=a : Oo ae ee oe vot : oot e sty ~~ > et eh pigt ty : al 3; bye ord Lx te pees an yt we he - ‘ Srase sees tes 74 . eS Pe he ue + bog b aot a Se Se 3 1 we at Wh cos i Rechte ety we ae ot tt be te a ftp { 5 st pe Ox f ee bad ‘ ttt ss oe DS ee : pak Set vt oe dhs Se or bP ee to ~ et Fao wes at oe Ae SS at . + ~ z . : Paes 50" . : i pty xe y ae HS o oe os 4 4 tied ed © SP, 62 5 we 4 ee B - ‘ ed ae : ‘ 3 . ; j ; Phe ‘ - eee ee ee oe pr Sa : - : b : { rs Fins ae We ae | ; 3 ae THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY Wiuiiam E. Castur Jacques LoEB Harvard University The Rockefeller Institute Epmunp B. WILSON Columbia University Epwin G. CoNKLIN Princeton University Tuomas H. MorGan CHARLES B. DavENPoRT : wire Columbia University Carnegie Institution GrorGE H. PARKER HERBERT 8. JENNINGS Par vard! Umivereity Johns Hopkins University RAYMOND PEARL FRANK R. LILLIE Maine Agricultural University of Chicago Experiment Station and Ross G. HARRISON, Yale University Managing Editor VOLUME 28 1919 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. bine OI * ala he i ' i a a Prey i y nay) r Med, (et { 4h ‘i i i, i‘f Wy a wenn POA i qi oy hy AN Vie A eh y : ty ji 4 Uden i n j Ay ( \ i rua i 7 aby, | li Hay " ee! Va AU i a La v) i wit > id Ay a " ih i r cr i | ‘ VAN oe Ain vi y If iv M by a) Vay By = CONTENTS No. 1. APRIL J. Percy Baumpercer. A nutritional study of insects with special refer- ence to micro-organisms and their substrata. Eighteen figures......... 1 H. D. GoopaLe AND Grace MacMuturen. The bearing of ratios on theories of the inheritance of winter egg production ...................02000.2+. 88 No. 2. MAY Wi.ram B. Krrxuam. The fate of homozygous yellow mice. Two figures. 125 Cart R. Moors. On the physiological properties of the gonads as con- trollers of somatic and psychical characteristics. I. The rat. Five IPOS Sao oH 07 Gd CER ROEREOIE So 0 60.0 S GU SETAC nes PAIS Men Ghd Moen eR. Bhd Mel bs ii Donatp Watton Davis. A sexual multiplication and regeneration in Sagartia luciae Verrill. Ten plates (forty-two figures)................. 161 Cavin B. Bripcres. The genetics of purple eye color in Drosophila........ 265 Epwarp C, Day. The physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to sensory responses. Five figures. 307 Nos, JULY Carvin B. Bripaes. Specific modifiers of eosin eye color in saat melanogaster. Two diagrams. . e: . dal C. H. Danrortu. Evidence that § sesne ores: are egal he pclecnicnt on ane basis of their genetic potentialities. . ae . 385 P.W. Wuitina. Genetic studies on the Mediterranean fleax nite Tphestia kiihniella Zeller. One figure and two plates. . sites Ate ake WHEELER P. Davey. Prolongation of life of Giielean eoreeauen appar- ently due to small doses of x-rays. Four figures...................0200% 447 Cart R. Moore. On the physiological properties of the gonads as con- trollers of somatic and psychical characteristics. II. Growth of gonadectomized male and female rats. One figure.................2-005 459 Davin D. Wuitney. The ineffectiveness of oxygen as a factor in causing male mroduction in Hydatimasembary, «i oascssad. saw eas heen have + Be blige 469 i ras Vike heer oO et 10 Laie pn iin Ati en wo Mh cs ee | a f ie OP ‘ib naga sory EA Cy i arama iat a nee koi Ra ms igi for ee Une “tee a i amie Pata Ng Kaede Cah page i 1 vad iy, ov ii Ry: d neil i | ana w fy | Wihgyn is } ] | ‘ . Plas! Pv i r ee a # Lh iia ‘ done : hae Pec tam | die Pete aaa DAT m “itwor int ode thai it a ci AERO TOG ek a ee | ) Dein ath Heirejungyat ‘ogi F el act) (4, -eagelenteiegtigy Wed | . oT i, Lal al Ry» soba Rhett (heeded o) Te Cycling ea nO ve a \ v4 J an i Ga Hot es \ Vay: Py tA ib Md ‘didenondlls ‘it tiie Ki yin ian end ee Bee i * wi: os) eR ¥ iy aye "VY RR 3 ecnianyiatte H me: a se, 8) ia et Ney A Fy Bae oe EET jth 2 i ‘ I) Aen Me ANE aC ye eu! TER pak bi4hM) iy . Evening ifn spill mlenizane te NUP aeed Mon nied wee. hea Chit Fo { Ria) 1. Wiss te Ny . shan’ dik is nH ai vie ia £7) sedieayuentinn i Nese caane en OP) RR 1 a a bait ith et Dilys bh a ae un iia aly ih f he eee fiphsheatat Ne i Ried UP aNerdaee SM yy rs cm aL alae spas o ri hata Hh hd Nut Na are Lh wi ga ne Beatin ‘i eerie if) Ray one bt " ty | | Bam vith | hae i AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 31 A NUTRITIONAL STUDY OF INSECTS, WITH SPECIAL REFERENCE TO MICROORGANISMS AND THEIR SUBSTRATA! J. PERCY BAUMBERGER Bussey Institution for Research in Applied Biology, Harvard University CONTENTS Introduction. .c2 hoc. os ana ok eee Re ORs epareaekr Oe ae 2 EUSPETIIN Misra tee sk Pas. 5 Se RRR a STD Senale nid Ng gas Smo Se 3 1. Food of an insect (Drosophila) living in fermenting fruit............. 3 A. Method and initial observations: a) Solid media for Drosophila; b) Preliminary observations on the food of Drosophila; c) Habits of adults and larvae; d)Ecology of cultures; e) Media for geneti- cal work; f) Are living yeasts present in the egg or pupa? g) Sterilization of pupae; h) Test of sterility..........:......... 3 B. Food of Drosophila: a) Growth of sterile larvae on sterile fruit; b) Is fruit the food for larvae or merely the substratum for yeast cells? c) Are products of fermentation essential food re- quirements of larvae? d) Is yeast a complete food for larvae? e) Can larvae complete their growth on any vegetable food other than yeast? f) Is yeast a more adequate food than fruit because of its higher rotein content? g) Conclusions.......... 11 °C. Discussion: a) Effect of food on larval, pupal, and adult life; b) Sugar requirement of adults and larvae; c) Natural habitat; d) Function of yeast in the ecology of Drosophila; e) Literature on” the faudmoierosopinilare (8 ee et ool Vet ee ain scene aoe ns 26 2. Experiments with a sarcophagous insect (Desmometopa).............. 43 3. Experiments with a coprophagous insect (Musca domestica).......... 43 4. Experiments with a mycetophagous insect (Sciara) and a mite (Tyro- glyphus) living in decaying wood: a) Experiments with Sciara; b) Experiments with Tyroglyphus; c) Association of wood-eating AaAeCHS with: [Unie ee «cine bts dE ee 47 Bxieniiet mycetophagy amoneinsects’... 4.) vey ish hielecaoanee oh bamt tee. 58 Microorganisms as liquefiers of the substratum........+...............+.... 64 Gootssmiraetive to. inseets.y.. eee rd eee cise Pe a.a/aph nie nie cialis ks 67 Macrooreanisms as food of other animals eee). i. ele nee ewes 69 Microorganisms as internal symbionts of insects........................00. G2 OSLNG SCT i OS Meee SE. 3. cial Ae a i tt: SOP ULCER UMN Rd ae... . . oatecnteys bl was oe ge tal Se 0)! alin Rot otelegs taehenare 75 * Contribution from the Entomological Laboratory of the Bussey Institute, . Harvard University. 1 pi J. PERCY BAUMBERGER INTRODUCTION Throughout the whole organic world the essential food element most difficult to acquire is nitrogen, as all nitrogen must ulti- mately come from the atmosphere and the power of combining with this gas is limited to a few microorganisms. Upon the nitri- fying bacteria, then, all higher plants and animals are dependent for their nitrogen which is handed from one organism to another, linking all together into one great interdependency which has Ox/DIZING BACTERIA hawous PACTERIA ia a ELL a \) == eh a NITROBACTER- — Se ~~ ; a — aaa RLS/DUES, YREA, LETC. ——————e Fig. 1 The nitrogen cycle (from Bayliss). The accessory lines and circles in — -— -—are my additions based on evidence in this paper. been called the nitrogen cycle. I insert a diagram from Bayliss ? which clearly illustrates this cycle. The accessory circles and the lines that connect them are additions based on my experiments. The search of the insect for nitrogen is very complicated and has been, at times, obscure. Indeed, little definite information is at hand concerning the food requirements in general of these organisms, as the material consumed is often in large part merely the substratum for a small amount of assimilable food. This has led to many misunderstandings as to the synthetic power of A NUTRITIONAL STUDY OF INSECTS 3 insects. Since they are largely phytophagous, insects are amply supplied with carbohydrates, but have difficulty in obtaining sufficient protein. ‘The abundance of the former permits great activity, while the dearth of the latter limits the growth of the insect. This has led to a lengthening of the life-cycle in those species which must ingest large quantities of substrate in order to get enough nourishment to complete their growth. However, many insects that feed in decaying of fermenting vegetable matter of low protein content have an unusually short period of growth. The experiments and considerations which follow throw light on the protein supply of such insects and account for their rapid growth. These investigations were made at the Bussey Institution for Research in Applied Biology, Harvard University, under Prof. W. M. Wheeler. Valuable advice and assistance were received from Profs. C. T. Brues, W. J. V. Osterhout, I. W. Bailey, and Dr. R. W. Glaser. I am especially indebted to Doctor Wheeler for helpful suggestions and encouragement. EXPERIMENTS 1. Food of an insect (Drosophila) living in fermenting fruit A. Method and initial observations. a. Solid media for Droso- phila. While rearing Drosophila it was found necessary to de- termine the exact date of oviposition. As this is impossible in the ordinary culture tube of fermenting banana, a solid trans- parent medium was devised by myself and Dr. R. W. Glaser UOT 2) This medium is made as follows: Mash six ripe bananas in 500 cc. of water, allow to infuse on ice overnight, strain through cheese-cloth, and add 13 grams powdered agar-agar to each 100 cc. of the filtrate. Heat in double boiler till agar is dissolved, filter hot through absorbent cotton into test-tubes. Plug tubes, sterilize in autoclave, and allow to cool in inclined position so as to form solid slants of the medium. This medium is quite transparent, affords 15 to 20 sq. cm. area for oviposition and 6 to 10 ec. of substratum for the larvae. The 4 J. PERCY BAUMBERGER eggs, which are readily deposited by the female, are prominent objects on the agar. Bacterial and fungous growths occur over the surface, but I no- ticed that unless these become too-luxuriant before the larvae hatch, they are destroyed by the insect. The agar method has the advantage of permitting observation of the date of egg deposition and hatch and the details of larval habits. It also furnishes a method of making nutritional studies of various synthetic media. b. Preliminary observations on the food of Drosophila. In May, 1916, while rearing Drosophila melanogaster on banana agar, I noticed that molds and bacteria often completely covered the sur- face of the medium and killed the larvae. This was confined to cultures which had only ten or twenty instead of the usual fifty or a hundred larvae. The larvae congregated at the points where fungus was most abundant and caused the plants to dis- appear, apparently by feeding upon them.? An examination of the flora showed that Saccharomycetes were invariably present and often occured in pure cultures.’ This observation suggested an internal symbiosis between Dro- sophila and yeast. I found, nevertheless, that by washing the surface of the pupae with alcohol, the insect could be freed from all microdrganisms. The larvae of such sterile insects were not able to mature on banana agar nor could they mature on a syn- thetic medium of salts and sugars with ammonium tartrate as the source of nitrogen, as had been maintained by Loeb (715%), but were able to develop on either medium in the presence of yeast cells. c. Habits of adults and latvae. The Drosophila were intro- duced as pupae, usually three being placed on the side of the test- tube. The adults emerge after five to eight days, the time depend- ing on the temperature, and readily feed on the banana medium, 2 This interpretation was first suggested to me by Mrs. J. Jackson. 3 In 373 transfers of pupae, all descendants of adult Drosophila, taken from a stock bottle of fermenting banana, all tubes were infected with yeast cells carried on the bodies of the insects. 4 Loeb has since. corrected this view (716). Loeb and Northrop (’16 b). A NUTRITIONAL STUDY OF INSECTS 5 on which they leave little depressed spots where they have regur- gitated and sucked up the dissolved substance. If the medium has not dried enough to have taken on a hard, leathery crust, the. females oviposit after twenty-four hours and continue to do so for some days. The eggs are thrust into the agar so that the upper end with its two projecting floating structures is Just level with the agar; in this position they are prominent objects under the bi- nocular. After a period of one or more days, the minute larvae leave the eggs and move about over the surface of the medium. They are at this time usually 1.2 mm. in length. By the second day they have increased in size to 1.8 to 2 mm. in length, and be- gin to work in a vertical position, with the anterior end down, the full length of the body in the jelly, and the posterior end with its two projecting spiracles either in contact with the air or with a bubble of air which has been enclosed in a thin film of the medium and remains attached to the larva, thus enabling the latter to work the food material to a greater depth than its body length would permit. The head end of the larvae is merely a small pointed segment which served as a collar through which the pseudo-maxillary apparatus works. In shape the latter may be roughly compared to a plow with theshares prolonged posteriorly in- to two handles. Attached at the anterior end of this four-pointed median structure, is a pair of deflected falcate processes, sharp at the point and on the concave side, that work up and down con- stantly with a simultaneous backward and forward movement of the whole apparatus. The movements of these oral organs were observed ina drop of agar on a depression slide, and it was found that their constant movement continued without any appreciable rest periods. Occasionally the movement would stop without apparent reason for about two minutes, but there was no regu- larity in these periods of cessation. The larva might work for fifteen minutes without stopping or might stop several times at intervals of two or three minutes. Apparently the recovery from fatigue takes place in the interval between the movements. Progression of the larvaseems to be due to a series of protrusions of the anterior end with an accompanying circular contraction, the animal being held in place by the circles of spines on each seg- 6 J. PERCY BAUMBERGER ment, while the posterior end is drawn up. In more fibrous ma- terial, the mouthparts probably aid the larva in moving about. When fully grown, it leaves the medium to pupate on the side of the test-tube or the surface of the medium itself. d. Ecology of cultures. Drosophila is very extensively used by geneticists in breeding experiments. The insect is reared in small glass bottles or milk jars, plugged with cotton and containing fer- menting banana covered with absorbent paper. Quite often these ‘cultures go bad,’ i.e., smell strongly of acetic acid or be- come putrid or covered with mold, so that the imsects are de- stroyed and the breeding experiment terminated. The method commonly employed in making the culture media is to boil skinned bananas, to cool the mass and to add two cakes (24 grams) of Fleishmann’s bread yeast (bottom yeast) per dozen bananas. ‘This is allowed to ferment and is used as a stock sup- ply from which to prepare clean culture bottles. In this manner the medium is kept fairly sweet, probably due to the great develop- ment of the yeast, with an accompanying production of alcohol which retards® the development of molds and bacteria. If pupae are taken from a bottle that has gone ‘bad’ and placed on banana agar, a number of different bacteria or molds may develop around then, prominent among which are a mucor, Rhizopus nigricans Ehrenberg, the bread mold, Aspergillus, the green herbarium mold, Penicillium glaucum, the blue mold, and the acetic acid bacillus. If pupae are taken from a good culture tube with yeast alone or yeast and the acetic acid 5 In this connection Lafar (’10, II, 2, pp. 288-240) writes: ‘‘From the stand- point of the oecological theory of fermentation, the alcohol produced by yeast should be regarded as a weapon capable of hindering the appearance of other fun- goid competitors in saccharine nutrient media. However, when accumulated in the medium during the progress of fermentation, it also restricts the further de- velopment and action of its producer. In this case, as with yeast poisons in gen- eral, the first result is the cessaticn of cell reproduction, a larger quantity of alco- hol being necessary to arrest fermentation and a still further quantity to kill the cells.’”’ Reproduction of yeast cells ceases at a 6 per cent and fermentation at a 5 to 24 per cent concentration of alcohol. It should be also remembered that most bacterial or fungus cultures have a tendency to become pure, probably owing to the production of some definite antagonistic substance, or to better adaptation to the medium by the successful form (Hiss and Zinsser, 710). — A NUTRITIONAL STUDY OF INSECTS ( bacillus, the Drosophila larvae grow rapidly, the fungous growth soon diminishes and is visible at only a few points on the surface. If the flora contained molds, the whole surface of the medium is soon covered and the Drosophila eggs are killed, or more often hatched and the young larvae die. If the mold does not completely cover the surface, many larvae survive, and upon increasing in size, are able to destroy the mass of mold hyphae and form a fairly clean surface. The larvae are able to do this only when they are in large numbers and have reached a size of 3.5 to 4.5 mm. before being covered by the molds. It would seem, therefore, that the destructive action of the molds is mechanical rather than toxic. It was also observed that molds seldom gain a foothold on media in which large numbers of larvae are feeding. This observation explains why ‘strong’ cultures of Drosophila (as usually reared on fermenting banana) remain ‘sweet’ and seldom go bad. Banana-agar culture tubes in which the Drosophila pu- pae have been placed on the glass, rather than on the medium it- self, often remain sterile till the adults emerge. The latter spread the spores over the surface of the agar at the same time that they deposit their eggs. Thus the molds and bacteria have little time to grow before the larvae are at work. The development of molds and bacteria is not apparent in the presence of large numbers of larvae and a strong culture of yeast in the proper nutrient medium. e. Media for genetical work. In selecting the best medium in which to rear Drosophila the most important considerations are abundant food for the yeast cells and a moist jellylike consistency of substratum to which the larvae are adapted. Transparency and solidity of media will add to the convenience of the investigator. I have obtained the best results by using Saccharomyces ellip- soideus, in the stock bottle of banana, as the fragrant odors of fermentation produced by this yeast stimulate oviposition by the fly. The two following media have proved most satisfactory: 1. Fermented banana agar. Ferment one dozen mashed _ ba- nanas for 48 hours, strain through cheese-cloth, add agar, steril- ize and slant. 2. Pasteur’s culture fluid agar. 8 J. PERCY BAUMBERGER 10 grams ashes of yeast 10 grams ammonium tartrate 100 grams rock candy 1000 grams water Add agar, sterilize in Arnold sterilizer, slant. Into sterile tubes of these media the introduced adults or pu- pae carry living yeast cells which are distributed through the me- dium by the activity of the larvae. f. Are living yeasts present in the egg or pupa? In the follow- ing experiments undertaken to show that microéganisms are not transmitted through the egg of Drosophila, the first precaution was to free the insect from external microérganisms. Usually eggs are used for this purpose, but the small size of Drosophila eggs makes this a difficult procedure. As it is well known that the lining of the digestive tract of larvae is thrown off upon pu- pation, pupae were selected for sterilization. The pupae from a culture strong in yeasts were submerged in 85 per cent alcohol for ten minutes and then introduced asep- tically into sterile slant culture tubes of agar-agar and fermented banana filtrate. If no yeast developed around the pupae which were placed on the food, the tube remained sterile after the emergence of adults, oviposition, and hatching of larvae. The sterility of the tube was later tested by introducing a few loop- fuls of the medium into a sterile tube of similar food. It had previously been determined that yeast developed readily on fer- mented banana agar. 2. Larvae which had been feeding on media contaitane living yeast cells were submerged and washed in 85 per cent alcohol and . then introduced into sterile culture tubes. In all cases yeast de- veloped on the new media. Cultures from the digestive tracts of the larvae gave similar results. Apparently, many cells es- cape digestion in the stomach, as is the case with seeds or insect eggs in birds. 3. Eggs were sterilized by soaking i in 85 per cent alcohol for ten minutes. The larvae which hatched were always sterile. From the foregoing experiments we may conclude that living microérganisms are not present in the eggs or pupae of Drosophila. A NUTRITIONAL STUDY OF INSECTS 9 However, a loose symbiosis exists between yeast and the insect. As mentioned above, surface fungous growths disappear in the presence of larvae which often seemed to be more numerous at this point. From these observations I inferred that the larvae fed upon the microérganisms present. g. Sterilization of pupae. ‘The sterilization was accomplished by the use of ethyl alcohol. Asa precaution the operator’s hands were washed in alcohol, and a lighted burner, clean forceps and platinum loop as well as sterile culture tubes were ready ona TABLE 1 ae pesos Pea ALCOHOL TREATMENT 2 = LARVAL PERIOD CONTAMINATION 2 i) A 8 2 50% 10 seconds 2 | None Yeast cells A 9 2 50% 5 seconds 2 | None Yeast cells A 10 3 50% 20 seconds 3 | 12 days p! | Yeast cells All 3 50% 2 seconds 3 | 11 days p | Yeast cells A 12 3 50% 2 seconds 3 | 14 days p | Medium brown, yeast coccus, rod AN 118} 2 50% 2 seconds 1 None . Yeast cells (?) Al7 3 85% 2 minutes 5) 26p Yeast cells A 18 8 85% 5 minutes 6 28d! A 19 5 85% 6 minutes 5 | None A 20 6 85% 7 minutes 6 26p Yeast cells A 24 5 85% 10 minutes 4 25d A 25 7 85% 10 minutes 6 44d 1 d indicates larval death p indicates pupation table also washed with alcohol. Pupae were taken from a tube having a strong growth of yeast, but uncontaminated by molds'® and placed in a sterile watch-glass. Alcohol was then poured in till the pupae were submerged. All floating pupae and all larvae were removed. ‘The results of this treatment for different periods of time are shown in table 1. The pupae are able -to withstand a treatment of 25 minutes in 85 per cent alcohol if applied when they are about two days old. 6 The frequency with which pure yeast growths occur in Drosophila cultures has already been mentioned on page 4. 10 J. PERCY BAUMBERGER Treatments of five minutes seldom kill the pupae, and in 90 per cent of the cases render them sterile. The sterilizing effect was not entirely understood till pupae were used which came from a Drosophila stock bottle of fermenting banana contaminated by molds and bacteria. These pupae when washed with 85 per cent alcohol saturated with HgCl were sterilized in less than 50 per cent of the cases as shown in table 2. This indicates that the sterilization involves two stages, 1) de- struction of molds and bacteria by feeding of the larvae and a good strong yeast growth and, 2) killing of yeast by alcohol. The toxicity of alcohol for yeast cells is shown to be high by the following experiments: Three grams of yeast were separated in 25 cc. of sterile water and two drops of this fluid were added to each of ten watch-glasses filled with 85 per cent alcohol and to ten sterile banana-agar tubes. TABLE 2 No. NO. No. ADULTS NO. TUBES NO. PUPAE CULTURES PUPAE SHE os fey EMERGED | CONTAMINATED| CONTAMINATED 15 150 | 85 per cent alcohol 15 8 83 = sat. with HgCl. After 1, 5, 10, 15, and 20 minutes, respectively, two sterile banana agar tubes were inoculated with two drops of yeast from the watch-glasses of aleohol. The tubes were kept under observation for 21 days. The results in table 3 show that a five-minute ex- posure to 85 per cent alcohol is fatal to yeast cells.” h. Test of sterility. The sterility of a culture tube could usu- ally be judged by the fact that no growth occurred, 1) around the pupae which were placed on the medium, 2) at spots where adults regurgitated on the medium, 3) at adult fecal spots, 4) upon ovi- position, 5) upon emergence of larvae, 6) upon pupation of larvae. If no growth occurred in the first case, i.e., around the pupae, the medium showed no sign of contamination throughout the life- cycle. Bacterial. growths visible around the pupae might dis- appear during the life of the larvae, but usually reappeared when 7 Paine (’11) showed that yeast cells are highly permeable to alcohol which readily and permanently plasmolyzes them. A NUTRITIONAL STUDY OF INSECTS et the larvae pupated. Apart from visible growths, the ster- ility of the tube was tested by introducing loops full of the medium on which larvae were working or had pupated into a sterile tube of one of the following media; potato agar, banana agar, Pasteur’s agar, nutrient gelatine, nutrient bouillon and yeast agar. One or all of these media were used to test the environment of the larvae for the presence of microérganisms (fig. 18). Usually crushed adults, pupae, or larvae were also introduced into the test culture tube. The method of inoculation was by stab or streak; in the former case semianaérobes could develop. Ba- nana agar was used most often, as it more nearly resembles the natural environment of the fly and its associated organisms and also can support vigorous growths of a large flora.’ TABLE 3 = . om TIME OF EXPOSURE TO 85 oe? NUMBER OF INOCULATIONS Sin GRA RTE NUMBER OF CONTAMINATIONS min. 10 0 10 2 1 2 2 5) 0 2 10 0 2 15 0 2 20 0 Smears of the media were examined after staining in the usual manner with eosin or Loeffler’s methylene blue. This examina- tion was made with a 1.6 mm. Zeiss objective. Fresh smears were examined, before staining, with dark and light field illumination. . B. Food of Drosophila. a. Growth of sterile larvae on ster- ile fruit. From the foregoing experiments it is clear that yeast is always present in the habitat of Drosophila larvae and is usually imported into sterile media on the body of the adult or pupa. Pupae and eggs do not contain living yeasts and any yeasts on the external body surface can be killed by alcohol. 8 For example, Cocobacillus acridiorum, Bacillus prodigiosus, B. coli, B. aceti, Streptococcus dispar, Saccharomycetes cerevisiae, 8. ellipsoideus, 8. anomalus, Penicilium glaucum, Rhizopus nigricans, Aspergillus, Fusarium, Gliocladium, etc. iby J. PERCY BAUMBERGER If sterile pupae are placed on a sterile medium of banana agar and protected from contamination, the adults emerge and ovi- posit, but the larvae that hatch develop very slowly and finally die before pupating. The great difference in rate of growth be- tween sterile and non-sterile larvae on the same food is shown in figure 2. In cultures A 10, 11, 12, amd 17 living yeast cells were present and the larvae grew at a normal rate, reaching the full length of 8 mm. in eleven to twenty-six days when pupation took place. In cultures? A 18 and 25, on the other hand, the sterile larvae reached a size of only 3 mm. after twenty-eight to forty- Z / lang th 2 Sp millimekers Pe Duper Age im Days. D= Death of larva , nny 7e zie te 30 az 4 Fig. 2. Larval growth on banana agar. A 10, 11, 12, 17, growth in cultures infected with living yeasts; A 18, 24, 25, slow growthof larvae in sterile cultures; - A 24, infected with living yeast on twenty-sixth day, causing an increase in growth. four days, when they died. In culture A 24 the sterile larvae reached a length of 2 mm. in twenty-six days when the medium 9 The size of the larvae on different media was determined by placing the tubes and a millimeter scale on the stage of a binocular microscope and measuring the length of five to ten of the larger specimens while ‘crawling’ at full length. The larger specimens were selected for measurement because, although female adults were allowed to oviposit for only one day, the eggs showed considerable variabil- ity from one to three days in their date of hatching, depending on the readiness with which the female oviposited on the medium. The cultures were kept in a steam-heated room in which the maximum tem- perature for the entire period of experimentation varied between 96° and 71°F. and the minimum between 73° and 56°F. As compared experiments were run parallel in time, the error due to temperature differences should not be great. It should be kept in mind that each point on a curve of growth is the average of the whole culture of larvae, i.e., usually twenty or more individuals, thus a single curve has considerable weight. A NUTRITIONAL STUDY OF INSECTS 13 was inoculated with living yeast. This caused a rapid increase in size and ended in pupation six days later. The acceleration which takes place on infecting a sterile me- dium with living yeasts indicates that the alcoholic treatment in sterilizing the pupae does not cause the decrease in the rate of growth of the sterile larvae. Other cases of acceleration which occurred due to accidental contamination of a sterile medium quite often bore out this conclusion. Therefore, it is certain that sterile larvae grow more slowly than non-sterile larvae on sterile food, and that the rate of growth can be increased by infecting the medium with the living yeast. b. Is fruit the food for Drosophila larvae or merely the sub- stratum for yeast cells? As sterile larvae grow so slowly and do not pupate in sterile fruit, but develop normally if it is infected with living yeasts, the question arises as to the true position of the fruit in the ecology the insect. By using a medium con- taining the inorganic salts and the sugars and ammonium tar- trate necessary for yeast growth, the starch, oils, fats, proteins, and other substances of the fruit were eliminated from the experiment. The composition of the medium was as follows: PNGAn=AO AP. . < . «ce semen 4.0 grams KGEIR Oss ae eee Auer ee 0.165 grams Grape-su gar. «5... cseeee | 16.5 grams IMigS @Oktes tes rau eee 0.165 grams @ane-sueary.: iv cones 16.5 grams oO MCR: 6 APART 5 A Rare 200 ce. Ammonium tartrate...... 3.3 grams Sterile larvae lived only five days on this sterile medium and showed no increase in size; but if the medium was infected with living yeasts, the larvae grew at a normal rate, reaching their maximum size in ten days, and pupated normally. The adults which emerged from these pupae were sexually fertile and of large size. Thus, in the presence of living yeast, Drosophila larvae grow normally in a synthetic nutrient medium for yeast with ammonium tartrate as the only supply of nitrogen. There- fore the simplest nutrient medium for yeast if infected with liv- ing yeasts is equavalent to fermenting fruit in the ecology of Dro- sophila larvae. 14 J. PERCY BAUMBERGER The nutrient medium for yeast in itself is not an adequate sub- stitute for sterile fruit, as Drosophila larvae live longer on the latter, e.g., Medium Increase in Length Longevity Stertlevbamiam ate Gar iy. )0: «fe cis vee eee 1.8 mm. 26 to 44 days Sterile yeast nutrient medium............ ope 0 mm. 5 days Therefore sterile fruit has greater food value for sterile larvae then the simplest ‘nutrient medium for yeast.’ Fruit is mainly the nutrient substratum for yeast cells, but has some food value - - for Drosophila larvae. c. Are products of fermentation essential food requirements of Drosophila larvae? In the preceding experiments living yeast cells had an opportunity to develop and form products of fermentation in the media. As these products may have food value for the larvae, the essential difference between a septic and a sterile food might be the absence of these substances. If this were the case, the larvae would be dependent on yeast not as a food, but as a chemical agent. By boiling yeast before adding it to yeast nutrient agar, the formation of fermentation by-products was prevented. Fleish- mann’s bread yeast was used for this purpose and 6 grams were added to every 100 cc. of yeast nutrient agar. On this medium sterile larvae grew at a normal rate, reaching their full size in ten days and pupating normally. This proves that Drosophila lar- vae grow normally on dead yeast in the absence of any by-products of fermentation. d. Is yeast a complete food for Drosophila larvae? In the media used thus far various substances besides yeast were present. To eliminate these and determine whether or not yeast alone is a complete food for Drosophila larvae, media were made up of Fleishmann’s compressed bread yeast, water, and agar-agar.'° It was found that sterile larvae on a medium of 6 grams yeast per 100 cc. water grew as rapidly as non-sterile larvae and many times faster than sterile larvae on banana. In figure 3 cultures W 3, 4, and 5 show that larvae on dead yeast grew to maximum 10 Sterile larvae live a maximum of five days on sterile 13 per cent agar and water medium, showing no increase in size. A NUTRITIONAL STUDY OF INSECTS 15 size in four or five days, while larvae on sterile banana (A) did not reach their maximum size in twenty-eight days. The minimum requirement of yeast was*found by the use of media consisting of 1, 2, 3, 4, 6, 9, 12, and 24 grams of yeast, re- spectively, separated in 100 cc. of water and thickened with pow- dered agar-agar. On sterile 1 per cent yeast the larvae grew very slowly for twenty days, dying at a length of 4 mm. without pu- pating. On sterile 2 per cent yeast larvae pupated when 5.5 mm. in length, reaching this size on the 10th day. On 3, 4, 6, 9, and 12 per cent yeast media the results were much alike, the larvae Larva! /ength ee 4 te rHllimeters. P 7 | P S| |e ics [uw Average CF Sterile Bonara. P= Pupation. D=Death of larva. Age in Days. Fig. 3 Larval growth on dead yeast. W 3, 4, 5, show rapid growth on dead yeast; A, shows slow growth on sterile banana. reaching a size of 7.5 to 8 mm. in length on the third, fourth, or fifth day and pupating before the eighth day. On 24 per cent yeast the larvae often reached a length of 6.5 mm. on the first or second and pupated before the sixth day. Records of the growth of cultures of larvae, on yeast media of different strengths, follow in table 4 and the mean larval periods are included in table 5. In figure 4, curve 1 shows the rapid growth on 24 per cent yeast; curves 2 and 3, the maximum. and minimum rates of growth on 3 to 12 per cent yeast, and curves 7, and 9, the slow growth on 2 and 1 per cent yeast. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 1 PERCY BAUMBERGER J. 16 | a d 8 d L | 8 d dele 8 dis d |g q L d=) glee d d d 8 gg |g-¢ [ge |o-a Iga Ig Calmee |) 82 1G eal a 9| g Gz |g f shop ur avasn) fo aby b WTaVL ia a Bp eels 2G ZL See Pen reralfes ) LEAD NS fs Seales hk V (2) peste er laiee ees Sa eecullers EN TNE LS 8 6 G| G G G € | <@ ie.) Ye) NaN © il OC acne ee 4svok 4ua0 GR eee + 4svok 4u90 a © 1a) oa Kernen “aSBok 4ued @octteeeee 4svak 4uW99 qotteet ysvok yu00 poccccs4svad ques Ge Cee rec et! 4sBvok 4u00 Doe 4svok yuo OP SR aces 4svok quad pene ae ee a 4svod quad Q@citteete 4svak qua0 Zi sane nase 4svad quedo L 2 SaecnsroGoueo Qo asvak 4u99 (Ammer cies 4svoA 4ued Jod FZ Jad ZI Jad ZI tod Zz] dod ZI ied 9 Jad 9 Jad 9 lod F Jed 7 lod ¢ Jed ¢ Jed ¢ Jod Z Jad J eee > iim Conia staysodureg ae ATG jos vuvueq poe }Ue ULteyun snoonbe JOR vidaw A NUTRITIONAL STUDY OF INSECTS 17 TABLE 5 2 oes 3 Se Fe Z aH A a aa) ica] < fa) iS} a = q p a < Ze MEDIA y a8 aS 5 < 5 as Be 5 < bce eeriene | age! Ge (Wee g a 4 es A 9 > g g vy a a 9 > Hotaqueoussol. unfermented | 8 | 20. 3/5 banana 4 | 28.75) 1.29 | 4.5 9 |5.5 Banana mash.......... Pised oo tte 0 Bammer mashep seit conan ea) Agaricus campestris.......-.. 1 | 14 }12.43 | 1.89 | 15.2 | 14 |3.85) 0.6221 | 16.1 Il OG CEM ELISE cbc oops esoue 2} 0 |20. ZOO CEN WELSH cone cs ouaosc 1) 10 11.4 | 1.42 | 12.4 SIPet Cent VeASt ter. 22: 1 | 29 | 7.14 | 0.27 | 3.78) 29 |4.75| 0.2318 | 4.8 3 PERICEMbsVeAStpacts teen. = 2] 10 | 6.00 | 0.14} 2.038 SIPCRICEMURVCASty-yatociee cha: 3 | 50 | 6.32 | 1.93 | 30.5 | 40 [4.33] 0.0409 | 0.9 Apericent yeast............ 1 P28e SR055 le Soe) laid AWeMGent yeasts. a aa. 05. 2 | 36 | 6.47 | 1.70 | 26.0 | 33 |6.63)11.66 25.0 A percent yeast ............ 3 | 40 | 7.20 | 1.208) 14.00} 30 [5.58] 2.89 43.59 Gpercentiyeast......4--... Le a OON ROO 0.0 | 20 |4.0 | 0.0 0.0 Gpenicentiyeast'. 2... 4. ..- 2|21|6.0 | 0.308) 5.1 | 20 |4.4 | 0.154 3.5 GPETICEMtAVEASbi.s-e.- e 4 | 61 | 6.46 | 0.157} 2.4 | 60 |4.37| 0.2454 | 5.6 2 mencent-yeast . s. ... cn: 2.) 40) 5.8 1 0.5 8.8 | 40 [5.45] 1.07 19.6 i2pencent yeast...-.-...... 3 | 51] 7.0 | 0.713] 10.2 | 51 [4.46] 0.6361 | 14.2 IPC CENGWCASG 2). 2 ar eee 4/17/6.3 | 0.525) 8.3 | 17 |4.17| 0.41 9.8 12 per cent yeast ............ 5 | 50] 7.8 | 1.095] 14.0 | 25 |38.68] 1.2 32.6 24 percent yeast ..........-. 1 | 60 | 5.21 | 1.61 | 30.9 | 60 |5.44) 0.527 9.6 24percent yeast ..........:. 2 | 12 | 4.83 | 0.188) 3.8 | 12 |3.93} 2.02 56.0 3 per cent yeast ......average | 89 | 6.55 69 |4.50 A percent yeast ...... average | 76 | 6.85 66 15.85 6 per cent yeast ......average | 79 | 6.11 100 |4.30 12) percent yeast ...--. average |158 | 6.87 133 |4.07 24 per cent yeast ......average | 72 | 5.14 (2 |b. 18 1F, from adults reared on Agaricus campestris. These experiments show that dead yeast is an adequate food for Drosophila larvae when in a concentration of 2 per cent or more. e. Can Drosophila larvae complete their growth on any veg- etable food other than yeast? Bacteria and fungi other than yeasts appear to have some food value for Drosophila larvae, as in microscopic examinations of the digestive tract bacteria often form the bulk of the contents. The following experiments show * 18 J. PERCY BAUMBERGER that’ these microdrganisms are not as valuable to the insect as yeast cells. A few larvae were reared on vinegar-plant agar, pupating on the sixth day. On manure agar growth was slower and pupation took place on the fifteenth day. On lactic acid ba- cillus and on Rhizopus nigricans agar no growth took place, but the larvae died in three to five days. On plain agar infected with a semianaérobic bacterium a few larvae pupated after twenty-six days. Therefore yeasts are a more complete food for Drosophila larvae then other bacteria or fungi. I have already shown that fruit (banana) is of some food value for Drosophila larvae, as it will keep the insects alive for periods Lorre fength jo millinveters. 2 = Pupotina. 2= Death of larva. tn Days OSE ie ea a a EH et Fig. 4 Larval growth on various media. 1, 24 per cent yeast;2,maximum 3 to 12 per cent yeast; 3, minimum 8 to 12 per cent yeast; 4, vinegar plant; 5, mushroom; 6, yeast nucleoprotein, sugars, and salts; 7, 2 per cent yeast; 8, hot aqueous extract of banana; 9, 1 per cent yeast; 10, cold aqueous extract of banana. of twenty-eight to forty-four days and permit them to increase in size to a limited extent. The activity of the larvae and analysis of the banana indicate that the insect is abundantly supplied with carbohydrates (20 per cent sugar in ripe fruit). The protein content, on the other hand, is relatively low (1 per cent) and is probably deficient. The long life of the larvae on sterile banana with the accom- panying slow increase in size, indicates that all the food elements required for maintenance and repair of tissues are present, but the protein content is either too small or lacking or deficient in some amino-acid necessary for growth. There is also the possibility . ~ A NUTRITIONAL STUDY OF INSECTS 19 that some vitamine may be absent or may have been destroyed by the high temperatures of the autoclave. Some light is thrown on these questions by a comparison of the rates of growth of larvae on banana media which have been more highly concentrated by partial desiccation or by extraction with hot water. The growth of insects on these media is shown in table 6 and figure 5. On sterile food consisting of mashed whole bananas, especially when they have dried out slightly and are thus concentrated, an occasional small pupa is formed. On a hot aqueous extract of ba- nana a larger number is formed from which small adults emerge. TABLE 6 : LARVAL PERIOD NO. NO. : g MEDIA (20 ee IN See) |i 2 3 o Bamamaemmasia i. 22.12) vaeys « a x Pedietinage:(:...... x x x Limnophilinae...... x Limmnophila:... .:.... x x Eriopterinae........ elobpitapscte 24: ? x Gnophomyia........ X Hexatominae...... x x Trichocerinae....... Trichocera........ x Ptychopteridae..... x x x Rhyphidae........ x ? ? x Boletophilidae...... e Mycetophilidae..... x ? x x etapere yee a: 22.) sex ° IIRC CHIS). avis gis oo ts x x Bietarigae...//,.;.....|. x Platyuridae........ x x Psychodidae........} x x Xx x x % Blepharoceridae.... x x x Chula 28 cs a x x x x x Dixidaenw. er eee x x Ceratopogonidae... x x Chironomidae...... x x x Orphnephilidae..... x i Bibiontdge.<..-..-- x x Scatopsidae........ x x x x Simmulitdae:........- x x Stratiomyiidae..... i Xx x x Stratiomis.......... xX x x x Odontomyisa.7)....|. x x me X Oxvcersmerr ne. ts2| xX x x Geosarguseces...Js:| x Microechrysai.,..: 4s). x Eupachygaster..... x x Xylophagidae...... x x x Xi Coenomylidae...... x x x x Tabanidae......... lh Rape. Xi x Coniops..... a ete xx ? x 59 GROUP eptid aes eer + I Gherixet eer cers ae oe Chrysopilain:.. 1.482 Crater eek oe Asiloudes: a.62545 2a Mrydiaidaes ace oe ASUGAe 2.8 see F Dasyillltsiprasce cones Bombyliidae........ Therevidaew +) ..cae Scenopinidae....... Platypezidae....... Bipunculidae....... Synmpaidaese ee see Conopidae....... noe Psiliditiecen.: 2.c.see oe Sepsidaenenaceme Pry petidae...: aa Sapromyzidae...... Agromyzidae....... Geomyzidaes sane Drosophilidae...... Ephydridae......... Oseinidae: ete. cca Phycodromidae..... Borboridae......... Heteromeusidae.... Helomyzodae....... Cordyluridae....... Anthomyidae....... Mirrseidset 3... a0 Oestridaes...2.5.5..: Sarcophagidae...... Dexidaes akc. Hippoboscidae...... Streblidae.......... Nycteribiidae... . Sciomyzidae....... Saprophytic “~ w “nw LARVAL FOOD Algae Higher plants Animals wan KK KK OM 4 |: | 9 °¢ | | sian ED ad & 6 9°¢ aay OL ay ore ie Gate Ie lige iincuale ‘aay qdVvV ¢ || 8:8] P:al 0:91 8:8 aa! 8:8 aa 0:91 0:91 | ; 6 ‘8 | | Gesticea (Gee Shee (la lames e 9% O29 TAGAANG Qefoe| 9a Cn 7 | 8B Gg G'S qqey | F || OT: 0.) CL=P Bag 91:0 8:8 91:0 01:9 aaa OO = | Ga8 6 ‘8 9°9 6 ‘8 LaGG ged } | Gupte deel see. | Shee 26. O59 VEE ICL See PI Gao? Gal Yee oT Wee Oe. LEG & Gy: qadey | § Zl: h “| O19 8:8 8:8 aa 01:9 1:6 aa 0:91 | Ai | GaGa Coca a 46 G8 OG GED Ni Be GaG — Gh GerGog el sae BT adey Gil 8:8 8:8 8:8 0:91 0:91 PI aor | PGI 0:91 | g 8% 6 g cy! GB NG eagle lees Geed I \ | I 0:91 0:91 0:91 0:91 0:91 0:91 0:91 0:91 0:91 ie aia J Se Tae aave q4vV qavv qqey qdeVv dae ad VV Seas SSVTO xI IIIA IIA IA A AI Ill Il I SsSVTO Sole, * So[BULA YT SIYD]L UL pajursd alD sassppo ay, “fsoay) aaypuUsazD ay} WO pajaadxa sassp)I PUD soLMY 1 4TaVL INHERITANCE OF WINTER EGG PRODUCTION Merete moe Oml st T'S 20 Oh 0, Wl

6 lO GUE Cc. colle seee ec ose :oge | & : PanuyUOj—UOTPCIIUIS (Wf) PAIG-SSO1o PUODS oY} Jo SBUIPVPT N ls z| 0€ 0€ 0€ 0€ g 08 0€ © 08 0€ ad Ms 0€ 0€ 0€ 0§ Japuy | JaAQ |tepug | 10AO © | Jopug } IAQ | o |Zepug] 10AD a5 oi Japuy) | JeaAQ |topuy) | IAQ na 5 ef ; ones poqoodxg | o1yv1 poarosqag oye poyoodxnT OIPel poAIOsqO ‘S ie oye poqoodxgy | O1}BI PoArTOsSgG, oa q 5 s ° AUNUOAHL AAILVNYALTV AUOAHL 8, THV ad fe 5 ANUOGHL AAILVNYALTV panuyuog—p WIAVL AUOGHL §,THV ad sud ~HLOW AO HAHN ON LI LS GE YAGNON § YaHLVA 96 INHERITANCE OF WINTER EGG PRODUCTION 97 years, a good fit between observed and expected ratios is ob- tained, although such a result is hardly to be expected. More- over, the data for 1914-15, if combined with that of whatever other year may be involved, give satisfactory ratios except for one male, viz., male no. 271. The ratios for this male are, on Pearl’s theory, 3:32:12 observed, to 0:35.50:11.50 expected. On the alternative theory the ratios are 3:44 observed, to 11.50: 35.50 expected. The various years are not wholly comparable to one another. Certain changes in management, described in later sections, have been forced upon us. Selective matings have also been made in various years. Some were made for high production, but others were made primarily for low production, late maturity, hatching quality of eggs, vigor, broodiness, and size. Thus, a portion of the matings each year are made at random as far as egg production is concerned. In compiling these tables the progeny of each pair was first distributed into the three (respectively two) groups of pro- ducers required by each theory and the expected ratio that agreed most closely with. the observed ratio determined by inspection (compare table 3). It was of course necessary to select a gametic constitution for the male that would give suitable ratios for all the females with which he was mated. An exact fit for all females cannot be expected, but in only two instances has the deviation been greater than three and one- half individuals, i.e., a change of three individuals and one-half from one class to another makes the observed and expected ratios agree perfectly. Changes of three or three and one-half individuals are rare. The sum of the observed and expected ratios of the various mates of each male constitute the data given in table 5. Within the limits designated, we have not hesitated to choose that theoretical ratio which gives the closest fit in the total. In some instances it is possible to assign more than one theoretical ratio, but we have not done so. By accepting the discrepancies mentioned between observed. and expected ratios, we avoid classifying an individual breeder Gees. OF) Oe FE a) 2b Gog Ble Ge Bl Our, Pa SO Ad Gee call c eeeGy ko 1) Oe causa) 0-2 gece “9 0.9L°6% 6°99] 0 09°22 FF Fa + : filge°r :94'91| 0 :98°8: 94°F] 0 :09°8!09°91 Peay eee ON Oe ee NOG ie aOR 0. 09°96 09°99) 0 0 98°09 pe Ne Gee O9 fal 0. 8 Se ee 07 0: ag epieae) le Th ts Del Share a eet ee Onley 0.0922 09's] 00902 Ges 8 16s6 <99'°¢ 00> € -09 ¢| O- & & eo Nia |i aT 2 Si Peat cae ee Sn I9AO0 I9AO I9AO 1I9AO sopug| P2 fropug | PE | 1% lropug] PY | OH bropua) Pet OF 0€ OF 0€ 9350 UI 3 UIOd UOISIAIC, « sso UT JUTOd UOISTAIC, AUOTHL AAILVNYALTV AUOGHL §,TUVad SUAHLOW AO UAAHWON P2jD]S VSUMNAGYIO SSayUN ‘spay pwH)s] apoyy SUD] UL pajvadxa ‘UwDWoYy UL SO’DL padsasggd poop f-yon7q ur dnowb yooa sof uoyonpoid bba woop pup ybry waanjaq syurod worsrarp jouaaas yyun pup fisoay) aayyousayy 9y72 pup h1oay) 8 J4DIq YIOG UO paspg ‘pajvad xa PUD padLasgo ‘soUvY 0 &F 02 92 8h 0 L9 LT 9F GILOEV |196°8 = 9h°& 9° 9) 0 = 9 + DUS O- 9 = SES OOTEY cle i L GOO teenme hs OL os OS Sig 0 19°9T 09° FF) 0 L9°9T 09 FF 9/S0EV Goes 8) GS OG Ne Oia ee Ome needy | CIE TSN, Dea iG Oe 2G). Olsee SOME Ge OMeenO) serine 0 09°42 02 2% O GT LO &F zicoev log 7 097 09 094) O : & = 9) OF F&F HI? HCV ive came! Gar cr ie) Po =< Gili O52 Gr ame 0 $1 8% LT 6%) 0 OS GT 98 OF FErLV 41 : £|09°9:09°71| O -O7ST:094| Of 9 = Gir LOTEV Tees i) Z Allen Senn di are) oT |e) Som i! 7161-8161 e IdA0 IdA0 IadA0 IaAO | 4 he ia OF 0& OF 0€ ie B z repuy ae iepuy, in 0197 |rapug ae 0197 | rapug a : E i a 7 = OF 08 OF 0& 2 iS) $399 Ul yuIod UOISTATG 8399 UI JUIOd UOISTATC 3 i) : a] 3 s AUOTHL AAILVNUALIV AUOTHL S,TUVAd 5 ‘adh} G WIGVL ‘swaonpoud a20vpau 98 0 29ST 0| 0 29°ST 0 0 ST &h| 0 TT 92 °'0F OP TIT: 09'T \94°6 7968S | O: SI - O} O° SI : O}} GEeceVv |l94'S : 98° \94°8 :96°E | O -098 2009'S | O -:09°S :09°S |> Z6I9 Sl = 0 Shee Wer eb ooo 0a) SOL IST 0 Viera as Sel Nee (Ym PMs oh 0 SG 0 IT Go| 098'L 09 °2F 0 086 0| 0086 0 09°01: 09'T 6 SalaeGe Ole ONES cum One ay MEP ee GE PNGG Cle GET 4s Ss Ms = OSS 67 GLE De == Ol FOGGY Fl ote SH) Or Gi lrwc Greene Ot Ol eee eae iss oe ib ©3{0) PISO Pele an Paes Ol, a0 : 0 €8 02 0| O OF LT 88 0° 6h Ou 0! 6T 0 GG EOL Ge 9. oT! OO: £ & ONOE TOG G = 0.1) ZISZ EaedO Ls ON nOo Seen Oui Onan a= = OM Gue yO £O LAE a RY SF aa TY ES (la a Pmst Quali Alert 0)! Ot: Te ee Opera rei) (ame O10 y) 0 0 8T 02 Opaear 0 94°92 96° T 1099'S 2 0G T | 66°15 94g 0 | Ge I° 4 °¢ 0 | + SlE8z Pe ONESOR E60) RO te DP -* 400), (Oe ah iaee amon inal Oar es) TERETE (4) rpc Pieceaacate (5 Pere Hem) a leaner) oe Oulemas OncQres . Tl) say ON me ORemiy eemuae() 3 o> 0 §8'8 0; 0 §€8'8 0 0 L9°8T OF} O FT og oy S681 GLO \96 SI59L'01 96'S: 9L01 - 0 196°S:4L°01 - O01 910G% Grape gel CRA) OO. DUG (OAS Tare 2S SITY: tI : O Vale ieee] 5 Soir ld} OiieGee 6 - © GxAre Sale Ga FP? eS Se Te 0 aes 0 G18 0; 0&2 8 0 0 &2 9T 0s} O OT Gh GEE 2 9L'0 |96'8 G40 | O fees PON One et 0|> @Zt 96° S947 Lo GA OOF N0G86 4. On0G GOs. 2/208 yr: (OO ta ON! Oe" F Ol 0 =F 0 Gots 4G Plea (Sal) Score “meee A] as: cree ns SI61-V16T 0 09°23 Og Ts} oO Pe OF 0 7o 09'TS| O FT ST LF Re TO MOG TG ONO Oo 8 te Sol NNO) WOM aOG=¢ sy riPCaVi ip 2) é SiO po AO @ =< 8 lp Lely ae A ¥:9 Ae eet Sha ech rl en ts rd ans) eae 0) Cire Se AP cee allen tee Ge es 0 06°92 92 6F 0 61 88 SP 0 §€9'22 §T'§9} OGL TT 92 °€9 OF ¢ 61969 -94'91| 0 706° 4°09 71| 0-09°9 -O9 SIl> SISOSV |148'6 °SI'2 109°9 -OF TIGL 1:0¢8 -46°4| O-: 7 : SIs 9ezEeV Olea cl Ono (je tet() cei Mec Ao) MEPs (o) | Guns G6) Teor S88 | Pe Poe . 0). 8 - "2h Os 0)08'-6E : oeeer 0 Ose st 0 | iors 54 Le oF (Tae Comets | O° O = @1>TI6I9 (|l09°T7-09°0 Ile 7 NOG 0G B= O09 F092 = Only SiSZiy [ceeoieeOeee Gel Ou 1 -< ) fle cOpse tone PGi: MeO0e GT: 0) eeeeGees 0) |) 8 = iereet 9161-161 eect) eh Oe (el Oe 0 || [eprom (0) OF Sen, Oe ee OL TSGs: ibraneeally De pre) Oe O LO T(0) se 0 SLT TE 009°8T 9% OOS 8T 9 96°86 GL 9 \GL 16: G6 k 9° SF 5 ONG6'9 GL G6 = OP LITLEV |194 7-96 7 \8h I-96 7 ey | 0: OFT -O08°T | > Glee 66 : 0 8% : iT Li 6G 30. He Wigan |b Go eal Go te! OReesGn mek Di Cauca O88 4T OF O66 9ST 8E€ 00s0e 0 00g0e 0 6 : f[ \09°£ -09@ Setar aS) O:- £ + & |¢ SIPCEV G = OGL & «G6 UE | Nees O- G = O04? G/692z Gree SG [2 cto | St eaearm6 eI) ey 0) Tee eee BAO) eh 0 €€ 8 0 0 €€'8 0 0 eF6 2 0 &h6 02°8 -09°0 |02'8 -0¢°0 Gets a0 Git dee. OL EIS STL 2 h8°T 1969 2 9L'6 109° 2099: 0 |96°b:09'F - 96° | > E|F08 Lie era) Gi. OF S28 se ee gence eee ) Sa ae Sica ak {a | Sareea E OF I9A0 og IdA0 OF I9AO oe I9AO0 4 ‘a OF I9A0 oe I9AO OF ISAO 0g I9A0 4 Kd Jopuy) Ce Jepuy ae 0197, |\ropuy ae °17Z |ropug Lee : ee lopuy ee repuyQ Pat °Z | apuy ee 017 |rapug res i Be cs a F *. OF 0& OP 0& ¢ = OF 0€ OF : 0g g wo wo Bol eee ple 83239 UI yUIOd UOISIAT(T S550 UL JUTOd UOISIATC 5 S Ss50 UI JUIOd AOISIAIC, S550 UT JUIOd UOISIAIC, A i-} io} rH I Zz AUOAHL PAILVNUGLIV ANOWHSL S,THV Ad a z AUOGHL AAILVNUALTV AUOTHL 8, TUVad i z panuyuojg—sG TIAVL 100 eS 0 02°F% 98°6b| 0 26°6T 02 97 91 : gilo¢'st:09'0¢| O-: £E : £1} 0°09°6 :09°Tél) SiPLZV O67: Ell Prema eeUe-nOcera til 0.:. ble met0g Ois-O> <= 0el > Oren 0 (ystu O12°9T Lb OOL FT 9°68 Pee sO) of 2-220! “Faas TONG ICO IGEBIOOL O US Bisel lll = OxOu ti OG ale ssO) sgh ce ea eOlbOny, neo FP (cha ea Ole (0) iba ane 0 FOV Bleek Wh See Pesala y= aes ey mean ecm cs 0 0b 6 6s} 0076 §& 009°%%2 09} O O88 FI BE 6 Ca 4 $ |0¢ 1. : 09°9: 7 [£81099 -S1'7 G189 68° ¢ °89'¢ \09 6 -08 & Oe Sse One en 6 Cae ataaler Yu Ge sana Ai ee oe Gi Siw Clee Gy ~ he 84) - ele uae aca eee a namees ac: 00F?T O| O86 GP 009°8T 99} O09 8T 99 OF: — | Sl + 6 \9L'9 = 8 3 98'T \98-4:98'6 :09'@ |} TePLV 9 SG a Oe BNC ot SOLE ap ogee saa Gece Glee. Os Piewasee|, LOS Ol Re Oe| nO ates. ara g Pee os oye etek 0) Sembee elton aman 0. 2090 GL 3 98°T \48°9 :81's 09 L094 - O| 0:09°9:09'S |} Sle9cr |l09°0 :09°0 \o¢'0 :09°0 090 - T:09'0| O- O : T}> TIezeV Ga.= = 0 Saal [Wea fiptoseay 0 ear one CR Gy 2 cuca | Tee OTs a ec Ol el norsk OE arial seme | | | | | | 0 L0°€% 26'€9| 0 bh LT 69°09 0 L8°'T% 88°F} «60 BE ST EL TF 98ST: 9L°8¢\09 OF: 09°TE| 0:09°S1:09°98@| O° FL°6 5 96° GE) PTTILIOE ||09 17! OF TT09 LE! OF 8I|9e ST: OF 96: GE ET\GL TT: 0996: $271) > B|LOTF ace ac Dieetaca| mle ert «2a het Oe 9 ~ 6S WW’: 6 SS: oo Gill) (Gis Cet ee Gal AGL Vo 0G. ee SI 0 88 °€% OL 6% 0 9% GT SL EF 0 99 LT Th} 0 98'9T 09 9E 09 88: 09 TI|09 SI: 09° 8I|96°9: 09 LI: G8 O1| 02961! GL°06| > SISTSS |]9L S15 967 |GL ST GOT \96°7 2h GL ET O96 7 © 6 HGLT |? HILOV vo .01) 9OL-- 8T i 2 lire 3 {OI Send tele! tell 20: J OT eas ino él oI Spa Nh ere 0 9% 0 0 92 °ST GL PE 0 16 0 0 99 TZ Ne Cresoina Vi 9 196 6 Ghd 0 \§F LT. §¢ -£8°& |? €|/P16 G - 0 \09 ¢ -096 Os Gc 0) nS Ge 9 €66V Or 0 9 P G+: 8 0 G Vv v G0 a 0 Go ie At) Ole ear 1% O aT 6T €h| 0 FT LT $9 0 0 9%| 0 0 916-968 \48 6 -&9'& 0: 6 : F |GL° T5099 5 9L°9 | + HIEESE Oe a 0 I Oe 10 T OO Lites Ol = € G21. ap G:'8 + § Gand sdb. 0 I 0 I 0: 0 I OSee0 101 OF 05 28769) 0-0" 16: 6a 06986 TL GS} 0 08 EE Lg 0-27 8 ORES A a este OR 8.) OM sO tees Se en LG h - 861968 -G4°18| O - FI 2 Tél O 296°E- 94° T8|} SlZT9e Qos OF Sa 0 Ors Sr Oke a0 8 |) DE 0G) Gh OG a Oma 1c) Ores 108 QrOg 2h Si see 0 eh rE 19 099 TE 09° 8h 0 L6 82TH 09° £0946) GF = O8 O -94L9-96'8¢) O° F -: TS) HPGeP |148°e -SI'7 1940-5969] O 096 7096 | O :e4:0:9¢'9 1 elrzzZ 8: 2 9 OC Oe eas) = ALCP OLS OF t= 6G Ge oF BS Oc) OR see ee ees) OR Ie 098° PE 919%} 0 OO O08 TF O T& 98°69| 0 8ST #9°8¢ OG 4 - 09h OE Oe OSs Cue OSG One OR ee SG aah Gy ah a0 G)) (O8 = Oe ECO ONE: =. G0) -2E)G99 Lae 8 DS es There SHE ME {0m mey ei) Shes SLl) AL Ee VIGO! = ee ee CeO eee nee 069 € O€ 8& OTT LT 89°3¢ 0 9% LG LS) 009°ST ZS 48 88: §188\96 61! 96° L7| O 20948 09°6E| O + $8°9I: G4 °08\ > 9ISZLF 7S AIST CE HLE°L | OF09°'S 2094) O 594° 986 |} SIL20S Reo LOG S620! Cee OG: ae ecOG|! Gays) «Si Leeaney |) Vv ZL Ge OSC CO) Ces a 2a) Ob Gee ee AG 0 61ST 09 6% 0 09°2T FI SF OF OL 98°63) 50>= Or ee G6 16: GT \8E 61/696 G6 £20971: 96°L |69' 9° OF TE: £88 | > FISLEF @ 2 S097 SOS 2 | OF Ge UP) ORs Ga Aan Oey Vioe 35 CO eeeaieeseemOL 3 Gel Sos PL.: 2 6 L Gy Zt Oy ee een Owe eee Mall) 0 LO LT Gh LS] O TT OT €T'E9 0 8& 09°86 0 OO LT'#8!) 