solution is also a factor. EFFECT OF TIME-RATE OF CHANGE IN CHEMICAL CONCENTRATION On December 5 numerous colonies in 1 ec. of solution were taken from an old culture from which considerable water had evaporated and exposed in diffuse light. The colonies were strongly negative. Hydrochloric acid was now added in sufh- cient quantity to make the solution n/1000 HCl on the basis of pure water. The HCl was now increased step by step through the following concentrations: n/500, n/333, n/25Q, n/200, n/166, n/142, n/125, n/111, n/100. During all this addition of acid there was at no time any indication of a reversal to positive reactions. The colonies were continuously definitely negative. But when colonies were taken from the same jar and put directly into the n/200 solution or into any of the solutions above this, they became strongly positive, remained so for several minutes, and then became negative again. These results show clearly that the effect on reversion in orientation in Spondylomorum of slowly adding a certain amount of acid to a given amount of culture fluid is very different from the effect produced by rapidly adding the same relative amount of acid. We have previously demonstrated that the effect of acid on reversion is in all probability associated with the accompanying reduction in hydroxyl ions. If this is true, it may be concluded from the results now under consideration that it is dependent upon the time-rate of change in the concentration of the hydroxyl ions. ; 516 Ss. O. MAST Whatever the processes within the organism may be that are induced by the environment in the production of changes in the sense of orientation, the same processes may, at least in part, occur without any immediate action of the environment as the following evidence shows. EFFECT OF CHANGE IN PHYSIOLOGICAL STATES As previously set forth, if the proper amount of acid is added to a culture fluid in which Spondylomorum is negative, it becomes positive, but ordinarily it remains positive only a few minutes and then becomes negative again. If now fresh negative speci- mens from the same culture jar are added to the solution con- taining the acid, they respond just as the first specimens did when the acid was added. They may be momentarily negative, but they soon become positive and remain so for a few minutes, after which they become negative and remain so. This is clearly seen in the experiment described on page 508. It was observed repeatedly in other experiments. This demon- strates conclusively that the reversion from positive to negative orientation is not due to a change in the solution, but to a change within the organism, for if it were due to a change in the solution, the second lot of individuals, when added, should have responded just as the first set was responding at that time. EFFECT OF ANESTHETICS As previously stated, the addition of chloroform, ether, or chloral hydrate like the addition of acids causes negative colonies of spondylomorum to become positive, and the reactions of the colonies to these substances in reference to orientation are in detail essentially the same as their reactions to acids. If a trace of chloroform, for example, is added to a culture solution containing negative colonies, they become, in the course of a few moments, strongly positive. Then after a few minutes they become nega- tive again. If now more chloroform is added, they again become positive and after a few minutes again negative, etc., until the chloroform becomes sufficiently concentrated to produce anes- thesia. The reversion from positive to negative orientation is REVERSION IN ORIENTATION TO LIGHT Bry. not due to evaporation of chloroform, for if, after this reversion has occurred, fresh negative specimens are added to the solution, they become positive at once and then after a few minutes nega- tive. That is, they respond just like the original colonies did immediately after the chloroform was first added. The results obtained with chloroform and ether were quite as definite and conclusive as were those obtained with acids, but those obtained with chloral hydrate were much less definite. In reference to the acids, it was concluded that their action on orientation. is probably associated with the time-rate of change in the concentration of the hydroxyl ionsin the culture solution. The action of the chloral hydrate may have been due to the same cause, for it was slightly acid in reaction, but the action of the chloroform and ether could not have been due to this for both gave distinct alkaline reactions with neutral red. Moreover, it was found that chloroform produces reversion in negative col- onies in chemically pure water, 1.e., in a solution practically devoid of free hydroxy] ions. EFFECT OF TEMPERATURE AND ILLUMINATION The effect of illumination and temperature on the sense of ori- entation was not intensively studied. But the observations made indicate that strong light tends to induce negative, and weak light positive orientation, and they demonstrate clearly that an increase in temperature tends to make Spondylomorum positive and a decrease tends to make it negative. The relation between temperature or illumination and the sense of orientation is, however, not specific. For example, under certain conditions, colonies were found to be positive in all temperatures above 14° and negative in all below about 10°, while under other conditions the same colonies were found to be negative up to nearly 42° and positive above this temperature. Similar results were re- peatedly obtained. Moreover, under certain conditions, rever- sion in orientation could not be induced by changing the tem- perature, and under other conditions it was observed without any change in temperature. Changes from positive to negative orientation in constant temperature are particularly prevalent 518 S. O. MAST after the opposite reversion has been induced by an increase in temperature. Similar results were obtained in observations on the effect of light. Thus, it is evident, that the temperature or the illumination required to produce reversion in orientation in given colonies varies greatly. It should be emphasized here that the effect of light and temperature are opposite, 1.e., an in- crease in temperature produces the same effect as a decrease in light and vice versa. DISCUSSION 7 The experimental results presented in the preceding pages indicate that reduction in alkalinity, increase in anesthetics, increase in temperature, and decrease in illumination, all have the same effect on the sense of orientation in Spondylomorum. They also indicate that the same change in the sense of orientation may occur without any appreciable change in the environment. It is consequently probable that reversion in orientation is due to some specific change in the physiological processes of the organ- ism which can be induced by changes in any one of the environ- mental factors mentioned, 1.e., alkalis, anesthetics, temperature, and light. What these processes are is not known. They may involve electrical tension and polarization or permeability or absorption. We have demonstrated that reversion depends upon the time- rate of change of effective factors in the environment. This, however, does not prove that reversion is dependent upon the time-rate of change in the physiological processes involved rather than upon the state of the physiological processes as such, for we have also demonstrated that reversion induced by a given environmental condition is rarely, if ever, permanent. That is, if negative colonies are put into a solution containing the proper amount of chloroform, they become positive, but they usually remain positive only a few minutes and then become negative again. This indicates that there is rapid adjustment on the part of the organism to the new environmental condition. The fact, therefore, that reversion is not induced if the effective factors in the environment are slowly changed, may be due to adjustment REVERSION IN ORIENTATION TO LIGHT 519 proceeding at the same rate as the environmental changes and not to lack of sufficient speed in the rate of change of physiological processes. SUMMARY 1. Spondylomorum orients fairly accurately in light. It is negative under certain conditions and positive under others. 2. Chloroform, ether, chloral hydrate, all of the acids tested (carbonic, hydrochloric, nitric, sulfuric, formic, boric, chromic, tannic, tartaric, oxalic) when added to the culture solution cause negative specimens to become strongly positive. They have no effect on positive specimens except perhaps to make them more strongly positive. 3. Ethyl aleohol, ammonium chlorid, and pure water have no appreciable effect on positive colonies, but they probably cause negative colonies to become slightly positive. 4. Formalin, sugar, oxygen, hydrogen peroxid, magnesium sulfate, calcium chlorid, potassium nitrate, and all of the alkalis tested (sodium, potassium, and ammonium hydrate) have no appreciable effect on the sense of orientation. 5. Increase in the concentration of the culture solution pro- duced by adding culture solution part of which has evaporated or to which sodium hydrate has been added causes positive colonies to become strongly negative. Decrease in concentration produced by adding a less concentrated culture solution causes negative colonies to become strongly positive. 6. Increase in temperature and decrease in illumination cause negative colonies to become positive. Decrease in tem- perature and increase in illumination cause positive colonies to become negative. 7. The sense of orientation is not specifically related to the concentration of chemicals in the environment. Spondylo- morum probably may be either positive or negative in any solu- tion in which it orients at all. 8. The effect of acids on the sense of orientation is probably due to the reduction of hydroxyl ions produced in the culture solution by the acids. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 26, NO. 3 520 S. O. MAST 9. Reduction in the concentration of hydroxyl] ions, increase in anesthetics, increase in temperature, and decrease in light, all produce the same reversion in the sense of orientation, and this reversion may also occur without any change in the environment. It is, therefore, probably due to some specific change in the phys- iological process in the organism, which may be induced by a number of different factors. 10. Reversion depends upon the time-rate of change in the concentration or intensity of the effective factors in the environ- ment, but it has not been demonstrated that it depends upon the time-rate of change in the physiological processes which are involved in reversion. LITERATURE CITED? ALLEE, W. C. 1918 Reversal of phototaxis and carbon dioxide production in May-fly nymphs. Anat. Rec., vol. 14, pp. 95-96. Arisz, W.H. 1915 Untersuchungen iiber den Phototropismus. Ex. d. Ree. d. Trav. bot. Neer., vol. 12, pp. 44-2138. Esterty, C. O. 1917 Specificity in behavior and the relation between habits in nature and reactions in the laboratory. Univ. of Cal. Pub. in Zool., vol. 16, pp. 381-392. Ewatp, W. F. 1914 Versuche zur Analyse der Licht und Farbenreaktionen eines Wirbellosen (Daphnia pulex). Zeitsch. f. Sinnesphysiol., Bd. 48, S. 285-324. GAMBLE, F.. W., AND KEEBLE, F.° 1903 The bionomics of Convoluta roscoffensis, with special reference to its green cells. Quar. Jour. Mier. Sci., vol. 47, pp. 363-431. Homes, 8. J. 1901 Phototaxis in Amphipoda. Am. Jour. Phys., vol. 5, pp. 211-234. 1905 The reactions of Ranatra to light. Jour. Comp. Neur. and Psye., vol. 15, pp. 305-349. 1906 Studies in animal behavior. Boston, 266 pp. Lors, J. 1905 Studies in general physiology. Chicago, vol. 1, 423 pp. Mast, S.O. 1907 Light reactions in lower organisms, II. Volvox. Jour. Comp. Neur. and Psye., vol. 17, pp. 99-180. 1911 Light and the behavior of organisms. New York, 410 pp. Ostwatp, W. 1907 Zur Theorie der Richtungsbewegung niederer schwimmender Organismen, III. Arch f. d. ges. Phys., Bd. 117, S. 384408. WasHBURN, MarcaretF. 1917 Theanimalmind. Secondedition. New York, 386 pp. ? For a fairly complete list of references to the literature on reversal in the sense of orientation, see Washburn (’17, pp. 200-214), Holmes (’16, pp. 116-119), and Mast (’11, pp. 264-287). : AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JULY 19 RELATIVE EFFECTIVENESS OF FOOD, OXYGEN, AND OTHER SUBSTANCES IN CAUSING OR PRE- VENTING MALE-PRODUCTION IN HYDATINA! A. FRANKLIN SHULL Zoological Laboratory, University of Michigan CONTENTS RirntimachuiC CLOT ees ts rn a Are eS occ ns Sl ace oe RS eur act 521 BRE PEREMEDUS MA tars oe chem net ares Met. cols 2d DARIN ASB Ae stars soe Biteciwor GO perecentsoxy. enya ae asf: a OO eae hee thE 522 Amount of oxygen dissolved from atmosphere and from Euglena....... 