jehebat nw ‘ethetl Oe eo. - ‘ r 4 wre > Gh es ee ae te oe + ee ee te? oe? abate De® tet eretete terete. 4a tetet~ . a eer cours Het WAH 9h 0 be oh alah nk atinttot ERAS amr Guuesesstntetanetetenubah Haale thelarets ee ee eee nee nerd atidnrahssetepreDepaselnliaiaias gAgmanchittamnet weiaNeadngtecnedeay-nob risen! )-vseeattmertaarannivervedvernanawtpanayawhebabungecsocesoetrmemen ee mee iree chek meke ee eee ee ae ean ta asataaaiy ey try tee Nee aabdbawhsratute¥ed-tersratete tasenesa hos , neta aparnede frre STO TTT ee Pabehieres ste pose aa stated = ho 8 medaae rrachosabyaehenpteapecdbeaebenanearsuheneschd eee OO I a IEP ee te rT th ya A Ob ee ee Oe Ry » LCC APA ees PPO Gy + pert ata eA. ap teprenepesraertaasoapere pereney - rane sBs fies ’ ode i taetiall o nome - ¥ Rt Sra WW stomp Ran matin Digitized by the Internet Archive in 2009 with funding from University of Toronto htto://www.archive.org/details/journalofexperim28broo ‘Bi THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY Wint1amM E. Castie Jacques LoEB Harvard University The Rockefeller Institute EDWIN G@ CONKLIN EDMUND B. WILSON Princeton University Columbia University Tyomas H. MorGan CuHarRLes B. DavENPOoRT E Neer’ Columbia University Carnegie Institution GEORGE H. PARKER HERBERT S. JENNINGS Harvard University Johns Hopkins University RAYMOND PEARL FRANK R. LILuIE Maine Agricultural University of Chicago Experiment Station and Ross G. HARRISON, Yale University Managing Editor 0 ee l i | q” VOLUME 28 3 | 1919 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. we f te: aan oa ae) Pa ; wy a f , i =. N an ‘a ASH J; i dS wy (\ WZ, ch i | oe | shayoanc a aan cay ma 0 RT share ; : ) y eerie Bs SB aaa Pk Maal H idee a &~. 5 4 THT ROA #1 nt "ta >? iy oS tie ~~) sqeigd admit ie : 1 An " . aT . cis Sai Yaseen = » > Vigisaaet HT sont) es are : a %, ‘ iting!) Pharr Hl . core Pee 2! ," usin «ciel eel dmoirat . ase a weet s pliowviatt nldv pwotigall c+ ead wil gotgece | : ry 7m 1 ai 7 j ne ah Poa | a ye ’ 7) \ i i) - 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. Goopate anp Grace MacMu.ien. The bearing of ratios on icone of the inheritance of winter egg production .........-.-. 0.2002 ee seer ees 83 No. 2. MAY Wiuuam B. Kirxuam. The fate of homozygous yellow mice. Two figures. 125 Cart R. Moore. On the physiological properties of the gonads as con- trollers of somatic and psychical characteristics. I. The rat. Five TECTUTT ES ae GOR ce eee I HES SORT CE retin OIC ce cc 137 DonaLp Watton Davis. A sexual multiplication and ee Yl in Sagartia luciae Verrill. Ten plates (forty-two figures). . Beco one loa Catvin B. Brinass. The genetics of purple eye color in Toeceentilan. .. 265 Epwarp C. Day. The physiology of the nervous system of the nie I. The relation of the nerve ganglion to sensory responses. Five figures. 307 No: 3. JULY Catyvin B. Brivars. Specific modifiers of eosin eye color in eee melanogaster. Two diagrams. . a Bo eey C.H. DanrortH. Evidence aa germ- saat» are Peenieen to eelecuen! on one basis of their genetic potentialities. . sBe LE ss sl OOO P.W. Wuitine. Genetic studies on the Nicditeranoan iow atop i Ephesti a kithniella Zeller. One figure and two plates. . aa ; .. 413 WHEELER P. Davey. Prolongation of life of Tribolium ‘eonfusum appar- ently due to small doses of x-rays. Four figures. ‘a Ree et Cart R. Moore. On the physiological properties of the gonads 2 as con- trollers of somatic and psychical characteristics. II. Growth of gonadectomized male and female rats. One figure.................+.00. 459 Davip D. Wuitney. The ineffectiveness of oxygen as a factor in causing Malewroduction initHydatina senta. ...... 0100-0 vebeees+ «+++ oe aemeies S00 has t : a ; ert ipeinys tie BF rad Ya i fipridciddia eve a a i } ee). ear Tae aa ed easter iit fi Agar at ar 5 poy PD ‘ean; hip Y Hen tye a" te ty LAT tas se RMS oo 63.002 ane a ot long 3 of ut ¥ 7 We: te a f Ny f : - m Ve 7. i NE * ‘ : Pon ee ry) , a ie te Re RL Tay ‘ i) a , ; , é f ma ait owt: lyst ttt walls) f Crna) eatin af te otal ar wed ‘aires, "eal i pi Peaecen tes Lav glaze step Oat t : Cody ay T. veniielinteasiily, lastdiggei bow dita 4g) iv S8bi: , Mee Peg Cbkcrdytes ost MO | URES Gem eee eo) et : ck sei rcmccnn e's fue menhasllefialion: rae ‘ae mime ee iu ie airs siMgil BRR) He aM: Aid vagn ‘doe shai i ni. Ci gig putea aad Nea Jitaeeie tf he Wee PREIS ee i ; ih a at : i. we ek: urging a a Eo Ney Cpe PDR i pte rei oye yes ‘Nt ey i vale ae im) mt } : t | 3 ig a i -_ ghidyocnutt qs shay oh ea hy ; . ee ree Seman a PK Hs! " ciauaitige ae i aids cm Wot iealee gt. ty riclern aa atl yon sikh qt hi Ne ns ee porta Hakone tis va eee i sh Yan Pies italy, 1 stn inh sean yas ifyeilySpaat Hing de iter) BL Paes ome). 0 arial ow! bre FTUA0 Fey ent t tet Reg mrowsrtnary inipitodtin’ , Le nie y On ai, ee phe mica REY Te ee ees fb Day. bons ry Ne ‘ nt Tah me # ae spaletusl,. {acai ay a: ¥ ie AC Le ee | erea eh Mts: “han eis ju ae THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 1 APRIL, 1919 Resumido por el autor, J. Perey Baumberger. Estudios sobre la nutricidn de los insectos, con especial mencién de los microorganismos y sus substratos. El autor ha demostrado mediante experimentos que la ad- quisici6én de los alimentos por las Drosophilas que viven en frutas ‘fermentadas depende de los poderes sintético y absorbente de las células de levadura. De un modo semejante, sus estudios sobre las relaciones de Musca domestica con el estiéreol, de Desmo- metopa con la carne en putrefaccién y de Sciara y Tyrogliphus con la madera podrida, demuestran claramente que estos artro- podos también se alimentan de microorganismos. Asi mismo, ha intentado trazar el origen y desarrollo de estas costumbres, la extensi6n probable de su ocurrencia y considerar las asocia- ciones conocidas de los animales con los hongos en general. Los experimentos y consideraciones tienden todos a establecer el principio de que los insectos que viven en substratos en fermenta- cién y putrefaccién que contienen escasa cantidad de protein as, se alimentan generalmente de los microorganismos que enellos existen y de este modo obtienen beneficios gracias a la propiedad que poseen los hongos de extraer, absorber y sintetizar muchos compuestos nitrogenados no proteinicos. Translation by José F. Nonidez Columbia University ° 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 [isnfrvarBhivenvoiaiaes:! Meta) weed ae coins Gente ieeo See: mig RG Ooi Oke ero te Para ‘sobs 2 Bre DOrinenitae eee Meters oases hats a as nae cele = « alo = siei’s Seeing 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- eal 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.......... it 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 OUaGheHOods ote rosophila ae sas: oes fees) Sule oes as oe ese 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; ¢) Association of wood-eating ARAE CHA WIRD P LUNE. 5d...) wove dees Meta eek. 5 6 ROORERR eieiale 7 Extent of mycetophagy among insects................. Sia « s' eeeeeectae 58 Microérganisms as liquefiers of the substratum..............-.....-.---+-5- 64 Odors attractive to insects... . 0.2... 1... eke cece cen e tere eee teen e eens 67 Microorganisms as food of other animals......... 0.00.0... . 8p cee eee nese 69 Microérganisms as internal symbionts of insects..................0- 25s ees 72 (Clei@ WENoinian =. ott sk se SAS ene ROR ROREIRE aye eo iio Aas aeeae 7 RMI RERD ELS MMe Ri oro Beis kbp eee boxs:'s ove Slo . =| Vaan Hh, A or 4as. P= fupation Al Age 17 Days. D= Death af ferva rary 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 ° 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: (AT = 20 rr 4.0 grams 1 SS) 8 2 7 arty aap” an ge 0.165 grams Grape-stigaregs..... 2)... 16.5 grams Mig S Omar He: eee 0.165 grams Cane-sugare. 77s... : 16.5 grams Bi ON. tte Seg re se, 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 Sterile bananatacany ere eer sre en. 16-0 cee 1.8 mm. 26 to 44 days Stenile yeast nuimentamedmmes.-. 0485... .s. 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 1} 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 ce. 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 ita ralllireters. 2 Larva! /ength 5 A Vm A a Sterile Banana- = y yerage 26) G - P= Fupostion. D=Death of larva. Age t7 Days. 1] 2) 3) Py 6 7 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 eat aes ‘asvak quod sod 77 shivp ur avai) fo aby b ATAVL | d 2h eee) | T. d S| AA SSE eames G sre: qsvok yuoo 10d Z] d |e |e Gl € ISL | py 7°° 3° 4seod quedo dod ZT d Si[.- 2 | 9 1G pe cet qsvok yuoo 10d z{ | d 8 9 Z cts ¢*-4sved 4uo0 tod ZT d | 8 Le QUNGSS 51 Can ae nee 4svok yuoo aod 9 Gaon |peron |S: 9° (eg Cees ysvok yuo0 aod g cli ies 7 |¢°S | [9+ -4svoX gue od 9 1 P| 9) | (So ale ese Aeo O Dm Om a7 d GUNNS 9 |9°S || lc ees oda rods d Saieacel Ss g@:-7s7>*-asBek 4ue0 Jed ¢ al a OMG eG Go nseek ued ted ¢ d she) al) Gulla "+> -qasvak yuoo tod ¢ d Se ee Ge Gee eG cGal GaGa a Gulla LD eae, Fe < ysvok yuod rod Z : i7 & 6 Set Set Sa Shea Lee Z Po ROSE ois os ysvok yuao sod T d 9 G GZ Ie°z Tr laa staysodurey d Daldeg |S 9 1S°¢ Gg GP if alll Xe ese ee ae [OS BRE USC pojuoutezun snoenbe 4077 S1OJOUIT] [UL UL YYSUoryT |u| oz | or | or | | ox a | | o [8 a as ae A NUTRITIONAL STUDY MEDIA Hot aqueous sol. unfermented | banana Bananamash: ses. --: css. 1 Bamamaynaa sys Wives. -e)- =)slsfe= Agaricus campestris......... 1 ipencent yeast... .2-ses05: Me 2MWer Cent yeasb »-. a. se ace 1 SD Per Cenitayeas uri eeeeee 1 3 percent yeast...... Pacis 2 SD PELGenuyeaSte- cree = - 3 ALDER OMIH WEED ood oao dooee 1 AMeRCeMiViCasl er ace sea: = 2 APeMeCeMinVedstea-. eis 6 = 3 Gipencentyeasti...-.4-6- 4. - 1 Gpencentyeasta.. 52-54... 2 Gpercenbweast .-.. = - 2: 4 IZ PELCeME VeASt si. -52 255. .: 2 WARDET COMVERSE i. acc cis + eee © 3 percent yeast:........-... a i2per cent yeast ».....2-.... 5 ZEereent yeast <. 2. <0... 1 24 per cent yeast ............ , 3 per cent yeast ......average 4 per cent yeast ...... average 6 per cent yeast ...... average 12 per cent yeast ...... average 24 per cent yeast ......average NO. PUPAE (0/0) 76 79 158 02 OF INSECTS 7 TABLE 5 4 2 =) & : 4 me a Ve ce < ae ae & | > a ae Bre lenres S a I a nee od 20 3 (5. 28.75) 1.29 | 4.5 9 15.5 11.0 0 13.0 12.43 | 1.89 | 15.2 | 14 |3.85] 0.6221 | 16.1 20. heAS fear 19% 4 7.14 | 0.27 | 3.78] 29 |4.75] 0.2318°| 4.8 6.00 | 0.14 | 2.03 6.32 | 1.93 | 30.5 | 40 [4.33] 0.0409 | 0.9 13.05!| 1.55 | 11.1 6.47 | 1.70 | 26.0°| 33 |6.63]11.66 25.0 7.20 | 1.208] 14.00] 30 [5.58] 2.89 43.59 5.00 | 0.0 0.0 | 20 |4.0 | 0.0 0.0 6.0 | 0.308] 5.1 | 20 |4.4 | 0.154 3.5 6.46 | 0.157] 2:4 | 60 [4.37] 0.2454 | 5.6 58) 1 O85 8.8 | 40 |5.45| 1.07 19.6 7.0 | 0.713] 10.2 | 51 [4.46] 0.6361 | 14.2 G3) ROes25i0 8.3 | 7 70.41 9.8 7.8 | 1.095] 14.0 | 25 13.68] 1.2 32.6 5.21 | 1.61 | 30.9 | 60 [5.44] 0.527 9.6 . 4.83 | 0.188] 3.8 | 12 |3.93] 2.02 56.0 6.55 69 |4.50 6.85 66 [5.85 2 6.11 100 |4.30 6.87 133 |4.07 5.14 72 15.18 1 Fi 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? 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. Bacteria and fungi other than 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 od Lorre fangth Je willimeters. 2 =Pupatina. 2. = Death of Jarra. ie in Lays 3 OSE EA TE HEP a ee Fig. 4 Larval growth on various media. 1, 24 per cent yeast; 2, maximum 3 to 12 per cent yeast; 3, minimum 3 to 12 per cent yeast; 4, vinegar plant; 5, mushroonf; 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 desiceation 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 = z ¢ NO. NO. <5 MEDIA (20 a Nee IN eon ee 2 : o | S/Gia 2 Fe ys) 0 lO ee 11 days 1 0 — PARAM ASIN 2 oe his caciewis (ates One 7 days 16 16 + AH ADAMUIRS Nt. ss, cayesiyetssisiens nein ee 22 days 1 1 — Banana mash (slightly shrunken) 4..| 13 days 2 2 -- Cold aqueous extract of fermented/| 28 days (Average) 0 0 - banana 12 days (Average) 73 70 _ Cold aqueous extract of unfermented/{| 25 days (Average) 0 0 _ banana 13 days (Average) ’ 90 85 - 20 days 8 8 - eS extract of unfermented 20 dave 9 9 = 9 days (Average) 68 66 — These are potentially fertile, for when fed with yeast, the females deposit fertile eggs from which normal larvae emerge. This will be discussed below at greater length. ‘On cold aqueous extracts of unfermented banana, no larvae pupate. These results show that concentrated banana permits complete and more rapid growth of larvae. Therefore, the fruit is not en- tirely lacking in any amino-acid necessary for growth nor is any vitamine absent or destroyed by sterilization. It appears, more- over, that concentrated banana forms a complete food for Dro- sophila larvae, i.e., the protein deficiency of fruit is quantitative rather than qualitative. 20 J. PERCY BAUMBERGER As Agaricus campestris has a proteim content of 3.5 per cent as compared with the 1 per cent protein content of banana, this fungus was used as a food for larvae. Drosophila females oviposited read- ily on a medium of powdered Agaricus campestris, water, and agar. The sterile larvae grew rapidly to a size of 6 mm. (the maximum size on yeast is 8 mm.) and pupated after an average of 12.43 days. The curves of growth are shown in figure 6 and a typical record in tables 4.and 5. Theadults that emerged from these pupae were sexually fertile, but were quite small. A generation of larvae reared from some of these adults grew more slowly than normal on 4 per Larve/ sength 8! ip paliimelers. Pp = ayu 2 \) @® \z 2. P= Pupation. D= Death of larva Age 17 Days. i 3 ea, 9| Ho mt tz ye 9 5 Ne 7 Ne IS 2o 2 2 3 2144 2 6 27 2 Fig. 5 Larval growth on banana. 1, mashed whole banana infected with living yeast; 2, 3, 4, sterile mashed whole banana; 5, 6, hot aqueous extract of banana; 7, cold aqueous extract of banana. cent yeast, requiring thirteen days instead of seven days to reach maturity (fig. 6, curve 2, and table 5). Sexually fertile adults can be reared from larvae fed on mushrooms, but such adults are undersized. Agaricus campestris meets more nearly the food requirement of larvae than banana; this may be due to the higher percentage protein content of the mushroom or to a relatively higher content of certain necessary amino-acids. f. Is yeast a more adequate food than fruit because of its high protein content? The slow growth and small size of larvae and their failure to pupate when reared on sterile fruit are typical A NUTRITIONAL STUDY OF INSECTS 21 symptoms of protein deficiency. This deficiency is quantitative rather than qualitative, because more normal growth of the lar- vae is permitted when the fruit is concentrated. As these symptoms of malnutrition are not shown in larvae reared on yeast, we would expect to find a high protein content of adequate components in this food. This assumption is correct, for Atwater and Bryant (’02) have found by analysis that the percentage protein content of yeast is 11 percent. This is higher than the percentage occurring in any fruit or in Agaricus campes- tris, and Meisenheimer (’05) has shown that most monoamino- acids occur in yeast protein. 8) Lorva/ /ergth la tiillimelers. | | 2= Pupation. Age in Days. 1 7 a] 3] rr || 2] ens Fig. 6 Larval growth on mushroom. 3, 4, mushroom; 1, normal growth on 4 per cent yeast; 2, growth of larvae of mushroom-fed adults on 4 per cent yeast. Is the rapid growth of larvae on yeast due to this high protein content or to other substances present in the yeast cell, such as glycogen, fat, gums, hemicelluloses, etc.? By extracting the yeast nucleoprotein and making media with known sugars and salts, this problem was solved. The method employed to extract the yeast nucleoprotein and to make the media is described in a foot-note below."! 11 The pure yeast nucleoprotein was obtained in the usual manner (Hawk, ’16), as follows: the cells in 4 pounds of baker’s yeast (Fleishmann’s bottom bread 22, J. PERCY BAUMBERGER The growth of larvae on yeast nucleoprotein media is shown in figure 7 and table 7.. Table 7 shows that larvae do not grow on nucleoprotein in the absence of Pasteur’s solution, but, if these salts and sugars are added, do grow rapidly and form large numbers of normal pupae from which normal highly fertile adults emerge. ’ yeast) were killed by ether and broken open by grinding in a mortar with a quan- tity of pure white diatomaceous earth, adding enough water to keep the mass in a sticky, smooth condition. The grinding was continued till examination with a 1.6-mm. objective showed that many of the yeast cells had been cut into irregular rectangles. The yeast was then poured with the addition of 0.4 per cent NaOH . solution into a large bottle and 8000 cc. of the alkali were added. About 40 cc. of chloroform were then mixed with the solution to prevent the development of bac- teria. The contents of the bottle were thoroughly agitated several times a day. After forty-eight hours the supernatant fluid was poured off, leaving a great part of the yeast and diatomaceous earth in the bottle. This fluid was centrifuged in a large-sized electric centrifuge with a capacity of four 250 ce. bottles. Each liter was run for twenty minutes (fifteen minutes at maximum speed) and the super- natant fluid carefully poured off. This fluid was examined for yeast cells witha 1.6-mm objective and showed an entire absence of them. After all the yeast cells had been removed in this manner, the liquid had a clear opalescent color and proved to be rich in nucleoprotein. This was precipitated in great white floccules by adding 10 per cent HCl in drops. The precipitate dissolved in the alkaline or neutral solution, but remained in the slightly acid solution in which the largest amount was formed. The precipitate was separated pure from the solution by centrifuging and washing with acid alcohol and neutral alcohol in which it was insoluble. The white precipitate was then dried over HeSOq, forming a white powder. The remaining fluid was neutralized with N/10 NaOH and dialyzed for five days in running water, keeping the surface covered with toluol. No precipi- tate was formed. The neutral, salt-free solution was then heated to boiling and acidified with a drop of HCl or acetic aid and also acidified and then boiled. No marked precipitate was formed. Half saturation, complete saturation with (NH2)SO« when hot or cold and saturation with NaCl and with picric acid failed to bring down any precipitate. A heavy precipitate which appeared to be a pep- tone decomposition product of the nucleoprotein was produced upon the addition of phosphomolybdiec acid. The nucleoprotein was ground in a mortar’and then made into media as fol- lows: 1) Nucleoprotein moistened with tap-water was placed in test-tubes and sterilized in the autoclave; 2) nucleoprotein moistened with Pasteur’s nutrient solution (grape-sugar, Cane-sugar, ammonium tartrate, MgSO., K,HPO,, HO) in test-tubes and sterilized; 3) nucleoprotein and 1.5 per cent agar-agar tap- water solution autoclaved and mixed aseptically in sterile test-tubes, and, 4) nucleoprotein. Pasteur’s nutrient solution and 1.5 per cent agar solution autoclaved and mixed aseptically in sterile test-tubes. If mixture is made before autoclaving the protein adsorbs (?) the agar and on cooling jellation does not take place. ; A NUTRITIONAL STUDY OF INSECTS 23 Adults placed on media from which sugars were absent died after one to four days, whereas those placed on media containing Pas- ° teur’s solution lived for a much longer time. The larvae on the nucleoprotein alone live for several days, but do not increase in size and are not very active. It may be that a sweet taste is necessary to stimulate them to take food or it may be that carbo- hydrates are necessary to furnish energy “fuel.” The larvae on nucleoprotein and carbohydrates grow slowly at first, but quite rapidly after reaching a length of 3 to 4 mm. This may be due to the rather large size of the nucleoprotein crystals or to the depth 7\ Lorral senglb tn toillimeters. P= Pupation. D= Death of larva Age sa Lays. yo it Wea Ws 9 Fig. 7 Larval growth on yeast nucleoprotein. 1, 2, 3, 4, 5, yeast nucleopro- tein, sugars, and salts; 6, 7, yeast nucleoprotein and tap-water. to which they sink in the agar. In figure 7 are shown the curves of growth on these media. Curves 1 to 5 show the rapid growth on yeast nucleoprotein, sugars, and salts and curves 6 and 7 show the slow growth or diminution on nucleoprotein alone. One curve of growth on yeast nucleoprotein and sugar, etc., given (fig. 4, curve 6) in the same figure with curves for yeast media, shows that larvae grow more rapidly on yeast nucleoprotein and sugar than on 2 per cent but less rapidly than on 3 per cent yeast. It must be remembered that mechanical questions of ingestion and the question of taste or olfactory preference may largely affect the amount of material eaten and therefore the rate of growth. 24 J. PERCY BAUMBERGER TABLE 7 & a Zz 5 C z az =] = <0 A 2 m = MEDIA ao s z a : é +HAa < Dp a < Seay eee les eo? tae Geen ae se tenlsoke Nucleoprotein + tap water No. 1............. 2 6 3 3 =F Nucleoproteim -+- tap water No. 2.............] 2 5 4 + _ Nucleoprotein + tap water No. 3............. L2Gq 70 0 - Nucleoprotein + tap water No. 4............. 10d} O 0) _ Nucleoprotein + Pasteur’s sol. No.1.: ...... 10 ik 0 — Nucleoprotein + Pasteur’s sol. No. 2......... 0* 0 — Nucleoprotein + Pasteur’s sol. + agar No. 1.| 5+ Nucleoprotein + Pasteur’s sol. + agar No. 2.]| 6+ MeN ew BNE AN ox -_—_ : Nucleoprotein + Pasteur’s sol. + agar No. 3 2 - Nucleoprotein + Pasteur’ssol. + agar No.4.| 7 30 30 = Nucleoprotein + Pasteur’s sol. + agar No. 5 3 0 _ Nucleoprotein + tap water + agar No. 1..... 5d] 0 0 _ Nucleoprotein + tap water + agar No. 2..... 21d| 0 0 - Nucleoprotein + tap water + agar No. 3.....] 3 16d} 0 0 ~ Nucleoprotein + tap water + agar No. 4 Kol |) O 0 - Nucleoprotein + tap water + agar No. 5 1 Hel ||, © 0 — Nucleoprotein + tap water + agar No. 6.....| 2.5 5d] 0 0 = Nucleoprotein + tap water + agar No. 7 4 0x5) 70 0 |}*- x No eggs laid. d indicates death. * Media dried. Hence the objection that there is no exact mathematical corre- lation between the rates of growth and the protein concentration is not irrefutable evidence against the view that primarily such a relationship exists. For example, some farm animals will lose weight on a soy-bean diet of high protein value, but of a taste they do not like. Since Drosophila grows normally, pupates in large numbers, and develops into fertile adults of good size, it appears that a medium of yeast nucleoprotein, sugars, and inorganic salts is a complete food for this insect. It has already been proved by experiment (p. 13) that larvae do not grow on sugars and inor- ganic salts alone, so that the nucleoprotein is the substance which, if added, makes the medium equivalent to yeast cells as food for A NUTRITIONAL STUDY OF INSECTS PAD) the insect. As the sugars and inorganic salts are abundantly present in fruit?? and the addition of nucleoprotein of yeast is sufficient to make the synthetic medium a complete diet for Dro- sophila, it can be said that yeast is a more adequate food than fruit because of its high protein content. g. Conclusions. 1. Insects can be conveniently reared in a solid agar medium. 2. Larvae prevent the development of molds on the medium, but are always associated with living yeasts. 3. For genetical work fermented banana agar or Pasteur’s culture fluid agar is most convenient. 4, Living yeasts are not present in the egg or pupa. 5. The exterior of pupae can be sterilized by washing in 85 percent alcohol for twenty minutes. Yeast cells are more readily killed by this treatment than molds. 6. Banana agar is a good culture medium for fungi. 7. Sterile larvae grow more slowly than non-sterile larvae on sterile fruit; the rate of growth can be increased by infecting the medium with living yeasts. 8. The alcoholic treatment in sterilizing pupae is not the cause of the slow growth of larvae on sterile food. 9. The simplest nutrient medium for yeast, if infected with living yeast, is equivalent to fermenting fruit in the ecology of larvae. 10. Sterile fruit has greater food value for larvae than ‘‘sterile nutrient medium for yeast.”’ 12 Atwater and Bryant (’06) give the following analysis of the edible portion of banana: Total car- Water Protein Fat bohydrate Ash Fuel value per cent per cent per cent per cent per cent per lb. IMitoniaahbtans aot @eanhaene 66.3 1.0 0.0 16.3 0.5 330 IMM aninbiONs 5 hoo oc bee 81.6 WAG: 1.4 29.8 1 640 INSHGreeind DSoen ees Gn ee 1O.8 i) 83 0.6 22.0 0.8 460 Prescott (’17) gives this analysis of banana ash: per cent j per cent per cent SriCae ee fre: cee has 2.19 Phosphoric acid... 7.68 Potash............ 48.55 Waineecutere: oso. ab S2Mapnesia......... 6.45 Sulphur trioxide.. 3.26 monvoxide4.nt.3 4. -: OMS WR SOG atesno« «clo .ay Lon hl ones s eee 7.23 26 J. PERCY BAUMBERGER 11. Fruit is mainly the nutrient substratum for yeast cells, but has some food value for larvae. 12. By-products of fermentation are not necessary for larvae. 13. Dead yeast is an adequate food for larvae when in a con- centration of 2 per cent or more. 14. Yeast is a more complete food for larvae than other fungi. 15. Concentration of banana by hot-water extraction or drying makes it an adequate food for larvae. 16. The protein deficiency of fruit is quantitative rather than qualitative. 17. Agaricus campestris meets more nearly the food require- ments of larvae than banana. . 18. Yeast nucleoprotein, sugars, and salts are an adequate food for larvae. 19. Yeast is a more adequate food than fruit because of its higher protein content. C. Discussion. a. Effect of food on larval, pupal and adult life. On page 15 and in table 5 and figure 4 it has already been shown that the concentration of yeast affects the length of the larval period. This effect can be seen more clearly if we plot the larval period on the axis of ordinates (vertical), of a graph, and the number of grams of yeast per 100 cc. of water, in the media, on the axis of abscissae (horizontal). A curve drawn through the points established represents the effect of the concen- tration of yeast upon larval life. This curve is drawn in figure 8. It changes its direction very suddenly at a point between 2 and 3 per cent of yeast, going up from a larval life of 6.55 days on 3 per cent to a period of 11.40 days on 2 percent yeast. If we continue the curve in the same direction we approximate a period of twenty days representing the larval life (which ends in death) on 1 per cent yeast medium. This great change in the direction of the curve indicates that there is a definite concen- tration (2 per cent) of yeast mecessary for the completion of the larval period. Between yeast concentrations 2 per cent and 3 per cent there is a difference in the larval period of 4.95 days, whereas between concentrations 3 per cent and 12 per cent there A NUTRITIONAL STUDY OF INSECTS : 27 is only a variation of 0.76 day and between 3 per cent and 24 per cent—a difference of 1.11 days. It appears that the normal condition for larval growth is in a medium of yeast concentration between 6 to 24 grams per 100 ce. This is shown in the size of pupae in table 8. : jl : Agorreus Caommpesirrs. Fig.8 The effect of the concentration of yeast on the length of the larval and pupal periods. A con¢éentration of 1 per cent of yeast appears to be sufficient to furnish the energy, repair, and some growth requirements of the larvae for a considerable period without furnishing quite enough food to allow the necessary growth changes or storage preliminary to pupation. The larval period must be considered as a nutrition unit. In the case of larvae on 2 per cent yeast, the insect obtains in 11.4 days sufficient food material to give 28 J. PERCY BAUMBERGER TABLE 8 Effect of larval food on size of pupae MEDIA LENGTH OF PUPAE ADULTS mm. Bananamash........ eo bene 2.0-2.5 HNP COMIWV CNS o555000n00000000 800008 2.5-3.0 Hot aqueous sol. unfermented banana. 3.5-4.0 Ac anicuscamMpesunisn, yacec ait ee 3.5-4.5 Undersized SOCIMCE NAV CHS. coggueeothsooccpdunces 4.0-5.5 ZBVOLSIO CEM NASM oaoien cio We ao nome en oce 9.0-5.5 HOEMCRMY WANs daoc0a0cponoeeseone ed 5.5-6.0 l2pericentsyearste- scenery cee 5.0-6.0 Normal size PAN VOIP CEN G WEDS poo ce clo coocesosganeet 5.0-6.0 the energy, wear and tear, growth and storage requirements. On 3 per cent yeast the larval period approached normal at the expense of the reserve stuffs in the pupa, for the latter is under- sized. This is also true of 4 per cent yeast. On 24 per cent the larvae usually reach a size of 6.5 mm. in length on the first day and are therefore three days ahead of all larvae on media of 3 per cent yeast; still pupation occurs only 1.41 days earlier. Ap- parently there is a certain periodicity in the larval life, since there is a tendency for the larva to pupate after a certain length of time whether it reaches the maximum size before this period or is still undersized. The probable explanation of this phenom- enon is that certain changes go on in the larva, since a meta- morphosis of the nervous system and digestive glands is known to take place during this period, at a definite rate if the mintmum necessary food substances are available." The pupal periods (table 5) of Drosophila, fed during larval life on different concentrations of yeast, are also plotted in figure 6. The figures show that there is no consistent variation be- tween the pupal period of larvae which lived for a long or a short 18 Mendel and Judson (’16) studied the proportional weights of skeletons of retarded rats and found that the skeleton grows at a normal rate in retarded in- dividuals. On normal food the growth of retarded. individuals is accelerated, but that of the skeleton is retarded till equilibrium between tissue and skeleton weight is established. A NUTRITIONAL STUDY OF INSECTS 29 period before transforming. Therefore, the pupal period is not correlated with the length of the larval life, i.e., it has, also, a fixed periodicity. The growth of insect larvae may be retarded on sterile fruit and then greatly accelerated by adding living yeast cells to the food. Three curves showing this effect are drawn in figure 9 and curve A: 24, figure 2 has already been referred to (p. 13). The insect can be maintained for a long period of time on a minimum of protein (banana) at the same size or slowly increasing pas, lerglb | tn rollimeters. z P= Pupation Age tn Days et Ts) bah 2s eas leap ye te a eh cle as ek ce ae ar ce ce So sh sk Fig.9 The growth of retarded larvae. 1, larvae from hot aqueous extract of banana placed on 4 per cent yeast on eighth day; 2, larvae from hot aqueous ex- tract of banana placed on 9 per cent yeast on fourteenth day; 3, larvae from cold aqueous extract of banana, living yeast introduced on twenty-sixth day. size, after which it may be made to develop to normal size by placing on an adequate diet (6 per cent yeast). Except in one case (fig. 9, curve 3) the acceleration in the growth of retarded individuals was not above the normal rate of larvae on an ade- quate diet. This may be due to the fact that the length ofthe larva is a poor criterion of the extent of its development and that units of one day are not accurate for an animal with a normal larval period of about six days. In this connection it is of interest that Osborne and Mendel (14) have shown that rats can ,be kept for a long period at the 30 J. PERCY BAUMBERGER TABLE 9 Lengthening of life cycle by retarding larvae AGAR MEDIA EGG PERIOD a ae tech LIFE-CYCLE IN DAYS days days days 24 pericent:yeast.... 1252 sr acre 1-3 5.14 5.18 11.32-13.32 l2percentyeasts-)..4556 erates 1-3 6.87 4.07 11.94-13.94 Gipericent yeast... ae 1-3 6.11 4.30 11.41-13.41 4 pericentiyeast.....-.-0- aoe 1-3 6.85 5.85 eso 76 =115),'7/ So pencent yeastua-.-- sae i 1-3 6.55 4.50 12.5 -14.5 28.75 5d 35. 25-37 .25 Hot aqueous extract barney 1-3 20.00 5.0 26. -28.0 Retarded on cold extract, ba- nana. Accelerated on 26th 1-3 32.0 5.0 38. -40.0 day with living yeast...... J same body weight by underfeeding of protein or by feeding on proteins lacking in the amino-acids necessary for growth. In this way the ‘‘menopause was postponed long beyond the age at which it usually appears.’”’ The capacity to grow after adult age is not lost if the rats have been retarded (15), since after being stunted for 100 days beyond the normal growth period, they may reach full size when put on an adequate diet. It “appears as if the preliminary stunting period lengthened the total span of their life’ 17). In March, 1916, these investigators produced evidence from their experiments that ‘‘after periods of suppression of growth, even without loss of body weight, growth may proceed at an exaggerated rate for a considerable period.” Larvae which have been retarded in their growth by an inade- quate diet and then given the proper amount of yeast food develop into normal pupae and adults. The pupal period re- mains the same for larvae living twenty-eight to thirty-two days as for larvae that have lived five to seven days. This is shown clearly in figure 8 and table 5. Thus the total span of the life- cycle could be increased from eleven days to forty. days or more by retarding the rate of larval growth. This is shown in table 9. b. Sugar requirement of adults and larvae. The preceding experiments on synthetic media show that both adults and larvae A NUTRITIONAL STUDY OF INSECTS B31 TABLE 10 AGAR MEDIUM LARVAL LIFE IN DAYS ADULT LIFE IN DAYS Nucleoprotein + sugars + salts + water. . 10 to 21 (pupate) 5 to 7+ Nucleoprotein + salts + water........... 5 to 21 (die) 1 to 4 Sugars + salts + water.......-....------ 5—(die) 5 to 10+ require sugar as food. On nucleoprotein and sugars larvae live ten to twenty-one days, grow to full size and pupate normally and adults live more than five days. On nucleoprotein and water, however, the larvae live five to twenty-one days without marked increase in size, but all adults die in one to four days. These results are shown in table 7 and figure 7. On plain sugar and inorganic salts, adults live five to ten days, while larvae die in less than five days without increasing in size. The facts are summed up in table 10. These results show that larvae require sugars for successful pupation, but live longer on pure™ nucleoprotein than on pure sugar. Adults, on the contrary, live longer on pure sugar than on nucleoprotein alone. From general observations it appears that adults oviposit more readily on sugar agar than on nuleoprotein agar, but de- posit most eggs and over the longest period on nucleoprotein- sugar agar. Therefore it would seem that sugars stimulate ovi- position and nucleoprotein increases egg production and that both sugars and proteins are necessary for the normal activities of both larvae and adults. c. The natural habitat of Drosophila. I have already shown by experiments that Drosophila requires sugars and protein and that these substances can be supplied in a most normal form in yeast cells, so that it isof interest to study the conditions that exist in the natural environment of the insect. Schultze (’11) has recorded the fly as occurrmg im many dif- ferent fermenting and decaying fruits, vegetables and fungi, fer- menting tree sap, vinegar, tumors, and animals preserved in for- mol. Sturtevant (16) describes a number of new species of 14 Carbohydrates are formed as decomposition products of nucleoprotein. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 1 ae J. PERCY BAUMBERGER Drosophila from decaying fruit, vegetables and fungi, and from feces. Henneberg (’02) suggested that the larvae probably live on the microérganisms present in these media. It is apparent that micro6rganisms are usually present in great abundance in the larval environment, and from my experiments it appears that they normally serve as food for the insect, since the adults are at- tracted by the odors of fermentation products (p. 68). Micro- organisms, known to occur on the exterior of fruits, probably usually contaminate the substratum before the flies oviposit. Flies assist in establishing a suitable flora by accidentally carry- ing in the digestive tract or on the minute hairs of the body many yeast cells. This can be seen in cross-sections of the body (figs. 10 and 11) and has previously been shown to be the case on page 8. The yeast growth takes place more rapidly in the presence of larvae, because the latter spread the cells throughout the medium. At first sight, this would appear to throw doubt on the ability of the larvae to digest the yeast cells, but serial sec- tions show clearly the process of disintegration. Figures 11, 12, 13, and 14, are microphotographs of successive sections through the stomadaeum, mesenteron, and proctodaeum showing the proc- ess of disintegration which takes place mainly in the middle digestive portion of the tract. However, as in the case of birds which feed upon insect eggs and seeds, many living cells pass through the alimentary canal (page 8). In nature Drosophila larvae are usually found in a substratum suitable as a nutritive medium for microérganisms and abound- ing especially in yeasts. In this environment the insect has available as food both the substratum, usually fruit, and the micro- organism. But why have the larvae become dependent on the latter? Three reasons immediately come to mind, viz.: 1. As all fruits which are soft enough for Drosophila larvae to live in are always infected with yeast from the air, the larvae would unavoidably ingest fungous cells with the fruit. The nu- tritive value of the food would soon affect the life-cycle of the insect and bring about a close adaptation to a yeast diet. 2. Larvae feed upon microérganisms and by their constant movements carry the spores throughout the substratum. In A NUTRITIONAL STUDY OF INSECTS ao Fig. 10 Microérganisms on exterior of adult Drosophila (section through leg) (De a2)). Fig. 11 Cross-section through abdomen of adult Drosophila showing ingested microorganisms (p. 32). 34 J. PERCY BAUMBERGER Figs. 12, 13, 14 Serial cross-sections through the digestive tract of a Drosophila larva showing the digestion of microérganisms (p. 32). A NUTRITIONAL STUDY OF INSECTS 30 this manner yeasts often become predominant and prevent the development of destructive molds. These habits of the larvae may have come about through selective killing, for a mortality of 75 per cent was shown in cultures where the activities of the in- sects were suspended by low temperatures. This high mor- tality occurred only in cultures contaminated by fungi and was not. due to the low temperature itself, but to the uncontrolled growth of the microérganisms. 3. Another explanation may be drawn from a consideration of the relative food value of the substratum and the microorgan- isms. The fruits or vegetables of the substratum (omitting beans, corn, spinach, etc., which are unlikely to form an appreci- able number of breeding places)have less than 2 per cent protein and are relatively rich in carbohydrates. Microérganisms, on the other hand, have a protein content of over 10 per cent, but are poor in carbohydrates.!* This is shown in table 11. The protein content of yeast cells is readily available, as the enclosing cell membrane has been shown to be closely allied to pectin (Casagrandi). Salkowski (94) finds that the membrane is composed of two layers, one of which forms on hydrolysis d-glucose and the other glucose and mannose. The membranes are readily destroyed by the digestive juices, as yeast is exten- sively used as protein food for farm animals and even for man (Salomon, 716). The components of the yeast nucleoprotein in- clude almost every known monoamino-acid cleavage product, 15 Kappes’ (’89) analyses of Micrococcus prodigiosus scraped from the surface of solid media gave on an average: Water, 85.5 per cent, and dry matter, 14.5 per cent, the latter having a percentage composition of protein (N X 6.25) of 71.2 per cent, i.e., 10.3 per cent for the whole cell. Nucleoprotein has been separated from cholera, bubonic plague, anthrax, and diphtheria bacilli and from B. pyo- cyaneum megaterium and Staphylococcus pyogenes aureus (according to Ben- ecke, 12) and from yeasts by Hoppe-Seyler (71), Kossel (’79), and Stutzer (’82). Clautrian (’95) showed the presence of glycogen in dried Boletus edulis (20 per cent), Amanita muscaria (14 per cent) and yeast (31 per cent). In 1866 Hoppe- Seyler found 0.25 gram lecithin and 0.44 gram cholesterin in 81 grams of dry yeast, and later Naegli and Loew (’78) found 5 per cent fat (stearic and palmitic acid) in the yeast cell. Yeast gum (mannan) makes up 6 to 7 per cent of yeast (by dry weight), according to Salkowski (’94). Later the following gums were isolated from different yeasts, viz., mucin, dextran, laevulan, mannan, arabin and galac- tane (Lafar, ’03). 36 J. PERCY BAUMBERGER TABLE 11 FOOD MATERIAL WATER PROTEIN PAT preg as ASH AUTHORITY INU GS ese ees oes 85.9-37.5] 1.0-0.2) 1.2-0.0)14.4-2.7| 0.6-0.1) Atwater (17) Vegetables.......| 94.3-44.2] 1.8-0.4) 0.6-0.1)/21.9-2.2) 3.2-0.4| Atwater (717) 8.65 Stutzer (’82) VOOR Re access coes|| Goll al. 2 0.4 21.0} 1.8 Atwater (’02) B. prodigiosus...| 85.45 10.33 0.70 eae Kappes (’89) Putrefactive bac- Nencki and (eM s oq gangeoe =| eee 13.96 1.00 0.78 Scheffer (’80) Mushroom....... 88.1 Bodh 0.4 6.8 2s en | Acti wieemn Calize) 1 Probably largely starch in the compressed yeast. Glycogen makes up 31 per cent (by dry weight) and gumes 6 per cent of the yeast. viz., glycocoll, alanine, valine, leucine, proline, phenylalanine, as- partic and glutamic acids, tyrosine, tryptophane, and probably serine and cystine (Meisenheimer, *15), and it is therefore not surprising that it forms, with sugars and salts, a complete food for Drosophila. As the preceding experiments show that Dro- sophila larvae require more concentrated protein than is present in the substratum, it is apparent that the habits of the insect are for this reason, adapted to the use of the richly protein micro- organisms as food. d. Function of yeast in the ecology of Drosophila. The func- tion of yeast in a Drosophila culture is clearly defined by the fol- lowing two experiments: 1. Larvae grow slowly on a weak, cold-water extract of asep- tic unfermented bananas and remain at about the same size for a period five times the normal life and then die without pupating. If this culture is left open for a few minutes, in such a position as to allow a few fungous spores to fall into the medium, the larvae will increase their rate of growth and pupate in a few days. 2. Larvae on sterile 1 per cent yeast agar grow to a length of 4 mm. in twenty days and die without pupating. If the culture is inoculated with a minute quantity of yeast cells the larval period is only seven days and is followed by pupation. In both cases the yeast cells remove, by adsorption from the medium, the amino-acid molecules in their immediate neighborhood. As this goes on a steady diffusion of amino-acid molecules occurs A NUTRITIONAL STUDY OF INSECTS 37 towards the place of lowest concentration, and thus the yeast finally adsorbs and builds up into its own protein all the amino- acids of the substratum. The yeast grows at the surface or just below it where it is carried by the larvae and therefore brings within reach of the larvae nitrogen that had been distributed throughout the medium, many parts of which could not be reached. In the experiment with banana agar the yeast not only concentrated the amino-acids of the substratum, but prob- ably synthesized them into more complex molecules; in the second example, the living yeast merely concentrated all amino- acids at the surface of the medium without increasing their com- plexity. In a synthetic medium of sugars and salts, yeasts would concentrate and synthesize into protein, the ammonia of the substratum. It has already been shown that concentrated banana permits larvae to pupate, but the rate of erowth is not normal. There- fore it is apparent that while the banana is not entirely lacking in the substances necessary for complete growth, it is not as ade- quate to these demands as yeast or yeast nucleoprotein. There- fore we may conclude that the function of yeast in the ecology of Drosophila larvae is to concentrate at the surface and synthesize the ammonia" and aminoacids of the substratum into nucleopro- tein, which fills the protein requirements of the larvae. e. Literature on the food of Drosophila. Valuable contribu- tions to our knowledge of the food relations of microorganisms to insects have been made by Delcourt and Guyénot. These au- thors reported in 1910 experiments with Drosophila in which they showed that the larvae could be reared on a potato medium free from all microérganisms except yeast or a complex of yeast and acetic-acid bacilli. Microscopical examination showed the yeast cells in the digestive tract in all stages of digestion. ‘This paper was followed by a second in 1913 (a) in which the authors deter- mined whether the insect fed on the products of the yeast’s chemical activities or upon the yeast cell itself. In order to ob- tain this information, it was necessary to operate with larvae 16 Yeast can also synthesize protein from a urate source of nitrogen. 38 J. PERCY BAUMBERGER that were sterile or with which only a single species of microér- ganism was associated. This was accomplished by means of an ingenious method for the aseptic transfer of adults from one flask to another and by the use of various media adverse to different species of molds and bacteria. This method of sterilizing the larvae is much less direct and requires more time than my method of sterilizing pupae with alcohol. The flies were finally found to be sterile except for the presence of yeast cells and these were eliminated by a rapid transfer of females from bottle to bottle, thus permitting aseptic oviposition in a few cases. The sterile larvae which emerged were then fed on a medium of potato and dead. baker’s yeast or dead baker’s yeast, water, and cotton. The authors at this time made no definite statement about the function of the microérganisms, but left that for later papers. In 1913 (a) Guyénot reported that he had been able to raise fourteen generations of Drosophila in the absence of living organ- isms. The larvae were reared equally well on potato and living yeast, potato and dead yeast, and on dead yeast alone, but did not grow normally on sterile potato. Guyénot (’13b) there- fore concluded that in nature the larvae nourish themselves principally on living yeast and other microérganisms. The work of Delcourt and Guyénot was unknown to me until after I arrived at similar conclusions’ by different methods. The experiments with Drosophila as reported above are therefore in part an independent corroboration of the work of these authors. Loeb ('15) reared Drosophila on a medium of salts and sugars with ammonium tartrate as the only source of nitrogen and there- fore concluded that this insect has as great synthetic power as bacteria. Later (’16) he pointed out that yeasts may have been intermediate in the synthesis of protein, and in a third paper (Loeb and Northrop, ’16 b) showed that yeasts serve as food for Drosophila and are required for the growth of the larvae." These authors were unable to isolate the substance in the yeast 17 My experiments extended over the entire period between May 1, 1916, and June 1, 1917, and were partially published in three papers (see Bibliography). 18 Eggs were sterilized by washing in 0.1 per cent HgCle for six to seven minutes. A NUTRITIONAL STUDY OF INSECTS 39 on which larval growth depended, but found that the micro6r- ganism when extracted with hot alcohol could no longer serve as food for the insect. The addition of those special substances nec- essary to higher animals did not take the place of the substance extracted from yeast. The insects could not be reared on the normal salts, sugars, and amino-acids or proteins sufficient for ‘higher animals, viz., cane-sugar, MgSO,, NaCl, and CaCl., with casein, edestin, egg albumin, or a mixture of leucine, alanine, gly- cine, asparagine, tyrosine, tryptophane, and histidine, or with milk. Twelve successive generations of the flies were raised aseptically on yeast, water, and citric acid. It should also be mentioned that Loeb and Northrop raised aseptic flies on aseptic unfermented banana, but were unable to secure a second genera- tion from them even after feeding the adults on yeast, as both sexes were sexually sterile. My experiments show that Drosophila can be reared normally on yeast nucleoprotein, sugars, and salts, therefore any ‘special substance’ required by the larvae must be present in this mixture. As previously mentioned, I have been able to rear sterile larvae on sterile hot aqueous extract of banana agar and obtain adults which appeared to be sexually sterile, as they did not oviposit on the banana during six days (the usual preoviposition period being twenty-four to forty-eight hours), but when half of the number were transferred to an aseptic 4 per cent yeast-agar medium, the females oviposited in one to three days. The larvae that emerged reached a length of 5 mm. in three days; the females remaining on the banana did not oviposit. Guyénot (’13, b) has explained this as a nutritional phenomenon. He observed that normal females from yeast-fed larvae placed upon a poor food, such as carrot, after a few days deposit eggs which though fertilized no longer develop to maturity, but die as partially developed em- bryos. If the same female recopulates after a period, it at first deposits normal fertile eggs, then abnormal fertilized eggs, and finally unfertilized eggs. The following experiments of Guyé- not’s (’13 d) will serve further to illustrate this point. He reared adults from aseptic larvae fed on sterile potato, but found that most of them were almost sexually sterile. Oviposition did not 40 J. PERCY BAUMBERGER begin till the females were seven to twelve days old (normal pe- riod thirty-six hours) and the number of eggs was 117instead of the normal 576 (24 per day). Only five larvae emerged, 49 em- bryos died owing to deficiency of the sperm, and 63 eggs were unfertilized. Anatomical examination by the author (’13 e) showed that only 20 to 40 eggs are normally formed in the body of the female at the time of emergence from the pupa. Thesé are deposited in forty-eight hours, and after that all the stored material in the eggs, normally 24 per day, must be derived from the body and food of the insect. The effect of the food of the adult upon fecundity is very marked, thus ‘non-fertile’ sister adults from potato-fed larvae were placed, 1) on potato, where they laid one egg per day for 7 to 13 days and, 2) on potato and yeast, where they laid 10 to 15 eggs per day after 5 days and then 24 eggs per day. The converse experiment was to place sister adults raised from larvae fed on potato and yeast on 1) potato and yeast, where after 24 hours, 20 to 27 eggs per days were de- posited for 10 to 17 days and, 2) on potato, where after 24 hours 20 to 27 eggs were deposited for 3 days and after that but 1 egg per day. These experiments all account for the death in the embryonic stage of eggs of a normal female, but the following ex- periment shows clearly that it is due to resorption by the female of the sperm cells in the bursa copulatrix (Guyénot, ’713b). 1) Adults from larvae reared on potato when placed on yeast laid from the 4th to 15th day 300 normal eggs, on potato after 7 to 13 days, 2 to 3 fertile eggs, later 20 eggs which died without hatching although fertilized, and finally 30 unfertilized eggs. 2) Adults from larvae raised on yeast, when placed on yeast deposited 24 eggs per day after 36 hours, and on potato, behaved the same as adult bred on potato, but the effect was slightly postponed. The foregoing considerations show that the fertility of adults is a question of gross nutritional requirement and that it is diff- cult to interpret the yeast requirement in these cases, as a need of special substances. This is especially true since the accessory factors or vitamines which have been studied by Funk (11), Os- borne and Mendel (’13), Hopkins (’12), and others are necessary only in extremely minute quantities and are not used up in a A NUTRITIONAL STUDY OF INSECTS 41 short period, as would have to be assumed in the case of the female Drosophila. Guyénot (’17) has summed up all this work in a thesis and has added some experiments concerning the exact constituents nec- . essary for a synthetic diet for Drosophila. In this he is success- ful to the extent that with one exception the components of an adequate diet are discovered. These are peptone, lecithin, inor- ganic salts, water, and an extract of yeast, the composition of which is unknown but appears to be a part of the yeast protem. molecule. This extracted substance is most completely removed from yeast by boiling in 60 to 70 per cent alcohol and can be rec- ognized by its solubility in boiling absolute alcohol, cold 70 per cent. aleohol, and boiling and cold water. Attempts to substi- tute amino-acids, cleavage products of nucleoprotein, nuclein, carbohydrates, salts, organic acids, and fats for this special sub- stance were all failures. Experiments with peptone gave best results when 4 per cent was used, but no larvae pupated unless lecithin was added, which permitted the storage of fats and pupation, but not the emergence of adults. The addition of bouillon to peptone also permitted a few abnormal pupae to be formed, but no adults emerged. Completely filtered autolyzed yeast, together with lecithin and peptone, made a complete and normal food for the insect. Liver autolyzed or extracted could be substituted for the yeast extract with equal success. The author also studied the formation of reserve fats and found that this process depended mainly on lecithin, but could go on to a slight extent at the expense of the protein derivatives in the yeast extract. These results of Guyénot do not necessarily conflict with my own, as the special substance extracted by boiling alcohol is probably included in the nucleoprotein used in my experiment,” as Guyénot has pointed out. No fats were present in the yeast nucleoprotein used in my work, as I had extracted these with ether. As Guyénot found that lecithin is required, there would appear to be conflicting results in this regard, however, he also 19 In drying, the nucleoprotein was washed with cold alcohol, but the special substance of Guyénot is not extracted unless the alcohol is boiling. 42 J. PERCY BAUMBERGER found that certain crystalline solids left on the filter after filtering autolyzed yeast could be substituted for lecithin. These crys- talline substances are probably also constituents of yeast nucleo- protein. Loeb and Northrop (’17) have recently used glucose beef agar for maintenance of larvae and adults so that the temperature coefficient of the duration of life could be determined, and Northrop (17a) has shown that the total duration of the life of Drosophila can be increased by retarding the growth of the larvae, as the pupal and imaginal periods do not seem to change with the increased larval life. These results are entirely com- parable to those given on page 30. Northrop (’17 b) describes experiments which have led him to the conclusion that yeast sup- plies a special substance necessary for the growth of Drosophila larvae. This author finds that banana, casein, and sugar sup- plement yeast as a food for larvae and permit the development of a larger number of adults than could take place on yeast alone. The optimum mixture contained 33 per cent yeast, and as the amount of yeast decreased the number of adults reared became less and growth of larvae slower until at a proportion of yeast of 1:128 the growth of larvae became abnormal. Kidney, liver, and pancreas of dog were adequate foods for larvae, but spleen, heart muscle, muscle, blood, adrenal, and thyroid were not a complete diet for the insect. The author concludes that the special substance required for growth cannot be obtained from protein or carbohydrates. From my experiments I have evidence (p. 14) that banana and sugar have food value for Drosophila larvae, and to this extent my results are in accord with North- rop’s, however, since the insects can develop normally on yeast nucleoprotein, sugars, and salts it seems probable that the special substances required for the growth of Drosophila are included in nucleoprotein. | In summing up the results of my experiment I conclude: 1. Drosophila normally feeds on fermenting fruit, obtaining a large part of its nourishment from the microdrganisms, especially yeasts, which are in a loose symbiosis with the insect. 2. Dead or living yeast is a complete food for Drosophila. A NUTRITIONAL STUDY OF INSECTS 43 3. The larvae are dependent on the nucleoprotein of yeast for special substances necessary in their growth. 4. The function of yeast in the ecology of the insect is to con- centrate at the surface of the medium and to synthesize into nucleoprotein, the urates, ammonia, or amino-acids of the sub- stratum. 2. Hxperiments with a sarcophagous insect A pair of adult Acalyptrate muscid flies of the species Desmo- metopa m-nigrum Zett. (determination by Mr. C. W. Johnson) were received through the courtesy of Dr. W. M. Mann. They had emerged from some poorly dried snail shells, collected in the Fiji Islands, on the decaying flesh of which the larvae had fed. The adults were placed on banana and yeast agar, where the female deposited about forty eggs, most of which died owing to a thick mat of a black mucor that grew over the surface of the medium. The six larvae that emerged fed readily on the rich yeast food, and in about twenty-two days reached a size of 12 to 15 mm. in length. The black fungous mat was not destroyed and did not seem to injure the larvae. The six pupae formed were normal, and six adults emerged after three to five days and oviposited on the medium. The usual manner of interpreting the normal feeding habits of this species would be to state that the larvae fed on decaying ani- mal tissue. This, however, is open to doubt in view of the above experiments, and we must now consider the probability that all decaying or fermenting substrata are merely the media on which the fungus and bacterial food of the insect is growing. 3. Hxperiments with a coprophagous insect An investigation of the food of the housefly (Musea domestica) also gives support to this theory. The insects were obtained in winter by placing bran mash in the greenhouse. The mash was prepared by boiling “Educator” bran with an equal volume of water with constant stirrmg for twenty minutes. It was placed in a large porcelain dish in the hothouse, where it soon became 44 J. PERCY BAUMBERGER covered with Rhizopus nigricans and Penicillium glaucum molds. These growths were turned under each day and the bran thor- oughly wetted till the mash had a sour smell and no longer be- came covered with the molds. Thus far no flies had deposited eggs on the mash, although large numbers of Lucilia caesar and Musca domestica fed on the moist surface. Several lumps of ammonium carbonate were then added, giving the medium an odor indistinguishable from manure. No eggs were observed, but larvae were soon found feeding Just under the slightly in- crusted surface. They gradually worked down in the medium as they became larger and the upper portions of the mash were dried out or the nutritional substances “burned” out by the fungus due to over-aération. When this same medium was used a second time, larvae were seen to penetrate into hard lumps of the bran which still retained a visible white powdery appearance of fungous mycelia. Two-day-old (5-mm.) larvae were transferred to bouillon yeast, banana and yeast, bran and Pasteur’s agar media. The larval life was as follows: Bouwlllontagar. fe .e.0 268 ee 5 mm. to pupation 13.5 days Pasteur ssa gar can seer ee ne 5 mm. to death 2.0 days Brana PaTsetese eee se) Pee ee ee a 5 mm. to imago 8.0 days WeaStra any mest ives nee bet Meroe 5 mm. to pupation 4.0 days Banana and yeast agar................ 5 mm. to pupation 4.0 days Tati als Mea tan Maren ye career asl Ney: 5 mm. to pupation 13-21.0 days The larvae showed signs of great disturbance when placed in the Pasteur’s medium where the sugars seemed to act as a poison to them. On bouillon agar the larvae were not very successful in completing growth and usually formed abnormal pupae; on the other agar media growth was more rapid than in the bran mash. This was due to the luxuriant growth of yeast, molds, and bac- teria which the bran, banana and yeast agar supported and which served as food for the larvae. Sections through the larvae from the bran mash showed a complete absence of all material except bacteria, fungous spores, and yeast cells in the digestive tract. The microphotographs in figures 15 and 16 show the process of digestion of the microédrganisms and leave no doubt that they A NUTRITIONAL STUDY OF INSECTS 94} SUIMOYS BAIG [ BoTJSouIOp % vs NI, ® Jo yor > 1 OATJSVSIP oy} Y snoly} “(FF “¢ I ) so.1ods sn BUN JO UOTJSonIp SUOT}IOS-SSO1D [BIIOQ OT ‘CT “Sst 46 J. PERCY BAUMBERGER serve as food for the msect. Furthermore, the larvae are never found in any materials that are not infected with micro6rganisms and are in a process of decomposition or fermentation, and it is doubtful that the larvae find in these substrata nutritional sub- stances which will be of great value to them, for the food mate- rials are rapidly changed to decomposition products which are either absorbed by microérganisms or are too simple in composi- tion to be available for the insect.22. The real function of the microorganism is to synthesize protein from ammonia, urates, ete. Female flies seldom deposit eggs in substances which do not have the odor of ammonia, which as products of the action of yeasts, molds, and bacteria indicate the presence of the fungous foods of the larvae. An attempt was made to sterilize Musca domestica pupae by washing in 85 per cent alcohol or in 85 per cent alcohol saturated with HgCl, for periods of from five to forty minutes, but in each of the 200 pupae used in the experiment, Aspergillus or Penicil- lium mold developed around the pupa. It would seem that the molds are carried within the pupae, although this is not definitely proved. In this connection the following quotation from Bogda- now (08 b, p. 199) is of interest: Wenn Calliphoralarven bei Anwesenheit von Bakterien nur von Albumosenlésung ernéhrt werden kénnen, so kénnen die Larven von Musca domestica umgekehrt mit Stérkekleister oder mit Gelatine ohne Zusatz anderer Stoffe gefiittert werden, aber, soviel ich beobachtet habe, nur dann, wenn Schimmelpilze und Bakterien da sind. Prowazek (’04) has found that Apiculatus fungi are usually present in the intestine of Musca and Sarcophaga larvae. A number of experiments have been performed to determine whether or not the pathogenic bacteria, with which housefly larvae be- come contaminated in their natural habitat, survive in the fly during the pupal period. Most of the results (Graham-Smith, 13) show that only such spore-forming bacteria as anthrax pass through the pupa alive. ebbutt (113) m this connection raised houseflies on agar, a little human blood, and living bacillus dysen- 20 See page 48. A NUTRITIONAL STUDY OF INSECTS 47 teriae (type ‘Y’) and B. typhosus. The fly eggs were sterilized by washing in 3 per cent lysol for two to three minutes. The pupal contents were plated and found usually to be sterile. Teb- butt does not mention the fact that the larvae probably obtained their nourishment from the bacteria. From these experiments it appears very probable that the larvae of Musca domestica feed on microdrganisms and are as- sociated with them in the same manner as Drosophila and yeasts. 4. Experiments with a mycetophagous insect, Sciara, and a mite, Tyroglyphus, living in decaying wood a. Experiments with Sciara. Through the courtesy of Mr. A. M. Wilcox who turned the material over to me, I was enabled to work on another fungus-eating insect found in twigs of the mountain ash apparently affected by ‘black knot’ or some dry black-rot disease. Under the bark and in the cambium of the wood slender white worm-like larvae, 12 to 15 mm. long, with a shining black head could be seen working in a glossy gelatinous sheath which they appeared to spin or secrete. As determined by Mr. C. W. Johnson, the larvae proved to belong to a species of Sciara, a genus of fungus gnats which feed in decaying vegetable matter and are pests on cultivated mushrooms. The larvae were transferred to a medium of bran agar which they infected with a mucor, a Gleocladium, anda Fusarium. The larvae moved on the surface of the agar through the field of verti- cal sporophores with their black globular sporangia overhead. Occasionally one would raise its head and sieze a sporangium between its mandibles. The disintegrating sporangia could also be seen in the digestive tracts of the semitransparent larvae. The mandibles are peculiarly fitted for such feeding, as they are quad- rate in form and having three large and several small interlocking teeth. The flat surface which the mandibles form would make it impossible to seize any structure not raised above the surface of the substratum. The larvae are also very fond of the mucilag- inous secretions or exudations which appear as brilliant globules on the sporophores or sporangium walls and as a sheath around the larvae. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 1 48 J. PERCY BAUMBERGER As mentioned above, the larvae appear to move in a gelatinous sheath over the wood. This habit has been observed in a number of fungus-eating Diptera and has been described by many authors as a secretion of the larvae. The following extract from Malloch (17) will illustrate the present interpretation of the habit: Nearly all of the larvae (Mycetophilidae) spin webs in the galleries they make in their food; in the case of species that live externally upon fungi the web is slimy, rather loose and irregular. I have paid particu- lar attention to some species I have reared, and find that the larvae of this group do not pass over the threads but through them as in a tube, the body enclosed except anteriorly. The threads-are slimy in nature and the presence of the larvae may be detected by the glittering sur- face of the fungus which appears as if a slug had crawled over it (p. 250). My larvae, feeding on mold in bran media, could be observed very closely under the binocular microscope. It was seen that in passing through the ‘field’ of sees the larvae usually took a cer- tain course, thus forming a ‘runway’ similar to that made by a rabbit in high grass. The sporangia of mucors are converted into a mucilaginous mass when the spores are discharged and the sporophores also secrete a sticky fluid, both of which stick to the surface of the larvae as they pass through the fungous growth. Thus a shining gelatinous sheath is formed through which the larvae pass. When moving over its course the larva ‘flows’ along in a large drop of liquid which completely surrounds the insect and assumes the same form. The surrounding drop, if stained with eosine and examined under the 1.6 mm. objective, proves to be a mass of spores of various kinds all ‘arranged as though embedded in a clear unstained substance. A larva will often reach out and eat a portion of the surrounding drop of another larva that is passing. The sheath when stained shows a mass of mycelia growing from the spores embedded in the gelatinous matrix of fungous mucilage. Brues (’02) has described the ‘web’ of Neoglaphyroptera opima Loew. and believes that it is spun by the larva which is. found under the bark of fallen trees. As the insect is quickly killed by evaporation, he believes the web to be a protection against this danger. The larva was described as at times moving its head towards the web as though eating it. A NUTRITIONAL STUDY OF INSECTS AQ From the foregoing observations on Sciarid larvae, it is apparent that they do not spin the web or secrete the gelatinous tube which surrounds them, but merely become covered with the exudations and spores of the fungi on which they live, and these spores, exu-. dations, and hyphae serve as food for the insect. Upon pupation the enveloping drop of mucilaginous material surrounds the last larval skin and the pupa forming a cocoon of spores from which mycelia grow out. If the larvae are placed on a smooth paper under the binocular they are unable to move their long footless body. The mandibles with their flat surface cannot grasp the small particles of fiber which do not stand out above the surface. If a drop of water is placed upon the larva it immediately moves about actively by means of a ripple of circular contraction which starts at the posterior end and rolls a collar of integument over the anterior end. The anterior end is then protracted and the process repeated. It is apparent that in such a method of loco- motion an enveloping fluid of high surface tension would be of ereat assistance. The function of the ‘accidentally’ accumulated mucilaginous envelope is twofold, first, to serve as a protection against evaporation and, second, to assist in locomotion. The larval period is about twelve days and the pupal period four days. Adults are very active and run about rapidly, the male when in pursuit of the female flapping its wings vigorously. Adults may be seen to eject a hard white gelatinous body com- posed of fungus hyphae, ete. The adults and pupae seem much more immune to fungus attack than Drosophila. In the pupal stage this protection may be due to the complete envelope of gelatinous substance. The female deposits several separate piles of light yellow spherical eggs on the medium. These likewise seem to be immune to fungus injury as the mold often completely envelops them without causing death. The development and movement of the embryo can be observed through the egg which hatches in three to four days. An attempt was made to sterilize the eggs and pupae, but death always resulted, probably owing to the soft exterior of these stages. The insects grew equally well on bran agar, yeast agar, and banana agar, feeding upon the luxuriant fungous growth always present. 50 J. PERCY BAUMBERGER The consideration of main interest in the present paper is the peculiar relation of substratum, microérganism, and insect which again finds an example in the food of this animal. It is is well known that molds contain enzymes capable of dissolving cellu- lose and hemicellulose, i.e., celluloses and cytases, which enable them to extend hyphae throughout the woody tissue of trees, etc., thus extracting all the nutritional substances. The nitrog- enous natter is largely stored in the form of protein in the spores of the fungus, whereas the excess of carbohydrates may be excreted in the sticky drops of the sporophores. The insect feed- ing on the fungus, the wall of which it can dissolve, derives the benefit of the enzyme activities of the mold. If a section is made through larvae which have been feeding in the wood, it is seen that the great quantities of wood that pass through the digestive tract remain unchanged in structure. On closer examination fungous growths of an exobasidiomycete can be seen in the tissue cells (fig. 17). These fungi are dissolved out by the insect diges- tive enzymes and serve as food. The wood eaten by Sciarid larvae is therefore merely the substratum in which the fungous food material is embedded. This type of relationship is quite common among ‘wood-eating’ insects and is quite comparable to the symbiosis of Drosophila and yeast. b. Experiments with a mycetophagous mite living in decaying wood. A mite of the genus Tyroglyphus (determination by Mr. N. Banks) was also found on decayed mountain-ash twigs and bred upon bran agar like the Dipterous larvae described above. Five mites added to the tube climbed about on the thick growth of fungus, apparently eating the spores in the muclaginous sporangia. As the mites rapidly multiplied the growth of the molds was checked and they were cleaned off the surface till only a few blisters or pustules of Fusaria remained. The mites could be seen to feed in large numbers at the edge of these pustules which served as food for two months allowing the mites to in- crease enormously in size and number. Many of this same genus of mites are known to feed on cheese, ham (on which powdery molds grow), and dry molds of various kinds, and manure, de- caying fungi, and vegetable refuse are always inhabited by mites A NUTRITIONAL STUDY OF INSECTS ae Fig. 17 Cross-section through the digestive tract of a Seiara larva showing the fungus mycelium on ingested woody tissue (p. 50). Fig. 18 Agar Drosophila cultures (from left to right). Ist tube inoculated from second tube. 2nd tube, large larvae and pupae on non-sterile banana. 3d tube. small larvae of same age as 2 on sterile banana. 4th tube, inoculated from 3, showing its sterility. 5th tube, large larvae and pupae on 6 per cent dead yeast of same age as 2 and 3 (p. 11). ny, J. PERCY BAUMBERGER of various species. It is quite probable that mites inhabiting de- caying and fermenting material, feed largely on the microdrgan- isms present. c. Association of wood-eating insects with fungi. The reason for the association of Sciara and Tyroglyphus with fungi is probably because of the indigestibility of the cellulose walls of the wood and the small amount of protein contained in them. The composi- tion of wood varies with the season and with age, species, loca- tion, and tissue, so that'it is difficult to make any general state- ment. Haberlandt (15) has recently studied the digestibility of wood and concludes that unless the cellulose is destroyed or changed, wood has little food value for mammals, as the nutrient substances are inaccessible. Birch-wood was found to have the highest nutritive value, giving the following analysis: water, 4.56; nitrogen, 0.108 (protein, 0.675); ether extract, 0.45; nitrogen-free extract, 61.56; crude fiber, 32.2, and ash, 0.46. In general it may be said that, except in the living phloem and cambium, the protein content of wood is extremely low. Carbohydrates are usually abundant and wood is therefore used in the manufacture of sugar, some processes yielding as high as 25 per cent. These com- pounds are probably quite maccessible to insects because of their chemical and physical nature, but are readily dissolved and con- verted by fungi due to their notable enzyme activities. The low nutritive value of wood causes the insect to either lengthen its life-cycle so as to be able to extract a greater amount of wood or it leads to association with microérganisms either as food or as symbionts. In the first class belong the large majority of heartwood-boring Lepidoptera, Coleoptera, and Hymenop- tera, for example, the moth larvae Zeuzera pyrina and Sesia api- formis with a two-year life-cycle; the larvae of the beetle Saperda populnea with a two-year, Elaterid larvae with a three-year, Melolontha vulgaris with a four-year life-cycle, and Sirex (Hy- menoptera) with a larval period of one year. The next step which may lead to the habits of the ambrosia beetles and termites might be the reingestion of material already passed through the digestive tract as described by Escherich (’95) for the beetles of the family Ipidae (Bostrychidae) which he believes to be adopted A NUTRITIONAL STUDY OF INSECTS 53 for the purpose of extracting all the nutriment possible from the food already comminuted. ‘This would lead us to the case of the Cecidomyiid, Asphondylia prunorum, which was studied by Neger (08 a,b; 710). The adult deposits with the egg on the prune tree a mass of fungous spores and mycelia which serve as food for the larva and finally grow on the tissue of the gall formed. The fungus itself is not concerned in the gall formation, but merely serves as food for the gall inhabitant. Upon the emergence of the adult insect the fungus breaks through the gall and can be seen as a white growth from the outside. The fungus, a Macro- phoma species, is very similar to the fungus fed upon by the wood- boring beetles (Xyleborus, Xyloterus, etc.) and has never been found elsewhere; the galls have therefore been called ‘ambrosia- gallen.’ At the pinnacle of this development may be placed the ambro- sia beetles and termites. Schmiedberger (’36) first gave the name ‘ambrosia’ to a protein-containing white substance which he found to be the food of the insect rather than the chips of wood cut by them. This was made certain by the observations of Hartig (44) and it was also decided that the white substance was a fungus which grew on the wood cuttings. The subject was further studied by Hubbard (97) in America and Neger (’07) in Germany. Though these two investigators do not agree in all their observations, they have made certain that different spe- cies of fungi are associated with different species of beetles and that these associations are constant for the same species in spite of changes of host plants or parts of plants eaten (Xyleborus saxisenil). The fungus is independent of the food plant, but de- pendent on the products of the insect. Hubbard maintains that the female consciously carries the spores of the fungus to the new gallery and sows them. Neger, however, believes that the spores become attached to the highly sculptured wing cases of the female as it leaves the larval gallery, the walls of which are coated with the fungus. The fungus that grows on the walls of the galleries is different for different beetles, but in general is composed of either a chain of round cells which are assembled in an irregular heap or of upright threads with a round corpuscular cell on the D4 J. PERCY BAUMBERGER tip. The latter condition is termed conidial by Hubbard. Escherich observed that these conidia served as food for the beetles and their larvae. As a certain degree of moisture is re- quired by the fungus, the insects never select dried-out trees for their galleries, but always bore in wood which retains some sap. Other species of ambrosia beetles (Corthylus) are able to live in sapwood in the absence of the fungus, perhaps because of the abundance of protein in that region. The fungus is usually prop- agated in a little bed of chips, prepared by the female, m which the egg is deposited. In other species the woody tissue, after passing through the digestive tract of the larvae, has a yellow mustard color. In this condition it is plastered on the walls of the gallery and serves as a medium on which fungus grows. Un- doubtedly, under these conditions the fungus may be considered as a chemical collaborator digesting food indigestible to the im- sect and furnishing it in a tender luscious form to the larvae. Hedgcock (06) studied fungi from various species of Ambrosia beetles and was able to refer them to the wood-bluing (Cerato- stomella), wood-blackening and -browning (Grophium, Hormo- dendron, and Hormiscium) and to the wood-reddening (Penicil- lium and Fusarium fungi. According to Escherich and Neger, the fungus in the absence of the larvae may assume a slightly dif- ferent form. Neger has also found fungi in the galleries of Ce- rambyx and Tetropium luridum, but is unable to decide whether or not it serves as food. The fungus-growing habits of the white ants or termites are principally known through the work of Hagen (’60), Holte ((99), Haviland (’02), Triigardt (’04), Doflein (05, ’06), Petch (706), and many others. These wood-eating insects build subterranean nests, the ground which they excavate being placed in a pile which itself is later used to form chambers. Shafts may be left in this superstructure with an outer chimney; these are used as permanent scaffolds and have little effect on the ventilation of the nest. The fungus gardens are either on the floor or suspended from the ceilings of the chambers, and consist of a mass of com- minuted woody tissue which has passed through the digestive tracts of the workers and is then built into the comb. The fun- t It - ~~ A NUTRITIONAL STUDY OF INSECTS eus which grows on this comb (as in the case of ambrosia beetles, really upon food cut by the insect, but largely indigestible to them, is described by Petch (06) as follows: The mycelium on the comb bears small, white, stalked or almost ses- sile ‘spheres.’ These consist of branching hyphae bearing either spher- ical or oval cells. The spherical cells do not eerminate. The oval cells verminate readily but it has not been possible to reproduce the ‘spheres’ from them. When the comb is old an agaric grows from it. This agaric appears In two forms, one of which has been assigned by various mycologists to Lentinus, Collybia, Pluteus, Pholiota, and Flammula, and the other to Aymillaria. It develops in a cartilaginous, almost gelatinous, universal veil and is a modified Volvaria. Other fungi which grow on combs removed from the nest include Mucor, Thamidium, Cephalosporium, and Peziza. As these are not found in the nest, though some of them are capable of development underground, it 1s probable that the termites ‘weed out’ foreign fungi from the cultiva- tion of the comb. The comb material is probably sterilized by its passage through the alimentary canal2! That the spheres form the food of the termites is probable, as in the case of the leaf-cutting ants; neither case ean be considered definitely proved. Termes redemanm and T. ob- scuriceps undoubtedly prefer fungi, or wood which has been attacked by ILO D, Phone wet It is most probable that the ‘spheres’ in the ter- mite comb and the ‘Kohlrabihéufchen’ of the leaf-cutting ants investi- gated by Moller are parts of a normal mycelium, and that their shape 1s modified by the insects only in a very shght degree, if at all stake The available evidence appears to show that the ‘spheres’ are part of the mycelium of the Volvaria, but it has not been possible to connect these forms experimentally. The fungus gardens are found in all chambers except the royal chamber: here the queen lies in state and is fed (Doflein) with a concentrated and easily assimilated food consisting of mycelial spherules by the workers. The larvae, according to Petch, do not show the presence of any spherules in the digestive tract, but may be fed on some regurgitated or predigested food furnished by the workers, which in turn feed on decaying wood. It is of interest that the queen termite is the only known adult msect which increases in size. The queen is usually the center of a pool of fatty secretions on which the workers feed with great satisfaction. 21 My italics. 56 J. PERCY BAUMBERGER Undoubtedly these complicated habits have come about by taking advantage of the enzymatic and possibly synthetic power of the microdrganisms. The type of association is the same as with Drosophila, but is further complicated by the physical prop- erties of the wood and the habits of the insect. Some insects that feed on wood are apparently not associated with any microdrganisms, for the burrow does not appear to be discolored. However, as a general thing, larvae are in symbiosis with some microérganism when boring in woody tissue. When feeding on leaves larvae are often completely or reasonably ster- ile, and this is what would be expected from the foregoing as- sumptions, for the tissue of the leaves is very soft and readily digested. Portier (05) has thrown some light on this subject by his studies on the caterpillar of Nepticula, a small lepidopterous in- sect that feeds in the parenchyma tissue of rose leaves. The eggs are deposited on the leaf, which he supposes is sterilized by the sun’s rays, and the larva bores directly down into the leaf, sealing up the entrance; the feces are not thrown out as in the case of Tischeria. The exterior of the leaves were sterilized by Portier and the whole leaf, with the excavation cut open, was covered with bacterial media. The fifteen cases investigated were perfectly sterile?? whereas all cases of species which throw out the feces (Lithocolletis and Tischeria) were contaminated with bacteria and fungi, especially Aspergillus niger. Later (11 a) Nonagria typhae larvae, that live in the trunk of Typha latifolia, were investigated, and it was found that they were associated with a pseudobacillus present in all tissues of the body, having passed through the chitinous peritrophic membrane during ecdy- sis. The bacteria were in all stages of decomposition in the phagocytic cells of the blood. A second paper (11 b) showed that in Nonagria typhae a more complex situation exists than at first deseribed, in which two microérganisms are in symbiosis: a micrococcus and a fungus Mucidium (Isaria). This fungus must 22 During the summer of 1917 I examined the digestive tracts of some thousand Porthetria dispar caterpillars, pupae and adults, and found microorganisms pres- ent only in pathological cases, therefore this insect is not associated with fungi. A NUTRITIONAL STUDY OF INSECTS ar be held in check by the larvae, since if allowed to sporulate, as occurs after death, injury would result. Portier believes the se- cretion of the labial glands have this function. To lnk these observations with the case of Nepticula, in which the larvae are aseptic throughout life, Portier (11 ¢) describes the condition in Gracillus syringella, which at first feeds in an aseptic condition on the soft interior of the leaf, and then, after feeding on the ex- terior of the leaf and becoming associated with a digestive flora*® capable of dissolving woody tissues, bores into the twigs, the organisms being in part absorbed by phagocytosis as food for the larvae. Internal symbionts have also been found in a beetle, Anobium paniceum, by Karawaiew (’99) and Escherich (00). These symbionts always occur in definite cells in the anterior end of the midgut. Karawaiew thought he recognized a vacuole in them and therefore considered them to be Flagellates. Escher- ich, however, studied them in hanging drops of sugar solution and determined that they were Saccharomycetes. As these yeasts always occur in the same cells and pass through the pupa into the adult, it is quite likely that they are transmitted through the egg from one generation to another. Escherich found that the number of yeast cells varied with the amount of nourishment taken in the different stages of the insects’ metamorphosis, thus they were very numerous in the larva, rare in the pupa, and few in the adult. He therefore concluded that the fungi are inti- mately concerned with the nutrition of the insect. As the Ano- biid feeds mainly on very dry house timbers, the symbiosis with a fungus could very well be of value to the insect in the extraction of food from the wood. In general we may conclude that insects overcome the disad- vantages of the chemical and mechanical composition of wood by association with microdrganisms either as food or as internal symbionts. *3 Henseval, M. (compare Biedermann) ascribes an antiseptic property to an essential oil secreted by Cossus ligniperda larvae. This oil has the property of making the wood more workable (‘angreifbar’). DS J. PERCY BAUMBERGER EXTENT OF MYCETOPHAGY AMONG INSECTS As a corollary to the foregomg conclusions we may assume that the foods of many insect larvae feeding on dead, decaying, and fermenting vegetable and animal matter are the mico6rganisms which live upon the substratum in which the insects are embedded. The extent of this habit among insects is very great, includ- ing a large number of Coleoptera and an especially large number of Diptera. This habit is usually apparent from the habitats selected by the insect, thus Metealf (16) lists the following habitats for the scavenger short- and long-tailed filth larvae of the flower-fly (Syrphidae), viz.: In decaying parts of trees and herbaceous plants, diseased or flowing sap, heaps of turf or soft mud containing vegetable matter, and im stagnant or putrid water, sewage, manure, or human feces. The larvae also occur as accidental body parasites, causing intestinal, nasal, auric- ular, and vaginal myasis. Some species serve as scavengers in the nests of termites, ants, wasps, and bees: It is apparent that microédrganisms abound in all these environments, with the possible exception of the animal body In the latter case, how- ever, it is well known that a foul odor, indicating some bacterial action, always precedes infestation. In more normal habitats the microodrganisms so completely outweigh the other nutritive materials that it is quite likely they (the bacteria) serve as food.?! Townsend (93) lists the following habitats for some of the scav- enger Acalyptrate muscid larvae: dung, decaying wood, under bark, plants, leaves, roots, tubers, and fungi; in salt or alkaline water and mud; urine, vinegar, sap of wounded trees; cheese and animal fats. Again in this ease all habitats selected by the fly normally abound in microérganisms, and it is quite safe to assume that they (the fungi) serve as food for the insect larvae. The great extent of the use of microérganisms as food among insects is shown in a table of the feeding habits of larval and adult Diptera. In this table I have assumed that the food of insects, that always inhabit substrata of a fermenting or decaying nature, 24 Osborne and Mendel (’14) showed that the bacterial content of feces was 20 to 40 per cent. GROUP Mrpulidae...: 6.6.45. Ctenophorinae...... eR Wim aes eisai = Limnobiidae........ Cylindrotominae.... Limnobiinae........ ILynnONO) ONE), ceoangeoue Pediciunae..25...... Limnophilinae...... hima ophillare ss. +s - Eriopterinae........ iEvelobrae sane eae Gnophomyia........ Hexatominae...... Trichocerinae....... (irichoGeraseneeaee Ptychopteridae..... Rhyphidae........ Boletophilidae...... Mycetophilidae..... LUGS Ce See a ey ae Platyuridae........ Psychodidae........ Blepharoceridae... . Culididaes ee Dixidaeneneree are Ceratopogonidae... Chironomidae...... Orphnephilidae..... Bibionidae......... Scatopsidae........ Simuliidae.......... Stratiomyiidae..... Stratiomia.......... Odontomyia........ Oneycerna. si: aso. Geosarhus..c....e. Microchrysa........ Eupachygaster..... Xylophagidae...... Coenomyiidae...... Tabanidae......... ' Coniops..... Stee eee LARVAL FOOD fungi Saprophytie a eM A tal Lo a ADULT FOOD 2 a 2 gf E 3 3 2 2 8 {si a] S2jK5) 75 5 a | ea | 8s | 58 | 8 | 22 3 = I o4 a9 ier = Ba Sp ‘ef 3 Ae a, | ae 3 BE = Fa . (0 emmmOme = 21. 1 (° We" lel ates | Ol: 1 \Onogamee clear < meseieet- yoo! A ieq A STUIO ' ‘ 66 Yooy yynowdAlg powvg Aq © ystus0y = 0 §9°9T OZ 26 iF Mi 0 seer 6F1\ 0¢°ST: 09°02 | OT: 0% (0 :09°ST :0S*0% [T: ei : 02 | f T | 799 Gung ay Cae eae: =e & : ae: ‘ J 0 90°ET 10°T3] |. | noc 0 2z'St 60°95 \ 0 ST: OS OF JOS ST: 0G OE |O + OS°ST + 0S°OE |Z + 0S" 9T: OS "O8] f Rieter see : ae oa | Nias . OT | 6g¢ OF SINOS Houlendces MeCN OS GIA0G°rilc. = SI + &2 Us G G G Og 's ai VGH 8 G if 0 L9'IT 88°69 le aa 6 6 Ystutoy Aq YY Yooy yANourdATg posieg OSnS 05ST 29 SE 10) 25 OSS OG aes Ores aero cel cm) 0) SITRGTan GSN. 0 $9 ST 96°9¢ (@) a 1s ee es 6 G02 Ie 6 p>) 822 lige preeonte |_¢- = Te |lp 109 en 0G cou |Ouemeymmemeeme aula 0.988 28) |, oe 0 99°F 87°09] |. | ccc Sie. Mee 40 Ne 62 Ore | Se eae ee || TCE SOS IG 8 re I) SOD 6 eal Batra i ; SSUTIYVIY OWIBY) URIPUT YSsiu10Dg SSUTYVU YOoY YINowATA polvgd og og | og oc |S! og o¢ |S | o¢ | of | wa] = |l oe og | og oe |s| og | seao {S| o¢ | of | wad | & Tepuy) | eAO |JopuyQ | JeACQ | Oo | Jepug |} 10AGQ | 6 |z9puQ| 19040 ais = Japuy | JaAQ jt9pug | 1aAGQ | d| Jopug |: pue gg |o] Jopug} I0AO aS = S| g bles SI 3 cs Se ry se oye poyoodxg | o1yva poarosq¢ oye poqoadxny 01481 poaslosqa fs iS O1YB1 payoodxny| O17R1 poArosqaG, oOlyel popodxnT OLVBI PIAIOsqC, E yi ge] 3 B | & ANOUHL AAILVNUGLIV AUOTHL S$, Tav ad 5 e AUOGHL AAILVNUALIV ANUOGHL $,Tuvad 3 s GL, “DAG WoLf UayD} vg “fiLOIY} BQ DUL9]D ay) pun hioay) § )4Dag Yj0g Uo pasng ‘pajaadxa puv paasasqo ‘souvy % WITAVL 94 (Tt | 29g T | 29¢ 0 69°21 9T 29] }T | 79S TOT: 28°61 log 91: 0S'6T [0 : ST : 8ST [e :OS°STOS°GT) |T | £98 I | 7g¢ T | zg¢ eS ae ae am G0) GueesOne ice mda scour I¢. 8" 0 See |. re eee ee ey le | zee 1e-% :29'% |0¢'F: OS's 0 : OS's : OS'S [0:09 F : 09% 0 €8°3% 8L ‘9% \e ong m8°G :et'or 10¢'9: 096 0 : 8 : 8 j0:089:096| | 0 GZ TT 8°98 ¢ | 11g GTede oO \imecisece (A SW = 2S OP Gis 0 &0°ZT 182] \ | = ee rar ate ll L | 928 ea) 1 | Oc: TL lOO F0G°8T 06 eT\9 : Oc > TH / ‘uoTyVIOUES (J) POIg-Sso1o PUODIS oY} JO SSUTYVIN 0 6 oF ye a COUT) 00 8) 2 OO: 2 tig 4 0 st TL 2 \oy ies Glam MeG? 6: Ce 0) gee. OT. > Zach LE 61: 9 GI GL TE: $e °9S OG TT: 0S 0& GT OG 6T- 66S |/,9°g : eT Or |0¢'8 : §1 6 €1 66 9T 0S OT ee iis pela) 0:92 7: 96°6 06 :% 2G: 01'6T Cots Ohenes OS LOG FL OL SIT ce'P - 91 -S9 TT 0 : 0S TT : 09 0& 0g €1 0g TI N Ve) 0. GL 8T OF 6F 0 09°6T ¥:1G6 9F 61 0 92 °9T €F LF Loh Zz 0 61 9T 69 SF GeV ooo. 0 09 2 89 €S P09 ST: 09 GI 0 T8°&T 9T 8% 6:66 +: Lg 0 0&9 °9T 91°89 Ler Giees 166 0 LT 18°99 0 TL ST 08 6% IT :0¢°2 -0¢ OT Téd €9¢ 9¢ €L¢ THE JOURNAL OF EXPERIMENTAL ZOOLOGY. VOL. 28, no. l 96 0 T 92°19 ar GO: Sf 0S f :0S°op 0 ome S he |02 OSeT = 09 Oly 0 9% 3T 09°09 £9 OT: LE°ZT (09°02: 0S'°S + |Se°F 0S TT: SZ OT|S :0S'ST: 09'S | { 0 03 19°29] |, 99-9) 9 20) 0G) aa eso eames ia 0 %9°9T 19 SP a a, We : ete Te y | Sz¢ . Cee Gl 09 2a0c36 d 0 :0¢° 22: 0¢°6 On seeacl aLy le ame ae) et pe (Veo eee Semeeriemts. (0) 243-3 18 if 0 19°6 Wales “leap ee : € | gzg Bex fia Ce Oem eeanon ee atl a: 0) We. 6 .: T 0 92'€2 09°79] |, IS¢ 69 | p> 9 10 3 Goo) “GesOm sya f 4 0 19 ST 09 9F le Bae . GR Aros Mitten me = Nl) so MODS Aker derce lO Masemts mace 0 € FT 09°29] |. OG ee Onp ers \areets myer |0,.0S bear sO0 Caleta ac ay f PANUIYUOJ—UOTYBIIUID (yf) POIG-SSO1d PUODOS oY} Jo SSUTPBIY paINU1jUo sy) — SBUTZVUL yooy yynoudTd polieg 0g 08 0g 08 2 0g 08 5 0g 08 ad - 0g 0g 08 0g > 08 yao |&| og 0g ad » Jepug) | JoAQ |tepuy | 10AQ Oo | Jepug | IAQ | Oo |Lopug] 10AGQ Bute = Jopul) | tAQ |Aopu~ | JeAQ | Oo] Jopuy | puBoEg |o jfepuy)) JOAO |] 2S i oi Si a ry Es ~ a oy o1jel poyoodxi OLJBI PIAIISq(¢) oryed peyood xn O1YVI PoALOBGO g y ores poyood xg OI}BI POATISG CE O1yes poqoodxiy OL}EAI POAIOSG() S 4 re -_ 2 S| 5 2 ANOGHL AAILVNUALTIV AUNOGHL 8, Tuv dd 5 . B AYOUHL GAAILVNYUALIV ANOUHL §,Tuv dd re a panuyuojg—p WLAV.L 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 compaaable 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 ege 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 ratio, but we have not done so. By accepting the discrepancies mentioned between observed and expected ratios, we avoid classifying an individual breeder 0 €€ 92 9% 0 oT LE O €F 02 92 8h 0 19 LT 9F 09°9 -09°6 Go) Ob\ TOSI On| AO Serene lies | ZOGN ell SSI eno: Od) 0 tO On CSO 90 6 Sha SZ OTe Vy Wt 3G G2 29) OMS Ofer Gale irate 0 Sy LO oP = Vee pal asst Oe aaC 0 92°62 £699} 009 22 HF FS 0 19°9T 09 FF 0 L9:9T 09 FF y 2 FTGeT 9h 91] O 29685 9L 71) O 209° @:09 1) > 9\S08V Gas 3G) G8 TSO S57 eee ea er OMe enol aChen ey, [etl tA oo UW) ee ANS aR oe ict ie) DG ee SaaS OR) oo. ey | Ne te 009°9€§ 09°99, 0 O 9808 0 09°42 02 2h O BT LOS} f TOG G09) OV Ge 9 0 OFS 8 | CICOSY 097% 097 102 F092 |. (0) = Sess On OR mene amr On | anIGL en Seeeeed [ESSE el CMR] Negro Oceans jl ee aCe aL ey 4 (Po i Got hs05 ey Sea Ome eee 0 09°42 09°€9) 0 09 02 bP 0 tT 82 LT 6b 0 09 GT 98 0F §& < $148'@ -69'& 09° O- € -09¢| OF & © FI? HEMIV 71 = £109°9 09°71) O : 0781-092) 0 =O = Sr 201eyi V2 SG CPt eOpewenbas. Ge Ore (Gos. OF ie co 4!) Lo? Wile La 2 PTS OU eo ener f 161-816] “ya I9A0 og IOAO OF IaAO 0g J9AO 4 Ke OF I9A0 0g I9AO OF I9AO 0g 12A0 2 rs Japuy a, rapuyQ a O392, Jopuy) ae A Iopuy ie 5 5 z riepuy) Le iepuy) La 2297; opus) Ls 0407; riepuy) be 5 s & SI 5 =— iS J oe OF og OF s 0 Cae 0 og 0 og see a eel) ee Guilin 2 8550 UI JUIod UOIsIAIG $390 UI 4UIOd UOISIAIG 3 iS) S359 UI JUIOd UOISIAICT 8399 UI JuIOd UOISIATG | o = = & z &S cA AUOTHL AAILVNUGLIV AUOGHL §,1uvad a 2 AUOGHL AAILVNUALTV AUOGHEL 8, Tuvad s z r 7 PID]S WSUMNLIY]O SSaJUN ‘spay pwunjpsy apoyy adh} paovf-yonjq ur dnosb yova sof uoyonposd Bba una py pup ybry uaanjaq sjurod uUorsrarip pDLadas YUN puD fisoay) DAYDULAYDY 9Y7 pun hisoay) 8)4DIg YIOG WO pasng ‘pajyvadxa puy padsasgo ‘soUuDvy “soUDIT UL pajvadxa ‘UM WoY UL SO’jDL paasasqg $ GTAV.L “suvonpoid aLoovpaut 98 0 29 ST 0| 029 °8T 0|| 0 ST 6h .0. TT 9% OF) | O9 TT: 09'T \92°6 *96°8 | O ¢ ST O| O° ST <: O | GISZEV NOLS 66-8 \el°9 70s 'S | 0 7099-0959 los 0g 9 07 o> cielo Cl 10 Sl On Ons ter o| 0: et: of] im pts g leer A te) Se | Cael Zo} 098'L 09°2P 0 08'6 0| 008°6 0 09° OT: 091 (3) atts Sefer CR 3 ONS Te 1G) GT Wale Hes GL Nes Gl Ge Tae eee 0 198°@: GL IT +: 01> F1060V Aieee(0) Ole Calc Ol () leper sto mee at mT : 0 PI 0! | Pe 01 Onl? vee Un eo 0 €8°02 0| 0 OF LT 8 0 GE. Of; Ov pF 0 PAD ETN SSG ADE IEN (OS SON OPS SAI) SA tered eae me I 0| 0 I Ol) Os a Oke ee (0) a am Ee PN alo SHE pV PR ga RE I 0 | Db -0) | Ot ieet Ott O pee eee (iy ft OF Oy 2 0 | 0 ST. “ONS 07 = oT 0 GG Ge T1089 5091 | Ges PG GF 0\ 98 1 946°9 0 |} glesz ES PRO SO || Ore 7 el eO cee, 0 |} TOSI 0 Piece UO) sGarer Caen lec enG 0 || [eRe aiees ON (0) I On cOn: TE o || 0 ee°'8 0| O69 0 01 20sTs 07 0) = FE) eine = Co SI: GL 0 \96 GI: 90.0 | Ge :S°OL:01 « TONGGeGIGL-On sanOn ie 90g Ge ji g Bi| cOace We Bil Deep S |> Zi16cV Pleeeran() rT 8-0), (GS Rb ee 0) (Ck teem) aio |i Z Zi 0 g Tale Onsme Zz 0 91°8 0| 0&2'8 0 0 92°91 0a] 0 OT “ul Ga SiG OLGETG. 8G0508| 0 ay 0: 30) oa arava OLIN) SAI vA Gil 6 0G 066 | 0:05°9 094 21208 7 8-0) ie & 0] @ 8 4p One Oeaear 0 Goo cal ny. CUM ieue tee oS | Te eee GI6I-7 161 0 09°32 OG'TS} 0 TT LT OF 0 $2 09 TS] O FT ST LF CureetemOg 2S <09°9)| 0 ss /S se | 0 8097209 7G) vince: VA 9 g la) Me Ta ee tee oe | Courant Ga ee r:9 nace || Ae ap Pr oy eure alee) r Orin i Cees 0 06°$% 93 6P 0 61 88 °SF 0 9°32 ST'§9) OGL IT 92°E¢ OL - BIG6°9 = 4°91] 0 :09°2:08' TI] 0:09°9 -09°91|> SISOEV 148°6 °S2'2 |09°9 -O9 TIS 1-098 :46°2| O: * <: &Il> O1ezEV Ol Clee Omol Bes A Se eS to 6:8 ¢ 6l| TS ee Bele ae? ral 0 LE cy OO 0 0S 6E 0 €€ &T 0 0 €€ &T 0 fees at eee): (Q) oi I Os Oo “Gl pity 0g TT-0¢°0 He 7 09-7094 © 0 \09'7:09°'4 : O|> SlszIP Ucesciee aE OG 0 Loo 1 Oc UW) 5 6 (2 een) cl (OO Seon (0) © o @ o 9T6I-GI6T eee ye i | ty i Pal oe 40) Oo ft so @ Ooo fs OMe ss [ee (WPS eae} Wo la 0) 0 G0 ST 0 0 GL ST Té 0 0S'8T 9% 0 09'8T 9F G6 86° GL'G \GL 16. G6 °k 9° SF + OE OGL CO: OFF LITLEV GL 7 2967 \@h 1 2961 Qo Go IE 0: OFT - OFT | > GLze 66 : 0 86: I LGC) jy DG eh G3 I Ge - I (e853. =o |! OLS Gea 0 88 LT OF} 0 62 9ST 8€ 0 0S 02 0 00g02 0 CO N09) £ 20956 Go & ao @& O* L + §& \¢ SHCEV GI ONGISE GG. [aes =U) O° G < O01} Zj69GS G3 II SaeemaG Ibo yo a Ih Tete to Seo 0) goo [Seca Ze atoe 10) gent seco} 0 €€'8 0 0 8 0 0 &)6 37] O &b6 Fl) 02° 8 -0¢°0 |0¢°8 -0¢°0 Oo th 2 0) Go fo 0 |? 