48°61: &1 O6\77 6 7 99°84} O 2 OFSI:08°6I| O 098 09°76) >} FISL9 Ocean 0 BP 8} OF ONO = O05 © 9 ORES ie Ol OL Sole L Glee) eee On mane I ¢ 0 On ihOe Salk S369) 05-210; :_ Agu O0GL°8 €€ 9% OGL8 8 °9F 0 €€°92 OG Lb} 0 02 GT Sb SF 0G G6: 0G GF We! £)09°8: 7I-09'¢) £: 7E + 2 |) Spee |196°8 :92°8 log 7 :09°@1| OF 2: Ol] O 2949 :96'TI|} FlezE9 SGr aac SG Galo e-— al or [ertsl eu Sra! & rere ae 6) Gee onl) (052s 6 | O Se GE 416T-9T6T OF ees 0g tons a F ENO : oe eae A a OF 9A0 oe zeAD OF I9AO0 oe IdA0 4 Z lepuy ay rapu oe DZ |rapug Pe onZ | repay ae 5 Be pepe | rae oly tack 0197 lapug PS °Z | rapug aaa 5 Be | ee = gj “ OF 0& OF 0& 2 OF | 0€ OF 0& 2 eee Palece 3550 UL JuIOd UOIsIAIG, S550 Ul JUIOd UOISIAIC] 3 3 $550 UL 4uUIOd UOISIAIG S559 UI JUIOd WOISTAIC, ’ 5 o & Zz & 2 AUOGHL AAILYVNUGLIV AUOUHL §,Tuvdd a 2 AUOGHL GAILVNAALTV AUOUHL §,THVad a 2 panuyuojg—G ATAV.L 102 0 8T 0| OTL TT GE 09 £3096 \0F 2-096) 0: Of © OGG Ts @ = 92°86 | TIGose OL = 0 Gir ARGU ene AM) |e [oes eee ee On” Ce Saez 0) 0 gZ09 ‘8 IIGBT, 999 1616 ||64°@ -96'°9 0 8 | 0 09:5 -09-7.| 0 = 0 = 98 | ¢eizars ‘9 98GB, 999 ¢008 g G 0 Sulies0 CX Gal Oeitin Oia <8 0 68°26 80°89) 0 Zh 8ST 12 2b 0 Go 8h 191 0 9 0°09 4E°87°69°ST| 7S : &&| O OF TT/09°9é| O -09°88:09'S8E|> TIZGLZ6 |l9L'6 :96°98| O -: 88 O 2646 :96°9S) 0: oO : géis siz608 Poe elle Oc Oo “Psi: EL cOh SEs -6z Gor Cg) TL ANZSI OY -peGee POSE iec dean 0 €89% 99°09} 0 OF LT OF TS 0 T§ 9T 68°89} 009 TT 16°99 8 - £109°9-509'8 | O :094 /09°L| O ELE ° G6 TT|> S/E2E9 ||96°ST- 92° 07/09'6 09°77) O 2 eLIl Se @t| O :09°6:09't4\s 9I1gzc8 i} G Ol S20 ter Gerre One OnraGee mae MOT fected (0)) emits 1] feae( 9 eisai 3 eine /()) Pei e S 00S°SE 82°29| 0 £OT OG 2g 0 €T €9°99) 0096 9&8 §9 GLE i968 0 Il} O :09¢-:09°'¢| Of O ¢ TT> GLeLL |1G4°9 ° 96 6IGL Si 96°66 O 2099 50961] O :09°S:09' Gels FIT928 Pes 2 a0 LE 02 Speer ORS Oke Pon Gere 6h) Gare eG) 10) -289) -- ecGL) (Om ecment Oe 0 26°23 09° 2b 092 ST 98°GE 0 99 0% €9| 0 99°02 €9 Ci NOTES SOGc9N Oo Ol | O «ORL “09° L |> FOLFS SS L0G Fa O98 | 0). 39s at | OM OG 09 "S| seiGry le te Pee sO Chae gO. gs 2 Gr iinet): Say eA Oe ee dae | ee ectaec ole we Oy 0 61% 68°F9} O €L°8T 99°09 0 8T'08 §2°€9| 0 82°%% 98°89 91 += 9877 O1'99' 17} O - OT: 98) O 2946 509° ST) LAPIS ||68°7T- 89° 891989 -9L°19| O :09°8I°09'9S| O °96°9: 94°19] 812208 eae ier Os OF? O8 Ob TL —-TF Oiler ae BEG we Bek SOON ie MG Tate lhe WAT ace a ale ; SI6I-LI61 0 al 0/| O tT 0 | » USTU 0 49° FT OG'TS| O19 FT OG 'TS Geen 0 8 Obi! ees aor Osher ea 0 |) S| ~20D 196 'OF 94° 1946 °96°7 19465 £ 296%!) O 209°L509'9 |> FCISF Cee Oull (Ol easy Name Dale ieee ell Ole: at OU Pall glee sm One sce Palle sige er Gta s)- mney 0 GZ 16°€9| OLO ET FILS 96°71: GL'Y \09°6 09'6 \48°S:096 °S1'L) O 2096 -09°6 |} T/Z88F 09° SI:09 Lelee'y °92°96| O :09°8T:09'Lé| O :09°9:09°98)$ 9IIT8G¢ Pls: Qin Gallic ss ea eGul g SG ere OG CT) 82) OS GS) Ose SI eRe) Gh Ome eee 103 “1gA90SIeUM LuUdSoOAd AUB 9INDES 0} JOpPAO UI ATQvOSUBYIAIZUI SoTVUI OMY 9SN 09 AIBSSID0U SBM q Wa) q [ So] 41 x ae 0 69'6€ €9 FL; OLS TE TH 89 0 GL TE T9 ZL) 099 € OG 69 09° Té- 0G 61 ST: 861 O /09°08:09'02| O ‘09ST: 09° 48> WLFI8 1]89°6 2 LE°GT/GL'8 296° 91] O -OF6T-09 EI) O -09° 4-09 LI) ¢ FI9L8 GG: 6T Fl: Key O « 6G 8 ill> Oo WPL 2 = KG Ge wana Gen i Wo gilas tsi) We cS) ; 0€T Lb 02 91) O op 8T 99 0 LG LE 08} 0 92°92 $8 99} | 69 8b: L871 TT: l@| 0: 6G - 61) O- OF : 8é@l> GL608 |1\94'9 - 96° \89 & = 46 9 (Oi om EES g (GQ) TSE 3 d EISPLZ 6G: ST IT: /pAl 0) oe SG 8 SSI (LL a Le fh ts NS | este) Oca 7 is Oy: 9 0 PE SE 61) O 93 Th &L 0&6 2h 96°92) OTT SE 20 L9 LG: dG 61: GG} O: Le: Le) O : GG: TE)? GI8cS8 6&: 6819488: 96'77| O :09°86:09°66| 0 - 8% : GF) 812208 8G: 9G OG: ig] @ s- te 9 Gal Woo Wee sins Gy +: «9G VG Fy PS OC a GGr = Sh (soupy YOu) 8161-L16T 09 I9A0 0¢ zeX0 09 IdAO 0¢ I9A0 A es = 09 IdAO oe I9AO 09 IdAO 0¢ IdAO 4 cz Re riepuy Lain Jopuy) ae DAWA lopuy) aa ORY s ropu() aa eS B i iepu Lie rIapuy) x Ore7, ropuy cae ENS iepug oe 5 _ : = Es ce, 5 de 09 0g 09 og a 7 09 0g 09 0g 3 z z e z > S550 UT JUIOd UOTSIATG . $350 UT JUTOd UOISIAIG 5 S Ss50 UT JUIOd UWOISIATC, S590 UL JUIOd UOISIATC 5 g sl Z a Z AYOAUHL GTAILVNUALIV AUOGHL §,TUVad a 2 AUNOGHL AAILVNUGLIV AUOGHL §,TUVAd a g papnpuej—G ATAV.L 104 INHERITANCE OF WINTER EGG PRODUCTION 105 as a ‘somatic’ high, mediocre, or zero producer, save in a very few instances, and these all clearly result from some peculiarity of management, such as birds hatched late in May with records of nearly 30 eggs. Such exceptions are apparent rather than real. In one or two instances females, such as female no. 6067 mated to male no. 5584, table 3, have been encountered where the number of high producers does not meet expectation. In the case just cited, the progeny were obviously subnormal in vitality, but were kept because they came from a high-hatching line. We can readily understand that cases may arise where a bird is genetically high although its record is mediocre, but it is hard to see how a bird genetically mediocre should markedly transcend the division line between classes. Some records, especially records not far above the division line are to be expected, but since it is a universal experience with Mendelian ratios that individual families frequently deviate markedly from expected ratios, although the average fit closely, it has seemed wisest to take this way out of such a difficulty, rather than assume that the genetic constitution is not represented by the somatic record. : In compiling the tables, the genetic constitution of each male— is first determined by the ratios in which his offspring occur. When this result is checked against the parents of the male in question, it may happen that a male of the class indicated by his progeny could not have arisen from such parents. It is possible, however, to adjust all such discrepancies where a division point of 30 eggs is used. Doubtless adjustments can be made in the case of the other division points, but it does not seem necessary to pursue the matter further. Inability to make such adjust- ments would indicate that neither theory has any basis in fact. The same care has been taken in classifying the females. To do so, however, is not as simple as it seems, for one change often involves others, and a long chain of changes is often necessary in order to reduce all the data to a harmonious whole. We have gone over the data with this end in view, for both theories with the division point at 30 eggs. With the single exception 106 H. D. GOODALE AND GRACE MACMULLEN of male 3003, as given elsewhere, every case encountered falls into line. The data, therefore, are reduced to a harmonious whole—a fact that speaks strongly for the validity of both theories. It is also noteworthy, where families are large, that they fit the scheme with practically no difficulty. We do not feel entirely confident, however, that it would be possible to accomplish the same results with a really adequate series of data, in which the families are of sufficient size. On the other hand, it is certain that one could use either scheme as a guide in breeding only if one knew definitely the gametic constitution of the birds he started with. This knowledge can be obtained either by a series of breeding tests extending over several years or else one must have available a progeny from each mother of twenty or more daughters. It is impossible to start as we have done and make the progeny and parental tests agree except by constant shifting of birds among the various gametic classes. One family, sired by male no. 8027, when a division point at 60 eggs is used, fails to show a good agreement between observed and theoretical ratios on Pearl’s theory, due to a deficiency in the expected number of birds laying over 60 eggs among the daughters of high producers. Such a deficiency is explicable, in part at least, because of the ease with which a record can be depressed below the division point, through environmental or managerial factors, for 60 eggs is at or near the maximum pro- duction for birds beginning to lay December 1. Thus, it is easy to understand why a portion of the daughters do not reach 60 eggs. : The only real exception to the application of Pearl’s theory to our Rhode Island Red data is the case of the family sired by male no. 3003. The detailed data of this mating are given in tables 6 and 7. In some respects the production of the daughters of this male is similar to that of the daughters of the Cornish male described in another section. In 1915 male no. 3003 was mated to several poor-producing birds, primarily in order to secure a flock of non-broody Rhode Island Reds. As will be observed from the table, all his offspring are either mediocre or zero producers. 0 GL’8T 09 LF | 0 08 FT 809% | uorjzonp 092°0T0 |0 2 0TO; | worjonp -01d 359 OSVIOAY -o1d S30 odvIDAY 41:01 109°6 09°41 é é “*syej0q poqoodxgy || 09 01-09°@ | O1-§ é é “*"s[830} poyoodxy POL | Sah Gh eG a Ol} G Olle: s210 Mar. 6 June 29 Mar. 13 Feb. 16 Mar. 18 Mar. 27 Mar. 9 July 2 Feb. 20 Feb. 21 Feb. 18 Mar. 28 = aS | Eggs to Mar. 1 ow Mother’s band number D> cS) > = 6982 5832 3180 2453 Daughter’s band number B1023 B1881 B 749 B1035 B1036 B1037 B1038 B1209 B1210 B1713 B1714 B1885 B1886 B 347 B 348 B 476 B1020 B1239 B1562 B1907 B1909 B1910 B2474 B1211 B1785 B2385 B2466 1917-1918 Date hatched Apr. Apr. Apr. Apr. May May Apr. 1915 Mar. Mar. Apr. Apr. Apr. Apr. May May May May 191 May Apr. 29 6 6. 25 29 6 6 6 10} 4 20 5 1914 Apr. May May June 22 6 20 3 Age at first egg 221 Date of first egg 295 | Feb. 21 Nov.22 20 eels) Eges to Mar. 1~ f=) INHERITANCE OF WINTER EGG PRODUCTION 109 In 1917 he was bred again for the same purpose as in 1915, but this time was mated to other birds also. The results for 1917 are in striking contrast to those of 1915, although the records of the new mates are only slightly better than the old. How- ever, if the two years are combined, the actual ratio is 12 high: 14 mediocre: 11 zero, which may represent a theoretical ratio of 13.75 high: 18.50 mediocre: 4.75 zero. Cornish male by Rhode Island Red female cross The detailed data of this cross are given in table 8. The results of this experiment are of prime importance for theories of the inheritance of winter egg production, because it demon- strates, in the Rhode Island Red breed at least, that high pro- duction descends from mother to daughter. It is important to realize that the Cornish stock used in our experiments was ob- tained from the same source as Pearl’s. Further, our data on the Cornish are in good agreement with the numerical results obtained by him (Pearl, 712). It is likewise of importance to note that the Cornish females bred to the same male as the Rhode Island Red females give a strikingly different result. If we attempt to reduce these data to the form of Pearl’s theory, we at once become involved in difficulties. In the first place, no high _ producers at all are to be expected, yet we find that twenty- eight of the thirty-one individuals are high producers. While the average winter egg production (49.2 eggs) of the cross-bred pullets is not equal to that of their mothers (71.6 eggs), it is nearly equal to that (52.5 eggs) of the mother’s families, i.e., the mothers and their ‘sisters. The average pro- duction of the mother’s families seems a fairer basis for com- parison, as the mothers were a group of individuals with production much above the average, and regression would cer- tainly be expected. The average hatching date of the five mothers is April 16; of their families, May 2; and of the cross breeds, April 19. The differences in time of hatching may 4 Unless otherwise indicated, a high producer is a bird that lays 30 or more eggs before March 1 of her pullet year. 0 ST €@ TL Gg 0 PIT FT 6E TS | °° UoTjonpoid 380 ostIDAYy 09°81: 09°46 SL FZ | 98°S1/ 94°86 | ZI: 0E |]0 “OF LT/09'7e\G > EL: FZ \O°FL 6°96 |G + 1: OCR assy 81870. G OM Gan) G 0 Go WW0 oc Gs Whe > M > WO [O- [Po PNG oN) 8 (Os OA ELGP § I 6 I é é Go GN os js Gis. Goa Ik (OS Ge GH ike flo = [fhe ANDOU €806 g 0) || Sie 0) G 0 COON Ges ON GeeG. - O MOOS CG 09 OG |G Ge Ome OG tec 8666 YSIUu.0r) 0: Y | Oo & 0 WW > 8) Ie OFS Om Oy Ok Ole A read ees Dirake seen OF 1940 6EE8 Gs GG a I iho ty Ie Oe BO NOS AG wie amine coferen | Teme Ore a ec mes OF 19AQO 668 &-: CamleCaueya OS" fee X09 7 | So Go 810 2a = Sy0- 09 7 ae 10 oo & OF 194Q GL6L 0: Ge Mello 7 0 8 | 0: 8 |/0 Oe Pi o fs io Osa |0 Ds saa te} OF 1940 OOFS OG Ge OLae Wee Ove Ge" 9 T= 9/0" -0956 -0956 \0): Gy WOeSZ = 66910 Leon) OF 1940 6608 Spey pues] 9pouy s SS) oS eSloS| wo o8 wcloS SCIPS! SoS en eo Se eaallecdSe leer Sues oS oryer poyoodxy “07 Ee re o1jei poqyoedxy “0g Bey Nits O1jyel pozood xy “Of | OlFBI[VNJOYW “OF jOTJeI poyoodxG 0g} OBIeNzyoy “Og waonaoud MaaWON 40 GNIX aNyd §,uaHLOW S539 UL JUIOd UOISIAIG 8390 UI JUIOd UOTSIAIC, AUOGHL AAILVNUALIV AUOTHL 8,THVad fijajpavdas uaarb si sayjou yova fo fiuaboud 04, “[6]6 Ou ayou ysiusoy fo iiuaboud fo uoyngrsig 8 ATAV.L 110 INHERITANCE OF WINTER EGG PRODUCTION iiss introduce a slight error in the comparisons. The error cannot be estimated, because two lots of birds, hatched a week apart may differ by as much as 30 per cent in average egg production. However the regressions from the mothers’ families’ average is about the same proportionally as Pearl observed in the case of the reciprocal cross, viz., Barred Rock males by Cornish females. Several cross breds laid more eggs than their own mothers. This happened in two out of the five families. with an average excess of best daughter over mother of 215 eggs against 153 eggs of mother over best daughter in the other three families. If the Cornish are not restricted to the three gametic classes, viz., 5, 6, 9, as is done by Pearl, it becomes possible on Pearl’s theory, to secure a set of ratios that fit the observed fairly well, if it be assumed that the zero producers are physiological zeros. This assumption is plausible, because most of the Cornish are late maturing. Without this assumption it is impossible to find theoretical ratios in Pearl’s theory that fit the observed ratios. Nevertheless, it will be observed from table 8 that considerable violence is done to the observed ratios in the case of one Cornish female, when an observed ratio of zero high to three mediocre to two zero producers is referred to a theoretical ratio of 23 high, 24 mediocre, and no zero producers! It is also necessary in other years to make similar changes in fitting observed to theoretical ratios, on the assumption that male no. 9191 belongs to class ITT. There is still another difficulty to be explained on Pearl’s theory, and that is the occurrence of a single Cornish female that is unquestionably a high producer. It is difficult to believe that she is a mutation because she is the daughter of a bird that though late hatched and somewhat slow about maturing, laid 27 eggs in a single laying period before March 1, i.e., has the characteristics of a late maturing high-producing female. On the alternative theory, no difficulties are encountered in fitting observed to theoretical ratios. Some of Pearl’s observed results are in agreement with the results of our cross, particularly the progeny of Cornish male no. 578 (Pearl, ’12, p. 373). Here. a number of high-producing THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 1 112 H. D. GOODALE AND GRACE MACMULLEN birds appear where none are expected. Moreover, Pearl’s data shows that 15.5 per cent of the F, offspring from Cornish males by Barred Plymouth Rock females lay more than 30 eggs, the average excess production being 15.7 eggs or more than 50 per cent. Now, the average winter production of these over-30 birds is al- most .exactly the same (viz., 45.7 eggs) as that (viz., 46.2 eggs) of the over-30 birds produced by Barred Plymouth Rock males on Cornish females. It should be observed further that in most cases the mean winter egg production of the over-30 group, in cases involving the use either of F, or pure Barred Plymouth Rock males is also low, ranging from 35 to 57.2 eggs, the average being 45.9. While it is not feasible to pursue the analysis further, because of the form in which the data is presented, yet in view of the facts presented above, it is not hard to believe that the results of these crossing experiments do not necessarily involve sex linkage. THE BEARING OF MODIFICATIONS IN MANAGEMENT ON THE RESULTS When these experiments were started, it was with the intention that the methods of management (i.e., the environment) to the smallest detail should remain constant throughout the work. Unfortunately, however, it soon developed that something was radically wrong with the method of rearing the chicks, which made changes here imperative. Later, results of the experi- ments themselves made certain changes in time of housing the pullets necessary. The various changes made are given in appendix 1. Of the various changes there is only one, viz., a change in method of rearing the chicks, that appears capable of affecting the results to an appreciable degree, though it does so in two ways. First, this change now gives us with certainty normal adults to place in the laying houses. Second, it makes possible the maintenance of the laying flocks in a state of freedom from infect ous disease, particularly roup. Roup, althougn very capricious in making its appearance, is favored by certain envi- ronmental conditions, expecially the weather. Moreover, the INHERITANCE OF WINTER EGG PRODUCTION as birds react variously. Some are apparently immune, others take the disease in a mild form and without apparent detriment to egg production, while others take it with various degrees of severity (our Cornish, for example, are particularly susceptible). Such individuals may cease laying entirely for a time or lay spasmodically. The presence of roup, therefore, complicates matters greatly. The degree to which these changes in management have af- fected our records is a matter of surmise only. We have no precise data on this point. The evidence from the records themselves indicates that it may be of small moment as attested by the presence of low records in clean years. Nevertheless, the presence of these changes in management introduces an element of doubt, especially because of the variability in susceptibility of individuals and families to roup. Size of families There is a difficulty,» common to Pearl’s data and my own, that renders it impossible at present to decide between the two theories and their several modifications and which renders it somewhat doubtful that either scheme has any foundation in fact. This difficulty les in the fact that the adult female off- spring of each pair are so few in number that it is almost always possible to refer any observed ratio to some theoretical ratio that will bring each mother into line with the rest of her group. Thus an observed 2:2 ratio will fit any of the following theo- retical ratios, viz., 2:2, 1:3, 3:1, 3:5, or 9:7. Since the progeny of a single male in numerous instances is fairly large, it may be urged, by the law of errors, that the agreement between observed and theoretical ratios is adequate proof that the scheme correctly represents the actual mode of inheritance of winter egg produc- tion. This would be true only under certain conditions, which are: First, there must be no bias. Second, the sample of females mated must be a representative sample of the population. 5 This difficulty is not confined to egg-production statistics, but unfortunately is encountered in much Mendelian work, particularly with mankind. ae? H. D. GOODALE AND GRACE MACMULLEN Third, the sample must be sufficiently large to include a proper proportion of all kinds of females. Fourth, the various females must produce approximately equal numbers of progeny. Not one of these points is fulfilled. The first, because it is evident that a bias exists which is the endeavor on the part of the in- terpreter of the data to secure a good fit between observed and theoretical ratios. No method exists by which this bias can be overcome. The second point is not met because the females are selected samples. The third, because the samples are com- paratively small. The fourth, because some females produce many more progeny than others. It must also be remembered that the application of ratios is made to the progeny of each mother separately. The importance of size of family is not merely an academic consideration, but very real as anyone who has worked with Mendelian ratios will readily appreciate. Even when dealing with the simple monohybrid ratio of a morphological character the progeny of a single pair often deviates widely from the expected ratio even when fairly large numbers are secured. The average number of adult offspring per mother in Pearl’s experiments with pure Barred Plymouth Rocks is only 2.85, while in my experiments it ranges from 2.5 in 1913 to 6.97 in 1916. The largest family (i.e., offspring of one mother) of pullets I secured, hatched in April and May, is 19. One can only guess at the size of Pear.’s largest family, but it cannot be much greater, for with the sexes evenly divided and with maximum production and perfect hatchability, the maximum number of April- and May-hatched pullets that can be expected in any one year from a single mother can hardly exceed 25. If 10 pullets per mother are available and if their records appear in the ratios of 4:6, 5:5, 3:7, or even 2:8, they would not be im- probable deviations from 1:3, 3:5, or 1:1 ratios, for three birds moved from one side to the other in the extreme case of the 2:8 ratio changes it to equality. Thus, if one encounters a series of ratios which corresponds in general to that of a particular male mated with females of several types, but among which one female occurs that gives a ratio of 2:8, when equality is required, the INHERITANCE OF WINTER EGG PRODUCTION LS bounds of probability are not exceeded in considering the 2:8 ratio a chance deviation from equality. The same sort of reasoning applies if we recognize three sorts of winter layers instead of two. In order to secure families (i.e., progeny of a single pair) of sufficient size, hatched at the proper time, and fulfilling the other conditions necessary, it is necessary to repeat identical matings of a critical nature through a period of years. To do this would require huge physical facilities. It will therefore be difficult to secure suitable data. There is another consideration that affects the application of small observed ratios to theoretical ratios. Any individual is placed in one of two (three, respectively) classes. It is obvious, then, that ratios such as the following 1:0, 0:1, 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2 and so on will occur as a matter of chance.§® It is obvious, furthermore, that only when it is possible to obtain a really adequate number of female progeny, not less than 20 from each pair, will it become possible to determine definitely whether or not either theory has any basis in fact. As Castle (15, 716) has maintained, we may be dealing with a character that is purely continuous in its variation. The ease with which any observed set of ratios in small families can be made to fit at least one theoretical ratio is emphasized by the changes in management as given in an earlier paragraph. Thus, no difficulty is experienced in securing a close fit between expected and theoretical ratios in 1914, although the egg pro- duction of the progeny of the males used that year was entirely different from that of the same males in other years. Of course it is necessary to assign the males to different gametic classes in the different years. Under such circumstances, it is obvious that Mendelian ratios may not express the true mode of inheritance of fecundity. 6 The series for three classes are 1:0:0, 0:1:0, 0:0:1. 1:1:0, 0:1:1, 1:0:1, ene EO ZIOs 1, O22: 1), andisorons 116 H. D. GOODALE AND GRACE MACMULLEN RATIOS AMONG THE PARENTS NEEDED TO GIVE THE OBSERVED MEAN WINTER EGG PRODUCTION OF AMERICAN BREEDS WITH RANDOM MATINGS The mean winter egg production of several American breeds of poultry is about 36 eggs as has been pointed out by Pearl (15b). This average results from random matings. To secure this average requires that the ratios in which the various classes of females appear shall be 12 high, 8 mediocre, and 1 zero, if the average winter production for the high group is about 55 eggs. The high-producing Barred Plymouth Rocks studied by Pearl (’12) averaged 53.08 eggs, the mediocre 15.58, and occur in the observed ratio of 360.5 high to 252.5 mediocre to 30 zero, which is close to that expected. If 36 eggs or thereabouts is the general winter average of American breeds that are properly managed, it follows that the three classes of females should occur in the proportions given above. ‘To produce females in this ratio with random matings requires that the various classes of males also appear in certain definite ratios. We have de- termined one set of ratios (percentage), viz., 14.7 : 15.5 : 25.0: 3.9 : 3.3 :5.3:11.0:19.2 : 0.2, that with females in the ratio of 12 high: 8 mediocre: 1 zero (percentage ratios, 57.1 high, 38.1 mediocre, 4.8 zero) reproduces very nearly the initial ratio among the female offspring. It does not, however, exactly reproduce itself among the males. Whether or not it is theoretically possi- ble to secure a ratio among the males that with random matings will yield females in the proper proportions and reproduce both itself and the proper female ratio in each generation we will leave to those who have the necessary taste and attainments in mathematics. It is evident, however, that such theoretical ratios must exist if either theory has. any basis in fact. CRITICAL VALUE OF MALES OF CLASSES I, II, AND V FOR PEARL’S THEORY There are three classes of males, viz., J, JJ and V, that have a critical value in determining the validity of Pearl’s theory, because the ratios in which their daughters appear is the same whatever the mother may be. Thus, class J males throw all INHERITANCE OF WINTER EGG PRODUCTION 117 high females, class JJ males throw half high and half mediocre, while class V throw only mediocre. Neither Pearl nor myself have many satisfactory data on such males. Pearl notes but one class J male, and this one mated to 10 females had an average progeny per mother of 1.8 daughters. There were four class IJ males. They have a total progeny of 35 daughters and an average of 2.7 daughters per mother. These families are wholly unsatisfactory. No class V male appeared. Several Rhode Island Red males have appeared that throw all or nearly all high producers, but not all kinds of females were mated to them. Two class JJ males are recorded. The number of progeny of one is small while one of the mates of the other gave a ratio of 11 high: 4 mediocre: 0 zero against an expected ratio of 73 high: 73 mediocre: 0 zero. No class V male appeared. It is somewhat difficult to ascertain even roughly the propor- tion in which the males of the various classes should be encoun- tered, though it is clear that class V males are to be expected only rarely. If we assume that the proportions given in a preceding paragraph are approximately correct, then classes I and JJ should appear at least as frequently as VJJ and at least half as frequently as JJJ. They do not do so, however (table 9). It appears probable that the validity or non-validity of Pearl’s theory could be demonstrated beyond doubt by first obtaining males supposed to belong to classes J, JJ, and V and then breed- TABLE 9 Observed numbers of males in each class as determined by Pearl’s theory CLASS OF MALES FLOCK pe eee Barred Plymouth Rocks (earlelGi 2) is. js.cek Sa 1 + 2 4 0 0 Lisle 0 Rhode Island Reds, Mass. Agricultural Experiment SUEDE NS es ae ee ean meee TF Ne hoes 5 0 i> | ile 5 1 * This includes males inadequately tested because mated to high producing females only. 118 H. D. GOODALE AND GRACE MACMULLEN ing them extensively to the several sorts of females, to see whether or not large families could be obtained having the re- quired ratios. SUMMARY OF CRITICISMS OF BOTH THEORIES In several places we have indicated difficulties encountered in applying Pearl’s theory of the inheritance of egg production to our data and which affect the validity of that theory. Sufficient data, however, are not at hand to wholly disprove the theory, though the Cornish male by Rhode Island Red female cross demonstrates that the theory is not of universal applicability. These difficulties mentioned, however, render necessary a state of suspended judgment even in rspect to the applicability of Pearl’s theory to his own data. These difficulties may be summarized here: 1. The extremely small size of the individual families and the various consequences that, flow therefrom. 2. Results of the Cornish, Rhode Island Red cross. 3. Occurrence of high producers where none are expected. 4. The too great ease with which abnormal production may be made to fit theoretical ratios. 5. The lack of class J, JJ, and V males mated to all sorts of females with families of adequate size. ; 6. The facility with which the several division points between high and mediocre producers may be employed. Difficulties 1, 4, and 6 apply also to the alternative theory. CONCLUSIONS The conclusions to be drawn may be stated as follows: 1. The mode of inheritance of winter egg production remains to be determined. 2. The validity of Pearl’s theory can be settled only by breed- ing operations conducted on a large scale, with disease and the necessity of practical considerations eliminated. It should be borne in mind, however, that as Pearl’s mediocre producers are birds that lay at a slow rate, irregularly and spasmodically, INHERITANCE OF WINTER EGG PRODUCTION 119 while mine are mainly birds of late maturity, that the numerical results obtained may be wholly valid for his strain. Moreover, in the Barred Plymouth Rocks the winter cycle of production is a characteristic feature of nearly all records. It, therefore, remains entirely possible that Pearl’s theory is fully applicable to his particular strain of Barred Plymouth Rocks, in which case it should be stated as a theory of the inheritance of rate of production during the winter cycle. Looked at from this angle, then, it is apparent that Pearl’s theory may be a complete explanation of this phase of the inheritance of egg production. Although our numerical data are similar to those of Pearl, we believe that the methods used in this paper are wholly inade- quate for the solution of the problems of inheritance of egg pro- duction. On the contrary, the problems should be approached from an entirely different angle, namely, that of the inheritance of the several factors, whose combined action results in a given number of eggs for the winter period (compare Goodale, ’18). SUMMARY 1. The data furnished by the flocks of Rhode Island Reds at this station furnish ratios that agree with those expected on Pearl’s theory of the inheritances of fecundity. 2. A division point at 40, 50, or even 60 eggs gives ratios that agree well with the observed ratios. 3. An alternative theory that recognizes two classes of winter egg production, viz., high and mediocre, and assumes two genes, both of which must be present in the female zygote in order to have high production, and which are inherited according to the Mendelian dihybrid scheme, will also account for the observed ratios. 4. Reasons are presented, showing that the applicability of these schemes may depend upon the small average number of offspring produced by each pair of parents. Methods by which the falsity or truth of these schemes may be established are given. 120 H. D. GOODALE AND GRACE MACMULLEN LITERATURE CITED CastrLE, W. E. 1915 Some experiments in mass selection. Am. Nat., vol. 49. 1916 Can selection cause genetic change? Am. Nat., vol. 50. DrypdEN, J. 1916 Poultry breeding and management. Springfield. GoopaLte, H. D. 1918 Internal factors influencing egg production in the Rhode Island Red breed of domestic fowl. Am. Nat., vol. 52. Peart, R. 1912 The mode of inheritance of fecundity in the domestic fowl. Jour. Exp. Zool., vol. 13; also Maine Agri. Exp. Sta. Bul. 205. 1915a Seventeen years’ selection of a character showing sex-linked Mendelian inheritance. Am. Nat., vol. 49. 1915b Mendelian inheritance of fecundity in the domestic fowl and average flock production. Am. Nat., vol. 49. INHERITANCE OF WINTER EGG PRODUCTION JeAlL APPENDIX 1 Changes in management At the outset of these experiments, it was determined that methods of management should remain constant; but, unfor- tunately, serious difficulties in rearing the young stock appeared and radical changes were necessary. As later events demon- strated, the trouble was disease, and not methods of manage- ment in the narrow sense. We were much more fortunate in the methods selected for handling the adult stock and in the selec- tion of our methods of incubation. It is the purpose of this section to describe briefly such changes as have been necessary in order that the reader may be able to understand their bearing on the results of the breeding work. Adults. Feeding. Rations and methods of feeding have remained constant, except in the sort of green food fed. When- ever possible, green cabbage has been fed during the winter, but on a few occasions it has been necessary to substitute mangels. At other seasons any green feed available has been fed. Housing. Large open-front houses have been used, except in 1913-14 when about half the pullets were placed in six small pens (accommodating twenty-five to thirty birds each) of a long open-front laying house. The large houses are of two types, but very similar. It has never been possible to observe any difference in production attributable to the differences in housing. Numbers in flock. One type of the large house was built for 72 birds to a pen, the other for 100. The partitions are solid so that each pen is virtually a separate house. During recent years, in order to accommodate the birds, it has been necessary to place more than the theoretical number of birds in each pen, as high as a 70 per cent increase having been made in one in- stance. This theoretical overcrowding, if it has any effect at all, should result in decreased egg production, and therefore in a direction opposite the observed results of the experiments. Time of placing birds in the laying houses. At first the pullets were placed in the laying houses late in October or in November, according to age. No eggs, however, were laid on the range. 122 H. D. GOODALE AND GRACE MACMULLEN As the pullets have matured earlier and earlier each year it has become imperative to get them into the laying houses much earlier, if possible before laying commences. The earliest hatched pullets now go in early in September, and the later hatched, by the first week in October. To make room for the pullets, the birds of the preceding generation are moved into outside roosting sheds sometime during the summer. Chick rearing. In 1913 and 1914, the chicks were brooded in the long-pipe brooder house and grown on the College range. The mortality,” however, was so great that in 1915 resort to the small hovers was made. This season was devoted to the estab- lishment of a satisfactory method of brooding, which, with one or two final adjustments, has been kept as uniform as possible since then. Chick rations. During 1916 and 1917 the rations were con- stant. In 1915, they varied from flock to flock. Those em- ployed in 1914 and 1913, though unlike each other and unlike any of the other years, were constant throughout each season. However, all these various rations were fully adequate to pro- mote rapid growth and eannot have had any effect on subse- quent egg production. Disease. Adults. One of the most difficult problems to con- tend with in the management of poultry is the appearance of disease. It is a formidable difficulty in the way of securing con- sistent results year after year and is unquestionably the most important factor in preventing one from keeping poultry under uniform environmental conditions, even more important than the weather. During the course of these investigations, a method of eliminating infectious disease has been developed. The records for 1916 to 1917 and the winter of 1917 to 1918 are virtually free from the influence of infectious disease. Chicks. a. White diarrhea. Although bacteriological exam- inations were made on dead chicks for this disease in 1913 and 1914 by the Department of Veterinary Science, it was not dis- 7 This mortality was due mainly to white diarrhea, and not to the brooder house, as later event proved. INHERITANCE OF WINTER EGG PRODUCTION 123 covered until 1915. Later, when the agglutination test was applied to all breeders, reactors belonging to the original flock were found. Thus most of the mortality during the first three weeks after hatching in 1913, 1914, and 1915 must have been due to this disease, as no difficulty is experienced at present in rearing chicks. During and since 1915, the experimental flocks have been free from this trouble. b. Filth diseases. There is a group of diseases about which very little is known, but which we infer are spread by filth. It has been the common practice for years to clean and disinfect brooders or colony houses, place baby chicks therein, take pains to clean the brooders frequently while the chicks are growing, but no attention is paid to the ground over which the chicks run, which may be the adjoining hen yard, nor does the attend- ant make the slightest effort to avoid the transfer of filth from the adult birds, especially by means of his feet. If, however, adequate measures are taken to prevent such contagion, trouble from this source disappears. In addition to bacillary white diarrhea, trouble from this class of diseases was experienced in 1913 and 1914, and in that part of the flock of 1915 which was not isolated. In 1915, most of the chicks were reared under isolation. In 1916 and 1917, all the chicks were reared on clean ground by a special attendant. APPENDIX 2 Remarks to practical poultrymen Theories of egg production are of little interest to the prac- tical man unless they can be turned into actual eggs. The poultryman has been told first one thing and then another about breeding for more eggs until the time is approaching when he will believe that it is all humbug. The situation is, indeed, con- fusing. This is because the various theories that have been and will be developed are necessarily attempts at reducing the available facts to order. They should be regarded as reports of progress and subject to revision. ‘Theories, however, are ex- tremely useful, because they stimulate further investigation 124 H. D. GOODALE AND GRACE MACMULLEN and add new facts and new points of view. Pearl’s theory has been very valuable, because it has drawn attention to the im- portance of the male’s influence on the egg production of his daughters. That influence is demonstrated, even if it is not all important as it seemed at first. Fortunately, in spite of the confusion of theories, there is a solid basis of fact on which the poultryman may proceed who wishes to improve his egg production through breeding. Pear] and Dryden have shown that flock egg production is increased by suitable methods of selecting and testing the breeders. We, too, find that selection gives results. The method of selection, moreover, is so simple that anyone can use it who is willing to use trap nests and keep the necessary records. The essentials of the method are these, assuming that the start is made with unpedigreed stock. First, the stock must be per- fectly sound and vigorous. Second, select breeders that equal or exceed some definite number of eggs and mate them to the strongest males available. Third, put only strong (but not necessarily large) healthy pullets in the laying houses. Fourth, when the daughters’ records are available, group them together by mothers. Fifth, pick out the best and second best families. Select the breeding males from these families, taking care to pick the strongest and most vigorous individuals. Mate these to the best pullets of the best families. Proceed in this way testing each year the results of each mating by the records made by the progeny. One should not depend wholly on the numerical record in selecting the breeders. Look carefully to the age at which each pullet commences to lay. Note whether or not she lays con- tinuously or takes vacations. Note whether she lays rapidly or slowly, or whether she lays late into the fall, and the number of times she becomes broody. On the commercial plant it would probably be unwise to trapnest more than 200 birds annually. The best 10 or 15 per cent of these should be used to produce the next flock to be trapnested, and the remainder used to propagate the general flock. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 31 THE FATE OF HOMOZYGOUS YELLOW MICE WILLIAM B. KIRKHAM Osborn Zoological Laboratory, Yale University TWO FIGURES INTRODUCTION The failure of breeders and investigators to obtain any homozy- gous yellow mice and also the observed smaller average litters born to yellow parents (Cuenot, 05; Castle and Little, 710) in- dicated the advisability of attacking this problem from the embryological side, a conclusion justified by the satisfactory outcome shown below. 600. appears, the lumen of the uterus is closed, and all this goes on even though the blastula itself is being rapidly dissolved. This is in marked contrast to the fate, mentioned above, of abnor- mal blastulas in healthy non-yellow mice, which perish without effecting the uterine reaction. However, the same phenomenon of a normal, fully prepared implantation site with no trace of embryonic cells likewise occurs in about 50 per cent of the preg- nant white females examined, which had either produced still- THE FATE OF HOMOZYGOUS YELLOW MICE eit: births or had eaten the young at the time of the preceding par- turition. An intensive study of such examples from both yel- low and white mice has failed to reveal any differences between Fig. 2. Abnormal blastula from the same set as the one shown in figure 1. This individual developed further than the one illustrated above, being larger, as well as more normal in form, and progressed further toward implantation. Phagocytes have overwhelmed one pole, and all the cells of the blastula are under- going cytolysis. X 240 them. In both kinds the uterine lumen remains closed until the fifteenth day of gestation, although the destroyed epithelium is regenerated a day or two before this. No ‘wandering’ cells 32 WILLIAM B. KIRKHAM appear in these specimens, which offers some additional support to the statement of Asai (’14) that this type of cell is embryonic in origin. The similarity in the histological details of the absorbed em- bryos in white and in yellow mice might be taken to prove the identity of the underlying causes in the two cases, but the as- sociated facts tend to modify any such view. The degenerate embryos from white females were obtained, with possibly one exception, from animals which were more or less pathological, and the one possible exception is quite likely not such, as the preceding litter in that instance was removed at birth and other- wise might not have survived for long. Thus in white mice the absorbed embryos might all be accounted for on the basis of pathologic uterine environment which selectively disposed of the weaker members of the sets of blastulas, the set mates sur- viving. The same factor may be present in all yellow females, but in these animals, instead of the unfavorable environment being abnormal, we should have to assume, on account of the universality of the phenomenon, that it is actually a normal correlation with yellow coat color, an assumption further sup- ported by the marked tendency in yellow mice of both sexes to fatness and sterility at a relatively early age. This matter will be subjected to further investigation through a projected study of the non-yellow offspring from yellow matings which, if the above assumption is correct, should be differentiated in general vitality from control animals, offspring from non-yellow parents. Parental abnormality, however, cannot account for all the facts connected with the failure of homozygous yellow mice to be born, and we must further assume an inherent weakness, or lethal factor, in all the homozygous yellow embryos which invariably brings about their destruction during implantation, while their fellows in the same ovulation and environment, but endowed with factors associated with a different coat color, im- plant and develop normally. The assumption is warranted that the degenerate embryos found regularly in pregnant yellow mice mated with males of the same coat color are the missing homozygous animals, for their THE FATE OF HOMOZYGOUS YELLOW MICE 133 occurrence agrees numerically with the requirements of such a case, but a definite proof that such an assumption is valid must await future experimental inquiries. If it is possible at some future time to transplant the ovaries from a yellow female to a mouse of another coat color, it is conceivable that a subsequent mating with a yellow male might, under the assumed more favorable uterine conditions thus obtained, produce viable homo- zygous yellows. The statistical evidence from this research is presented in table 2, where embryos less than three days old are omitted, owing to the failure of abnormalities to become evident before the morula stage. TABLE 2 Showing percentage of degenerate embryos in yellow as compared to white mice YELLOW MICE WHITE MICE Total sets of embryos from healthy non-suckling females 3 to 20 days pregnant...:....0....... 21 26 Total mormal embryos.) |... 2. ./!00se. Se 94 189 Total degenerate embryos....................5. 39 2 Percentage of degenerate embryos.............. 29-+- 1.0+ The data show clearly that the degenerate embryos occurring quite regularly in healthy yellow females mated with males of the same coat color must be considered as of a quite different nature from those occasionally found in white mice, for the latter are found almost only, if not invariably, in unhealthy females. The evidence thus strongly indicates that the former are the missing homozygous animals. The proportion of degenerate embryos in normal yellow females is, including additional material obtained since the publication of the preliminary report, 29+ per cent, which is but little higher than the Mendelian expecta- tion of 25 per cent, and is quite within the limits of probability when the total amount of the material is relatively small. 134 WILLIAM B. KIRKHAM CONCLUSION Three points stand out in a survey of the results attained in this investigation. First, all mouse embryos encounter a crisis at the time of im- F Whar JV plantation of the blastula in the wall of the uterus, and in un- usually large sets of blastulas one or more appear always to perish at this time without producing any uterine reaction. In healthy mice other than yellows, however, those blastulas which induce a swelling of the mucosa uniformly complete their im- plantation, while the blastulas resulting from yellow matings al- most always lose at least one of each set after the mucosa has reacted. Second, apart from this loss of certain blastulas during im- plantation, the embryonic and early postnatal history of yel- low mice is exactly the same as that of mice of other coat colors. Third, the evidence that the blastulas lost in yellow females during implantation are the missing homozygous yellow mice consists on the one hand of the absence of any like phenomenon in healthy white mice, and on the other hand of the statistical correspondence of the percentage of embryos so lost with the Mendelian expectation of homozygous yellows. THE FATE OF HOMOZYGOUS YELLOW MICE 135 LITERATURE CITED Apter, L. 1912. Versuche mit ‘‘Mammimum Poehl’’ betreffend die Function der Brustdriise als innerlich sezernierende Organ. Miinchner Med. Woch., 59. Asat, T. 1914 Zur Entwickelung und Histophysiologie des Dottersackes der Nager mit Entypie des Keimfeldes. Anat. Hefte, Ist Abt., Bd. 51. Caste, W. E., anp C. C. Lirrtz 1910 On a modified Mendelian ratio among yellow mice. Science, N.S., vol. 32. CuHaruLtTon, H. H. 1917 The fate of the unfertilized egg in the white mouse. Biol. Bull., vol. 33. CurEnot, L. 1905 Les races pures et leur combinaisons chez les souris. Arch. Zool. Expér. et Génér., 4° Series, T. 8. 9. FRAENKEL, L., anp F. Coun 1901 Experimentelle Untersuchungen iiber den Einfluss des Corpus luteum auf die Insertion des Eies. Anat. Anz., Bd. 20. IsBsEN, H. L., anp E. SrercLepER 1917 Evidence for the death in utero of the homozygous yellow mouse. Amer. Nat., vol. 51. Kirxuam, W. B. 1916 The prolonged gestation period in suckling mice. Anat. Rec., vol. 11. 1917 Embryology of the yellow mouse, Anat. Rec., vol. 11. Lors, L. 1908 The production of deciduomata and the relation between the ovaries and the formation of the decidua. Jour. Amer. Med. Assn., vol. 50. MarsuatL, F. H. A., anp W. A. Jotty 1907 Results of removal and trans- plantation of ovaries. Trans. Roy. Soc. Edin., vol. 45. Meyer, A. W. 1917 Intra-uterine absorption of ova. Anat. Rec., vol. 12. Soporta, J. 1911 Die Entwicklung des Eies der Maus vom ersten Auftreten des Mesoderms an bis zur Anbildung der Embryonalanlage und dem Auftreten der Allantois. Arch. f. Mikr. Anat., Bd. 78. Resumido por el autor, Carl R. Moore. Sobre las propiedades fisiolégicas de las gonadas como reguladores de los caracteres somaticos y psiquicos. I. La rata. Un estudio de las modificaciones que siguen a la gonadectomia en las ratas jévenes y el transplante ulterior de la gonada opuesta en cada animal (repeticién de los experimentos de Steinach), demuestra que los efectos aparentes de tal experimento no son tan marcados como podria esperarse después de leer las comunica- ciones de Steinach. Es muy dificil demostrar la presencia de cambios somaticos debidos en absoluto a la presencia de la glan- dula transplantada. Bajo el punto de vista psiquico los resulta- dos son mas definidos. Los machos jévenes han sido transfor- mados aparentemente en hembras tan tfpicas que el instinto maternal de protejer y criar a los pequefios puede notarse in- mediatamente. Las hembras jévenes han sido también trans- formadas en machos hasta tal punto que se conducen psiquica- mente como tales y los imitan, de una manera muy exacta, en el acto de la cépula. Los cortes histolégicos de los ingertos demues- tran que el ovario ha persistido aparentemente funcional, mien- tras que el testiculo ha sufrido cambios marcados que conducen a la destruccién de los espermatocitos y espermatozoides. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 31 ON THE PHYSIOLOGICAL PROPERTIES OF THE GONADS AS CONTROLLERS OF SOMATIC AND PSYCHICAL CHARACTERISTICS 1. THE RAT CARL R. MOORE Hull Zoological Laboratories, University of Chicago FIVE FIGURES In several papers during the past few years E. Steinach has described many changes occurring during the development of mammals, both of a somatic and of a psychical nature, which he affirms are due to a secretion of the sex gland.!_ He reports that an ovarian graft in a completely castrated young male rat or guinea-pig will so modify the subsequent development that the animal becomes somatically and psychically a female. These animals are proportionately lighter in weight, shorter in body length, hair smooth and fine, and pelvis smaller than that of males of the same age. Also these ‘feminized males’ react more like females than males (more docile, absence of male instincts toward female rats, reactions toward young characteristic of that of a mother). In guinea-pigs growth of the mammary glands and milk secretion was also reported. In the reverse experiment, i.e., if pieces of testicular tissue are grafted into completely spayed young females, the animals become mascu- linized as maturity is reached; they resemble males instead of females both somatically and psychically. Steinach supposes that a secretion from the interstitial cells of the grafted testis and ovary in each case is the controlling factor since the secondary characteristics of the opposite sex do not appear unless the implanted gonad obtains vascular con- 1 See Steinach, 710, ’11, 712, 713. 