525 ID yiReroteollionieraiae COpatecins J, 55s! Pawan eee atettn Feo ves beams cob es 526 Oxycen derived trom photosymebesiss. 0.0). esses ee ee fens ce 529 Englena versus oxygen and other substances as male-producing or male- RAPLessin ee a Ments ei ss > Ae Ree aia) Ses 4 oe tee deeb bere Sheed 532 inespoure tovoxymeennan: Nebraska limeG@ es. . ja. o 65s .ciycct re «eee ie eee 538 TD ISCUSHTO TPE eas reer cee va at. fc! ora RMR EES See et ae n,n ea cc ream eR ge 539 SVERTTMT SL, AN BRR EO PMR lt 2 <4 ht a eee ee eee A 543 IBY GIIMO PDO NN: Oh EID cle EEmCRER eT oLo E ois sl ola oie aan ots Oka Ehscicn alae ced Ao De 544 INTRODUCTION A paper on the effect of oxygen on the life cycle of the rotifer Hydatina senta by Shull and Ladoff (16) left certain matters in doubt. Whereas water saturated with a mixture of air and oxygen which contained 40'per cent of oxygen caused an increase in the ratio of male-producers to female-producers. saturation with a mixture containing 60 per cent of oxygen was attended by practically no results. Two explanations seemed possible. Either different concentrations of oxygen had different effects or oxygen was incapable of increasing male-production when male-production was already high from other causes. The 1This investigation was aided by a grant from the Bache fund of the National Academy of Sciences, to the trustees of which the author desires to make grate- ful acknowledgment. 521 Epp ‘A. FRANKLIN SHULL experiment referred to above, in which 60 per.cent oxygen pro- duced no effect, was performed at a time when the rotifers were going through one of their well-known ‘epidemics’ of male-pro- duction (Shull, 715 b), whereas the experiment with 40 per cent oxygen came at a period when male-production in the control line was low. No opportunity was found to test these two possible explanations, owing to the loss of the stock cultures of rotifers, and the experiments were published while this question was still unanswered. The important discovery by Whitney (14) that rotifers fed upon a green organism, Chlamydomonas, produced many more male-producing offspring than those fed upon other material, raised new questions. Whitney took no account of certain initial chemical differences between the contrasted lines nor of the fact that green organisms yield oxygen as a by-product of photosynthesis. Shull and Ladoff pointed out that although nutrition might on a priori grounds be expected to have the effect which Whitney ascribed to it, the other obvious factors should be eliminated before attributing the residual effect to nutrition. It remained a question, therefore, how much of the increase in male-production which Whitney observed upon feeding with green organisms was due to food, how much to.oxygen or other factors. In the experiments described in this paper an answer to these questions is found. The initial chemical differences referred to above have been eliminated. The effects of oxygen and green organisms have been measured and compared. In addition, the effects of these two agents, which increase male-production, are compared with two of the principal agents known to decrease male-production. EXPERIMENTS Effect of sixty per cent oxygen To show whether saturation with a mixture of air and oxygen of which 60 per cent was oxygen caused any change in the amount of male-production, the following experiments were performed: Experiment 1. On each of the days named in table 1 three female rotifers were placed in each of two dishes. In one was MALE-PRODUCTION IN HYDATINA 523 poured some spring water and scum from a manure solution was added for food. The water put into the other dish was first saturated, by continued agitation, with an air-oxygen mixture produced by removing half the volume of air and replacing it with oxygen. Such a mixture must have been composed of about 60 per cent oxygen and 40 per cent nitrogen. Manure scum was added for food, as in the control dish. The dish was then placed under a sealed bell jar, from which half the air was withdrawn by a filter pump and replaced with oxygen. The air-oxygen mixture under the bell jar must have consisted of about 60 per cent oxygen and about 40 per cent nitrogen.? After twenty-four hours all the parents were removed from these dishes. The young hatching from the eggs laid in that period, experiment and control alike, were reared to maturit ’ in untreated water with manure scum as food. Whatever differ- ences they exhibit, therefore, are due to differences in the condi- tions which affected the mother or early developmental stages. Table 1 shows the results. The excess of male-production in the presence of oxygen is fairly marked. Experiment 2. This experiment was designed to contrast the effects of 60 per cent and 40 per cent oxygen mixtures by two simultaneous tests. In the absence of a second suitable bell jar, one test was made with 60 per cent oxygen, the next test with 40 per cent oxygen. It was hoped that by this alternation an epi- demic of male-production, if one occurred, would occur in the control of both series of tests. Since, however, it was not feas- ible to make one test each day, and several days usually elapsed between tests, that object was only partially attained. It happened that on the whole the line used for control produced more male-producers on the days when 60 per cent oxygen was employed than on the days on which the 40 per cent mixture was used. How much this circumstance vitiates the conclusion to be drawn from the experiment is not known. ?The method of procuring air-oxygen mixtures of a given composition is here stated again specifically, since Whitney seems to have misunderstood the expressions ‘60 per cent oxygen’ and ‘40 per cent oxygen.’ 524 A. FRANKLIN SHULL TABLE 1 Showing the effect of saturation of the water with an air-oxygen mixture, of which 60 per cent was oxygen, upon the ratio of maleproducers (o Q ) to female- producers (9 9) in the rotifer Hydatina senta AIR OXYGEN DATE Number of o&?}|Number of 9 9; Number of oc? |Number of 2 9 HGU RUA YE 2Os 6k, oaks Wiehe 2 6 Pe 16 ebruanyacGe er foes ne 1 16 2 12 Ber ayee lees b es widectae 6 30 1 21 Behmuanyi27 a Ace. Fi hl aa 9 27 15 14 elpuary 20n eha. cst oe meee 4 28 17 19 Mitrehoies ye anc. ne wet 0 20 2 24 0 WI ay Penn A MRA rel ges rope 22 127 39 106 Percentage of @9......:.. 14.8 26.9 The method of carrying out the tests was precisely the same as described for experiment 1, except that in the 40 per cent oxygen tests the mixture used in saturating the water and the mixture under the bell jar were made up of three parts air and one part oxygen, instead of equal parts of the two as in the 60 per cent oxygen mixture. Table 2 shows the results of the experiment. In both of the tests the oxygen mixture produced more male-producers than the control. If the absolute difference between experiment and control be taken as the measure of effectiveness of oxygen, the 60 per cent mixture appears to be a little more potent than the 40 per cent. If the ratio of male-production in the experiment to male-production in the control is the proper measure of effec- tiveness, then 40 per cent oxygen produced greater results than 60 per cent oxygen. In other experiments the impression has been gained that the relative difference between experiment and control is a better measure of the effect produced than is the absolute difference. In this experiment the ratio of the per- centage of male-production in oxygen to the percentage of male- production in the control is 1.51 in the case of the 40 per cent oxygen, 1.41 in the case of the 60 per cent oxygen. The fourteen tests made are too few to use statistically to determine whether MALE-PRODUCTION IN HYDATINA 525 TABLE 2 A comparison of the effectiveness of a 60 per cent oxygen mixture and a 40 per cent oxygen mixture in increasing male-production in the rotifer Hydatina senta. The tests covered the same period of three months, but were made on alternating days ge Meee a4) ee DATE Eo. | Bor | Bor | Sloe DATE Bo | So | So | So Boece ao esha | ule Bel ee | Zh | Ze Be} Ze | ge | 28 March iGaacane a: |) ood 10 20n ie Viarehn eee 1 25 1 27 Marchol3 4s se 0 22 Ane 3S) leiViarch alee 6 | 22 2| 38 ManchvliSieseus oe. 0 29 0 So eMiarchi2 Ose 0 25 3 21 Miar:chy2o esses er 0 1 2 Guiearchy2 7 12 11 355 2 AME Atay ye ais. 3 1l7/ 8 GR | PeAgO THIN (eee 1 4 0 6 April 10 0 16 2 || QO Aho. 0 14 17 9 Anrl Ties Viet. 14 18 18 Ze | ceNore ee ke 1 30 Qin 29 Sao A (Oe ey, ea | | 22S Saleecon eA ri 20s. ener 3) 38 Su) 45 Py Oe) PADS eae Bic 3) | 20) 040525) | April 28 1| 44 On ead May 2 4 15 of TIMI eTe oa bone « 17 23 12 17 May 9 iL eal i |] alas | IME oe 1 1 Oneal, IMaiys VO ets cases sc 8 | 28 1 18} Wiley? BOs se e6o8: 0 10 Onl 323 June 4 0 9 Ones s uanet*5e0 5. eee Bil) Be Sul) 22 PME See sew 1125 OnlmtGaleune 12.) 8. HOPES OAL ocala wanes: c: 62) 1° 257) 103) 273 46) | 287 | 83) 312 Percentage of SONS CARRE 19.4 27.4 13.8 21.0 the difference between these two ratios is probably significant or not. Amount of oxygen dissolved from atmosphere and from Euglena After repeated tests had shown that the increase of male-produc- tion in rotifers reared in oxygenated water was no mere accident, it became a pertinent question whether, and to what degree, the increase of male-production among rotifers fed upon green organ- isms is dependent upon the liberation of oxygen in photosynthe- sis. As a preliminary problem, it was necessary to discover how nearly the amount of oxygen dissolved from the air-oxygen mix- tures described in this paper and in that of Shull and Ladoff 526 A. FRANKLIN SHULL equaled that dissolved in water in which green organisms were kept as food. A number of tests to determine this relation were made, some of them closely resembling the experiments, others differing more or less, but bearing directly on the oxygen content of treated water. The oxygen in solution was measured by the Winkler method. In this paper all measurements of oxygen are given in number of cubic centimeters per liter of water. This quantity is com- puted from the tests by means of the formula 5D.820 b n i tov in which ‘b’ is the number of cubic centimeters of potassium bichrcmate solution used in standardizing the scdium thiosul- phate; ‘t’ is the number of cubic centimeters of sodium thio- sulphate required in titration against ‘b’ ec. of potassium bi- chromate; ‘v’ is the volume in cubic centimeters of the water sample tested; and ‘n’ is the number of cubic centimeters of the thiosulphate required for the water sample. In the tests here described ’b’ was always 25, the other factors being variable. Direct solution of oxygen.—In table 3 are described the tests to determine the amounts of oxygen dissolved in water when the water is agitated in contact with an atmosphere containing an excess of oxygen. ‘Test No. 1, consisting of seven parts, was most nearly like one of the oxygen experiments, and was furthermore -most complete. This test indicates that the water used in one of the rotifer experiments contained 6.81 ec. of oxygen per liter at the outset (a). After it was shaken with an atmosphere of which 40 per cent was oxygen, it contained 11.33 ec. of oxygen per liter (b). When the food (manure scum) was added, the water con- tained only 7.58 ce. of oxygen (c). After standing a day under a bell jar, in an atmosphere of which 40 per cent was oxygen, the oxygen content of the water had risen to 8.35 ec. per liter (d). Had the dish been kept in air instead of under a bell jar, the oxygen content would have fallen to 5.43 ec. per liter (e). — TABLE 3 Showing the effect of agitation of water in contact with an atmosphere containing an excess of oxygen in increasing the oxygen content of the water. The columns of figures were obtained by Winkler’s method of determi- nation of dissolved oxygen ° NUMBER OF CUBIC ry WN POR COMPUTATION |. CENT eon TREATMENT OF WATER (SEE TEXT) Bees PER LITER (n) (t) (v) (0) l-a Untreated water, without food or any- thing else, tested at once for oxygen...| 8.9 | 20.7 | 88 6.81 b Water shaken with 40 per cent oxygen at- mosphere, tested immediately for COURIC dete as erate ot «0-615 e Oe eee 14.8 | 20.7 | 88 11533 c Shaken with 40 per cent oxygen atmos- phere, manure scum added, immedi- ately filtered and tested for oxygen....| 9.9 | 20.7) 88 7.58 d Shaken with 40 per cent oxygen, manure scum added, put under bell jar in 40 per cent oxygen atmosphere for a day, then: filtered*and tested a...) .....05.2 KOE) |) PAVE A |] Yets) 8.35 e Shaken with 40 per cent oxygen, manure scum added, kept in aira day, filtered AMOR GEStCO. meaty ists A eRe eke ore tes). Moll | AWE A |) ste: 5.48 f Untreated water, manure scum added, then immediately filtered and tested...| 7.6 | 20.7 | 88 5.82 g Untreated water, manure scum added, kept in air a day, then filtered and GESPEU EME Areal ee, Sb) SEIT R aA ees he 8 t: 5.4 | 20.7} 88 4.13 2-a Shaken with 60 per cent oxygen atmos- phere and tested immediately......... 