89 S14 Sh8 T1969 GL 6 109% 2099: 0 \G6 6097 °S6°@ | > E1F08 (@ - —={0) Cee OU ite ee a Ore coma) 8 MB XO) Somme. chop 1! [ko yi = I Eo 3 Il OF 7086 ¢ SUN) OF sono 0g CLO a Ke OF LOKO 0g LOO OF IdA0 Ge I9AO0 a fe jopuy De iapuy ae YE Japuy De oro 7; Japuy ae : BE Jepuy) Pe depuyy ae O17; lapuy We C197; Jepuy) ie : Bg F m “| 3 se OF 0g OF 0g 2 4 OF 0g OF 08 g 7 Lae Bi & $320 UI JUIOd uOoIstAIg $559 UI JUTOd UOISIAIG, : io) S550 UI YUIOd MOISTAIG, S559 UT JUIOd UOISIATC, 5 o & Zz : a Z AUOGHL AAILVNUALTV ANOGHL §, TV Gd a z AUOGHL AAILVNUGLTV AUOGHL S,TUV ad a : panuyuUuojg—G WTAV.L 100 pe Te ae 9 le a ree 0 02' $2 98 6% 80 26 GT 02°95 | OZ wtieePi eres 0c\) Ore ie see ; eL tO 0- 0 0 ay (0) 0 | (stu | 0OL2°9T Lb O OL FT 9 66) | ¥ 0 YS (0) Ve (0 ats ¥0) PED (0) ee (Aol eSO WN fst aT Ma PA OP VAM ROE of 0:09 TT 0g°T 0: 8 : G1? 9169%V (ese (0) iy Saal i aeeoie((). (te <)al| acl arena ol! ev Gl sacle ee teas Loe CL Soe area || 0 0 6 69] 0066 6&9 009°%3 09] O OL FT €F/) Ghar ce es, US eMGuING Th eOG. Gnag CONE O9Ne 6 Re ae oo hoe ee hg 9 eee) (Nee (ete DitalG «an 00F FT O 0186 9°38 009'8tT 99] 009 °8T 99)) Gia 1p GI ° G \G2°9 ° 8 5 O8°T 196-7: 86-6 °09°@ |> TISFLV Gis) 4) OL 1G (Goal eae ge Aime 8 Al oe 0 leas 7 eB (Ie 8 <0) Oi car Qe ste g LE | ee ema [sce aa eran Toa eek 0 o/S8T 0 0L9°9T FE 0 ye 0 07 0 01) (QOD BUEN VAS ere mOr RO TSI) G (0) 0:09°9 -0¢°& |> glgoze |109°0 -08°0 |0¢°0 :0¢°0 \0¢'0 : T :09°0 OS 0s aie | LIEZEY: Gen ee (N) Qu Sa ae] OSL PSE Ot eee eT [Me COVA ene Coes = 0h eye (Oj cm a0) 1 en ee of Seas a) 0 L0°€Z 26'S} 0 FF LT 69°09 0 L8°'1% 88°F%| 0 8E ST EL TF | 98°81: 94°8¢\09 01:09 Is| 0:09°S1:09'98| 0°94°6 ° 98 'eslsatlzI9e@ |l09 Tt: 09 TI09' 48: 09 SIS ST: 09°98: Ge'SI\Gd TT! 0996: GL 11) > 6|LOTF ) iptcrs Wd Olemecel. tae G22 Be eeat ce tr.-2 6 COmeAGT| Malem eGo ure Ot Meal sen OGaem il 0 88° €Z OL 6b 0 9F ST ZL SF 099'LT Th 0 98°9T 09°98 09°88: 09 T1l09° $1: 09811989: 09 21:96 O1| 0:96°81:94°08|$ SleTes |lee°SI: 96°F |GL°SI: 96°F G67 294° EL ST-0\96'7 2 6 FELT |> FILEV HCE OT MO less ST: ce Sete OL Ges eS er Titers, “il oy aed Cece oul Seay i ace 2 : Op Cc eO 0 9% GT GLE OnoicaeO 09912 &€ GO Ha OCG Ce AOL Le OLIGTAl< 99 ESIGN clog Oo ONOGe 6 09 -& (eee (0) ay SY Sa cera Oe 8 0) Opate cape ens. 8) 3.0 IME FE epee 0) aor wy 4 Oy Ga sae O. Sy see eal) OLGIuGhe 1S 0sO PLL Lb ara 0 oO 9%, 0 O oF G16 98'S |48°6 ° 89'S 0: 6 : *F |96°T:09°9 :94°9 |> Hesse OL ese ies ay: Oi Oey Os 0 JRRTIGLGY: On 2 ¢ Ge ee ip Chae I A Ol Ons Tl .O.? Oeste Oe Ont eee 101 d est sy || (0) oI 48°86: ST’ 88\96 GI: 96° L7 [ose aQe ls meOcs teeeh 9646: GLY 188 61-69 6 joc) tet aay 4861: ST 0677 6 5 99 Ss HES = 26] Ol +: & 0G GG: OF ¢ Lé « & GG € Go. he OF 3580 UL yutOd UOIsIAICT AUNOWHL AAILVNUGWVIV 0 0 L869) 0 0 18°69 0 €9 80 TL Z| 0 ae €& Lg Ob => OM ae SO. Oa ans T9225 &£ - 86968-9416 O : TI: Tél O te a GZT9€ (i (0) vc st) (0) 8 ey Il + ¥ fo OGNOm Ih = fo) OG: i, 0 09 LT ST €9) O ol FT 19 0 99°'TE 09°8F O LG 86 TP O $949 596°8é) O: F +: TE; HPSeh |148°e -SI'7 19405969) O -098 -09'S |. O “he 0 Go" é EIPLS eek Spe (0) Ae Oks map Mite ee ay i Oo) Oe meres (Oe 098 F& GL'9F O 0 06 TP 0 TE 98°39} 0 8T 79 8g 0 094 - 09d (O) (8) STN Veg AS G * 6&1 DD FIN ho i SIN) GO) GIi &|99c9 OR eee aoe Ome Oke Sil 0 (Gl Waa dl Were arr te i() alti my |k 029 €& OF 8&| OTT LT 89°29 0 9% 19°29) 009°ST FF 2g O -09° 46-09 66) O -9e°9T 4°09) - SISGLP if 416) 6-48 4) OF08'& -09°4)| O ©9413 966 | > 12208 Gi OGEa-me OG Guia sce tome ey Wo 4 COMMON Vino 2. ari o/h OCG oe 8 0 6T ST 09 6% 0 09 oT VT Gb 0 OT 98°69) 0 OT B08; 69 GELS OG TT: G6 L \89'° 9° OF TT: £8°8 | > FSLEP Bo B09 7° OG4G-| (Ole 8 sa el Oe 4 |? C9827 toy A) oes CEFN teh soma io 72 te a) A Coed Os AS Lee O aia L 0 LO°LT op LG} O TT OT &T 9 0 8€ 09 €6) 0 O LI 8 O 0961: 0¢'6I| O 2098 :09°7é| > FI8L9 O19 Ooo IN OO SO Oo OM 6. TOPS [Loe tale oe Ol | sa (6) so fer [hoe tS soe 8) Wo [fe SN A) to os 8) 0918 €&§ 9% OGL8 && 9F 0 €€ 9% O09 Lb 002 6T GP SF OG SRST OG GE | Soe wi eae nae a GIPSSo |196°8 -94'8 |0g'7 : sO GUO Le OT ON Ge CRT eT GL69 pe GIL Cea ots bio Sally eo 483 ho G) So GM (On oe tee oe 0) o5 AY eas i! 4161-9161 | I9AO 4 = OF IOAO 08 IOAO OF Id9AO 0g 19A0 a rs In C220 ae z a 2 |laopuy ion lopuy eae PZ |rapuy rae PZ lropug eed 5 5 Ei Ob 0€ | < OP 0g OF 08 5 ul i 8 | 2 72 rn s| 4 So80 UL 4UIOd UOISTAIC] I o Sp00 UL JUTOd WOISIAICT S500 UL JULOd UOISIATCT =| ie} - —— ———|#} 4 ||-—— : S| 4 7 ¢q i a q AUMOMHL S$, TXV Ad z AUOUHEL GAILVNUGLTV AUOTHL §,TU Vad cA panuyuog—o WTAVL 102 48°87°S9°81| 7S <= && PG. so STs BBG = 6c 98\77 OL: 99° TF eos)! «Tt * slh G6TT GL 7G v1 OF aar6 rr >_~ve.—EKS > 09°6 -09'6 \L8' #5096 - SI 2h o i2) hen! OPO TL Te GE 09°4 -09°@ 109° 4 -09'% | O ¢ OT 0\96'T: 9 :94°8 1) TI¢0E6 Oleere0) ae sees ye enna 0° 866Ze 82 £1) .0' OF 92,09 "8 9IIGVT, 00g 1616 |192°@ $989 0 S| 0-078 -097 | 0: 0 > Bl\ elec ‘9 IGRI, 90g S008 ee ee | 0 BE Ohya > eet nf) eee 0 68°33 80°SS| 0 Zh 8ST LZ ZF 0 & “ep T9). 0 9 20'09)) O TIZGZ6 \|GL°6 °96°9&| 0 88} O °9h6 °96°98| 0: O : 8&1} SiZ608 Oro eee lie « Se 6a a oo) 1S LEO = Reo 2G SO tee rel 0 €€'93 99°09] 0 OF LAT OF TS 0 T€'9T 68°89} 009 TT T1699 0 $094 -09°'4 | O 2GhE 2 Se TI|> SISLZ89 9s Si- 64 071096 - 08°77! O 2 GL II: 96°@t| O :09°6:09'747\$ 9I9z¢8 Ono Cy we pec. (Oi el tT) 20 Fl 0S Tee bene Fis 0 09'SE 82°29} 0 £OT 092g 0 st €9°99/ 0096 9&8 €9 0-099 -0¢9'9| 0: Of TH} ALLL |194'°9 - 96 61/9L°@ ° 96°86] O 099 -09°6I| O :09°&:09°e4s\$ HI9L8 Ol YP ee Oks 0 eee aL Oe. 61" Sitges. Oster Ol ss me Oe 0 26°23 09'Zh| 093 ST g8'gE 0 99°02 €9| 0 99°02 £9 OO: OL = Gi O09 L0G ee | TOlLFs 8S NOE TD S098 Onn 1G = | O097 1509 F | > sierzZ Oa ser GiliOrer see sea Sonia Sete BOOPs seat eel 20 se Gut me) 0 61'S 68°69] O €2°8T 99°09 0 8T' 08 €2§9! 0 82°22 98°89 0: 91: 98) O *GL6 509°@7|> LILFIS 68 715 89°89\96°9 24°19) O 20961: 09°96) O 26°92 94°T9|$ 812208 ee Ol 9¢| 0 [Mime ee 1, OL Ss CS Page eOON at 2 Sb +S). [i .caeeee . SI6I-LI6I 0 FT Os0) PT 0 » YStu 0 49°FT OS'TS| O19 FT OG TS Gay UE 0| @ jl 0 |> S| 10D 19601 94°8 1946 2967 \GL'@: L f96 47 | +O 209°L509'9 | FZISF Les Oni ales, 8% 0 |} Ol : > OReraral Tes! 6s ear eaten oem 0 9% 12'§9| OL9 ST FTL O :09°6 -09-6 |> TIZS8F |]09°SI: 09 Le\96'7 °24°98| O -09°ST:09'Lé| O :09°9:09°98|' OII8G¢ Gareeeile iow Gi Hemi Bt) 8G CLecmecie Oa: CS) Os CR seemed Pdi Gn) orice 103 ‘IoAoosyeyM Auoso1d AUG 9INdaS OF Iops1O UT A[GVaSUBY1ZUT SoTVUI OMY OSN 04 ATBSS9d0U SBA YT y 069° 6€ €9 F1; OLS TE TH 89 0 G1 TE 19 ZL} 0 99°€Z 0S '69 0G Té: 096 SI: 86) O - 09°06: 09°06) O OF ET 09 £6) > OILVIS ||69'°6 - LE GIGL'8 - 96°91) O -OFET-09°61| O -09°L:0G LI| > FII9L8 Co SON evs eC OCG sa OI Or PT 2c Glo ST) Oe ON 0 a> Gee Ci (een One O€T Lb 02 94) O FF 8T 99 0 LG LE 08} 09% °92 &€ 99 89°88: LE °F] Hi. ge) 0: GE = Gil O- OF = 86); S/2608 182 9°96 § 69'S 2489) O° GF = GIO se Sa eeIGEZe Some ee CemOmemncG ca GO) =. [lor LG lee SS Do OPO) eee Fee ae ee) 0 vE 8E 6L) 0 9% Th &L 0 €6 GF 96 SL) 0 TT SE 1029 der Ol SOL 9G) Ole Le 2 26) 0 5 S65 Ts\\) Gleacs 68: 66|9L 86: 96'7t| O :09°86:0¢'68| 0 : S@ : gF\} 312208 SG Oc Oca pClenQes ea "9¢| Oo “Ocern Pe Ch OG Fo 2 FP LT TP 9c 1 eee (soyrmoy YOrH) 8T61-LI6T I9A0 IdAO 1dAO0 ioao | 4 fe I9AO0 IdAO I9A0 IdAO | 4 2 09 UR 0g UB ¢ 09 3 0g 2 | Oo ‘a 09 2 0¢ 09 0g qd a Jepug Pa ropuy ee °PZ lapug Pi 0197, \rapug ae : Ba ropuQ a gopuy) ae 9197 |rapug Fae °Z lropugq Se : Be c Ed r L 09 0g 09 og S : 09 0g 09 0g 3 3 s 5 5 B §o59 UL JuLOd UOISIAIG So50 Ul JUIGd UOISIAIG 3 o $590 UI yUuIOd UOISTATG S550 UI JUIOd UOISIATG 5 o is 2 Bi 2 AUOUHL GAILVNUALIY AUOGTHL §,TUVad a 2 AUOGTHL WAILVNUGALIV ANOgHE 9,T4Vad a 2 papnpUeg—G ATAVL 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 whére 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 Lb | 0 08° FT 80°90 | uotjponp 092010 |0 92:0TO| uoryonp -o1d 399 OBVIOAY -o1d #0 oDdvIOAY #1: 01 \09°6 +0941 é é “-gyujoy poyoodxg || 09°01:09°@ | OF & é é “syegoy poyoodxyy PUTMOT | Clk tee CL leCie eGlueteen OLE fda a) oa 8/8401 Qs ie HST Oa) Oho ten 05 iia sos Oil scree s]B}0} [BNYOV Ce Pal eee a i 0: T: 9 |06 pun | sess 9: ¢| 9 g fio 1 g Poo ¢ |0e Jopuy) | Z869 Bast) 0) \hOe OMiGue Oi. UalGus 0: 0 \08 Jopun | LLEV jee (el a! 0 0 I 0 (Oa 9! 0 jog topuy | 2969 rome a Oe Ue ae OP Nese bearie AME cc Ic 0 jog opun | SLE een) 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 al 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 l. 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- voing cytolysis. % 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 oo WILLIAM B. KIRKHAM appear in these specimens, which offers some additional support to the statement of Asai (714) that this type of cell is embryonic in origin. The similarity in the histological details of the absorbed ¢m- 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 environ nent 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 genera! 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...........----- 21 26 Total normal embryos............----+2225++55> 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- 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 Niger mit Entypie des Keimfeldes. Anat. Hefte, Ist Abt., Bd. 51. Caste, W.E., anp C. C. Lirrte 1910 On a modified Mendelian ratio among yellow mice. Science, N.S., vol. 32. Cuaruton, H. H. 1917 The fate of the unfertilized egg in the white mouse. Biol. Bull., vol. 33. Curtnot, 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. Conn 1901 Experimentelle Untersuchungen iiber den Einfluss des Corpus luteum auf die Insertion des Eies. Anat. Anz., Bd. 20. Issen, H. L., anp E. SrerauepER 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. Logs, L. 1908 The production of deciduomata and the relation between the ovaries and the formation of the decidua. Jour. Amer. Med. Assn., vol. 50. MarsHalL, 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. Soporra, 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 fisiol6gicas 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 gldn- dula transplantada. Bajo el punto de vista psfquico 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 dela 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.1. 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, ie., 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 ease 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, 711, 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 femalcs. 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—.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 > 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 corium 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 aleohol, instru- ments sterilized in carbolic acid solution, and the table covering, towels, gown, etc., 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, etc., 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, 12. 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 4 Complete references of work done on rats have been compiled by Donald- gon 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 is “ 5 BODY JMS USE ee ies He eh ae s, mee ren ie ficat’ 261 DAYS grams | grams) grams| grams | grams | grams | grams ecm. I | Female with testis....... (ol 92a LOT) WA MAO RAS eGo el poliee() II | Female with testis....... Cfo I OB | A) ase all Teel Me WE |p LG III | Female (normal)......... 95199) | 27 Nae 25s SZ SA bm seO IV | Male with ovary......... 80 | 99 | 187 | 158 | 178 | 164) 4 1 V | Male with ovary......... 97 | 106 | 133 | 150 | 159 | 160 | 180 | 18.5 VI | Female with testis....... Sor Oa eA ale ON ale ciranl el ece2anl alts yl elt) VIN Male @ormal)s.20 2... 22. | 1027) 109") 1399 166s 1808) L90N 235.2085 VIII | Male with ovary......... 89 | 104 | 138 | 160 | 168 | 166 | 193 | 19.25 IX | Female (normal)......... 100 99 | 125 | 195 | 2a el GOs pee) | eligese 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 (I, I, 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 IX 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. 6 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 ™ Steinach seems to have used partly wild rats, partly tame (white) ones, and crosses between these. § 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 nestie 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 coveréd 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 a.m. 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 oppcsite 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, IV and n/male in opposite corner. 4.00 p.m. Feeding. 5.30 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 am. 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."3 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 rat 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 isclated 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- 4 The 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. 15 TIn 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 somejecases it has we vas . S iN fae 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 295 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 ‘6 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. 2 156 CARL R. MOORE Fig. 4 Section of testicle graft showing seminiferous tubules and part of epididymis, from muscle and fascia of female rat (6, AIBI - II). Graft same age as figure 3. st., seminiferous tubules; e., epididymis. 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 onto = oe ; ~ “S enw y ro ~ CIES Needing, qe Oa ¢ (Ce faeay es iG } Se 4G, 0 a an Ore XQ Kt 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 Donatpson, H. H. 1915 The rat. Memoirs of The Wistar Institute of Anat- omy and Biology, no. 6, Philadelphia. Goopatn, H. D. 1916 Gonadectomy in relation to the secondary sexual char- acters of some domestic birds. Carnegie Institute of Washington, no. 243. Kina, Heten D. 1915 See tables in The rat, Donaldson. Srernacu, E. 1910 Geschlechtstrieb und echt sekundare Geschlechtsmerk- male als Folge der innersekretorischen Funktion der Keimdriisen. Zentrl’bl f. Physiol. Bd. 24, S. 551-566. 1911 Umstimmung des Geschlechtscharakters bei Séugetieren durch Austausch der Pubertitsdriisen. Centrl’bl f. Physiol., Bb. 25, S. 723-725. 1912 Willkiirliche Umwandlung von Séugetier-Mannchen in Tiere mit ausgepragt weiblichen Geschlechtscharakteren und weiblicher Psyche. Pfliigers archiv. f. d Gesammte Physiol., Bd. 144, 8. 71-108. 1913 Feminierung von Miannchen und Maskulierung von Weibchen. Zentrl’bl f. Physiol., Bd. 27, S. 717-723. SrotsenBuRG, 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. * Seeariny 4 hd "Oratlint ole if s , A : ne eres Al Malte ») a tases tt) at : ye ‘ ra aon ta) “wbitsviises4 1 SPURL Ske ‘0 : ha Lis Vane ian Bik ney ik padintel nu obese ADE Aaa 4 ee Pee 1 baal ty ysheyt wiv nar: fyi 2? A YY at 1 Sie i \ g Olathe igi 4! Sh eotoib , pees ) oy ape valli cts Sty obpiatgraon, ous ne re . 4 vcr y non # “epi iba 2 3 Son omis tie rhc 7 sie fo al ioIt ti phi its oe BOT omar 9) apts i As 1 i! a : er 2 bvils 1 ‘shi icine fi, ‘4 ee ¢ i ; ; v i b« wT core hive ih ’ a 4er(yet ‘jel tania Tin vtrv ito) co ition lo 2% aon ! | gr gat is hal 4" 4 i i wal ive 4 ef Pent oA predduala, : : uleerial! ctate ; rs ! Ls 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 biparticién en sentido aboral-oral, seguida de regeneraci6n. E] 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 bipartici6én son generalmente desiguales. Después de la bipartici- én los bordes de la pared del cuerpo se encuentran y fusionan. Los del esdfago se reunen también y en el centro de la regi6én 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 mds proximo sea incompleto; pero si es com- pleto y no directivo se regeneran un par de mesenterios comple- tos no directivos mds un mesenterio completo tinico, 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 afade una serie de estructuras bastante definida y fija que no guarda relaci6n alguna con el tamafio inicial y la forma de los productos de la biparticién. Antes que la regeneracién se haya completado pueden llevarse a cabo otras biparticiones. Los individuos resultantes son, por consiguiente, diferentes en extremo con relacién al ntimero 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 regeneracién que proba- blemente es el medio mds 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 | ASEXUAL MULTIPLICATION AND REGENERATION IN SAGARTIA LUCIAE VERRILL DONALD WALTON DAVIS FORTY-TWO FIGURES CONTENTS Mb rOUMICULONG. fecha a ein sev 5 fo EAS a lee MT NE De ec Sec epeeee tte Sethe Deseriptionvof material : 20. 885k Ve REE ENS Pir kt RUAS Sian foe: Bassion snayrecenerption ay tcrs by. te, eo c4.ad sleet ces. bs oe. 3 aed: Bilan CYORSHSS1O Mice hook: Ee te pe er os er Oe eRe ee A ano Gourserob rezenerailone sa. sees Hat oe oe as ROR e caste tens ee Hormationioi mew siphonoglypusyes steel nade -eeseees aes Development of new mesenteries.............22.000cee eee eee Number and arrangement of mesenteries in regenerated regions........ Order of development of mesenteries..:::-... 2.5 oars ns nce dee ce pe Results of repeated fission and regeneration........................ Discussion of regeneration in hexactinians.........................2005. Resornptronyofoldumesenteries:* | 3.04 alas seas «cles yieak oe des feo Sequence of new mesenteries.................... BP SARE SS tr aro Siphonoglyphs and directive mesenteries...................00000000: Influence of old bounding mesenteries...............-..000-eeeceees Pairing oremesenberiestis tess f53'- BRIE. Ve. e248 ofade ceys SanewQags ~< Orne Sante sa a INCI CS 608 So oy hs, iA cee Orange stripes in fission and regeneration.......................---. External indications of internal structures.......0:...0.5..25...0....000% Coniposttton ofinatural eroups id. 72rd). U/l. as othe OL Siphonoerly phot - Pee lene sehc eke oe Rey ka 8s Ati Boat arise ke WER OR GET CR shi conse tty teak cE eae SUAS ae Bis a A ee WEATRE BULLER. oles eal NEMS Sere ARNG eR ares etc retin: cena net ene Form resulting from ontogenetic development.......................... Summing? ics bw «cvdker iid. hath ibe ieee Fie ae a pee es LEUNG (ct 0] 0h a oe eae re eee ee SeeReT Ag, Sa PRT ney oe ek a Re Symbols used in tables and plates.............. tee: eh th 3 SMe Ne Nae General explanation’ of tables'3 to 8)..000. 0). . ASR SNCT Mxhlestayio Mar pes ees BE Sep chad §. 04. Ca ed ste Be earns wil SEPA SP ba IFA pa orp ese Saya, 2, wh apalrerll DE eS Gabel che 217 218 219 220 221 223 wi 225 231 232 233 234 244 1 Contributions from the Zodlogical Laboratory of the Museum of Compara- tive Zodlogy at Harvard College, No. 318. 161 162 DONALD WALTON DAVIS INTRODUCTION This paper deals with the processes of fission and regeneration and their variations in the sea-anemone Sagartia luciae Verrill, together with the relations of these processes to the distinguish- ‘able morphological types and to the life history of the species. Many descriptions of isolated instances of multiplication by asexual methods in Actinians have been published. Among these are somewhat detailed accounts of the external aspects of the process of fission. ‘The present paper stresses the internal phases of this process, to which little attention has heretofore been paid. A few studies have been made of the regenerative processes following the separation, but no connected account of the external and internal features of division and regeneration in a single species has hitherto appeared. Moreover, in species of Aiptasia, where our information is most nearly complete, the processes are markedly different from those in 8. luciae. Variations in the details of form of Actinians have long been the subject of passing comment or of more or less serious study. That these variations, more specifically variations in the num- ber of siphonoglyphs, are related to processes of reproduction was first suggested for Metridium dianthus by Parker (’97) and has been demonstrated in that species by Hahn (05). A similar suggestion as to 8. luciae was made by Davenport (’03). This suggestion has been confirmed in the observations here reported. Owing to the presence in 8. luciae of external mark- ings which are not found in other anemones and which give evidence of fundamental internal structtres, it has been possi- ble to determine the relation of the processes of asexual repro- duction to these internal structures. The significance of studies of the asexual processes of S. luciae is increased by failure to obtain thus far any stage in its ontogenetic development. For various reasons, more complete knowledge of the life history of this species is desirable. Its appearance at New Haven about 1892 has been reported by Verrill (98); its rapid increase and gradual dispersal have been described by Verrill and by Parker (’02). It is now the most abundant sea-anemone ASEXUAL REPRODUCTION IN SAGARTIA 163 found between tidemarks on the coast of southern New Eng- land. Prof. H. B. Torrey called my attention to its occurrence in San Francisco Bay as early as 1906. Hargitt (714) refers to a report of its presence at Plymouth, England, in 1908, and records its identification at Naples in 1911. It may be col- lected at any season and lives well in the laboratory for long periods. A few liters of sea-water is all that is necessary to maintain many specimens indefinitely; a finger-bowl with a liter of water for changes serves well for a few. It may be shipped in moist seaweed without injury. It is therefore avail- able not only at the seaside, but wherever desired. I wish to express my hearty thanks to Prof. W. E. Castle for his kindness in directing the work in its earlier stages, and to Prof. H. W. Rand, whose encouragement and suggestive criti- cisms have been most helpful in the later stages. The work has been done chiefly in the Zoélogical Laboratory of Harvard Uni- versity and in the Woods Hole Laboratory of the United States Bureau of Fisheries. To the authorities of these institutions I am indebted for the facilities generously afforded. To these and to others who have in various ways lent aid, I here make grateful acknowledgment. DESCRIPTION OF MATERIAL Sagartia luciae Verrill is one of the smaller sea-anemones, but highly variable both in size and in relative dimensions. While each individual has its particular tendency toward a more elongated or a broader condition, the same individual under different circumstances may be elongated and narrow or short and thick, and owing to the varying amount of fluid contained in its cavities its total bulk changes greatly from time to time. In still water the animal tends to elongate and, in a state of normal expansion, shows a length of column one and one-half to three times its diameter. In strong currents the tendency is toward a shorter and thicker body, the length of the body being even less than the diameter. Specimens attain rarely a length of 20 mm. or a diameter of 15 mm., but in no case are 164 DONALD WALTON DAVIS these extreme dimensions approached simultaneously in the same individual. In the majority of specimens the more ex- treme measurement does not exceed half the stated maximum. The ground color of the column is olive-green. A variable number of orange or yellow stripes extend along the whole length of the column from the oral dise to the base. Less con- spicuous vertical lines along the column, darker than the ground color as seen by transmitted light, mark the attachment of the mesenteries. In oral view the siphonoglyphs, commonly two, may be distinguished with the aid of a hand lens. Except in very small specimens or in early stages of regeneration, each siphonoglyph is clearly marked by a chalk-white bar which extends from the siphonoglyph radially across the disc. The white bars appear with approximately normal prominence in the photographs shown as figures 1 and 2. In its internal arrangements this anemone follows the general Hexactinian plan, yet with great variation in detail. The number of siphonoglyphs and associated pairs of directive mesen- teries is commonly two, but may be one, three, or rarely four (tables 11 and 13). In one specimen five pairs of directive mesenteries were noted. The number of pairs of complete non- directive mesenteries varies approximately from four to eleven; although specimens may frequently be found, at an early stage of regeneration when the new mesenteries have not become established, with less than four pairs of non-directives in evidence. The form resulting from ontogenetic development has not been determined with certainty, but is probably (p. 218 ff.) a regular hexameric one with two siphonoglyphs, two pairs of directive mesenteries and four pairs of non-directives (fig. 26). In the animals studies, one, two, three, or even four (fig. 25) cycles of incomplete mesenteries are present. These are quite regular in number and arrangement in an area formed at one time, but vary greatly between newer and older regions of a regenerated animal. I shall use the terms ‘endocoel’ and ‘exocoel’ in their accepted significance, the former indicating a space between the members ASEXUAL REPRODUCTION IN SAGARTIA 165 of a pair of mesenteries, the latter a space lying between two adjacent mesenteries not members of the same pair. It will be convenient to distinguish between the endocoels partially enclosed by complete mesenteries and those bounded by incomplete mesenteries. The former I shall call ‘complete endocoels,’ the latter, ‘incomplete endocoels.’ Obviously, the spaces between members of a pair of complete mesenteries may be further dis- tinguished as directive or non-directive endocoels, and incom- plete endocoels may be designated according to the cycles of the incomplete mesenteries bounding them. The white bar on the disc extending radially from a siphono- glyph marks the position of a directive endocoel, with which the cavity of a so-called directive tentacle communicates. Or- dinarily the orange stripes seen on the column of the living animal mark all the complete endocoels and the endocoels of the first cycle of incomplete mesenteries (p. 207). In very small specimens these orange stripes may be wanting entirely. In regenerating regions they appear rather late and, so far as has been determined, without regularity (p. 214). In such regions no dependence can be placed upon the number of orange stripes in determining the number of complete mesenteries. By the time the regenerating area, which is at first paler than the older portion, approaches closely to the old region in depth of color, the orange stripes have developed in full number and, almost without exception (p. 213), accurately denote the number of mesenteries of the first two cycles. The specimens used were collected at various places and at all seasons. They were kept in the laboratory, for as long a time as desired, in small dishes of sea-water. When first taken to the laboratory the water on them was changed every few days. After algae became abundant on the walls of the dishes, changes were necessary only in order to compensate for loss by evapora- tion. Three or four times a week the slime which accumulates over the surface of the animals was washed off with a pipette. Feeding was attempted by placing minute shreds of fish or other meat upon the tentacles. This was effective, but very laborious. It was also hazardous, since any particles of food left in the 166 DONALD WALTON DAVIS dishes after feeding or disgorged by the animals tended to foul the water. It was found, furthermore, that the animals thrived quite as well without this attention, doubtless being supplied with sufficient food in the form of minute organisms. Under these conditions specimens decrease in size somewhat at first and after a time become lighter in color. In other respects they appear to remain in perfectly normal condition. In order to study the internal structures, specimens were fixed and sectioned. ‘The larger mesenteries can be seen in sections cut by hand, but smaller mesenteries are sure to be overlooked if dependence is placed upon that method of examination. My specimens were therefore stained, embedded in paraffin, sec- tioned, and studied with the compound microscope. For narcotizing, A. G. Mayer’s method of immersion in a 3/8 mol. solution of magnesium chloride has proved entirely satis- factory. After a half-hour in this solution the specimens are thoroughly stupefied and almost invariably well extended. Animals in good active condition tend to expand in this solution, even though they may at first contract from the mechanical stimulation incident to immersion. Unless fine histological fixation is desired, 4 per cent formalin is a convenient and satis- factory killing agent. From this the specimens may be trans- ferred immediately to 70 per cent alcohol. Staining in toto for eighteen to twenty-five hours in Kleinenberg’s haematoxylin brings out well the mesogloea, which gives the best evidence of the position of mesenteries and of the longitudinal muscles upon them. : Some of the specimens whose regeneration is recorded divided under observation in the laboratory; others were discovered, either in their natural habitat or in the laboratory, at the close of division or immediately following separation of the parts. Nearly all of the other specimens examined have shown exter- nally some portions of regenerated material. ASEXUAL REPRODUCTION IN SAGARTIA 167 FISSION AND REGENERATION The frequent occurrence of fission in this species has been noted by Davenport (’03) and the process has been well de- scribed by Hargitt (14). A very similar process has been de- scribed for a closely related form by Torrey and Mery (’04). The process apparently consists simply in the tearing of an individual into two parts by the migration of two portions of the basal disc in opposite directions. The first suggestion of an approaching division is seen in the elongation of the base from its usual nearly circular outline to an oval form. In the course of a few hours or days this elongation becomes more extreme. The whole body becomes flattened closely against the sub- stratum, evidencing a state of extreme tension. Eventually a rent occurs in the base. This rent widens, gradually extends up the column and finally involves the oral disc and the esopha- gus. Division may result in the production of more than two pieces almost simultaneously; although, as I shall show (p. 173), this occurs really by successive fissions rather than by simul- taneous division into three or more parts. I believe that the mouth and esophagus are invariably cut by the plane of fission. There is no suggestion of a division resembling that known as basal fragmentation. Any other method of asexual repro- duction than that described above must be very rare in the adult form of this species. . Certain external features of the process of reconstruction following division in 8. luciae have been described by Davenport (03), and in 8. davisi by Torrey and Mery (’04). After fission is accomplished, the two resulting individuals re- main in a limp state closely pressed to the substratum, and the torn edges of each slowly roll inward and fuse. Within a day or two the separate pieces, their wounds having closed, acquire the normal upright position and typical cylindrical form, and the tentacles become expanded. For some days after this the region of fusion of the edges shows as a narrow vertical streak of much lighter color than the adjacent old parts of the column. At this stage in less brilliantly colored specimens the new tissue 168 DONALD WALTON DAVIS may easily be mistaken for an orange stripe (cf. fig. 3, 4). In the course of weeks this regenerating part increases both in absolute size and in proportion to the old tissue until finally it forms, In some cases, much the greater part of the polyp. Also the newer region becomes darker until it is no longer distinguish- able in color from the older. Meanwhile tentacles and a faint white bar are formed in the new region of the oral disc, and pale orange stripes appear in the new area of the column. lag behind the other members of their respective pairs in becoming attached to the esophagus. A similar condition is represented in figure 18 and constitutes the sole evidence that regeneration is in progress. In the latter instance as well as in one half of the specimen illustrated in figures 15 and 16 the exact position of the plane of division cannot be