137 138 CARL R. MOORE nections and remains in a living condition after the transplan- tation. It is thought the modification is affected by a hormone produced by the interstitial cells, that its action is a chemical one, and that it may sensitize the nervous system to react in a new capacity. These results have been criticised severely by many investi- gators, and to the writer’s knowledge the observations have not been corroborated in other laboratories. In regard to criti- cisms the following annotation from Lancet (vol. 193, no. 18 of ii 1917, p. 687) is of interest: It is a drawback to the experimental method, as practised on lines of Baconian induction, that anyone may make a few random experi- ments and with the results lay some sort of claim to general attention. Hence we should preserve a carefully critical attitude towards claims to medical discovery until some circumstance evinces the likelihood of some truth in them. Lately (Zeitschrift fiir Sexualwissenschaft, Aug., 1917) the physiological work of Steinach, Foges, and Lode has come in for repeated discussion. Steinach described having changed the sexual disposition of small mammals by implanting, as the case might be, an ovary into a young male or testicular substance into a young female. When the necessary operations were successful the treated animals in their behavior showed reversal of the natural con- ditions, males attempting to mate with males and females with females. But (a very important point) such was the case only if the young animal so treated had been first deprived of its own primary repro- ductive gland—i.e., if it had first been castrated or spayed—other- wise the implantation had no feminizing or masculinizing effect. It was, in short, as though a clear field was necessary for the exogenous influence. Around these findings the theory has been constructed that the products of testicular and ovarian secretion—that is the specific reproductive hormone of the two sexes—are sharply antago- nistic the one to the other. Their effect on the brain, from which the sexual impulse proceeds, is described as an ‘erotising’ one, in the direction of masculinity or femininity respectively. The mode of action is supposed to be bio-chemical. The conclusions want more evidence to back them.? It was during the process of some studies on sexual modifica- tion in rats and guinea-pigs that Prof. F. R. Lillie suggested the desirability of repeating Steinach’s experiments. To Doctor 2 The article referred to has not yet been obtained by the writer; possibly it has not reached this country. GONADS AS CONTROLLERS OF CHARACTERISTICS 139 Lillie I am greatly indebted for many suggestions, abundant material, and a constant interest in the problem. The results agree, in some respects, with those of Steinach, but are not so far reaching as might be expected. This paper contains an account of the observations made on the white rat; a report of similar experiments on guinea-pigs will follow later, as the obser- vations are not yet complete. MATERIAL EMPLOYED The white rat (Mus norvegicus) was used for the experiments and the operative procedure described. by Steinach has been repeated almost exactly. It appears, however, from differences to be noted later that Steinach must have employed a slightly different strain of rats than those used in these experiments. Animals of the same age and almost invariably of the same litter were selected for the cross transplantations. The method was as follows: A brother and a sister rat were etherized at the same time, and after operative conditions were observed the peritoneal cavity of each was opened.? The ovary, after removal and sometimes accompanied by small pieces of oviduct, was cut in half with scissors to aid in the establishment of a vascular con- nection. A piece was placed on either side of the midventral line of the male between fascia and the abdominal musculature or imbedded more deeply in the substance of the muscle. The genital cord of the male was severed above the epididymis and both testes removed. Pieces of these were similarily placed in the body of the females and the muscle layer with peritoneum, and the skin were sutured separately. Usually a slight injury of the fascial layer of the external oblique muscle as well as the eorium of the skin was made with the point of a knife to aid in the establishment of a vascular connection. Aside from the homoplastic transplantations with complete removal of the normal gonad, some of the animals were merely castrated and spayed without subsequent transplantation; also, ‘3 The abdomen was shaved, treated with Lugol’s iodine and alcohol, instru- ments sterilized in carbolic acid solution, and the table covering, towels, gown, ete., sterilized. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 2 140 CARL R. MOORE several cases of homoplastic transplantation were made in which the normal gonad remained undisturbed. Parts of fifteen to eighteen litters were used in the experiments, comprising over fifty operated animals. All cases of grafting were not suc- cessful, and a few animals died before their mature cond:tion could be studied, but the successful cases were sufficiently dis- tributed to afford a wide range of conditions for study. It has seemed advisable to give in detail the history of one litter and allow this to serve as an example of the general conditions in such experiments. This litter fulfills very adequately the con- ditions sought from the fact that the transplanted gonad ob- tained vascular connections and persisted in each operated animal. Furthermore, there is need for constant comparisons between operated animals and the normal ones under similar conditions. The one litter eliminates such ordinary differences as age, number of rats in litter, age of mother, hardiness, eic., and since the conditions are the same separate histories will not have to be repeated. The litter (no. 6, AIBI) was born September 22, 1917, and consisted of four males and six females. Operations were carried out at the age of 25, 28, and 35 days, and all were cases of homo- plastic transplantation with complete removal of the normal gonad, i.e., ovaries were removed and pieces of testis placed sub- cutaneously and vice versa. Of the ten rats in the litter trans- plantations were made on three males and four females; unfortu- nately, however, one female escaped from the cage within three days after operating, which left three castrated males containing transplanted ovaries and three spayed females containing trans- planted testis, one normal male and two normal females as controls. On November 12, 1917 (51 days old) these were so marked that each one could be recognized and each was then given a number making possible a complete and separate record of each individual rat during the course of the experiment; the experiment was terminated June 3, 1918, at the age of 254 days. At that time the rats were killed and the grafts preserved for cytological study. Since the different rats will be treated in- dividually, the following numbers will serve to identify them: GONADS AS CONTROLLERS OF CHARACTERISTICS 141 Litter 6 AIBI I. Spayed female with testis graft. II. Spayed female with testis graft. III. Normal female (control) IV. Castrated male with ovary graft. V. Castrated male with ovary graft. VI. Spayed female with testis graft. VII. Normal male (control) VIII. Castrated male with ovary graft. IX. Normal female (control). BODY WEIGHT AND BODY LENGTH It is unfortunate that the distinguishing somatic characters of the male and the female rat are not more sharply marked than they are. However, the studies of Stotsenburg (’09, 712, ’13), King, and others! show that within certain limits the growth curves (body weight) and the body lengths normally afford criteria for a distinction between the two sexes. Steinach hsa placed considerable emphasis upon these weight and body-length relations of his feminized males and masculinized females as being indicative of maleness and femaleness. It is the opinion of the writer, however, that such slight differences in weight are but poor criteria of maleness or femaleness under abnormal conditions. It is true that the normal curve of growth for male rats is considerably above that for females, but it is decidedly unreliable to choose two or three rats at random and classify them sexually on the sole basis of weight; there is too much tendency for variation even among rats of the same litter for it to be reliable. Also a slight pathological difference may pro- duce a relatively great change in weight even though the rat is in apparently good health. Too many operated animals are required to make the factor a convincing one. But of more importance than this, perhaps, is the fact made known by Stot- senburg that early castration of male rats does not influence the subsequent growth curve, while early spaying of female rats re- sulted in an increased growth curve over that of the normal female * Complete references of work done on rats have been compiled by Donald- son in a book ‘The Rat’ (Memoirs of The Wistar Institute of Anat. and Biol., no. 6, 1915). 142 CARL R. MOORE of 17 per cent, 24 per cent, and 30 per cent in three lots observed. The tendency of growth in the total absence of gonads would then be to equalize the weight of the two sexes, but whether this condition would actually be realized is uncertain. At any rate a spayed female with grafted testis would increase in weight above the normal for females not because of the testis, but because of the absence of the ovary. These factors alone would tend to make the weight of an animal a very unsatisfactory test of maleness or femaleness. TABLE 1 BODY 66 80 108 138 165 192 261 | LENGTH, DAYS | DAYS| DAYS] DAYS] DAYS | DAYS] DAYS AGE 261 DAYS ANIMAL grams| grams| grams| grams| grams| grams | grams cm, I | Female with testis....... 75 | 92 | 119 | 187 | 149 | 143 | 165 | 18.0 II | Female with testis....... G58 955 SAS e Sel eT OM etd aeLO Gelmeidco PEI) Pemale (mormal).:.2...:. 95 | 99 | 127 | 147 | 125 | 182 | 145 | 18.0 IV | Male with ovary......... 80 | 99 | 187 | 158 | 173 | 164) 21 1 V | Male with ovary......... 97 | 106 | 183 |} 150 | 159 | 160 | 180 | 18.5 VI | Female with testis....... 88 | 92) |) 124 a 5O aL 77 heb a Sub VII | Male (normal)...........| 102 | 109 | 139 | 166 | 180 | 190 | 235 | 20.5 VIII | Male with ovary......... 89 | 104 | 138 | 160 | 168 | 166 | 198 | 19.25 IX | Female (normal)......... 100 | 99 | 125 | 195 | 1320) 51605) G7 oe ab 1 No. IV, killed at age of 238 days. Many factors also enter in that tend to discount the apparent specificity of length as a determinant of sex. Stotsenburg’s early spayed females were found to increase in body length over that of the normal females. Here again, if these animals had possessed transplanted testis their increase in length could not be considered as a result of the secretion of the testis, but rather of absence of the secretion of the ovary. Nevertheless, to ob- tain whatever evidence possible a careful series of weights has been kept on the litter in question and is given in table 1. At the termination of the experiment the total body length of each was recorded.®* > The length of the body and tail, often employed, could not be accurately measured, as a small piece of the tail had been removed from some of them as a distinguishing mark for those individuals. GONADS AS CONTROLLERS OF CHARACTERISTICS 143 Thus it will be realized that these physical factors (weight and length) are very poor barometers of the conditions here repre- sented, that of an intersex condition. If, for instance, rats V and VIII (males containing growing ovaries) be compared with VII (a normal male), those containing the ovaries are both lighter in weight and shorter in body length than the normal male, but do not fall to the level even of the heaviest female (IX). We may perhaps, with all justice, refer this decrease to the presence of the ovary. Also if the females containing testis tissue (J, II, and VI) be compared to the mean weight of the two normal fe- males (III and IX) there appears, only in case of rat VI, an in- crease in weight which by no means approaches the weight of the normal male; and had only female LX been used as control there would be an actual decrease in comparison of I and IX. The former (I) having had the ovary removed should have been heavier than the latter (IX) which had ovaries present. The question also arises whether we should refer these weight modi- fications to a variation in the intensity of the sexual condition, or whether they may not be merely the result of disturbances in the regulators of metabolism which we know may produce vari- ations. It may be possible that the elimination of some secre- tion of other glands may affect the final result as well. If this were true, surely we could not consider this secretion as a factor in determining the sexual condition of the animal. Weight and length are then very unsatisfactory criteria for determining the changes associated with cross transplantation of gonads. HAIR, MAMMARY GLANDS, SKELETAL CHANGES, FAT DEPOSIT Steinach has used a few other criteria as tests for the result of the sex hormone (Pubertitsdriise) in its powers of modification, ' but the writer also finds it impossible to consider these as valid support for the hypothesis. ® Some experiments under way at the present time indicate a potential weight difference in the two sexes that appears to be independent of the gonad. Even though these experiments are not yet complete, the indication is that early spayed female rats, though they increase in weight over that of the normal females, do not reach to the height of the growth curve of the castrated males. 144 CARL R. MOORE It is possible that the differences of the male and female hair coats of Steinach’s rats were more pronounced than in the strain used in these experiments.’ It is true that a slight difference can be noted in normal healthy white rats of the same age. The male hair coat appears slightly rougher, the hair being a little more coarse than that of the female; this in a general way gives a softer, smoother appearance to the female than to the male. But this also is subject to so many variations that it is decidely unsafe to use it as an indicator. The variations at different ages are considerable, and a slight metabolic disturbance also gives entirely different appearances to the hair. Numerous instances have been noted in which the female coat was rougher in appear- ance than that of the male. Indeed, the writer has often found it entirely impossible to choose the males and females from a cage of normal and apparently healthy mixed rats by this means alone. This being true, it would be entirely impossible to note the changes in an intersex condition and to place properly these changes as quantitative determiners of a modified sex condition. If one were a decided advocate of the idea, it would be a simple matter to record differences that would support the hypothesis. It is possible, however, that Steinach’s material showed greater differences than the rats used for these experiments. In relation to mammary glands Steinach has already pointed out the fact that rats offer very poor material for study of their changes. The primordial teat is not produced in the male so that little influence from the implanted gonad upon the primor- dial mammary gland can be seen.® Steinach (’12) has reproduced radiograms of feminized male rats to show the difference in size of the pelvis between these and normal or merely castrated male rats. These radiograms show very clearly the comparatively small size of the pelvis in femin- ized rats, but they also show, to a like degree, the reduction of 7 Steinach seems to have used partly wild rats, partly tame (white) ones, and crosses between these. 5 Guinea-pigs afford much better material for study of possible changes in the mammary gland due to internal secretion of gonads; experiments are now under way on these animals and will be reported at a later date. GONADS AS CONTROLLERS OF CHARACTERISTICS 145 all the other bones in the body. No x-ray examinations of the pelvis of the modified rats have been made by the writer, but it appears that probably these characteristics, also, are not specific nor distinctive. In a rat of smaller size one would naturally suppose that the pelvis would be smaller as well as all other bones of the body. And it would seem probable that the condition of intersex, as one encounters it in these cases, would present the same difficulties for discrimination as would weight, length, hair coat, ete. The fat deposit featured by Steinach is a poor indicator of sexual conditions. It is generally true that the tendency for fat accumulation in the normal female rat is more pronounced than in the male. For this to be constant even in the normal condition presupposes a continued, uniform metabolic condi- tion. The question of intergradations in sex again arises as well as the difficulty of recognizing the quantitative amounts of the fat deposit. ‘To illustrate from this litter: Rats I and Vi showed a greater amount of fat deposit than did rat no. V° but the two former rats possessed implanted testis,!° while the latter possessed the implanted ovary, and the fat deposition should have been reversed. Rat VIII, on the other hand, possessed more fat de- posit than either I or VI, which should be the case if only the implanted glands were to be considered. This affords us little evidence for or against the assumption of a modification follow- ing implantation of the corresponding sex gland. OBSERVATION ON BEHAVIOR The behavior of these rats has given more evidence to support the idea that the sex gland regulates the characteristics of the animal than any other set of characters which has been observed. A. Feminized males These behavior observations were carried out both while the animals observed were in the cage with other members of this * The amount of fat was not actually determined quantitatively, but merely noted from macroscopic observations. 10 For conditions of these grafts, see section on microscopic observations. 146 CARL R. MOORE litter and while separated for observation. It is beyond ques- tion that the early castrated male rats which have received im- planted ovaries display a maternal behavior towards the young. The two normal female controls (III and IX) gave birth to lit- ters of young during the course of the experiment and often the mother with young was allowed to remain in the common cage. It was repeatedly observed that the feminized male rats would enter the nest with the mother or without her, would nestle the young and repeat exactly the behavior of the mother when the young attempt to suckle. If the litter is a large one and the young from seven to ten days old, the mother will assume a peculiar position to enable the young to suckle; the abdomen is arched and both the fore and hind-legs are widely separated as the young wriggle around underneath in search of teats. The reaction is quite characteristic. This reaction was displayed absolutely typically by the feminized males. The normal male rat and the masculinized females are seldom if ever, seen on or near the nest,!! and apparently they take no interest in the young. The following observations from the note-book will illustrate the phenomenon and its frequency: April 30, 1918. Normal female gave birth to litter of six young. May 1. Normal male and masculinized females had been removed from cage leaving mother and three feminized males. Feminized male V showed all apparent reactions of mother—persistently occupied nest ‘of young with mother and apparently young were attempting to suckle, no teats developed, however, and young could not suckle. Fem. male licked young, tucking same under him, when attempt to remove him from young would attack and bite. Evident mother instincts, would arch abdomen for young to attempt to suckle. Prof. Lillie sees behavior. May 2. Observed several times during day (normal male now in cage), feminized male IV several times on nest with mother—it lies down with young allowing them to search for teats, arching abdomen as they work around in search of teats. Normal female (mother) had to lie across body of fem. male to get to young, three suckling mother, 11 The cage in which the litter was confined was 26 x 18 x 12 inches, made of galvanized wire, sides and top, and a movable bottom. The nest was made from paper torn up by the mother and placed in one corner of the cage. When the mother leaves the nest, especially during the early life of the young, she almost invariably covers it with small loose pieces of paper from the edge. *GONADS AS CONTROLLERS OF CHARACTERISTICS 147 three attempting to suckle fem. male, remained so for forty-five minutes changing position slightly if disturbed but immediately lying down again with young searching around underneath. If young are displaced from nest—mother immediately picks them up in mouth and returns them to nest—fem. male not observed to return them to nest but allows them to remain at the edge where placed. Normal male never seen in nest with mother when latter is suckling. May 6,2 p.m. Fem. male IV found on nest with young, on all fours with abdomen arched and young attempting to suckle. Thought at first was mother, reaction so characteristic, but examination showed mother away from nest. Reaction could not have been different in mother, so characteristic, no question whatever as to appearance’ of same. Young of course could find no teats but were trying very hard. Best and most conclusive reaction yet observed. 3.30 p.m. Feminized males V and VIII removed from cage leaving fem. male IV, normal male and mother. Mother on nest: IV and n/male in corner of cage opposite nest, n/male almost invariably occupies this place and has never been seen to show any interest in young. 4.30 p.m. All three old rats in end opposite nest, young covered. 5.30. All three old rats in end opposite nest, young covered. 6.00. Same. 8.00. Mother and IV on nest, young attempting to suckle both, both removed from nest—n/male not near nest. 8.30 p.m. Mother had covered young, all three old rats away from nest. 9.30 p.m. Fem. male (IV) on nest with all six young underneath abdomen attempting to suckle, abdomen arched, legs spread, all young searching for teats; absolutely normal female reaction. Mother and n/male away from nest, IV removed from nest. May 7, 8 a.m. Mother in nest with three young suckling, IV in opposite corner of cage attempting to suckle two young—nest had been torn up during night and one young one nestling under n/male but he showed no reaction to it and was only asleep. Placed all young in nest with mother. 11.00 am. IV and mother both on nest, mother suckling four young, IV covering two. Norm. male in end opposite nest. IV taken out of nest—five minutes later IV had returned to nest with mother and young, n/male in opposite end of cage. IV again removed from nest, and mother occupied it. 11.45 a.m. Mother and IV on nest with young, one young under IV, n/male in end opposite [IV removed from nest. 2.00 p.m. Mother suckling young, IV and n/male in opposite end of cage. Mother removed from nest. 2.40 p.m. IV onnest with young attempting to suckle, removed from nest. n/male in end opposite. 3.15 p.m. IV on nest with young, n/male in opposite corner of cage, IV removed from nest. 148 CARL R. MOORE 3.30 p.m. Mother on nest, [V and n/male in opposite corner. 4.00 p.m. Feeding. 5.380 p.m. Young covered up in nest—all three old ones away from nest. 8.45 p.m. Same. May 8 a.m. 8.00 a.m. Mother and IV both on nest, IV removed, n/male in opposite corner. 9.30 a.m. Mother and IV on nest with young, n/male in opposite corner. Mother removed from cage, leaving in cage only fem. male (IV) and n/male. IV removed from nest. 10.00 a.m. IV on nest with young, watched reactions for fifteen minutes with Prof. M. M. Wells. IV trying to suckle young—abdo- men greatly arched for young to get under, hind legs spread apart when little ones approach that region of abdomen from beneath, changes position slightly as young search from place to place for teats. Proclaimed by Prof. Wells as non-questionable maternal reaction. Young displaced from nest, IV replaces them—picks up young in mouth—suddenly picking up one at a time carries four to opposite end of cage placing them at side of sleeping n/male whose reactions are wholly passive—young begin to crawl under him but he continues to sleep. IV covers remaining two in nest, remain so for five minutes. Young taken from under n/male and placed in center of cage, IV comes off nest, picks up all and returns them to nest, IV disturbs n/male apparently in search of young intrusted to his care, pushes him out of corner. IV moving about cage—mother returned to cage. May 22. Litter in cage 37—cage contained normal male, normal female, two castrated males, two spayed females. Observed many times daily for twelve days (until young would leave nest) never was normal male, castrated male, or spayed female seen near young which were in end of cage in nest made by mother. Were never seen to give any attention to young in period of twelve days observed. Two points in particular are established by consideration of such a set of observations, especially those made at intervals over a period of two consecutive days: ,these are, 1) that the feminized male (in this case rat no IV) does not merely display a sporadic interest in the young rats, but that it is a continued interest and apparently as characteristic as that of the mother, and, 2) it gives not only a comparison of the feminized male be- havior and that of the mother, but it also brings to attention very forcibly the absolute passive reaction of the normal male. Also the series of observations continuing over twelve days fails GONADS AS CONTROLLERS OF CHARACTERISTICS 149 to reveal the slightest interest displayed by either spayed females or castrated males.!” B. Masculinized females In order to observe better the reactions of the masculinized females, these were placed alone in cages for a day or so before subjecting them to the experiment. The various rats were then put into the cage and the reactions of the masculinized female noted. - Practically any two strange white rats placed together in the same cage immediately show interest in each other, and whether they are two normal males, two females, or male and female they almost at once begin to nose around the external genitals of each other, and many times if they are both males a fight begins. These reactions are general for practically all rats when placed together. But in case the two are a normal male and a female in heat, the act of copulation begins immediately and, though of very short duration, may be repeated a great number of times.2 But in all the writer’s experience among the rats of the colony used (amounting to several hundred) he has never observed a normal female attempt to imitate the male in the act of copulation. One of the most, perhaps the most characteristic feature of the whole process of copulation in regard to the male, is that after each attempt almost invariably the male assumes immediately a position that allows him to lick the copulatory organ before the next repetition. It is a very interesting fact that the masculinized female would attempt to imitate the male in the act. of copulation in an ab- solutely typical set of reactions. Of course, no male copulatory 12 The spayed and castrated rats of cage 37 were at this time over six-months old and had been operated on at the age of about thirty days. These rats nor- mally would have been sexually mature. 13 The males are not especially keen discriminators, for if a female in heat is placed in the cage with several males the excitement is very great and repeatedly results in one male attempting to copulate with another. This attempt is also often made if a male is placed in a cage with a single male, especially so if a female in heat has just been removed from the cage of the male. 150 CARL R. MOORE organ had developed, but despite this fact, each time the at- tempt was made the masculinized female repeated exactly the male behavior by licking the normal position of the male organ. The following extract from my notes represents the type of behavior: May 29. Normal female in heat, placed in cage with masculinized female (rat I). Mr. F. L. Dunn and the writer watched four un- questionable attempts at copulation, each time masculinized female licks region of penis of male, same as normal male reaction, though no penis is present. Not so enthusiastic as normal male, repetition not so frequent. Same normal female in heat placed in cage with an early spayed female on which no transplantation of testis had been made— no attempt at copulation. June 3. Normal female in heat, placed in cage with masculinized female (rat I)—reactions perfect and characteristic of norm. male. Copulation attempted eight or ten consecutive times with only short intervals between attempts, masc. female entirely as enthusiastic as norm. male—each time licks external genitals as does norm. male, before next attempt. Normal female taken from cage for fifteen minutes, returned to cage again. Copulation attempted instantly, just as normal male would have acted. Copulation attempted time and time again, absolutely normal reaction. Spayed female without testis showed none other than passing interest toward female in heat, no attempt whatever to imitate act of copulation. Same true of old female, did not attempt to copulate. There is absolutely no question as to the occurrence of these reactions, for they could not be more characteristic had the fe- male in heat been placed in a cage with anormal male. The nor- mal female reacts also in an absolutely typical manner towards the masculinized female, just as she would have acted had the masculinized female been a normal male and capable of fulfill- ing the act of copulation. And it is significant that, so far as the writer has observed, neither a normal female in the height of vigor, an old female, nor yet an early spayed female that had not received the testicular transplantation has ever been seen to attempt the act of copulation by assuming the rdéle of the normal male rat. | It may be remarked that the only psychical reactions that appear to be of value to the writer are those illustrated above, those of the sexual reactions of one rat to another and the ma- GONADS AS CONTROLLERS OF CHARACTERISTICS 151 ternal behavior. Steinach has described at great length the docility of the normal female rat (does not fight, is easily handled, not so apt to bite or to resist handling, etc.), but here again the variations are too great to be of any practical value. Many females of this colony used are decidely more pugnacious than males. In several cases, these, after repeated handling would bite, scratch, and resemble any other than a meek and mild- tempered female, and at the same time the males show entirely as mild and even-tempered disposition as any female of the colony. As for the sex reactions, it is true that, in cattle especially and in some other animals, the females often attempt to imitate the role of a male; but among rats this tendency has, so far as known, never been observed. As noted above, female rats have been placed in cages with normal females, early spayed females, and old females (each of these having been in isolated cages for some days), but in no case have they attempted to imitate the male. Aside from the general reactions exhibited by any two strange rats when placed together, the reactions except in cases of transplanted testis, have been negative. MICROSCOPICAL OBSERVATIONS The litter from which most of the data for this paper has been taken afforded excellent material for these considerations on account of the fact, mentioned previously, that one or more of the transplanted gonads were retained for the 225 days between the time of operation and the termination of the experiment." The animals were killed with ether and the transplanted glands removed from the place of growth and development; the tissue was killed in Bouin’s fluid, sectioned, and stained with haema- toxylin and eosin. The young ovary successfully transplanted into a young male animal persists and undergoes its differentiation in quite a nor- 144The blood circulation in these transplanted pieces of tissue was estab- lished in different cases, either principally through blood connection with vessels of the muscles or from cutaneous vessels in the superficial fascia. 152 CARL R. MOORE mal manner.® Inasmuch as the ovary at the time of trans- plantation was cut into two pieces, and these probably never of the same size, it was impossible to ascertain the amount of growth in any specific piece. The stroma tissue is quite characteristic of the normal ovary. Both mature and immature Graafian follicles are abundant and some contain ova apparently ready for ovulation. That ovulation really does occur is shown by the presence of a few corpora lutea, some of which still contain a blood clot in the center. Many Graafian follicles indicate an evident tendency for degeneration, and atretic follicles are often - found in the preparations. Reference to figure 1 will convince one that the ovarian cortex, stroma, and contained follicles as well as the medulla of the ovary are retained in apparently a normal condition. This section is a piece of the original young ovary placed subcutaneously in the male at the age of thirty days and allowed to grow until the animal was killed 230 days later. Vascular connections may be established either from cutaneous blood-vessels or in case of a deeper transplantation by vessels supplying the abdominal muscles. This graft was imbedded in the superficial fascia. The section passes through the ovum of two mature follicles and each contains a very dis- tinct nucleus. Figure 2 (from same graft as fig. 1) shows not only the more mature follicles, but a young follicle as well as a corpus luteum. The graft from which these two sections were prepared had persisted throughout the entire period of the ex- periment in almost a normal condition. It was physiologically active, since the germinal epithelium shows an apparently normal condition, young follicles are present also mature follicles evi- dently almost ready for ovulation, and corpora lutea are present, showing that a previous ovulation has taken place. The con- dition of this graft can represent the general condition of the persisting grafts, since no other features worthy of mention have been noted in other grafts. 15Tn several cases of transplantation of both ovaries and testis, more espe- cially testis, the glands failed to persist and underwent resorption, leaving little if any traces of the original implantation. GONADS AS CONTROLLERS OF CHARACTERISTICS 153 The implanted testicular tissue does not persist in the same degree of normality as does the ovary, and in somejcases it has a ten senty yen ALAA Fig. 11° Section of an ovarian graft in superficial fascia of male rat (6, AIBI- V). Rat 30 days old when graft was made; growth removed after persisting for 225 days. bv., blood-vessels; gf., Graafian follicle; ge., germinal epithelium; m., medullary region of ovary; p., peritoneum of female transplanted with ovary; s., stroma; sf., superficial fascia. failed to persist at all. In the majority of cases the tissue was placed subcutaneously in the female near the midabdominal \ Drawings made by Kenji Toda. 154 CARL R. MOORE line. Usually the seminiferous tubules are found scattered about either in the muscle tissue or in the fascia above and are not bound up into a compact mass; figure 3 (a section of a 225- day testicular graft in female 6, AIBI-I) shows the dispersed condition of the tubules imbedded in muscle. In figure 4 (same age graft as in figure 3, but taken from female 6, AIBI-II) part Fig. 2 Section from same ovarian graft as figure 1. bv., blood-vessels, cl., corpus luteum, ge., germinal epithelium, gf., Graafian follicle, s., stroma. of the epididymis as well as the degenerated seminiferous tu- bules is present. These tubules, as Steinach has described, are decidedly different from those found in a normal testis; they contain only an irregular lining of cells of large size (interpreted as Sertoli cells). Spermatozoa and spermatocytes are entirely absent. The lumen of most of the tubules is filled with a retic- ular substance possibly representing products of degenerated GONADS AS CONTROLLERS OF CHARACTERISTICS 155 Fig. 3 Section of testicle graft showing scattered seminiferous tubules in the abdominal muscles of female rat. Graft had persisted for 225 days before its removal from the female (6. AIBI-I). Tissue transplanted from male to female at age of thirty days. m., muscle; st., seminiferous tubules. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 CARL R. MOORE 156 iferous tubules and part of le graft showing semin 1¢ Fig. 4 Section of test epididymis, from muscle and fascia 0 Graft same age f female rat (6, AIBI - IT). 1S. . é., epididym ferous tubules; . .. Semini st as figure 3. GONADS AS CONTROLLERS OF CHARACTERISTICS 157 cells. Figure 5, a more highly magnified part of the graft from female 6, AIBI-II, shows the condition of the cells that remain within the tubules as well as the large number of cells between the tubules. The latter evidently represent the interstitial gland described by Steinach and supposed by him to be the seat of the production of the hormone of the testis. Between the tubules are usually to be found a great number of leucocytes scattered indefinitely throughout the tissue, but the significance of these is not entirely clear. The writer purposely refrains from an elaborate description of the cytological findings, as Steinach has discussed them ade- quately. The figures included will enable one to realize the condition of these pieces of transplanted gonads. DISCUSSION These observations corroborate parts of Steinach’s experi- ments and tend to support very strongly his ideas of the trans- forming power of the gonad of one sex over, at least, the psychic nature of the opposite sex. While the writer is entirely unable to interpret the observations of the possible somatic differences of these evidently modified rats as a distinction of maleness and femaleness, nevertheless the psychic behavior of the animals, absolutely distinct in itself, lends great weight to the idea of a transformed sexual nature. On the somatic side the early spayed female rat with implanted testis increases in weight, but it has been shown that the testis has no influence upon growth, but that elimination of the ovary does allow a relative increase in weight; the same is also true in regard to body length. The presence of the ovary tends to re- tard the growth of the animals, either male or female, but it is very difficult if not impossible to interpret these changes intelli- gently in relation to a modified sexual condition. The testis is entirely without effect in this regard. Neither the hair coats, fat deposition, nor temperamental be- havior outside of sexual reactions and maternal instincts appears to the writer to be of any deciding significance, and the size of the pelvis and other bones, for reasons already given, appear at 158 CARL R. MOORE WA ant 46 “WO ute Nese ina? \, Ay a | = o 3 \ Poo) } Saweesges LESS Fig. 5 Section of testicle graft (same as fig. 4) more highly magnified to show interstitial cells, and degenerated condition of tubules. st., seminiferous tubules; Ic., interstitial cells. GONADS AS CONTROLLERS OF CHARACTERISTICS 159 best to be very poor criteria of maleness and femaleness. The negative influences, such as failure of the penis and seminal vesicles of the male rat to grow, are nothing more than we could expect in any castrated form. It has been shown repeatedly that many structures of this kind depend upon the presence of the testis for their growth and development. And since the primordium of these are absent in the female, we could not ex- pect their development. Guinea-pigs as well as rats afford good material for these con- siderations only from the ease with which they are handled and with which they withstand operation, but they afford very poor material from which to draw demonstrable conditions and defi- nite conclusions in regard to sex modification of this kind. They possess no distinct sexual differences, aside from the internal and external sexual organs, that will specifically classify them as a male or female, and hence are decidedly inadequate for experi- mental purposes to decide the question at hand. The writer purposely postpones further discussion of the bear- ing of these results until the observations on guinea-pigs are complete. Hull Zoological Laboratories, University of Chicago 160 CARL R. MOORE BIBLIOGRAPHY Donaupson, H. H. 1915 The rat. Memoirs of The Wistar Institute of Anat- omy and Biology, no. 6, Philadelphia. GoopaLe, H. D. 1916 Gonadectomy in relation to the secondary sexual char- acters of some domestic birds. Carnegie Institute of Washington, no. 243. Kine, Heten D. 1915 See tables in The rat, Donaldson. Sreinacu, E. 1910 Geschlechtstrieb und echt sekundire Geschlechtsmerk- male als Folge der innersekretorischen Funktion der Keimdriisen. Zentrl’bl f. Physiol. Bd. 24, 8. 551-566. 1911 Umstimmung des Geschlechtscharakters bei Siugetieren durch Austausch der Pubertitsdriisen. Centrl’bl f. Physiol., Bb. 25, S. 723-725. 1912 Willkiirliche Umwandlung von Siugetier-Minnchen in Tiere mit ausgepragt weiblichen Geschlechtscharakteren und weiblicher Psyche. Pfliigers archiv. f. d Gesammte Physiol., Bd. 144, S. 71-108. 1913 Feminierung von Minnchen und Maskulierung von Weibchen. Zentrl’bl f. Physiol., Bd. 27, 8. 717-723. StorsensurG, J. M. 1909 On the growth of the albino rat (Mus norvegicus var. albus) after castration. Anat. Rec., vol. 3, p. 233. 1913 The effect of spaying and semispaying albino rats (Mus nor- vegicus albinus) on the growth in body weight and body length. Anat. Rec., vol. 7, p. 183. 1917 Observations on the influence of isolated ovaries on the body growth of the albino rat (Mus norvegicus albinus) Anat. Rec., vol. 12, p. 259. ee wee j ea : fr : fis. if yi my wa ee thin 1H 91g Biiarrs HS aye. Polbritn 4 “Meera: ¢ me nda: Li NaN Auticsts x) Resumido por Donald Walton Davis, por el autor, Herbert W. Rand. Multiplicacién asexual y regeneracién en Sagartia luciae Verrill. Sagartia luciae se reproduce asexualmente por medio de una biparticion en sentido aboral-oral, seguida de regeneracidén. El plano de esta biparticién es vertical y tiende a ser perpendicu- lar al eje mayor de la boca, tendiendo a cortar los endoceles a preferencia de los exoceles y en este caso suele cortarlos endoceles completos a preferencia de los incompletos, seleccionando los no directivos mas bien que los directivos. Los productos de esta biparticién son generalmente desiguales. Después de la bipartici- 6n los bordes de la pared del cuerpo se encuentran y fusionan. Los del eséfago se reunen también y en el centro de la regién fusionada se desarrolla invariablemente un sifonoglifo. A cada lado de este plano directivo nuevamente establecido se regeneran dos pares de mesenterios completos no directivos, siempre que el mesenterio originario mas pr6ximo sea incompleto; pero si es com- pleto y no directivo se regeneran un par de mesenterios comple- tos no directivos mds un mesenterio completo unico, el cual se dispone en pareja con el mesenterio impar originario para formar un par no directivo. Los mesenterios incompletos se regeneran conservando las posiciones caracteristicas. El autor describe el orden de desarrollo de los nuevos mesenterios; en relacién con ellos se desarrollan nuevas bandas anaranjadas conservando su posicién caracteristica. La regeneracién anade una serie de estructuras bastante definida y fija que no guarda relacién alguna con el tamajfio inicial y la forma de los productos de la biparticién. Antes que la regeneracién se haya completado pueden Ilevarse a cabo otras biparticiones. Los individuos resultantes son, por consiguiente, diferentes en extremo con relacién al nimero de sifonoglifos, mesenterios y bandas de color anaranjado. Sa- gartia luciae es, probablemente, fundamentalmente hexamérica. La escasez de individuos regulares hexaméricos depende proba- blemente de este proceso de biparticién y regeneracion que proba- blemente es el medio mas efectivo para mantener 0 aumentar la poblacién de una regién determinada. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 1 ASEXUAL MULTIPLICATION AND REGENERATION IN SAGARTIA LUCIAE VERRILL! DONALD WALTON DAVIS FORTY-TWO FIGURES CONTENTS BIR ENGAONE © Roouge se Pek cae sts ee ROE E TS Elie etl pense eee 162 Mesempmon of materials. : .J)i) Fk... eae beeps S741.) QoL! See 163 iadiommanhrereneraiien :t).7. sib. bayer pee ee ees 2/4. bores eccpehien. 167 MERE. 05, FAST OW. 2 Soc at ss Og 2 ca Et a 5 ote gS Nl 169 GSMEAe Ol TEPENETALION!:, < .\2..'s » a txy- Hae SF. Peale ln Oe Hoe: DS 1 gee ate te: 207 Oramberspripes and cineliges sls af. es he eh eae ee oe ee 207 Grance stripes in, fission andiregeneration.......22..<.5. 2s ees ee Meng Ok ee ea ee eee eee 220 PEAS OETIDOS. 5 oiscs.4 1 te See enters o nLereia exe nGee/a ee ree anes 221 Form resulting from ontogenetic development....................20.-0000 223 Beers) 703. feo oes a. Aaah iiss: Be ee eyes ae ae 225 MTR UT 5 00355. 2 hun dT git, d aorqualeane acd un eA ee eee 231 Peenriolgcused in tables and plates. tend omen a d, cic?, c3 c RRS 0h) RE ee ar seer d, ci c ees 12) DR SYS ee Se d, clc?, c3ct, (1) (1) les eee ee eae aR Da Sa. ih oe d, c\c? (1) DUST ay eeias Meio tile ys eee See. ad, Ge, e%c%, ad d ITI gs oR TR i eset eS Gh C0, (E- Table 2 gives, in a form favorable for reference, the normal regeneration formula of one side of the new area when the adjacent old bounding mesentery is a complete non-directive, c; an incomplete mesentery, (1); a directive, d; or when no un- paired old bounding mesentery is present. The symbols used are those adopted for the formulas of tables 3 to 8 (see explana- tion, p. 232), except that the complete non-directive mesenteries are numbered with exponent figures. Mesenteries represented by the same symbol in different regenerating regions correspond, in completeness or incompleteness, in location with respect to the new directive plane, in position of the longitudinal muscle bands, and, as will appear later, in size relations from a short time subsequent to their first appearance until the new com- plete mesenteries reach a degree of development equal to that of the old mesenteries. It must not be assumed, however, that the complete non-directive mesenteries designated by the same exponent are in all cases homologous (cf. p. 197). When less than the maximum number of mesenteries repre- sented in the above formulas are regenerated, two non-directives on either or both sides of the new directives are commonly lack- ing. The missing mesenteries are probably those designated c? and c?. The reduced number is found in approximately 50 per cent of all cases of regeneration following division in an endocoel of the first incomplete grade, but rarely following division in other planes. This is brought out in the right half of table 10. 192 DONALD WALTON DAVIS It will be seen that the full number of mesenteries (4) in the half of a regenerating zone adjacent to a complete non-directive bounding mesentery is found in 93 cases, the reduced number (2) in 8 eases. On the side of old incomplete bounding mesen- teries of the first grade the number of instances of the reduced number (3 mesenteries) is greater (22 to 21) than the number of cases of fuller regeneration (5 mesenteries). ‘Toward incom- plete bounding mesenteries of second or third grade, no reduced numbers are found; and the same is true of regeneration follow- ing divisions in exocoels. No reductions of the type here. con- sidered are recorded for regenerations after division in directive endocoels. The instance of regeneration of but five complete mesenteries under these conditions is no. 10b of table 3. This specimen shows an incomplete new bounding mesentery in place of the directive that would be expected to pair with the old bounding directive. It may be well to call attention here, on the one hand, to the rather fixed character of mesenteries and other structures added in regeneration—a set of structures only slightly variable except as modifications near the edges are necessary to enable this set to fit into normal Hexactinian order with old parts adjacent to the boundary—and, on the other hand, to the highly variable result of such a regenerative process superimposed upon the process of fission previously described. The extremely variable number of complete mesenteries is thereby fully explained. From what has been said of the number of mesenteries found in the new region, it is clear that the production of new mesen- teries does not continue indefinitely. On the contrary, the result of the process is strictly limited. In the following pages it will be shown that the new mesenteries appear in a quite definite order. Order of development of mesenterres For my study of the order of development of mesenteries in regenerating regions, I have used such naturally divided speci- mens represented in tables 3 to 6 as showed sufficiently early stages of regeneration, and also some specimens that had been ASEXUAL REPRODUCTION IN SAGARTIA 193 artificially cut. Of the latter I have used only such as show clearly the distinction between new and old sectors and are not complicated by the presence of mesenteries extending through only part of the length of the column. About three days after fission has become complete, two mesenteries appear approximately in the middle of the space between the two old bounding mesenteries. This space may appear much like an ordinary endocoel (fig. 4), but is usually somewhat wider. At this stage the two new mesenteries are sometimes (as in the specimen referred to) united by their inner edges forming a loop. Occasionally they retain this connection until they are complete orally and have well-developed longi- tudinal muscles; but usually they soon separate at the tips, as indicated in figure 5, which represents a more aboral section of the animal shown in figure 4. Very soon after this two other mesenteries appear between the first two. This establishes a set of four mesenteries, which retain the same relative size nearly up to the time when they become complete. This set of four is a striking feature of the regenerating zone for a considerable period even after other mesenteries have appeared. It probably corresponds with the group of four mesenteries found in §. davisi by Torrey and Mery (’04) and represented in figure 5 of their paper. ‘These four mesenteries appear in their character- istic relations in the photographs shown in my figures 7 and 8. The inner members of this set of four ordinarily become com- plete slightly in advance of the outermost ones. They form the directive mesenteries. The outer members of the set of four be- come complete and remain the nearest complete non-directives on either side of the new directive pair. They are the mesenteries referred to as c!in the account of the mesenteries of the new region, and are so labeled in the figures. The first four mesenteries of the new region are formed in the order described, a0 matter in what spaces division has occurred. The order of development of additional complete mesenteries, as well as their number, de- pends chiefly upon the old bounding mesenteries. As shown above (p.186), on the side of a regenerating region adjacent to a complete old bounding mesentery, two additional 194 DONALD WALTON DAVIS mesenteries destined to become complete almost invariably appear. One, lying nearer the bounding mesentery and de- veloping for some time slightly in advance of the other, becomes the mate of the old complete bounding mesentery. This is designated c? in the figures. The other complete mesentery becomes the mate of c!. It is referred to as c?. On the side of the new directive plane toward a complete old bounding mesen- tery the order of formation of the complete mesenteries is there- fore as follows: c!, d, c3, c. This order is apparent, through differences in size of the mesenteries, in figures 9 and 10, repre- senting sections at different levels of a single animal. The order in which the new mesenteries become complete is some- what different from the order of their appearance, the directives usually being first to reach the esophagus, followed very shortly by cl. c very soon equals c? in size, and these two become con- nected with the esophagus at a somewhat later time. For a long time their inner ends are free from the esophagus near the aboral end of the latter. Occasionally the inner ends of c? and c® are united as described for the new directives. Instances of the stage where all of the new complete mesenteries except c? and c® are attached to the esophagus are represented in figs. 15, 18, and 31. As has been previously remarked by Carlgren (04, p. 52), it was probably this stage in regeneration which the Hertwigs (’79, p. 82) took for a stage in ontogenetic development in the case of two specimens of Adamsia. On the side of the first set of four toward an incomplete old bounding mesentery, the next mesentery to appear is likewise destined to become the outermost complete mesentery of the new piece. The longitudinal muscle, when it appears, faces toward the new directives, whereas the muscle of the outermost mesentery adjacent to an old complete mesentery faces (see above) toward the old part. This mesentery is c’. The next to appear is the one designated c?. It is destined to pair with cl; c3, the mate of c!, appears very slightly after c?, or even simul- taneously with it. An incomplete mesentery, (/), appears ad- jacent to the incomplete bounding mesentery at a later period, as described below. ~ ASEXUAL REPRODUCTION IN SAGARTIA 195 Whether the old bounding mesentery is complete or incom- plete, soon after the appearance of the mesenteries destined to become complete and about the time these reach the esophagus, pairs of the first order of incomplete mesenteries appear in the exocoels between d and c! and between c? and c®. This is shown in an early stage on the right sides of figure 13 and of figure 11, and in the older regenerating regions of figures 16 and6. In some cases these pairs appear simultaneously, but commonly the pairs nearest the directives are slightly in advance of the pairs nearer the boundary. Considerably. later, pairs of a second cycle of incomplete mesenteries appear in their characteristic positions, alternating with the pairs of both the complete and the incom- plete mesenteries of the first order. See, for instance, figures 17, 27, and 28. In cases where the old bounding mesentery is an incomplete mesentery of the first grade, the new one (J) mating with it appears about the same time as the first cycle of incomplete mesenteries in the regenerating part, but in nearly all cases as the first of the mesenteries of this order. It remains the largest mesentery of its cycle for a considerable period. This mesentery is shown in figures 13, 11, 16, and in the older re- generating region of figure 6. When, on the other hand, the old incomplete bounding mesentery is of the second order, the new incomplete mesentery (JJ) pairing with it appears much later, at about the same time as the incomplete mesenteries of the second grade in other parts of the new region, usually as the first representative of this cycle. Two such new incomplete mesenteries may be seen in figure 17. An apparent exception is noted in connection with no. 80, table 7. It thus happens that one old and one regenerated mesentery, constituting a pair, may be definitely assigned to a certain cycle. Furthermore, the bounding mesentery is the only one of this cycle to be produced between the boundary and the nearest pair of new complete mesenteries, no incomplete mesenteries of higher grade being formed in this space. Consequently, after regeneration follow- ing division in an incomplete endocoel of the second order, the space including the boundary and lying between two adjacent pairs of complete mesenteries will lack incomplete mesenteries 196 DONALD WALTON DAVIS of the first cycle unless a pair of these were included in the old part. As a result, normally divided and regenerated specimens occasionally lack a pair of Incomplete