14.50) 20.8 | 55 17.67 b | Shaken with 60 per cent oxygen, poured into six watch-glasses, kept under bell jar in 60 per cent oxygen atmosphere one tday, then tested: an2 a= 9.95) 20.8) 55 ieal3 c Poured into six watch-glasses, set be- side bell jar one day, then tested...... 6.05) 20.8 | 55 7.88 3-4, Untreated water in twelve watch-glasses, manure scum added, left twenty-four MOUS sien tes teas: ase weet ten e f 5.97) 20.7 | 88 4.57 b | Shaken with 80 per cent oxygen atmos- phere, then tested at once for oxygen. .| 29.25) 20.7 | 88 22.40 c Shaken with 80 per cent oxygen, put into twelve watch-glasses, manure scum added, kept under bell jar in 40 per ; cent oxygen twenty-four hours, then {ICSE 0 sry RE Be REE + 2 en 10.23) 20.7 | 88 7.73 4-a | Untreated water tested for oxygen........| 18.3 | 20.7 | 120 7.47 b | Shaken with pure oxygen and tested im- mediately moroxycen oy neue. 3 40.5 | 20.7 | 120 22.75 c Shaken with pure oxygen, kept forty- eight hours in atmosphere of 60 per COMGLORV GONE rare oh siete wel oa He eats SYA POC ype leze) 18.14 528 A. FRANKLIN SHULL The control started with 6.81 cc. of oxygen per liter in the un- treated water (a). Upon the addition of food (manure scum) the oxygen content fell to 5.82 cc. per liter (f); while after standing a day it further fell to 4.13 cc. per liter (g) . Assuming that the oxygen content of the water in both halves of the experiment changed uniformly during the twenty-four hours in which eggs were being laid, the mean oxygen content of the oxygenated water was 7.08 + 8.35 5 = 7.965 ce: per liter; while the mean oxygen content of the control dishes was 5.82 + 4.13 5 = 4.975 ce. per liter. Absolutely, the difference between experiment and control was about 3 ce. per liter. Relatively, the oxygenated water contained 60 per cent more oxygen than did the untreated water. Test No. 2 (table 3) is less complete than the foregoing, but indicates approximately the amount of oxygen involved in the 60 per cent oxygen experiments. No test of untreated waterat the beginning of the twenty-four hour period described was made and no food was added. It is likely that under these circum- stances the untreated water exposed to air would not lose oxygen, but might gain it, so that the average content of such water dur- ing the twenty-four hours would be not greater than 7.88 cc. per liter (c). The water shaken with 60 per cent oxygen must have averaged about 15 ce. per liter (a and b), or nearly double the content of the untreated water. This test was not very relevant, but since the comparison was to be made between the 40 per cent oxygen experiments and those in which green food was used, a more accurate test of the 60 per cent oxygen experiments was not made. Tests 3 and 4 have little bearing on the experiments, and are recorded for whatever interest may attach to the solubility of oxygen at higher pressures. ‘ MALE-PRODUCTION IN HYDATINA 529 Oxygen derived from photosynthesis. Euglena was used in these tests. Cultures of a species that formed a sheet of green animals in a somewhat inactive, though not encysted, state along the sides of the aquarium were maintained. A considerable quantity of Euglena was usually available, and it was possible to make nearly as thorough a test as was made with manure scum. Table 4 states in tabular form the nature of the tests and their results. Test No. 5, of six parts, gives the best idea of the Kuglena ex- periments described in this paper. If the results of this test be applied to one of the Euglena-rotifer experiments, presently to be described, they indicate that the water used in the experiment contained 6.05 ec. of oxygen per liter at the outset (a). When Euglena was added as food, the water contained 6.58 cc. of oxygen per liter (b). After four hours in direct sunlight, it contained 9.27 ec. per liter (c), but on standing overnight the oxygen content fell to 7.35 ec. per liter (d). In the control the water contained 6.05 ee. per liter at the outset (a); 5.51 ec. after manure scum was added (e), and 4.63 ce. per liter after the dish had stood twenty hours (f). It should be remarked that probably neither the Euglena nor the manure scum used in test No. 5 was as abundant as in one of the rotifer experiments, though the difference could not have been great. Accepting these results as typical, it appears that the mean oxygen content in one of the Euglena cultures was z kK ee + 9.27 ¥ 1) (= " 1.00 y 16) 2 (BETO 725082,06; 5 20 « 20 8.233 ce. per liter while the mean oxygen content of the control was 5.01 + 4.63 5 = 5.07 cc. per liter. Thus the absolute excess of oxygen in the Euglena dishes over the control was about 3.16 ec. per liter. Relatively the Euglena dishes contained about 62 per cent more oxygen than the manure-scum cultures. As was pointed out above, the quantity 530 A. FRANKLIN SHULL TABLE 4 \ Showing the effect of Euglena in increasing the oxygen content of water, particularly as contrasted with water containing manure scum. The columns of figures are the results obtained by Winkler’s method of determination of dissolved oxygen NUMBER VARIABLES IN FORMULA baal She FOR COMPUTATION TEST a METERS OF SEE TEXT NoMEee TREATMENT OF WATER ( ) paar PER LITER (n) (t) (v) ©) 5-a Spring water, without treatment, tested ALONG TOL ONV LCE eer emery Tame cis; | PADS vets) 6.05 b Spring water, Euglena added, then im- mediately filtered and tested foroxygen.| 8.6 | 20.7 | 88 6.58 c Spring water, Euglena added, kept in direct sunlight four hours, then filtered SN GES tS Cees Ae yee re iene ak wat cae ie PN AADETE | ists: 9.27 d Spring water, Euglena added, kept four hours in direct sunlight, sixteen hours in diffuse light and darkness, filtered AN GUsteSUCHM aceate nae Pee eer vice scien tet 9.6 | 20.7 | 88 7.385 e Spring water, manure scum added, then immediately filtered and tested....... e2aeZ0R(@ RSs broil f Spring water, manure scum added, left twenty hours, then filtered and tested. .| 5.4 | 20.7) 88 4.63 6-a | Spring water in flat dish, Euglena added, kept cool in direct sunlight four and one half hours, then filtered and tested HOP LOK CEMA AA Are MR, oe es oc eels 8.8 | 21.5] 88 6.49 b Spring water in flat dish, manure scum added, kept in diffuse light four and one half hours, then filtered and tested.| 3.0 | 21.5] 88 Py PAL 7-a Untreated water in ten watch-glasses, Euglena and a little manure solution without scum added, kept cool in di- ' rect sunlight five hours, then in diffuse light and darkness fourteen hours, filtenedsancdateste Gere weiss oe lace Sema 20kto| cod Heth b Untreated water in ten watch-glasses, manure scum and a little water from Euglena culture (but without Euglena) added, kept in diffuse light and dark- ness nineteen hours, then filtered and VESLEC ete a. coum re oie ers Leiner sel eee 3.83] 20.75) 67 3.84 MALE-PRODUCTION IN HYDATINA Dol TABLE 4—Continued NUMBER OF CUBIC CENTI- METERS OF VARIABLES IN FORMULA FOR COMPUTATION TEST (SEE TEX aD TREATMENT OF WATER T) OXYGEN PER LITER (n) (t) (v) (0) 8-a Spring water in bottle, Euglena and a little manure solution without scum added, bottle sealed without air bub- ble, agitated on clinostat for two hours in direct sunlight and twenty-two hours in diffuse ight and darkness, fil- ered and Gestedecmemane Miiveees )cok a Meow 16.91} 20.75} 115 9.89 b | Spring water in bottle, manure scum and a little Euglena water (but without Euglena) added, bottle sealed without air bubble, agitated on clinostat two hours in direct sunlight and twenty- two hours in diffuse light and darkness, filtered andstestedse esses. 5. 4.1 | 20.75) 115 2.39 9-a Spring water in bottle, Euglena added, bottle sealed without air bubble, agi- tated on clinostat thirty-two hours in diffuse light and darkness, filtered and CES UEC Vey cis Byer bs anne Peers sail OD ZOE ces lets 11.84 b Spring water in bottle, manure scum added, bottle sealed without air bub- ble, agitated on clinostat thirty-two hours in diffuse light and darkness, MIGeLeds and eStedl sae eenine seria: oi Sap08 20 edie lle 2.08 of Euglena used in the rotifer experiments was probably greater than in the oxygen tests just described, so that the excess of oxygen in the Euglena dishes in the experiments was probably somewhat more than 62 per cent. The decrease of the oxygen content of the water over night, in the Euglena culture, may be in part due to other organisms, since my Huglena culture was not quite pure. In this respect the tests described here apply more correctly to the rotifer experi- ments of this paper than they do to Whitney’s experiments. Probably, however, a greater part of this overnight decrease is 532 A. FRANKLIN SHULL due to escape of oxygen into the air. Quiet water in contact with ordinary atmosphere will not, as a rule at least, retain as much as 9.27 ec. of oxygen per liter. The remaining tests in table 4 (Nos. 6 to 9) are so unlike any of the rotifer experiments that it is unnecessary to explain them further than they are explained in the table. All of them con- tribute, however, to the proof that Euglena under various cir- cumstances markedly increases the oxygen content of the water in which it lives. Euglena versus oxygen and other substances as male-producing or male-repressing agents Since the quantity of oxygen produced by Euglena and dis- solved in the water is about the same as that dissolved directly from the atmosphere in the rotifer experiments, the effects of Kuglena and of oxygen in increasing male-production may be directly compared. Any excess of male-production in the Euglena experiments over that in the oxygen experiments may therefore be attributed to another factor, probably the food. The following experiments were designed to compare the effectiveness of Euglena with that of oxygen among the male- increasing agents, and with manure solution and creatin among the male-repressing agents. Experiment 3. Euglena versus oxygen. In each of three dishes was placed the same number of rotifers, the number rang- ing from four to eight on different days. Two of the dishes were filled with spring water; to one of these Euglena was added as food, to the other manure scum. In the third dish was placed water that had first been saturated with an air-oxygen mixture composed of about 40 per cent of oxygen and 60 per cent of nitro- gen. Manure scum was added to the water in the third dish, which was then set under a bell jar in an atmosphere of which 40 per cent was oxygen. ‘The chemical composition of the medium in the three dishes at the beginning of the experiment was equalized by adding to the Euglena-fed culture as much manure solution (without scum) as was added to the other two dishes with the scum, and by adding to the manure-scum cultures as MALE-PRODUCTION IN HYDATINA 533 much water (without Euglena) from the Euglena stock as was introduced into the first dish with Euglena. It was easy to obtain this water practically free of Euglena, since the Euglena was a quiescent form that produced an incrustation along the sides of the jar, while very few individuals were actively swimming. The parent rotifers were removed from their dishes from twenty-four to forty hours later. All young rotifers hatching from eggs laid in that time were reared to maturity, in all three lots alike, in spring water with manure scum as food. The male- producers and female-producers are recorded in table 5. TABLE 5 Comparison of the effects of Huglena and manure scum as food, and of oxygenated and untreated water, upon the proportion of male-producers in the rotifer Hydatina senta EUGLENA AS FOOD IN | MANURE SCUM IN OX¥- } MANURE SCUM IN UN- UNTREATED WATER GENATED WATER TREATED WATER DATE Number of| Number of}Number of|Number of|Number of|Number of omy 99 oe? fot tS) INarche2 Onmiscry-tcnese 13 43 1 22 2 37 Wiamchao Zeer nage 21 33 15 33 1 21 INNO PBs Sey Gen oe ooe 14 18 0 22 0 21 Vim che2baeenne sess ac 29 28 1 29 0 18 IMGT CHRO Se sesh nia 3 Des 5 9 0 1 Mime 298 eek tn. ee 25 55 2 at 3 74 hy STN > 7 a RR 4 62 4 41 10 44 lbetre Sos) NC 61 4 64 0 59 yoy er ld oe ae Neate 1 46 0 28 0 35 7 Ny over MLC} Men epee ew 1 16 23 45 13 58 April Oe ere 3 68 0) 53 3 57 Avpril 23) SR akin rs 6 27 0 20 0) 5 April 255s: san ieee ae 0 47 & 51 4 30 ATT 265 9.5, Reeth eae 0 34 3 47 0 54 April 30s ieee fk: 0 57 0 31 0 40 May (224s OR) &. 0 32 1 86 0 69 May Sa cyurs. 1 25 9 26 3 46 May 7s eee Wea 2 43 0 47 1 18 May 9:0 tekst) 4 67 0 41 0 50 Miaiy: 14, 8 nes = 0 29 1 21 0 20 Ao tales. <0: wens 141 818 74 793 40 (57 Percentage of 79. 