determined. In figures 19 to 24 the condition of the incomplete mesenteries gives evidence that different regions are of different ages without indicating the precise position of the boundary between newer and older parts. It is obvious that only such specimens as reveal the exact boundary between new and old structures should be included in table 6, but not even all of these can fairly be counted. As indicated by the figures just cited, evidence of division commonly persists longest in the incomplete mesen- teries. Hence at a late stage of regeneration, following a division which does not involve a pair of incomplete mesenteries, there may be doubt as to whether one or both of the members of a pair of complete mesenteries lying near the boundary belong to the old region. If only one is old, the division was in a com- plete endocoel; if neither is old, the division must have been in an exocoel. See, for instance, the older regenerating region of the specimen shown in figures 15 and 16, where the mesenteries marked ¢ are probably (but not certainly) old. Such a case cannot be counted, of course. But specimens in a correspond- ingly late stage of regeneration after a division between incom- plete mesenteries, although they do show the precise position of ASEXUAL REPRODUCTION IN SAGARTIA 171 the plane of division, should also be excluded from a group on which to base a study of the relative frequency of divisions in different spaces. Such a case is illustrated in the newer re- generating region represented in figures 15 and 16. Therefore only such specimens have been included in table 6 as show the characteristic differences between new and old complete mesen- teries. This introduces a slight error through the exclusion of such a specimen as that represented in figure 17, which has no old complete mesenteries. Furthermore, since the specimens of table 6 are to be regarded as a random sample, so far as concerns the position of the division plane, there have been excluded from it all animals selected for sectioning because of the relation of the new area to the directive plane or to the orange stripes which mark the complete endocoels and certain incomplete endocoels. It is to be noted that the plane of division is in almost all cases strictly vertical. Occasionally a mesentery is found in an oral or aboral region, but not throughout the length of the column. Such mesenteries, as would be expected, are usually small. That they represent mesenteries torn during fission is probable, in spite of the fact that not a single case has been noted in which both paired individuals show parts of the same original mesentery. Since such partial mesenteries are rare in older parts, it is likely that they either are normally completed by regeneration, or are absorbed. As I have seen nothing what- ever to indicate that a torn mesentery ever grows up or down the column, and as there is some evidence (p. 172)that absorption of mesenteries does occur and that regulative processes correct certain abnormal adhesions, it is likely that the partial mesen- teries referred to are removed by absorption during regeneration. Certain other indications of departure from a_ perfectly simple vertical tear may be mentioned. In pairs 3 and 10 of table 3 there is disagreement between the bounding mesenteries of the related individuals. In no. 3a an incomplete mesentery of the most advanced cycle (I) stands in the position of a mate to a complete mesentery, c, in no. 3b. In no. 10a an incomplete mesentery of doubtful grade, (1) is similarly opposed to a direc- tive, d, in no. 10b. It is entirely improbable that in these cases 172 DONALD WALTON DAVIS the unlike bounding mesenteries were adjacent to each other in the original animal, since pairs of mesenteries of different character are rare. The apparent lack of agreement may be due in some cases to a tearing from the esophagus of mesen- teries previously complete or, in a single case (no. 18b, table 5), to the fusion of an incomplete mesentery with the esophagus during closure of the wound following fission. That complete mesenteries should occasionally have their connection with the esophagus broken is not remarkable considering the nature of the process of fission. The possibility of the abnormal union of an incomplete mesentery with the torn edge of the esophagus is supported by the fact that, in early stages of regeneration of cut specimens, atypical adhesions of parts are common. Such adhesions seem much more rare in later stages, indicating that regulation probably occurs. It is conceivable that an attach- ment of a normally incomplete mesentery with the esophagus might persist even though adhesions of other parts should be eliminated. There remain four cases (pairs 38, 10, 15, and 22) in which the most probable explanation of the disagreement is the complete elimination of one or more mesenteries. It is possible that this loss of mesenteries is due to the detachment during fission of a minute piece which was overlooked and lost; or it may be that such a piece was partially separated from the larger ones, or otherwise extensively damaged, during division and subsequently absorbed. In support of the latter hypothesis may be mentioned certain mesenteries found in a few specimens not represented in the tables. These are mostly pieces regenerating after being separated by artificial cuts. 'The abnormal mesenteries are at- tached to the column wall, but not to the esophagus. They extend through only part of the length of the column. The mesogloea of these mesenteries is thick and stains heavily, but the longitudinal muscles are feebly represented. They are cer- tainly injured old mesenteries and are probably in process of elimination. Summarizing the occasional irregularities adjacent to the plane of fission, we may say that small mesenteries may be cut in two ASEXUAL REPRODUCTION IN SAGARTIA 173 by the fission plane and later resorbed; complete mesenteries may be torn from the esophagus and persist as incomplete mesenteries; incomplete mesenteries may adhere to the torn edge of the esophagus and thus appear as complete mesenteries, and one or more mesenteries may be either wholly eliminated by tearing during division or by this process combined with ab- sorption during the early stages of regeneration. It should be emphasized that the irregularities just referred to are distinctly exceptional. In general the plane of fission is strictly vertical; and, considering the apparently mechanical tearing of the tissues in fission, there is remarkably general agreement between the mesenteries found in the separated parts. Ordinarily an individual separates into two parts, which re- generate for a considerable period before a second division super- venes. Occasionally one finds instances, such as are represented in table 5, of division resulting in the early formation of three or more regenerating individuals. The case of no. 19 is typical. It was a large diglyphic specimen with twelve orange stripes and giving no external evidence of a previous division. It was found upon Fucus, where the slimy foothold may have operated to prevent division. The presence of well-developed gonads, as shown in the photographs of products of this division (figs. 27 to 32), may also be associated in some way with a delay in fission under natural conditions. Upon being brought into the labora- tory, like many other specimens, it promptly migrated on to the glass surface of the container, and in about a week showed an early stage of fission. One week later a division was completed, resulting in one large and two smaller pieces. After three days more the largest part had divided into two. There were then four not very unequal pieces. Thirty-four days thereafter three of the regenerating products of the division were killed, the fourth having been lost. In no. 22 the interval between suc- cessive fissions was eighteen days; in no. 17, twelve days; in no. 18, two days. In nos. 20 and 21, as in the first division of no. 19, the observations suggested simultaneous production of three parts. The number of regenerating zones, however, as 174 DONALD WALTON DAVIS indicated in the table, are such as would result from successive simple divisions. The condition of no. 21 is conclusive in this respect, the evidence from nos. 19 and 20 being less positive because of loss of parts. I have found no instance of division of an animal into three parts giving rise to but three regenerating regions, as would be expected in case of a single compound division. One might expect to find that specimens dividing into more than the usual number of pieces would be shown to have had an unusually high number of old mesenteries. Such, however, is not the case. Only no. 17 possessed, before division, a high number of complete mesenteries. The average for specimens included in table 5, omitting nos. 19 and 20 on account of missing parts, is 13.5 complete mesenteries. The average for specimens dividing into two parts, included in tables 3 and 4, is 16.4. Evidently, then, the close recurrence of fission is not correlated with a high number of complete mesenteries. Indeed, it seems to be related rather to a number lower than the average; and it eventually tends through generations to raise the number of mesenteries higher than does the more common simple division. As will appear from the statements below, multiple fissions are not associated with particular numbers of siphonoglyphs and pairs of directive mesenteries. Tables 3 to 5 record the divisions of twenty-two polyps. Seventeen of them were diglyphic before division, three were apparently triglyphic, one was monoglyphic and one tetraglyphie. In thirteen cases a diglyphic individual separated into two parts, each receiving a pair of directive mesenteries. One of the di- glyphic specimens (no. 17) divided into three parts and another (no. 19) into four parts. In both of these cases one pair of directives passed to each of two of the resulting pieces, and one or two pieces contain no old directives. Another of the di- glyphic specimens (no. 16) divided into two parts, one with two pairs of directive mesenteries, the other with none. Still an- other of the diglyphic individuals (no. 21) divided into three parts, one part with one pair of directives and a single bounding directive, one part with the corresponding bounding directive ASEXUAL REPRODUCTION IN SAGARTIA 175 mesentery, and one part lacking directives. One of the tri- glyphic animals (no. 10) divided in a directive endocoel on one side, each of the two resulting individuals receiving a pair of directives and an unpaired directive (in one part the directive on the boundary being absent). A second triglyphic specimen (no. 22) divided into three parts, one receiving a pair of directives, another receiving a single directive mesentery on one boundary, and the third receiving a pair of directives and an unpaired directive. The third triglyphic animal (no. 20) divided into three parts. One part received one pair of directives; a second part possessed no old directives, and the third, which was not successfully preserved, must have received two pairs of directive mesenteries. The monoglyphic animal (no. 18) divided into three parts not through the directive endocoel. Consequently one part possesses a pair of old directives and two parts have none. The tetraglyphic specimen (no. 15) divided into two, giving one part three pairs of directives, the other one pair. Thus in fifteen out of seventeen cases of the division of di- elyphic individuals the plane or planes of division cut the major transverse axis, giving the pair of directives at the extremities of this axis to different pieces. The division of the monoglyphic specimen occurred in a corresponding plane. One of the triglyphic specimens divided along a similar plane as regards two of its pairs of directives, the division passing through the third directive endocoel. In three of the four remaining cases (nos. 16, 21, and 22) some of the directives are regenerated ones whose imperfective development at the time of division may give occasion for the unusual position of the fission plane. The fourth specimen (no. 15) also has some regenerated mesenteries, but the limits of the regenerating zones are obscure, owing, at least in part, to faulty preservation and to sectioning in a some- what oblique plane. What part the directives play in de- termining the position of the division plane may only be sur- mised. This question is discussed below. Here it should be emphasized, first, that the number of directives does not de- termine the number of parts into which a polyp shall divide; secondly, that in division there may be separated a piece which 176 DONALD WALTON DAVIS before regeneration possessed no directive mesenteries, and, thirdly, that the division plane cuts the directive plane. We may now consider more specifically how nearly per- pendicular to the major axis of the mouth is the plane of division. In other words, is there any tendency toward bilateral symmetry of the structures of the old piece with respect to its directive plane? The simplest cases in which this problem may be studied are those of originally diglyphic specimens with sym- metrically placed non-directive complete mesenteries. Eight specimens of this nature are recorded in table 3. Three of them divided in spaces’ other than complete endocoels. Of two regular hexameric individuals in this class, one (no. 4) divided in two incomplete endocoels forming two pieces, each sym- metrical with respect to its directive plane; the other (no. 6) divided in one complete and one incomplete endocoel forming pieces as nearly symmetrical as could result from division in such spaces. A third specimen (no. 1), with eight pairs of complete non-directives, divided likewise through one complete and one incomplete endocoel into two nearly symmetrical parts. Five originally symmetrical polyps divided in two complete endocoels—assuming that in two cases a bounding mesentery recorded as doubtfully incomplete was really complete at the time of division. Four of these (nos. 7, 11, 13, and 14) were regular hexameric specimens and one (no. 9) was regularly octameric before division. Each of these five divided into two symmetrical pieces. Neglecting the possibility of division in directive endocoels, the chances are even that a regular hex- americ individual dividing in complete non-directive endocoels will produce two symmetrical or two asymmetrical pieces. The number of symmetrical and asymmetrical pieces produced by such divisions should be approximately equal. Actually four divisions of this sort gave eight symmetrical pieces. The chances in such a division of a regularly octameric individual are two to one in favor of producing two asymmetrical pieces as against two symmetrical pieces. One such division gave two symmetrical parts. As far as these few cases have any signifi- ASEXUAL REPRODUCTION IN SAGARTIA 177 cance, they point to a tendency on the part of symmetrical individuals to divide into symmetrical parts by a vertical plane perpendicular to the directive plane. Since a precisely symmetrical old piece has like mesenteries adjacent to its edges, whereas an asymmetrical piece may have like or unlike bounding mesenteries, any tendency toward the formation in division of accurately symmetrical pieces would be indicated by an excessive number of like bounding mesenteries. This gives us an opportunity to discover the limits of the tendency. The numbers of like and unlike bounding mesen- teries of tables 3 to 6 are summarized in available form on the left of table9. We may best disregard the less common classes and consider only the twenty-eight old pieces with two com- plete non-directive bounding mesenteries, thirty-two cases of one complete and one incomplete mesentery, and nine specimens with two incomplete bounding mesenteries. If we arbitrarily distribute into two equal groups the thirty-two cases of pieces with unlike bounding mesenteries, we have data suitable for the use of Yule’s ‘Coefficient of Association’ (’00, p. 272) asa measure of a possible tendency of the division plane to pass through similar spaces on opposite sides of the directive plane. The distribution of the divisions in these cases is, then, as given in table 1. Complete positive association, the invariable association of two incomplete or two complete mesenteries on the two boundaries of a piece, would be represented by a value for Q of + 1. Com- plete negative association, the constant association of a complete mesentery on one boundary with an incomplete mesentery on the other boundary of a single piece, would be indicated by a value for Q of — 1. Merely chance association would be indi- cated by a coefficient of 0. The calculated coefficient with its probable error is — 0.008 + 0.02. The coefficient, being less than its probable error, is practically zero. Our figures, there- fore, give no indication of a departure from the chance associa- tion of two incomplete mesenteries, an incomplete and a complete one, or two complete mesenteries as old bounding mesenteries lying in one piece adjacent to a single division plane. 178 DONALD WALTON DAVIS TABLE 1 Complete Incomplete Complete a= 28 b= 16 Incomplete c= 16 d=9 The coefficient of association by the formula, ad — =- sie at is — 0.008 a The probable error of Q by the formula, 1— Q? : DS Weasel Boe : |e De an Sea is = 0.02 This implies that the demonstrated tendency of {the plane of fission to lie perpendicular to the major transverse axis of the mouth (p. 177) is not so strict as to regard the difference be- tween two adjacent endocoels, the one complete, the fother incomplete. The same facts as were summarized in the left part of table 9 have been represented in the corresponding part of table 10, counting separately each bounding mesentery. The bounding mesenteries indicated by symbols in the column headed B in tables 3 to 6 are represented in table 10 in the column headed ‘Frequency’ and in the appropriate rows. Thus each unit in the frequency column of table 10 indicates a position of a division plane on one side of one individual resulting from a division, and the total number of units in the column of frequencies is equal to the number of symbols in all the columns of bounding mesen- teries in tables 3 to 6. In reckoning the total number of divisions in complete and in incomplete endocoels I have sub- tracted the numbers of mesenteries whose apparent complete or incomplete character is in doubt. I have, of course, included ASEXUAL REPRODUCTION IN SAGARTIA 179 in the total number of incomplete endocoels those mesenteries undoubtedly incomplete but of undetermined grade. The per- centages have been figured on the revised totals. The table shows division planes as follows: complete endocoels, 62 per cent; incomplete endocoels, 34 per cent; exocoels, 4 per cent. The significance of these figures, in showing tendencies of the division plane to pass in spaces of a certain character, is af- fected by the relative numbers of spaces of each character in the animals at the time of division. Possibly the relative width of the spaces should also be taken into account. At an early stage in the formation of a cycle of incomplete mesenteries the adjacent exocoels are so reduced that the sum of the widths of endocoels considerably exceeds the sum of the exocoels. T his condition gives way to a state of practical equality between endocoels and exocoels so soon that the temporary inequality probably has little if any influence on the relative frequency of passage of the plane of fission in different spaces. Certainly, the inequality between endocoels and exocoels is not sufficient to account for more than a small fraction of the excess of endocoelic divisions recorded. The number of exocoels in an anemone is precisely equal to the sum of all endocoels. According to chance, neglecting in- equalities in size, 50 per cent of the divisions should occur in exocoels. As shown in table 10, division planes occur in exocoels in only 4 per cent of cases. It is clear, therefore, that there is a strong tendency of the division plane to pass through endocoels. The proportions here found do not agree with those given by Torrey and Mery for S. davisi. Their results, stated by the method I have used, give twenty-nine cases of division planes passing through exocoels to seventy-three passing through en- docoels. These results, as Professor Torrey informs me, were obtained from hand sections of the anemones. Some incomplete mesenteries might be overlooked by that method. This might be responsible for the recording of a somewhat greater proportion of divisions in exocoels than I have found from the study of microtome sections, but it seems scarcely possible that it can ac- count for the whole discrepancy. Even if the divisions in incom- 180 DONALD WALTON DAVIS plete endocoels below the first grade are added to the exocoels as shown in my table 10, these form only 12 per cent of all the divisions, whereas Torrey and Mery found 28 per cent of all the positions of division planes to be in exocoels. Apparently, S. davisi and 8. luciae differ in the proportion of divisions in exocoels. As between incomplete and complete endocoels the chances cannot be accurately stated, since the cycles of incomplete mesen- teries are variable in number. Probably division occurs very rarely if at all in a region where the first cycle of incomplete mesenteries is not yet established. The members of this cycle are equal in number to the complete mesenteries. Usually in- complete mesenteries of the second grade are also present. These are twice as numerous as those of the first cycle. Later cycles are sometimes present, at least in the older parts of a specimen. We may say, therefore, that probably not less than 50 per cent nor greatly more than 75 per cent of endocoels are incomplete; and that while, according to chance, divisions should occur in incomplete endocoels in at least as many cases as in complete endocoels, the actual numbers found are 57 incomplete to 103 complete endocoels, or 36 per cent of divisions in incom- plete endocoels. There is, then, a marked tendency to divide in complete endocoels. Since in regeneration the full number of complete mesenteries to be formed appears very promptly and the directive or non- directive characteristics are obvious at an early stage, the chances of a later division plane passing in directive or non- directive endocoels may be quite accurately determined. The old parts of animals represented in tables 3 to 6 show 572 com- plete mesenteries. Of these, 154 are directives. Therefore approximately 27 per cent of the complete endocoels of these animals are directive endocoels. Of 103 divisions in complete endocoels (as calculated for table 10), six, or approximately 6 per cent, are in directive endocoels. That is, of the divisions in complete endocoels there were less than one-fourth as many divisions in directive endocoels as chance would demand. This is only another expression of the fact, already demonstrated ASEXUAL REPRODUCTION IN SAGARTIA 181 (pp. 176, 177), that the plane of fission is more or less nearly per- pendicular to the directive plane. That directive endocoels are avoided, gives additional emphasis to the tendency to divide in complete non-directive endocoels. The inequality of the parts resulting from a division is strik- ing. The relative size of the parts may be roughly measured by the number of complete mesenteries. Of fourteen divisions of diglyphic specimens into two parts each, no case of equal distribution of the complete mesenteries between the two parts is found, whereas cases of extreme inequality do occur. The other specimens show in general a similar lack of equality of parts. One might explain the inequality in the products of division of a regular hexameric individual on the ground that the tendency to divide in complete endocoels and the tendency to divide in a plane perpendicular to the major axis of the mouth overcome any tendency to divide into equal parts. But this does not hold for the unequal division of a regular octameric specimen such as no. 9 (table 3). In this specimen the tendencies for the fission plane to pass through complete endocoels and to pass per- pendicular to the major mouth axis would both be satisfied by a division which should produce equal parts. Yet this division was very unequal. In a number of other cases the inequality of parts is much greater than is demanded by the tendencies to divide in complete endocoels and along the favored plane. It should be noticed that, in every case where a recent regeneration zone is indicated in the old part (tables 3 to 6), the piece con- taining this regenerating part possesses the greater number of complete mesenteries. It is possible that there is a tendency toward an equality in which different regions have values dependent upon age. The tendencies of the planes of fission so far as analyzed are summarized on pages 226, 227. It-remains to be pointed out that, even among specimens having the same number of siphonoglyphs and complete mesenteries, the im- mediate products of division are quite diverse in form. The same processes, dealing with specimens already varying as shown in tables 3 to 7, lay a foundation for the most extreme numbers of mesenteries and siphonoglyphs found in the fully 182 DONALD WALTON DAVIS regenerated individuals. Regenerative processes may begin with pieces varying in number of siphonoglyphs from none to three, and in number of complete mesenteries from zero to fifteen. COURSE OF REGENERATION The course of regeneration as seen in the living animal has been described above (p. 167 ff). The account which follows is derived from the study of sections. It should be noted that the white lines and orange stripes fade out immediately upon killing by any of the methods I have used. ‘The external distinctions between new and old tissue are also lost in the process of fixation by some methods, but in formalin the newer area remains dis- tinguishable by its lighter color, at least in those cases where the color differences in the living specimens are great. There is, nevertheless, little uncertainty as to the relation of external and internal features. After examining a series of stages in regeneration the boundaries between new and old material can be quite as accurately determined in sections as in living indi- viduals, and the external markings and internal structures can in almost all cases be satisfactorily correlated. In sections of specimens killed a few days after division the column wall in the region of fusion appears much like the old column wall. In some specimens this region is distinguishable on account of its being slightly thinner and having the outer border of the ectoderm less scalloped. This condition occa- sionally extends slightly into a part of the column which, as indicated by the mesenteries, is old. In the greater number of cases, practically no evidence of the region of fusion is found in the column wall a few days after division. In the esophagus evidence of the division persists much longer, especially at its aboral extremity. Closure of the esophagus orally occurs within four days. At this time its torn edges are contracted, the esophagus being much shorter on that side-than on radii distant from the plane of division. Its full length on the regenerating side seems to be reached only as the new complete mesenteries become ASEXUAL REPRODUCTION IN SAGARTIA 183 attached to it. Because of this contraction and the conse- quent oblique position of the aboral aperture of the esophagus, the normal edge is often cut in transverse sections of re- generating individuals. The torn lateral edges of the esophagus before their union appear precisely like the normal free edge, and no clear boundary between these edges can be distinguished. Sections therefore show the esophagus open laterally, and it is impossibe to determine, in some cases, whether or not fusion has been completed. Certainly, the union proceeds down the column somewhat in advance of attachment to it of the develop- ing complete mesenteries. Formation of new siphonoglyphs Following closure of the esophagus, a siphonoglyph appears along the line of fusion of its edges. A siphonoglyph differs from the remainder of the esophagus most obviously in its evenly rounded outline, the other portions showing in sections scallops with their convexities toward the cavity of the tube. A fully developed siphonoglyph shows also, in sections well preserved and stained, more abundant cilia than are to be seen in other regions and usually a more prominent heavy line marking the basal bodies of the cilia In consequence of these distinctive features, siphonoglyphs may usually be readily distinguished from other regions of the esophagus. The time at which differ- entiation of the siphonoglyph becomes clearly evident varies greatly, but is usually about ten days after separation. The undifferentiated esophagus continues to grow laterally on both sides of the new siphonoglyph, leaving the latter permanently in the middle of the regenerated region. When division has occurred in the plane %f a siphonoglyph, this structure persists and an additional siphonoglyph is formed. Three clear instances of this are recorded as nos. 21a, 21b and 44, thefsiphonoglyphs being indicated in tables 5 and 6 by the pairs of directive mesenteries with which they are invariably associated. 