mesenteries of the first cycle between two adjacent pairs of complete mesenteries. As will be shown later (p. 210), this involves the loss of an orange stripe, giving rise to an uneven number of these externally ob- servable features. An unexplained lack of a pair of incomplete mesenteries of the first cycle between two pairs of new complete mesenteries is evident in the specimen represented in figure 6. Some instances have appeared which show two pairs of incom- plete mesenteries of approximately equal size unseparated by mesenteries of a higher grade. I have among my sections perhaps half a dozen examples of this anomalous condition (p. 212). I have no explanation for it. Neglecting incomplete mesenteries except the single bounding one, the order of appear- ance of new mesenteries on the side of the new directive plane toward an old incomplete mesentery may be indicated as follows,cly GWE er CF. 3(1). Soon after c! becomes complete orally, ct becomes equal to it in development. c? and c’, as in the case of division in a com- plete endocoel, become equal in size, but lag considerably behind their mates in becoming complete. Various stages in regenera- tion on this plan are represented in figure 6 (older regeneration), 11 and 13 (reduced regenerations), and 15. I have few examples of early stages of regeneration where the old bounding mesentery is a directive. ‘Two such cases indicate that the new bounding directive appears at a stage very slightly in advance of the paired incomplete mesenteries, i.e., at the same stage as an incomplete bounding mesentery. Examples of later stages show the bounding directive fully as well developed as the pair of directives in the middle of the new area. It is possible that the new bounding mesentery in such cases develops more rapidly in the intermediate stage about the time when it reaches the region of the oral disc and esophagus. I have no clear cases showing early stages of development after division in exocoels. One would expect the order of de- velopment to be the same in such cases as when division is in ASEXUAL REPRODUCTION IN SAGARTIA 197 incomplete endocoels, except for absence of new ‘bounding’ mesenteries. In the foregoing account, the determining influence of the old bounding mesenteries has prominently appeared. This influ- ence appears early, and finally results in an adjustment of old and new parts that restores in the bounding region the normal pairing of mesenteries of a given cycle, and, usually but not invariably, the regular alternation of pairs of different cycles. In connection with the mesenteric formulas of regenerated regions, a word of caution was given concerning assumptions of homology between mesenteries bearing the same designation. The reason for this may now be made clear. Mesenteries indi- cated by the same symbol in different formulas (table 2, p.191) are similar in character, as directives or non-directives, in posi- tion of the muscle banners (toward or away from the directive plane), and in location with respect to other mesenteries. If we retain these designations but place them in the order of their development, we have the following as the chief formulas: Novia, seh dec No: tb. eld No. 2a, ¢, 4, ¢, &, 6, (1) No. 2b) eh. d.c* (1) Formula no. 1b differs from no. la in the absence of mesen- teries c? and c?. In formula no. 2b the third mesentery is desig- nated c?, but may really be homologous with c! of formula No. 2a. In that case the reduction here also consists in the suppression of c? and c*. If this is correct, strict regard for homology would require that the mesenteries labeled c? in the lower part of figure 11 and on the left of figure 13 should be labeled c‘. In one or two instances I have found mesenteries c? and c? in a very early state of development when the other com- plete mesenteries were united with the esophagus through the greater part of its length. Whether these would have attained full development or would have disappeared cannot be de- termined. In either event these cases may represent a condition intermediate between the more complete development and the 198 DONALD WALTON DAVIS reduced regeneration, and may point significantly to c? and c’ as the mesenteries omitted in the reduced type of regeneration. Comparing formulas no. la and no. 2a as given above, we find that the third mesenteries to appear are different. Under these circumstances it is impossible to decide which mesenteries of the two formulas are truly homologous. | It will be noticed that, as described above, the members of a regenerating pair of complete mesenteries, except directives, do not appear simultaneously. Although attaining eventually to a condition of approximately radial symmetry, complete mesen- teries arise in regeneration in bilateral fashion, a member on one side of the directive plane corresponding in degree of develop- ment with one in similar position on the opposite side of the directive plane. There is no reason to suppose that the same is not true of the development of mesenteries in the metamorphosis from the larval state. The complete mesenteries of S. luciz, therefore, are all to be regarded as primary mesenteries, belong- ing to the first cycle of mesenteries, which is very generally found arising in a bilateral manner in the ontogenetic develop- ment of Hexactinians (cf. p. 207). Perfect bilateral symmetry of a regenerating region is often prevented through the influence of unlike old bounding mesenteries, or by unknown factors which cause the suppression, on one side of the directive plane, of mesenteries present on the other. Results of repeated fission and regeneration It has been shown that the products of fission in a group of specimens of 8. luciae vary greatly in numbers of siphonoglyphs and of complete mesenteries. The former were found in speci- mens recorded in tables 3 to 7 up to three, the latter up to thirteen. It has been shown, further, that there is added in regeneration a set of structures including a siphonoglyph and a pair of directive mesenteries, together with other mesenteries including mates to the old bounding mesenteries. The com- plete mesenteries on either side of the new directives may vary from one to five, making possible a total addition in regeneration ASEXUAL REPRODUCTION IN SAGARTIA 199 of four to twelve. As a matter of fact, the smallest number added in any case of complete regeneration as recorded in these tables was five and the largest number eleven. The average results of repeated divisions and regenerations may be derived from the data at hand. On the right of table 9 are given the numbers of mesenteries regenerated in zones of different types. The same data are given, considering sepa- rately each lateral half of a regenerating region, on the right of table 10. As explained in connection with the latter table, the average number of complete mesenteries in regenerating zones of all types is 8.4. It is obvious, therefore, that repeated divisions into two parts followed by regeneration will tend toward an average of approximately 17 complete mesenteries. Table 11 exhibits the number of complete mesenteries after regenera- tion for the specimens recorded in tables 3 to 6, and the average number of such mesenteries for each siphonoglyphic class and for the whole. The latter average is 15.8. The fact that this is below the number toward which repeated division and regenera- tion tend, points to a still lower number of mesenteries in the form resulting from ontogenetic development. The average number of complete mesenteries, before division, of specimens represented in table 4 (including only those indi- viduals all of whose fission products were available for record) is 13.5, a number considerably below the average for all the specimens of tables 3 to 6. In each of the cases given in table 4, the two closely succeeding divisions resulted in rapidly increas- ing the number of complete mesenteries. Although it may be purely a coincidence that apparently multiple divisions have, in these observed instances, occurred in specimens with a low num- ber of complete mesenteries, it is possible that such divisions serve in an adaptive way to bring about a rapid increase. Whether this is correct or not, certainly the variations described for the processes of fission and regeneration are adequate to account for much wider variations in form of regenerated speci- mens than have been encountered. This indicates clearly the probability that there are correlations in these variations that are as yet unproved. 200 DONALD WALTON DAVIS DISCUSSION OF REGENERATION IN HEXACTINIANS The only previous detailed accounts of radial regeneration of sea-anemones comparable with the foregoing are those of Carl- gren (704 and ’09) and Cary (’11). I shall now review the work of these writers in so far as it bears upon the problems here considered. Carlgren dealt with the regeneration of Sagartia viduata, Metridium dianthus, and Aiptasia diaphana. Of the last-named species only naturally produced basal fragments were studied. In Metridium, the material considered of natural frag- ments, artificial pieces of the same character as those separated naturally, and pieces cut from the base of the parent polyp in such a way as to exclude, so far as possible, all tissue of the column and of the mesenteries. I shall refer to these last as ‘basal pieces.’ Of Sagartia viduata, which does not reproduce naturally by asexual methods, artificial fragments of various forms and sizes were used, including some ‘basal pieces.’ Natural fragmentation in Metridium and Aiptasia consists in the separation from the parent polyp of a small portion of the base and adjacent wall of the column with the adhering parts of mesenteries. The fragment thus receives only a very small proportion of the material of the parent polyp. The products of division in §. luciae, while they may be far from equal, are, so to speak, of the same order of magnitude, and each contains some part of the base, column, esophagus, circle of tentacles, and set of mesenteries. Resorption of old mesenteries In Metridium and S. viduata degeneration of mesenteries is evidently a prominent feature of the process of reconstruction. In these forms, however, degeneration does not commonly go to the extent of eliminating all of the old mesenteries. In Aiptasia, as stated by Andres (’82), rearrangement of mesenteries begins before the separation of the fragment from the parent. Carl- gren believes that degeneration of the mesenteries, also, begins before separation is complete. He leaves in doubt the extent to which degeneration may go, since he was unable to determine ASEXUAL REPRODUCTION IN SAGARTIA 201 with certainty, in his sections, any boundary between old and new regions. His account is supplemented in this respect by that of Cary (’11), who studied the regeneration of three species of Aiptasia. According to Cary, all of the old mesenteries are resorbed, first at the oral extremity of the piece and progres- sively down the column until they have disappeared entirely.* Cary studied the process of regeneration following pedal lacera- tion in another species (from Beaufort, North Carolina) that has been incorrectly known as Cylista leucolena. Presumably de- generation of the old mesenteries occurs in this species as in Aiptasia. Resorption of old mesenteries, with the possible ex- ception of occasional members torn during fission (see pry does not occur in S. luciae. Sequence of new mesenteries A brief account of the various types of arrangement and order of appearance of the mesenteries in regenerating anemones, as described by Carlgren and Cary, will be followed by a discussion of Carlgren’s theories concerning the relation of these types to one another. Figure 35 shows the types of arrangement of complete mesen- teries found by Carlgren. Parts enclosed in the dotted lines indicate old regions, without an attempt to represent the number 3 After examining Carlgren’s original figures (’04, Taf. IX, Fig. 4 and 7) and text, I cannot agree with Cary (’11, p. 94) that ‘‘it seems very evident that all of the mesenteries shown in Carlgren’s Fig. 7, Taf. IX, are old ones which have come over in the fragment from the parent individual and which will never come to be a part of the permanent system of mesenteries of the actinian arising from the laceration embryo.’’ His interpretation of this figure may be correct, but the evidence for it is by no means conclusive. Criticism of another sort is due for Cary’s treatment of Carlgren’spaper in otherrespects. Glaringinaccuracies in interpreting the statements concerning the figures mentioned above are fortunately largely exposed by his quotation from the text. His mutilation of Carlgren’s excellent figures is not so obvious to anyone not having the latter’s paper at hand. Comparison with the originals of the figures (Cary, ’11, pp. 92, 93) purporting to be copied from Carlgren’s paper (’04, Taf. IX, Fig. 4, 7), reveals amazing discrepancies. To alter a figure in such fashion is a violation of the privilege of copying, even if modification be acknowledged; to do it without such admission is an offense against both the author, whose work is thereby misrepresented, and the reader, whose confidence is abused. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 2 202 DONALD WALTON DAVIS or character of the old mesenteries. Differences in size of the members of a pair of mesenteries in the diagrams of this figure indicate the order in which the mesenteries of the pair become complete. Table 12 sets forth the frequency with which these types occurred in regenerating pieces of different species. Cary’s regenerating specimens of Aiptasia and Cylista (?) belong to type IV. The cases I have given of regeneration in §. luciae which do not involve matching up of new with old complete mesenteries are of Carlgren’s types III, VI, or II. For con- venience in indicating the order of appearance of these mesen- teries, I have added to diagrams I and III the symbols by which I have designated mesenteries occupying corresponding positions with respect to the new directive plane. Using these symbols for the different mesenteries, we may construct formulas for the order of appearance as given by Carlgren in these different types, thus: (I) Ge cy d, C’, c (i) eka. 7073 (TED: clr dratvesc’ Carlgren’s statements indicate that in (I) c? and c* appear nearly simultaneously. In (III), ct appears only slightly after c', and c® very soon after c?. It will be seen that the sequence of mesen- teries in development given for type (III), as well as the order of becoming complete, is the same that I have found to occur in S. luciae except in cases where a complete old bounding mesentery is present. Carlgren has proposed (’09, p. 41) an ingenious theory con- cerning the relations of the types of arrangement of mesenteries demonstrated in his regenerating specimens. For the arguments he produces in its support, the reader is referred to Carlgren’s paper. I shall here simply outline the theory. Types I, VII, IX, and X show clearly bilateral arrangements of the mesenteries in development. Type VIII is typically biradial in form. Type I Carlgren takes to be the most primi- tive. It parallels the Edwardsian method of ontogenetic de- velopment except in that the presence of the old part inhibits ASEXUAL REPRODUCTION IN SAGARTIA 203 the so-called ventral pair of directives. Type II might be considered as having been produced by reduction from either a bilateral or a biradial condition, but Carlgren regards it as a modification of Type I through the failure to develop of the innermost mesenteries (d) there represented. According to this view, then, d of Type I is lacking in Type II, and c! of Type I becomes d of Type II. Type IV is a combination of Types I and II, and Type X is the result of a doubling of Type I. Type VII represents an extreme expression of the tendency to the bilateral arrangement of mesenteries shown normally in Type I. Type IX represents normal development, such as Type I, ina basal piece having no part of the old column to interfere with the formation of the ventral directives. In explanation of Type III it is assumed that there are here two regenerating regions, each similar to that of Type II. In one of these regions the directive mesenteries are replaced by mesenteries of the old part. Type VIII is similar to Type III, but lacks old parts which might occupy the position of one pair of directives. Type V is a mixture of Types II and III; Type VI, of I and III. It may be pointed out that the explanation for Type III is not entirely in harmony with the order of development of that type, but would demand the order c! and c‘, d, c? and c*. It seems to me, however, that we cannot profitably consider at present the more abstract questions of the relation between the biradial and bilateral types of development. We cannot expect to solve these problems of form determination from examination of data collected for other purposes and assembled into tables and diagrams. The most that can be expected of such material from this standpoint is that it may present definite problems and suggest favorable points of attack. One of these problems concerns in a concrete way the relations between the biradial and bilateral plans of regeneration. Carl- gren has shown that both types may occur in the same species or even in the same regenerating pieces, and he has made a beginning in ascertaining the conditions determining the plan of development. He has shown that in Metridium larger pieces containing part of the base, column and mesenteries develop, 204 DONALD WALTON DAVIS almost without exception, according to the biradial plan; but pieces containing material from the base alone frequently de- velop on a bilateral plan. In S. viduata the same influence of the character of the material in the regenerating piece is found with a greater tendency, whatever the nature of the piece, toward a bilateral plan of development. It may be that the rounding up of the more homogeneous material of basal pieces results in conditions similar to those influencing the development of mesenteries in ontogeny. Siphonoglyphs and directive mesenteries We cannot go far in considering the relation between these different types of mesenteric development without being con- fronted with the more general question of the determination of form in ontogeny and in regeneration. Carlgren’s discussion of the relation of Types I and II suggests one of these, i.e., the governing influence of the siphonoglyph. He puts forward (09, p. 43) the idea that, upon the formation of the siphonoglyph during regeneration, this structure immediately determines that the pair of mesenteries latest formed in the same plane shall be directives, and that no other mesenteries shall be formed in this plane. According to this view, the difference between Types I and II is due to the earlier stage at which the new siphonoglyph is established in II. The latest bilateral pair of mesenteries formed at the time of the appearance of the siphono- glyph becomes the pair of directive mesenteries, and no new bilateral pairs arise except those mating with mesenteries already present to form radially placed non-directiye pairs. My observations lend some support to this hypothesis in so far as it involves the determination of the directives by the siphonoglyph; but I believe that in 8. luciae delay in formation of a siphonoglyph does not lead to the production of more than two bilateral pairs of mesenteries adjacent to the potential directive plane. As I have already indicated, the siphonoglyph appears shortly after closure of the wound in the region of the mouth. At about this time some of the new mesenteries reach ASEXUAL REPRODUCTION IN SAGARTIA 205 the esophagus. Previously the longitudinal muscles of the new mesenteries are not apparent, but about the time these mes- enteries reach the esophagus the muscles rapidly develop and the directives are distinguishable from the non-directives. The order of events suggests that the development of mesenteries into directives is determined by the presence of a siphono- glyph. In a number of cases I have found a group of four small mesenteries with the characteristic proportions of the first four regularly formed in regeneration, but extending only a short distance up and down the column. Elsewhere the column wall and mesenteries gave no evidence of a division which might have given occasion for such a regenerating area. I interpret these as regions of regeneration following comparatively slight injuries to the body wall. In most of the instances the mesenteries are small and have no indication of longitudinal muscles. In no case do mesenteries in these sets show the characteristics of a pair of directives. In one case apparently the two inner mesen- teries of the four have the characteristics of non-directives. In another instance two mesenteries only are formed. These are long and slender and show no longitudinal muscle bands. They reach the esophagus in a region where histological evidence of a siphonoglyph is not present, although there was a slight groove at that region of the mouth, as seen from the exterior and in sections, and although a narrow white line, indicative of the presence of a siphonoglyph, was to be seen in the living animal extending part of the way from the groove toward the tentacular zone of the disc. These observations suggest that the first four mesenteries which form so constant a feature of the regeneration of S. luciae arise independently of the siphonoglyph, and that con- trary to Carlgren’s hypothesis, no more mesenteries are formed - adjacent to the directive plane even in the absence of a siphono- glyph. The order of events in S. luciae is as follows: fusion of edges of the column, appearance of a set of four new mesenteries, extension of these to the esophagus, formation of a siphonoglyph, development of longitudinal muscles in positions which mark out the inner members of the first set of four mesenteries as a 206 DONALD WALTON DAVIS pair of directives. This order of events may be interpreted as an epigenetic form-determining series, in which event A leads to event B, ete. Influence of old bounding mesenteries A most striking difference between the regeneration I have described for 8. luciae and that of all the types given by Carlgren and Cary consists in the total absence from the latter of any variation ascribed to the influence of old mesenteries. In Aiptasia, where old mesenteries are resorbed, and even in cases of Types IV and IX in §. viduata, where little or no old tissue belonging to mesenteries or column is present, the lack of influence of old mesenteries upon regeneration is not surprising. But in many other cases, in both 8. viduata and in Metridium, well developed mesenteries apparently exert no influence over regeneration. The variations in arrangement of new mesenteries in the species - described by Carlgren are of wholly different character from the variations seen in 8. luciae, which are governed almost com- pletely by the mesenteries on the torn edges of the old piece. The influence of these bounding mesenteries in §8. luciae is apparent in the earliest stages of regeneration. They have no obvious effect upon the directive mesenteries nor upon the first non-directive mesenteries, which precede the directives. The character of the bounding mesentery may, however, determine the nature of the third mesentery on either side. If the bound- ing mesentery is complete, this third mesentery becomes its mate and a fourth mesentery, which becomes the mate of the first non-directive, very soon appears. If, on the other hand, the bounding mesentery is incomplete, the third mesentery is followed by a fourth and a fifth. The determination of the number of mesenteries is effected before these mesenteries have their longitudinal muscles developed. My impression, here as in connection with the determination of siphonoglyphs and directives, is that we have to do not with one, but with a num- ber of form-determining influences successively brought to bear. Some of these influences have been suggested, but we are far ASEXUAL REPRODUCTION IN SAGARTIA 207 from having any complete list of the factors concerned, and from understanding fully their order of effectiveness, much less their fundamental nature. Pairing of mesenteries The tendency to form, ultimately, unilateral pairs of mesen- teries, i.e., pairs whose members lie on the same side of the directive plane, is apparently very strong (except in the case of directives) in the regenerating regions of all these species of anemones. This is especially evident in 8. luciae, where it usually involves the matching up of regenerating mesenteries with old ones. There is no evidence, however, that non- directive mesenteries destined to become complete ever arise in regeneration as unilateral pairs. In every case described, one member of the pair precedes in development. According to their manner of development, then, the complete mesenteries are what have been commonly referred to as ‘primary’ mesenteries. They correspond in order of appearance with the ‘Hauptsepten’ of the Hertwigs (’79, pp. 81, 88), with the ‘protocnemes’ of Duerden (’02, p. 388), and with both the-‘protocnemes’ and the ‘deuterocneme’s of MeMurrich (’10, p. 4). The first cycle of mesenteries arising as unilateral pairs in the primary exocoels are, then, secondary mesenteries. They are ‘metacnemes’ according to Duerden’s terminology and ‘zygocnemes’ in McMur- rich’s. As already stated, there is no evidence that the latter ever become complete in the forms whose regeneration has been studied. ORANGE STRIPES AND CINCLIDES Some observations on the relation of the orange stripes to the processes of fission and regeneration in 8. luciae have been made by Davenport (’03, pp. 140, 148). Her statements will be re- ferred to in connection with the evidence I have collected bearing upon the points involved. In a well-expanded living animal with brightly colored orange stripes it is not difficult to determine the relative positions of 208 DONALD WALTON DAVIS siphonoglyphs, directive mesenteries, and orange stripes. An orange stripe is invariably found opposite the siphonoglyph be- tween the lines of attachment of the members of the pair of directive mesenteries. As Davenport has stated, the orange stripes occur only in endocoels. Comparison of the number and position of the orange stripes as seen in the living animal, with the number and position of the mesenteries as found in sections, indicates that the orange stripes lie between the mem- bers of pairs of complete mesenteries, and also, contrary to the statement of Davenport, of incomplete mesenteries of the highest grade. For instance, the specimen a section of which is represented by figure 17 showed ten orange stripes; those repre- sented by figures 19, 21, 22, and 26 showed twelve stripes each. In a regenerating specimen in which all the mesenteries and stripes of the new region are formed, the above statement holds (with rare exceptions to be mentioned later) for both new and old parts. The frequency with which division occurs in the endocoels occupied by orange stripes suggests the possibility that these stripes have some functional significance in the processes of fission or regeneration. It is clear, however, that the presence of such stripes is not essential to the normal progress of these processes. Specimens divide spontaneously in other regions, and the resulting pieces, as well as similar fragments artificially produced, regenerate readily. Furthermore, the related species, S. davisi, reproduces freely by the same method, yet is destitute of any such stripes. Davenport (’03, p. 143) calls attention to the fact that the cinclides occur on the stripes. I find, however, that they are not confined to the orange stripes. Figures 33 and 34 show three cinclides, clearly marked by protruding acontia, none of which are in positions occupied by orange stripes. In figure 19 at x appears another cinclis, likewise not situated on an orange stripe. Under favorable conditions of lighting, the cinclides are readily visible on the living specimen with the aid of a hand lens, and their distribution may be accurately determined. They are found in vertical rows in all positions in which fission planes may ASEXUAL REPRODUCTION IN SAGARTIA 209 pass. That these openings affect the position of the division plane is possible but quite unlikely. If they are of significance in the location of the fission plane, their distribution should show some relation to the frequency of fission in different planes. As to this, no evidence is at hand. — ' The fact that cinclides are commonly located in other positions as well as on the orange stripes is evidence against the idea suggested by Davenport (’03, p. 148) that the presence of the stripes may ‘“‘be considered as a case of warning coloration?”’ Orange stripes in fission ahd regeneration ‘During the process of fission, an orange stripe lying in the endocoel cut by the plane of division has been observed, in a few cases, to be divided by the tearing of the column wall. After the division the narrow border of orange along the edges of the pieces must be promptly absorbed, for in no case observed has there been found early in regeneration an orange stripe on the boundary between the old tissue and the new. On the contrary, nine specimens in early stages of regeneration which were killed and sectioned after counting the orange stripes give clear evi- dence of lack of stripes in positions almost certainly containing them before the division. At the time these were killed none showed any orange stripes in the newest regenerating area. Of the nine individuals, six represent the paired products of three divisions (nos. 11, 13 and 14, table 3) and three are unpaired specimens (no. 30, table 6, a specimen represented in figure 9, and one other). All of the paired and two of the unpaired specimens showed division through two complete endocoels. The third unpaired specimen had divided through endocoels of the first incomplete grade. All divided in planes where orange stripes were to be expected. Furthermore, the paired specimens showed orange stripes in such numbers and positions as to occupy all regions, outside of the new area, in which they nor- mally occur, indicating that the orange stripes were probably fully formed before the division occurred. In none of these cases was there any evidence of an orange stripe marking the 210 DONALD WALTON DAVIS endocoel on the boundary between new and old parts. In all of them the number of observed orange stripes corresponds with the number of complete and first order of incomplete endocoels lying wholly within the old part. Thus the specimen of which a section is shown in figure 9 had three orange stripes in the positions indicated diagrammatically in figure 36. The other unpaired specimens showed five and seven orange stripes, respectively, corresponding with the number of undisturbed endocoels normally bearing stripes. Each of the pairs had three orange stripes In one member and seven in the other similarly situated. The indication’ given by these nine specimens is that an orange stripe lying in a space cut by a plane of fission is lost. In regeneration following division through an endocoel occupied by an orange stripe, a new stripe is finally developed in the bounding endocoel in addition to those formed in endocoels lying entirely within the new region. Were this not true, com- plete endocoels lacking orange stripes would be common, whereas they are actually exceedingly rarely, if ever, found. The formation of new bounding mesenteries, when division occurs in one of the lower grades of incomplete endocoels, has already been described (p. 195), the result being that the mesen- teries of a pair enclosing the boundary between new and old regions are of the same grade. The bounding endocoel conse- quently may be definitely designated as belonging to a certain cycle. When this bounding endocoel is of the second incomplete grade or of a lower order, no orange stripe is formed in it. The endocoel on the boundary between new and old regions, is, then, no exception to the general rule that orange stripes are formed in endocoels of the complete mesenteries of the first cycle, but not of lower cycles.. In support of this statement I may cite a number of examples. Anemone no. 82, table 7, a section of which is shown as figure 17, had divided through in- complete endocoels of the second grade or possibly lower. This specimen, when killed, had ten equidistant orange stripes, equaling the number of endocoels of the first two cycles. Evi- dently there were no orange stripes in the bounding endocoels. ASEXUAL REPRODUCTION IN SAGARTIA ort Specimen no. 80, table 7, gives similar evidence. The number of orange stripes and their position in relation to the new area as observed before killing are shown in figure 37. It was noted that possibly one of the orange stripes, here represented as lying within the new area near one boundary, might be an old stripe. Study of sections shows in the old part a pair of di- rective mesenteries and an adjacent pair of non-directives of the first incomplete grade, the plane of fission passing through secondary incomplete endocoels lying lateral to these. The two old and nine new orange stripes correspond in position with the complete endocoels -and the incomplete endocoels of the first cycle. Evidently orange stripes were not produced in the secondary incomplete endocoels occupying the boundary. As a result of this position of the division plane, on one side there is a space between two pairs of complete mesenteries that does not contain a pair of incomplete mesenteries of, the highest grade. In all probability there was no orange stripe in this region—two adjacent orange stripes occupying complete endocoels. This illustrated the seldom realized possibility of the normal pro- duction of a fully regenerated specimen with an odd number of stripes, one less than twice the number of pairs of complete mesenteries. Another specimen showed exactly the same ar- rangement of complete mesenteries, incomplete mesenteries of the first order, and orange stripes, but with the position of the bounding planes no longer evident. Another specimen with twenty-one orange stripes and eleven pairs of complete mesen- teries is similarly explained. Sections of this individual give evidence of two regenerating regions, both in a very late stage. Probably three of the four division planes involved passed through complete endocoels; one evidently cut an incomplete endocoel of the second order. Both old and new mesenteries bounding the latter endocoel are clearly of the second order. This is the only region bounded by pairs of complete mesen- teries in which representatives of the first cycle of incomplete mesenteries are lacking. Undoubtedly this accounts for the lack of one orange stripe from the number usually found in a specimen with eleven pairs of complete mesenteries. Anemone 212 DONALD WALTON DAVIS no. 54, table 6, showed ten orange stripes and (internally) six pairs of complete mesenteries. On one side the boundary between old and new evidently lies in an incomplete endocoel of the second or lower grade. This accounts for the lack of one orange stripe between two adjacent pairs of complete mesenteries. On the other side an undisturbed old incomplete endocoel of apparently the first order is present, and an orange stripe would be expected there. In the absence of an orange stripe in this . region, one is driven to consider the possibility of this pair’s belonging really to the second order of incomplete mesenteries rather than to the first cycle of which it is apparently a member. In that case four cycles of incomplete mesenteries must have been present. If this is so, division must have occurred in an exocoel. Both the presence of four cycles of incomplete mesenteries and divisions in exocoels are uncommon occurrences, and this com- bination is consequently improbable and the explanation un- satisfactory. In the specimens just described, orange stripes were not found on the boundary between old and new regions when this lies between members of the second cycle of incomplete mesen- teries. If this always held true and the incomplete mesenteries were invariably regularly placed, we should never find orange stripes associated with two pairs of incomplete mesenteries not separated by a pair of complete mesenteries. It has been shown previously (p. 196.) that irregularities in the first cycle of incomplete mesenteries are occasionally found. Thus two pairs of incomplete mesenteries of the largest size with no complete mesenteries separating them were found in three specimens on which counts of orange stripes gave odd numbers. Regeneration was apparently complete in these cases, and the number of stripes is such as to lead one to suppose that in each case one occupies the additional incomplete endocoel. In these instances a reduction from the typical number of orange stripes, which is double the number of pairs of complete mesenteries, or an excess over the typical number, corresponds with a departure from the normal number of incomplete mesenteries of the first order. ASEXUAL REPRODUCTION IN SAGARTIA 213 For the sake of completeness, two specimens showing still wider variations from the typical conditions should be described. One of these on which ten orange stripes were counted shows mesenteries that would account for but eight. The other two stripes may have been situated in a broad expanse of the body wall between two pairs of complete mesenteries, in which are seen one pair of large incomplete mesenteries and an irregular lot of small ones, of doubtful history. They vary in number and in size at different levels. I have no explanation to offer for the presence of this group of mesenteries or of the stripes associated with them. The other specimen showed a single stripe in its older part and none in the newer region. Sections revealed five pairs of complete mesenteries and three complete mesenteries whose mates were not yet complete. A specimen with mesenteries present in such numbers and degrees of completeness would ordinarily show at least seven orange stripes. In the preceding paragraphs are described all of the specimens I have observed and sectioned which constitute exceptions to the general statement that the number of orange stripes in an undivided or fully regenerated specimen is twice the number of pairs of complete mesenteries. Of the nine such exceptions, three, showing one less than the typical number of stripes, are explained on the ground of division in an incomplete endocoel of second or lower grade resulting in elimination of one endocoel normally occupied by an orange stripe; three with an extra orange stripe show, each, an additional unexplained pair of in- complete mesenteries of apparently the first order in whose en- docoel the supernumerary stripe probably lay; while three show wholly unexplained deviations from the normal relations of stripes and mesenteries. Over against these exceptions must be urged the significance of forty-nine specimens on which the number of orange stripes as determined before killing was exactly double the number of pairs of complete mesenteries as ascertained by examination of sections. Undoubtedly in these cases the orange stripes occupied the complete endocoels and incomplete en- docoels of the first order only. These forty-nine specimens show 214 DONALD WALTON DAVIS great diversity in features other than the relation of stripes to mesenteries. Thus two have eight stripes; eight have ten stripes; twenty-three have twelve stripes; five have fourteen stripes; nine have sixteen stripes; one has eighteen stripes, and one has twenty stripes. Also thirty-six are diglyphic; twelve are monoglyphic, and one is triglyphic. Four, of which one is represented in figure 26, show no evidence of having undergone division, while the others give more or less evidence of regenera- tion, and a few show clearly the precise position of the plane of fission. Of the latter, some represent divisions in complete en- docoels, some in incomplete endocoels of the first grade, and one, of which a section is represented in figure 17, in incomplete endocoels of the second order. Most of the specimens are biradially symmetrical, but a number depart from this con- dition. The triglyphic specimens and diglyphic individuals with an uneven number of pairs of complete mesenteries cannot be strictly biradially symmetrical. It is clear, therefore, that in nearly all cases undivided specimens, or those in advanced stages of regeneration, have orange stripes corresponding in number and position with the complete endocoels and the highest order of incomplete endocoels. The number of incomplete endocoels of the first grade being almost invariably equal to the number of complete endocoels, the number of orange stripes is twice the number of pairs of comp'ete mesenteries in the overwhelming majority of cases. Nevertheless, in any given population of this species one finds a large proportion of specimens showing an odd number of orange stripes. These are to be accounted for, in small part, on the basis of the irregularities in arrangement of the incomplete mesenteries described above (p. 211). Thus the exclusion of a pair of incomplete mesenteries of the first cycle from a primary exocoel owing to a division in an incomplete endocoel of second or lower grade adjacent to a pair of complete mesenteries, would reduce by one the number of orange stripes in the fully re- generated individual. Again, the unexplained duplication of necomplete mesenteries of apparently first grade was shown to be associated with an extra orange stripe. By far the greatest ASEXUAL REPRODUCTION IN SAGARTIA 25 number of specimens showing odd numbers of stripes are speci- mens which have recently divided and have either developed no new orange stripes or have produced, at the time of observation, less than the full set to be acquired. The number of the new orange stripes gives some idea as to the completeness of the regeneration. From what has been said of the number of mesenteries regenerated and of the relation of orange stripes to mesenteries, it is evident that the number of new orange stripes may vary from five to eleven. In case of division in one or two directive endocoels, twelve or thirteen are possible numbers. Numbers below seven or above eleven are, however, distinctly uncommon for single completely re- generated areas. Below seven it is likely that the stripes of the new region are not fully formed. Above eleven—unless one of the orange stripes on the boundary between new and old lines is opposite a white line on the oral disc and therefore occupies a directive endocoel—it is almost certain that the new area con- sists in fact of two regenerating zones of not very different age. The number of mesenteries regenerated is such that repeated fission and regenerations (p. 198) wou d tend toward an average condition with about seventeen complete mesenteries. Since the complete mesenteries are always paired and correspond, with rare exceptions, with the orange stripes, we may say that the tendency is toward approximately eight pairs of complete mesenteries and sixteen orange stripes. Division followed by complete regeneration would rarely give rise to individuals with fewer than seven orange stripes, and never with fewer than five. The great majority of individuals with small numbers of orange stripes have not completely regenerated, while even among specimens with high numbers of stripes many have not yet completed the regeneration. As to the order in which orange stripes appear in the new region, I have little information. Frequently the presence of stripes may be ascertained while as yet they are so faint in color and so close together that the number cannot be de- termined. At a somewhat later stage there is some reason to believe that stripes are present in complete endocoels when as 216 DONALD WALTON DAVIS yet none can be seen between the members of pairs of incom- plete mesenteries. One of my specimens indicates this. It was described, while living, as a triglyphic individual with three old orange stripes and eleven on the boundaries, or within, the new area (fig. 41). Sections show normal positions for the three old stripes. The sections also demonstrate that the regenerated area is composed of an older and a newer part. The three division planes are all in complete endocoels. Including the bounding endocoels there are in the new tissue eight complete endocoels. In the older regenerated part there are three pairs of incomplete mesenteries, all of the first grade, no mesenteries of a lower grade being present anywhere in the regenerated portion. In the most recently formed part there are four pairs of incomplete mesenteries. The relative position as well as the number of the orange stripes makes it probable that the eleven new orange stripes occupied all of the complete endocoels and the incomplete endocoels of the original piece and of the older regenerated area only. It is probable that new orange stripes would later have appeared in the incomplete endocoels of the first grade of the newest region. Another regenerating specimen showed externally ten orange stripes, whereas sections revealed six pairs of complete mesenteries and six pairs of incomplete mesenteries of the highest grade. It is probable that two of the second grade of incomplete endocoels in the new region lacked orange stripes at the time of observation. I have no clue as to which of the endocoels lacked the stripes. The formation of orange stripes first in the complete and later in the first grade of incomplete endocoels would correspond with the order of development of the mesenteries of the regenerating region. This order of development of the stripes is in harmony with the statement of Davenport (’03, p. 143, and fig. 2) that new orange stripes appear between old ones. It cannot be too positively stated, however, that this formation of new stripes, as well as the production of new complete mesenteries, is con- fined to regenerating regions, and that the process is strictly a determinate one. After the formation of the group of complete mesenteries heretofore described (pp. 185 to 192) and of the first ASEXUAL REPRODUCTION IN SAGARTIA PAW cycle of incomplete mesenteries alternating therewith the number of orange stripes is absolutely limited and their positions are determined. The only occasion thereafter for the formation of new complete mesenteries and new orange stripes is a division initiating a new regeneration. EXTERNAL INDICATIONS OF INTERNAL STRUCTURES Much information concerning internal structures may be ob- tained by a consideration of all the external features of indi- vidual anemones. Differences in ground color of the column wall or in the width or intensity of color of the orange stripes per- sisting in spite of changes in the state of expansion of the speci- men, especially if associated with unevenness in length, breadth, or whiteness of the white bars, give indication that fission has occurred and. that regeneration is in progress. These features may, especially in the earlier stages of regeneration, present differences sufficiently sharply marked to indicate the precise boundary between old and regenerating tissue. An obstacle of no little importance is the liability of unequal contraction in different parts of the wall of the column to alter for the moment the intensity of all colors and the relative positions of the stripes. The intensity of coloration of the new as compared with the old part, the width of the new sector, and the presence or absence in it of a new white bar or of new orange stripes, give basis for judging within broad limits the stage of regeneration of internal parts. Frequently the presence of three or more sectors of different ages may be readily determined, and the different sectors may show various stages of regeneration. When two regenerating areas of nearly the same age lie adjoining each other, it is often difficult to recognize them as two. In certain cases it is practically impossible to distinguish by external observation the constituent parts of such a double regeneration. Whenever the number of orange stripes of an apparently single regenerating region excéeds eleven, it is almost certain that two regenerations are involved. When the orange stripes are faint or when they are less than seven in number in a given regenerating THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 2 218 DONALD WALTON DAVIS sector, there is ground for suspicion that not all of the orange stripes of the zone have appeared, or that a part of the regen- erating region has been separated off in a later fission. As previously stated (p. 167), there is danger, especially in faintly colored specimens, of counting as an orange stripe the narrow space representing the union of the torn edges shortly following a division. This is especially likely in faintly colored indi- viduals. It should be constantly guarded against when count- ing stripes. When no slightly developed regenerating area is present the internal structure may in most cases be accurately — inferred from the number of orange stripes. Thus Davenport’s (03) figure 11 represents probably a specimen possessing material of four different ages. The oldest shows three orange stripes; the next oldest, two; the third, six, and the most recently formed area, nine. All except the last formed sector have been partially removed in the divisions which initiated the later regenerations. If, in addition, the number and position of white bars on the oral disc were known (thereby locating the directives), still other details might be surmised. Under favorable conditions even the incomplete mesenteries of second and later cycles may be identified through the column wall after complete mesenteries have been located. It should be clear, then, that careful exami- nation of living specimens of 8. luciae enables one to distinguish recently divided specimens and to estimate approximately the stage reached in the regenerating region; and, in specimens that have not recently undergone fission, to infer quite accurately and in detail the number, character, and positions of the mesenteries. COMPOSITION OF NATURAL GROUPS We may now look into the composition of colonies of this species and consider to what extent this composition is affected, or accounted for, by the processes of asexual reproduction de- scribed. This examination should also yield information con- cerning the sexually produced form of the species. With these questions in view, I have studied natural groups of individuals taken at different seasons from a variety of situations around Woods Hole. ASEXUAL REPRODUCTION IN SAGARTIA 219 Siphonoglyphs The condition of these groups as regards number of siphono- glyphs is represented in table 13. These are referred to as ‘unselected’ specimens, but the conditions under which the collecting was done and the counts made allow of a considerable degree of selection in addition to the normal errors of random sampling. Each lot consists of a portion of a natural group of _ individuals collected at one time from a closely restricted ’ locality, such as a single stone or a few stones of similar quality lying near together. Each lot was either taken immediately into the laboratory and examined or, when collected during the winter, taken to Cambridge and studied there. In order to determine the number of siphonoglyphs it was, of course, necessary to wait until the animals expanded, exposing the oral disc. Naturally, also, many of the specimens were expanded at one time and the removal of a few whose condition was noted caused many others to contract, necessitating much delay in completing the count. There was opportunity for uncon- scious selection because many were open at a time and because some show their condition clearly at a glance while others require close examination to reveal the state of their siphonoglyphs. In a few cases sections were necessary to determine the number of siphonoglyphs. Experience shows that external examination is not wholly reliable for this purpose and that, in case of any irregularities in the mouth region (cf. fig. 1, 4), only sections can give certain information. In each lot some specimens per- sistently failed to expand; some left the stones and were lost; and some underwent division probably as a result, in part, of the change in environment. The errors from all of these sources together undoubtedly render this table useless for any precise statistical analysis. Furthermore, the great variation among different lots indicates that, even if the accuracy of the numbers shown could be depended upon, the total numbers are much too small. Nevertheless, some idea of the relative magnitude of diglyphic, monoglyphic, and other classes may be obtained from the numbers given. As shown in the table, these groups gave 220 DONALD WALTON DAVIS 1101 diglyphic, 112 monoglyphic, 61 triglyphic, and 4 tetra- glyph‘e individuals. It will be understood that all stages of regeneration were represented in specimens of these lots, and that consequently many specimens in early stages of regeneration failed to show siphonoglyphs that nevertheless would certainly develop n the middle of the new region (p. 183). Such potential siphonoglyphs were, of course, counted just as if they were actually completed. Mesenteries No attempt has been made to determine directly the number of mesenteries in numbers of individuals. Remembering the relation shown to exist between mesenteries and orange stripes, the counts of stripes presented hereafter have some significance from this standpoint. Consideration has been given, however, to the number of mesenteries of apparently undivided diglyphic specimens. Of the 1101 diglyphic individuals in the groups represented in table 13, 63 failed to show in the living state satisfactory signs of division. Forty-one of these were sec- tioned and studied for internal evidences of division. Among these, thirty gave unmistakable evidence of the sort already described (pp. 169, 170), five showed slight irregularities such as are commonly associated with fission and regeneration, while six were forms without any