14.7 805 5.0 534 A. FRANKLIN SHULL Comparison of the second and third divisions of the table indi- cates that oxygenation of the water increases the proportion of the male-producers by 3.5 per cent. The first and third divisions show that Euglena increases male-production by 9.7 per cent, of which 3.5 per cent is presumably attributable to the oxygen lib- erated in photosynthesis, leaving 6.2 per cent to be caused by some other factor, probably the food itself. The food, if this assumption is correct, is a little less than twice as effective as the oxygen. Experiment 4. Euglena versus manure scum, spring water versus manure solution. The effects of the four agents named, in conjunction with one another and in opposition to one another, are here tested. On cach of the dates named in table 6 four rotifers were placed in each of four dishes. In two of the dishes was placed manure solution, the food in one of these being Eu- glena, in the other manure scum. Into the other two dishes was put spring water, the food in one of them being Euglena, in the TABLE 6 A comparison of the effects of Euglena and manure scum, arid of spring water and manure solution, wpon the proportion of male-producers in the rotifer Hydatina senta pre AND EUGLENA AND MANURE seum MANURE SCUM Raton SAPEUIE ROLuTtON abi whee DATE ig Num- |! Num- |} Num- |} Num- | Num- | Num- | Num- | Num- ber of | ber of | ber of | ber of | ber of | ber of | ber of | ber of ee ae cots) fof) ie} 48) eee) Rebruanyel OP eree eee 4 14 5 14 0 10 0 23 Bie lorirariyoel ke aie te she cette 10 13 15 0 24 3 24 Rebuy anya cde yoni eo 20 16 16 0 11 9 8 HiGloAwee ia AOS, sgosoodars oFe il 10 7 12 1 15 5 21 Webruary 28. Ase Sale ab 41 10 Ate ln 27 3 30 Miamclis alan: s cae. t35e) 4 2: cee 1 33 5 29 4 23 2 31 Mame hi 274 see cae 0 9 10 a 0 0 0 13 VCD ee he che: Aa are 5 47 29 23 0 16 12 26 Maire hy TAS ees eee 3 5 33 1 32. 0) 38 8 37 MarchilGeecsnieeeeneate. 0 34 2 14 8 16 0 25 Totals ees Wyk on He tec lA! 251 98 200 15 180 42 238 Percentage of @...... 17.70 * 32.88 7.69 15.00 MALE-PRODUCTION IN HYDATINA 535 other manure scum. The two dishes receiving manure scum as food received also a small amount of water from the Euglena stock (without Euglena), and the dish receiving Euglena as food in spring water received also a small amount of manure solution (without scum). This was done in order to equalize the initial chemical composition of the media in the four dishes, except as it was intentionally made different by manure solution and spring water. After from twenty-four to forty hours the rotifers were removed from their dishes. All offspring hatching from eggs laid in that time were reared to maturity, all four lots alike, in spring water with manure scum as food. Comparison of the third and fourth divisions of table 6 shows that manure solution reduces the proportion of male-producers by 7.31 per cent, but this reduction is more than offset by feeding with Euglena, as in the first division of the table. That is, Eu- glena increases male-production more than manure solution reduces it. If, however, as indicated by experiment 4, over one- third of the total effect of Euglena is due to the oxygen liberated in photosynthesis, then Euglena as food is less powerful in in- creasing male-production than manure solution is in decreasing it. The maximum male-production was obtained, as was to be ex- pected, by using both Euglena and spring water (second division of table 6). The minimum is obtained from manure scum and manure solution together. This minimum is approximately doubled by substituting either spring water for manure solution or Euglena for manure scum. Making both substitutions prac- tically doubles the results of a single substitution. Experiment 5. Oxygen versus creatin, Euglena versus creatin. On each of the dates in table 7, three to five rotifers were placed in each of four dishes, the number in each dish being the same on any one day. The four dishes were treated as follows: 1) One lot was immersed in a dilute solution of crude creatin, of the concentration indicated in table 7; and fed Euglena. 2) One dish was filled with a measured quantity of water which had first been saturated with an atmosphere of which 40 per cent was oxygen. Enough of a more concentrated creatin solution was added to produce the same final concentration as in the preceding dish. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 3 536 A. FRANKLIN SHULL TABLE 7 Showing the effects of oxygen, Euglena, and creatin upon the proportion of male- producers in a Nebraska line of the rotifer Hydatina senta . _. | MANURE SCUM | MANURE SCUM pres) Ceanart 8 | co oan | ok nee | ee CREATIN SOLUTION SOLUTION pare SOLUTION EN) ee Num- | Num- | Num-| Num- | Num-|} Num- |} Num- | Num- ASSIS ber of | ber of | ber of | ber of | ber of | ber of | ber of | ber of a? Qe ae 29 fos 29) ope) 2° IMiary 2116 sens yes 0.01 0 18 0 23 0 26 0 26 Mla Ge cokes ec 0.02 0 21 0 2 0 ig 0 9 Miaye2 ees an 0.01 4 14 0 29 0 25 0 21 IMlaig28 Saar aes 0.01 0 19 6 24 1 32 0 35 May e2Aeane ant 0.02 18 14 0 13 1 21 0) 19 Misia = vere re 0.02 ilzé 24 0 19 0 41 2 42 Main 30 eae 0.08 0 10 3 43 4 27 9 43 IMiatya3 lee 0.025 9 30 0 40 4 55 3 63 Jume:Asces ee. 0.02 + 17 3 21 3 47 10 61 Rotalsx-easer eee al mee, 167 12 214 13 291 24 319 Percentage of o'@ ...... 23-7 Dee 4.2 6.9 Manure scum was added as food, and the dish was set under a bell jar in an atmosphere of which 40 per cent was oxygen. 3) A third dish was filled with the dilute creatin solution, manure scum was added as food, and the dish set in air. 4) The fourth dish, as general control, was filled with untreated spring water, manure scum was added as food, and the dish was set in air. After from twenty-four to forty hours the rotifers were re- moved from the dishes. The offspring hatching from eggs laid in that period were reared to maturity in spring water, being fed with manure scum. The rotifers used in this experiment were received from Prof. D. D. Whitney, who collected them at Lincoln, Nebraska. Experiment 6. This is a duplication of experiment 5 in every respect, except that the rotifers used were part of a line received from Prof. A. L. Treadwell, who in turn obtained them from Prof. D. D. Whitney, then at Wesleyan University. They are presumably from the same New Jersey stock as was used in Dr. Whitney’s former work. All the preceding experiments in this MALE-PRODUCTION IN HYDATINA 537 paper, with the exception of experiment 5, were performed with these New Jersey rotifers. Table 8 shows the results of experiment 6. While the last three divisions of table 7 show differences of the kind expected, these differences are small. Comparisons based on the first three divisions of that table would indicate that Euglena is about 18 times as potent in increasing male- production as is oxygen. Or, deducting from the total effect of Euglena the 1.1 per cent which a comparison of the second and third divisions of the table would indicate was due to oxygen, the effect of the Euglena as food would be nearly 17 times as great as that of oxygen. A different result, however, is found in table 8. The first three divisions of the table show that oxygenation increases the proportion of male-producers 5.1 per cent, Kuglena 20.2 per cent. Deducting from the total effect of Euglena the fraction charge- able to oxygen alone (20.2 — 5.1 = 15.1), Euglena as food is nearly three times as effective as oxygen in increasing male- production. TABLE 8 Showing the effects of oxygen, Euglena, and creatin upon the proportion of male- producers in a New Jersey line of the rotifer Hydatina senta : MANURE SCUM | MANURE SCUM 4 ue =nRnEenS cal epee AND OXYGEN- |AND UNTREATED Raireeean OF SOLUTION ANSON) (NOEL) CITE ES SPRING WATER CREATIN SOLUTION SOLUTION een SOLUTION ers Num- | Num- |} Num-| Num- |} Num-} Num-| Num- |} Num- c ber of | ber of | ber of | ber of | berof | ber of } ber of | ber of Je? ke} fe) oun) 9) one) 92 Se 22 IwtaVey Ae so4d00 0.02 19 11 2, 29 2 19 0 28 UNE? ee eee 0.02 3 D, 5 44 5 33 8 Dik dias Io ssakooce « 0.02 3 19 0 21 0 6 0 itch JUN eels eeeeeeee ne 0.02 2 22 6 15 1 29 0 12 June 14 eke 0.02 4 28 6 26 1 29 4 23 Totals eee eel eel 82 19 135 9 116 12 101 Percentage of W@...... 27.4 1.8) Uo? 10.6 On wy) CO A. FRANKLIN SHULL Response to oxygen in the Nebraska line The failure of the Nebraska line in experiment 5 (table 7) to show any marked effect of oxygen is the only inharmonious result obtained in this study. It seemed possible that this line was not as sensitive to oxygen as the New Jersey line used in the other experiments. If this were the case, any other response to oxygen might prove less marked in the Nebraska line than in the New Jersey line. Such a response to oxygen in this species was found in earlier experiments. In a former paper (Shull, 715a) it was shown that races of rotifers might differ very markedly in the place of egg-laying. One race laid its eggs very largely on the bottom or sides of the dishes; another race chiefly attached to the surface film of the water. In connection with later oxygen experiments (Shull, 718), it was shown that placing the dishes in an atmosphere containing an excess of oxygen caused the eggs to be laid more largely on the bottom than when the dishes were kept in air. If the Nebraska line lacked responsiveness to oxy- gen, this lack might be evidenced by failure of oxygen to alter the place of egg-laying. This possibility gave rise to the following experiment. Experiment 7. On each of the days named in table 9 several rotifers of the Nebraska line were placed in each of two dishes of water and fed manure scum. One dish was set under a bell jar in an atmosphere of which 60 per cent was oxygen, the other placed under a bell jar in ordinary air. After a period of from sixteen to eighteen hours, the eggs laid on the bottom and at the surface film in each of the dishes were counted. As a check upon the results, a similar test was simultaneously made upon the New Jersey line. Table 9 gives the results. While the Nebraska line is here shown to be responsive to oxy- gen, the controls show that it normally laid more of its eggs at the bottom than did the New Jersey line. In view of this normal difference between the two lines, it is difficult to judge of the rela- tive effect of oxygen upon them, and the discrepant results in experiments 5 and 6 probably remain unexplained. MALE-PRODUCTION IN HYDATINA 539 TABLE 9 Showing the number of eggs laid at the surface film of water and at the bottom of the dish by two lines of rotifers, one from Nebraska, the other from New Jersey, when kept in air and in a 60 per cent oxygen atmosphere. The excess of oxygen causes more eggs to be laid at the bottom, but it 1s questionable which line is most affected this way, since their behavior is different under the same conditions NEBRASKA LINE NEW JERSEY LINE Air Oxygen Air Oxygen DATE Eggs Eggs Eggs Eggs Eggs Eegs Eges Eggs laid at | laid at | laid at | laid at | laid at | laid at | laid at | laid at surface |bottom | surface |bottom | surface |bottom | surface |bottom December! Stik .ise cia); fer 5 25 0 22 16 12 2 29 Mecemibershh eh Clete oae 28 23 1 44 28 16 3 22 Decemberilop sere ato. 4 24 0 28 1 5 1 3 Decembers20be ev. eee 6 24 0 13 32 26 (0) 2p Rata Sex eye Acer eee 43 96 1 107 77 59 6 79 Percentage at surface... 