184 DONALD WALTON DAVIS Recalling the types of division described above (p. 174 ff.), it is now clear how specimens with various numbers of siphonoglyphs may arise. Diglyphic specimens most frequently give rise to diglyphic forms, but may produce triglyphic (nos. 16b, 21a, 21b) or monoglyphic (nos. 16a, 21¢) ones. Triglyphic individuals may produce diglyphic (no. 22a), monoglyphic, triglyphic, or tetraglyphic specimens. A tetraglyphic form (no. 15) did give rise to one individual with four siphonoglyphs, and one with two. Other possibilities are at once evident. It is wholly prob- able that any form may give rise, by one or two divisions, to any form. Since a new siphonoglyph is invariably formed in the new region at an early stage of regeneration, it is obviously more appropriate to refer specimens in process of regeneration to their ultimate class rather than to the one to which they appear at the moment to belong. It is therefore no longer admissible to say, as Davenport (’038, p. 141, line 20) has heretofore done, referring to the usual type of division of a diglyphic individual, that “‘a monoglyphic 8. luciae is the result of longitudinal division of a diglyphic form,” or to refer to the ‘‘metamorphosis of a monoglyphic to a diglyphic type’”’ (ibid., line 33). The true monoglyphic type is a fixed type which does not develop a new siphonoglyph, thus transforming into a diglyphic type, although it may produce one or even two diglyphic individuals by a single division. The division of a diglyphic individual in the great majority of cases results in two diglyphic forms, never two truly monoglyphie ones. Development of new mesenteries By the time the siphonoglyph is clearly differentiated, all of the complete mesénteries to be formed may have appeared, but they do not reach the esophagus until somewhat later. New mesenteries show first in the lower half of the column. The earliest evidence of a mesentery is a very thin sheet of mesogloea extending from the mesogloea of the column wall inward into the endoderm. When the sheet has pushed nearly through the endoderm, the latter becomes projected inward in advance of it. ASEXUAL REPRODUCTION IN SAGARTIA 185 The new mesentery is then similar to the old incomplete mesen- teries of the latest cycle except that the regenerated ones are usually thinner and have a much narrower and more lightly staining plate of mesogloea (for instance, figs. 7 and 8). They rapidly elongate vertically, extending downward to the base and upward to and across the oral disc; but the increase of a mesentery in thickness and in radial extent is slower until it has become complete by crossing the oral disc to the esophagus. It then rapidly extends centripetally, apparently as a result of tension, since a mesentery that has recently become complete is often thinner than one of the same age that has not quite reached the esophagus. At about the time a mesentery becomes com- plete in the extreme oral region, its longitudinal muscles become clearly evident (figs. 9 and 10). Subsequently these mesen- teries grow thicker and their longitudinal muscles become more fully developed (fig. 13). They become attached to the esoph- agus farther and farther aborally. For a long period, however, at least some of the new complete mesenteries fail to reach the esophagus at lower levels where the old ones are attached (figs. 15 ane DSistc2-e): The incomplete mesenteries arise in similar fashion, but they develop more slowly and retain much longer the slighter degree of development which distinguishes them from the incomplete mesenteries of the old region. Number and arrangement of mesenteries in regenerated regions Upon examination_of the mesenteric formulas of the new regions of the specimens represented in tables 3 to 6, it is apparent that the variations in number and arrangement of the new mesenteries are determined almost wholly by the mesenteries on the boundaries of the old regions. Specimens listed in tables 7 and 8 give additional data upon which to base a study of regeneration formulas. Some of these were selected, in col- lecting, upon a basis which would interfere with their usefulness in considering the relative frequency of fission planes in different regions. Others were excluded from tables 3 to 6 on the ground TH® JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 186 DONALD WALTON DAVIS that, had the natural division occurred in some other plane, adequate evidence of the position of the plane of fission would not have persisted to the time of killing. Still others are re- generating individuals resulting from artificial cuts. The last mentioned are pointed out in the explanations of tables 7 and 8. It will be seen that they regenerate in the same general manner as if they had divided naturally in a corresponding plane, but ap- parently show wider variations in the number of new mesenteries. Table 9 shows the frequency of different numbers of new com- plete mesenteries, based upon the specimens represented in tables 3 to 7. I include here as complete mesenteries not only those actually reaching the esophagus, but also mesenteries now incomplete but destined to become complete. These are, in nearly all cases, readily distinguishable from true incomplete mesenteries whatever the stage of regeneration. For incomplete mesenteries appear only after those destined to become complete have reached relatively large size, and they develop more slowly than the complete ones. Furthermore, the members of a pair of “new incomplete mesenteries appear simultaneously and remain of equal size during their growth; while members of a pair of complete mesenteries reach such a state of equality only late in regeneration. Since all of the complete mesenteries destined to be formed in a regenerating region appear quite promptly, their number may be safely judged almost from the beginning. These features are illustrated in figures 23 to 30. The numbers of new mesenteries actually complete is indicated in the tables of mesenteric formulas in connection with each regenerating zone. Where complete old non-directive mesenteries occupied both boundaries, regeneration involved the production of eight com- plete mesenteries in twenty-five cases and six complete mesen- teries in two eases (table 9). The character and arrangement of the eight mesenteries commonly formed may be made clear with the aid of figure 15, representing specimen no. 64 of table 7. The older of the two regenerating sectors, occupying about two-fifths of the area of the cross-section (lower left), is limited by two old complete bounding mesenteries c, c. On either side of this regenerating region, adjacent to the old bounding mesentery, is ASEXUAL REPRODUCTION IN SAGARTIA 187 a new mesentery, c’, forming with the bounding mesentery a non-directive pair. In the middle of the new area is a pair of directive mesenteries, d, d. Between the directive and the new bounding mesenteries, on either side, is a pair of non-directives, cl, e, Figure 16, representing a more aboral section, shows the first cycle of incomplete mesenteries appearing in the primary exocoels. Other cycles of incomplete mesenteries are formed later. Where, under thege conditions (as in specimen no. 26, table 6), six mesenteries are regenerated, only one complete non-directive is formed between the new directives and the old bounding mesentery, and this one mates with the latter. I have no specimen with the reduced number of complete mesenteries on both sides of the new directives, but no doubt such instances may be found. Only four new complete mesenteries would then be developed. When division has occurred in two incomplete endocoels leaving an incomplete mesentery adjacent to each boundary of the old region, ten complete mesenteries are usually regenerated. This is the case in the later regenerating zone of the specimen (no. 81, table 7) shown in figures 15 and 16 (upper, right). In the middle of the new area is a pair of directives, d, d. On either side of this pair of directives are found two pairs of non- directives, c!, c2 and c?, ct. Permanently incomplete new bound- ing mesenteries, (1), (J) (fig. 16), are formed paring with the old incomplete bounding mesenteries. These new incomplete mesenteries appear later than the complete mesenteries, and in many specimens represented in the tables, they were lacking at the time of killing. In later stages of regeneration they are invariably found. They are frequently present in the aboral region while lacking at levels nearer the oral disc. This is the case in the specimen represented by figures 15 and 16. Other incomplete mesenteries develop in pairs in the exocoels of the new region. In three cases where both of the old bounding mesenteries are incomplete (nos. 4a, 4b of table 3 and no. 79 of table 7) only eight complete non-directive mesenteries were regenerated, and in one case (no. 76, table 7), but six. In these specimens only one pair of complete non-directive mesen- 188 DONALD WALTON DAVIS teries was formed lateral to the directive pair on one or both sides. When one of the bounding regions is occupied by a com- plete and the other by an incomplete mesentery, nine, seven, or five new complete mesenteries are formed. When nine. are regenerated (as in no. 19a, table 5 and fig. 27), the new mesenteries consist of one pair of directives, a pair of non- directives on either side of these, a second pair of non-directives on the side toward the incomplete mesentery, and an additional complete mesentery pairing with the latter on the side toward the complete bounding mesentery. When but seven complete mesenteries are formed after division in one complete and one incomplete endocoel, the arrangement is the same as when nine are produced except that on the side of the incomplete mesentery two fewer complete non-directives are found. Sections, at two different levels, of a specimen (no. 66, table 6) regenerating in this way are shown in figures 12 and 13. Other specimens are represented in figures 11 and 19. In the case recorded as show- ing only five new mesenteries, two fewer are formed on each side. Where a directive mesentery lies on the boundary, the new bounding mesentery is a directive mating with the old one, and two pairs of non-directives are formed on that side of the wholly. new directive pair (for instance, nos. 21a and 21b, table 4). In one case (no. 10b, table 2) the new bounding mesentery adjacent to the old directive was not, at the time of killing, developed to a stage where it could be distinguished from an incomplete bounding mesentery. On a side where an exocoel is involved in the division (for instance, nos. 54 and 84, table 5), regeneration is precisely the same as in a case where an incomplete mesentery occupies the boundary, except that no new bounding mesentery is formed. In none of the cases I have observed are there less than two pairs of non-directives on such a side. Two peculiar and in- teresting cases (nos. 18b and 22b, table 5) possibly illustrate the influence of the old bounding mesenteries on the regeneration, and suggests its limits. No. 22 was a specimen regenerating ASEXUAL REPRODUCTION IN SAGARTIA 189 from an experimental cut. Mesenteries of this regeneration are represented in table 5 by Roman type. The specimen divided into two parts, a and be, in incomplete endocoels of the first grade, but with the loss of one of the bounding mesenteries in a. Eighteen days after this division, fission again occurred in be, producing moieties b and ec. On one side (the left in the table) this new fission plane passed again adjacent to the same old incomplete mesentery of first grade, which in this division passed to c. Meanwhile the normal number of mesenteries had been established, including the mate to this old bounding mesentery, now represented on the far right of the formula of b. The case of no. 18 is somewhat different. Here the division into a and be was followed after but two days by the separation of be into b andec. This occurred on one side along the boundary between old tissue and that just beginning to regenerate. The early state of this regeneration at the time of the second division may have had something to do with the irregularities indicated by the queries as to original completeness or incompleteness of two of the mesenteries. Certainly, in the two days that elapsed between the divisions, few mesenteries could have been estab- lished in the earlier regenerating zone, and none would give any indication of their final state of completeness if this had been determined. It appears that, while a regenerated normally, pro- ducing a mate to its incomplete bounding mesentery, the corre- sponding regenerating zone of b shows no corresponding incom- plete bounding mesentery (which would be represented on the far right of the upper row of the formula for no. 18b). Rather, the two regenerating zones have matched up together, with two complete mesenteries from either side of the boundary forming a new pair. It must be noted that the uncertainties of identi- fication in this case are such that great dependence must not be placed on the interpretation here suggested. Such a juxtapo- sition of two regenerating areas may account for the large number of mesenteries (including two pairs of directives) recorded in an apparently single regenerating region in specimens nos. 88 and 89 of table 8. 190 DONALD WALTON DAVIS From these statements it appears that the chief variations in number and arrangement of mesenteries on each side of a re- generating area are wholly dependent upon the old bounding mesentery on that side, except for two complete non-directives, which may or may not be present whatever the character of the bounding mesentery. Five specimens showing exceptional features in their regenera- tion are represented in table 8. Two of these specimens show an abnormally large number of new mesenteries,—one (no. 88) a second pair of directives, the other (no. 89) a second pair of directives and two additional pairs of non-directives. A third (no. 87) shows a new region which is normal except for the weak stage of development of the muscles in the pair of mesenteries in the position ordinarily occupied by the directives. These mesenteries extend to the esophagus, but their longitudinal muscles are so weakly developed that the directive or non-direc- tive character of the pair cannot be determined. ‘The fourth (no. 86) shows this pair of mesenteries extremely reduced. They do not reach the esophagus and show no evidence of longitudinal muscles. The fifth (no. 85) shows a reduction from the usual number of mesenteries, there being only one new bounding directive and two pairs of non-directives. An old regeneration zone in no. 40 (table 5) shows an excess of two mesenteries over the usual number. Neglecting the few exceptions (six in all) just referred to, we may say that in regeneration a new pair of directives is formed, approximately in the center of the new zone, and that on each side of this pair of mesenteries are formed one or two pairs of complete mesenteries plus an odd mesentery if required to mate with a bounding mesentery on the edge of the old part. Treating separately each lateral half of a new zone, the tendency in full regeneration is to produce on each side of the middle plane a directive and three non-directive mesenteries; if a complete old bounding mesentery is not present, additional mesenteries are formed to mate with the odd new non-directive and with any old incomplete bounding mesentery. ASEXUAL REPRODUCTION IN SAGARTIA 191 TABLE 2 SS a Sa REGENERATED MES- OLD BOUNDING MES- FORMULA ENTERIES ENTERIES N sh Bin co oS SETS RAE ae aT se eee the (HOR (e c Peas. hee hy. SON. po, REG d, c c Cs UR ee cece eens d, cle?, c3ct, (1) (1) No. 2 oe (1) TREE at nn tea ara ee GCC" INOS SH eee kereettesa me. gee rabies hep Baa? d, cic?, c3c4, d d ISIC Loan oo ete Ree ane Moen ake onc al (re Cie 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 c3. 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 3 eases. On the side of old incomplete bounding mesen- teries of the first grade the number of instances of the reduced number (8 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 mesenteries 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 S. 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, no 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. © 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 ct. The next to appear is the one designated c?. It is destined to pair with cl; c’, the mate of c!, appears very slightly after c?, or even simul- taneously with it. An incomplete mesentery, (1), 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 and 6. 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 (7) 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 (I7) 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:nesedsct, efve, Dk 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 ee 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: INOnta. GO. CC No. Ib c,d IOs 2a 6s OE, CGS, 1CL) No Zp ed: €..(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 1s 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 (’04 and ’09) and Cary (711). 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 8. 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 8. 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 p. 171), 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. [X, 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 S. 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: (onck. chvdartiwe (I). cc, d, c. CRED Wek rede etre ae: Carlgren’s statements indicate that in (1) 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, LX, 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 Il. 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-directive 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 S. 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. Ina 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 8. 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 §. 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 eom- 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, 1.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 S. 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 McMurrich (’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 MeMur- 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, 143). 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. 143) that the presence of the stripes may ‘‘be considered as a case of warning coloration?” Orange stripes in fission and 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 Zig 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 Die 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 imcomplete 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 ncomplete mesenteries of apparently first grade was shown to be associated with an extra orange stripe. By far the greatest ASEXUAL REPRODUCTION IN SAGARTIA 215 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—aunless 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 ate 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 ZW 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 exceeds 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 Stphonoglyphs 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- glyphic 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 P74, 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 endocoel 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 223 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 cases 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 Jarge 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 S. luciae or M. marginatum have developed their characteristic colors. Whatever its proper classification, this anemone was symmetrical and diglyphie 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 5S. 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 completemesen- teries and fewer than sixteen orange stripes. It appears from the statements of Davenport (’03, p. 143 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 Dav 1port or myself is twenty- two. It seems likely that Dave.iport 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 disc 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 monoglyphic, 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). Succeeding 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 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 (p. 177). 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 Zak 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. 1938, 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 incomplete 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 (p. 216): 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, 213, 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. College of William and Mary, August, 1917. ASEXUAL REPRODUCTION IN SAGARTIA 231 BIBLIOGRAPHY AnpreEs, 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. CARLGREN, OsKAR 1893 Studien iiber nordische Actinien I. Kongl. Svenska Vet. Akad. Handl., no. 25, 10, pp. 1-148. 1904 Studien tiber 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. II. 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, Gertrupe Crorry 1903 Variation in the number of stripes on the sea-anemone, Sagartia luciae. Mark Anniv. Vol., pp. 137-146. DicquemareE, 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. Gossr, Puture Henry 1856 Tenby: A seaside holiday. London, John Van Voorst. 1860 Actinologica Britannica. London. John Van Voorst. Gossz, Puiniep Henry 1865 A year at the shore. London. Alexander Strahan. Hammarr, M. L. 1906 Reproduction of Metridium marginatum by fragmental fission. Amer. Nat. 40, pp. 583-591. Harcirr, 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, ALFrrep GoLtpsporoucH 1905 Seashore life. New York Aquarium Nature Series, I. New York Zool. Soc. McCrapy, Joun 1858 Instance of incomplete longitudinal fission in Actinia cavernosa Bose. Proc. Elliott Soc. Nat. Hist. of Charleston, 8S. C., 1, pp. 275-278. McMovreicu, 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. 259-278, 1 pl. 1902 Notes on the dispersal of Sagartia luciae Verrill. Amer. Nat., 36, pp. 491-493. Tuynne, 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 DONALD WALTON DAVIS Torrey, Harry Beat, AND Mmry, JANET RurH 1904 Regeneration and non- sexual reproduction in Sagartia davisi. Univ. Cal. Pub., Zoél. 1, pp. 211-226. VeRRILL, A. E. 1869 Oursea-anemones. Amer. Nat., 2, pp. 251-262. Yuue, 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!, e?, ce’, 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. (1), 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, ete. (tables 3, 5, 6), number of small mesenteries in a acre rede 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 in detail, except that definitively incomplete mesenteries are represented in the formulas only when they occur as bounding mesenteries. In cases 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. For explanation of symbols, see page 232; For a discussion of certain irregu- larities shown in table 3 to 5, see pages 171 and 172. OLD MESENTERIES NEW MESENTERIES here ce NO. BEG ‘ B wp) |epel 32p | es Seelneph pel xen) Be |ClDayel 3 in plete ; 1/2 ¢ cc ddalccesccs|# (ib) four 0 | 20 os Cc eesce \ddi|ice;ce) 1G) six One208 67 ake (ai ec\cemitddh|kccsces| dy) four On eS9 7 oss) (0) cc ddiiiccycc|) (a) four 0 | 14 3 ja. Gi) |jcesee |.ddlicesce' |) twelve very small OF LOS ealtbaee MDa cc.ccmiadd c four very small o | 10 42 (II) dd GOV @AN| ees hele eevee (U0) || 33-1) ul see CL) icckce® kd dkiceyccule Ch) min) eee" | ddaltce cosa (Gs) 3 | 51 5 )2.. (Onitcesccxcclrddaliccsces|yac oO \Ce. nol! || Ge 6 | 50 .< Sagartia- ieee aos renee 26| 48} 15) 7 |2+-1? 10 Heneanis Metridium, natural...... 1 1s ep 2 Metridium, artificial.... 319 4 | Aiptasia 5 ses eae Ec be pcedse aD 1 21 P Sagartia..2; ener eee Bale Lah et? 1 2 Baer pieces Metridiumagaa. eae 2 | 2 1 1 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. Srey lie | Plea es hae LOT DATE LOCALITY ae eh Nese ah 2 G9] & |aPlR oO! & = A |e & a Pe t900, July. 12s aye) 2 Fame lala ate ok oor 2 alan lees Zee So0o duly 18.52... 3! U. oO. De be Whatis.cs.... +s. cue Ole 44a OU nas 3 | 1909, September 22..) U.S. B. F. Wharf.............. 16} 69| 11] 3] 99 Ae LOLOMNovemben2oa.|) Lenzancemsnrre seer ones oe 15} 182] 10] 0} 207 OWL OMe March, Sin: Sv CenZancee same see. coca «2. 19] 134) 7| 0} 160 On) | LOM eMiay 80's. case Gut of Cancer, smooth stones..| 13] 118} 4} 0} 135 Co UIE IM ER CDE wk oie Pine Tree Island, smoothstones.| 5} 94} 8} 0} 107 8 -t9l te gune 2950523. 9. Gut of Cancer, smooth stones..} 14| 129] 2] 0| 145 Se LO se un C2 Oars te Gut of Cancer, rough stones...} 28] 254) 8} 1] 291 LOUALS ere ema US ss 5 re Mere ole PA b's Leslee 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 ni hba le |e lb Lie laa [GAllble SL LC SDL GIG GRIT Le: 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} 0} 1 3 1} 0} 2) 2) 2) 5) 1) 7} 6] 4) 7| 8/13) 4/10} 5] 9) 1 2} 0} 3 OUR ach satay Se. 1} 6} 8/11) 5)15) 9)30/18)22/16/16)26} 9/15) 6]11] 3} 2] 3} O} 4 xX 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, c1,c!. X 60. ASEXUAL REPRODUCTION IN SAGARTIA DONALD WALTON DAVIS PLATE 1 245 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 (1) to (I) 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. X 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. In the upper part of the figure a fold of the oral dise 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. 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 ct of a region regenerating the typical number of mesenteries (p. 191, L197)... X94. 246 ATE 2 PL AGARTIA g ASEXUAL REPRODUCTION 1} DONALD WALTON DAVIS el Se (ay = Seem ABR 4 i yah WEEIG ae 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, (1). 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). X 45. 14 Section of a triglyphie specimen (no. 63, table 6). Externally there were seen, not quite opposite each other on the oral disc, 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. ASEXUAL REPRODUCTION IN SAGARTIA PLATE 3 DONALD WALTON DAVIS ») 19 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, Vor. 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 ¢ 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 (I) 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 ASEXUAL REPRODUCTION IN SAGARTIA Tp DONALD WALTON DAVIS PLATE 4 251 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 c3 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. X 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 PLATE 5 DONALD WALTON DAVIS ASEXUAL REPRODUCTION IN SAGARTIA PLATE 6 EXPLANATION OF FIGURES 23 and 24 Sections of a specimen which showed no certain external evidences 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. 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 cinelis in an incomplete endocoel of the second grade ata. A pair of nesenteries of the fourth incomplete cycle is present near the directives below. & 8. 26 Section of a perfectly regular diglyphie specimen with two white bars opposite each other on the oral dise and twelve orange stripes symmetrically placed and evenly spaced on the column. Internally no evidence of division. Pairs of directives are indicated by d. & 14. 254 ASEXUAL REPRODUCTION IN SAGARTIA PLATE 6 DONALD WALTON DAVIS 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 37 days. For a full account of this case see page 173. X23: 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 cycle and one of third eyele. 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. ASEXUAL*REPRODUCTION IN SAGARTIA PLATE 7 DONALD WALTON DAVIS PLATE 8 EXPLANATION OF FIGURES 31 and 32 Respectively more oral and more aboral sections of no. 19¢. 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. 43. 258 PLATE 8 ASEXUAL REPRODUCTION IN SAGARTIA DONALD WALTON DAVIS 9 3 9 « PLATE 9 EXPLANATION OF FIGURES 35 Diagrams I-VI, VIII, and X have been modified from those of Carlgren (09, p. 39, fig. I]). Diagram VII has been constructed from his description (09, p. 35) of a specimen of Sagartia viduata (no. 15a7). Diagram IX 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, IT 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 N SAGART 8 c} C2 ae OS mM. sce. o-% acee pote sd clusnis astysie sae =e 285 PRU UBER ACIRUSTSLDRNC TEN ert Cae (hig es halon ec) Gm Less Sivts die Vale Sa 286 A three-point back cross, black purple vestigial with balanced inviability... 288 AC OPMCLU EME a eee th ee). ALL ROC. Od. ee A. Re EE. 291 The relation between coincidence and map distance..................-.000- 292 The use of purple in mapping other genes, curved, streak, etc.............. 292 ALCOR BLOGs NaCle GY ORBEA ae isi. daria Gere sess ns Bokeh dhe onl oe s.4 siciakle acldlea mare 295 A summary of the linkage data involving purple....................0.e00e- 296 Special problems involving purple—age variations, coincidence, temperature variations, crossover mutations, progeny test for crossing over.......... 297 Sanmnanyiand valuation. <./4.3<¢,25> bis ce bee eF de} Bea teeta ark. 302 PUL UEP UMONCLLOG «sini uh ain pened & Aion <0 eal ieee cs eae) elie ee le 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 ight 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 2 32, # 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; B1) 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 outecrossed 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 Bit 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 = a an) i Srarilion) | iscveemition) | Veraiion) | « (Fermalion BOSE fae fs ANd ya om ays 90 186 0 0 71 202 0 0 13 0 as ee 72 197 0 0 72 206 0 0 BP 2 ers settee eters 45 126 0 0 65 195 0 0 51 88 ff 3 BU poe tiaee../28.. 2 alae ais. 98 178 27 2 43 72 4 0 Boe meme eent. vakigs:s 54 191 0 0 37 70 0 0 Mopaltyss ess sete 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 Fi (vermilion) sons and daughters from the outcross of (vermilion) purple vestigial males to vermilion females (VERMILION) = : 1912, suNE 17 (VERMILION) PURPLE CMON) en) SSP Men yy aches ss Petes eat 200 23 9 5 188) 5 7a AONE IEEE SAAS. Reo 88 21 3 4 BBE OUTS. BAI TROL 255 66 25 5 Ont Re ek ccc bie tt cyte ets as 368 19 3 t 1B PA eee Pols ais eee eens oe 346 17 19 9 272 CALVIN B. BRIDGES the results of the first were fully known. A purple vestigial male outcrossed to a wild female produced wild-type sons and daughters (page B39; + @ 15, + o@ 10). Four of the Fi females were back crossed each by two or three purple vestigial males from stock. In this case Fi 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, suLyY 16 Pt io A ee Purple vestigial) Wild type Purple Vestigial BoGeeeaa 3 eee rntte se otebae 82 163 12 15 BS6S2 ers se Soense a etatenebens 80 6%} 14 10 BBO lege ties Recs are voce tosh 32 53 3 rl B3022 sakes eo ocorsborernant an: 62 141 9 9 ‘Total. 2 eee aries 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 Dike 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 vestigial 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 iA) ably = EP SAE 35 Seven of the eight daughters tested had this expected compo- sition, but one (no. 3464) gave only offspring corresponding to = = -. = 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. A significant feature of this case is that while the change must be described superficially as double crossing over, daughters should all have the composition the composition > The symbols above the line represent the genes in the chromosome derived from one of the two parents, those below from the other. The + signs represent’ simply the wild-type allelomorphs of the mutants dealt with, and these signs may usually be omitted. GENETICS PURPLE EYE COLOR DROSOPHILA Zt this double crossing over occurred within a region only ten units long—a space shorter than that in which double crossing over of the ordinary type has ever been detected even in certain re- gions of the autosomes in which double crossing over is relatively most frequent. MUTATIONS Two new mutations were found and two old ones reoccurred in these back-cross experiments on the linkage of purple and vestigial. ‘Kidney’ eye shape, a third chromosome recessive, was found in B. C. culture B10.2, June 26, 1912 (table 1). This mutant, the first affecting the shape or texture of the eye, was con- siderably used in the early days (Morgan, ’14, and Bridges, 15), but has now been superseded by mutants less variable and easier to classify. In culture B39.2 it was noticed, July 26, 1912, that several of the wild-type flies had more dorsocentral bristles on the thorax than the regular number, four. Later it was found that such extra-bristled flies were occurring in small proportions in all four sister cultures, from which it would appear that the muta- tion was a recessive, introduced through the purple vestigial stock used twice in the experiment. The extra bristles oc- curred among all classes in the experiment indifferently, which would seem to indicate that the gene were not second chromo- some, since if it were the extras should have been relatively more frequent among the purple vestigials. The number of extra: bristles varied from one to four, the highest total bristle number observed being eight. Extra bristles were also observed to be frequent in two or three other stocks. A stock throwing extra thoracic bristles derived from B39.2 was maintained by careless mass selection for some time and was finally given to Mr. E. C. MacDowell to be used as the basis of rigorous selection experiments (MacDowell, 715). As the result of a survey of all stocks known or suspected to contain extra bristles, MacDowell chose a certain wild stock as the most favorable starting-point for his selection. 276 CALVIN B. BRIDGES In culture B9 a jaunty (jJaunty 4) appeared which gave rise to a stock similar to the original jaunty, but so far as known of separate origin. In three or four of the cultures, for example, in B9.1, are wings (are 6) appeared, and these were indistinguishable from the original arc, though quite certainly of different origin. Since these early experiments many other mutations have arisen in experiments involving purple, but these need no special mention here. THE INVIABILITY OF VESTIGIAL—PREMATURATION, REPUGNANCE, LETHALS One of the most striking features of these crosses involving purple and vestigial was the failure of vestigial to appear in as high a proportion as expected. In the F, (table 2) where 25 per cent of the flies were expected to be vestigial, only 12 per cent were vestigial; in the back crosses where half of the flies were expected to be vestigial, only 29 per cent (table 1) and 36 per cent (table 3) were vestigial. That is, only about half as many vestigials as were expected appeared in these back crosses. Such a condition is usually described by the blanket term ‘inviability;’ but a consideration of the ‘inviability’ met with in the case of rudimentary (Morgan, 712) had just led to two: new conceptions: first, that the power of fertilization possessed by a given gamete is influenced by its genetic environment prior to maturation; second, that a given type of gamete is less likely to produce a viable zygote with one than with another of two classes of sperm. The conception of ‘prematuration’ was used to ac- count for the fact that a rudimentary-bearing egg from a pure rudimentary female is much less able to give a viable offspring than a like egg from a mother only heterozygous for rudi- mentary. The principle of ‘repugnance’ was exemplified by the cross of rudimentary by rudimentary, which gave no offspring whatever though repeated several hundred times and although both the male and female give offspring when outcrossed. The shortage of vestigials in the above crosses was thought to be parallel to the results given by rudimentary, except that in the GENETICS PURPLE EYE COLOR DROSOPHILA Dad case of vestigial the effects of prematuration and repugnance were not as great in degree. On the basis of these results, an analysis of the extent to which each of these principles con- tribute to the ‘inviability’ of vestigial was undertaken by G. L. Carver (results not yet published). In Carver’s investigation it was assumed that the shortage in these experiments had been largely due to a cause intrinsic to the vestigial itself, for which reason any stock of vestigial should be equally valid for the test. The stocks used in the above experiments were not used because they were full of odds and ends of mutations which might lead to confusion. The tests of Carver showed that very little pre- maturation or repugnance is inherent in vestigial, the ratios being exceptionally close to Mendelian expectation. Wherefore it seems probable that the shortage met with in the purple vestigial experiments was due to some cause peculiar to the stocks used or to the culture methods used in the experiment. Later tests with stocks descended from these original stocks have failed to give such aberrant viability. Another explanation, that has been more recently applied to particular instances in which a character ordinarily of excellent viability has not appeared in the expected proportion, in that a lethal gene is present. Thus, an autosomal lethal in the second chromosome and quite far to the right of vestigial (i.e., close to speck) would give results roughly comparable to those observed. The difficulty with such an explanation in this case is that the uniform results given by all the cultures would re- quire the lethal to be present in nearly all the individuals—a frequency entirely out of the question both from a priori con- siderations and from the results of subsequent tests made with these stocks. Such lethals as have been found in the vestigial stock (by L. J. Cole and by students at Columbia) should not give results like those observed. 