irregularities to indicate unlike ages of different parts. Of the six last mentioned specimens, five were biradially symmetrical with six pairs of complete mesenteries, and one had eight pairs. One with six pairs of complete mesen- teries is represented in figure 26. Of the five slightly irregular and therefore possibly undivided specimens, two possessed six pairs of complete mesenteries, two had eight pairs, and one showed ten, all being strictly biradially symmetrical. These numbers are small, and their interpretation must be modified by the fact that, among fully regenerated forms, regularly hexa- meric individuals with twelve orange stripes (and by infer- ence six pairs of complete mesenteries) are more numerous than other forms. It can, at most, be said that, among apparently undivided specimens, a regular form with six pairs of complete mesenteries is the most common type. ASEXUAL REPRODUCTION IN SAGARTIA Zit Orange stripes The factors concerned in the determination of the number of stripes apparent on a given individual at a particular time are such that any mere enumeration of orange stripes in a set in- cluding recently divided individuals is of highly questionable significance. Nevertheless, a study of a tabulation of counts of orange stripes may yield suggestive results and may be made an occasion for pointing out further the effect of processes of asexual reproduction upon the external appearance of specimens. Counts of orange stripes made upon four lots of specimens collected at Woods Hole are given in table 14, and the totals obtained by adding the first three are plotted in figure 42. In figure 42 are given, also, data published by Davenport (’03). If we look at the solid line in figure 42, representing a summary of my counts of July and September (lots 1, 2 and 3), we see that, beginning with twelve, the higher even numbers of stripes are represented by more individuals than the odd numbers. This is in harmony with the greater number of completed regenerations in the classes with higher numbers of stripes. The odd numbers here probably represent chiefly uncompleted regeneration. Be- low twelve, the odd numbers of orange stripes are most abundant. This is particularly true below eight. Among those lower num- bers, recently divided specimens form probably the greater pro- portion of individuals. The predominance of odd numbers among these classes is in all likelihood chiefly due to those speci- mens which have formed no new stripes. With the relative positions of orange stripes and mesenteries of different orders in mind, it is obvious that divisions in two complete endocoels or two incomplete endocoels of the first grade would give usually odd numbers of old stripes; divisions in one complete endocoél and one incomplete endocoel of the highest grade would ordi- narily give even numbers of old stripes; while divisions in other planes might give either even or odd numbers. The relative frequency of divisions in different planes (table 9) is such as to make the expected ratio of odd to even numbers of old orange stripes approximately 4:3. It is possible that during the 222 DONALD WALTON DAVIS progress of regeneration of the stripes there is a tendency further to increase temporarily the proportions of instances of odd numbers. Comparison of the data for the separate lots is rather sug- gestive. Lots 1 and 2, collected July 12 and 18, respectively, differ in that the earlier lot includes a slightly larger proportion of specimens with a very low number of stripes indicating re- peated, rapidly succeeding divisions, while the later set shows a few more individuals with high numbers of stripes indicating more nearly completed regenerations. Lot 3, representing a group collected September 22 from the same place as lot 2, shows a relatively much greater number of individuals with twelve or more stripes. In this group the mode is at twelve with prominent secondary modes at fourteen and sixteen and only lower modes at seven and five. ‘This indicates progress in regeneration with less frequent divisions since the collection in July. Lot X represents a group of specimens selected for their large size. The distribution as shown in the table suggests that the group is composed of specimens that have, for the most part, completely regenerated. Examinations of my records, which show for each individual the number of stripes in areas of all different ages, confirms this suggestion. Davenport’s curve ap- parently represents a group of individuals of which a very large proportion have recently divided. It includes a few specimens with more than twelve orange stripes and a large number with twelve, but the great bulk of individuals show fewer than ten. Davenport gives no indication of the time at which these speci- mens were collected. The large number of individuals with few orange stripes, in connection with my curves, suggests the prob- ability that they were collected early in the summer. Among the groups below ten Davenport’s curve in contrast with mine shows an excess of specimens with even numbers of stripes. A possible partial explanation of this lies in the readiness (p.167) with which the line of fusion of cut edges soon after division may be mistaken for an orange stripe. One stripe thus added in a certain proportion of cases would alter the relative numbers of specimens with odd and even numbers of stripes from the state ASEXUAL REPRODUCTION IN SAGARTIA 228 shown in my groups to that given by Davenport’s series. The large number of recently divided specimens in the latter’s series would make this error possible in a large number of cases. The eases figured by Davenport (’03, figs. 3 to 7) are evidently in too early a stage of regeneration to permit of any difficulty on this point; but it would be encountered in dealing with slightly later stages. If we assume twelve to be the number of stripes typical of specimens resulting from ontogenetic development (see p. 225 for discussion of this point), the large number of specimens with twelve stripes as compared with those having higher numbers suggests the likelihood that there were included in Davenport’s collection an unusual proportion of undivided specimens. In this connection the length of time the species has inhabited the region concerned may be significant. Possibly the entrance of S. luciae into a region and its establishment there is accomplished by migration of larval forms, while the propagation of the species in a region already occupied is much more largely brought about by the asexual method. That asexual repro- duction has been in progress for a shorter time in the group studied by Davenport than in my lots is indicated by the rela- tively small number of specimens with more than twelve stripes, since I have shown that repeated regenerations tend toward the production of an average condition with about seventeen stripes. FORM RESULTING FROM ONTOGENETIC DEVELOPMENT From what has been said it is evident that no form as regards number of siphonoglyphs, mesenteries, or orange stripes cer- tainly distinguishes regenerated individuals from those that have not undergone fission. The development of transforming em- bryos must actually be followed to get unquestionable evidence as to the form resulting from that process. I have made a rather careful search of certain restricted localities at all seasons of the year for such specimens with little success. The dates of my special searches at Woods Hole were January 21, 1905; July 12 and 18 and September 22, 1909; November 28, 1910; 224 DONALD WALTON DAVIS March 31, April 30, May 30, and June 29, 1911. I have had ‘opportunities at various other times for less careful observations of localities in which this species occurs. The only recently transformed anemone I have found is a very small and nearly colorless specimen collected at Woods Hole August 4, 1909. This specimen showed no trace of the green color or orange stripes of the column, or of the white line on the oral disc char- acteristic of S. luciae. Whether it belongs to this species, to Cylista leucolena, to Metridium marginatum, or to still another of the species of anemones found at Woods Hole, cannot be stated. In the absence of color it resembles C. leucolena, but we have no reason to believe that at this stage either 8S. luciae or M. marginatum have developed their characteristic colors. Whatever its proper classification, this anemone was symmetrical and diglyphic and had nearly attained a regular hexameric con- dition of the mesenteries. Each pair of complete non-directive mesenteries, however, had one member incomplete—that mem- ber lying in all cases toward the same end of the chief transverse axis. It is therefore in a stage intermediate between the Ed- wardsia condition and the regularly hexameric form. Even were it determined that this specimen represents a. stage in the development of 8. luciae, it is by no means certain that the attainment of the condition with six pairs of complete mesen- teries would mark the end of its ontogenetic development. This specimen therefore throws no direct light upon the present problem. It does, however, suggest possible directions in which evidence may be found, and the circumstances connected with it indicate some of the difficulties in the way of a complete solution. The apparent scarcity of developing embryos em- phasizes the impression that the extraordinary method of asexual reproduction is the chief and highly successful means of perpetuating this species and increasing its numbers. In Hexactinians six is the most common fundamental number of pairs of complete mesenteries, as two is the typical number of siphonoglyphs. I have shown (p. 220) that such a form is the commonest one among apparently undivided specimens of 8S. luciae. I have shown further (p. 198) that while successive ASEXUAL REPRODUCTION IN SAGARTIA 225 divisions and regenerations in this species tend toward a condition with an average of approximately eight pairs of complete mesen- teries and sixteen orange stripes, the actual averages in the populations studied are less than eight and sixteen. This is presumptive evidence that the form resulting from ontogenetic development possessesfewer than eight pairs of complete mesen- teries and fewer than sixteen orange stripes. It appears from the statements of Davenport (’03, p. 148 and fig. 1) that twelve is a common number of pairs of complete mesenteries. This is clearly an error. Among several hundred sectioned specimens I have found but two with twelve pairs. These are both tri- glyphic individuals and both show clear evidence of having divided. I have not seen a single biradially symmetrical speci- men with twelve pairs of complete mesenteries. Such a speci- men would normally have twenty-four orange stripes, whereas the highest number recorded by Davenport or myself is twenty- two. It seems likely that Davenport mistook incomplete mesenteries of the first order for complete ones. This is the more probable since she indicates that the specimens described as having twelve pairs of complete mesenteries have twelve orange stripes, the number which I have shown to be character- istic of a form with but six pairs of complete mesenteries. Con- sidering all the evidence at hand, it is probable that the sexually produced form is a diglyphic one with six pairs of complete mesenteries, six pairs of incomplete mesenteries of the first order alternating with these, and twelve orange stripes occupy- ing these two sets of endocoels. Specimens of this form may also be produced by regeneration following fission, and all normal departures from this typical form are due to asexual reproductive processes. SUMMARY Sagartia luciae is typically Hexactinian in form and structure except for the wide variation in number of siphonoglyphs and of pairs of mesenteries. The number of siphonoglyphs and of associated pairs of directive mesenteries varies from one to five. The number of pairs of complete mesenteries including the 226 DONALD WALTON DAVIS directives varies from five to twelve. Three or four grades of incomplete mesenteries may be found (p. 164 and tables 11 and 13). ~ Asexual reproduction in this species occurs by a process of aboral-oral fission with subsequent regeneration (p. 167). In this process movements of parts of the basal dise in opposite directions are initiated, centering in two or more isolated re- gions. These movements place the intervening tissues of the base under strain and result in a rupture of the basal wall. The rent progresses until base, column and esophagus are suc- cessively involved, and complete separation of the individual into two or more pieces finally results. Details are recorded of the divisions of one abies one tetraglyphic, three triglyphic, and seventeen diglyphic speci- mens (p. 174 and tables 3 to 5). The resulting pieces may possess one or more siphonoglyphs and associated pairs of directive mesenteries or they may lack these structures until regeneration has occurred (pp. 174, 175). Sueceeding divisions may be delayed until regeneration is completed or they may occur at any earlier time. They may even follow so rapidly as to give the appearance of simultaneous division into more than two pieces. In every such case a re- generation zone is formed in each piece for each successive fission (p. 173). In passing upward the plane of fission rarely cuts a mesentery, i.e., the plane is a strictly vertical one (p. 171). There 4 is no tendency toward strict equality of the products of a division (p. 181). Strictly diradially symmetrical polyps tend to divide into parts which are themselves symmetrical with respect to the original directive plane, i.e., the plane of division tends to be approxi- mately perpendicular to the major axis of the mouth (p. 176), but no tendency to divide in spaces of the same kind on opposite sides of the column was detected G cme Weare While division may occur in any vertical plane, it tends to occur in endocoels rather than exocoels (p. 179), in complete endocoels rather than incomplete endocoels (p.180), and in non-directive rather than directive complete endocoels (p. 180). ASEXUAL REPRODUCTION IN SAGARTIA 227 No obvious change occurs in the old part in consequence of division, except that possibly mesenteries injured in the process of fission are eliminated by absorption (pp. 168, 171, 172, 200). Regeneration processes begin with the rolling in and fusion of the torn edges of the body wall (pp. 167,182). Inthe region of fusion new structures are gradually differentiated, eventually constituting a large proportion of the bulk of the individual— often far the greater part (pp. 167, 182). The torn edges of the esophagus also grow together and a new siphonoglyph invariably becomes differentiated in the region of fusion of these edges. As regeneration proceeds, the new siphono- glyph occupies the middle of the new region, thereby marking this as a new directive plane (p. 183). In those instances where a siphonoglyph is cut by the fission plane (p. 183), a siphonoglyph occupies the corresponding boundary between old and new parts of the regenerated animal and a wholly regenerated siphonoglyph is formed in addition. Origin of specimens with different num- bers of siphonoglyphs from any of the common types is com- pletely explained by the manner of division and regeneration (p. 183). Four new mesenteries, constituting a very characteristic group, become established in the middle of the new region of the column and grow across the oral disc to the esophagus (p. 193). Longitudinal muscle swellings appear on each of the first four mesenteries on the side away from the directive plane (p. 194). Additional mesenteries follow in a bilaterally paired manner lateral to the first set of four (pp. 193, 198, 207). Among these later mesenteries certain members mate with the two outer ones of the set of four, forming unilateral pairs. The two inner members of this set of four constitute a pair of directive mesen- teries. Other of the later mesenteries become paired with the old bounding mesenteries or with each other, so that eventually all of the mesenteries, with the exception of the directives, are present in the unilateral pairs characteristic of the Hexactinians (pp. 193, 207). Variations in the number of mesenteries formed depend almost wholly upon the character of the old bounding mesenteries. On 228 DONALD WALTON DAVIS the side of a new directive plane which is toward an old incom- plete bounding mesentery are usually produced two pairs of complete non-directive mesenteries. On the side of the new directive plane which is toward an old complete non-directive bounding mesentery are commonly formed a pair of complete non-directive mesenteries and a single complete mesentery which forms a non-directive pair with the old bounding mesentery (pp. 186, 188). The only common exception to the numbers of complete mesenteries as stated consists in the omission of a pair of complete mesenteries lateral to the new directives (pp. 187, 188, 191). The number of complete mesenteries formed in a regenerating zone is thus strictly limited and almost invariable, except through the influence of the old mesentery adjacent to the boundary between old and new tissue. This mesentery, if unpaired, exerts a perfectly definite determining influence upon the course of regeneration, an influence which makes its appear- ance soon after regeneration has begun and effects a normal pairing of the mesenteries of the bounding region and usually a regular arrangement of the pairs of different cycles in harmony with those of other regions (p. 193). The order in which the new mesenteries appear and the order in which they become attached to the esophagus (pp. 193 to 188, 201; see especially pp. 193, 196) do not correspond, and neither agrees with the order of ontogenetic development described for any Actinian whose transformation has been completely followed. At a late period of regeneration a stage is passed through corre- sponding with that described as a stage in the ontogeny of Adamsia by Hertwig; but this is probably correctly interpreted by Carlgren as a stage in regeneration (p. 194). In the course of regeneration incomplete mesenteries appear in pairs in the normal positions (p. 194). Where the old bounding mesentery is an incomplete one a single new incomplete mesen- tery of the same cycle is formed to pair with it. When division has occurred in an incomplete endocoel of the second or lower order close to a pair of complete mesenteries, no new incomp!ete mesenteries of higher cycle are formed in the region limited by the complete mesenteries and including the boundary between ASEXUAL REPRODUCTION IN SAGARTIA 229 the old and new parts. Consequently a pair of incomplete mesenteries of the first grade and its accompanying orange stripes are sometimes lacking in a region where they ordinarily occur (p. 195). Certain other irregularities in the occurrence of incomplete mesenteries are found (p. 196). The orange stripes seen prominently on the living specimens normally occupy the complete endocoels and the incomplete endocoels of the first order (p. 207). Since spaces of these grades almost invariably alternate regularly, the orange stripes are commonly present in even numbers in individuals that have not divided or have completely regenerated following a division (pp. 114, 122). When a fission plane passes through an orange stripe, that stripe disappears (p. 209). As regeneration proceeds orange stripes are formed in their characteristic positions, including the bounding endocoels, provided these are not incomplete endocoels of the second or lower grade (p. 209). Probably orange stripes arise, in the new area, first in the complete endocoels and only later in the incomplete endocoels of the highest grade (p. 215). The number of orange stripes is strictly limited by the number of mesenteries of the first two cycles, no new stripes being formed either in the old tissue or in the new tissue after the characteristic spaces are occupied (p52. lG)4 Odd numbers of orange stripes are usually to be explained on the basis of incomplete regeneration, occasionally on the ground of irregularity in the cycle of mesenteries of the highest incom- plete grade Certain cases of unusual conditions of orange stripes remain unexplained (pp. 195, 210, 2138, 221, 222). The number, position, breadth, and color of the orange stripes of any individual give significant indications of its internal con- dition (pp. 215 to 217), but mere enumerations of orange stripes are of little value. The numbers of stripes in individuals of groups examined by the writer varies from zero to twenty-two, the plotted curve showing modes at seven and twelve with a tendency toward minor modes at even numbers above twelve stripes and at odd numbers below ten. This is interpreted as 230 DONALD WALTON DAVIS an indication that among individuals with higher numbers of stripes a preponderance of specimens have completely re- generated; while among the lower numbers a great proportion have recently divided and regeneration of stripes has not begun or has at least not been completed. Such counts as have been made further indicate with some degree of probability that divisions are more frequent in the spring and less frequent toward fall (p. 222ff.). By the processes of regeneration described, a rather definitely fixed set of structures is added to a piece resulting from division, regardless of the form of that piece (pp. 185 to 192). The outer- most mesenteries of this set are modified to match up with the old bounding mesenteries. Since the forms of old pieces are quite various (p. 181), the resulting individuals are likewise diverse in the number of siphonoglyphs, mesenteries and orange stripes (pp. 184, 192). With rarest exceptions the variations from the typical Hexactinian form may be accounted for on the basis of the processes of asexual reproduction described. Re- peated fission and regeneration tends to produce specimens with approximately eight pairs of complete mesenteries and sixteen orange stripes (pp. 198, 215). The fact that actual counts of mesenteries in specimens of the groups studied show a distinctly lower average than this, indicates that the sexually derived form possesses fewer than eight pairs of complete mesenteries. The fundamental form of the species, i.e., that resulting from ontogenetic development, has not been ascertained. It is prob- ably biradially symmetrical and hexameric, with two siphono- glyphs (pp. 218 to 220) and six pairs of complete mesenteries (pp. 220, 221), two pairs of which are directives. Sexually mature individuals of both sexes have been found, but the development of the fertilized egg has not been followed. The failure to obtain developmental stages, together with the abundance of stages in asexual reproduction (p. 220ff.), suggests the probability that the latter constitutes the chief means of maintaining or rapidly increasing the population of a given region. 5 College of William and Mary, August, 1917. ASEXUAL REPRODUCTION IN SAGARTIA 231 BIBLIOGRAPHY AnprEs, ANGELO 1882 Intorno alla scissiparit& delle attinie. Mittheil. Zool. Sta. Neapel, 3, pp. 124-148. Boun, Georces 19.8 Scissiparité chez les Actinies. Compt. Rend. Soc. Biol. Paris, 64, pp. 936-939. CaARLGREN, OskaR 1893 Studien iiber nordische Actinien I. Kongl. Svenska Vet. Akad. Handl., no. 25, 10, pp. 1-148. 1904 Studien iiber Regenerations- und Regulationserscheinungen. I. Uber die Korrelationen zwischen der Regeneration und der Sym- metrie bei den Actiniarien. Kongl. Svenska Vet. Akad. Handl. no. 37, 8, pp. 1-105. 1909 Studien iiber Regenerations- und Regulationserscheinungen. IT. Ergiinzende Untersuchungen an Actiniarien. Kongl. Svenska Vet. Akad. Handl. 43, no. 9, pp. 1-48. Cary, Lewis R. 1911 A study of pedal laceration in Actinians. Biol. Bull. 20, pp. 81-108. DavENPORT, GERTRUDE Crotty 1903 Variation in the number of stripes on the sea-anemone, Sagartia luciae. Mark Anniv. Vol., pp. 137-146. DicquUEMARE, J. F. 1775 A second essay on the natural history of the sea anemones. Phil. Trans. Roy. Soc., London, 65, pp. 207-248. DuerpEN, J. E. 1902 Relationships of Rugosa to living Zoantheae. Ann. Mag. Nat. Hist., ser. 7, 9, pp. 381-398. Gosse, Puitirp Henry 1856 Tenby: A seaside holiday. London, John Van Voorst. 1860 Actinologica Britannica. London. John Van Voorst. Gossr, Puitip Henry 1865 A year at the shore. London. Alexander Strahan. ak Hammart, M.L. 1906 Reproduction of Metridium marginatum by fragmental fission. Amer. Nat. 40, pp. 583-591. HaraittT, CHartes W. 1914 Anthozoa of the Woods Hole Region. Bull. Bur. Fisheries, 32, pp. 223-254, pl. xli-xliv. Hertwic, O. unp R. 1879. Die Actinien. Mayer, ALFRED GoLpsBoroucH 1905 Seashore life. New York Aquarium Nature Series, I. New York Zool. Soc. McCrapy, JouHn 1858 Instance of incomplete longitudinal fission in Actinia cavernosa Bose. Proc. Elliott Soc. Nat. Hist. of Charleston, 8. C., 1, pp. 275-278. McMorricy, J.P. 1910 Actiniaria Siboga Exped. Pt. I. Ceriantharia. Parker, G. H. 1897 The mesenteries and siphonoglyphs in metridium mar- ginatum Milne-Edwards. Bull. Mus. Comp. Zoél. Harvard Coll., 30, pp. 259-273, 1 pl. 1902 Notes on the dispersal of Sagartia luciae Verrill. Amer. Nat., 36, pp. 491-493. TuyNnne, Mrs. 1859 On the increase of madrepores. Ann. Mag. Nat. Hist., (3) 3, pp. 449-461. Torrey, Harry Beat 1898 Observations on monogenesis in Metridium. Proc. Cal. Acad. Sci., (3) 1, pp. 345-360. 232 J DONALD WALTON DAVIS Torrey, Harry Brat, AND Mery, JANET RutH 1904 Regeneration and non- sexual reproduction in Sagartia davisi. Univ. Cal. Pub., Zoél. 1; pp. 211-226. VERRILL, A. KE. 1869 Our sea-anemones. Amer. Nat., 2, pp. 251-262. Yue, G. Upny 1900 On the association of attributes in statistics. Phil. Trans. Roy. Soc., London, (A) 194, pp. 257-319. SYMBOLS USED IN TABLES AND FIGURES The following symbols appear as subordinate headings under Old and New Mesenteries in tables 3 to 8: B, bounding mesentery, i.e., a mesentery adjacent to the boundary of the new or old region in which it lies and without a mate in that region. The mesentery occupying this region is represented whether incomplete or complete. D, directive mesenteries, which occur in pairs. N-D, complete mesenteries which are not bounding mesenteries nor directives. With rarest exceptions they are found in pairs. The symbols below are used in table and figures to designate specific mesen- teries. In bold face they represent mesenteries formed during ontogenetic develop- ment or in the oldest regenerating region distinguishable in that specimen. In italic type they indicate mesenteries regenerated after the recorded division. In roman type they represent mesenteries of intermediate age. Some of these may have been incompletely developed at the time of the latest division. d, a directive mesentery. c, a non-directive mesentery, actually complete or destined to become so. c!, c?, c3, ef, regenerated non-directive mesenteries, actually or potentially complete, whose positions with respect to the new directive plane are indicated in the figures by exponent numerals. c?, a mesentery now complete, but possibly derived from an incomplete one through abnormal fusion with the esophagus. It is consequently doubtful whether this was complete when fission took place. (I), a member of the most advanced cycle of incomplete mesenteries repre- sented in an individual. (II), (III), members of the second and third cycles of incomplete mesenteries found in the specimen. (1), a permanently incomplete mesentery of undetermined grade. (1)?, an apparently incomplete mesentery which was probably derived from a complete one by being torn from the esophagus. * symbol indicating absence of any unpaired mesentery adjacent to the boundary between old and new parts, i.e., the boundary lies in an exocoel. + symbol used in table 5, no. 18b and no. 22b (and in table 10 referring to the same specimens) to designate the bounding conditions when a second fission plane has passed through the boundary between old tissue and tissue regenerating following a shortly preceding division. Four, five, etc. (tables 3, 5, 6), number of small mesenteries in a regenerating region at a stage of development too early to indicate the final formula. ASEXUAL REPRODUCTION IN SAGARTIA 233 GENERAL EXPLANATION OF TABLES 3 TO 8 Mesenteric formulas of regenerated polyps are here represented in tabular form, but in a manner which has to some extent the significance of a diagram. In the first column of each table are given numbers designating different indi- viduals or groups of related individuals. Letters in this column refer to different regenerating animals derived from a single anemone by fission. In the columns included under ‘old mesenteries’ are indicated the complete mesenteries and incomplete bounding mesenteries of the old part which persist without obvious change. Similarly, under ‘new mesenteries’ are represented mesenteries of the region formed after division. In all cases where regeneration is sufficiently advanced to give conclusive evidence of the final formula of the regenerated region, this is given indetail, except that definitively incomplete mesenteries are represented in the formulas only when they occur as bounding mesenteries. In eases of earlier stages only the number of new mesenteries present is indicated. The formula of each individual at the time it was killed is given between two adjacent horizontal lines. With the exception of a few complicated cases, specially explained in connection with table 5, the arrangement of complete mesenteries in any specimen whose mesenteric formula is given in detail can be readily pictured if, reading from left to right, one imagines the indicated mesen- teries distributed in pairs around the body of the anemone with the last of the new mesenteries adjacent to the first of the old ones. Thus figure 9 shows a cross-section of a specimen having the same arrangement of complete mesenteries (disregarding the distinction between old and recently regenerated parts) as represented in the formula of no. 7b (table 3) after regeneration. Similarly, figure 18 may serve to illustrate the arrangement of complete mesenteries of no. 7a after regeneration, and figures 19 and 21 of no. 7 before division. In the following three columns are given respectively the number of mesen- teries in the regenerated region actually reaching the esophagus at the time of killing, the number of days during which regeneration had proceeded, and the numbers of the illustrative figures. Specimens represented in tables 3 to 6 may be fairly regarded as constituting a small random sample of the specimens found near Woods Hole, Massachusetts (pp. 170,171). Those represented in tables 7 and 8 may not be so considered, but are useful as showing further instances of regeneration following division of the types they represent. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 TABLE 3 In this table are given the mesenteric formulas of fourteen pairs of polyps resulting from natural fission and regeneration. for interpretation of the formulas, page 233. larities shown in table 3 to 5, see pages 171 and 172. i 1 9 i 4 5 6 7 rhs 8 9 a. 10 i {2 For explanation of symbols, see page 232; For a discussion of certain irregu- 234 OLD MESENTERIES NEW MESENTERIES eco crs FIG- Bo) ox) pene ee Nel |onep fs Cleta| Pays es c cc dds \cc,cc | (ID) four 0 | 20 c ce,ec | dd} ce,cc | (II) Six ON 20 67 (@)3|Gecsee. addti cc,ec"}) 2) four 0} 9 (I) cc dd /|cc,cc} (I) four 0 | 14 (Dalcereceiieddiiecesce.| (DL) twelve very small Op tO teas (it) cesce dd c four very small o | 10 a aCH) dd CG | Sces| ides iec.ceo| (i) aaa coll (Ey ieccrec Weddaltec ces (yin) neen laddalicerco alec 3 | ol (I) | cc,ce,ec| dd | ce,ec | c @ || ce || Ga Ge 6 | 50 @)ice,ces)-ddi| ce c € | €e4\ da"| cerces| (Gn) 9" e50 (II) dd c @ || Ge \l ole) || eexee || CHD), & |) 4 (II) | ce,ce | dd} ce Cc GlnGelndadm|ecerccu| "Cae Gn eg c cc ddiinicc c @ | @e | cal || ee 8 | 45 Cc dd Cc Can CGaInG Ca |aRGG 8 | 45 (Ds fcexceaitdd’|| ec CU ilye dd) ce. |) 5.256 (I) cc dd Cc GulGculnddalance (J) a\eoom| eae c dd c four very small 0 | 24 Cc ce,ce | dd | ce,ce | (1)? siz small 0 | 24 Qs ccsecm| dd c | ¢€ | ce | dd | ce,cc Sales CCSCCmsdGulcG Cc CleGCm RaGe\nccacom |) 3 | 34 (ye dd c Gan eCCaladalnccecculan() 8 | 26 c cc ddujimicc c @ | @@ | ake || ee c 5 | 26 T) cc dduiimice c seven small 4 | 20 Dy Scexcems dd! ee,ces\ecc six very small 0} 20 c dd eight small 0 | 20 c ce dd ice eight small 0 | 20 cc ddaiiicc c two O) || = dd c two () | = ASEXUAL REPRODUCTION IN SAGARTIA Dao TABLE 4 In this table are shown the mesenteric formulas of two additional natural pairs. No. 15 showed some evidence of a preceding regeneration, but the limits of the regenerated region were not clear. OLD MESENTERIES NEW MESENTERIES HURLED AS TION NO | y y |B|N-D| D|N-D| D | N-D] D B |B] N-D| D | N-D| B ee Days 15 fan e | ce | dd.| ce~| dd icc,ce| dad) (i) COrCElidds Geel Gull an50 ae c |cc,cc| dd |cc,cc a) weluce.cc| dd iii ce |e |) S150 1622-7 2 ICC Ke (III) | cc,cc| dd | cc,cc 6 |22+ b...| * |cec,ce| dd |ec,ec| dd (III) cc¢,cc| dd | cc.cc 6 |22 TABLE 5 This table presents the mesenteric formulas of six groups of polyps, each group derived from a single individual by natural fission. The number of zones of regeneration indicates that the multiple divisions involved are successive rather than simultaneous. Certain regenerated polyps show two new areas. In these cases only bounding mesen- teries are represented in the old region on the same horizontal line with the new mesenterics. Two such specimens (nos. 17, 19) have a pair of complete mesen- teries in the old region between the two division planes. This pair is represented in the table in a horizontal space interposed between the lines showing the bound- ing mesenteries. To arrive atthe complete formula of a polyp with two regener- ating regions, the symbols of the first new area must be read, in reverse order, between the extremes of the old area represented on the same line. Similarly, the second new area must be inserted in reverse order between the limits of the old ‘region shown on the same line with vt. This may be most conveniently illustrated by no. 19b, of which a section is represented in figure 32. No. 19a is repre- sented in figure 27, no. 19d in figure 30. The mesenteric formulas (disregarding incomplete mesenteries and distinctions between old and new mesenteries) of this and the other available fully regenerated products of division of no. 19 may be represented as follows: NOMS COG GOW CG GGGG CGNugeaen es NGI DMO, “CO CCACGCONCG dm CONCE MCG, CC LOE, oe) No. 19d _ dd, cc, cc, dd, cc, cc (fig. 30). No. 19 was described before division as having two white bars on the oral disc and twelve orange stripes. Its formula was probably, like that of the specimens shown in figure 25, as follows: dd, cc, cc, dd, cc. cc. The lost part was very probably like No. 19c in having no old siphonoglyph or directives. No. 20 was triglyphic before fission, as determined by the white bars. The lost part must have possessed two old siphonoglyphs and pairs of directives. Daggers (+) serve to call attention to boundaries in two cases, Nos. 18b and 22b, where apparently a second plane of fission passed in the plane of junction of old tissue with new tissue regenerating after a shortly preceding divison. For dis- cussion see page 188. 236 NO. 20) b ! | 21) b Cc. aN SS ee eS a an DONALD WALTON DAVIS TABLE 5—Continued OLD MESENTERIES D | N-D dd| cc cc cc,cc,cc| dd cc Ge5cc cc,cc dd dd dd dd} ce,ce dd| ce,ce NEW MESENTERIES ago eae B B |N-D|D|N-D| B Ls ES on| 8 oO QA c c ce | ddj| cc,cc 2| 13 c c cc | dd\ cc,cc| (1) | 4 | 18 c four On| c four 0| 6 c c Ce Nhl Te |\| (UO) Ne 42 GLC) veo" kdl | dd cc,cc Calan ech) dan icc Cues 35 | Cc dd c six very small 14 36] c dd c c six very small Sieh. Le dd CoG) Cm ee, .cc ae) 38 | (I) dd cc,cc 3 ce,cc | dd | cc 2 39] c ce | dd CC,cCc Gules cow aa.) .ce cus 40 | c | ce,cc | dd cc,ce ddi'c e>\ eco da’ ce c | 8} 30 Ai ences |-CC.cC Gd) sce, cc.ce c two very small 42/1 c ce | dd cc,cc dd) t'e."\-¢..|) ce || -@a"\ cc CAs 43 | (1) ce Eonen| (cose daainces Gn) 4|' va 40 44) c d | d | cc,cc| dd| cc c 16 4550) sicc,ce.| dd cc Cl alle NeGih || ae || (ED) Net 46 | (I) | ce,cc | dd cc,cc (I) two 3,4,5 47 | (I) dd cc C6 | eel nad. \ece We co 48 | (I) cc |dd CC Clelncalece nad |ceml|n i) malat 491 (11) | ice ‘|. dd| ““cc,ce (1) four 50 | ¢ dd ce c|e| cc | dd| cc | ¢ |8 1 ae dd c eight + Be ce} ce \dd! “-ce,ce dd) c |}c]| cc | dd| cc | c |8 5a |) CL) | ee,cc | dd c siz small 54 | (II) dd + | cc,ec] dd | ce,cc| (IZ) |10 237 238 DONALD WALTON DAVIS TABLE 7 This table shows the mesenteric forms of thirty-three specimens. Some of these were selected, in collecting, on the basis of the position of the new area or on other grounds that prevent their consideration in connection with some maiters for which specimens given in tables 3 to 6 are available. Others were excluded from table 6 on grounds which have been stated in connection with that table and in the text (p. 169 ff). Still others are animals regenerating after artificial cuts. In the case of these last, the length of the regeneration period is given in the last column of the table. Since the period of regeneration ts unknown for all other specimens listed in this table, the cut specimens may be identified by the presence of figures in the last column following their formulas. Dashes in the columns for Old Mesenteries indicate that some part of the old material present at the time of the recorded division was removed by a later division or cut and the full formula of the old part is therefore unknown. In the old regions of these specimens only bounding mesenteries are ‘recorded. The case of No. 80 is discussed on page 211. Attention should be called to the fact that a new mesentery apparently of the first incomplete cycle mates with an old one of the second incomplete cycle. It is clear that no orange stripe was formed in the included endocoel. OLD MESENTERIES NEW MESENTERIES mod NO. Cie ee oan) ce PE in poe lane ea we ra hoe FIGURE B | 32D |p|] ND [aD |B | Bol NED. | Del BD hie ee] 2 onl 3 oO 1A a Gc) \)dda) cc c c Coys saayl! occ ¢ |'8 Ke dd c c Coma tadalice Chola VI dd c c CChadGa|\ece c |8 98. tC CC) | Adal sice Cc c CCONIRGGalmece cc |8 59 | c | ce,cc | dd | ce,ce Cc c Cena ada\ieree c | 8 60") %c — | == —— | I Cc Wyteseh || Be (taal Oi 61 Cc — |j—| — | — c Cc cc das) (ce c 4 6OZiE ac — |/—|{| — |—! c Cc COou| Goa|ence CL ono 63"|) xc — |—|]— |—! ¢c Cc COMM Gdn mntce B |\s 14 64] c — |—|/] — |—! c Cc Coun Gaulaec es 15,16 G5nimuc dd c Cc CON Ra aleece Colne 66 | (1) Go dau icc Cc c oe Wale) Il sae (A) | 12,138 67 4) cc | dd c c cc' |dd| ce CA 68 | (I) dd |cc,ce) dd| ic c COME alte (J) | 5 | 56 69 | (1) re | Rw (eg a c ce ad. \ cence 1) ol, 70 | (1) — |—| — |/—]| ¢ c Cet, |idds| cee. Nh) ale (2) )- \see,ce | dd. W) occ c c CCpalvagalpecc GAY ire WN hs ec. f|\'dd \'ce,cec'|.dd |) @)-1 (1) 4 cece)! deal vce 6 9 (BC ddijiace (I) | (@) COMM GGamCC haa |\ea, 19,38 74| (I) | ce c c co |dd} cc | QZ) }7 | | ASEXUAL REPRODUCTION IN SAGARTIA 239 TABLE 7—Continued OLD MESENTERIES NEW MESENTERIES ion NO. Te | cae) Apes ee ll a RI PP (a paees FIGURE Biwi eNeD. | Doi) Netnele Eh aly Be MND. | ND | 4B 4 be 2 8714 hee eck) rec,ce |dd i'ce,cc | Cys Cy" Wee. ec" dd" |"ed,ce |") 10 76 | (I) — |— | — |}— 1 OD ce aa | ce (1) | 6 TH 08) — f= |G) | ee, ce || dd | ec;ee | () |10 78 | (I) — |—-}] — |—!|] @ IW | ce,ce | dd | cc,ce | (Z) {10 79) (Ly ‘ce (I) | ~~) ce | dd | cc,ce | (1) | 8 | 49 80 | (II) dd Gai Ch), ec.ce’ aa | cece (1) 110 37 81 | (I) | ce,ec | dd | ec,cc Cy Chyr Reece: da \cece FZ) 10 15,16 82 | (II) CUE)“ CERY Weevce da \'cevec™) CITY 140 37 Somer ce yiCG.cc a CONCCA AG |\nGC.CG 10 | 56 Saepree d @ \ee.ce laa | 'cc.cc 11 | 49 TABLE 8 This table records the mesenteric formulas of five specimens with apparently unusual - regenerated parts. These were excluded from table 7 only for convenience. For description see page 190. A possible partial explanation of these irregularities as given on page 190. OLD MESENTERIES NEW MESENTERIES Seas NO | () By} oD |) NED.) Dy |) B | Bolo NED) Oe ene: |) D | cNeDe |B ee 2 om S 6) A So. | dd |icc,ce d @ | Ce,ee 5?| 49 yoo * (dd hecicc | dd |. “e c cc ? | ce,cc 7 | 54 Sr | + + dd |\ec,cc'| dd’ | ** cc,cc | dd?'| cc,cc OP 88 | * cc Ch i) cc dd | cc,cc | dd 10 | 56 89:1) * cc c c ce dd | cc,cc | dd | ce,cc 15 | 60 240 DONALD WALTON DAVIS TABLE 9 This table summarizes the facts presented in tables 3 to 6, on the left with respect to character of the spaces cut by the division plane, and on the right concerning the total number of complete mesenteries formed in the corresponding regenerating regions. On the right, specimens of table? are included, but those of the preceding tables which had not regenerated sufficiently to show their final mesenteric formulas are necessarily omitted. Only those cases are included which show clearly the final formula of the regenerating region, but they are included whether or not the potentially complete mesenteries were actually complete at the time of killing (p. 186). On the left are indicated the classes into which the regenerating regions are grouped on the basis of the old bounding mesenteries. The ‘frequencies’ here given are the numbers of cases in each class found in tables 3 to 6, and may be taken to rep- resent fairly well the proportions in which divisions of the different classes naturally occur. (See p. 177 for discussion.) Figures in italics in the right half of the table represent for each class, as given on the left, the frequency of the numbers of complete mesenteries (indicated at the heads of the columns) found in tables 3 to 7. NUMBER OF COMPLETE | MESENTERIES (5-11) 4 WHICH HAVE BEEN OLD BOUNDING MESENTERIES A REGENERATED g Q = | 5] 6] 7] 8] 9} 10] 11 r lete TWO MON=GITCCHIVES: 706. eer ee 28 2 26 delay as Fear One directive, one non-directive..| 4 1| 3 One complete......... Non-=directived: csc.sc} cocci seioen B24 Dh Wy ee , {One directive, one incomplete non- One incomplete..... : : AO GITECHIVEs 20 drcdu et Etna ame 1 Two incomplete non-directives..........-.-..eeceeeeres- 9 1 3 a One complete non-directive, one doubtfully incomplete. .| 3 1 : One complete non-directive....... if 1 No unpaired mesen- : ; t eas Onerdirectivenjcccs leone ee oes 0 1 Se ae ‘** | One incomplete non-directive..... 5 i 4 Unpaired mesenteries lacking on both sides............. 0 1 One complete non-directive....... 1 1 One non-directive doubtfully com- (See Wate ites). Sneacis d Suplete.4 icine cote oe Sept eee 1 1 @Onevdirectives.cokor eee eee 1 One incomplete non-directive..... 1 1 Notre. On one side a new division has passed through the boundary between old and regenerating regions (p. 189). TABLE 10 This table deals, in its right and left halves, with the same material as the corre- sponding parts of the preceding table; but in the present instance each boundary and associated half of a regenerating zone is considered separately. On the left side of the table are summarized the facts represented in tables 3 to6in so far as these concern the character of the spaces through which the planes of division pass. The symbols in the column headed ‘B’ are the same as those in the columns of bounding mesenteries in those tables and have the same significance. In the column headed ‘Frequency’ are given the number of instances of the bounding mesenteries indicated on the left recorded in tables 3 to 6, and the swm of these frequencies is the total number of symbols inthe columns of bounding mesenteries in those tables. Since each old bounding mesentery is separately considered here, the sum of the numbers in the column headed ‘Frequency’ is twice the sum of the numbers found in the corresponding column of table 9. Omitting doubtfully complete endocoels, doubtfully incomplete endocoels, and the four examples indicated in the last row, we find the ratios and percentages of locations of the fission planes to be as follows: complete endocoels, 103 (62 per cent): incomplete endocoels, 57 (34 per cent): exocoels, 6 (4 per cent). The ratio of complete to incomplete endocoels, is 103: 57, or 64 per cent to 36 per cent. On the right are given the frequencies of the numbers of complete mesenteries (at the heads of the columns) regenerated between the new directive plane and the bounding mesentery indicated at the left of the rows. This part of the table includes speci- mens from table 7 in addition to those of tables 3 to 6. Of necessity, it excludes such specimens as did not clearly show the final number of regenerated mesenteries. Because of these inclusions and exclusions the number of individuals repre- sented in the two halves of this table are not the same. The total number of re- generated mesenteries represented in this table is the same as in table 9. The average number of mesenteries added in the course of a division followed by regeneration may be obtained as follows: take the average number of mesenteries indicated in each horizontal row of the right half of the table (omitting doubtful rows) and multiply by the frequencies given onthe left; add these totals and divide by the sum of the frequencies (omitting doubtful ones). This gives 4.2 as the average addition for a lateral half of a regenerating zone, or 8.4 per regenerating area. Note. The four cases, in the last row, designated by a dagger refer to no. 18b and no. 2b, table 5, in which specimens two successive fission planes have apparently cut away first one and then a second old bounding mesentery. (For fuller ex- planation see p. 189.) COMPLETE ate MESENTERIES POSITION OF DIVISION PLANE B QUENCY REGENERATED 2|3|4|5/6 Non-directive........... c 97 | 3) |93 Complete ..4 Directive............... d 6 1|\ 4 Doubtfully complete....| ¢? 1 1 epee First grade..............} (1) |. 42 | |24] |22 Second grade............] (II) 8 10 Incomplete. ; Third grade.......... (IIT) 4 4 Undeterminable grade...| (1) 3 3 Doubtfully incomplete. .| (1)? 3 1 TEXGCOCIS Eee re Ps one etd oe ee ac rae < 6 9 T 4 2| 1 (SCETTOTE VR sete ee ee een ere ee TABLE 11 This table shows the number of mesenteries reaching the esophagus in individuals represented in tables 3 to 6 upon completion of the recorded regeneration. Speci- mens are distributed according to the number of siphonoglyphs and the number of complete mesenteries. The figures in the body of the table represent numbers of specimens. Average numbers of mesentertes are given for each siphonogly phic class and for the whole. The triglyphic specimen with nineteen complete mesen- teries had five directives, one lacking a mate. NUMBER OF COMPLETE MESENTERIES AFTER REGENERATION AVERAGE MESENTERIES 10 11} 12} 13] 14} 15) 16} 17) 18} 19} 20] 21) 22) 23) 24 Monogly phic. ..¢.05. -6< pistons 2 1 11.3 Diglyphiess)..065 «5 sh areeeah-el Nita) mola ASI uo 2 3 15.0 ari pdivphaier.s cate: eos ease ae 1) 1) 2) 1/3 20.5 Tetraglyphite fib. 4) caer a 1 24 BH OUAIS. J.75 nei eRe a: 4) |14 6) 1) 8} 1) 6 1] 4) 1) 6 1 15.8 TABLE 12 This table is taken, with some modification, from Carlgren’s table (’09, p. 40) sum- marizing the results of his study of regenerating anemones at stages when the order of development of the mesenteries could be determined. Roman numerals at the head of different columns represent the types of arrange- ment indicated by corresponding numerals in figure 35. In the horizontal rows are indicated the number of cases of each type found in regenerating pieces of different kinds. Carlgren used, in this study, artifically cut pieces of Sagartia viduata; natural fragments of Metridium dianthus, and pieces similar in character artificially separated from the parent polyp; and naturally produced basal fragments of Aiptasia diaphana. Some pieces of Sagartia and Metridium were cut in such a way as to leave only basal tissue to form the new polyp. These are designated as ‘basal pieces.’ In all of the ten cases of Type VIII in Sagartia the new directive plane is perpendicu- lar to the plane passing through the middle of the piece before régeneration. In at least some of these specimens (and I presume in all, since otherwise the relation of the new directive plane to the old middle plane could probably not be determined) some old tissue was present, on one side, between the mesenteries I have desig- nated as c? and c8. The single ecample of Type II in Aiptasia shows a second esophagus. Type VIII is the one found by Cary in three species of Aiptasia and in ‘Cylista leucolena’ (?). Regenerating specimens of S. luciae, when old complete bounding mesenteries are absent, are of Type III, V, or II. DABS ONTO Ne CIAO EN OU by D-<}p><¢ iSagamtaay.! oi vee 3) Sly 26| 48] 15) 7 |2+1? 10 Eragmenta Metridium, natural...... 1 1D 4 Metridium, artificial.... 3/9 + Alptasia. }. . Seahacknseee 1 21 DAMATUAY NS eke a 65 poll Be fae bi 1 2 Metrics). ee eee 2 | bo — p_ Basal pieces { ASEXUAL REPRODUCTION IN SAGARTIA 243 TABLE 13 This table indicates the numbers of specimens in lots collected at different times and from different localities, all at Woods Hole, Massachusetts. The specimens are classified according to number of siphonoglyphs. For an account of the methods of collecting these specimens and recording the data here given, see page 218 ff. LOT DATE LOCALITY ok EI z ak CI eo 2 |aels | 6 a A/a |a I dy rebaoo, aly: 12. e200. amplis lan Genser tesce cise acc 7) et WA ed ee) es) ze masne, July 18.0052 2¢: US Bebe Where cee. c! OCS) Ol Aan are gl: 48 3 | 1909, September 22..| U.S. B: FP. Wharf......02..250..) 16) 69] 11) 3] 99 4.) 1910, November 28. .| Penzance. <. 022.220 000 0. 2 15} 182} 10} 0} 207 5 WilGil> March 3... ERenZancem treet iter i 19] 134; 7} 0} 160 Go lorhWiaveaioue § . o Gut of Cancer, smooth stones. .| 13] 118} 4] 0] 135 7 | 1911, May 30........| Pine TreeIsland, smoothstones.| 5] 94} 8! 0} 107 S) agit, June 29) 2.60... Gut of Cancer, smooth stones..| 14} 129] 2} 0} 145 9 | 1911, June 29........} Gut of Cancer, rough stones...} 28} 254} 8] 1] 291 ARTE Se 6 ha eR ee si Pe > Ca ae a 112/1101| 61} 4/1278 TABLE 14 This table shows, for four lots of individuals, the number of specimens recorded as having the indicated numbers of orange stripes. Lots 1 to 3 are the same as those of the same designations in table 13. Lot X is a selected group of unusually large specimens collected at U. S. B. F. Wharf, Woods Hole, Massachusetts, Sep- tember 11, 1909. ‘ These data are subject to most of the limitations stated for table 13 (p. 219). The totals of this table (excluding lot X) are plotted in figure 42. For a discussion of the data, see page 221. NUMBER OF STRIPES LOT wh Tn | cs IL lc 1 5 eT ST 0} 1] 2] 3] 4] 5] 6] 7| 8} 9] 10} 11) 12} 13] 14] 15) 16] 17| 18] 19] 20) 21 ' 1 5} 6) 6| 2} 6) 6/15} 9/10) 6} 5} 8} 4) 3) O} 1) 1) 1) 1 2 1) 0} 3} 1} 4} 2) 8) 3) 8) 3) 3) 5) 1) 2) 1) 1) 1) 0] OF Of] 1 3 1} O} 2} 2) 2) 5) 1) 7] 6) 4) 7] 8)13) 4/10} 5} 9} 1) 1) 2} 0} 3 LENT > ea ett ena 1} 6} 8/11} 5/15) 9)30)18}22)16)16/26} 9/15) 6}11} 3) 2} 3} O} 4 x 1} O} O} 2) 5) 1/13] 7}10} 2) 4) 1) 2) 1) 1 PLATE 1 EXPLANATION OF FIGURES 1 Colony of S. luciae photographed through the vertical sides of an aquarium. A number of specimens, such as that indicated at 1, are ensconced in barnacle shells. The characteristic vertical orange stripes on the column do not appear in the photograph. White bars may be seen crossing the oral discs. These mark the positions of siphonoglyphs. Continuations of the radial white bars frequently border the oral aperture. The individual at 2, has one radial white bar, a monoglyphic specimen. A triglyphic individual with three white bars may be seen at 3. The specimen at 4 has a minute radial projection of the white border which fringes the mouth, giving ground for suspecting some irregularity of internal structure. 2 A colony seen from above. 1, as in figure 1. 2, individuals each having but one prominent white bar, with extension partially surrounding the mouth, indicating a single well-developed siphonoglyph and, opposite this, a regenerating region in which a white line will eventually appear. 3, a specimen with one broad white bar and one narrower and fainter. The latter is newly formed in a regenerating region. 3 Microphotograph of cross-section of an anemone (no. 46, table 6) which had but one white bar and gave no external evidence of new tissue, being there- fore apparently monoglyphic. The line of new tissue marked by the loop at X was probably mistaken for an orange stripe. For explanation of symbols, see page 232. X 32. ' 4 Photograph at greater magnification of the new part and adjacent struc- tures of section shown in figure 3. X 60. 5 Similar photograph of the regenerating region of a more aboral section of the same specimen. The loop has opened and appears as two separate mesen- teries, c!, cl. X 60. 244 PLATE 1 ASEXUAL REPRODUCTION IN SAGARTIA DONALD WALTON DAVIS 245 PLATE 2 EXPLANATION OF FIGURES 6 Microphotograph of a section of a specimen (no. 1b, table 3) killed twenty days after division. The newest area lies between the two old bounding mesen- teries, (II) and c, and contains six small mesenteries—c is complete orally. The greater part of the section is occupied by an older regenerating area, from (I) to (1) exclusive. Mesenteries labelled c? and c? and the mesenteries in corre- sponding positions on the opposite side of the directive plane of this area are complete orally. X 27. 7 Photograph at higher magnification of the newest area and its bounding mesenteries from the same section as shown in the preceding figure. > 56. 8 A regenerating area (from no. 3a, table 3) slightly more advanced than the newest part shown in figure 7, though actually younger (ten days). Bounding mesenteries, (I), both incomplete. In all, twelve new mesenteries can be counted in the sections of this specimen. X 60. 9 Far oral section of a specimen described when living as having three orange stripes and one white bar in the old region, and one white bar in a regenerating area which occupied approximately one-fifth of the circumference of the speci- men. See diagram, figure 36. Typical regeneration following division in two complete endocoels. The characteristic size relation existing among the re- generating mesenteries is well represented. On each side a new mesentery, c’, mates with the old bounding mesentery, c. The appearance of a siphonoglyph in section is shown at S in the new region and in the groove on the opposite side of the esophagus. In the latter the cilia, though distinguishable in the photo- graph, are not reproduced in the figure. Inthe upper part of the figure a fold of the oral disc appears. X 40. 