30.9 0.9 56.6 nO) DISCUSSION While this paper was in preparation an article by Whitney (17) on a similar subject, the relative effectiveness of oxygen and food as sex determiners, was published. These two papers, however, in no sense duplicate, since they approach the problem from different angles and with different criteria of judgment. A considerable part of Whitney’s paper is devoted to experi- ments to show that green organisms as food increase the pro- duction of males in other rotifers than Hydatina. It is to be regretted that many of these experiments were performed with mass cultures without controls. If these rotifers are like Hyda- tina in producing males in ‘epidemics,’ the lack of controls is most unfortunate. It is well known to every student of Hyda- tina that, especially in lines that produce only a moderate pro- portion of males, these males often appear in well-defined ‘waves,’ and it has been shown that these epidemics not infrequently have a rather regular periodicity (Shull, 715b). If, in an experi- ment without control, a line which regularly produces numerous males once a month, the experimenter, ignorant of the interval 540 A. FRANKLIN SHULL between periods of many males, should attempt to alter male- production by introducing something into the culture five or six days before the beginning of the period of males was due, he could use any one of a dozen agents and obtain the same result, namely, an increase of male-production a few days later. Notwithstanding the lack of controls, Whitney’s conclusion that the use of green organisms for food in these several species of rotifer increases male-production is probably correct; it is un- doubtedly correct in the case of Hydatina, as Whitney (14) sat- isfactorily showed in an earlier paper and as I have confirmed in my experiments. The chief question is how much of this effect of green organisms is due to nutritive differences, how much to other agents associated with the green organisms. One of Whitney’s experiments bears upon this point, but the results are not easily understood. That experiment is repre- sented by his table 5 (Whitney, 717). Information on which one could base a judgment of the significance of this experiment is not given. The cultures were apparently of some size, all of them containing green organisms. These cultures had been in existence from five to thirty days, presumably in the sunlight (either direct or diffuse), before the beginning of the experiment. Some of them were then excluded from the sunlight, but for how long a time is not stated. If the duration of the darkness was not prolonged, it is to be expected that the average oxygen content would be appreciably greater than in a culture which had never been kept in the light. The experiments described in table 4 of this paper, test No. 5, indicate that oxygen accumulated in photosynthesis is only very gradually lost. It is not clear, then, that the oxygen content of the cultures in the darkness was as much lower than that of the sunlight cultures as Whitney sup- posed. It is almost certain that, unless they were excluded from the sunlight for a considerable period, these dark cultures contained appreciably more oxygen than a culture never kept in the sunlight would contain. It is a pertinent question, therefore, whether the smaller amount of oxygen in the cultures in darkness was as effective as the larger quantity in the sunlight. Some of Whitney’s cultures MALE-PRODUCTION IN HYDATINA 541 in the darkness were aerated, others not, and he appears to have suspected that if oxygen were a male-producing agent, the more oxygen there was present the more males there should be. In the experiments described in this paper there is little difference, in its effect on male-production, between water saturated with a 40 per cent oxygen atmosphere and water saturated with a 60 per cent oxygen atmosphere, although it was shown that 60 per cent oxygen atmosphere left the oxygen content of the water very plainly greater. What difference'there was indicated that the lower concentration of oxygen was the more effective. It can- not be accepted without question, therefore, that aeration of the water, in addition to the use of green food with its incidental oxygen, should further increase the male-production. Disregarding the above objections to the experiments on which table 5 of Whitney’s paper is based, we may examine the results of those experiments. Instead of comparing only selected parts of his table 5, the whole table may be included. Since in some of the cultures more mothers were used than in others, merely adding together the offspring of parents that were treated in the same manner would give to those cultures having many mothers undue weight. A more just method of combining like cultures is to average the percentage of male-producers in all cultures; each culture is then as weighty as the others, regardless of the number of parents used. By this method of combination, Whitney’s table 5 is converted into the following: Per cent male-producets Suniiaht ewithederation,): ah waa ee Pils os eet Sona cts yi Bons tenh et 54.5 Darkness: wiwhiaenratlonn | vrs ere pastes hi. Ley Ou etie Bal Mek Ee 19.0 Darkness: wilh WOUbrderatlOl)- yew ts i tot Ales a50. Gsekah bce a rgete 40.6 It is difficult to see what these results mean, though Whitney concludes from them that oxygen is not a male-producing agent. Just why one is to infer that the quantity of food is the cause of the differences shown, is equally obscure. The argument for oxygen is at least as good as the argument for food. The remainder of Whitney’s paper is devoted mainly to the effect of oxygen upon the protozoan and bacterial food supply of the rotifers in experiments like those of Shull and Ladoff. He 542 A. FRANKLIN SHULL finds that the increased oxygen content of the water favors more rapid multiplication of these food organisms, and infers that the rotifers were therefore better fed. That there is a certain low concentration of food at which an increase of available food would mean an increase of nutrition is probable. If food were in greater concentration than this, it is not clear that further increase in the amount of food at hand would cause an increase in the amount eaten. A man who sits down alone at a table with a hundred pounds of steak and fifty pounds of potatoes eats no more than if only fifty pounds of steak and twenty-five pounds of potatoes are set before him. If the above criticism is not valid, and it is after all merely quantity of available food that determines male-production, that fact should be very simply and easily discovered by controlled experiments, involving large numbers of individuals, and the next desirable step in the investigation is perfectly clear. The criticisms of Whitney’s methods and conclusions in the foregoing paragraphs may seem to indicate a wider divergence between his conclusions and mine than really exists. Certainly the conclusions reached in this paper and in that of Shull and Ladoff differ less from Whitney’s conclusions than Whitney’s recent article would lead one to suppose. Thus when Whitney (17, p. 114) writes “According to Shull and Ladoff, one would expect many male-producing females to be produced in sunlight and Chlamydomonas with the accompanying excess of oxygen, but no male-producing females would be expected to be produced in darkness and Chlamydomonas with no excess of oxygen,” he overdraws the indictment somewhat. The quoted statement could be correct only if Shull and Ladoff had denied that food had any effect on male-preducticn, and had asserted that all of the increase of male-production obtained by Whitney was due to oxygen instead of food. No such idea was even suggested. Indeed, there are frequent passages in the paper of Shull and Ladoff that indicate the reverse. Thus on page 188 it is stated “our results may be interpreted as being largely in support of Whitney’s contention.” Again, page 156, “‘part of the increased male-preduction following the use of Chlamydomonas as food MALE-PRODUCTION IN HYDATINA 543 in Whitney’s cultures, was due to the oxygen liberated by the green flagellates as a by-product of photosynthesis.’”’ Likewise, page 157, “We suspect * * * Whitney’s conclusion that the food conditions influence male-production is correct.” All that was insisted upon by the authors of these passages was that (page 157) “‘all the factors obviously associated with Chlamy- domonas in the cultures should bé separately tested before any residue of influence is assigned to nutrition.”’ Two factors associated with Chlamydomonas in Whitney’s earlier experiments were 1) an initial difference in the chemical content of the water and 2) dissolved oxygen produced during photosynthesis. In this paper the former factor has been elimi- nated and the second has been measured. After deducting the measured effect of the oxygen from the total effect of the green organism used for food, there is a residual effect which it seems fair to attribute to nutrition. According to the results of ex- periments described in the foregoing pages, this residual effect due to nutrition is at least several times as great as the effect of the accompanying oxygen. SUMMARY Repetition of the experiments with an excess of dissolved oxygen, in which the excess was obtained by saturating the water with an atmosphere of which 60 per cent was oxygen, showed that oxygen is a male-producing agent. The experi- ments of Shull and Ladoff had demonstrated that a smaller excess of oxygen (obtained from a 40 per cent oxygen atmosphere) increased male-production, but left in doubt the effectiveness of a 60 per cent atmosphere. Rather direct comparison of the two concentrations of oxygen, in this paper, shows that there is little difference between them. The lower concentration may, per- haps, be a little the more effective in inducing male-production. The amount of oxygen dissolved in the water in these experi- ments has been measured by the Winkler method. The cultures which were first saturated with an atmosphere of which 40 per cent was oxygen and were then enclosed in a vessel in such an 044 A. FRANKLIN SHULL atmosphere, contained on the average about 60 per cent more dissolved oxygen than did the untreated controls. Cultures in which Euglena was used for food were also com- pared, relative to their oxygen content, with cultures in which manure scum was used as food. The Euglena cultures under the conditions of the experiments, contained on the average about 62 per cent more oxygen than the manure scum cultures. Pre- sumably, therefore, when Euglena increases male-production in Hydatina, as much of that increase is due to oxygen as is directly produced by saturation with a 40 per cent oxygen atmosphere. In experiments with Euglena as food, after deducting the in- erease in male-production presumably due to oxygen liberated, it was found that Euglena was two or three times as effective as the oxygen, and in one case many times as effective. The experi- ment in which Euglena appeared to be many times as effective as its liberated oxygen was performed upon a different line from that of the other experiments. This line was tested to discover whether it was responsive to oxygen in another way (laying eggs at surface film or bottom of dish). While it was plainly responsive to oxygen, there is some doubt whether it was as responsive as the other line used. Euglena as a male-producing agent was compared with manure solution as a male-repressing agent. The repressing effect of manure solution was a little more than offset by the Euglena (including the effect of the oxygen liberated by the Euglena). BIBLIOGRAPHY Suuuu, A. F. 1915 a Inheritance in Hydatina senta. II. Characters of the females and their parthenogenetic eggs. Jour. Exp. Zodl., vol. 18, no. 1, January, pp. 145-186. 1915 b Periodicity of male-production in Hydatina senta. Biol. Bull., vol. 28, no. 4, April, pp. 187-197. 1918 Effect of environment upon inherited characters in Hydatina senta. Biol. Bull., vol. 34, no. 6, June, pp. 335-850. Suutt, A. F. anp Laporr, Sonza 1916 Factors affecting male-production in Hydatina. Jour. Exp. Zodl., vol. 21, no. 1, July 5, pp. 127-161. Wuitney, D. D. 1914 The influence of food in controlling sex in Hydatina senta. Jour. Exp. Zodél., vol. 17, no: 4, November, pp. 545-558. 1917 The relative influence of fgod and oxygen in controlling sex in rotifers. Jour. Exp. Zodél., vol. 24, no. 1, October. pp. 101-145. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, AUGUST Ds CONTRIBUTION TO THE STUDY OF EPITHELIAL MOVEMENT. THE CORNEAL EPITHELIUM OF THE FROG IN TISSUE CULTURE SHINICHI MATSUMOTO (KYOTO) Osborn Zoological Laboratory, Yale University NINE FIGURES CONTENTS I. Behavior of epithelium cultivated im vitro............)../:0.02....0: 546 1. Description of types of epithelial movement in plasma............. 547 he) INMoiermSaNH Than iGy alae ono hUINARMD Semone nye eo eee cb acce se ote 548 Dan Vo vementinontISSUCR asses orc asi Meteora tee tee eee eee 549 GLOtherhtypestoL mowvemrentic. 2). 5 leant Gack ieeewn tener - 549 2. Movement of epithelium cultivated in serum..................... 550 Sev elociinmot cellimovenrenitie sae. -04 <4 a) sR eee 551 II. On the reaction of epithelial cells to certain solid supports............ 555 (Pe No mennent.on! iat SUA CES Hears. rsg sy exe LE Oe OOD Fie MONA, (Alesis) havo herorel MaKe lia < SOA Rn nA, eS Sate 555 be Onudead cornea) oye earn: lea. Liles ea Ue ily aioe boa do eee pe 556 2 viovement oni berlike: supports... .. ..s..,. 1 sata soho sangre Ls: 557 SPE SPICeLRwie bras SUpPPOLGe an das scicks Rta eee ee eee eee pees 557 byiSulkatiber as} supporter. 2cs. 6 Gs SUE opel ea 8 yaoi eee 558 cu lassmvioolias Supp Onusiaecm. ack isle Ae a cataracts TOO earls 558 BIMASbestOs: Der AS SUPMORtgs a6. oc 5 o.< cert e avo hg ames he WOOD Cee VIO VELEN EON PIL MMvebCwyeite eo cc tte homo eke ees Mates 559 CREB ITUA eee OE os J) BS. SOR ORO LL EET, Fe EES 560 The mechanism of the movement of epithelium, embryonic as well as adult, has been discussed by various observers, but still further investigation is needed by the use of the tissue-culture method, where direct observation is possible. The experimental work reported in the present paper is con- cerned with a study of the movement of corneal epithelium, and especially with its behavior with reference to certain mechanical supports. 