278 CALVIN B. BRIDGES THE PURPLE ‘EPIDEMIC,’ ‘MUTATING PERIODS’ Shortly after the discovery of purple, purples or eye colors closely resembling purple began to be found in stock and experi- ments everywhere. In the interval of six months following the discovery of purple such occurrences numbered fourteen and furnished the first as well as the most striking of the ‘epidemics of mutation’ that seemed to sweep over our material at this period. From later and well-authenticated cases (e.g., ver- milion, cut, notch, ete.) it appears that certain mutations do recur, and in the case of cut, four independent occurrences followed one another so closely that the term ‘epidemic’ is descriptive of the condition observed. However, in the early cases (purple, jaunty, arc, etc.) it is certain that a large majority of the apparent cases were not true reoccurrences of the mutative change, but were due to several other conditions. Thus, the Ist, 5th, 6th, and 13th apparent purples proved to be maroon, a third chromosome eye color practically indistinguishable from purple in appearance. That is, ‘mimic’ mutations were not at first distinguished from the original type. Nor were new mutant allelomorphs distinguished from types already known unless the difference was striking. Certain others of the occur- rences were proved not to be of independent origin, thus purples 8 and 9 were both shown to have been descended from a certain common stock, and purples 10 and 11 were traced to a second common stock. It is undoubtedly true that in many cases where no connection can be traced such connection really existed, especially in the case of recessives, which might be distributed without giving sign of their presence. The psycho- logical element, too, is important—it is exceedingly difficult to recognize a mutative change, even a striking one, before one becomes ‘sensitized’ to that particular mutation. Some of our mutant characters had long been present in stocks or experi- ments so that many flies showing the character must have been seen before attention became sharply focused upon the differ- ences shown. Contamination and errors of one sort or another have also inflated the number of apparent reoccurrences of mutations. It is therefore to be doubted if more than two of the apparent reoccurrences of purple were genuine remutations. GENETICS PURPLE EYE COLOR DROSOPHILA 279 REPETITION OF THE PURPLE VESTIGIAL BACK-CROSS TESTS Because of the number of disturbing conditions that had been met with in the first set of tests of the linkage of purple and vestigial, a second and more extensive set was started. These second experiments were carefully planned, and in the results ob- tained approach present standards of uniformity and reliability. The viability of vestigial was excellent, and the equality of con- trary classes throughout the experiments speaks for the favorable culture conditions. The new experiments were conducted with a purple vestigial stock descended from that used in the experi- ments of table 3, but cleared of mutations and perhaps other disturbing factors by outcrossing to wild and by selection started among the F; progeny and maintained for several generations until it seemed probable that the stock was clean. Also, from the progeny of table 3 some purple (not-vestigial) crossovers were selected, and from them was secured in a few generations a simple purple stock free from vestigial and from the other mutant characters known to be present. A preliminary test of the qualities of this purple stock was made by outcrossing a male to a wild female and carefully examining all F: flies (table 4). The F, showed only purple (150) and wild-type (300) flies as expected, but the ratio was 1:2 instead of 1:3. While this deviation was significant (4.1 times the probable error), it indicated a peculiarity of the wild parent rather than of the purple, and was not further regarded. The vestigial stock used was that from which purple itself was derived. It had been examined frequently and seemed to be clean. TABLE 4 The F2 offspring from the cross of a purple male to a wild female WILD TYPE PURPLE 1912, NOVEMBER 25 Females Males Females Males WilinSaeee eae Tay Pibnts cic ele kee 118 81 35 32 Oy 7/) Sek epee oe ee nan AT 54 33 40 Tota prea inate cck alin acts. 300 150 280 CALVIN B. BRIDGES TABLE 5! The B. C. offspring given by the F, wild-type sons, from the outcross of a purple vestigial male to a wild female, when back crossed to purple vestigial females NON-CROSSOVERS CROSSOVERS 1913, suLY 7 ee ae ae ee S| eee eee en Purple vestigial} Wild type Purple Vestigial DO: 2tbae eee eee 62 52 0 0 IDR. xste sett kate cae 113 141 0 0 DS .caee eee ae lee aera 131 96 0 0 POSH Aa, Seas 34 28 0 0 13d har te Ae Seen tr 89 68 0 0 1B Pe aimee tg fe 33 22 0 0 iD Up eae REA Oi os Lei A oan 90 112 0 0 ARG UAL ine ie Raion ae Oee 552 519 0 0 1 This table and the next (table 6) were included by Morgan in his paper on ‘No crossing over in the male of Drosophila . . . .,’’ Biol. Bull., April, 1914, pp. 200 and 201. The question of crossing over in the male was the first point attacked. Complementary P; matings were made (June 13, 1913) by crossing purple vestigial to wild (‘coupling’) and by crossing purple to vestigial (‘repulsion’). F, males from these matings were back crossed singly to purple vestigial females from the stock. The parents were in several cases transferred at the end of ten days to fresh culture bottles and second broods then raised. The offspring from the ‘coupling’ experiment (table 5, 5 pairs, both broods) gave a total of 1071 flies, not one of which was a crossover, and the ‘repulsion’ experiment (table 6, three pairs, both broods) added 704 more (total 1775), not one of which was a crossover. Since these were back-cross experiments, there was no masking of results possible, and crossover gametes had every opportunity to reveal themselves had any been formed. There- fore, each fly recorded above is a true non-crossover. While the total absence of crossovers in these repetitions of the male test cannot prove that the apparent crossovers in the original test were not genuine crossovers, it added to the already large body of evidence which showed that they were aberrations from the normal condition. GENETICS PURPLE EYE COLOR DROSOPHILA 281 TABLE 6 The B. C. offspring given by the F, wild-type sons, from the outcross of a purple male to a vestigial female, when back crossed to purple vestigial females NON-CGROSSOVERS CROSSOVERS 1913, JuLy 7 ESS RAN ER SA ef Purple Vestigial Purple vestigial Wild type JD ied CAR en Sei oe te ha a 62 42 0 0 1D) Cais Cor Meee Shs Lt Memes Baa 70 78 0 0 Wega renior ta oes cd ua 61 53 0 0 EXEPT MAREAG KS! REE LAG. 66 103 0 0 Dee. Saw ighta rer. absence 79 90 0 0 Movale. = se Last Nieurachaers crs le 346 358 0 0 The second point attacked was the amount of crossing over in the female between the loci purple and vestigial. F, daughters from the same two complemetery crosses that had furnished the material for the male tests just given were back crossed singly to purple vestigial males. BALANCED INVIABILITY—COMPLEMENTARY CROSSES The reason why both ‘coupling’ and ‘repulsion’ experiments were made is that by combining the two sets of data one can calculate a linkage value more nearly free from the errors due to disproportionate inviability of any class (Bridges, 715; Muller, 16). Within each back cross the inviability effects due to a given mutant form are largely neutralized. Since the inviable form occurs both as a crossover and as a non-crossover, both of these classes are lowered, but lowered proportionately, so that the linkage ratio remains practically undisturbed. This internal balancing holds less well for combinations of characters, for any given combination occurs in an experiment either as a crossover or as a non-crossover, but not as both, and should any combi- nation have an inviability disproportionate to that of the com- ponent mutant forms, then the crossover value would be disturbed. The remedy for this condition is to balance the experiments in which a relatively invariable class occurs as THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 282 CALVIN B. BRIDGES a crossover by an equal amount of data in which this same class is a non-crossover. It is often not convenient or possi- ble to have complementary crosses of equal weight; but what- ever is done in that direction, however little, is of advantage, and even a partially balanced result is to be preferred to one from only one type of cross. With improvements in culture methods, inviability effects have been very much reduced everywhere. Also with the great increase in the number of mutations, there is now provided an abundance of forms which show only negligible inviability. Our regular werk utilizes only these viable forms, and except for very special purposes those mutants which show more than a slight inviability are avoided. The first back crosses of purple vestigial had shown a marked inviability for vestigial and a slight inviability for purple. The new back crosses’ showed practically no inviability for purple and a very moderate amount for vestigial, but still enough to repay the added labor required by the balancing cross. As in the first back cross test of the female, the linkage shown was fairly strong. Since the linkage shown by second broods proved to be different from that of firsts, only first broods will be con- sidered for the moment. The ‘coupling’ experiment (table 7) gave a total of 2839 first-brood flies, of which 305, or 10.7 per cent, were crossovers. The ‘repulsion’ first broods (table 8) gave a total of 2335 flies, of which 303, or 13 per cent, were crossovers. When the first-brood data from both these experi- ments are combined so that the inviability is balanced, the crossover value is 11.8 (table 9). These two component crossover values differed slightly from each other and from the value (9.1) obtained in the original experiment. It may be questioned whether the difference in the crossover values was entirely due to inviability. Slight differ- ences of this order, but many of them undoubtedly significant, are continually appearing in our work. Other known causes of linkage variation besides inviability are: differences in the age of parents (Bridges, ’15), or of the temperatures at which the experiments are conducted (Plough, 717), or mutant ‘crossover’ genes (Sturtevant, Muller, and Bridges), and probably to several GENETICS PURPLE EYE COLOR DROSOPHILA 283 TABLE 7 The B. C. offspring given by the F, wild-type daughters, from the outcross of a purple vestigial male to a wild female, when back crossed to purple vestigial males NON-CROSSOVERS CROSSOVERS Sees 1913, suLy 5 OF Skea sai re Wild type Purple Vestigial |CROSSOVERS 1D Sa ee et Sar 178 202 16 16 7.8 PVAMpre SEN ee. 152 227 13 14 6.6 —1.2 DD) Bee. aE ca. 91 100 18 13 14.0 WOR eee eck ae Bees oan 69 104 12 8 10.3 —3.7 ID OSG oan So ee 165 150 7 19 10.3 1D Oi Fen ae oe Se 191 216 18 17 7.9 —2.4 DIDS atts ocr eke 140 149 20 15 10.8 1D) eae tees oot SNE 116 122 9 4 Nays —5.6 1D) De Spee area oh Sek a 191 214 20 19 9.0 DIR Re 5 RE RE Ct 196 229 11 22 lens —-1.8 DR cote tt ae. 202 226 20 22 8.9 Deere ke 197 228 25 20 9.6 +0.7 DG Ass: Fete Ss 105 158 il7/ 17 11.4 WMreeriee satncatarc : sets 188 232 17 14 6.9 —4.5 G21 8 esp nhc SM lt pep te aR 123 140 26 30 17.6 DI en. PRO Me: 129 179 11 20 9.1 —8.5 FIStS acdsee ae. oe | eo 1339 154 151 10.7 Seconds: sy..5.0.... 1238 1539 116 119 7.8 —2.9 other internal and external factors not yet analyzed. The best that can be done in correction is to calculate mean values from as many experiments as possible where none of the recognized causes of variation are especially active and thus obtain a sort of composite picture of the ‘normal’ condition. 284 CALVIN B. BRIDGES TABLE 8 The B. C. offspring given by the F, wild-type daughters, from the outcross of a purple male to a vestigial female, when back crossed to purple vestigial males NON-CROSSOVERS CROSSOVERS af ie a a CHANGE Purple Vestigial ee Wild type crossovers| “1TH AGE Dies a aioe Peace 157 178 26 21 1253 DIG Ree moe eee 200 165 12 14 Deis —5.6 DSc s seis Bees 198 176 23 23 11.0 DIS Reet oe ce ree 242 195 19 26 9.3 —1.7 DI ee ee 252 227 34 38 Sys DK ge sige Stree Bee ae 198 178 26 20 10.9 —2.2 1D eae ae dnd oae & 205 158 27 32 14.0 DING Bie ke eee de 213 246 14 23 7.4 —6.6 DINE Nata See tte 66 54 6 i 12.4 DINGS eaist sate Pe: 66 64 4 i 7.8 —46 DORR Ais Merete cas 189 172 30 32 14.6 DOV thie. ee eae 217 225 13 18 Gas —8.1 FANSUS. teva geeete te 1067 965 146 157 13.0 Seconds. seen ase 1136 1073 88 108 8.1 —4.9 TABLE 9 Linkage of purple and vestigial with balanced inviability crossovers | CROSSOVERS TOTAL | op CROSSOVERS Purples.) Moen sections 1067 154 Vestigial cas) Grbac beiegee 3 965 151 Purple vestigial ncn. Gas 1195 146 Wild type. a s.o,.s-0 cae sweet: 1339 157 (otal... 0% otetreeateee ene 4566 608 5174 11.8 GENETICS PURPLE EYE COLOR DROSOPHILA 285 THE VARIATION OF CROSSING OVER WITH AGE The reason for raising second broods in these experiments was to obtain more offspring from each female and thus secure a more trustworthy index of the genetic behavior of each indi- vidual. This practice was extended to all the work at this time, and was continued until a comparison of the crossover values of the first and second broods brought out a remarkable relation in the cases involving the second chromosome. There was found to be a change in the amount of crossing over so that both in the totals for each experiment and in a great majority of the indi- vidual cultures the crossover value had fallen significantly. Equally surprising was the fact that there was no such change in the case of the first chromosome, and this added another proof of the distinctness of our linkage groups, that is, of the individuality of the chromosomes involved. The first case in which this decrease for the second chromosome was clearly seen was that of the back cross tests of the purple vestigial linkage given in tables 7 and 8. Of the eight females whose tests are given in table 7 seven showed a decrease in the percentage of crossing over and only one (F) showed an increase, which, how- ever, was smaller in amount than the smallest of the decreases. In the complementary case ‘repulsion’ (table 8) all six females showed a decided drop. The totals likewise reflected this same change; the decreases were 2.9 and 4.9 units, respectively. The crossover value calculated from the balanced second broods was 8, a decrease of 3.8 units, or, compared with the corresponding crossover value (11.8) from the balanced first broods, a 32 per cent decrease from the normal amount. Many other experi- ments have confirmed the fact of change in crossing-over fre- quency with the age of the mother, and some slight analysis has been made of the mechanism behind the results (Bridges ’15). THE LOCUS OF PURPLE—A TWO-POINT MAP The repetition of the purple vestigial back crosses was not carried out until the summer of 1913; meanwhile considerable progress had been made with the mapping of the second chromo- 2586 CALVIN B. BRIDGES some. ‘The test of the amount of crossing over in the female between the loci purple and vestigial (table 3) had given a crossover value of 9.1 units. The next crossover value to be worked out was that of black vestigial as about 20 units (Mor- gan, 712). A THREE-POINT MAP With these two values alone it was not possible to determine the relative order within the chromosome of the three loci in- volved; it was apparent that black was farther away from ves- tigial than from purple, but it could not be told whether it lay on the same or on the other side of vestigial from purple. This value was expected to be one of two values depending on the order of the genes; it should be an approximation to either the sum (20 + 9 = 29) or the difference (20 — 9 = 11) between the black vestigial and the purple vestigial values. To carry out a back-cross experiment for black and purple it was first necessary to make up the double recessive. No easy task was anticipated in this, for it has just become known that on ac- count of no crossing over in the male no double recessive could be obtained in F., and in fact none was obtained (table 10). As expected, the F, ratio approximated 2:1:1:0. Three sorts of F; mass culture matings were made: black x black, purple X purple, and black x purple. Of these matings the last type is by far the most valuable, since in case one of the flies happened to come from a black purple crossover egg X a black sperm it would give some purple offspring when crossed to purple; and these inbred, would give the required black purples as a quarter TABLE 10 P; mating, purple 7 X black 2; Fi mating, wild-type 99 and Jo Fo, 1912, ocroBER 24 PURPLE BLACK PURPLE C68. ee Sei TAS ee ee 136 0 COO eh Rte 157 0 O70: ASe ec 78 0 Totals ee c cee 371 0 GENETICS PURPLE EYE COLOR DROSOPHILA 287 of the next generation. Likewise, if one of the purples had come from a crossover black purple egg the black xX purple cross would produce some blacks that would give the required black purples upon inbreeding. If both the black and the purple chosen happened to have come from crossover eggs, then the double would be produced in F; directly. In case none of the parents proved to be from crossover gametes then at least the F; wild-type flies are equivalent to the F; and would save a generation in the repetition. The other two types of crosses would give a favorable result only if both parents happened to be from crossover eggs, in which case the double would appear among their progeny. It so happened that one of the black x black crosses gave a few black purples in F; directly, and from these a stock was made for use in back crossing. At the same time a P; mating of a black male to a purple female was started to furnish the required F; heterozy- gotes. A single test of the F, male showed, as expected, no crossing over whatever in the male (table 11). Two back-cross tests of the female gave a total of 773 flies _of which 38, or 4.9 per cent were crossovers (table 12). Of the two expected values, that of 30 is excluded entirely, and that of TABLE 11 P, mating, purple 7 X black 9; B.C., Fi @ X black purple Q NON-CROSSOVERS CROSSOVERS MALE TEST B. C. OF MALE 1912, DECEMBER 14 Black Purple Black purple Wild type 1 | DRE eePete el 2 Bore Neer Rae? acre 74 71 0 0 TABLE 12 F, mating, purple @ X black 9; B. C., Fi 2 X black purple 3 NON-CROSSOVERS CROSSOVERS BoC. OF FHMAL® 1012, DECEMBDR 12 °|_ 2 ee Black Purple Black purple Wild type OU Are di Aostetsig chereisin areas 320 339 13 18 TP Rre nee Ria Roe aA NAS old 33 43 3 4 288 CALVIN B. BRIDGES 10 is approximated, though not very closely. On this basis, the order of these genes is black purple vestigial, and not black vestigial purple. A THREE-POINT BLACK CROSS, BLACK PURPLE CURVED, WITH BALANCED INVIABILITY Most of the linkage experiments up to this time had involved only two loci, as the three just cited, namely, purple vestigial, black vestigial, and black purple. It was now realized that a more complex type of experiment involving all three loci at once would yield returns whose value far outweighed the greater labor entailed. Thus, a multiple back cross for black purple vestigial would give linkage data upon all three crossover values simultaneously, and these values would be strictly comparable, since there would be no possibility of discrepancies due to different conditions of culture or parentage. Accordingly, the simple black purple back cross was done on a scale only large enough to decide between two possible values and thus show what was the order of the three loci. A knowledge of this order is of great advantage in synthesizing the multiple recessive. It was found, as already stated, that black and vestigial are the two farthest apart and the mating was accordingly arranged so that a crossover anywhere within this whole distance would give the required triple form. That is, black purple and purple vestigial were mated together and the resulting purple offspring oD were inbred. The F, black purples and purple vestigials were crossed together in several mass cultures, and in F; some triples occurred, showing that some of both kinds of F, flies used had come from crossovers eggs. A better method would have been to back cross the F; female by a black vestigial male. In this case every black vestigial crossover would be b pr v f known to be of the composition ules , and these inbred b + vg would give the pure triple without the chance of failure that the method, actually used, ran. A stock of black purple ves- tigial was made from the triple recessives that hatched in F's (March, 1913). GENETICS PURPLE EYE COLOR DROSOPHILA 289 In carrying out the triple back cross, the principle of balancing the inviability by complementary crosses was applied. To com- pletely balance a three locus experiment required four types of crosses, so that every class may appear in each of the four crossover categories, namely, (0) non-crossovers, (1) crossovers in the first region, that between black and purple, (2) crossovers in the second region, that between purple and vestigial, and (1, 2) double crossovers, the simultaneous occurrences of crossing over in both regions. Thus, the cross of black purple vestigial by wild and the back cross of the F; wild-type daughters by the triple recessive male gave one of the four types of crosses (table 13). The other types of experiment carried out were black by purple vestigial (table 14), black vestigial by purple (table 15), and black purple by vestigial (table 16). In order TABLE 13 P, mating, black purple vestigial # X wild female; B. C. mating F, wild-type 9 X black purple vestigial ¥ b pr vg _b bivpe }|L" db) eae pr vg vg pr | 1914, JANUARY 9 ee fey etn oe penne 5 Black | wild Purple | Black -«.7| Black Se ibe Black Seatieal paegle vestigial Teas Purple OTA TRESS 795 EC 420 24: 89 140 3 1 11 12 1 1 TAD ett. fsts tase 118 137 3 3 15 9 ~- — TAS ae ey site 61 78 2 at 12 11 — = Total? ef: ar 268 355 8 8 38 32 1 hee TABLE 14 P,, black X purple vestigial; B. C. test of Fi: 2 2 singly 1914, b blpr__ve b vg bj pr ! MARCH 9 pr ve” pr wy ail lvo 88 114 96 10 10 fi 15 1 1 103 92 81 15 5 12 16 1 0 104 99 98 Mil 15 8 17 0) 0 116 97 66 2 12 14 13 0 1 116 164 (i 6 15 12 19 2 0 Total...) 566 418 tt 57 53 80 4 2 290 CALVIN B. BRIDGES TABLE 15 P,, black vestigial X purple; B. C. test of F; 2 2 singly 1914, b vg _ bj pr b cpeay vg MARCH 10 pr vg pr vg 101 85 126 10 7 . 23 15 1 0 102 137 133 13 13 16 23 0 0 112 67 68 7 7 12 8 0 1 113 92 137 13 8 21 9 1 1 Total...| 381 464 43 35 72 55 2 2 TABLE 16 P,, black X purple X vestigial; B. C., of Fi 9 2 singly 1914, Deer mb) vg b_opr [re b OCTOBER 28 vg pr pr vg 654 130 152 10 5 14 10 0 0 670 124 111 10 8 11 14 1 1 671 137 138 15 6 12 30 2 2 672 131 154 7 10 18 12 0 il 673 162 151 11 a 12 13 2 0 674 159 162 8 u 16 24 0 2 Total...| 843 868 61 43 83 103 5 6 that these four crosses should balance closely, the same number of cultures (six) was started in each case. A few of these cul- tures failed, and the total data in the separate experiments are consequently not in equal amounts. The balance is for this reason not perfect, though such partially balanced results are far better than an equal amount of data secured from only one of the four possible types of experiment. Additional cultures could have been raised until a balance was reached, and such a practice has been followed in other cases, for example, the ver- milion sable forked case reported by Morgan and Bridges (’16). A summary of these complementary crosses appears in table 17, from which the following balanced crossover values are calcu- lated: black purple 6.4, purple vestigial 10.8, and black vestigial 16.3. GENETICS PURPLE EYE COLOR DROSOPHILA 291 TABLE 17 A summary of the four types of black purple vestigial black cross, with inviability balanced COMBINATIONS | 0 1 2 12 TOTAL Boba ive 268 355 8 8 38 32 1 1 3 ST ET ISOS SECT (pee eo ed eso SO ae as ae 711 623 16 70 2 i f 566 418 44 57 33 = 80 4 2 Se ete ert on SE sia Niieeaat na: fol eet 307 Rachile reactions of normal Ascidiaymentulla.:.45.550.5-s946>456-645e ae. 310 A. Observations on unstimulated individuals....................... .. 310 iBeEResponsestuolachie stimmullationiyy. ee eee: oe aL aes aise Cee mrold Effect of operations upon the tactile response...........:.......+.......... dll ae SH phe ee GHC GID NONS Ss vega esi: Boho: ote clad arc sbecte the cot: oll PP UA EO MOL SP MOLE s tetera se setae cle enti Conch ee ices tek 312 Gylixtinpationiof the gangions t.o48h). Aes. Otel, Oy al MIS Fos! 315 Reactions of Ascidia meftula to chemical substances...................... 317 Reactions of Ascidia, mentula to vibrations.......:... 2560.00 .20ceceeecseee 319 SuMmMaAryromreachonstonASClaia mentula-.....-.0+s.2 22 40. den. sche ec ee 320 Sensory reactions of Ascidia atra Lesueur.................... 0.0000 eee ee eeee 322 Experimentsron;Giona intestinalis, 4 i. 2c). eed vad). ogeke 32. ei 323 eaction: ossuunicates amdeehins “sew. shee lee) areola ye, tea 327 Summary of results of previous investigations....................00000000s 333 EA DIO GLAD Yih 15 he x Joe lee CALGON he ERR a eetele CA eto ee ON eee Bete eo ee ee 335 INTRODUCTION When one beholds a solitary ascidian in normal surroundings, one comes to the conclusion that such an organism needs no complicated nervous system to cope with the exigencies of its environment. Unless the observer disturbs the ascidian by stamping on the sand, he may watch it for a considerable length of time without detecting any movements which would indicate it to be an animate object. From time to time, however, it is 307 308 EDWARD C. DAY seized with convulsions. The attacks, though violent, are brief, and soon the animal is as erect and motionless as before. These sudden contractions at irregular intervals are the only vigorous muscular movements which the tunicate makes. They are performed in order to clear the pharyngeal sac of foreign particles, or to expel faeces, sperm, and mature ova from the atrial chamber. When the animal is strongly stimulated around either siphon, this characteristic vomiting reaction occurs. Stamping on the sand or clapping two stones under water will call forth a similar response. The lobes of either siphon are capable, however, of individual response and react to feeble stimulation with a local curling in of the lobe involved; a stronger stimulus causes puckering of the other lobes as well, while a still more vigorous application of the stimulus elicits an almost simul- taneous closure of both siphons, the one stimulated being usually the first to respond. This response of the siphons being an easily observed reaction, it served as a good indicator of the normal functioning of the nervous system. ‘The circlet of tentacles just inside the aperture of the oral siphon which forms in some species a hairy strainer to prevent too large bodies from being swept into the pharynx with the ingoing current, could likewise be used as a reaction indicator, though with less success. The lip-lobe reaction was by far the most delicate of the two, and through it the general irritability of the animal and the conductivity of the ganglion could be studied under varied ex- perimental conditions. Through this reaction the relation of the central nervous system to the sense organs and musculature could be fairly well ascertained, but as an index of the relation of the ganglion to the activity of the visceral organs this reaction was useless. ; While a few preliminary studies were made upon Ascidia atra Lesueur, a species found on the coral reefs of Bermuda, most of the investigations were carried out on Ascidia mentula, Ciona intestinalis, and Ascidia mammillata, but chiefly on Ascidia mentula, at the Naples Zoological Station. NERVOUS SYSTEM OF THE TUNICATE 309 My interest was first aroused in the subject of tunicate re- actions during a zoological expedition from Harvard University to the Bermudas in 1910, and it led to later investigations carried out in 1913 at Naples. For the opportunity of enjoying the facilities of the Naples Station I am indebted, on the one hand, to Harvard University for a traveling fellowship, and on the other, to the Smithsonian Institution for the use of its research table there. I wish here to express my thanks to the various members of the Naples staff for their services and nee kind- nesses during my sojourn. In how far Ascidia atra differs from Ascidia mentula in the finer details I cannot say, but in the gross anatomy they are very similar, although in color the former is bluish black and the latter a milky white. Both are from 4 to 6 inches in length, have a smooth outer tunic unornamented with protuberances, hairs, or other local modifications. The margins of incurrent and excurrent siphons of both species are lobed. The only record of the number of lobes which I have for Ascidia atra is a general statement that the number on the incurrent siphon ranges from 7 to 11, while that for the excurrent is 5 to 6: Hecht (18) says 8 on the oral and 6 on the atrial. For Ascidia mentula the average number for the former is 9 and for the latter 6. The interlobular margins of both siphons of Ascidia mentula are edged with red pigment, while the lobes themselves are, as a rule, uncolored. Just proximal to these red edges are situated isolated pigment spots; a single one usually between each lobe on the oral siphon, and two or more in a cluster be- tween each lobe on the aboral siphon. Occasionally an animal was brought to the laboratory which was totally brick red. Ascidia atra, on the other hand, is a velvety bluish black all over, due to the presence of pigment in the outer tunic. TI never ran across one, however, where pigment was restricted to only the margins of the inane Ascidia mammillata, like Ascidia mentula, is whitish except for the pigment spots, in this case black, around the incurrent and excurrent apertures. Ciona intestinalis has red pigment spots on the margins of the siphons. Mention is made of these details of pigmentation because they 310 EDWARD C. DAY will later be referred to in connection with the sensitivity of the animals to light. The procedure of experimentation was first to study the activities of the normal, undisturbed animal under as natural conditions as possible; second, to record its reactions to stimu- lations of various kinds, and, third, to discover the effect of various operative procedures such as incisions, amputations of siphons, and extirpation of the ganglion, upon the sensory reactions and the beat of the heart. Studies made upon the heart-beat will be given in a subsequent paper. TACTILE REACTIONS OF NORMAL ACIDIA MENTULA A. Observations