10 Part of a more aboral section of the same specimen as shown in figure 9. The order of size of the new mesenteries is maintained. X 40. 11 Part of a section of an anemone (no. 8b, table 3) killed fifty-six days after division. On one side of the new directive plane only two potentially complete non-directives (c!, c?) formed where more commonly four are re- generated. Attention should be called to the probability that the mesentery c? in such a region of reduced number of new mesenteries corresponds with the one designated c! of a region regenerating the typical number of mesenteries (p. 191, LO ioe. 246 ASEXUAL REPRODUCTION IN SAGARTIA PLATE 2 DONALD WALTON DAVIS 247 PLATE 3 EXPLANATION OF FIGURES 12 Section of a specimen (no. 66, table 6) which showed one prominent and one faint white bar, the latter lying in a sector of new tissue constituting one- fourth of the bulk of the animal. A fold of the oral dise (or of the column) appears on the side of the section occupied by the new mesenteries. As in the individual represented in figure 11, less than the typical number of regenerating mesenteries is present on the side of the incomplete bounding mesentery, (I). x 45. 13. Part of a more aboral section of no. 66. Size relations indicate the order of development of the new mesenteries. c? and c* are usually more nearly equal in size. The new incomplete bounding mesentery (J) appears here though lacking farther orally (fig. 12). > 45. 14 Section of a triglyphie specimen (no. 63, table 6). Externally there were seen, not quite opposite each other on the oral dise, two white bars lying within the limits of a darker zone forming the greater part of the specimen. On one side was a small lighter colored (newer) sector. There were apparently ten orange stripes of two ages in the older part and beginnings of orange stripes in the newer zone. The most recently formed part, between c and c, shows a typical regeneration following division in two complete endocoels. Two mesenteries on the left of the section between the old bounding mesentery, c, and the pair of directives show a peculiar condition strongly suggesting the position of an old division plane. An anomalous set of four mesenteries is seen at X (p 205). X 25. 248 ASEXUAL REPRODUCTION IN SAGARTIA PLATE 3 DONALD WALTON DAVIS 249 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 PLATE 4 EXPLANATION OF FIGURES 15 and 16 Sections of a very small and faintly colored specimen (later division recorded as no. 82, table 7; an earlier division as no. 64, table 7) having two white bars, but with no orange stripe and showing no external evidence of a division. Figure 15 a more oral section; figure 16 one taken below the aboral end of the esophagus. The sections indicate that the specimen is composed almost entirely of two regenerated areas. The older division plane passed through two complete endocoels as indicated by the size of mesenteries c and c. The later division occurred in incomplete endocoels, as is shown by the incom- plete as well as the complete mesenteries. The new incomplete bounding mesen- teries (1) and (J) of the later regenerating area may be seen in figure 16. Both figures X 60. 17. Section of a small individual (no. 82, table 7) with one white bar (mono- glyphic) and ten evenly spaced orange stripes. There was no external evidence that a division had occurred. The sections make clear that somewhat less than one-fifth of the specimen, including no complete mesenteries, is old material, the remainder having regenerated following a division in two incomplete en- docoels. Old bounding mesenteries (II) are incomplete and certainly not of the first incomplete grade. X25. 18 Section of a small specimen which has certainly regenerated, as shown by mesenteries c? and c*, but which gives no evidence as to the precise position of the plane of division. X 65. 250 PLATE 4 ARTIA x 1 AL REPRODUCTION IN SAG ASEXU DONALD WALTON DAVIS 25 PLATE 5 EXPLANATION OF FIGURES 19 Section of an anemone (no. 73, table 7) composed of an older region con- taining a white bar and five more widely separated orange stripes, and a newer region with a slightly fainter white bar and seven less widely separated orange stripes. (Cf. diagram, fig. 37.) A cinclis is visible at X. X 20. 20 Section of a small specimen probably having undergone two regenerations as indicated by the character of the incomplete mesenteries. Precise positions of the division planes not certain. In the latest formed area (below), c? and c* on either side of the directive mesenteries are incomplete aborally. Possibly on either side the mate to c is an old bounding mesentery; if new, the latest division occurred in exocoels. XX 40. 21 and 22 Two sections of an anemone which showed, externally, slight irregularities in the distribution of its twelve orange stripes. These may have been due to unequal distention of different parts of the column wall. No certain external evidence of a division was present. Internally, as shown in these two figures, the greater development, in the upper region, of the incomplete mesen- teries, especially those of the second grade, indicates that regeneration has occurred, but does not give any assurance as to the exact position of the boundary between new and old. Both figures X 25. 959 aVJe ASEXUAL REPRODUCTION IN SAGARTIA PLATE 5 DONALD WALTON DAVIS PLATE 6 EXPLANATION OF FIGURES 23 and 24 Sections of a specimen which showed no certain external evidenees of division. The slight irregularity in the incomplete mesenteries shown at X in figure 23, and especially the irregularities in the incomplete mesenteries of the corresponding region in figure 24, indicate that two regenerations have occurred. The older is represented by the upper portion of the figures. The greater de- velopment of c, figure 24, shows it to be an old bounding mesentery of the older of the two divisions. X 35. 25 Section of a diglyphie specimen with two white bars and twelve orange stripes distributed as shown in figure 39. Sections showed no certain evidence of division, although the slightly greater development of incomplete mesenteries near the directives below suggests that this may be an older part. Pairs of complete non-directive mesenteries are here indicated by c. An acontium may be seen protruded through a cinclis in an incomplete endocoel of the second grade at a. A pair of nesenteries of the fourth incomplete cycle is present near the directives below. 8. 26 Section of a perfectly regular diglyphice specimen with two white bars opposite each other on the oral disc and twelve orange stripes symmetrically placed and evenly spaced on the column. Internally no evidence of division. Pairs of directives are indicated by d. x 14. 254 ASEXUAL REPRODUCTION IN SAGARTIA PLATE DONALD WALTON DAVIS a : v x Qs : ba th i BARK ey). ye Es 2 MA 4 ~ el PLATE 7 EXPLANATION OF FIGURES 27 to 32 represent two sections from each of three of the four parts into which specimen no. 19, of table 5, divided. No. 19, before this division, was a very large diglyphic specimen with twelve orange stripes and showing externally no evidence of a previous division. Old regions are indicated by a greater number of cycles of incomplete mesenteries, by the greater size of all mesenteries, and by the presence of gonads attached to the larger mesenteries. The position in the original polyp of the three specimens figured is indicated in table 5. The regeneration period was from 34 to 87 days. For a full account of this case see page 173. X 23. 27 A far oral section of no. 19a. The fission plane passed through one com- plete endocoel and one endocoel of the first incomplete grade. An unexplained irregularity is seen in the incomplete mesenteries of the old part near the com- plete bounding mesentery, c. 28 Amore aboral section of no. 19a. 29 and 30 Respectively more oral and more aboral sections of no. 19d. Division occurred in incomplete endocoels, one of first eyele and one of third cycle. The small old bounding mesentery is in the position of one of the third incom- plete grade and far aborally is similar in size to others of that cycle. Its new mate, barely indicated in figure 30, is smaller than the other new incomplete mesenteries, all of which are of the first incomplete grade. At present, there- fore, it appears to belong to the second incomplete cycle and it is so indicated by the label. Probably its relation to the third cycle would be evident later. 256 ASEXUAL'REPRODUCTION IN SAGARTIA PLATE 7 DONALD WALTON DAVIS 29 Pen - SS 257 PLATE 8 EXPLANATION OF FIGURES 31 and 32 Respectively more oral and more aboral sections of no. 19¢e. Two regenerating regions are represented. One of these follows a division in one complete endocoel and one endocoel of the first incomplete grade. This matches up with no. 19a (figs. 27 and 28). The other division occurred in one complete endocoel and one endocoel of the third or perhaps second grade (fig. 32). At some levels there is a mate to the incomplete bounding mesentery of this region, (III), which appears to belong to the second cycle of incomplete mesenteries. 33 Part of a section of an anemone fixed with protruded acontia, a. One lies in an endocoel of the second incomplete cycle and one in an exocoel. A third cinclis not occupied by an acontium is distinguishable at x. & 8. 34 A more highly magnified view of an acontium penetrating a cinclis lying in an exocoel or possibly in an endocoel with exceedingly slightly developed bounding mesenteries. X 43. 258 ASEXUAL REPRODUCTION IN SAGARTIA PLATE 8 DONALD WALTON DAVIS 259 PLATE 9 EXPLANATION OF FIGURES 35 Diagrams I-VI, VIII, and X have been modified from those of Carlgren (09, p. 39, fig. II). Diagram VII has been constructed from his deseription (09, p. 35) of a specimen of Sagartia viduata (no. 15a7). Diagram LX has been taken from his figure (’09, Taf. 4, Fig. 47) representing a section of a specimen of the same species. These diagrams show different arrangements of mesenteries found by Carlgren in regenerating specimens of Sagartia viduata, Metridium dianthus, or Aiptasia diaphana at stages giving indication of the order of development of the mesenteries. Only mesenteries that would eventually be complete are represented. Portions enclosed in dotted lines represent old material. Some mesenteries were present in every such region and, usually, one or more of the old mesenteries were complete. The two mesenteries shown here are not to be understood as indicating the number or character of old mesenteries. I have labeled the new mesenteries in one-half of each diagram I, II and III with the symbols given in my figures and text to mesenteries occupying corre- sponding positions with respect to the new directive plane. Carlgren’s diagram 5 (’09, p. 39, Fig. II) shows mesenteries c? and c? on the left side bearing muscles facing each other instead of facing the mesenteries with which they commonly form non-directive pairs. The same arrangement is represented on both sides in his figure 37 (’09, Taf. 3). Since I find no mention in the text of so remarkable a condition, I am led to think that an error was made in both places. I have therefore changed the positions of these muscles in the corresponding diagram (III) of my figure. Certainly this coincides with the conditions in 8. luciae. For the frequency of these types of arrangement of mesenteries in regenerating pieces of the different species, see table 12. For a discussion of the relations of these types, see text, page 201. 36 to 41 Diagrams representing arrangements of orange stripes and white bars, as seen on living specimens. The inner part of each diagram represents the oral dise of an anemone with mouth in the center and with one or more white bars, shown by stippled lines, extending radially across the dise. The tentacular ring is not represented. In the outer part orange stripes are indicated by solid lines, Broken lines mark the boundaries between recently regenerated and older tissue. The distinction between these in living specimens showing early stages of regeneration is indicated by differences in color; in length, width, and density of the white bars; and in breadth, depth of color, and closeness of the orange stripes. The newer area is toward the lower edge of the plate except in figure 39, where it is above. 36 Diagram of specimen at a stage of regeneration when all of the complete mesenteries have formed but not all have reached the esophagus. A new white bar but no orange stripes have appeared. A section of this anemone is shown in figure 9. 37 A specimen (no. 80, table 7) in an advanced stage of regeneration. Two old and nine new orange stripes (the original record noted a possibility that one of the mesenteries close to the boundary might be old). For a description of internal structures, see page 211. (Continued on page 262) 260 VII ES GS) GS) SS M18) Ons) & ay aS. GS oe (Continued from page 269) 38 Specimen (no. 73, table 7) a section of which is seen in figure 19. The correspondence between external features and internal structures is typical, the distribution of orange stripes accurately indicating the positions of pairs of mesenteries of the first two cycles. 39. Specimen a section of which is shown in figure 25. Three days before this diagrammatic sketch was made, the orange stripes were noted as being equidistant. Internal structures give little suggestion of regeneration. The unequal distribution of the orange stripes shown in the diagram may be due to the momentary state of expansion. 40 Diagram of a monoglyphic specimen (no. 48, table 6). Color difference did not clearly locate the boundary between old and new parts on the left. In sections the mesenteries gave clear evidence of its position. 41. Diagram of a triglyphic specimen with three old and eleven new orange stripes. Apparently the eleven orange stripes and two of the three white bars belonged to a single regenerating sector. Sections make it clear that there are here two regenerating zones of different age. See page 216. PLATE 10 EXPLANATION OF FIGURES 42 Curve (solid line) representing the number of stripes in individuals of combined lots 1, 2, and 3 of table 14. The similar curve (broken line) given by Davenport (’03) is added for comparison. For a discussion of the data, see page 221. 262 ASEXUAL REPRODUCTION IN SAGARTIA DONALD WALTON DAVIS 120 aaage 7a ee FEEEEEEEEEEEEEEEEEEEE i] SSEEEETHUERESEUESEE ae is i ea [eae a ty ot HH] | +4444 HHH Pa ee ERS 5 a a eas | ea ERE BRA hs SARS eee fl i ee en a fafa [es ah ee a La OR a | aN Rc Pa Gat i OF oS a LT a we te en Fe fe a a a | 7 Jee SES e eee fal SE AT ee a a in i | ttt a et i] a pe ea GER eEAsa ee | a a eee ah eS Se ee See 70 60 50 40 30 10 PLATE 10 Resumido por el autor, Calvin B. Bridges. La genética del color purpura de los ojos de Drosophila melanogaster. El color ptirpura es una de las primeras mutaciones halladas en esta mosca (descubierta en 20 de febrero de 1912), la cual ha resultado especialmente Util. Es un caracter estrictamente re- cesivo, facil y rapidamente separable del tipo salvaje, completa- mente viable, fértil y productivo. Su locus estdé situado en el segundo cromosoma, en un punto distante 6.2 unidades a la derecha, del locus del color negro y 52.7 unidades a la derecha del de la mutaci6n estrella. La regién indicada corresponde a la mitad del cromosoma, coincidiendo con lo que se pens6 al planear- le, puesto que esta regién se caracteriza por un doble entrecruza- miento! anormalmente elevado, una sensibilidad especial a la accidn de la edad, calor y frio sobre la cantidad de entrecruza- mientos y por una limitaci6n especial sobre la accién de ciertas varlaciones genéticas originadas por el entrecruzamiento de los cromosomas. El color purpura se ha utilizado para el desarrollo de muchos aspectos importantes de la genética de Drosophila; con el color bermell6n produjo “‘intensificacién”’ o ‘‘modificacién desproporcionada.’’ Ha servido también como modelo para la mutacion ‘“‘mimica” repetida, y ha sido también ‘‘recurrente.” Finalmente, este color ha sido utilizado muy intimamente en el andlisis del ligamiento autosomal (acoplamiento F., cruzamiento retrogado del macho y hembra para comprobar el entrecruza- miento, planeado de dos y tres puntos etc.). La curva de coin- cidencia correspondiente a la variacién de la edad es, en términos generales, la imagen de la curva de entrecruzamiento para la variacion de edad, mientras que la curva de coincidencia de la . variacion de temperatura parece limitarse a una linea recta inde- pendiente de la temperatura. Estas dos variaciones dependen, aparentemente, de dos factores fisioldgicos independientes que afectan a la “‘longitud internodal”’ y al ‘‘coeficiente de entrecruza- miento,’’ respectivamente. Translation by José F. Nonidez Columbia University 1Con esta palabra traducimos la correspondiente inglesa ‘‘crossing over’’ entendiéndose que se refiere a los cromosomas exclusivamente. ‘‘Acoplamiento”’ puede servir para traducir la palabra ‘‘coupling”’ y ‘‘ligamiento”’ para “linkage,” a falta de otras mejores qué sugieran la idea expresada por las palabras inglesas, que tal vez convendria conservar sin traducir. N. del T. AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 31 THE GENETICS OF PURPLE EYE COLOR IN DROSOPHILA CALVIN B. BRIDGES Columbia University, New York City CONTENTS MEM UCLIONI Sn Cee Coles oem OR ye cise ak ote MP Re era a x EE Dr aS ce 8 265 LD Nae ee MRR oS i 00 IS Se Re Da eae thea |’ Oe ee a ee ee OS eae 266 BUMMCEVCARCOS. ae Aen.) s s IKELS SeEL4 OSG. PUES CRUSE De Oe ws See dees 266 TOES ere (OLS TS AS Oe A Be CREPE MMT ARRAS ac rie Wane ls 6 eS ERR ey ed 267 The differentiation of purple by vermilion—disproportionate modification... 268 iBhe elation of purple to pinks’. $35 sees oak ete ale eo eens be bo Sha nn cote 269 PRewinkwee ol purple and restigralisy.s/ ini). Sera ees Sas Fee 2 ee Ss. eee 269 Back cross tests of males, purple vestigial ‘coupling’....................... 269 Back cross tests of females, purple vestigial ‘coupling’..................... 270 MNOrORUN sine OVEr Li bie DIGIC Y. s/h: Sa apie OR ac ae eo leon e's anh ie See 273 Mutations s:.5.25. 02... EL UARE OEE CAPO EOOES. SON tas Set N : REET etre 275 The inviability of vestigial—prematuration, repugnance, lethals........... 276 wietnurple- epidemic,” “mutatmMe: perous .... sees puss As. agetiloniend se sets 278 Repetition of the purple vestigial back-cross tests. ..................e eens 279 Balanced inviabilty—complementary CrosseS................0e cece cece eens 281 Phe wariations of crossing Over: withy ages inicsis- cng a. ak. sl 0)) oa ble Siw seize 285 ihe tects OF nutples 4 two-pouit map, jor. he nets Nie a siss oe Giecto ame pate oF 285 PURGE C EAM GOU TAA y of oy ok Sa tetera arte cic se iis Sea Ae wa ts Sad + rene en eee 286 A three-point back cross, black purple vestigial with balanced inviability... 288 AP OMICEMETICEWR, ce iat ie ete eke oe EC Eh Reta s rst rae tae 8h ba EaaPa Ek, AEE 291 The relation between coincidence and map distance.....................05- 292 The use of purple in mapping other genes, curved, streak, etc.............. 292 PRE OEE HIRO COSSESLis s'h5 31, cred cath aia 5 otis vie coal tooisne Pate ae MS cat ke 295 A summary of the linkage data involving purple....................000000- 296 Special problems involving purple—age variations, coincidence, temperature variations, crossover mutations, progeny test for crossing over.......... 297 pana eA ST Ch yAA VEL TON), «2/5. 2 - Bp sicyhcalbeseiig cl aba isu oe SE Gaade se a eee dew Pus 34 Tht peg 302 LUA PeT esi E NES. COU EG be SEE 5 at nee ters Stra eae eit 8 AG ARR re a a ue 305 INTRODUCTION Of the two hundred or more mutations of Drosophila, ‘purple’ ranks high among those that have proved especially useful because of their ease of identification or other excellent characteristics, and 265 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 266 CALVIN B. BRIDGES because of their favorable location in the chromosome. Purple has an even higher interest because of its connection with the development of several new fields in genetics and of principles that are now made use of in every Drosophila experiment. ORIGIN In a stock which was supposed to be simply vestigial there was found, February 20, 1912, a single male which had an eye color much like that of the well-known double recessive ver- milion pink. The color of the vermilion-pink eye is about that of the pulp of an orange, and the early papers accordingly referred to this double recessive as ‘orange.’ The new color. was seen to differ slightly from vermilion-pink in that it was of a brilliant ruby-like transparency, and lacked the flocculent or slightly cloudy appearance of vermilion pink. This difference seems to arise partly from a difference in the distribution of the pigment. In vermilion pink the pigment looks as though it were mainly in the spaces between the radially arranged om- matidia with: a clearer zone just under the surface of the eye. One sees in the vermilion pink eye a light fleck which travels over the eye as it is turned. This seems to be due to a de- ficiency of pigment in the deeper parts of the eye, and-the light fleck is this light center seen through the small group of facets whose axes are in line with the eye. The pigment in the case of the new eye color gave the appearance one would expect if it were uniformly distributed or even in solution throughout the eye. INHERITANCE This single male with the orange-like eye color was outcrossed to a wild female, and in F, gave only wild-type males and females (wild-type @° 382, & 33; reference no. Bl) which showed that the color was recessive. In F, the orange-like color reappeared, but in addition the sex-linked eye color vermilion emerged, and also a new eye color, ‘purple,’ which appeared equally among the F, females and males, and therefore was known to be an autoso- mal (not sex-linked) character. It was now evident that the GENETICS PURPLE EYE COLOR DROSOPHILA 267 orange-like color resembled the old ‘orange’ (vermilion pink) genetically also, for it was proved by this F; to be a double recessive, vermilion purple, in which purple corresponds to pink. It seems probable that the two eye-color mutations, ver- milion and purple, present in the male first found were not of Simultaneous or related origin. There was a vague report that the vestigial stock had contained vermilion at some time pre- vious to this discovery. No vermilion or purple was found in it subsequently, however. DESCRIPTION! The purple eye color passes, in its development, through an interesting cycle of changes closely parallel to those seen in the ripening of a ‘sweet’ cherry. In the pupa the eye is at first colorless, then it assumes a creamy tone, which in turn becomes pinkish, passing progressively through a yellowish pink to pink and to ruby. When the flies hatch the color is a transparent rather deep ruby. This color rapidly deepens to garnet and then passes on to a purplish tone. The typical purple color at its maximum development—in flies about a day old, while retaining much of its transparency, appears darker in tone than the red of the wild type, purple being the first of such ‘dark’ eye colors. As the fly becomes older this ‘ripe-cherry’ color is progressively obscured, apparently by an increase in a flocculent red pigment like that of the wild fly. The eye color thus be- comes somewhat lighter than red again, though always dis- tinguishable by a lesser opacity and by a light ‘fleck’ in place of the hard dark fleck seen in the wild eye. With extreme old age the color approaches still closer to red, but does not become strikingly darker, as do pink and sepia, for example. In purples of the same age fluctuations in color are not great. The sepa- ration of purple from red is easy if done while the flies are mostly | under two days old, though the climax in the development of the purplish tone offers the most favorable stage. 1 For a colored figure of purple see plate 5, figure 8, of a forthcoming Carnegie publication (No 286) by Bridges and Morgan. 268 CALVIN B. BRIDGES THE DIFFERENTIATION OF PURPLE BY VERMILION—DISPRO- PORTIONATE MODIFICATION While the difference between the color produced by the purple gene and the color produced by its wild-type allelomorph (red) is distinct, it is neither great nor striking, since in tone purple is first slightly darker and later somewhat lighter than red. How- ever, in classifying the eye colors in F, from the cross of ver- milion by wild, it was observed that the difference between vermilion purple and vermilion not-purple was not only con- stant in direction, but also conspicuous in extent. The sepa- rability of purple versus not-purple is favored by the presence of vermilion, which may therefore be called a ‘differentiator’ of _ purple. Regarded in the converse relation, namely, the effect of purple on vermilion rather than the effect of vermilion on purple, purple is a much stronger modifier of vermilion than of not-vermilion. Purple may be described as a ‘disproportionate modifier’ of vermilion, since from the small amount of its effect on eye color when acting alone one would not have expected the great effect it produces when acting in the presence of vermilion. This type of intensification—disproportionate modifier and, conversely, differentiator—stands midway between the normal relations where combination effects are roughly proportional to the separate effects so that both genes may be called ‘general modifiers,’ and the special relation where a given gene, ‘specific modifier,’ produces by itself no visible effect whatever, but which gives a more or less marked effect when acting in con- junction with some other gene, its specific base, sensitizer, or differentiator. | In order to make full use of this differentiation of purple versus not-purple by vermilion, it is necessary that all flies used in the experiment should be made homozygous for vermilion. This is often inconvenient, and accordingly only in the early and comparatively simple experiments was this method em- ployed. It was soon found also that the separation of purple from red was not causing any trouble, so that the differentiation in this case has little net advantage, though it is still of interest as being the first example in Drosophila in which intensification was recognized and deliberately made use of. GENETICS PURPLE EYE COLOR DROSOPHILA 269 THE RELATION OF PURPLE TO PINK Some of the first purples which emerged in the F, were crossed to pink, to test whether these two somewhat similar eye color were allelomorphic or not. Four such pair matings produced only wild-type males (134) and females (137), which showed that purple is not an allelomorph of pink. THE LINKAGE OF PURPLE AND VESTIGIAL It was observed (April 2, 1912; Bl) that in the F, from the cross of the original male to wild nearly all of the flies that were purple were also vestigial. This observation, following on the heels of the black-curved case, furnished a second example of autosomal linkage, this time one of so-called ‘coupling,’ the black-curved case having been ‘repulsion.’ No full counts were made of the proportion of purples that were vestigial. Indeed, at this early stage the linkage relations were receiving less attention than eye-color ‘series.’ BACK-CROSS TEST OF MALES, PURPLE VESTIGIAL ‘COUPLING’ The advantages of the back-cross method of testing linkage and the amount of crossing over had only begun to be appre- ciated. This method had been applied to a few cases in the X chromosome, and the general attack upon the linkage of all autosomal mutations planned by Sturtevant and Bridges (March 5, 1912) contemplated its full use. Thus far only two autosomal back crosses had been completed—those by which Sturtevant showed the absence of linkage between the second chromosome and the third chromosome (balloon ebony, May 10, 1912, and black pink, May 12, 1912). Because of the difficulty of getting the necessary double recessives no back cross which in- volved autosomal linkage had been possible until purple arose in the vestigial stock and thereby gave the required double recessive, purple vestigial, with which such a test of the amount of crossing over between purple and vestigal could be conducted. From the F, described above, matings were made which gave two stocks to be used in this test. One stock was the simple 270 CALVIN B. BRIDGES purple vestigial, and the other was purple vestigial pure for vermilion. The special advantage of this latter stock lay in the fact that the presence of vermilion accentuates the difference in eye color between the flies that are purple and those that are not, that is, vermilion purple is easier to separate from ver- milion than is the case in the equivalent separation of purple from red. This latter stock was accordingly used in the P,; mating for the first back-cross test. Vermilion purple vestigial males were outcrossed to females of vermilion stock (May 25, 1912). Both parents were homozygous for vermilion. and the F, flies were all vermilion as expected. Both purple and vestigial are recessive. When the back-cross matings came to be made, the culture bottle happened to contain no virgin F, females, since the Pi mating had been made at Columbia and the F, progeny used had hatched en route to Wood’s Hole. The back cross was therefore made in only one way—by mating the F,; males to virgin vermilion purple vestigial females of the stock kept for that purpose. Five back crosses were started by mating in each case a single F, vermilion male by two or three stock ver- milion purple vestigial females. At the end of ten days the parents were removed from the culture bottles and were put in fresh bottles in which second broods were raised. In one case a third brood was raised (table 1). The linkage results of these back crosses were somewhat unexpected, for in four of the lines no crossovers at all were obtained, and in a fifth only a few. In the original F, culture several crossovers had been noted, and five F, cultures raised from the brothers and sisters of these back-crossed males were giving in the neighborhood of 15 per cent of crossovers (table 2). The apparent crossovers had all appeared in one culture of the first and of the second broods, and for this reason a third culture was raised from that particular set of parents and it also gave apparent crossovers. A second back-cross experiment, using the simple purple vestigial stock instead of the vermilion purple vestigial, was started (June 25, 1912) a month later than the first and before GENETICS PURPLE EYE COLOR DROSOPHILA Zee TABLE 1 The B. C. offspring given by the F: (vermilion) sons, from the outcross of (vermilion) purple vestigial males to vermilion females, when back crossed to (vermilion) purple vestigial females First and second broods given separately. NON-CROSSOVERS CROSSOVERS 1912, sunE 241 7" 6 ri = Ped atraerail) ermilion)) eV Senne” ilk Cranial BLO eae Sys elected 90 186 0 0 71 202 0 0 [0 eee pee ie am 72 197 0 0 72 206 0 0 Poti CA BW Lot obe 45 126 0 0 65 195 0 0 51 88 7 3 PUREE ihc ia ih wicca 98 178 27 2 43 72 4 0 Eminent eat oss he 54 191 0 0 37 70 0 0 Movabten: <.vrdtokes sau: 698 1711 38 5 1 Date on which the cultures of the table began to produce offspring. TABLE 2 The F2 offspring given by the F: (vermilion) sons and daughters from the outcross of (vermilion) purple vestigial males to vermilion females (VERMILION) 1912, sunn 17 (VERMILION) Ee (vaRMLELOR) Gaesere leo lsd BR ee ene 200 23 9 5 oe 74s outs cid tartare aaa 88 Al 3) 4 BS rary es Sieh kidd kote: 255 66 25 5 TEV ST 2 Benen eee ae ee 368 19 3 ff 15X8) 2 tate Di oe ee eee 346 17 19 9 he. CALVIN B. BRIDGES the results of the first were fully known. A purple vestigial male outerossed to a wild female produced wild-type sons and daughters (page B39; + 9 15, + o& 10). Four of the Fi females were back crossed each by two or three purple vestigial males from stock. In this case F; females happened to be chosen because, as is usually the case, they hatched somewhat earlier than their brothers in the same culture. These back-cross cultures (table 3), in common with the previous F, cultures (table 2), showed a fair amount of crossing over between purple and vestigial. A calculation showed that the percentage of crossing over was 9.1. TABLE 3 The B. C. offspring given by the F, daughters, from the outcross of a purple vestigial male to a wild female, when back crossed to purple vestigial males NON-CROSSOVERS CROSSOVERS 1912, suLy 16 potatoe e ee Purple vestigial Wild type Purple Vestigial BBO ele: : ae eee e Hee ae 82 163 LZ, 15 1B ROS s Kesey SE AMG 2 5 CMMI ame 80 133 14 10 BS SOE. heres co ein: bees 32 53 3 7 BOSD eee Fone Ne Sis fo ee 62 141 9 9 ABCC) sts eee ees ee 256 490 38 41 This was recognized as being of a different degree from the apparent percentage of 1.8 calculated from the first back cross (table 1). It was now realized for the first time that the two back crosses had differed in the sex of the F; flies tested by the back crosses—that the first back cross was a test of the amount of crossing over in the male and the second was of crossing over in females. Up to this time there had been no suspicion that the result of a back cross could be in any way dependent on the sex of the F; parent used in the experiment. From this evidence it was concluded that there was crossing over in the male, but that it was of different degree from that in the female. In Septem- ber, 1912, Morgan showed that in the case of black vestigial no crossing over whatever had occurred in the male, while in the female there was even more crossing over than had been found GENETICS PURPLE EYE COLOR DROSOPHILA ate ‘in the case of purple vestigial. Subsequent tests, including hundreds of thousands of individuals, have shown that ordi- narily there is no crossing over in the male for any chromosome and that the few cases that have occurred were probably not brought about by the same mechanism as that by which cross- ing over is ordinarily effected. NO CROSSING OVER IN THE MALE A clear conception of the fact of no crossing over in the male was prevented in the original vermilion purple ve&tigial back- cross test by the apparent occurrence of crossovers in one of the five lines. No tests were made of the apparent crossovers because there was at that time no evidence, aside from the in- consistency within the experiment, to suggest that they were highly unusual. Against the supposition that some clerical error might have been made is the strong internal evidence pre- sented by the aberrant cultures. Thus, the cultures could not have been F,.’s that were mislabeled, since the proportion of purple vestigials in this line is the same as that in the other back- cross cultures and is much larger ‘than that in any of the F, cultures. Also, the parents were carefully examined when they were transferred to the third culture bottle and were seen to be a vermilion male and vermilion purple vestigial females only, which is the back-cross type of mating. The examination of the parents also excluded the supposition that the line may have been a back-cross test of the female rather than of the male. Perhaps some unknown peculiarity of the stocks used may have been responsible for the apparent crossing over. Thus, it has been suggested that some other eye color resembling purple, such as ‘maroon,’ had been present, probably only in hetero- zygous form, in the vermilion purple vestigial stock. Such an explanation would account for the crossover class classified as vermilion purple, but entirely fails to account for the comple- mentary class of exceptions—the few but carefully attested ves- tigials that were not-purple. In fact, none of the suggestions that have been made have offered a satisfactory escape from the alternative of some kind of crossing over in the male. 274. CALVIN B. BRIDGES If these were true crossovers, it is possible that their pro- duction had no relation to the mechanism by which crossing over is ordinarily effected. Thus, Muller (16) reported a case of crossing over in the back-cross test of a certain F,; male from the mating of truncate to black. However, all of the gametes of this particular F; male proved to be crossovers, so that crossing over must have occurred, once for all, in an early cell of the embryo, and, as usual, no crossing over whatever occurred during spermatogenesis. The spermatozoa, all of which were descended from this embryonic crossover cell, simply inherited the cross- over combination. In the case of purple vestigial, a like ex- planation would apply, except that in this case the crossing over occurred in a somewhat later stage of the embryo, and in conse- quence only a part of the spermatogonial cells carried the crossover combination and only sperm decended from _ these particular cells produced crossover progeny. That somatic crossing over has little analogy to the ordinary type is proved by a similar case of embryonic crossing over in the female, which was then followed by crossing over of the ordinary type. A mating was made such that a certain class of lL+4+ B 3 = Mi al See ello Seven of the eight daughters tested had this expected compo- sition, but one (no. 3464) gave only offspring corresponding to i = * = 2 That is, the gene for lethal 9 was found to be not in the chromosome in which it entered the zygote, but in the homologous chromosome derived from the other parent. As in the truncate x black case, this transmi- gration took place after fertilization and so early in the em- bryonic history that all the germ cells were descended from this altered cell. hi ip ‘nt ' ee ie ie om i% - Fy we $ ets .. babes, td beeper bh A i stale r co bal ral anteaters bm a aie Ais e xh 4 eee 6 | { ath mak it ih 4 Aart | ea ff lc, Wir, ed Fag Vi ¥ may iy » tat EYE COLOR IN DROSOPHILA MELANOGASTER 351 homozygous whiting had no visible effect in the absence of eosin have proved that red, the normal allelomorph of white, eosin and cherry, is unaffected. In order to observe the interaction of white and whiting, two (eosin) whiting males from M114 were out-crossed to white females of pure stock. The F, females were white-eosin compounds of normal color, and the sons were white as expected (table 9). F2 from the cross of eosin male by white female gives in equal numbers the four classes: white-eosin °, white 2, eosin o, and white &. The cross of (eosin) whiting TABLE 9 The F, and Fy, offspring from the cross of (eosin) whiting males to white females 1914, 1/20 WHITE-EOSIN 9 WHITE IMMUNO 5 deste ata ch ORS Ain PSE Deedee 113 107 IIT Rance eS tas ARs ote 46 46 JRO SSNS Sela ra Ren ean 159 153 1914, 2/3 eee nots) WHITE 9 (wine) EOSIN ey) | WHITE (of (oommso 9 44 81 54 75 31 37 65 24 55 32 18 Pal 25 28 hotale 99 167 103 158 -male by white female should give this same result with the fur- ther genetic subdivision of each of the four classes into 3 not- whiting to 1 homozygous for whiting. Since all white flies are incapable of further dilution the white whiting double recessive should be white even in the absence of effect of whiting on white. But in the white-eosin compound females we shall have a chance to observe the interaction: of white and whiting. If white and eosin react in the same fashion toward the whiting gene, then the substitution of one white allelomorph for one eosin in the white-eosin compound should not prevent the whiting from diluting the compound to a colorless eye. But if the white cannot thus be substituted, the white-eosin compound should THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 3 352 CALVIN B. BRIDGES be either unaffected or less affected by the action of the whiting gene. Should the white-eosin whiting flies be colorless, then the colorless classes should be 3 white + 1 whiting white + 1 eosin whiting (eosin whiting if males, white-eosin whiting if females), and these classes should total 2, or 62.5 per cent. The colorless flies did in fact constitute 61.7 per cent. What is even more significant, the proportion of colorless flies among the females was just as great as among the males (table 9). The colorless males are known to include the (eosin) whiting flies, and the conclusion is justified that the like class of colorless females included white-eosin whiting flies. TABLE 10 The F, and Fy, offspring from the outcross of (eosin) whiting males to cherry female 1914, 1/28 EOSIN-CHERRY @ CHERRY ©’ 1 RR i SO RS 64 75 VETO cee ae Rees. tes cede oe 64 63 OLAS. MPR toss elas 128 | : 138 EOSIN- | 1914, 2/9 Gees to (earnna) CHERRY 9 (we) EOSIN o oe) CHERRY <' (waimvo) fot °) 27 159 64 12 93 28 169 48 24 73 Total... 328 112 36 166 The most interesting of these three experiments was that in which (eosin) whiting males were outcrossed to cherry females. The F, females were eosin-cherry compounds, and the males were cherry which is of nearly the same color (table 10). F» of the cross of eosin male by cherry female gives in equal numbers the F, classes eosin-cherry compound @, cherry @, cherry <, and eosin @. The eosin-cherry female and the pure cherry female are so nearly the same color that they form one phenotype, but the eosin males can be easily separated from the cherry males. The cross of (eosin) whiting male by cherry female will give further subdivisions of each of these classes into 3 not- whiting: 1 homozygous for whiting. EYE COLOR IN DROSOPHILA MELANOGASTER Soe. The result obtained in this F, was entirely unexpected; for cherry, which is allelomorphic to eosin and so closely similar as to be distinguishable only in the males, gave a totally dissimilar reaction with whiting, the double recessive, cherry whiting, being indistinguishable from cherry both in males and females (table 10). This result is very clearly shown by the males of this F, which closely approach the ratio 4:3:1. 25 imee | Ehe eS Oum. |steheee4 ame 2 65 34 2h ova, |)2eh. | Ohm 2ahn eloeme 3 58 63 4h roms |) lo | Oo 5 || PA lriy | 7 sea 4 43 64 PANG ay ely xeon, Pl Serra 1 This is exclusive of one day when, in the absence of the writer, only a single alcohol treatment was given. and tail. His comb was large and corrugated, intermediate between rose and walnut. He was polydactyl, grade 1 (left foot only); booted, grade 2, and with a brachdactyly index of 88. In view both of his ancestry and his descendants, it is clearly apparent that this bird was heterozygous for brachy- dactyly and for polydactyly. ' The homozygous parents were thirteen pure bred white Leg- horn hens of a standard strain. There can be little doubt as to the purity of this stock, especially as regards the characters under investigation. In this experiment it was to be expected that the germ cells produced by the females would all be of the same class, while those produced by the male would fall into several classes EVIDENCE OF GERM CELL SELECTION 395 depending on whether or not they contained determiners for brachydactyly, polydactyly, or both. Such a situation supplies the necessary conditions for a test of germinal selection. Since Leghorn white is dominant, and all germ cells of the females necessarily carried determiners for this trait, no critical data bearing on color selection were to have been expected from this experiment. TABLE 2 General summary of data from the four experiments EXPERIMENT EGGS EMBRYOS CHICKS Number | Poe | Suge fosinea] eR, | Sikes] ish | Nese 1 A 150 150 5 39 51 55 C 362 300 34! 108 72 86 9 A 191 180 2 ial 153 14 C 106 104 1 19 69 15 3 A 151 150 35 14 60 4] C 166 155 49 12 60 34 4 A 39 39 0 6 13 20 C 151 150 2 45 76 27 Totals: A and C A 531 519 42 70 277 130 separately C 785 709 86! 184 277 162 Totals: A and C * feaed \ A+C}] 1316 1228 128 254 554 292 1 As explained in the text, this figure is probably too large, owing to the method of recording in 1-C. C. Thecontrol part of the experiment, running from February 15 to April 18, was not originally intended for this purpose, but rather to supply data for the study of the normal heredity and embryology. of several traits, including those considered here. Consequently some of the data, especially those relative to fertil- ity, were not entered in strict accordance with the form adopted in the other experiments. However, this set of data is easily comparable with the other sets and, except for fertility, un- doubtedly furnishes a reliable control. Of the 362 eggs laid 396 Cc. H. DANFORTH only 300 were incubated. The 62 unused eggs were discarded from time to time in small numbers whenever the capacity of the incubator was exceeded, the incubator always being filled from the most recently laid eggs. ‘This method leaves no room for unconscious selection as to size, shape, etc. A. The alcohol treatment began on April 27 and continued until June 1. Beginning May 1, there were two daily treat- ments averaging a little over an hour in the morning and half an hour in the afternoon. In this period of thirty-six days male no. 8 was kept in the alcohol vapor a total of 63 hours and 11 minutes (table 1). Eggs were saved from May 3 to June 2. One hundred and fifty were laid, all of which were used (table 2). Experiment 2. (March 6, 1918, to June 19, 1918). The heterozygous parent in this case was a male (no. 27) hatched in March, 1917, from an egg used in experiment 1-C. He had the general bearing and many of the characteristics of a Leg- horn. His color was white, his comb large and walnut-rose. He was polydactyl, grade 3; booted, grade 2, and had an index of brachydactyly equal to 75. It will be apparent by reference to experiment 1 that this bird was heterozygous for the three dominant characteristics with which we are concerned, i.e., brachydactyly (including booting), polydactyly, and white color. The homozygous parents were six single-combed black Mi- norca hens purchased from a local dealer who gave assurance that they were pure bred and of a stable strain. Their somatic appearance as well as their racial purity showed them to be homozygous for the absence of the three above-mentioned traits, or, in other words, they exhibited the corresponding recessive characters, normal length of toes, normal number of toes, and black color. In this experiment, therefore, all the germ cells produced by the females were necessarily of one class, that tending to give black chicks with normal number and length of toes, whereas the germ cells of the male were expected to fall into eight classes, VizZ.: EVIDENCE OF GERM CELL SELECTION 397 Those tending to produce chicks that were . Brachydactyl, polydactyl, white. . Brachydactyl, polydactyl, black. . Brachydactyl, not polydactyl, white. . Brachydactyl, not polydactyl, black. . Not brachydactyl, polydactyl, white. . Not brachydactyl, polydactyl, black. . Not brachydactyl, not polydactyl, white. . Not brachydactyl, not polydactyl, black. It will be apparent that by the inspection of any chick it could be determined at once to which of these eight classes the sperm cell involved in its production had belonged. C. The hens began laying on March 6 and eggs were saved for the control experiment from that date to April 8, during which period 106 eggs were produced (table 2). Two of these were accidentally cracked and therefore discarded. One, laid April 3, was non-fertile. A. The period of alcohol treatment followed immediately upon the control period. The first treatments, of 1 hour and 2 minutes and 1 hour and 4 minutes, were given on April 8, and the final treatment on June 11. This bird was kept in the inhalation chamber for two periods of at least an hour each every day except May 5, when in the absence of the writer an attendant misunderstanding directions gave only a single one- hour treatment. In the period of sixty-five days no. 27 was kept in the alcohol vapor for a total of 143 hours and 56 minutes. He, showed the usual responses: hyperexcitability, occasional weakness in the legs, and crowing ability invariably impaired for many minutes. He was removed from the breeding pen on June 12. Eggs were saved from April 9 to June 19. During this period 191 were laid, of which eleven, unfortunately, were lost through breakage or otherwise (table 2). The two recorded as non- fertile were laid on April 21 and June 14, respectively. Experiment 3. (February 9, 1918, to June 12, 1918). The heterozygous parent (no. 28) hatched in June, 1917, from an egg used in experiment 1-A.. He was very similar to male no. 27 described above; in fact, these two males were selected OANA orR WN THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 3 398 Cc. H. DANFORTH because they were more nearly alike than any other two raised in the 1917 flock. They were at least half-brothers, possibly full-brothers. It may be noted in passing that no. 27 was pro- duced before the father had been alcoholized, no. 28 after the alcohol treatment had begun. No. 28 was white with a large walnut-rose comb. He was polydactyl, grade 3+, showing a trace of a sixth toe on the left foot; booted, grade 2, and with an index of brachydactyly equal to 70. It may be observed that brachydactyly and polydactyly were more pronounced in this specimen than in any of the other individuals used in the four experiments. The homozygous parents were six black Minorca hens similar to those used in experiment 2 and from the same source. In so far as the stock was concerned, the material for experiments 2.and 3 was as nearly identical as it was possible to make it. The two flocks were kept in the same room separated from each other only by a wire partition from floor to ceiling. A. Aleohol treatment of no. 28 began on February 9 with two one-hour periods and continued till April 7. The length of the treatments gradually, but rather irregularly, increased till on April 7 two periods of two hours each were administered. Throughout the whole time this bird was much more severely affected than either of the other males, even when the treat- ments were of the same intensity. ‘Weak’, ‘aimless,’ ‘walked in circles,’ etc., are among the notations. As another indication of the somewhat different reaction of this bird it may be men- tioned that no. 8, no. 27, and the alcoholized hens showed a tendency not to defecate during the treatments, at least after the first fews days. This was not the case with no. 27. No. 8 defecated only once during the whole period and no. 27 only three times, while no. 28 defecated thirty-six times. During a period of fifty-eight days he was kept in the alcohol vapor for a total of 141 hours and 32 minutes (table 1). The hens began laying February 19, and eggs were saved from that date till April 8, during which time 151 were produced. Of these, one laid February 26 was cracked, the other 150 were incubated (table 2). ' EVIDENCE OF GERM CELL SELECTION 399 C. The control period followed immediately, extending from April 9 to June 12, during which time 166 eggs were laid, eleven of which were lost for purposes of the experiment. The data relative to the remaining 155 are given in table 2. Experiment 4. (February 20, 1918, to June 12, 1918). The heterozygous parents were five hens (nos. 21, 22, 23, 24, and 26) from the 1917 chicks, and sisters or half-sisters of males nos. 27 and 28. Unfortunately, they were not uniform in type. All were white and all were brachydactyl and booted, but they varied in comb form and in polydactyly. Nos. 21, 24, and 26 were derived from 1-C, nos. 22 and 23 from 1-A. Nos. 23 and 24 were polydactyl, the others were not. The average index of brachydactyly was 91. The homozygous parent was an exceptionally fine black Minorea cockerel from the same source as the hens used in experiments 2 and 3. Experiment 4 is a reciprocal of experiments 2 and 3. In this case the male produced only one class of germ cells, while the polydactyl females should have produced the eight classes mentioned above and the non-polydactyl hens the classes numbered 3, 4, 7, and 8. Since individual records were not kept, this experiment yields critical data for only brachydactyly and color. C. For the control, eggs were saved from February 20 to April 24; 151 were laid, of which 150 were incubated (table 2). A. Aleohol treatment administered as described in an ear- lier paragraph was begun on April 27. It was planned to give about three hours a day, but this proved to be more than the hens could stand, and individual treatments were frequently cut down tolessthananhour. May 5, by anerror, only one treat- ment was given and on May 31, the morning treatment having been later than usual, it was not thought safe to risk an after- noon treatment. May 7, no. 24, which had recently laid, died immediately after being removed from the inhalation chamber. May 22, no. 22, died in the chamber, and a few days later no. 26 was overcome beyond recovery. Finally, on June 8, no. 23, which had been saved with difficulty on two previous 400 Cc. H. DANFORTH occasions, was lost. With only one hen left, the experiment was discontinued after the morning treatment of that day. Eggs were saved till June 12, the last four being laid by the sole survivor, no. 21. Of the six hens originally intended for this work one (no. 25) died February 20 at the beginning of the control experi- ment and four others were lost as a result of alcohol treatment. While an adequate number of eggs had been secured for C, there were, owing to these accidents, only thirty-nine available for A (table 2). These were all incubated and yielded enough data to justify the inclusion of this, the least satisfactory of the four experiments. RESULTS OF THE EXPERIMENTS In the fourth column of table 2 is shown the number of eggs incubated in each subdivision of the experiments. In the fol- lowing columns the results from these eggs are indicated. Num- bers in the fifth column, headed ‘non-fertile,’ refer to eggs in which no development whatever took place. At the end of three weeks of incubation such eggs were clear and full with a firm yolk; indeed on being broken they presented a more attractive appearance than the average ‘store’ egg. Only such eggs were counted as non-fertile, except in the case of 1-C where the data were originally secured for another pur- pose. Here, in a column headed ‘infertile or dying during the first few hours,’ were entered data that did not differentiate betweeen early death and actual infertility. The other figures in this column, those for 1-A and for 2, 3, and 4, are believed to be strictly accurate. The sixth column indicating embryos killed needs a word of explanation. The capacity of the incubators was not at all times adequate to care for the available eggs. In consequence, since the necessary data can be secured as readily from a fifteen- day embryo as-from a hatched chick, many eggs were opened between the fifteenth and twenty-first days. A few additional embryos were taken in earlier stages. EVIDENCE OF GERM CELL SELECTION 401 The next column, ‘died in the shell,’ shows the number of embryos that failed to hatch. This included embryos of from the first to the twenty-first day, but a considerable number. of them represent late stages, many having pipped without be- ing able to escape from the shell. Since fresh eggs were put in one or the other of the incubators every day, the temperature could not be varied to meet the requirements of embryos in late stages, and in consequence many chicks which were doubt- less originally strong did not hatch. It will be appreciated that the purpose of these experiments was not primarily to produce viable chicks nor to test the vitality of eggs, but rather to test the transmission of certain traits. To that end the main effort was concentrated on bringing as many embryos as possible to a stage where their peculiarities could be determined, the number that actually hatched (last column) being a mere incident. As indicated in table 2, a total of 1228 eggs were used, of which 1100 proved fertile. From these 1100 fertile eggs there were obtained in the manner just explained 808 embryos and 292 chicks. One hundred and ninety-four of the embryos died before the end of the seventh day and were therefore use- less for present purposes. This leaves 906 embryos and chicks which yielded data of value. For all of these the presence or absence of polydactyly was recorded. Eight hundred and- thirty-three of them reached at least the tenth day and fur- nished data on brachydactyly. Finally, 721 developed suf- ficient down so that their color could be determined. The distribution of these traits in the subdivisions of the several experiments is set forth in table 3.3 Table 3 summarizes all the pertinent data and calls for only a brief explanation. It will be understood that the sum of the 3 It seems unnecessary to extend this paper by the inclusion of the detailed protocols which would fill a number of pages. Data for eggs laid each day were recorded as well as the date and (known or estimated) age of every embryo that failed to hatch. For the chicks that did hatch the measurements of each toe were recorded, the index of brachydactyly determined and the grades of poly- dactyly and booting estimated. The data will gladly be put at the disposal of anyone who may wish to make use of them. 402 Cc. H. DANFORTH numbers under ‘polydactyly present’ and ‘polydactyly not present’ is always greater than the sum of the numbers under ‘brachydactyly present’ and ‘brachydactyly not present’ in the same series because of the fact that polydactyly can be determined at an earlier age than brachydactyly, and in all cases some embryos have died between the two critical stages. In like manner, color cannot be determined till a still later period, and in consequence the numbers under this caption are still further reduced. TABLE 3 Distribution of characteristics in embryos and chicks EXPERIMENT BRACHYDACTYLY POLYDACTYLY COLOR Number Part Present ey st Present alee White Black 1 A 56 64 37 89 88 0 C 69 107 73 147 129 0 9 A 52 68 48 79 55 55 C 26 48 30 47 38 35 3 A 52 45 52 48 53 37 C 38 51 37 53 48 37 4 A 18 14 (2) (81) My 16 Cc 48 Ce (28) (105) 63 55 Totals: A and C A 178 191 139 247 120! 