545 546 SHINICHI MATSUMOTO Several papers have already appeared which deal with the movement of epithelium of various kinds in vitro (Ruth, ’11, Champy, 714, Loeb, ’02, Osowski, ’14, and others). The observa- tion of frog skin in vitro has been described by Holmes (714) and Uhlenhuth (714) recently in detail. Previous to this, Harrison (10) Carrel and Burrows (’11), Lambert and Hanes (13), and others mentioned it briefly in their papers on tissue culture. “Among the works which deal with the culture of the corneal epithelium, that of Oppel (12), who made use of the cornea of certain warm-blooded animals in his investigations, demands special attention. Harde (16) made brief mention of an active lateral spreading of the corneal epithelium in the culture of vac- cinia with corneal tissue. However, the materials which were used by Oppel and Harde are evidently not very suitable for direct observation. For this purpose cornea of more simple structure is desirable. It must be added that a number of investi- gations on the problem of the wound healing of the cornea have been made, such as those of Peters (’85), Salzer (11), Lowenstein (13), and others, which must, of course, be taken into considera- tion. I. BEHAVIOR OF EPITHELIUM CULTIVATED IN VITRO The cornea of the adult frog (especially R. pipiens) was used. After thoroughly washing its whole surface with sterilized Ringer’s solution by means of a pipette, the entire cornea was cut out with a razor and put into Ringer’s solution (or serum), after which it was divided into small pieces with very sharp scissors so that the fragments showed sharp edges. Pieces cut radially were preferred. The cultures were all made by the hanging-drop method (Harrison, ’10), the technique of which need not be detailed here. The piece of cornea was taken from the Ringer’s solution and dropped on the surface of the cover-glass; excess solution was re- moved and a drop of plasma (or serum) run over the fragment; autoplasma was used in most of the cultures. The cover-glass was then inverted upon a thin glass ring and sealed on with vaselin (Harrison, 714). EPITHELIAL MOVEMENT 547 Though all precautions were taken to keep the cultures free from bacteria, it was sometimes necessary to throw away a whole series as a result of infection. This was due obviously to the difficulty of perfect sterilization of the tissue. In the aggregate, more than 1800 cultures were made, and in nearly all of the experiments two kinds of culture medium (plasma and serum) were used. The following descriptions are the results of the study of about 1500 cultures which were free from faulty technique. Fig. 1. Diagram showing various types of movement of corneal epithelium cultivated in plasma. g, cover-glass; c, tissue of cornea cultivated; ep, corneal epithelium; ed, endothelial surface; m, culture medium; ¢, cut end of the piece of cornea. Cell movement into the plasma (1/7), along the endothelial surface (EZ), on the epithelial surface (2’), along the cover-glass (G@), and along the lower surface of plasma (S). 1. Description of the types of epithelial movement in plasma (fig. 1) The cornea of the frog was very suitable for this purpose. For the observance of the intimate cell structure and a closer study of the mechanism of cell movement, high powers could be used. It was, of course, necessary to supplement the study of the living tissue by fixed and stained preparations of whole cul- tures and by serial sections. The observation of the living cultures was limited to the first week. 548 SHINICHI MATSUMOTO a. Movement in the medium. Whena culture of cornea, freshly prepared in vitro, was examined under the microscope, the edges of the piece were seen to be sharply defined. One or two hours later the epithelium on the cut ends gave the impression of be- coming a little translucent and swollen, a narrow clear rim ap- pearing along the edges. Here and there, around the edges of the fragment, isolated round epithelial cells were to be seen, singly or in groups, having been detached by mechanical injury during the operation. There was a short latent period before any movement was noticed, the first cellular activity appearing between the third and tenth hours. Accordingly, examination after twelve to twenty- four hours showed an active outgrowth of epithelial cells present- ing amoeboid processes. As arule, the corneal epithelium showed characteristic sheet-like extension during the period of active movement; the advancing edge was always furnished with an amoeboid border of hyaline ectoplasma, as has been described by Harrison (710), Carrel and Burrows (711), Lambert and Hanes (13) and Holmes (714). Some of the epithelium exhibited marked motility, recalling the movements of amoeba, so that an exact ecamera-lucida drawing could not be made; pseudopodia were formed, and through their activity the cells changed shape or moved from place to place. Some of the cells which showed filiform processes were over 0.2 mm. in length. The strong tendency of the cells to lateral spreading brought about the formation of a continuous membrane, extending nearly horizontally, usually slanting upward a little toward the end. The spreading membrane also changed its direction of move- ment, showing contraction under various circumstances. Growth in strands was also observed. The growing epithelium, usually two cells in thickness, sometimes covered an area a little larger than the original corneal piece, whereby the cells became flattened and the intercellular spaces grew wider. There was apparently little growth after the third (sometimes the fourth) day. The tissue itself then gave the impression of being less compact and more translucent than formerly. The epithelial cells showed a tendency to round off. The rupture ¥ = - Ee gs TL eae EOS eS es A Ss RSS Sree, SE Lene ee Py peed: Vea wn Onc 8 Oo ass are ~ Oo asi meen MENAE eC EOE = 5 ¢ Tey’ et PO eed wa (tea egy SOO Dee MS2e S . FHB2S Messed. | ine = We eS OS? wy SIO YD at gai a a “Sys Ce a “xaces* . noo ae = ig P ¢ ne { 2 ~ = asin ~ J Ae” 5 ai Me 23 r% vs - . eis Cc “Se a3. Z - al! Site =o Sew — a [Ss SS. as TE = ee = E se SA Se EPITHELIAL MOVEMENT 549 of strands or sheets of cells into isolated masses was of frequent occurrence. The number of round cells increased rapidly from day to day, and generally fat droplets grew in number and size with the age of the culture, until the cells were often literally packed with them. No marked increase of mitotic figures was seen. The activity of cell movement into the plasma depends on the consistency of the latter. Around the explanted tissue, lique- faction and retraction of the plasma were often observed which caused changes of the arrangement of cells. Fig. 2. Vertical section of corneal tissue cultivated in plasma. Experiment XXXVIII, 5, showing the epithelial movement (£) along the endothelial surface (ed). Age of the culture, four days. Drawn from one of the serial sections. t, cut ends; ep, epithelial, ed, endothelial surface; c, part of connective tissue of cornea. XX 98. b. Movement on tissue. The next important type of the epi- thelial movement is that on the corneal tissue. By virtue of this, the growing epithelium spreads over the edges of the frag- ment and along the endothelial surface. Figure 2 shows an ex- ample of this very clearly. A movement of this type occurred in some cases on all edges of a fragment and in others only on a part of it, combined with other types. It was also observed that part of the cut end of the epithelium might remain almost inac- tive, while in the other part marked activity occurred. c. Other types of movement. Whenever the growing epithelial cells came into contact with the cover-glass, they moved actively over it. The character of cell movement, however, was essen- 590 SHINICHI MATSUMOTO tially the same as on the endothelial surface. The advancing edge of the growth was composed of a very delicate sheet of protoplasm, showing amoeboid movement, with numerous branching hyaline pseudopodia. When the delicate membrane spread out to an extreme degree, it resulted in the breaking up of the sheet into isolated masses. Ina small percentage of the cultures the spreading of cells over the outer epithelial surface of the cornea (fig. 1, Z’) as well as along the lower surface of the culture medium (fig. 1, 8) was noted. The elements of the connective tissue showed neither active growth nor movement. The study of the lymphocytes found in the corneal tissue needs special investigation. 2. Movement of the epithelium cultivated in serum Generally, the cornea cultivated in serum showed some details of cell movement, different from those noted in plasma cultures. No amoeboid migration of the cells into the medium took place. During the first and second hours, the cells of the edges became clear and round, sometimes even swollen in appearance; they soon started to move. The most striking and constant phenomenon noted in the use of serum was the spreading out of epithelium over the tissue itself, especially on the endothelial surface. The epithelial rim extended usually parallel with the cut edges, showing amoeboid movement on the advancing border. The rapidity of spreading out of cells was sometimes quite remarkable. In a small per- centage of the cultures, movement of cells was also observed on the original outer epithelial surface of the cornea. If single cells or a part of the rim were in contact with the cover- glass, they clung to it and spread over the surface of the glass with marked activity. The boundaries of individual cells were often hard to distinguish under the microscope. When such prepara- tions are properly impregnated with silver nitrate (Lewis and Lewis, 712) the intercellular spaces can be clearly demonstrated (fig. 3). When the delicate membrane had extended to the EPITHELIAL MOVEMENT Holl utmost degree, the cells along the periphery became loose and isolated. The appended tables give some examples of the frequency of various types of epithelial movement in plasma (table 1) and in serum (table 2), respectively. Fig.3 Experiment XLIX, 2. Silver impregnation, demonstrating the bound- aries of individual cells. Drawn from a part of actively moving border closely attached to the lower surface of cover-glass. Cultivated in serum; three days’ growth. G, advancing border. X 450. 3. Velocity of epithelial movement The velocity of cell movement in vitro exhibits a considerable variation under different circumstances, and, as stated before, THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 26, NO. 3 D2 SHINICHI MATSUMOTO TABLE 1 Cultures in plasma CELL MOVEMENT SERIES NO. Cree On l On On Ppt ans ees MAMONS | medium fondothelial] epithelial | cover glaso| Sufs°® 50 a 18 14 14 1 1 50 bh 14 5 13 1 46 10 3 10 4 45 6 5 5 2 2 1 1 43 20 13 11 2 Ih @2) 52 6 5 4 1 3 57 10 9 3 2 4 61 10 8 6 2 2 1 62 2 2 2 2 2 63 34 24 33 102 33 18 31 106 15 15 14 104 15 1 15 107 11 6 11 110 6 8 111 9 4 9 212 18 11 17 Movall:..24| 239 149 206 10 20 3 2 combinations of several types of movement are often to be seen in a single preparation. During the period of activity, a change in the type or direction of movement was often observed, and such a change in one part was apt to modify the movement of neighboring cells. Single cells or small groups of cells, as a rule, moved more freely than those in the spreading membrane. The consistency of the plas- ma, the liquefaction and the contraction of the fibrin influence both the rapidity and the type of movement into the plasma. The temperature has both direct and indirect influences on the movement. There is also individual variation which cannot be attributed to any of the above factors. For these reasons, the velocity of the epithelial movement varies from hour to hour, and it is very difficult to analyze the factors influencing it. Usually, the most vigorous actiyity occurred during the first twenty-four hours. For instance, the epithelium on a piece of HPITHELIAL TABLE 2 Cultures in serum MOVEMENT Or On (JU) NUMBER CELL MOVEMENT NO SERIES NO. | OF PREPA- | 6 : 2 eCLONS bint. endotHelal epithelial t ORE, surface ae cass surface surface | = film 56 28 28 94 28 28 5 1 58 7 6 1 60 8 7 1 139 7 uf 138 7 6 1 140 5 5 135 8 s 165 14 14 166 12 12 167 15 15 208 23 23 1 219 1s 18 221 20 20 232 7 7 233 9 8) 234 9 8) otal ncee a-225 212 1 5 1 3 cornea cut radially into eight parts, spread out into the plasma within twenty-four hours in an extension which was even larger than their original area, though not in the same degree of rapidity at all eut ends. Similarly, in suecessful preparations, it was seen that the whole endothelial surface was covered within twenty-four hours by the epithelium spreading over the entire circumference of the piece. ment ceased. Some measurements are given in table 3. When the moving borders met each other, move- TABLE 3 MOVEMENT ON MOVEMENT INTO PREPA-| ENDOTHELIAL SURFACE PLASMA RATION aS ae 24 hrs. | 48 hrs.| 5hrs. | 24 hrs. | 48 hrs. Experiment 210, plasma culture, Lf) O14 | 0F4 | 0.50206) 0.37 085 ate ae: 2 ORIN KOR 1205 ORI | OFS ail ples Be OI Ok Oa Ob Nh Ib 4 | 0.14 | 0.65) 0.7 tonal OCMC ORS S| Mab LO | eee Oss POs |p 0 Oy 1 OME OES: 409 Zoe OPI ORGIE TOES SO | ORG anOzS Average movement in milli- IME TENS SMM chron siete cle OF12) 0.63 -OF835| 0208750) 7 EG 20 hrs. | 40 hrs.