108 separately C 181 283 168 352 1491 127 Totals: A and C added 1 These totals are exclusive of the figures from 1—A and 1-C. Since in experiment 4 some of the hens used were not heter- ozygous for polydactyly, the data entered under that head are not comparable to the corresponding data from the other ex- periments. They are included here, but in brackets. In table 4 the data presented in table 3 are converted into such a form as to enable more ready comparison between differ- ent parts of the experiments. There is some question as to how this could best be done, but the following method was adopted. EVIDENCE OF GERM CELL SELECTION 403 First, as to brachydactyly: it was found that upon adding ali the control data together there were 464 individuals of which 39 per cent were brachydactyl. Since this is about the fre- quency of booting in comparable crosses recorded in the litera- ture, it is assumed that 39 per cent represents the normal in- cidence of brachydactyly under conditions such as obtain in the control experiments. Next the total number of cases in each group is found and the probable error calculated on the assumption that 39 per cent represents the true incidence. This gives the fourth column in the table—headed ‘brachydactyl TABLE 4 Percentage distribution of characteristics. Compare table 3 EXPERIMENT BRACHYDACTYL POLYDACTYL WHITE Number Part| Observed Expected Observed Expected Observed Expected 46.7+3.1] 39+3.0 | 29.4+2.7| 36+2.9 100 100 39.2+2.5) 39+2.5 | 33.2+2.1) 36+2.2 100 100 43.3+3.1] 39+3.0 | 37.942.9| 36+2.9 | 50.0+3.2) 50+3.2 35.1+3.7| 39+3.8 | 39.0+3.7| 36+3.7 | 52.1+4.0| 50+3.9 53.6+3.4) 39+3.3 | 52.0+3.4| 3643.3 | 59.9+3.5) 50+3.6 42.7+3.5| 3943.5 | 40.1+3.5) 36+3.4 | 56.5+3.6| 50+3.7 56.3+5.9) 39+5.7 42.8+6.3) 50+6.4 38.4+2.9) 39+2.9 53.4+3.1] 50+3.0 48.2+£1.8) 3941.7 | 36.0+1.7| 36+1.7 | 52.7+2.2) 50+2.2 39.0+1.5) 391.5 | 36.0=1.6) 36+1.6 | 53.9+2.0) 50+2.1 ar lar Or sO of Totals { expected.’ The values in the third column represent the per- centages actually observed in each case with the probable errors calculated for the respective percentages and magnitudes. Thus in 1-A, for example, the chances, as it is generally expressed, are even that in a random sample of this magnitude the per- centage of brachydactyly would fall between 36 and 42 (39-3). The observed percentage is 46.7 with a probable error of 3.1, giving a range of from 43.6 to 49.8. In other words, in the hypothetical case the chances are 7.5 : 2.5 that the value would fall below 42, while in the actual case observed the chances are 404 Cc. H. DANFORTH the same that the true value (to be obtained from an infinite number of chicks produced under identical circumstances) would fall above 48. This may be taken to mean that the chances against the observed discrepancy being simply a chance occurrence are more than 16:1. In 1-C on the other hand the correspondence between observed and expected results is extremely close. Polydactyly and color are treated in the same manner as brachydactyly, it being assumed-on the basis of the controls — that the normal incidence of polydactyly is 36 per cent, and on the basis of genetic literature that the incidence of white color should be 50 per cent. DISCUSSION In the foregoing-sections the purpose and conditions of the experiments have been set forth and the data that they yielded have been presented. We may now examine these data and attempt to interpret their significance. The three characteris- tics especially investigated were brachydactyly, polydactyly, and color. These will be discussed first. Brachydactyly. Table 4 brings out the fact that in each instance the percentage of brachydactyly in A is considerably in excess of that in C, the differences in the four experiments being 7.5, 8.2, 10.9, and 17.9, respectively. The results of either of these experiments taken separately would point strong- ly to the conclusion that alcohol is capable of influencing the percentage of brachydactyly, and the fact that all four of them show such close agreement gives strong assurance that such is the case. If alcohol were without effect, the departures from the normal distribution would not be all in the same direction and the sum of the data from the four A’s should show a percentage approximating that from the four C’s. Such does not prove to be the case. In the C experiments there were 464 individuals of which 39 per cent were brachydactyl, while in the A experiments there were 369 individuals of which 48.2 per cent were brachydactyl. These figures are sufficiently EVIDENCE OF GERM CELL SELECTION 405 large and the difference between the two percentages is sufficiently great to warrant the conclusion that treating parents which are heterozygous for brachydactyly with alcohol vapor re- sults in an increase in the number of brachydactyl offspring produced. It is perhaps significant in this connection that in the first three experiments, those in which the males were treated, the increase in brachydactyly is roughly proportional to the aver- age daily dosage, the greater the dosage (table 1) the higher the percentage produced. In experiment 4, while the actual dosage was less, it was very apparent that the general physio- logical effects were much greater, and this fact seems to be mirrored in the more pronounced increase in brachydactyly in 4-A. It is also of interest that in the three experiments in which the males are involved there is a correlation between the magni- tudes of the C’s and A’s. In other words, if the percentage of brachydactyly is relatively high in one part of the experiment it is also relatively high in the other part, and vice versa. For example, in 3-C the percentage is several points higher than for all of the C’s combined, and in 3-A a similar condition obtains in reference to the total A percentage. In 2, A and C are both below their respective averages while 1-A is intermediate be- tween 2-A and 3-A, and 1-C likewise intermediate between 2-C and 3-C. This point will be reverted to in a later paragraph. The conclusion in reference to brachydactyly that seems justified is that by treating a heterozygous parent with alcohol vapor of sufficient strength the proportion of brachydactyl to normal offspring can be increased. Polydactyly. Three experiments are available for the study of polydactyly. The results of these experiments are not uni- form, and when the percentages for all the C’s and for all the A’s are computed there is found to be an exact coincidence. Such a group of data might well serve to illustrate fluctuations of percentile values in individual samples, and the tendency of these, in a sense provisional, values to approximate the true values as the magnitude or number of samples increases. Such 406 Cc. H. DANFORTH an interpretation here would imply that the alcohol was without effect on the transmission of polydactyly. While this seems to be the most probable conclusion, there are two points brought out in table 4 that deserve attention. One of them is a phenom- enon similar to that. mentioned in the discussion of brachy- dactyly, namely, a positive correlation between the percentage magnitudes of the A and C divisions of each experiment. In 1 are found the lowest values, in 2 intermediate values, and in 3 the highest values. This looks very much as if each of the three males had his own peculiar capacity for producing poly- dactyl offspring. The other point mentioned concerns experiment 3, where in A the percentage of polydactyl chicks was 52 which is a most unusual percentage in a cross of this sort. Here an effect of the treatment seems to be indicated, and such a supposition is strengthened by the fact that when the treatments were stopped the percentage dropped to forty. It has been pointed out that in the case of brachydactyly there is some evidence that the amount of rise in the percentage is dependent upon the strength of the treatment. It is possible that the same is true of polydactyly, but that the level required to produce results— the threshold—is higher. Experiment 3-A, which had the highest daily dosage and in which the treated male was most affected, seems to have been the only one that was sufficiently rigorous to produce an effect on the percentage to polydactyl chicks. For polydactyly it may be said by way of conclusion that the evidence is possibly negative, but that there is some indi- cation that when the alcohol treatment of the heterozygous parent is sufficiently intensive the relative number of poly- dactyl young is increased. . Color. In experiment 1 the female parents were homozy- gous for the dominant white characteristic of the Leghorns, while the treated male parent was a homozygous dark recessive. Since each parent could produce only one kind of germ cell, there was no chance for selection and the expectation of 100 per cent of white chicks was realized in the 217 individuals whose color was determined. In the other three experiments EVIDENCE OF GERM CELL SELECTION 407 the treated birds were all white and consequently, since they were derived from the cross made in experiment 1, heterozy- gous for color. These individuals should have developed two kinds of germ cell with reference to their color-producing potentialities, thus affording material for selection. Tables 3 and 4 show the results obtained. They do not reveal any ob- vious effect of the alcohol unless it be in experiment 3, where the percentage in A is considerably above what one would expect. 3-C also shows a high percentage, and when it is recalled that C followed immediately upon A, the probability that the high percentage in both cases is due to a common cause rather than to chance is somewhat increased. Except in 4-A, where the total number of individuals in- volved is only 28, the percentage of white chicks does not fall below 50 in any of the six separate experiments, the average being over 53 per cent. This is rather close to expectation, but the constant upward tendency of the white is at least no- ticeable and suggests a possible inherent superiority of the white producing germ cells. No very certain conclusion seems warranted in regard to the effect of alcohol treatment upon the transmission of color but in the experiment in which the treatment of the male was most severe there is some indication that the germ cells bear- ing determiners for the dominant character functioned more commonly. : In reviewing the four experiments, no. 3 will be seen to have yielded the most striking results throughout. In this experi- ment, not only did the three characters, brachydactyly, poly- dactyly, and color, show indications of the action of alcohol vapor upon the germ cells, but the fertility of the eggs was also markedly affected (table 2). This latter point is of interest in connection with Pearl’s results. One of the most constant features reported in his paper is a regular elevation of the per- centage of infertile eggs in his various alcohol series. When the present work was begun similar results were expected, but as will be seen by reference to table 2, they were realized only in experiment 3. 408 Cc. H. DANFORTH There are several considerations that may serve to explain this discrepancy. Pearl studied the effects of alcohol acting over long periods, while in the present work it was desired to have the treatments extend over the shortest possible periods consistent with securing a statistically sufficient number of eggs. Since Arlitt and Wells (’17) have shown that in the rat testis cells in different stages of spermatogenesis are affected differentially by alcoholic poisoning, it may be surmised that injury done to cells in early stages of gametogenesis may not show any effects in breeding tests for relatively long periods. It is possible on this assumption that Pearl’s data were taken mostly after the full effects of the aleohol had been established, while most of my experiments were stopped too soon to get the later results. The data from experiment 3 are in harmony with this supposition, for in C, which followed immediately after A, the percentage of infertile eggs was actually higher than it was during the time when alcohol was being administered. Another possibility that may be mentioned is differential susceptibility to alcohol. It might be inferred from Pearl’s paper that one hour is practically the maximum time that a fowl can live in an atmosphere saturated with alcohol vapor. The males used in experiments 1, 2, and 3 were able to endure much more than that, but of these three males no. 28 showed unmistakable signs of being most severely affected. It may be, therefore, that although stronger dosages were administered, the actual physiological reactions were less except in experiment 3. That male no. 28 was not naturally infertile is shown by his subsequent history. During July and till the 10th of August he was mated to no. 21, and this hen gave no infertile egg till August 26, sixteen days after having been separated from the male. Aug. 10, he was mated to eight black Minorcas, the remnants of experiments 2 and 3. Eggs from this mating . were saved beginning Aug. 19. Despite the fact that during much of the time the weather was unfavorable and the male was moulting, out of the 101 eggs laid only three were infertile, in marked contrast to the more than 27 per cent of non-fertile eggs during the experiment. Since the number of hens in the EVIDENCE OF GERM CELL SELECTION 409 original experiment was smaller than that used in the supple- mentary test, it does not seem likely that there was a dearth of sperm at any time during the experimental periods, but more probable that some of the eggs were entered by sperm too badly injured to develop a pronucleus or at least a viable conjugation nucleus. An egg ‘fertilized’ by such a sperm would very likely give no other reaction than an egg that had been reached by no sperm whatever, and would be recorded as non-fertile. Pearl’s statements as to the non-effect of alcohol on the trans- mission of Mendelian characters are clearly meant to apply only to the data he presents, and those data were derived from experiments obviously neither intended nor adapted for the solution of the problem attacked in this paper. Since he dealt with crosses between pure-bred homozygous strains in which all the Mendelian characters were in stable equilibrium, there was no chance for selection except on the basis of such characters as vitality and vigor. With reference to these characters, Pearl found selection to be possible. In the experiments now being reported the material afforded a chance for selection be- tween other traits, in this case Mendelian, and the results are believed to show that here also selection is possible. The question still remains as to the precise nature of the selection that takes place. Cole and Davis (’14) have pro- duced, evidence that with rabbits the sperm of one male may have greater fertilizing capabilities than that of another, even when conditions would seem to be more favorable for the latter. They have also shown that the fertilizing power of the sperm can be influenced by poisons administered through the male soma. But in their work also, homozygous males seem to have been used and no evidence is presented as to whether or not two kinds of sperm produced by the same male could be dif- ferentially affected. These and other results obtained by Cole and his colaborators, Stockard’s findings, and the conclusions reached by Pearl, all tend to suggest that the effect of poison- ous reagents is lethal rather than stimulating, if such is the case, we may assume that a germ cell or nucleus bearing a determiner for brachydactyly is more resistant to the effects of alcohol treatment than one not bearing such a determiner. 410 Cc. H. DANFORTH On this assumption it might be expected, although it need not necessarily follow, that germ cells carrying determiners for two or more characters of selective value would be more favorably circumstanced than those carrying only one such determiner. The available data is not sufficient to throw much light on this question, but so far as it goes it would seem to in- dicate that such may be the case. For example, in 3-A the. combination brachydactyl-polydactyl-white is represented by sixteen individuals where the expectation is thirteen on the basis of observed percentages and only six on the basis of expected percentages. The question as to the time in gametogenesis at which the selection takes place must also be left unsettled. The pur- pose of these experiments was to try to select between mature germ cells, and this seems to have been accomplished, but there is also some indication that the effects of the treatments have persisted for a period greater than the probable life of such cells. Moreover, there is some evidence, too meager perhaps to be given much weight, that ova, or their nuclei, are likewise selected. This may mean that the alcohol is effective as a selective agent as far back as the first maturation division and possibly determines at that time which nucleus will remain in the egg of the female or which will be the more effective sperm produced by the male. In this connection may be re- called the finding of Arlitt and Wells (loc. cit.) that, in the rat at least, the alcohol affects stages of spermatogenesis in the reverse order, attacking late stages first, and early stages last. There is one interesting by-product of these experiments which should perhaps be further emphasized, the indication of individual idiosyneracies in the transmission of traits. Among breeders a belief in ‘prepotency’ has occasionally had currency, but the possibility of this supposed phenomenon being real is often discounted by geneticists. When the data for these experiments were being tabulated it was observed with some surprise that they might be interpreted as furnish- ing evidence of something analogous to prepotency. It would EVIDENCE OF GERM CELL SELECTION 411 not be inconsistent with the conclusions of this paper if it should turn out that within certain limits heterozygous individuals may produce functional germ cells of reciprocal classes in dif- ferent proportions, the average ratio for all individuals being in most cases about 50 : 50, but in brachydactyly, for example, 39:61. Such a tendency, unless very marked, would be masked in ordinary breeding experiments since individual departures would be attributed to chance and lost in the totals. The data in the present paper happen to be so arranged that there are eight opportunities to note the correlation between per- centages produced by the same male parents in two differ- ent tests, and in all of these cases the correlation is positive. This is a matter that seems to be worth investigating and these data suggest a favorable method for attacking the prob- lem. To test the question properly, many matings between pairs or small groups of animals of proper gametic constitution should be made, and the data from these matings broken up into. blocks that could be compared with each other and with similar blocks from other matings. SUMMARY AND CONCLUSIONS These experiments were planned to test the possibility of selec- tion between the different classes of germ cells produced by a heterozygous parent. For a selective agent alcohol vapor, which, inhaled through the lungs, is believed to pass directly into the circulation and thence to the fluids surrounding the germinal tissues, was administered to fowls of the desired genetic con- stitution. As an index of any selection that might occur, the relative proportion of certain traits, brachydactyly, polydactyly, and white color, appearing in the offspring produced during periods of treatment was compared with the proportion of those traits produced during control periods. The results indicate that with at least some traits selection is possible and and that it is more rigorous the more severe the treatment. Since, under certain conditions, here artificially produced, it appears that germ cells with different genetic potentialities 412 Cc. H. DANFORTH react differently, a possibility of far-reaching importance is suggested: that, even under normal conditions, the genes which determine the genetic potentialities of a germ cell may have a real survival value for that cell and, moreover, that the prevalence of certain traits appearing in the adult may be in the final analysis largely regulated by the advantage or dis- advantages that the determiners for such traits confer upon the germ cells in which they chance to be lodged. LITERATURE CITED Aruitt, ADA Hart, AND Weis, H. Gipzon 1917 The effect of alcohol on the reproductive tissue. Jour. Exp. Med., vol. 26, pp. 769-778. Coun, L. J.. AND Bacnuuser, 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., vol. 12, pp. 24-29. Cougs, L. J., anp Davis, E. L. 1914 The effect of aleohol on the male germ cells studied by means of double matings. Science. N.S., vol. 39, pp. 476-477. Danrortu, C. H. 1919 The relation of brachydactyly to other characteristics in the domestic fowl. Am. Jour. Anat., vol. 25, pp. 97-115. Nice, L. B. 1917 Further observations on the effect of alcohol on white mice. Am. Nat., vol. 51, pp. 596-607. PEARL, Raymonp 1917 The experimental modification of germ cells. Jour. Exp. Zo6l., vol. 22, part I, pp. 125-164; part II, pp. 165-186; part III, pp. 241-310. Srockarp, C. R. 1913 The effect on the offspring of intoxicating the male parent and the transmission of the defects to subsequent generations. Am. Nat., vol. 47, pp. 641-682. Stockarp, C. R., anp Parantcotaou, G. 1916 A further analysis of the hered- itary transmission of degeneracy and deformities by the descendants of aleoholized mammals. Am. Nat., vol. 50, part I, pp. 65-88; part II, pp. 144-177. 1918 Further studies on the modification of germ-celis in mammals: The effect of alcohol on treated guinea-pigs and their descendants. Jour. Exp. Zodl., vol. 26, pp. 119-226. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 1 GENETIC STUDIES ON THE MEDITERRANEAN FLOUR-MOTH, EPHESTIA KUHNIELLA ZELLER P. W. WHITING Zoological Laboratory of the University of Pennsylvania ONE FIGURE AND TWO PLATES CONTENTS Eo Tntroductions acest ion desws 3c PRR Es 2 On, Fen? ar eae Ot ae 414 Aes Distribution and WGaxOn OM. 40+be eee ee erin elite 414 Be SOUPCe. OL gne-INALCHIAl «2... cas seis see ee ce eee itis seine s 416 Ge Teehniqued 20 6022.8 sn see ee EO es Pen siereee 417 II. Observations, experimental data, and conclusions..................... 419 A. Description and origin of variations noted....................-. 419 Ie Descriptionvolnvanianons seeeeer ce eerie sere 419 a. COLGES?...55 (a cee ee ae eee eee eee Ort skate 419 Os BiB iL IS Ady Re SO Aon as SE ET SSE AR 419 ¢ Fa ey | OVE mee Ge Se ee | ee a 419 df GemaGaliis oe.) lin ole eee MA tN ok nen eas apts oe AN da 420 €3, MOvEn: paris: vss cahcnieie some eee me ates erik w sae Ee 2. Analysis: of original stocker i400 Ae & aie ee aod erane widens 421 a. Bussey stock. Defective mouth parts.............. 421 b. Lowell stock. Defective mouth parts and sooty | a2 SCN Rtn As I ean I lr TR i lated ees Lisa nd 421 c. Calvert stock. Cleft tongue and sooty base........ 422 d. Washington stock. Defective mouth parts......... . 422 e. Strain A of Washington stock. Black color and dark PAT RECA ery eta cn & ie Seria es cpa ens an ate aaa 422 B. Tests for the hereditary nature of variations noted............. 423 [-orleredibys of colon. vaTniationse hea. scsi. actos cmb ee kleneete 423 a Blick ~ sa cene fs FERRI Pe ap Ds ctl esbaebeeed. clans 423 b>¢ Darke md =arears cree Shiite ued bso | ott aura reins 423 E=SOOLY DASC ar na eee ernie eens chy CER. ka ae oP RRE 424 2. \Heredity of oralidefectstiae, eee eae cee elaclslocte ne rok 424 a. Cleft tongue. The masking of a Mendelian differ- ence by environments)... jaccue io. eblas ss feet ao4 6). Defective palpo nd. 2AM CULE SRT Oe 426 C. Tests for linkage of cleft, black, and sooty..................... 426 1. Free segregation of cleft and black.......................- 427 2. Free segregation of cleft and sooty.................2.00008 427 3. Apparent complications of black and sooty................ 428 413 414 P. W. WHITING D. Analysis of the gradations of sooty base.....................00. 428 E. Reversal of dominance of sooty by black....................... 430 IL... Summary .andsceneral discussion. 52. 6/4020 «csc. lou. 2 Gs 6 ocetematee ene 434 A. The masking of a Mendelian difference by environment.......,. 434 | By By BESa cin GOLDIN AT CEs cp mjoreireneenioc kong s oboe ae SEE 435 C. Analysis of a case of continuous variation...................... 435 D:.. Dhetphysiology of color production: ...1.). 5 5:- N57 ae assay. vee ck 435 I. INTRODUCTION A. Distribution and taxonomy The Mediterranean flour-moth, Ephestia Kihniella Zeller, is widely distributed and very destructive to stored cereals. It was noticed in Paris in 1840 and in Constantinople in 1872. An outbreak of the pest occurred at Halle, Germany, in 1877, where it was supposed to have been introduced with some American wheat. The American origin was assumed by European writers for a number of years, but it is probable that the insect was rather widely distributed in Europe for some time, being noticed only when it became especially abundant and troublesome. It was not officially reported in America until 1889 (Canada). The insect belongs to the subfamily Phycitinae of the large family Pyralidae. Its nearest important relatives are the dried- currant moth, Ephestia cautella Walker, the chocolate moth, Ephestia elutella Hiibner, and the Indian-meal moth, Plodia interpunctella Hiibner. In June, 1877, Professor Kihn, of the university at Halle, sent a number of specimens to Zeller whose description appeared in 1879. It is thought advisable to quote the original description of the moth since the taxonomic type agrees with the genetic type as treated in the present paper. Color characters with which this paper is concerned have been placed in italics in Zeller’s description quoted below. Major, alis elongatis, ant. cinereis, strigis 2 obsoletis dilutioribus, obscurius marginatis: priore ante medium posita, oblique, subserrata, posteriore superne fracta, margini postico nigro-punctato admota, puncto venae transversae nigro gemino saepe in strigulam mutato, umbra subfasciata ab eo introrsus ad dorsum demissa; post. albidis, sub- hyalinis, ramis venae medianae griseis. o& @ GENETIC STUDIES ON FLOUR-MOTH 415 Grésse der Homeos. nimbella. Kopf und Riickenschild von dem lichten Grau der Vorderfliigel. Stirn gerundet. Taster aufgekriimmt, anliegend, hellgrau, an den Seiten des 2. und 3. Gliedes ausser an der Spitze schwarzlich. Sauger stark, auf dem Riicken hellgrau beschuppt. Fiihler grau, undeutlich geringelt. Hinterleib heller als der Thorax, mit weisslichem Bauch. Mannlicher Analbusch schmutzig weissgel- blich; Genitalzangen ansehnlich, langlich l6ffelf6rmig, mit gelblichen, verlingerten Schuppen reichlich bekleidet. Legestachel weisslichgelb, lang hervorstehend. Beine grau; Fiisse aussen dunkler mit weisslichen Spitzen der Glieder; Hinterschienen zusammengedriickt, am Ende durch Haarschuppen erweitert; vor der Spitze aussen schwarzlich. Vorderfliigel iiber 5 und bis fast 6’” lang (bei emem @ nur 4’), gestreckt, mit sanft gebogenem, nach hinten starker convexem Vor- derrand, wegen der Fransen ungefahr rechtwinkliger Spitze und sanft convexem Hinterrand. Grundfarbe hellgraw ohne andere beigemischte Farbe. Der erste Querstreifen, in der Mitte zwischen Basis und Querader- punkt, ist schrig, etwas nach aussen gebogen, undeutlich sdgezdhnig, auf der Subdorsale mit scharfem, einspringendem Winkel, heller als dre Grundfarbe, auswarts gewéhnlich nur bis zur Medianader brert schwérz- lich geséumt, darunter im Subdorsalwinkel mit einer deutlichen schwarzen Ausfiillung. Die ewer schwarzen, fast senkrecht tiber einander stehenden Queraderpunkte (der untere grésser und oft léngsstrichformig) vereinigen sich oft zu einem oben verdiinnten Querstrich; von dessen unterem Ende oder dem unteren Punkt lauft ein etwas breiter, nicht immer deutlicher, schwarzer Schattenstreif eimwarts zur Mitte des Innenrandes. Der zweite helle Querstreifen, dem Hinterrande ndher als der Querader und jenem parallel, macht in seinem obern Drittel einen scharfen, nach aussen offenen Winkel; er ist etwas sdgezdhnig und auf berden Seiten, auf der hintern vollstindiger als auf der vordern, mit schwarzen Aderstrichen gesiumt. Vor dem mit groben, schwarzen Punkten eingefassten Hinter- rande sind die Lédngsadern oft fast alle schwarz. Fransen einfarbig hellgrau. Hinterfliigel spitz, weisslich, durchscheinend, mit verloschener, grauer Hinterrandlinie, welche sich um die Spitze bis in den Vorderrand zieht; auch die Medianader mit ihren Aesten ist grau. Basis beim ohne Haarbusch. Fransen weisslich, an der Wurzel mit feiner, gel- blicher Linie durchzogen. Haare des Abdominalrandes oft sehr blassgelblich. Unterseite der Vorderfliigel einfarbig, schimmernd hellgrau, ganz an der Wurzel mit einigen einwirts verlingerten Haarschuppen des Vor- derrandes. Hinterfliigel weisslich, am ganzen Vorderrand bis zur Subcostale lichtgrau. Beide Geschlechter sind wenig verschieden; nur ist das 29 gewéhnlich das scharfer gezeichnete: In figure m is shown a specimen which closely approximates the type as described by Zeller. It is probable that his speci- 416 P. W. WHITING mens may have been a little lighter than the average of my own, for it was the transverse light bands (‘‘strigis 2 obsoletis dilutiori- bus’ and “Der erste Querstreifen,...heller als die Grundfarbe.. . . Der zweite helle Querstreifen,’”’) rather than the dark, that ap- peared especially to strike his attention. In general later de- scriptions emphasize the transverse dark bands. For example, Miss Ormerod (’89) describes the fore-wings as “pale gray with darker transverse markings.”’ The typical ground color of the fore-wings is gray or ashy (“ant. cinereis’”’ and “Grundfarbe hellgrau ohne andere beigemischte Farbe,”) in my specimens and according to all descriptions available, although Riley (89) mistakenly says “in the typical specimens raised by Zeller the ground color is pure yellow or nearly brownish.” I have not been able to correlate with sex any color character such as mentioned by Zeller. The markings of both sexes appear to me to be equally distinct. Numerous articles concerned chiefly with the economic im- portance of the moth are referred to in the bibliographies of American economic entomology by Nathan Banks. The experiments described in this paper have been carried on by aid of a Harrison Research Fellowship of the University of Pennsylvania. My thanks are due to the members of the Zoo- logical Department for their interest and suggestions. B. Source of the material The moths used in the experiments belonged to the following stocks: Bussey stock. On January 28, 1915, five adult moths were placed in a glass jar 4 inches in diameter and 5 inches high with tin cover screwed down tight. The cover must have admitted but slight circulation of air. The jar was half filled with white flour. It was set in a dark closet at the Bussey Institution, Boston, Massachusetts, and was not disturbed until June 2, 1916. At that time there were many insects in all stages of develop- ment. It is probable that the culture might have lasted much longer as the flour was by no means exhausted. GENETIC STUDIES ON FLOUR-MOTH 417 Lowell stock. A box of ‘Cream of Wheat” from Lowell, Mas- sachusetts, was found in July, 1916, to be infested. Calvert stock. A large tin box of flour heavily infested was given me by Professor Calvert in the fall of 1916. Washington stock. A culture was obtained from the Bureau of Entomology, Washington, D. C., in June, 1916. Strain A of Washington stock. A pair of moths was isolated from the Washington stock on July 11, 1916. The female was probably not virgin. A mass culture was made from the progeny. During July and the first part of August, 1916, the moths were bred and studied at the Marine Biological Laboratory, Woods Hole, Massachusetts. From then until the end of Sep- tember they were left in mass cultures. The results described below are from matings made at the Zoological Laboratory of the University of Pennsylvania. C. Technique Various methods of rearing the moths have been tried. It has been found convenient to use glass candy jars with an inside measurement of 45 inches in height by 4 inches in diameter. A small amount of cereal is placed in the jar and the etherized moths are set upon this. If progeny are produced, as may be easily determined by the appearance of webs, more cereal is added. Pupation normally occurs in silken tubes spun in the cereal, but overcrowding or lack of food sometimes causes the cater- pillars to wander. The moths emerge and rest upon the glass. They do not fly unless disturbed, so that it is an easy matter to collect them in a shell vial. After several moths have been thus secured they are turned into a wide-mouthed bottle containing. ether fumes. The etherized moths may be studied under a binocular and sorted out for recording and pairing. Once fertile eggs have been obtained from a pair, there is little difficulty in rearing the larvae. The jar has but to be set in a warm, humid place. The adult moths are, however, extremely sensitive to environmental conditions. Mating apparently occurs 418 P. W. WHITING at any ordinary temperature or humidity, but the females are very perverse about egg-laying. The proper conditions have not yet been determined, and consequently the results obtained have been a chance selection from a very large number of pairs set. Usually not more than 5 or 10 per cent of the pairs prove fertile, but I have occasionally had as good a ratio as 50, 60, or even 70 per cent. The most frequent condition of infertility is the failure of the female to oviposit. Examination has been made of a large number of such females. The abdomens have been found to be filled with large eggs apparently mature and normal. Another peculiar condition is the failure of the eggs to hatch unless almost all the eggs are laid. It has been found that if the female retains a large proportion of her eggs, the eggs which are laid do not hatch. The pairs have in this case frequently been observed to mate, but it is possible, nevertheless, that the eggs have not been fertilized. Records are being kept of all these conditions and further studies will be made. If conditions are warm and humid, moths begin eclosing five weeks after the parents have been isolated. Considerable vari- ation obtains in rate of development of the larvae from any one pair, so that moths of one fraternity are sometimes eclosing over a period of one, two or even three months. It is probable that this tends to compensate for the high sterility, for if conditions are not favorable for oviposition at one time they may be at another. ‘There is thus no difficulty in keeping mass cultures. Even extreme reduction in food with consequent reduction in size of the moths does not exterminate them. Despite the extreme sensitiveness of the adults, the species is adapted to tide over very unfavorable conditions. Since mating takes place as soon as the wings are dry, females are counted as virgin only when found in cultures before a male has emerged. GENETIC STUDIES ON FLOUR-MOTH 419 II. OBSERVATIONS, EXPERIMENTAL DATA, AND CONCLUSIONS A. Description and origin of variations noted 1. Description of variations. a. Color. During the studies at Woods Hole and later the moths were examined for vari- ations. It was soon noticed that there were many minor dif- ferences in wing color and pattern and that these appeared to be hereditary. Certain cultures produced moths darker than others, while some had the transverse bands of the wings very well marked. These differences persisted regardless of the nature of the food. A few color variations were very well marked. The variety shown in figure n has been called ‘sooty base’ or ‘sooty.’ The base of the primaries is black and the outer margin is much darkened. ‘There is also a decided tendency for the intermediate area to be lighter than in the type. The factor producing this variation, S, proves dominant to type. Another variation tends to darken the mid-area of the pri- maries, and possibly to lighten the base and outer margin. It has not as yet been studied satisfactorily. In some specimens it is very pronounced, but in others it grades into type. It is of interest because it has an effect the reverse of sooty base. It has been called ‘Dark mid-area’ or ‘Dark.’ A black variation, acting as a simple recessive, 6, is shown in figure o. The upper side of the primaries is black. The upper side of the secondaries is slightly darkened. The under side of the wings is light gray or white as in type, but some black ap- pears along the costal margin. The legs and body are black or eray. The homozygous sooty black, SS.bb, is shown in figure p. 6. Size. Variations in size are probably due to lack of suffi- cient food, since small moths have come out of certain crowded cultures. When virgin matings of these were made they pro- duced moths of normal size. c. Leg spines. ‘Two spines occur at the tips of both the middle and the hind tibiae and two occur in the middle of the hind tibiae. They were studied in many hundreds of moths from 420 P. W. WHITING mass cultures and individual matings. The number was always constant, and only slight variation occurred in length and divergence. d. Genitalia. No secondary sexual characters could be found. The sexes may be readily distinguished, however, by the claspers of the male, as shown in figure b, and the ovipositor of the female, as shown in figure e. Occasionally a moth was found that had peculiar genitalia. Examination showed that these were males in which the claspers were shortened, twisted, or lacking. In- ternal sexual organs were apparently normal and spermatozoa were present in all. e. Mouth parts. The adult moths do not feed. The mouth parts consist of the tongue, which is formed by the maxillae; the maxillary palpi, which are small and inconspicuous; and the three-jointed labial palpi, which conceal the maxillary palpi. The normal condition of the mouth parts is shown in figures ¢ and f. Many variations were observed in the tongue and in the labial palpi. The latter will be referred to hereafter as the palpi. Figure a shows a condition in which the palpi are fused in the median plane. The joints of each are fused with the correspond- ing joints of the other, so that there results a large flat three- jointed median palpus. The tongue arises from the normal position above the insertion of the palpus and appears to be normal in every way. Its coil is pushed aside by the palpus. Figure d shows a condition in which the palpi are asymmetrical. The two terminal joints of the left are lacking. The right is normal. Figure | shows a condition in which the palpi are both shortened symmetrically. All sorts of variations in the palpi may occur, due to loss or shortening of the joints. The tongue may be lacking altogether, as shown in figures 7 and k. This variation occurred in some of the mass cultures early in the work. Much variation occurs also in length. The elements of the tongue, maxillae, which are normally united to form a tube, may be separated to any extent. The separation may occur at the tip only or from the tip any distance towards the base. In some cases also the basal or middle part may be GENETIC STUDIES ON FLOUR-MOTH 421 divided while the distal part is normal. Failure of the maxillae to unite is apparently due to malformation. Figure g shows a condition in which the palpi are very small, the tongue is cleft to the base and a short distance from the base the elements diverge laterally. In figure 4 the palpi are normal; the maxillae are separated distally. Occasionally a moth fails to shed the pupal covering of the head. It may otherwise be quite normal. Figure 7 illustrates this, showing a ventral view in which also the maxillae are separated and coiled up at the sides. Other variations are swellings on the tongue or antennae, straight tongue and scaleless areas on wings. 2. Analysis of original stocks. Summaries of the earlier results will be given in order to show the origin of the variations studied. a. Bussey stock. Defective mouth parts. Nine non-virgin type females from the Bussey stock, when isolated, produced 841 type, 448 males and 393 females; 6 with defective palpi, 2 males and 4 females; 5 with deformed tongue, 2 males and 3 females; and one male with defective genitalia. Defects of palpi, tongue, and male genitalia, therefore, occur in the Bussey stock. Twenty virgin females from the Bussey stock were paired with males, either from the Bussey, the Lowell, or the Washington stock. They produced 473 type, 255 males and 218 females; 3 with defective palpi, 1 male and 2 females, and 1 female with cleft tongue. Each of these abnormalities occurred among the offspring of a different mating, so that no significant Mendelian ‘ratio appeared. b. Lowell stock. Defective mouth parts and sooty base. Ten non-virgin type females from the Lowell stock, when isolated, produced 917 type, 452 males and 465 females; 24 with defective palpi, 13 males and 11 females; 7 with deformed tongue, 4 males and 3 females. Defects in palpi and tongue therefore occur in the Lowell stock as in the Bussey stock. A pair of type produced 55 type, 25 males and 30 females; and 19 with cleft tongue, 9 males and 10 females. Another similar mating produced 27 type, 17 males and 10 females, and 8 with cleft tongue, 4 males and 4 females. This amounts to 422 P. W. WHITING 82 type and 27 with cleft tongue, suggesting that the latter may carry a factor acting as a simple recessive. Among the moths of this stock were also noticed several of the sooty base variety (figure n). A pair of these produced 2 sooty, 1 male and 1 female, and 1 type, afemale. This indicates that sooty may be dominant to type. c. Calvert stock. Cleft tongue and sooty base. The moths of the Calvert stock were rather light in color. Some of them had wings with sooty base. Two non-virgin females produced 101 moths with normal mouth parts and 2 with defective palpi. A sooty male was paired with a type female whose virginity was not certain. There were produced 7 type and 6 sooty. A pair of these type moths produced 9 type and 2 with cleft tongue, suggesting again the recessive character of cleft. The Calvert stock is of interest as it is the source of the factor for sooty base used in later experiments. d. Washington stock. Defective mouth parts. Sixteen non- virgin females from the Washington stock, when isolated, pro- duced 1523 type, 744 males and 779 females; 11 with defective palpi, 7 males and 4 females; 3 with deformed tongue, 2 males and 1 female, and 1 female with defective palpi and deformed tongue. Two other non-virgin females were isolated from a fraternity consisting of 44 moths with normal mouth parts. One of these produced 163 type, 85 males and 78 females; 7 with cleft tongue; 5 males and 2 females, and 2 with defective palpi, 1 male and 1 female. The other produced 116 type, 54 males and 62 females, and 40 with cleft tongue, 26 males and 14 females, thus closely approximating the Mendelian three to one ratio. [ee 2 a a en GH OR MATES Hamper gk Average weight em ber Average weight os VO a Be Litter 25 A?B? days grams grams per cent 26 3 28.6 3 Zhe 4.7 41 3 54.6 2 47.0 16.0 63 3 88.6 2, 78.5 12.8 90 3 130.3 2 95.5 36.4 120 3 Liv 23 pe 125.0 25.8 150 3 190.3 2 148.5 29.1 180 3 195.0 2 156.5 24.6 Litter 38, A1B! 31 2 35.0 2 31.0 12.9 44 2 59.0 2 51.0 15.6 60 2 89.0 2, 85.5 4.0 90 2 eS 2 111.5 oes 120 2 161.5 2 148.5 Sea 150 2 194.5 2 177.0 9.0 180 2 199.0 2 182.5 9.0 examination of this percentage column shows that only in four instances was the average female weight equal to or greater than the average male weight of a given litter. The weight of the males in these four cases, not only reached that of the females before the age of sixty days, but increased gradually as growth continued. By using the average weight of each sex of each litter at comparative ages as given in table 2, the average weight of the males and females of all seven litters has been expressed in the form of a growth curve in figure 1. The broken lines from thirty to sixty days is only approximately correct due to the fact that all litters were not weighed at exactly the same age prior to the sixtieth day. MOORE CARL R. 464 "4901100 Ajoyeurrxordde ATUO st 9AAND oY} Jo 4aed sty fad¥v ouT¥S OY} A]JOVXO 7B poYySIOM Jou 919M S19}4I] [][B ‘Buruurseq oy qu ‘yeyy JoRy OY} 07 Onp st skvp A4xIs 04 AZATY} UOAF S9AIND oY} JO 4Avd UOYyoIG oY, “SeAND oy} Aq poyuosoidor are pu poyndwm0d 919M SOB’ UOAIS 7V S10}} 1] [[V 1OJ XOS OVO JO JYSIOM OFVIOAG OY} SI} WAT f(Z o[Gu}) SAvp UT oBe poyeusisop OY} YB 10741] YOVI JO XOS YOO JO YYSIOM OSVIOAG OY} SUISN AQ PoJONI}SUOD 919M SOAIND OY, ‘“SeyvUey poAvds Jo ouo IOMO] OY} ‘SOTBVUT poyBIySBdO JO YZ MOIS OY} SjUosotdad 9AINO Joddn oy, ‘SABP UI STRUIIUG OY} JO OBB OY} OSSIOSqR oY} PUR qysiom Apoq jo survids yuosoIdod soPVUIPIO OYJ, *S}VA O]VULOJ PUB O[VUT pozZTUIOJOOpBUOS Jo YZMOIS Jo OAD T ‘Sy O8T 02) og! Osi Ovi oct Oz! oll oot 06 038 02 09 os Ov of 0c OF 0@ Ov : ble « \ x N\ \ ba ‘ x \ ‘\ 09 7 08 00} ost i 2] be 2 | a EE a3 ag EE 4 ¢ 4 7 ¢ 002 GONADS AS CONTROLLERS OF CHARACTERISTICS 465 DISCUSSION The primary object of these experiments was, first, to de- termine to what degree constant weight differences between the normal male and female rat were determined by the sex glands, and, second, to provide a basis for interpretation of weight differences in case of homoplastic transplantation of the gonads. In studying the effects of the influence of the gonad of one sex on modifying the somatic and psychical development of the opposite sex (a repetition of the experiments of E. Steinach), the writer has differed from Steinach in the interpretation of the results obtained. Steinach has placed considerable emphasis on modification of body weight of rats and guinea-pigs following removal of the normal gonads and the substitution of the oppo- site one by transplantation. If these transplantations were successful—i.e., if the graft persisted and grew—according to Steinach, the male became ‘feminized’ and, compared with normal males, relatively decreased in weight as development proceeds, while the female became ‘masculinized’ and corre- spondingly increased in weight in comparison with unoperated females. These changes from the normal weight for the sex he associates with the presence of the secretion of the implanted gland; the female increased in weight because a secretion from the testis was present in the female into which it had been placed, and having this male secretion the weight of the indi- vidual increases toward the normal weight of a male and away from that of a female. In case of a secretion of the ovary in a male animal, the weight of this feminized male approaches that of a normal female. Stotsenburg (’09), however, has shown for rats, that the presence of the secretion of the testicle has absolutely no in- fluence upon the growth of the individual. Also (13) he has proved that the mere removal of the ovaries of young rats results in an increase of from 17 per cent to 33 per cent compared with unoperated females. Considering these findings, the writer has found it impossible to associate weight differences with different degrees of maleness 466 CARL R. MOORE or femaleness in all cases after transplantation of the gonad of the opposite sex. The female increases in weight not because of the influence of the secretion from the transplanted testis, but solely on account of the removal of the ovary, which alone seems to have any influence upon the growth of the animal. There seems to be no doubt that the presence of the ovary does prevent the normal ascent of the growth curve. In order to know whether this sex difference in weight was due entirely or only in part to the influence of the secretions of the gonads, the pre- ceding experiment was carried out. The results show very conclusively that there is a real difference between the capacity of the two sexes to accumulate somatic materials when there are no secondary influences that may be attributed to the influence of the gonad. As table 2 shows, this difference has been exhibited at each stage by each of the seven litters used. It is interesting to consider this potential weight difference of the ‘determined male’ and the ‘determined female’ in their development in the light of Riddle’s theory of sex.’ Riddle and his co-workers have demonstrated actual differences in the chemical constitution of male-producing and female-producing eggs of the pigeon. He has not only shown that the female- producing egg contains a greater phosphatide content and a lesser percentage of water, but he has also demonstrated that these chemical differences found in the dimorphic ova of birds are carried over into the adult life of the bird. His idea is that sex determination is based upon a higher rate of metabolism of — the ovum producing a male than of that producing a female, and several researches are cited to show that the same relative rate of metabolism persists in the adult male andfemale. It has occurred to the writer that these basic differences in weight of the two sexes of rats may also indicate a possible difference in metabolism inherited from the original ovum from which each had been developed. But aside from the primary differ- ences that may exist in the determined male or female, the secondary influences that make the female a more apparent 5 Riddle (’17). 6 Lawrence and Riddle (16). GONADS AS CONTROLLERS OF CHARACTERISTICS 467 female and the male a more apparent male are due to the presence of the specific gonads, and these differences, in many _eases, have proved themselves capable of being controlled to a certain extent. CONCLUSIONS After early removal of sex glands the growth curve of the determined male is (without exception in these experiments) higher than that of the determined female. There is, there- fore, a real difference (of metabolism?) in the two sexes, which may represent an inherited difference from the original ova, but this difference may be accentuated by the presence of the ovary in the female. Hull Zoological Laboratories, The University of Chicago, February 17, 1919. BIBLIOGRAPHY Donaupson, H.H. 1915 Therat. Memoirs of the Wistar Institute of Anatomy and Biology, no. 6, Philadelphia. LAWRENCE, J. V., AND RippLE, Oscar 1916 Sexual differences in the fat and phosphorus content of the blood of fowls. Amer. Jour. of Physiology, vol. 41. Moore, Cart R. 1919 On the physiological properties of the gonads as con- trollers of somatic and psychical characteristics. I. The rat. Jour. Exp. Zoél. vol. 28, no. 2. RippLE, Oscar 1917 The theory of sex as stated in terms of results of studies on pigeons. Science, vol. 46, no. 1175, pp. 19-24. Srermacn, E. 1910 Geschlechtstrieb und echt sekundire Geschlechtsmerk- male als Folge der innersekretorischen Funktion der Keimdriisen. Zentribl. f. Physiol., Bd. 24, S. 551-556. 1911 Umstimmung des Geschlechtscharakters bei Siiugetieren durch Austausch der Pubertatsdriisen. Zentrlbl. f. Physiol., Bd. 25, 8S. 723- 125: 1912 Willkirliche Umwandlung von Sidugetier-Minnchen in Tiere mit ausgepragt weiblichen Geschlechtscharakteren und weibliches Psyche. Pfligers Archiv. f. d. gesammte Physiol., Bd. 144, S. 71-108. 1913 Feminierung von Mannchen und Maskulierung von Weibchen. Zentribl. f. Physiol., Bd. 27, S. 717-723. StrotsensurG, J. M. 1909 On the growth of the albino‘rat (Mus norvegicus var. albus) after castration. Anat. Rec., vol. 3, p. 233. 1913 The effect of spaying and semi-spaying albino rats (Mus nor- vegicus albinus) on the growth in body weight and body length. Anat. Rec., vol. 7, p. 183. 1917 Observations on the influence of isolated ovaries on the body growth of the albino rat (mus norvegicus albinus). Anat. Rec., vol. 12, p. 259. Resumen por el autor, D. D. Whitney. Universidad de Nebraska. La inefectividad del oxigeno como factor causante de la pro- duccién de machos en Hydatina senta. ’ Los cultivos del flagelado microscépico Chlamydomonas des- prenden cantidades considerables de oxigeno libre cuando se eolocan a la luz solar, hasta tal punto que el agua del cultivo puede contener hasta + 16 cc. de oxigeno libre por litro, mientras que en la oscuridad no hay desprendimiento de dicho gas. El autor coloc6é los rotiferos en dichos cultivos, a la luz solar y en la oscuridad, sirviéndoles de alimento los Chlamydomonas. Los primeros, aun a pesar de estar en un medio rico en oxigeno libre, originaron menos hijas productoras de machos que los colocados en la oscuridad, en los cuales la cantidad de oxigeno libre:es mucho menor. Este resultado se debe probablemente a la alimentacién. En la luz solar los individuos de Chlamy- domonas se reunen en la superficie del agua y en las paredes de la vasija de vidrio que los contiene y de este modo se hacen in- nacesibles como alimento de los rotiferos. Por el contrario en la oscuridad permanecen nadando activamente en el agua del cultivo durante 3 a 4 dias y los rotiferos pueden comerlos facil- mente. La cantidad de hembras productoras de machos result6 ser la misma en el agua de cultivo con una cantidad minima de -oxigeno libre (1 a 3 cc. por litro) y en la que contenfa una can- tidad mayor de este gas (2 a 8 cc. por litro). . Translation by José F. Nonidez . Columbia University ’ AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 2 THE INEFFECTIVENESS OF OXYGEN AS A FACTOR IN CAUSING MALE PRODUCTION IN HYDATINA SENTA? DAVID D. WHITNEY Depariment of Zoology, University of Nebraska, Lincoln, Nebraska Recent papers by Shull and Ladoff (16) and Shull (’18) main- tain that a sufficient amount of dissolved oxygen in the culture water will cause an increase in the production of males in the rotifer Hydatina senta. Whitney (’17) explained these results obtained by Shull and Ladoff as being due to the effect of oxygen upon the food supply which in turn effected the production of males. Later, Shull made further experiments which seem to demonstrate that oxygen is really a potent influence in causing males to be produced. Not being convinced, however, that the problem was finally settled, new experiments were undertaken by the author, which in their turn seem to show that oxygen in itself is not effective in causing males to be produced. METHOD In all of the experiments old culture water was used which had been made out of rain water and horse manure several weeks or months previous to the beginning of the present experiments. This culture water was such that when it was a few weeks old rotifers readily lived in it, and they also lived in it just as readily when it was several months old provided food was put into it. The best food supply in these experiments seemed to be a mixture of green flagellates, Chlamydomonas, and colorless flag- ellates, Polytoma. A pure diet of Chlamydomonas used for several successive days was detrimental to the rotifers. Many of them would gorge themselves on this food so that their stom- achs would burst and allow the contents to fill the whole body 1 Studies from the Zoological Laboratory, The University of Nebraska, no. 121. 469 470 DAVID D. WHITNEY cavity, thus causing death, while in others the Chlamydomo- nas would form a dense and compact mass in their stomachs which also caused death. By mixing a small proportion of. Polytoma with the Chlamydomonas, both of these troubles were avoided. The Chlamydomonas was raised in large quantities in bouillon solution in direct sunlight and the Polytoma was raised in stable tea in darkness. The details of rearing both of these flagellates have been published in former papers. In many of the experiments the amount of food was measured in a graduated pipette. The Chlamydomonas and the Polytoma were each separately centrifuged and all of the original culture water drained off. Then just enough old stable-tea culture water was added to allow the Chalmydomonas and Polytoma to be drawn up into the pipette. In this way the amount of food could be quite accurately measured and regulated at will in each experiment. It was found that the Chlamydomonas could be transferred from the sunlight to absolute darkness and would remain alive and active for several days, the duration of activity being somewhat dependent upon temperature. The amount of oxygen, number of cubic centimeters per liter, whenever determined in the experiments was determined by the Winkler method described in Standard Methods of Water Analysis published by the American Public Health Association of Boston. The sodium thiosulphate solution was standardized against potassium dichromate about three times per week. The following experiments are not arranged chronologically, but are so arranged as to present the evidence and data in a logical manner. EXPERIMENTS SHOWING THE PRODUCTION OF OXYGEN BY CHLAMYDOMONAS IN DIRECT SUNLIGHT It was considered desirable to determine how much free oxy- gen is given off by Chlamydomonas when the culture is in the direct sunlight for several hours. Varying quantities of Chlamydomonas were put into about 50 cc. of old stable-tea culture water and poured into stender dishes, 1 inch in diameter, OXYGEN AND MALE PRODUCTION 471 and piaced in a pan of running water in the direct sunlight. The running water maintained a temperature of 20° to 25°C. The old stable-tea culture water had been standing in a north light for several months and had only a small amount of free oxygen init. This made an excellent starting solution. Usually after a short time in the sunlight the Chlamydomonas would emit free oxygen in sufficient quantities to rise to the surface in minute bubbles. After several hours in the sunlight there was formed usually a frothy scum on the surface which was composed of these minute bubbles of oxygen. Table 1 shows some of the details and the results of these few experiments. Experiments 1 and 3 show the amount of oxygen generated in periods of one hour, two hours, and four hours, TABLE 1 Showing that in sunlight the green flagellates, Chlamydomonas, give off considerable quantities of free oxygen tn old stable culture water that is devoid of all food substances D 1 1a io 2 fie |eele =| LOTS LIGHT CONDITIONS TIME, 1918 Ft SI = a Zw ey De | S02 /be| oe : se | de | 52) 28 a 5 8 8 6 Wee hae bee ea ee, cc cc cc cc A | North light Several weeks 42 |} 0 42 | 5.31 1 B | Clouds and sunshine | 2-3 p.m., Oct. 31 50 42 | 6.14 C | Clouds and sunshine | 2-4 p.m., Oct. 31 50 42 | 8.55 2 Sunshine Several hours, Oct. 30} 50 42 |14.80 A | North light Several weeks 421 0 42 | 4.81 3 B | Sunshine 10-11 a.m., Nov. 1 50 42 | 6.6 C | Sunshine 10-12 a.m., Nov. 1 50 42 | 7.41 {| D | Sunshine 12 m.—4 p.m.,Nov. 1 50 42 |16.56 4 Sunshine 9 a.M.4 p.m., Nov.9 | 50] 1 42 |14.80 A | North light Several weeks 42 | 0 42 | 3.24 B | Fair 10 a.m.—3 P.M., Dec. 16} 50 | 0.25} 50 | 6.56 5 C | Fair 10 a.M.—3 p.m., Dec. 16} 50 | 0.50} 50 | 9.85 D | Fair 10 a.m.—3 P.M., Dec. 16} 50 | 0.75) 50 |13.13 E | Fair 10 a.m.-3 P.M., Dec. 16} 50 {1.0} 50 |16.42 472 DAVID D. WHITNEY while experiment 5 shows the amount of oxygen generated by varying quantities of Chlamydomonas during a five-hour period. It may be readily seen in the last columns of the table that the amount of free oxygen is greatly increased in the sunlight. TABLE 2 Showing that old culture water free from decomposing materials and containing only a small quantity of free oxygen gradually absorbs additional free oxygen from the surrounding air ' o = CULTURE | OXYGEN 2 LOT TIME, 1918 WATER PER WATER i = TESTED LITER cc. cc 7 | Tap-water 1 A 4p.m., Nov. 9 42 hs B G02 Tank rain-water 4p.m., Nov. 9 42 A 4 p.m., Nov. 7 42 1.51 | Old culture water | Unfiltered 4:15 p.m., Nov. 7| 42 2.27 | Old culture water | Filtered 9 C |} 5p.m., Nov. 8 42 4.55 | Old culture water | Filtered D | 4p.m., Nov. 9 42 6.83 | Old culture water | Filtered 11 a.m., Nov. 10 42 7.59 | Old culture water | Filtered F | 5 p.m., Noy. 11 42 7.90 | Old culture water | Filtered 9 a.m., Dec. 17 50 2.62 | Old culture water | Unfiltered 10 a.m., Dee. 18 50 4.92 | Old culture water | Filtered 10 a.m., Dec. 19 50 6.56 | Old culture water | Filtered 9 a.m., Dec. 20 50 6.56 | Old culture water | Filtered 3 3 P.M., Dec. 24 50 4.96 | Old culture water | Filtered 3 P.M., Dec. 25 42 6.64 | Old culture water | Filtered 3 p.M., Dec. 26 42 7.75 | Old culture water | Filtered A B Cc D (| Alt Seu. Dee 23 50 10 | Old culture water | Filtered B C D 3 P.M., Dec. 25 42 3.32 | Old culture water | Filtered 3 P.M., Dec. 26 42 4.80 | Old culture water | Filtered 3 P.M., Dec. 27 42 6.64 | Old culture water | Filtered 3 p.M., Dec. 28 42 7.38 | Old culture water | Filtered 3 P.M., Dec. 29 42 7.75 | Old culture water | Filtered on HOQWSe OXYGEN AND MALE PRODUCTION 473 EXPERIMENTS SHOWING THAT WATER CONTAINING A SMALL AMOUNT OF FREE OXYGEN WILL ABSORB ADDITIONAL FREE OXYGEN FROM THE SURROUNDING AIR IN DARKNESS It is not only important to determine the amount of oxygen generated by Chlamydomonas in the sunlight, but it is also equally important to determine the amount of oxygen that is absorbed from the air by the culture water when in darkness. When Chlamydomonas and Polytoma were added to the old stable-tea culture water which contained only a small quantity of free oxygen and then the culture was placed in darkness for several days the quantity of oxygen increased in the culture water several cubic centimeters per liter. This is shown in tables 2, 5, and 7. In the absence of light Chlamydomonas does not carry on photosynthesis and consequently does not give off free oxygen. Clear old stable-tea culture was taken, in some experiments it was filtered and in others it was used unfiltered, and the amount of free oxygen determined at the beginning of each experiment. Then several stender dishes containing about 50 cc. of this water was placed in darkness and at successive intervals of twenty-four hours the contents of a dish was tested for free oxygen. Table 2 shows that the culture water gradullay ab- sorbs free oxygen from the air throughout the three to four days’ exposure until it usually amounts to from 7 to 8 ce. toa liter. EXPERIMENTS SHOWING THAT FEWER MALES IN HYDATINA SENTA ARE PRODUCED IN SUNLIGHT, WHERE PRE- SUMABLY THE AMOUNT OF FREE OXYGEN IS HIGHER, THAN IN DARKNESS WHERE THE AMOUNT OF FREE OXYGEN IS LOWER In these experiments large-mouthed bottles about 13 inches diameter were used. Into each there was put a mixture of 50 to 60 ce. of filtered old stable-tea culture water, Chlamydomonas, and a little Polytoma. In the bottles in the sunlight about 1 to 1.5 ec. Chlamydomonas were put in order that there might be a large quantity of oxygen generated. In the bottles in 474 DAVID D. WHITNEY darkness not as much Chlamydomonas could be added because much of it would die, probably from lack of oxygen, decompose and befoul the water so as to prevent the normal growth and reproduction of the rotifers. In these experiments ten adult rotifers were put into each bottle of culture water and food and one bottle was placed in a pan of running water in the sunlight and the other bottle was placed in darkness at room temperature. Both of these bottles were left undisturbed for six days, then each was well stirred and a few drops of the liquid immediately taken out and the sex of the rotifers in these drops were recorded. In table 3 it is seen that in the sunlight the males constituted 8 + per cent of the rotifer population of 1736 individuals, while in the darkness the males constituted 28+ per cent of the population of 1654 individuals. TABLE 3 Showing that rotifers kept in darkness where there 1s no production of free oxygen by the Chlamydomonas produce more males than rotifers. do which are kept in the sunlight in the midst of considerable quantities of free oxygen that is given off by the Chlamydomonas 1 To 2 cc. OF CHLAMYDOMONAS AND A FEW DROPS OF POLYTOMA In 50 cc. CULTURE WATER IN EACH EXPERIMENT EXPERI- MENTS TIME, 1918 Sunlight Darkness Number | Number |} Per cent | Number Nee Per cent of 9 of # of A of 9 of fA to) 1 April 9-15 47 8 14+ 27 25 48+ Z April 15-21 266 28 9+ 284 78 21+ 3 April 17-22 fll 15 8-++ 152 67. 30+ 4 April 17-22 129 26 16-++ 65 18 21+ 5 April 17-23 175 31 15+ 223 85 27+ 6 April 18-24 37 12 24+ 38 12 24+ 7 April 20-27 40 2 4+ 61 9 12+- 8 April 28—May 5 110 9 7+ 70 24 25+ 9 April 29—May 5 200 f 1+ 42 42 50 10 April 30—May 6 150 3 2 83 56 40+ 11 May 1-7 150 2 1+ 71 32 31+ 12 May 2-8 65 3 4+ 14 8 36+ 13 May 10-16 50 3 5+ 50 18 26+ Rotel eis. ety Bee ay eee 146 8+ 1180 474 | 28+ OXYGEN AND MALE PRODUCTION 475 In a former paper it was shown how a high per cent of males could be obtained by feeding Chlamydomonas in the sunlight, and now opposite results are obtained! ‘These contradictory results are due to differences in manipulations of feeding and also to different conditions of the Chlamydomonas itself as a food in these two instances. In the former experiments the Chlamy- domonas. were put into the culture water with the rotifers and the rotifers fed upon them for only a few hours, during which the Chlamydomonas were kept actively swimming toward the lighted side of the dish as it was rotated upon a kymograph. In these later experiments the feeding conditions were quite different. No rotation of the bottles was made, and when this is not done many of the Chlamydomonas swim to the lighted side of the dish and adhere to the surface of the glass. This enables all such individuals to escape from being eaten by the rotifers. Sometimes the rotifers were able to pick up only a small number of stray Chlamydomonas in direct sunlight. On cloudy days and in darkness the Chlamydomonas are more active and are more available as food for the rotifers. In ex- periment 6 of table 3 cloudy weather prevailed throughout the last five days that the rotifers were in the experiment. Prob- ably the Chlamydomonas remained active in the diffuse daylight so that the rotifers were enabled to feed upon them as readily as in the darkness. This would explain why the per cent of males is equal in each lot. When the Chlamydomonas remain in the sun for several days in the old stable-tea culture water, which is devoid of all nutritional substances, very little, if any, reproduction occurs among them, but each individual becomes of full size and is covered with a tough covering which renders it indigestible for the rotifers. Consequently, if nearly all of the Chlamydomonas attach themselves to the side of the dish, the rotifers are unable to obtain a superabundance of food or, on the other hand, if some of the Chlamydomonas remain active they develop such a tough covering as to render them nearly unfit as food for the rotifers. In either case the rotifers are not overfed in the sunlight. While, on the contrary, in darkness all of the Chlamydomonas 476 DAVID D. WHITNEY remain active and small and do not develop the tough covering. Thus there is a fine food supply for the rotifers for several days. The matter of the food supply probably explains the sex ratio. When the supply is good as in darkness more males are produced and when it is poor as in sunlight fewer males are produced. In all of these experiments in the sunlight much free oxygen was given off during the day and rose to the surface forming a frothy scum. Table 4 shows the light conditions throughout these experi- ments. TABLE 4 Showing light conditions of the days during the experiments in lable 3 meee ee || eee vote, 101 eee April 9-14 Sun April 26-27 Cloudy April 15 Cloudy April 28-May 4 | Sun April 16-18 Sun May 5 Sun and clouds April 19-24 Cloudy May 6 Sun April 25 Sun Table 5 shows the details and results of another set of experi- ments in the sunlight and darkness in which the quantity of free oxygen in the culture water was determined both at the be- ginning and at the end of many of the experiments. The amount of the food was more accurately measured than in the experiments of table 3. Small amounts of Chlamydomonas, 0.10 ec., and of Polytoma, 0.05 cc., were put into 10 cc. of the culture water and well stirred. Then 3 ce. of this mixture was added to 50 ec. of the filtered old stable-tea culture water. To the lot that was placed in sunlight 0.2 to 1 ec. of additional Chlamy- domonas was added. Each lot was divided into three portions and put into three vials, 1 inch diameter, and allowed to stand twenty-four hours and then three or four rotifers were added. These experiments were started in the morning and thus afforded an opportunity for the Chlamydomonas to generate sufficient oxygen in the sunlight to quite thoroughly charge the culture water with it before the rotifers were added. It also enabled many of the Chlamydomonas to become attached to the sides *" OXYGEN AND MALE PRODUCTION AT7 TABLE 5 Showing that in sunlight where 2 to 18 cc. of free oxygen per liter is present about 43 per cent of male-producing females are produced, while in darkness where only 2 to 8 cc. of free oxyyen per liter is present about 78 per cent of male-producing females are produced DIRECT SUNLIGHT DARKNESS be s ra 2 Daughters & A 3s Daughters z PERIOD TIME, 1918 5 g FS = 5 By 8 a z = Be 2 logla [se 2 5 BEE. |Bslocloo| 82/231 8. Sele elae| go | 33] PS [55 BOISS| 2S i155 BN | 12) fo) A ee) NO) Kee re Ay cc cc cc. cc. ce Maint b 1 Beginning | 10 a.m., Oct. 17 12 12 End 10 a.m., Oct. 20 9) 11/55 4| 16/80 a Beginning | 10 a.m., Oct. 20 3 3 \| End 4 p.m., Oct. 23 29] 15/34-+ 12) 25/67+ 3 Beginning | 10 a.m., Oct. 20 4 4 End 11 a.m., Oct. 24 27| 13|82+ 4) 19|82+ vil Beginning | 11 a.m., Oct. 21 4 4 \| End 10 a.m., Oct. 25 28) 8|22+ 15} 21/58-+ 5 Beginning | 10 a.m., Oct. 22 5 5 End 10 a.M., Oct. 26 28} 12/30 7| 33|82+- 6 Beginning | 10 a.m., Oct. 24 4 4 || End 10 a.m., Oct. 28 9} 1/10 1} 787+ = Beginning | 10 a.m., Oct. 24 4 4 1 End 10 a.m., Oct. 28 10} 8/42+ 3} 17/85 8 Beginning | 10 a.m., Oct. 28 16 16 End 10 a.m., Oct. 31) 42] 5.64 17| 33/66 | 425.64 15| 35/70 9 Beginning | m., Oct. 28 12 12 End 10 a.m., Nov. 1} 42} 5.97 25] 25/50 | 42/4.81 19} 31/62 101 Beginning | 10 a.m., Oct. 29 12 12 (| End 4p.m., Nov. 2} 42/12.11 35| 15/30 | 4215.33 14) 36/72 u Beginning | 10 a.m., Oct. 30 12 12 End 5 p.m., Nov. 3] 42/11.72 10| 38|79+| 42/5 .86 9} 34/79+ 12 Beginning | 10 a.m., Oct. 31 12 12 End 5 p.m., Nov. 4| 42] 7.81 24] 26/52 | 4216.25 4| 46/92 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 28, NO. 3 478 DAVID D. WHITNEY ; TABLE 5—Concluded DIRECT SUNLIGHT DARKNESS s | g Daughters 8 a ES Daughters n 3 a 3 pes 5 PERIOD TIME, 1918 z g 5 Sap e eS 5 = = ius) eS 2.) 2 : 28 lee Boles ¢ [88 Zo = Bel Sn [eso 9[ro| 8" |35] &, (28/9 e/r9! 8 ey a8 \upentas BD ISS] PS IS8 5" Ss Oko. ie my IO 10 {4 my ce cc. cc cc 13 Beginning| 9 a.m., Nov. 1} 42] 2.15) 12 42/215] 12 \| End 3 p.M., Nov. 5] 42] 5.86 30] 20/40 | 42/5.33 2| 48/96 “4 Beginning | 9 a.m., Nov. 2] 42} 0.97} 12 42|0.97} 12 End 3 p.M., Nov. 6] 42) 7.81 33| 17/34 | 42/5.93 6} 44/88 15 Beginning | 10 a.m., Nov. 3} 42] 2.65] 12 42/2 .65] 12 \| End 3 p.m, Nov. 7| 42| 7.44) | 14] 3672 | 426.45) | 3] 47\94 16 Beginning | 8 a.m., Nov. 10} 42} 3.03) 13 42|3.03} 13 End ° M., Nov. 14 42)11.72 28) 22|44 | 42/7.42 12| 38/76 17 Beginning | 8 a.m., Nov. 11} 42) 4.31] 13 42/4.31] 13} 1 “\| End 3 P.M., Nov. 15] 42/11.72 32| 18/386 | 42/7.81 10} 40/80 ie Beginning | 11 a.m., Nov. 12] 42] 3.59] 13 42/359] 13 End 4 p.m., Nov. 16] 42} 7.81 32| 18/36 | 42/7.81 9} 41/82 19 Beginning | 10 a.m., Nov. 18] 42} 3.50} 13 42/3 .50) 18 \| End m., Nov. 17 | 42] 8.21 27| 23/46 | 42/8.21 11| 39/78 20. Beginning | 9 a.m., Noy. 14] 42] 3.12} 13 42/3 .12} 13 U End 4 p.m., Nov. 18} 42/15.63 36] 14|28 | 42/7.81 18| 32/64 PLAT. fs cccyes Toertiaions cls os arene g+* 483/373)/43-++ 6+* 178|649)78+- * Average at end. of the vials and thus render the available food supply scarce for the rotifers when they were added at the end of twenty-four hours. Three days after the rotifers were added about fifty young females were isolated in watch-glasses and the sex of their offspring recorded. In many of the experiments the Chlamydomonas in the sun- light had so spent themselves at the end of four days that no more free oxygen was found in the culture water than was found OXYGEN AND MALE PRODUCTION 479 in the culture water of the lots in the darkness. In others, however, a high per cent of oxygen was found at the end of four days. Several determinations of free oxygen were made in the mornings of cultures similar to the above and a small quantity of free oxygen was always found as in experiments 8 and 9. Thus showing that the excess quantity of free oxygen escapes from the culture water during the night and in the mornings no more is found than would have been found in such culture waters if exposed merely to the air. In the cultures in the sunlight much free oxygen was generated during the first three days, which was very evident by the frothy scum on the surface of the water. During this time the rotifers were subjected for many hours each day to a high per cent of free oxygen. During the night the excess of oxygen gradually escaped, but in the sunlight of the following day a new excess of oxygen was generated. The rotifers in the sunlight were subjected to perhaps 10 to 15 ec. of oxygen per liter for a period each day, while the rotifers in the darkness were never subjected to more than was absorbed by the water from the air, 7 to 8 cc. of oxygen per liter. Bear- ing this in mind, it is of considerable interest to compare the sunlight lots with the darkness lots in individual experiments as in 10, 17, and 20 of table 5 or to compare the average results of the total summary of all the experiments. In all the experi- ments excepting no. 11 the per cent of male-producing females is much lower in sunlight where there is an excess of free oxygen than it is in darkness where there is no more free oxygen than can be absorbed from the air. The results of experiment 11 are not clear, inasmuch as there was brilliant sunlight throughout the four days of the experiment. Table 6 shows the light conditions throughout these experi- ments of table 5. 480 DAVID D. WHITNEY TABLE 6 Showing light conditions of the days during the experiments of table & roe, 101 ae vie, 108 Feeegeriie October 17 Sun Oct. 29-Nov. 3 Sun October 18-19 Cloudy November 4 Clouds and sun October 20-21 Sun November 5-7 Cloudy October 22 Clouds and sun November 8-9 Sun October 23 Cloudy November 10 Clouds and sun October 24 Sun November 11-13 | Sun October 25-27 Cloudy November 14 Clouds and sun f| a.™., sun November 15 Cloudy egg ooo \ | p.m, cloudy November 16-19 | Sun EXPERIMENTS SHOWING THAT CULTURE WATER CONTAINING A VERY LOW PERCENTAGE OF FREE OXYGEN YIELDS AS MANY MALE-PRODUCING FEMALES AS CULTURE WATER CONTAINING A MUCH HIGHER PERCENTAGE OF FREE OXYGEN The experiments of the preceding tables 1 to 6 may be of interest, but the crucial test of the effect of oxygen in causing male-producing females to appear is really made in these experi- ments in table 7. All of these experiments were carried on in darkness and the food and culture-water conditions were the same as those in the darkness experiments of table 5 with the exception that the culture water containing the food was put into one 14-inch vial instead of being divided and put into three vials. The vials in lots A were not stoppered, but were kept open so that the surrounding air came into contact with the surface of the water, but the vials of lots B were closed with tightly fitting ground-glass stoppers. A small quantity of air, 3 ce. to 0.5 ec., was left in each vial. In some experiments the vials in lots B were inverted. The inclosed air bubble was changed every morning and evening, otherwise both the food and the rotifers would soon have died from lack of a sufficient oxygen supply. Thirteen female rotifers were put into each vial at the begin- ning of each experiment and allowed to remain three days undisturbed. At the end of that time, fifty young females were OXYGEN AND MALE PRODUCTION 481 selected at random and isolated in watch-glasses. In some ex- periments there were fewer than fifty young daughter females produced, and in such eases all of the young females were iso- lated as in lots B of experiments 13 and 14. An equal number of young females were isolated from the control lots A. The quantity of free oxygen in the culture water was de- termined both at the beginning and at the end of each experi- ment in both lots A and lots B. Old culture water was used which was made about the middle of the previous August and which contained only a small quantity of free oxygen. After this culture water was filtered it absorbed additional free oxygen from the air in lots A, while in lots B, in the stoppered vials, the quantity of free oxygen was diminished when the air bubble was made small enough as in lots B of experiments 9 to 17 and 19. In experiments 9, 10, 12, 13, 14, and 18 the oxygen supply became so low in lots B that none of the rotifer eggs hatched until the vials were opened and additional oxygen was supplied. In lots B of experiments 8, 11, 15, 16, 17, and 19 fewer eggs had hatched than in lots A, in all of which there were more than fifty young females at the end of three days. In lots A the free oxygen increased by absorption from the air to from 6 to 8 ce. per liter, while in lots B it ranged from 6 to 1 ec. per liter at the end of the experiments. However much the two lots, A and B, varied in their oxygen content, the per cent of male-producing females produced was about equivalent in each lot of the individual experiments and also in the general average of the summary of all of the experiments. In fact, the per cent of male-producing females in the summary of all lots A and lots B is practically identical. Such equivalent results in two parallel lots of rotifers, even under the same conditions, never previously have been obtained by the author. The most striking result of these experiments is the pro- duction of such a high per cent of male-producing females in culture water that was nearly depleted of free oxygen. The highest per cent among fifty young females was ninety-two. It is recorded in lot B of experiment 5. Only two or three lots exposed to the air exceeded this. Se a Ee ee psizr [8 |+os 0S'§| Z| Ol] & | 09 | 08 | 02 Is 2) 2 | Ystp uedo GG AON “W'd G pug \y 9 SI |0S'€| a | Ol a €1 [oo §| a | ystp uedg 61 “AON “Wa g | Suruurdeg | | z6oF |— |+0¢ 0g S| Ze] STi & |-F6 PLE] e Gh 2) ce | Yysip uedoO IZ “AON “Wd ¢ puq yi el (62 | Stl a@ €I |€6°2| Z| ysrp uedo 8ST ‘AON “Wa g | Suruuseg zsilt (6 |+0¢ 799 ch | STl a | $2] 281 St 18'2| Z| yYstp uedo 0Z “AON “Wd G pug |) a le SI 0 F 2h | Stl & EI |O$'F| Z| Yysip uedo LT ‘AON “WV QT | Suruulseg f PS\ZF (8 |+0¢ 80'S] ce | O'S} A | 96] SF] 2 0°24) ZF | ysipuedd| 61 “AON “Wad F puq vie s &I |2o | ce | Os a El |ZG°| a | ystp uedo QI “AON “WV 6 | SuruUIseg & A gojee IZT |+0¢ 98'S] 2 | O'S A | 881 WI” I8'2| cb | ysrp uodg SI “AON “Wa g pug ae = €l legs] a | OS! & €1 |2G'e| GF | Yystp uedg GI “AON “Wd fF | SuluUIseg & =e A 8216e ITT |+0¢ lv'S| ce | O€| A | 28] 116 182] Gb | ystp uodo LT “AON “N'V OT pug ie : A €I |so's| ce | o's] a SI [80'S] cb | ysip wedQ | FT ‘AON “Wa F | Suruuseg | / a 02 22 *09 929 22 , a —$ [| ——— | —— | | | ——_ | —————_ ———_s | —<—.$§s [J ———.-s J ——-s J ————_ | | | OE TF eens Be eke teel fle} S| Bo elt) s et eie |g 5 7/0/8581 608 pS + w a 5 40 +0 a8 =I ES eS g ZT/So| o 5 5 g So] 3 Ee Ss i= oe | oe) Bole 3 S Soot ae | 72 Ke. 2 eR Bolger eee S | 3 z Q, $a le lee o Q, Oo} = o “a oF 3 + a g + SIGI ‘WIL ‘qoruda SLOT Be | “6 a B a sioqyysneq | 8s = siozqsneq 5 ATddOS UIV AAHSINIWNIG SSANMUVAG NI IOULNOD és wayyy dad uabhixo aasf fo 99 4 02 @ BururmjUod sajom ainyna ur poanpoud ssafyos sv “uaa tad +z ‘saypwaf a 4Suronpoid-ajpw huni sp paonpoud 4334) tad uablixo aatf fo +99 ¢ 0} [ fo mnurmru oy) Bururzuod sayom ainzyno ur ssafyos 94) Burmoys st 4 ATaViL OXYGEN AND MALE PRODUCTION +9961 SFC 88/63 +€g\¢ +9\tF 9L/8E + E819 +OL|TE SE ial &t OL 92 ai &1 al ST 0Z vP +0¢ €1 €1 real é1 é1 €1 SI €1 &I rea re N NX o> = (=) a oo ‘0 S oo aa 44 SS joaee) feayyjea) Oa QQ OQ OO oo “. faaee) 98 OF cS 98 +£6 v8 08 9¢ OL cP 0G 61 &F cT 1G GV OF 83 GE O& él &% OT GG “ald §1 reat é1 &1 €1 &1 é1 é1 &I €1 GP ystp ystp ystp ysip ysip ystp ystp ystp ystp ystp ysip ystp WAY Wetp ysip ysip ystp ystp ysip ystp ueadgQ uadgQ usdg uedg uedg uedg uedg uedg uedg usdg usdg uadg uedg uedg uedg uedg usdg uedg uedg wedg ‘09q 09 20] “AON ‘20q “AON ‘20q] “AON "AON “AON "AON ‘AON “AON “AON "AON “AON “AON “AON “AON “AON “d oD OL IT pu sumuisoq Pun Suruuisogq pua Ssuruuiseg pad suLUUIsIg pud suuuiseq pu suluUlseg puq Suruuisog pud ‘Suruursegq pug Suruuisoq pug suuuisog cl iat €1 rat IT OL WHITNEY DAVID D. Ree ee ee ee ee ee ee Eiara rss © eve] e) 4;.0,.0ce/felvalsaelalinve (o)taini 6 .o)1s)\s)a a SLene ye) s) = a pua 4% WBBIDAY ogee fete] ‘lee [Koere ‘ae 0 callie!» @ 6 (sil = te aell| ee . . ere ene iwisleve aa elete Gowen sora «ele ‘aime 9s. e,e)8 Suruuis9q 48 WSVIOAV Arvuiuing ess @ ac8 .éils)) 0) @uetielie=s@ eille « ° . se'c wratteuehie se) al wid) bila’ W \etiets! =) eo!) (a) elle Bl @ ene ote ers! aver eo) ai.e. (e) Bete e evens, OSBBIOAYV +F2/819/Z1Z +H) 099 12e eee eeoevae OE i Me aga tgs OR NO) ZL|9E |FL | 8% 19°%| ZF |I-S'0| A | 92) 88 | ZT GIL) 2 | stp uedo 2 00q “Wd puq hy aI EI |S&°Z| Z |I-G'0| A et ieez| we | 6oysip uedg |. = f “90q| “WV IT] | Suluusog +esicg |Z | 0 90°€| zP |I-G'0| & [+22] 62 | TT Gt2| a | stp uedo g ‘00q “Wd T pugq hy SI EI |29°%| Gb |I-G'0] a SI |29°2| ce | stp uedo g 0eq “WV T] | Zuruuseg ZSiIF |6 | ZT OL'T| 2 | 0] a | O8| OF | OT 6r'2Z| ze | ustpuedg| = =¢ ‘90q “Wd fF puq hy a5 EI |29°2| | Sol a EI |29°2| ZF | Ystp uedo Z 00g “WV 6 | SutuUIseg 02 °02. 00 soo. 09 PlolcleZ zig tele |e|els)clezgi2| & Sere e ee eT] Se | ash tg CP Be eee eg = eelee) 2 | ¢ | 8 5 sm Bead | => wr | oA oa © i} + a7 co} ¢ sok g eR aig] ¢ 2 Billed q Q aa =) Ee g Q, ee c o pale g fe dite SIG ‘ANIL ao1mad SLOT Ss ge|oe| | & 0 pie ee oS 3 o B ey sioqqaneq ee e r s1oyqsneq e a ATddON8 HIV GCAHSINIWIA SSANYUVA NI IOHLNOO 484 a a es ee ee EE papnpuog—L ATAVL OXYGEN AND MALE PRODUCTION 485 EXPERIMENTS SHOWING THAT THE FREE OXYGEN CONTENT OF WATER IS NOT INCREASED WHEN CHLAMYDOMONAS IS ADDED After the preceding experiments had been completed. it was realized that perhaps there was a sufficient amount of free oxy- gen retained within the individual cells themselves of Chlamy- domonas to exert an appreciable influence when the Chlamydo- monas were transferred to other culture waters. In order to test this possibility, the experiments recorded in table 8 were performed. The Chlamydomonas was allowed to remain in the sunlight for several hours; then it was centrifuged, its culture water drained off, sufficient quantity of water added to liquefy it, and definite quantities of it added to various kinds of water. The water used was mainly rain-water, which varied widely in the free oxygen content. The rain-water that had been standing in the pipes from a large storage tank contained less than 1 ce. of free oxygen per liter, while rain-water in battery jars which had been exposed to the air for several days contained as much as 7 or 8 ce. of free oxygen per liter. Chlamydomonas was added to these two kinds of rain-water. Oxygen tests were made im- mediately with some unfiltered lots containing Chlamydomonas, while other lots were first filtered before being tested. In a few experiments Chlamydomonas was allowed to remain in the water about ten minutes, while in others it was allowed to remain about four hours before the tests were made. _ It was found that a considerable error was introduced by filtration. The quantity of oxygen was increased even in the most hurried filtration and was increased very markedly if the filtering process was prolonged for a few minutes, especially in the water that contained a very small quantity of free oxygen at the beginning of the experiment. In some experiments the water was decanted before testing for oxygen, but this method was only feasible when a sufficient time had elapsed to allow the Chlamydomonas to settle to the bottom of the bottle. In none of these experiments, when the error due to filtration was taken into consideration, was there found any evidence to WHITNEY DAVID D. 486 svuoul | -opAure[yy) Jo uorippe sanoy zy ref | Aep Zurmp ung | #2°¢ | eo] zP | JayJe SoyNUIUL OT potoz[iy | WI Sarpuvys 10yVM-urey AU Waa | | simoy ZZ rel \y Aep Zuiinp ung | e¢°¢ | GQ) at persy,guy) | Ur Surpurys aoyem-urey | Z-ure “wa ZF] g sanoy Z ref 88°L; O0| GF peisy[yuy) | UL Surpueys s0yem-urey L-wer way | Vv |] sAep [eiodes rel Aep Sump ung | of'G | OL! ZF petez[yuy) | Ul Surpueys 19zeM-ureyy g'uee “Wd fF] gq | sAep [eiados rel e 862; 00} @& perey[yuy) | UL Surpueys JoywAi-urey g-uee “way | V |) yystu Supa sanoy 9¢ ae -o1d jo ssouyaegq | 91'¢ | OL! a peie}[guy) | Ui Surpuvys 1ayVM-uley | g-uer “WV O&e | gq sinoy g¢ rel | é $99 0 GP potey[yuy) | Ul Surpueys soyvAt-urey | gue “Wy Css | V J sedid Aep sulinp ung | 2°90 | ¢@'o! a . petoz[yuy | Ul Surpueys 10yeM-ureyy fF uve “wasZ| g sedid T €2°0 0 CP pesey[guy | Ul Surpueys 10yeM-urley puee “Wa Z| V “09 929 i) 4 x = Slo SVNOWOdGAWV THO Sane & 3 e aqso ~ ‘ 5 8 dO SNOILIGNOO LHDIT NaD 3 z 5 UaALVM SNOILIGNOO HFLVM YaLy mM 616. AWIL SLOT Z Fs -Z£xXO | Ua 8 mee pasvaiour JOU $1 SLajom asay} fo JUajUoI vablixo ay) 1ayon ainyna PIO 40 1a}DN ULDs OJUL Jnd au SnoYy JDLaAasS Lof UNS ay] UL BurpuD{s Uaaq aaDYy YoIYN sDUOWOphD)Y,) fo sassDLU way yoy} Buinoyg 8 HTAVL 487 MALE PRODUCTION OXYGEN AND Aep Sump ung | Fe" T Aep Sutimp ung Aep Sutinp ung Aep Sutunp ung FLO 00 £0 ¢°0 00 00 0°0 00 00 0°0 0°0 0°0 0°0 GV GV seu -owopAure[yD Jo uortp -pB 19}JB 90U0 4B pot9} LT pers Ful) peed pers} gal) StBuoul -opAureyy) Jo uoryippe Io}J@ SoyNuIu G poaiez[l yy petoz[ gus) pore} guy) pet9y[ pato9 [TA pe19}, guy) pote} Jul) peso} peded tal peste; Gul) perez guy) sodid I9}@M-UlY sedid I9YVM-UTBY sodid I9}8M-ULRY sedid I9VBM-UTBY ul SurIpurys ul SuIpuRys ul Surpurys ul Surpurys sinoy 96 rel UL SUIPUBYS 19}eM-UTeY sinoy 96 aef UL SUIPUB}S 10}BM-UIRY sanoy 96 ref UL SUIpURys 10}eM-UTRY sinoy 7, ref UL SUIPURIS 10}VM-UIRYY sinoy ZZ rel UL SUIPUBYS 19}yVM-UIeY sunoy ZZ rel SUIPUB}S UI 10}BM-UTeY sinoy ZZ ref UL SUIPUBJS 10}BM-UIeY sedid 19} BM-UTBY sodid 10} BM-ULBY sedid 19} BM-UIBY sodid 10VBM-ULBY, ul Suprueys ul Surpueys ul Surpurys ul Surpueys 6 ue 6 “uBe (. 6 ‘uve * 6 ue“ 8 ue (. 8 uel (. 8 uee ce yp ‘ue te yi “uve ce. pb “uBr i 2 ‘uGe i p “uee (. Mn, “uBe (. 2 “ue ‘. 2} “uGr ce W'd G W'd G Oo SS SS ———__ ee oO a oy > 2 © WHITNEY DAVID D. 488 seuoul -opAue[yY jo uoryIppe sedid ‘wv Sulmnp ung | ¢¢°¢ | ¢G'O | ZF | 499JB Sanoy F poestog[Iq | Ul Surpuvys J9yVM-uUIvY II ‘uee “Wa Z| oO SBvuoUL -opAulv[yQ jo uoryIppe - sedid a ‘Wy Sulmnp ung | 91'G¢ | GO| Zp | 10938 sanoy F poyUBoeq | UL BuIpuvys 10};eM-UIeY IL ‘uve “wa Z| gq sedid 09:2 00) a pe1ey[yuy | UL SuIpuvys 10zeM-UIeY Il ‘uve “wayZ| V svuoul -opAule[yQ jo uolyIppe sedid ‘Wy Sulunp ung | 12°72 | GO| GP |aeyye sunoy F pedogiig | Ul Surpusjs J0yVM-UleY IL ‘uee “Wd ZF] q SBUOUL -opAure[yQ jo uolIppe sodid ‘W'v Sulmnp ung | 9¢°90 | GO| Zp | at9q;B sanoy F pojuBooG | Ul ZuIpusys 10,vM-UIeY Il ‘uve “Wa Z|] O or sodid 96% | OO} peroy[t | UL Surpueys 1oyvA-UIeY IT ‘uve “Wd 7 | @ sodid #20 | O'0 | “ZF persy[yuy | Ul Surpueys J0yeM-urey Il ‘uve “Wd Z| V SBUuoUL -opAUIB] YD JO WOIyIppe Aep Surmmp ung | 69°. | GO| Zp | 40938 SoyNuUTUL OT pe1eq[lT 10}8M 91IN}[ND PIO OL uve “Wage | Aep Sutnp ung | 98°90 | GO| ZF peyuBo0qd I0}BM 9IN}[NI PIO OL uve “Wwae| oO 6 cot] OO|} @& pe1ed la 19}BM GINn}[Nd PIO OL ‘use “Wage | 96'°0| 00| peyuReq 19}8M 9INY[NI P[O OL ‘use “Wae] V 99 ‘99 29 UddIT | pe | 2 y a tata ect oe e y s Bin tag's SNOILIGNOD UTLVM agLVM 6I6T ‘ANIL SLOT a8 -AxXO om 8 B panuyuojo—s AITAV.L OXYGEN AND MALE PRODUCTION 489 support the contention that appreciable quantities of free oxygen may be introduced into the new water with or within the cells of Chlamydomonas. In the experiments in darkness of table 7 only 0.05 or 0.10 ce. of Chlamydomonas was used which, of course, would have shown lesser results in regard to oxygen if they had been tested than the present experiments in which 0.5 cc. of Chlamydomonas was used. DISCUSSION In the recent work by Shull the summary of the results of the experiments under normal air conditions and under the 40 per cent and 60 per cent oxygen conditions show a higher per cent of male-producing females produced under the 40 per cent and 60 per cent oxygen conditions than under air conditions. If, however, one examines closely the individual experiments or lots in the tables 1, 2, and 5 of the results it is readily seen that the higher per cent under the oxygen conditions is produced in table 1 by the extraordinary results of two out of the six experi- ments. In table 2 three experiments out of fourteen experiments causes the higher per cent of male-producing females to be obtained. In table 5 two experiments out of twenty experi- ments of the oxygen-treated ones caused the total average per cent of male-producing females to be twice as large as it would have been without these two experiments. Some of the exceptionally favorable experiments under oxygen conditions were paralleled with similar results under air con- ditions in the controls. Thus indicating that the high per cent of male-producing females produced in parallel lots in air and in oxygen may have been due to other influences than an excess of oxygen. In some of the experiments under oxygen conditions no male-producing females at all were produced, while in many others very few were produced. If the oxygen was a real in- fluential factor in causing male-producing females to be pro- duced, many ought to have been produced in every experiment. Shull, however, does not claim that oxygen is the only factor that causes an increase in male-producing females, but that, 490 DAVID D. WHITNEY nevertheless, it is one of the potent factors in causing male- producing females to increase in number. The author takes the opposite view-point that oxygen is not influential in causing an increase of the male-producing females. In the experiments of Shull under air conditions, the rotifers produced 0 to 52 per cent of male-producing females in indi- vidual experiments and the average in the grand total pro- duction of 2334 females in tables 1, 2, and 5 to 8 was 10 + per cent of male-producing females. This per cent was of those females produced during the first twenty-four hours of the experiments. According to sample tests of such culture water as constituted these experiments, the quantity of free oxygen present during the twenty-four hour period was 5 + to 4+ ce. per liter. These results comprised of the production of about 10 per cent male-producing females in-culture water containing 5+ to4-+ ec. of free oxygen per liter should now be compared with the author’s experiments, lots B in table 7. In these lots with the diminished air supply the quantity of free oxygen at the end of the three-day period of the individual experiments was in some instances 1+ cc. per liter. The average quantity of free oxygen in‘all of the lots at the end was 3 + cc. per liter. This was a lesser quantity than was found in the experiments of Shull in the air. Shull obtained an average of 10+ per cent male-producing females in culture water containing 5 to 4 + ce. of oxygen per liter, while the author obtained an average of 74 + per cent of male-producing females in culture water containing 3 + ec. of oxygen per liter. In individual lots B of experiments 13, 17, and 19 in which the free oxygen was never more than 2 ec. per liter throughout the experiment and in lots B of experi- ments 13 and 17 in which the oxygen was diminished from 2 + ec. to 1+ ee. per liter during the experiment, the per cent of male-producing females ranged from 72 to 88. If these lots B of experiments 13, 17, and 19 are compared with the parallel lots A of the same experiments in which the quantity of free oxygen ranges from 3 + cc. to 8 + ce. per liter during the three- day period of the experiments, it is seen that in this increased quantity of free oxygen there is no increase in the percentage of OXYGEN AND MALE PRODUCTION 491 male-producing females. Furthermore, if the total averages are compared, it is seen that the high percentages of male-producing females are identical, although the quantity of free oxygen at the end of the three-day period averages in lots A at 5+ ce. per liter and in lots B at 3+ cc. per liter. Thus demonstrating that the production of male-producing females does not depend directly upon the presence of appreciable quantities of free oxygen in the culture water. SUMMARY 1. In the sunlight free oxygen in considerable quantities is given off by the green flagellates, Chlamydomonas. | 2. In darkness no free oxygen is given off by the Chlamy- domonas. ) 3. No appreciable quantity of free oxygen was found to be contained within the individual cells of Chlamydomonas when they were transferred from their original culture water into other water. 4, Culture water free from decomposing materials absorbs free oxygen from the surrounding air until its capacity of from 7 to 8 cc. per liter is attained. 5. In the sunlight fewer male rotifers and also fewer male- producing female rotifers are produced in culture water contain- ing Chlamydomonas which have given off much free oxygen than are produced in darkness in culture water containing less free oxygen. This is due to the fact that in the sunlight the Chlamydomonas become less available as food for the rotifers, while in darkness they remain more available for food through- out several days and nights. 6. Culture water containing the minimum quantity (in some cases less than the minimum quantity) of free oxygen, 1 cc. to 3 ec. per liter, in order to allow the normal activities of the rotifers, yields as many male-producing females as culture water containing from 2 to 8 cc. of oxygen per liter. 7. The general conclusion is that oxygen is a factor in causing a production of males except inasmuch as it is necessary for all life processes and activities of the rotifers. - 492 DAVID D. WHITNEY BIBLIOGRAPHY Suuuu, A. F., anp Laporr, Sonra 1916 Factors affecting male-production in Hydatina. Jour. Exp. Zool., vol. 21, no. 1, July 5, pp. 127-161. 1918 Relative effectiveness of food, oxygen, and other substances in causing or preventing male-production in Hydatina. Jour. Exp. Zool., vol. 26, no. 3, August 20, pp. 512-544. Wuitney, D. D. 1914 The influence of food in controlling sex in Hydatina senta. Jour. Exp. Zool., vol. 17, no. 4, November, pp. 545-558. 1916 The control of sex in five species of rotifers. Jour. Exp. Zool., vol. 20, no. 2, February, pp. 263-296. 1917 The relative influence of food and oxygen in controlling sex in rotifers. Jour. Exp. Zoél., vol. 24, no. 1, October, pp. 101-145. SUBJECT AND AUTHOR INDEX SEXUAL multiplication and regenera- tion in Sagartia luciae Verrill........... 161 AUMBERGER, J. Percy. A nutritional study of insects with special reference to micro-organisms and theirsubstrata.. 1 Bripces, Carvin B. Specific modifiers of eosin eye color in Drosophila melano- PASLET sls ac icusrerece asiapaietnetsteressieiers|s oycbalospeis 337 BripGEs, Catvin B. The genetics of purple eye colorin Drosophila i. ci... cccssies cess ELLS are subject to selection on the basis of their genetic potentialities. Evi- Genceitjhatirerm cso. ssscnu es eeeneece _. 385 Characteristics. I. The rat. On the physi- ological properties of the gonads as con- trollers of somatic and psychical.......... 137 Characteristics. II. Growth of gonadecto- mized male and female rats. On the physiological properties of the gonads as controllers of somatic and psychical...... 459 Color in Drosophila melanogaster. Specific modifiers of eosin eye........ ......+++0:- Color in Drosophila. The genetics of purple MO eerie ieraie she eiceiateciele niere SRO ontario ale 265 Bisset ORTH, C.H. Evidence that germ cells are subject to selection on the basis of their genetic potentialities...... 385 Davey, WHEELER P. Prolongation of life of Tribolium confusum apparently due to smallidoses Of X=C4y8. occ se cee esis cs cree e 447 Davis, Donatp Watton. Asexual multipli- cation and regeneration in Sagartia luciae Merrill een eran es aera ery eee 161 Day, Epwarp C. The physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to sensory TESPONSES: rca rice yer rinorel eaten Pct eiecice 307 Drosophila melanogaster. Specific modifiers OLeoRIn eye COlOLins.a.. cee ene oe 337 Drosophila. The genetics of purple eye color 281 Ee OMS OCU COT SURO Oe aa 265 GG production. The bearing of ratios on theories of the inheritance of winter.. 83 Eosin eye color in Drosophila. melanogaster. PecliGin odinersloly ase eee soe 337 Ephestia kiihniella Zeller. Genetic studies on the Mediterranean flour-moth......... 413 Eye color in Drosophila melanogaster. Spe- cific modifiersiof eosin’... 4...:-+.---.-- 337 Eye color in Drosophila. The genetics of UND LOS: i; «een ieee et eer: 265 ACTOR in causing male production in Hydatina senta. The ineffectiveness OUOXV LEN ABA. 5. hasan ee ee 469 Flour-moth, Ephestia kihniella Zeller. Genetic studies on the Mediterranean..... 413 Geen to sensory responses. The physiolony of the nervous system of the tunicate. I. The relation of the 1 (2) Ase R GUA OT OGG ot CODE GORE co ani. 307 Genetics of purple eye color in Drosophila PG ets fac eee eee ele tosdiolc ci5-ai0 som 26. Genetic potentialities. Evidence that germ cells are subject to selection on the basis OMUME ING MME racials tenia omeemeaeicen omnes 385 Genetic studies on the Mediterranean flour- moth, Ephestia kiihniella Zeller........ . 413 Germ cells are subject to selection on the basis of their genetic potentialities. Evi- CLCTICE LEE URE eee crore te ava) sats erstaketa heresecaveunts 385 Gonadectomized male and female rats. On the physiological properties of the gonads as controllers of somatic and psychical characteristics. II. Growth of............ 459 Gonads as controllers of somatic and psy- chical characteristics. I. The rat. On the physiological properties of the........ 137 Gonads as controllers of somatic and psy- chical characteristics. II. Growth of gonadectomized male and female rats. On the physiological properties of the.. 459 Goopate, H. D., anp MacMutien, GRACE. The bearing of ratios on theories of the inheritance of winter egg production...... 83 OMOZYGOUS yellow mice. The fate OL Geer alee PRE iG GD COR UOTE nei 125 Hydatina senta. The ineffectiveness of oxygen as a factor in causing male pro- GUC EL OME TUN a stiacae cistern ete iene oye evesareiens: sserevecs 469 NHERITANCE of winter egg produc- tion. The bearing of ratios on theories GEIPRORE ona est ante hrdteten es alter Sa varersbansia eve 83 Insects with special reference to micro-organ- isms and their substrata. A nutritional BCU VOL. Mee eeke ee sealer ee oie eeneeees 1 IRKHAM, Witii1am B. The fate of homozygous yellow mice................ 125 IFE of Tribolium confusum apparently due to small qoses of x-rays. Prolonga- PI OMPO Le ee ee seis rears biel theaors ielele clown foie exe Luctan Verriny. A sexual multiplication and regeneration in Sagartia.............. 161 ACMULLEN, Grace. Goodale, H. D., and. The bearing of ratios on theories of the inheritance of winter egg pro- GuchIGny ears ea een nel ates ete toeiamuae 83 Male production in Hydatina senta. ineffectiveness of oxygen as a factor in Gabel. Gan ORDO aa eo Alb demnoncde onan 469 Mice. The fate of homozygous yellow.... . . 125 Micro-organisms and their substrata. A nutritional study of insects with special TELETENCE LOwanany de esislsiieise me cinciees nigeria 1 Modifiers of eosin eye color in Drosophila melanogaster. Specific.. 337 Moorr, Cart R. On the ‘physiological properties of the gonads as controllers of somatic and psychical characteristics. Me Dhe maitice oasportion eet tue cote aac 137 Moorr, Cart R. On the physiological properties of the gonads as controllers of somatie and psychical characteristics. II. Growth of gonadectomized male and femaleratan sc ewes woseien peel eeees 459 Multiplication and regeneration in Sagartia luciae Verrlil, VAsexualliy. com ecaoaess 161 493 494 ERVE ganglion to sensory responses. The physiology of the nervous system of the tunicate. I. The relation of EG ih seers tte ere oizas tes colutatotere cveretaraeineys riers 307 Nervous system of the tunicate. relation of the nerve ganglion to sensory responses. The physiology of the........ 307 Nutritional study of insects with special reference to micro-organisms and their substrata, Ave 0555 Ge cere eireiteyersatsieivie.s sais XYGEN as a factor in causing male production in Hydatina senta. The INEMECTLVEMEBA Ole nie ae eielesitiels im ce eielase 469 HYSIOLOGICAL properties of the gonads as controllers of somatic and psychical characteristics. I. The rat. Ontthew:ccit guyana e ois Area hs or sieletemypiee nated 137 Physiological properties of the gonads as con- trollers of somatic and psychical char- acteristics. II. Growth of gonadecto- mized male and female rats. Onthe. . . 459 Physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to sensory responses. The..... . 807 Potentialities. Evidence that germ cells are subject to selection on the basis of their OTIC ULC EAE Sere obo hes a. ocoiatte Seapenese waneasteys eh okeaatlee 385 Production in Hydatina senta. The in- effectiveness of oxygen as a factor in CHUIBINE mallets 6s4.6 Seber elec tects aes 469 Production. The bearing of ratios on theories of the inheritance of winter egg........... Prolongation of life of Tribolium confusum apparently due to small doses of x-rays... 447 Psychical characteristics. I. The rat. On the physiological properties of the gonads as controllers of somatic and.............. 1 Psychical characteristics. II. Growth of gonadectomized male and female rats. On the physiological properties of the gonads as controllers of somaticand . . . 459 Rurple eye color in Drosophila. The genetics OLS eersteyerel syste rare) = erste aiafele, atolalaveiofetoiaterscoye ssi o 265 ATIOS on theories of the inheritance of winter egg production. The bearingof. 83 Rat. On the physiological properties of the gonads as controllers of somatie and psychical characteristics. I. The.. . 187 INDEX Rats. On the physiological properties of the gonads as controllers of somatic and psychical characteristics. II. Growth of gonadectomized male and female......... 459 Regeneration in Sagartia luciae Verrill. Asexual multiplication anl............. .. 161 Responses. The physiology of the nervous system of the tunicate. I. The selahien of the nerve ganglion to sensory.. BAe acity AGARTIA luciae Verrill. Asexual multi- plication and regeneration in............. 161 Selection on the basis of their genetic poten- tialities. Evidence that germ cells are Bubject tosses ey. eae eee ae eae eee 385 Sensory responses. The physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to....... .. 307 Somatic and psychical characteristics. I. The rat. On the physiological Rissa of the gonads as controllers of.. 137 Somatic and psychical characteristics. ITI. Growth of gonadectomized male and female rats. On the physiological prop- erties of the gonads as controllers of.... .. 459 Substrata. A nutritional study of insects with special reference to micro-organisms and: their? 05.202. cess Meee ee System of the tunicate. I. The relation of the nerve ganglion to sensory responses. The physiology of the nervous............. RIBOLIUM confusum apparently due to small doses of x-rays. Prolonga- tion of life of..... SoA DUR vated ate tone 447 Tunicate. I. The relation of the nerve ganglion to sensory responses. The physiology of the nervous system of the.. 307 ERRILL. Asexual multiplication and regeneration in Sagartia luciae.......... 161 HITING, P. W. Genetic studies on the Mediterranean flour-moth, Ephestia kiihniella Zeller.. ! 413 Wuitney, Davip D. The inefiectiv eness of oxygen as a factor in causing male pro- duction in Hydatina senta...............% 469 -RAYS. Prolongation of life of Tri- bolium confusum apparently due to small’ doses on seaaclf eee Stee 447 “~ ae F this Ni % + s 24) 2 t? * +s + + $ x4 ” i Sa ahets SSO . = Pete Nye ath 2 $i es as ~ - 2 1 fe Or de + be * * es eS + ee - > oe het te we $s = ‘Ae . Mad rae tetets 2 at coe ta SES eRe oe Se eaters ate ere 3 % roe a £ te SeF ote = 2 state? Pats - LESAAAS oP + 2's' e's *—) aia t sitet viekers! tet piyteletats eee 2 mo rhe > 2 +e ~) * =r eteielatyhetght nett h tether naeh Tee taiatste es” 3 3: HA . seretata tet te sett ? tee * sh : S ees : SL SESE SE 4 t 24,35 rat . 2 ee: SOa preletate a : BIS teks +: ate RS, es BEES DASE E NN at ets Tpfopatetae gates eteisteth ee sesh BE Ee ‘ : eee raietetetetatets ets ° Niatets Tetatetetgt : LOR ptatgretress te tete ts oi, : SEAS X ie pees : i tyes ateteia’ ee ; : : : tet of C = ere es a a - 2 Ree : ot A a : Soto Os = = Sietetstetete Bs & ut Ps ey os ta eR oe a et Se oes Sa ee yt om ae