| 60 hrs. Experiment 211A, plasma cul- 1 Ono ze 2h ez ture, at 20°C. 2 0.8 | 4.0) |) 5.4 3 ie) 2 2.8 Average movement in milli- NMC COTS ere aie ane ck enc N peers nga aat< || Bade 20 hrs. | 40 hrs. 20 hrs. b — Experiment 211B, plasma cul- Ose ale ture, at 20°C. NOs |) One Bi le eOle eke A SOS68N | ,05% Dy ORAS NOES 0.2 6. 7202628086 Average movement in milli- MICHELS Me sie aee wn See ee 0.65 | 0.78 20 hrs. Experiment 208, serum culture, 1 ORS at 20°C:* 2) 1054 3°, ORS 4 |0.8 5 | 0.6 GeO Average movement in milli- MVCLOLS = cus. Rocp dee te ee eee 0.6 *Pieces which measured: 1.2 x 0.6 mm.; 1.8 x 0.5 mm., 2.0 x 0.6 mm., 2.4 x 1.1 mm., and 0.8 x 0.8 mm., were entirely covered up with moving epithelium in seventeen hours. 554 EPITHELIAL MOVEMENT DIO II. ON THE REACTION OF EPITHELIAL CELLS TO CERTAIN SOLID SUPPORTS 1. Movement on flat surfaces a. On glass and celloidin. We shall first consider movement on the glass cover-shp.' There is difficulty by means of usual culture method in bringing the epithelial rim into contact with the cover-glass from the beginning. Attempts to do this by reducing the plasma failed, as the epithelium did not show activ- ity unless the medium was used in sufficient quantities. In the later experiments, the piece of cornea (cut off tangentially to eyeball) was placed on the cover-slip, mainly with the inner surface down, and subjected to slight pressure by means of thin silver wires or glasses and kept in that position for a certain period after mounting with serum, so that the cut ends came into contact with the glass. When the silver wires were properly placed and controlled under the microscope, it was possible in almost every instance to bring certain parts of the moving epi- thelium into contact with the cover. Of course, where even a minimal space existed between the tissue and cover-glass surface, the epithelium crept on the underlying tissue. Cover-slips coated with celloidin were also used. Thus the movement of the epithelium on celloidin and glass surfaces, respectively, could be compared. A comparison could also be made in the same preparation, by employing a cover-slip, only one half of which was coated with celloidin. Such preparations must be handled carefully, otherwise a change of cell form readily occurs. No characteristic difference whatever in the mode of movement on the two different surfaces could be noticed. If the epithelium grew out vigorously, closely attached to the cover-slip, each cell became flattened into an exceedingly thin layer. In some instances it was seen that the advancing border became very 1 Tt should be stated here that the cover-glasses used in these experiments were throughly cleaned and washed in the vapor of distilled water, as is usually done to remove any trace of alkali, in order to exclude any chemical influence from that source. (Ostwald-Ruther, Physiko-chemische Untersuchungen. ) 556 SHINICHI MATSUMOTO irregular. This was most marked when active cells on the border, showing unusual protoplasmic activities, tore themselves loose from their connection. Similar conditions, however, were often to be noticed in the actively mobile epithelial layer growing out into firmly clotted plasma. In the epithelial movement on the endothelial surface, such a condition but rarely occurred; as a rule, the cells moved more uniformly, showing a smooth advanc- ing border. In other instances small pieces of cover-glass or celloidin mem- brane were placed on the endothelial surface and pressed against it, so that a certain portion or the whole of this surface, which is a favorable support for moving cells, was covered. In these cases the epithelium was also seen to move over the celloidin or glass, if they were suitably mounted and the cells were not injured. Where the cell movement on such artificial supports failed, examination proved that the placing of the membrane was not suitable; thus, by control of pressure, it was possible to direct the moving cells to these artificial surfaces, though this was not always an easy matter. At any rate, it is an established fact that the epithelium is able to creep on objects of such nature that chemical influences are excluded. b. On dead cornea. If in a part of an explanted piece of cornea an epithelial defect existed, it was seen that the epithelial cells were able to cover the spot, spreading out from all edges, prac- tically with the same rapidity as on the endothelial surface. The same was true of a spot which had been killed by touching it with a heated needle. The cells injured by the latter procedure became round and detached, and the growing epithelium crept beneath them along the surface of the killed tissue. Preparations were made of large strips of corneal tissue (about 2x4mm.), across the center of which a sharply defined epithelial defect was produced, after which one half of the remaining corneal tissue was killed by means of a heated needle, making the tissue surface look opaque and wrinkled. In such preparations it was seen that the epithelial movement over the wound occurred uni- formly, covering both the live and the dead tissue with the same rapidity. EPITHELIAL MOVEMENT 55t Analogous experiments were made with pieces of cornea in which an epithelial defect existed at one extreme end and in which the tissue at the other end had been killed; similarly, the cell movement on a burnt wound was compared with a control prep- aration, in which the underlying tissue was left intact with simply an epithelial defect. The use of tissue, vitally stained by neutral red? or nile blue, facilitated the observation of epithelial movement on such wound surfaces. Analogous observations were made by using the surface of the cartilaginous layer of the sclerotic coat of the frog’s eye. This tissue could readily be isolated from the other layers after boiling, and it was then thoroughly washed before using. It was trans- lucent, showing characteristic convexity, which admirably fitted on the inner surface of the similarly curved cornea; even without any pressing, the two tissues would le so closely together in the culture medium (serum) that the epithelium from the cut ends of cornea crept over the cartilage. It proved helpful, how- ever, if the pieces were slightly pressed together. Out of twenty-six experiments, movement of the epithelium over the cartilaginous plate took place in twenty-one; in three cases no movement was observed owing to injury of cells. An example of this is illustrated in figures. 4 and 5. 2. Movement of fiber-like supports Next, the response of the epithelium to the various fiber-like supports were tested, such as spider web, silk fiber, glass wool, and asbestos. a. Spider web as support. Experiments were made similar to those of Harrison (14), using spider webs. In each preparation a sufficient quantity of serum was used as culture medium. Figure 6 represents one of the preparations of the series; many cultures were found where the cells clung to the fibers. Out of thirty cultures twenty-four were positive, four doubtful, two infected. 2S. Matsumoto. Demonstration of epithelial movement by the use of vital staining. This paper will appear in the next number of this journal. 558 SHINICHI MATSUMOTO b. Silk fiber as support. Other experiments were made, using silk fibers. In these a number of fibers of raw silk were stretched on the glass ring and cover-glass, imitating the experiment with the spider webs. Out of twenty-nine cases there were fourteen that gave positive results, one being infected. ch Aa d = = a3; > 2 eo OP) 50 ee et ren Ben A es t ~ 8 an ~ Fe oe 9299 Fig. 5 Experiment 233; 4. Same preparation as shown in fig. 4. Portion of vertical section, drawn from one of the serial sections. ep, epithelial surface of cornea; c, connective tissue; ch, cartilaginous plate, placed on the endothelial surface of cornea; ¢, cut end of cornea. Note the epithelial movement (#) on the sclerotic cartilage (ch). Forty hours’ growth. 98. Thirty cultures were made in this group, with positive results in twenty-two. Figure 8 shows an example of this series. e. Movement on pith, etc. Further studies were made with pith. Thin pieces were used, which were kept in distilled water which was changed every three days for a period of over three months. Experiments showed that the epithelium was able to cling to the cell walls of the thin piece of pith. Figure 9 shows clearly 560 SHINICHI MATSUMOTO the growing out of epithelial cells on the support. Outof twenty- seven experiments nineteen gave positive results. On the shell membrane of hen’s egg, used as support, after thoroughly washing and boiling, a similar condition was observed. J Ew 5 iS Se 8 ) - / é en Hla fii aS BO ys! 1 ate JE LT Fi aif if Sa S % SS ‘ ~ 4 ay : LD \ J | L \ 6 7 Fig. 6 Experiment 238, 11. Epithelial movement in serum on spider web. In the neighborhood of the explanted tissue (c) a good many isolated cells, round in shape, were noticed, which are omitted in this figure; they showed no active movement. ¢, edge of the piece; above, some cells move out in sheets on the cover-glass. Compare Jour. Exp. Zodél., 17, 521, figs. 4 to 7,12. X 98. Fig. 7 Experiment 49, Epithelial movement on glass wool (gw); culti- ile t, edge of tissue. vated four days in serum. x 98. Ill. RESUME This paper deals with the movement of corneal epithelium of the adult frog in vitro. Frog cornea is very suitable for this purpose, as it is so transparent and thin that it permits of direct observation of the cell movements. The corneal epithelium cultivated in plasma shows various types of movement, according to the nature of the substratum EPITHELIAL MOVEMENT d61 (fig. 1). The movement is of an amoeboid character. As a rule the cells have a strong tendency to cling to their own kind and thus extend in sheets, although under certain conditions active movement of isolated cells is also to be seen. In the majority of cultures movement into the medium or along the endothelial surface (or both) takes place according to the consistency of the culture medium. ‘The fact, that the epi- ‘al a Fig. 8 experiment 238, 1. Epithelial movement on fibers of asbestos; culti- vated forty hours in serum. At the extreme left cells are to be seen moving along cover-slip. Note the adaptation of single cells to asbestos fibers. > 450. thelium moves along the epithelial or endothelial surface is very important from various points of view; such a movement may easily be overlooked in the culture of non-transparent tissue, such as skin. In the preparations in which serum is used, no migration of the epithelium into the medium takes place. There is mainly a movement of cells over the tissue, especially on the endothelial surface. 562 SHINICHI MATSUMOTO Naturally, the question arises, whether the epithelial movement on the tissue, which is so frequently to be seen in the fluid me- dium, is chemotactic or thigmotactic in nature, It might even be that the type of movement is due both to mechanical and to chemical influences acting simultaneously. However, the epi- thelium is able to move with practically the same velocity both on a substratum where the covering epithelium has been simply Fig. 9. Experiment 239. Corneal piece cultivated in serum with pith; three days after explanation. Note the epithelium clinging to the piece of pith (p) moving out of the cut end (t) of cornea (c). X 98. scraped off and on one where the underlying tissue has been killed by heating. The movement in the latter cases cannot, therefore, be dependent upon chemotactic influences from the liv- ing tissue. The same is true of the movement taking place on the surface of the cartilaginous plate of the sclerotic coat, pre- viously killed by boiling. Furthermore, the epithelium is able to move on the surface of glass, on a celloidin film, and also ‘on such fiber-like supports EPITHELIAL MOVEMENT 563 as spider web, glass wool, asbestos, ete., when it is brought into contact with them. The cell movement on the glass and celloidin film is very vigorous, sometimes more rapid than on the endothelial surface. In the movement on such supports chemotactic influences are to be considered as excluded. It has been repeatedly observed by various writers that a suit- able support for the growing cells is an important requisite; Harrison (’14) demonstrated lately the importance of such fac- tors for the movement of embryonic cells very clearly. The facts above described confirm this view, that stereotropism plays an important réle in cell movement. The behavior of corneal epithelium in vitro serves to throw some light on the mechanism of epithelial growth in vivo. The experiments show clearly that the epithelium is able to extend from the cut end quite rapidly in sheets into the medium (plasma), or on the tissue (plasma and serum), and .can cover a large area, whereas mitotic cell divisions are not necessary at all. That Oppel (12) as well as Osowski (’14) did not observe amoeboid activity of the epithelium was perhaps due to the diffi- culty of direct observation. Special, careful observation on the movement of epithelium of warm-blooded animals, however, is necessary. In conclusion, I wish to express my indebtedness to Prof. R. G. Harrison for his direction and valuable advice. 564 SHINICHI MATSUMOTO LITERATURE CITED CARREL AND Burrows 1911 Cultivation of tissues in vitro and its technique. J. Exp. Med., 138, 387. 1911a An addition to the technique of the cultivation of tissues. J. Exp. Med., 14, 244. Cuampy 1914 Note de biologie cytologique. Quelques résultats de la méthode des cultures des tissus. Arch Zool. Exp., 54, 307. Harpe 1916 Some observations on the virus of vaccinia. Ann. d’l inst. Past., 30, 299. Harrison 1910 The outgrowth of the nerve fiber as a mode of protoplasmic movement. Jour. Exp. Zoél., 9, 787. 1914 The reaction of embryonic cells to solid structures. Jour. Exp. Loos, V7, p21. Houmes 1914 Behavior of ectoderm of amphibians when cultivated outside the body. Jour. Exp. Zodl., 17, 281. LAMBERT AND Hanes 1913 Beobachtungen an Gewebskulturen in vitro. Virch. Arch., 211, 89. Lewis AND Lewis 1912 Membrane formations from tissues transplanted into artificial media. Anat. Rec., 6, 195. Lors 1898 Uber Regeneration des Epithels. Arch. Entw.-Mech., 6, 297. 1902 Uber das Wachstum des Epithels. Arch. Entw.-Mech. 13. 1907 Uber einige Probleme der experimentellen Tumorforschung. Z. Krebsf:, 5. 1912 Growth of tissue in culture media and its significance for the analysis of growth phenomena. Anat. Rec., 6, 109. LOWENSTEIN 1913 Experimentelle Untersuchung iiber Regenerationsyorginge in der Kaninchencornea. Arch. Ophth., 85. Oppet 1912 Causal-morphologische Zellenstudien. V. Mitteilung. Die active Epithelbewegung. Arch. Entw.-Mech., 35, 371. 1913. Demonstration von Epithelbewegung von Froschlarven. Anat. Anz., 45, 173. Osowsxr 1914 Uber active Zellenbewegung in Explantat von Wirbeltierem- bryonen. Arch. Entw.-Mech., 38, 547. Perers 1885 Uber die Regeneration des Epithels der Cornea. Inaug.-Diss. Bonn. Ruta 1911 Cicatrization of wounds in vitro. J. Exp. Med., 13, 422. Sauzer 1911 Uber die Regeneration der Kaninchen-Hornhaut. Arch. Augenh., 69, 303, etc. UnLENHUTH 1914 Cultivation of the skin epithelium of the adult frog. J. Exp. Med., 20, 614. INDEX LBINO rat. Studies on inbreeding. I. The effects of inbreeding on the growth and variability inthe body weight of the 1 Albino rat. Studies on inbreedirg. II. The effects of inbreeding on the fertility and on the constitutional vigor of the......... 335 Albino rats. On several effects of feeding small quantities of Sudan IlI to young.... 101 Albino series of the rat. Ruby-eyed dilute gray, a third allelomorph in the.. 4 55 Albissima Vejdovsky. Reactions of the pro- DOSCisOlellanadniaeor new hee Taner ies 83 Alcohci on treated guinea-pigs and their de- scendants. Further studies on the modi- fication of the germ-cells in mammals: Thevetlectonerswe cease eee ae 119 Atuer, W. C., AND Srein, E. R., Jr. reactions and metabolism in May-fly nymphs. 1. Reversals of phototaxis and the resistance to potassium cyanide. II. Reversals of phototaxis and earbon di- Oxide;productionien..en ste ere eye 493 Allelomorph in the Albino series of the rat. Ruby-eyed dilute gray,athird...... 55 Antibodies. Studies on cy tolysins. 4b. Some prenatalietects:oL lensta cess usele eee 65 Arbacia and Asterias. Studies on the physio- logical significance of certain precipitates from the egg secretions of................. 459 Asterias. Studies on the physiological signifi- cance of certain precipitates from the egg secretions of Arbacia and................. 459 Axial gradient in the regeneration of Tubularia supported by facts? Isthetheory of ..... 265 ANUS, Manito Garcia. Is the thecry of axial gradient in the regeneration of Tubularia supported by facts?........... 265 Body weight of the albino rat. Studies on inbreeding. I. The effects ef inbreeding onthe growth and variability in the..... Bottom material ingested by holothurians (Stichopus). Theamountof.............. 379 Light re- ARBON dioxide production. ‘G in May-fly actions and metabolism nymphs. I. Reversals of phototaxis and resistance 10 potassium cyanide. lI. Reversals of phototaxis and........ ..... 423 Chemicals on reversion in orientation to light in the colonial form, Spondylomorum quaternarium. Effects of............... 503 Corneal epithelium of the frog in tissue cul- ture. Contribution to the study of epi- Ghenalanovenment. Whee .e. 4 sae eae 545 Crozier, W. J. The amount of bottom ma- terial ingested by holothurians (Sti- Chopus)) 98 sac. 20 sPoaeagn Monk varaohuriesaer 379 Culture. Contributicn to the study of epi- thelial movement. The corneal epithe- lium of the frop in’ tissue-.:..)...-- eeu. 545 Cyanide. Il. Reversals of BC and ear- bon dioxide production. Light reactions and metabolism in May-fly nymphs. I. Reversals of phototaxis and the resistance POMPOLASSIMINI oar ere Rear ere eer 423 Cytolysins. I. Some prenatal effects of lens antibodies. Studies on.................-. 65 IOXIDE production. Light reactions and metabolism in May-fly nymphs. I.Re- velsals of phototaxis and the resistance to potassium cyanide. II. Reversals of phototaxis and earbon........ 423 Ductless glands on the development of the flesh flies. The effects of the............. 255 FFECTIVENESS of food, oxygen, and other substances in causing or prevent- ing male-production in Hydatina. Rel- ALLV Oh ehh nein ee a Egg secretions of Arbacia and ‘Asterias. Studies on the physiological significance of certain precipitatestromiphe sy anon eee 459 Epithelial movement. The corneal epithe- lium of the frog in tissue culture. Contri- bution to the study Ole sa eee 545 Epithelium of the frog in tissue culture. Con- tribution to the study of epithelial] move- mente: Pie connea lars mae a een ieee me 545 EEDING small quantities of Sudan III to young albinorats. Onseveral effects OL ee Tae NCCE ee eee 101 Fertility and on the constitutional vigor of the albino rat. Studies on inbreeding. 11. The effects of inbreeding on the.......... 335 Flesh flies. The effects of the ductless glands on the development of the.. 255 Flies. The effects of the ductless glands « on the development of the flesh.............. 255 Fly nymphs. I. Reversals of phototaxis and the resistance 1o potassium cyanide. IT. Reversals of phototaxis and earbon dioxide production. Light reactions and metab- olismring May=s3>. one ore 423 Food, oxygen, and other substances in caus- ing or preventing male-production in Hydatina. Relative effectiveness cf..... 521 Form, Spondylomorum quaternarium. Ef- fects of chemicals on reversion in orienta- tion to light in the colonial............... 503 Frog in tissue culture. Contribution to the study of epithelial movement. The cor- nealepithelium ofthe..................... 545 Fundulus and mackerel. A study of paternal heredity in heterogenie hybrids. Hybrids Betweellieten se tear ttt etok eee ee vcs 391 ERM-CELLS in mammals: The effect of alcohol on treated guinea-pigs and their descendants. Furtherstudies on the MOcdincationOmuherscas st ntee eee ee ee 119 Glands on the development of the flesh flies. The effects of theductless................. 255 Gradient in the regeneration of Tubularia sup- ported by facts? Is the theory of axial. 265 Growth and variability in the body w eight ‘of the albino rat. Studies on inbreeding. I. The effects of inbreeding on the.. 1 Guinea-pigs and their descendants. Further studies on the modification of the eerai- cells in mammals: The effect of alcohol ODSELER Lede ye eran eee 119 Guyer, M. F., anp SmirH, E. A. Studies on eytolysins. I. Some prenatal effects of lenstantibOdresiie se acne cenit eon 65 565 066 HAs S. On several effects of feeding small quantities of Sudan II] to young BLD INO WAS. shes ake ern eS Oe 101 Heredity in heterogenie hybrids. Hybrids between Fundulus and mackerel. A Suc yrOmpavernalo or Reenter eee 391 Heterogenic hybrids. Hybrids between Fun- dulus and mackerel. A study of paternal Heredity As leo ce ele eee 391 Holothurians (Stichopus). The amount of bottom materialingested by............... 379 Horned toad Phrynosgma. ‘The physiology of the melanophores of the............... 275 Hybrids between Fundulus and mackerel. A study of paternal heredity in heterogenic hybrids aeapemi seen a ye simile ced | 391 Hydatina. Relative effectiveness of food, oxy- gen, and other substances in causing or preventing male-production in........... 521 NBREEDING. I. The effects of in- breeding on the growth and variability in the body weight of the albino rat. SEUGIES On sek. ae td ea tas 1 Inbreeding. II. The effects of inbreeding on the fertility and on the constitutional vigor ofthealbinorat. Studieson.. 335 Ingested by holothurians (Stichopus). ‘The amount of bottom material ............... 379 7 EPNER, Wiiu11am A., anp Ricu, Ar- NOLD. Reactions cf the proboscis of Planaria albissima Vejdovsky. 83 Kine. HELEN DEAN. Studies on inbreeding I. The effects of inbreeding on the growth and variability in the body weight | of the album Ouraitgene. ese eects eens ee ee ee ee 1 Kine, Heten Dean. Studies on inbreeding. IL. The effects of inbreeding on the fertil- ity and on the constitutional vigor of the SEU UO Mais se eh. epee et ar ee eS 8 335 Kane, Hreten DEAN, AND WHITING, P. W., Ruby-eyed dilute gray, a third allelo- morph inthe Albinoseries of therat....... 55 IXuNKEL, B. W. The effects of the ductless elanas on the development of the flesh ENS antibodies. Studies on cytolysins. _4 I. Some prenatal effects of................ 65 Light in the colonial form, Spondylomorum quaternarium. Effects of chemicals on re- version in orientation 1o.............. 503 Light reactions and metabolism in May-fly nymphs. I. Reversals of phototaxis and the resistance to potassium cyanide. II. Reversals of phototaxis and carbon dioxide DIOGUCHLON Gs site sets | Resi, none Sh came eer. 423 Me A study of paternal hered- ity in heterogenic hybrids. Hybrids between Fundulus.and.............2..... 391 Male-production in Hydatina. Relative effee- tiveness of food, oxygen, and other sub- stances in causing or preventing.......... 521 Mammals: The effect of alcohol on treated guinea-pigs and their descendants. Fur- ther studies on the modification of the germs-cellenns gw. I ada Be.. os ica 119 Mast, 8.0. Effects of chemicals on reversion in orientation to light in the eslonial form, Spondylomorum quaternaritum. .. 508 Materialingested by holcthurians (Stichopus). The amount of bottom. .....-:...:2..-+-- 379 Matsumoto, SHinicut. Contribution to the study of epithelial movement. The cor- neal epithelium of the frog in tissue CULT ie be te aero sees fore eee ed ear Deo INDEX May-fly nymphs. I. Reversals of phototaxis and the resistance to potassium cyanide. Il. Reversals of pkototaxis and carbon dioxide production. Light reactions and metabolismpetny 0s See oy ai eens 423 Melanophores of the horned toad Phryno- soma. Thephysiology of the............ 275 Metabolism. A demonstration of the origin of two pairs of female identical twins from tworowoaormhiehistoragess eye ance one 227 Metabolism in May-fly nymphs. T. Rever- sals of phototaxis and the resistance to po- tassium cyanide. II. Reversals of photo- taxis and carbon dioxide production. sihtreschionsiands.10 ene ee eee see 423 Modification of the germ-ceils in mammals: The effect of aleohol on treated guinea- pigs and their descendants. Further pastidie som pinie’s . 0h. asec. /see 2 ee Ewa ese 119 Movement. The corneal epithelium of the frog in tissue culture. Contribution to the study of epithelial. ..............-.... 545 EWMAN, H. H. Hybrids between Fun- dulus and mackerel. A study of pater- nal heredity in heterogenic hybrids.. 2/898 Nymphs. I. Reversals of phototaxis and the resistance to potassium cyanide. II. Re- versals of phototaxis and carbon dioxide production. Light reactions and metab- oligmiiny Misty flyer easier chee 423 RIENTATION to light in the colonial form, Spondylomorum quaternarim. Effects of chemicals on reversionin........ 503 Ova of high storage metabolism. A demon- stration of the origin of two pairs of female identical twins from'two.........:........ 227 Oxygen and other substances in causing or preventing male-production in Hydatina. Relative effectiveness of food,............. 521 AIRS of female identical twins from two ova of high storage metabolism. A de- monstration of the origin of two iy Dee PAPANICOLAOU, GEORGE N., AND SrocKARD, CHARLES R. Further studies on the mod- ification of the germ-cells in mammals: The effect of alcohol on treated guinea-pigs and. their descendamtse... 2. * ee SH AACA, A oe Te Ot : ota tate : Se eae eo wee © es tet! ; te eee P