on unstimulated individuals By the term unstimulated is meant cases in which I introduced no stimulating agent myself, although the animals were never entirely free from disturbing influences. The animals remain motionless for long intervals with both siphons wide open when they are in their natural environment of shore-water. JI have observed the same thing in the case of animals attached to the walls of large cement tanks in the ' Naples Aquarium, where natural conditions have been dupli- cated as nearly as possible. In the laboratory, although the tanks were large and provided with running sea-water, the animals in them were not entirely insulated from disturbing vibrations in the room. A protocol was kept of continuous observations on two speci- mens of Ascidia mentula for a period of about two hours and the time recorded for every closure of both incurrent and excurrent siphons. The animals were kept singly in large glass jars furnished with running sea-water. Both animals were very re- sponsive to vibratory disturbances and closed their siphons whenever a door shut or somebody walked across the floor. Hecht (’18) finds Ascidia atra also extremely sensitive to vibra- tions in the room. The response consisted simply of a puckering of the marginal lobes of the siphons; seldom did the whole body contract. When the room was quiet for a length of time, for one NERVOUS SYSTEM OF THE TUNICATE ok of the animals no reactions were observed, while for the other frequent contractions were noticed, especially of the atrial siphon. This siphon proved to contain one or two parasitic crustacea, 3 to 4 mm. long, partially embedded in the siphon near its tip, and whenever they kicked up a rumpus the siphon closed, and the closing of this siphon often induced closure of the other. The parasitized atrial siphon being in a quasi con- tinuous state of irritation responded more readily to the vibra- tions in the room than did the unexcited siphon. Sometimes small crabs were found living symbiotically in the pharyngeal sacs of the tunicates. B. Responses to tactile stimulation To penciling with a bristle the siphons give local responses, thus only the lip-lobe on the stimulated side of the siphon puckers in, provided that the stimulus is feeble; to a stronger stimulus, however, all lobes of the stimulated siphon respond, and if the penciling be of sufficient strength, the closure of the incurrent siphon is followed by closure of the excurrent siphon as well. The base and body of the animal are insensitive to penciling, while the necks of the siphons are slightly sensitive, though less so than the lip-lobes. The margins of the two siphons are there- fore the most sensitive areas of the entire body. As to relative responsiveness, little difference could be discovered between oral and aboral siphons; in one or two cases the aboral or excurrent siphon was the more sensitive of the two. EFFECT OF OPERATIONS UPON THE TACTILE RESPONSE A. Slitting the siphons When the oral siphon was slit lengthwise for half an inch, cleaving it in two parts, and the lobes of one side were stimulated, there was a response first in that half, then after a second’s latent period a response in the other half. A reversal of the 312 EDWARD C. DAY sequence occurred by stimulating the other half first. Thus the - wave of stimulation could travel in either direction around the cut. No effort was made to determine whether this stimu- lation wave was of the nature of a nerve impulse, or whether it was sunply a wave of muscular contraction which depends for its completion upon the continuity of the muscle tissues; nor further, whether the length of the latent period was dependent upon the intensity of the stimulus to any degree, or in how far it was affected by the depth of the incision. B. Amputation of siphons When an Ascidia mentula was narcotized with cocain, the oral siphon amputated below the circlet of oral tentacles, and both the animal and the amputated piece were put into fresh sea-water and allowed to recover, the aboral siphon showed a return of sensitivity in a little over an hour while the amputated siphon did not regain sensitivity until two or three days after the operation. When the operation was performed without first narcotizing, recovery of the amputated piece took place in twelve to fourteen hours, whether it was the oral or aboral siphon. Sunilar operations on the siphons of Ciona intestinalis revealed a greater recuperative power than that of Ascidia mentula, amputated siphons recovering sensitivity twenty-five minutes after the operation. The siphons possess, therefore, an irritability which is inde- pendent of any connection with the nerve ganglion. The thresh- old of stimulation lies higher for amputated siphons than for the intact siphons, it requiring a stronger stimulation to elicit a response. This diminution in sensitivity was probably oc- casioned in part by the insufficient blood supply due to the operation, with a consequent lack of oxygen and an accumulation of catabolic products, and in part by the interruption of the nerve reflex. The presence or absence of the rings of oral tentacles made no obvious difference in the general responsive- ness of the amputated oral siphon. NERVOUS SYSTEM OF THE TUNICATE 313 Not only did these amputated siphons exhibit independent irritability, but in the case of Ciona, an automatic rhythm was also manifested by the severed oral siphons. A siphon which had been cut off distal to the circlet of oral tentacles was observed two days after the operation to execute a series of rhythmical contractions which consisted of the periodic curling of the lip- lobes either inward or outward and the occasional constriction of the neck region. There was no special sequence in which the several lobes took part in the contraction, for a contraction would start on one side and be followed by a puckering, some- times of the opposite side, sometimes of an adjacent part. The proximal portion of the amputated piece, i.e., the region of the cut margin, played no réle in these contractions, the rhythmical movements being restricted to the oral margin and to the di- . rectly underlying portion of the neck. In fact, when the proxi- mal part was stimulated, the response occurred not in the proximal end, but in the distal after a short latent period, the impulse having traveled from the less irritable region of stimu- lation to the more irritable region of the sensitive lobes before producing a response. No attempt was made to analyze the underlying phenomena of this response or to seek an explanation of the rhythmicity. No such rhythm was observed for ampu- tated siphons of Ascidia mentula. As it has frequently been demonstrated that muscle tissue may be thrown into rhythmical contractions by the presence of various salts in the water, this rhythmical contraction in the Ciona siphon might have been due to the automatic response of either the muscle or the nerve tissue to the stimulating effect of the sea-water on the cut surfaces. Whether sensory cells are present in the siphon regions or not I have made no histological examination to ascertain, but nerve fibers can be seen extending out into the lip-lobes. Tunicates whose siphons had been amputated regenerated new ones in the course of about three weeks, but amputated pieces lived only five or six days and died without any sign of beginning to regenerate a new body. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 314 EDWARD C. DAY The three following protocols will give the history of such an operation with its subsequent effects both on the amputated siphons and on the desiphonated body: Protocol 1. Ascidia mentula nos. 10, 11, and 12. April 17, 1913. Narcotized all three animals with cocaine in sea-water and cut off the tips of the siphons at XX’ on the oral or incurrent siphon (distal to oral tentacles) and at YY’ on the aboral or excurrent siphon (fig. 1). Returned animals to running sea-water. April 18. Amputated pieces A and C are sensitive to tactile stimu- lation (B pieces not tested). April 19. A and C respond to tactile stimulation, but not to tapping on the jar containing them. One of the bodies B is sensitive to the tapping and responds by ejecting water from the siphons. Pharyn- geal sacs have fallen away from the cut surfaces of the siphons. April 23. A and C pieces are dead. (The supply of running water had got accidentally shut off.) The body pieces, B, are alive and appear to be regenerating new tips. May 9. New siphons have been formed on all three animals. These lie within the old cut tips and appear to have been developed from the pharyngeal sac. Distinct lip-lobes and interlobular pigment spots are visible on most of them: 7 to 8 lobes on 10C, 8 on 11A, 6 on 11C, 5 on 12A; on 10A and 12C the siphons are too puckered up to make out the number. Protocol 2. Ascidia m. no. 5. April 12, 1918. Cocainized the animal (400 cc. sea-water + 1 cc. 5 per cent cocain), and amputated oral siphon below the circlet of oral tentacles at XX’ (fig. 2). Pro- cedure was as follows: 5:10 p.m. Animal placed in cocain solution. 6:10 p.m. Benumbed; operated on, rinsed in fresh sea-water and returned to running sea-water. 6:45 p.m. Excurrent siphon sensitive to tactile stimulation; ampu- tated incurrent siphon A, unresponsive. April 15. A is responsive around marginal lobes; tentacles not. Lobes pucker in in response to tactile stimulation and to rap on sub- stratum. B is very sensitive both on cut surface and on excurrent siphon. April 16. A and B respond to tap on jar. April 17. B responds to tap on jar; A, too, if tap is strong. April 19. B responds to tap on jar; A no longer. May 9. B has regenerated a new incurrent siphon with six small lip-lobes and a new ring of pharyngeal tentacles. New siphon puckers shut upon tapping the jar. Protocol 3. Ascidia mentula no. 17. May 10. Incurrent siphon amputated to include the nerve ganglion. (Since the object of this experiment was to study the effects of the operation on the heart beat, no attention was paid to the amputated piece.) June 3. A new incurrent siphon and ganglion have regenerated; six lobes to the siphon (fig. 3). ~ NERVOUS SYSTEM OF THE TUNICATE 315 That the nerve ganglion is quite dispensable to the response of the siphons to stimulation is evident from the foregoing re- sults. Amputating the siphons, however, involved cutting off the blood-supply from the severed part, and the effect of break- ing all connections with the ganglion was so complicated with interrupting the circulation that no safe conclusions could be drawn with regard to the nervous control exerted by the gang- lion. By removing the ganglion, therefore, without at the same time injuring the circulation, its relation to the siphonal responses could be separately determined. C. Extirpation of the ganglion Four Ascidiae mentulae were operated upon by excising the ganglia, and they lived long enough for the ganglia to regenerate —a period of about one month. Immediately following the operation on two of the animals, the oral siphons began to open, but their irritability to stimulation was greatly reduced. Since the operation was performed with- out the use of a narcotic, this diminution of sensitiveness was probably due to shock. With the lapse of a few days irritability gradually increased, but it never attained the level of the normal animal until towards the end of the month when the nervous tissue had regenerated. Thus, although the ascidians responded to penciling and to tapping on the jar in which they lay, a greater strength of stimulus was required; the animals were no longer disturbed by those extraneous vibrations from the closing of doors and the treading of feet which had produced responses prior to the operation. Another animal which had had the ganglion destroyed by painting it with nitric acid, also opened its oral siphon imme- diately (1 minute) after the operation, but kept its aboral siphon closed for some time afterwards. Three days later, both siphons were open and responded to tactile stimulation, but the responses were of an inferior order. Two marked changes in the reactions of the animals were pro- duced by these operations on the ganglia: 1) a decrease in 316 EDWARD C. DAY general sensitiveness and, 2) a complete break in the coérdination of the siphons. While each siphon was capable of responding when stimulated directly, it did not join in the response when the other siphon was stimulated, as was the case before the operation. This state of incodrdination lasted until the gan- glionie tissue regenerated, about four to-five weeks later. A brief history of a single case will illustrate the course of events following an operation on the ganglion. Protocol 4. Ascidia mentula no. 2. April 12. Animal removed fromm water and ganglion extirpated. Oral siphon opened immediately after operation and before animal was returned to the water; aboral siphon remained closed. April 15. Siphons both responsive to tactile stimulation with a bristle and to tapping on table or jar with a scalpel. Nocodrdination between -the two. The tentacles are also sensitive to penciling, and when they are touched the oral siphon closes. April 26. Siphons both responsive to stimulation. No coérdination. May 5. Faint signs of recodrdination between the siphons. May 9. Co6érdination of siphons definitely reestablished. Out of ten trials in which the oral siphon was stimulated with a camel’s-hair brush, the oral siphon responded ten times and the aboral seven, the sequence of contraction being from oral to aboral with a latent period of about one-half second between the two responses. When the aboral was stimulated, out of ten trials the oral siphon responded five times to its ten, the sequence being aboral to oral. The regenerated ganglion can be recognized as a small whitish body lying in the region of the original one. The other three animals which had been operated upon at the same time as the one whose history has been given above, had also regenerated their ganglia, and, upon being tested, showed that the codrdination of the siphons had likewise been re- established. The fifth ascidian, whose ganglion had been de- stroyed with nitric acid, died a week after the operation. Up to that time, however, the siphons retained their independent sensitivity, but no coérdination existed between them. Two weeks earlier, but three weeks after the removal of the ganglia, none of the four ascidians exhibited any signs of co- ordination, although the siphons responded independently to direct stimulation. At the end of five weeks, however, the stimulation of one siphon brought about not only the closure of NERVOUS SYSTEM OF THE TUNICATE aw that siphon, but also the closure of the other siphon as well. The unstimulated siphon did not always respond, owing ap- parently to the exhausted condition of the ganglion which seemed to fatigue quickly with too frequent an application of the stimulus. As no method was employed either for measuring the strength of the stimulus or for graphically recording the response, no exact comparison could be made between the original normal responses and those subsequent to regeneration of the ganglion. The sensitivity of the animal as a whole, how- ever, did not seem so great after the restoration of ganglionic tissues as before the ganglion had been extirpated, for although the siphons responded readily to feeble penciling with a camel’s- hair brush, they did not respond to vibratory disturbances, such as the shutting of doors and the treading of feet. This in- ability to respond to vibrations may find its explanation in the ease with which the ganglion is fatigued, because, even though the ganglion were composed of nervous tissue of a higher degree of sensitivity than before, the depressing effect of frequent vibra- tory stimulations would prevent the state of irritability from attaining to a maximum. The latent period between the re- sponses of the in- and excurrent siphons was not measured, but it was about one-quarter to one-half second. Just what the histological condition of the ganglion is at the time of restoration of the power of coérdination it would be of interest to know. The neurogenesis, however, of the regenerat- ing ganglion and its relation to restoration of physiological function is a problem which still awaits investigation. REACTION OF ASCIDIA MENTULA TO CHEMICAL SUBSTANCES While the following are but a few incidental tests of a tentative nature, they throw a little light on the chemical sense of the animal. In the work of Hecht (’18) will be found an intensive treatment of the reactions of Ascidia atra to chemicals. Needle-like crystals of quinine (} mm. in length) were dropped on to the oral tentacles of three animals. After lying on them for a moment or two, they were ejected. A thread, knotted on 318 EDWARD C. DAY the end, weighted with a bit of tinfoil and soaked in a chemical solution of NaCl 2n, NaOH n/10, or quinine 10 per cent, and lowered on to the tentacles likewise evoked an ejection reaction. I operated on six Ascidiae mentulae as follows: three incurrent siphons were severed proximal to the oral tentacles (fig. 2) and three distal to them (fig. 1); the excurrent siphons of the latter Figs.land2 Two Ascidiae mentulae operated on as shown by dotted lines: in Fig. 1 the incurrent siphon (A) was amputated distal tothe circlet of oral tentacles, in Fig. 2 proximal to the tentacles; excurrent siphon (6) also cut off of the first animal. Nerve ganglion is shown as a black spot near the crotch of the siphons. Fig.3 Aregenerated Ascidia mentula. Twenty-four days after the incurrent siphon had been cut off carrying the nerve ganglion (gn) with it, a new siphon and ganglion had been regenerated and a codrdinate response of both siphons was obtained upon stimulating either siphon independently. three were also amputated a half inch from the tip. The am- putated pieces opened immediately after the operation, but were unresponsive to chemical stimulation. The following day when crystals of quinine were dropped on to the tentacles of the amputated siphons, it caused: them to rise up. Crystals dropped on the lip-lobes of both incurrent and excurrent amputated siphons caused local contractions of the NERVOUS SYSTEM OF THE TUNICATE 319 lobes. In the case of the incurrent siphon, the chemical stimu- lation of one lobe often induced closure of the whole siphon. On one of the excurrent siphons there was also a complete closure effected. The contraction in this case was progressive in both directions from the point of stimulation around the rim of the siphon. The lip-lobes were more sensitive than the tentacles to the quinine. The inner lining of the pharynx also proved sensi- tive to the quinine and when stimulated brought about a contraction of the pharyngeal papillae andof ten a closure of one or both siphons. The lip-lobes, tentacles, and pharyngeal lining were also sensi- tive to solutions of HCl, NaCl 2n, NaOH n/10, while to 50 per cent cane-sugar the same regions were unresponsive. The ten- tacles move about when stimulated, but do not contract, for they are stiff and apparently non-muscular organs. Often they gave an upward flip when stimulated. As this might have been due to the sudden ejection of water from the incurrent siphon, a siphon was amputated to include the tentacles, and the ten- tacles were again stimulated: they responded by rising up in a concerted reaction, but not so abruptly as before. When the ganglion was extirpated, this response of the tentacles could not be elicited. More data would probably have shown the con- trary to be true for this latter case, since the erection of the- tentacles occurs for an amputated siphon which is minus the ganglion and it ought also to occur for the intact siphon which is minus the ganglion. REACTIONS OF ASCIDIA MENTULA TO VIBRATIONS A normally sensitive individual closes its siphon to the slightest vibrations produced by disturbances in the room or by tapping on the jar in which it lies. If the incurrent siphon be amputated to include the circlet of tentacles, it is still capable of responding; but if it is excised distal to the tentacles, it no longer responds. The amputated excurrent siphon like the incurrent without its tentacles also proves to be inert to vibrations. If the ganglion be extirpated from an animal with both siphons intact, the 320 EDWARD C. DAY vibratory stimulus, if sufficiently vigorous, will still induce closure of both siphons after the effects of shock have passed. Since both siphons respond without the ganglion and since the excurrent siphon which lacks the circlet of tentacles is about as sensitive as the incurrent siphon which has them, it would seem that neither circlet of tentacles nor ganglion is the re- ceptive organ for the vibrations, but that the lip-lobes them- selves are capable of responding to the vibrations immediately. Hecht (18) finds the lip-lobes to be the vibration receptors in Ascidia atra. SUMMARY OF REACTIONS OF ASCIDIA MENTULA A. Normal animal to tactile stimulation 1. The margins of the siphons are the most sensitive part of the animal, and close when stimulated. 2. To feeble stimulation the response is local and restricted to a single lobe of the siphon. To a stronger stimulus the whole siphon closes. Further increase of strength of the stimulus pro- duces a closure of the other siphon as well. 3. There is a short latent period of about one-half second between the responses of the two siphons. 4. The sequence of closure is reversible; the impulse travels in either direction from one siphon through the nerve ganglion to the other, depending upon which siphon is stimulated. 5. Sometimes one siphon is more sensitive than the other and takes the initiative in the response when the two are stimulated simultaneously, as by vibrations in the room. B. Operated animals to tactile stimulation a. With siphons amputated. 1. Ina siphon partially slit longi- tudinally, each half responds locally to a feeble stimulus, while stronger stimulation sends the impulse around the cut and produces a response of the two halves in sequence. 2. Amputated siphons retain their sensitivity for five or six days and then die. NERVOUS SYSTEM OF THE TUNICATE a21 3. The desiphonated bodies of the operated animals recover sensitivity after the operation, live and regenerate new siphons. 4. The ganglion is not necessary to the process of regeneration of amputated siphons, as regeneration occurs even if the ganglion be amputated with the siphon. A new ganglionic mass appears in addition to the new siphon. b. With ganglion removed. 5. Extirpation of the ganglion has two main effects: @) an interruption of codrdination between the siphons; 6) a reduction of tone and general irritability of the animal. 6. The ganglion regenerates in from four to six weeks. 7 Coordination is reestablished with regeneration of the ganglion, and irritability is restored to almost its original degree. 8. The new ganglion is very quickly fatigued. C. Normal animal to chemical stimulation Lip-lobes, oral tentacles, and pharyngeal lining are sensitive to solutions of HCl, NaCl 2n, NaOH n/10, but are insensitive to 50 per cent cane-sugar solution. Quinine crystals applied to the lip-lobes cause siphons to close; applied to the oral tentacles, they cause these to flip up. D. Operated animal to chemical stimulation Quinine crystals applied to the lip-lobes or oral tentacles of amputated siphons produce the same but less vigorous response as for the normal animal. E. Normal animal to vibratory stimulation Both siphons are sensitive to disturbing vibrations in the room. F. Operated animals to vibrations Amputated incurrent siphons respond only provided they are cut off to include the circlet of tentacles; amputated excurrent siphons are insensitive. Deganglionate animals respond with closure of both siphons, but a more vigorous stimulation than normal is needed to elicit the response. o2n EDWARD C. DAY SENSORY RESPONSES OF ASCIDIA ATRA LESUEUR Since my experiments on Ascidia atra were of a preliminary character, there are only a few which may be mentioned here. For a good account of the physiology of this species reference should be made to the research of Hecht (718). The response to various forms of stimulation are briefly as follows: Tactile stimulation. When the outer surface of the test was stimulated with a bristle, the base and column of the animal were found to be insensitive, the necks of the siphons moderately sensitive, and the margins of the aperatures most sensitive of all. According to the strength of stimulation, the siphons respond independently or coérdinately. Vibratory stimulation. Tapping on the jar or, when the animals are in their normal habitat, stamping on the sand or clapping two stones together under water causes the siphons to close. Chemical responses. Acetic acid 1, 0.1, and 0.01 per cent strength when pipetted on the incurrent siphon caused it to close, but had no effect on the excurrent. A weaker solution of the acid, 0.001 per cent, and distilled water were both without effect. An animal put in one-half per cent solution of ether in sea- water became totally narcotized in eight minutes. It bent over double on itself during the process, closed its siphons, and became quite insensitive to tactile stimulation. When returned to running sea-water again, it revived in twenty to twenty-five minutes; the animal straightened somewhat, both siphons opened and responded to stimulation with a bristle. Either siphon could be locally anesthetized by pipetting a 1 per cent solution of chloroform in sea-water on to it. With both siphons rendered insensitive in this way, the animal still exhibited the ciliary current entering the oral and issuing from the aboral siphon, and it also gave vomiting reactions from time to time, forcibly ejecting water from both siphons and ‘thereby indicating that the body musculature was still active. NERVOUS SYSTEM OF THE TUNICATE 323 Operations. A few amputations and incisions were tried on Ascidia atra, but the animals did not live more than a day or two after the operations, due apparently to adverse laboratory conditions: owing to the presence of iron rust in the laboratory water, it was impossible to keep the tunicates alive for more than four or five days. No experiments were tried on ampu- tated siphons. An animal, however, from which the ganglion had been excised, recovered and displayed irritability of both siphons to tactile stimulation. As the animal died shortly after, no regeneration had time to occur. Light reactions. Both oral and aboral siphons of Ascidiaatra were tested for sensitivity to sunlight, but no evidence of a positive nature was obtained. When the animals lay in jars of running sea-water and kept in semidarkness, they closed the siphons periodically at approximately one-minute intervals, and when sunlight was focused on the siphon it could not be found to have any effect on these contraction-intervals. EXPERIMENTS ON CIONA INTESTINALIS For sake of comparison, a few operative experiments were performed on Ciona in which the siphons were amputated at various levels and records of reactions made for both amputated pieces and the bodies. The results are given in protocol 5 below, and by diagrams in fig. 5. 1. Effect of operative experiments on Ciona intestinalis, Animals 7, 8 and 9. April 17, 1913. Three specimens of Ciona intestinalis were operated upon, after first narcotizing with cocain, by amputating the siphons as indicated by the dotted lines in figure 4. It will be noted that in the operation on animal 9 the nerve ganglion (black spot at the crotch) was included in the amputated piece. From the protocol it will be noted that: a. Ten minutes after the operation all the amputated pieces except 7 B were unresponsive to stimulation; the amputated excurrent siphon 7 B showed automatic rhythmical contractions. 324 EDWARD C. DAY PROTOCOL 5 Siphons of three specimens of Ciona were amputated as indicated in figure 4 by dotted lines CIONA NUMBER 7 CIONA NUMBER 8 a | SSeS April 17, 1913 April 17, 1913 April 17, 1913 April 18, 1913 April 19, 1913 12.20 p.m. 12.30 p.m. 4.45 P.M. 12.15 P.M. 4.45 P.M. Siphons ampu- tated Siphons ampu- tated Piece A, no re-| Piece A, no re- action Piece B, irregu- lar automatic ~ contractions A, reacts to tac- tile stimulus action Piece B, no re- action Piece A, reacts to tactile stimulus; also automatic contractions N.B.—7A and 8A have long latent period. 8A reacts locally to stimulation of lip-lobe; often required repeated stimulations to elicit a response. A, automatic contractions A, no reaction B, feeble sponse to tac- tile stimulus. re-| B, reacts to tac- tile stimulus A, reacts feebly| A, reacts to tac- to strong tile stimulus; stimulus. also auto- matic con- tractions B, reacts to] B, reacts to strong stimu-| stimulus lus. Cut sur- faces of body also reacts CIONA NUMBER 9 Siphons ampu- tated Piece A, no reac- tion Piece A, no reac- tion A, reacts to tac- tile stimulus; also automatic contractions B, reacts to tac- tile stimulus A, automatic con- tractions B, reacts to tac- tile stimulus A, reacts to tac- tile stimulus B, also reacts or NERVOUS SYSTEM OF THE TUNICATE 32 PROTOCOL 5—Continued | DATE HOUR CIONA NUMBER 7 CIONA NUMBER 8 CIONA NUMBER 9 Ne ee OOOO April 23, 1913 | 3.30r.m. | A, feeble; con-| A, exhibits | A, shrunken and tracts only} rhythmical inert; outside locally on one} contractions peeling off, but side of the oral there are new margin lip-lobes ap- B, feeble con- parently regen- tractions erating in the center of the mass as indi- 2 cated by 8 new red pigment spots B, reacts to stim- ulation b. Four hours later all of the amputated pieces were responsive, and the amputated incurrent siphons 8A and 9A also exhibited rhythmical automatic contractions. . ¢. The latent period of response was lengthened as a result of the operation. d. 9A, the piece with the ganglion included, gave visible evidence of regenerating a new siphon in the center of the de- generating amputated one. e. B showed little progress in the regeneration of the lost parts for the same length of time. 2. Hxperiments with amputated siphon 8A a. When the amputated piece was stimulated at its proximal or cut end, the impulse traveled distally and elicited a response first at the margin. This response was of two orders depending on the strength of stimulus: first, a local puckering in of one or two of the lip-lobes, and, second, a constriction of the neck region just proximal to the margin, followed by a pursing of the whole distal end. These responses are shown by diagrams 25 and 24, respectively, in figure 5. The asterisk indicates the point stimulated. 326 EDWARD C. DAY b. Rhythmical contractions occurred every minute or two in- volving local portions of the oral margin. These periodic con- tractions are shown by the series of diagrams in figure 5. The letters D and P indicate the distal and proximal ends, re- spectively. The arrows indicate the points at which contraction occurred. No. Time. ¥ ip eer Tee 4. 2 30/5. 4615 ba IS. z: si 48: rae: 16. 4. 33:00 48:30 Up 17. 55: ey 50:00 Gi 18. 36: woot? 50:10 U 19. i “ST Aur JI: 350. 20. 3. eat 5:54: moa 21. 9 sesol! ayes 22. 10. 40: Te sso Js 23. 4 I]. ate aie 24 12 ANG *€ 25-15 3 abaot 5 Fig. 4 Three Cionae intestinalis operated on as indicated by the dotted lines. The nerve ganglion is represented by the black spot at the crotch of the siphons. In animal No. 9 the ganglion was amputated along with the incurrent siphon. Fig. 5 A series of diagrams showing automatic rhythmical contractions in an amputated incurrent siphon of Ciona intestinalis (SA of Fig. 4). These occurred every minute, more or less, as will be seen from time recorded in minutes and seconds to the left of each figure. The distal end of the siphon is uppermost in all figures except No. 25. D distal; P proximal. Arrows indicate the points and direction of contraction (puckering in, curling out or pursing). The asterisk means a stimulus applied at that point. fo NERVOUS SYSTEM OF THE TUNICATE oot There was a certain circularity in the movement of the local contractions around the oral rim, especially seen in diagrams 15 to 20, where the invaginated part occupies successively pro- gressive positions around the margin. These single local con- tractions seemed to summate in effect and to produce, after the completion of two or three, a more general reaction of the whole margin as in nos. 4, 6, 10, 21, 24. REACTIONS OF TUNICATES TO LIGHT From a consideration of the distribution of pigment one might well be skeptical of the photoreceptive nature of these pigmented regions. Asa rule, in Ascidia mentula the pigment is restricted to a narrow interlobular band at the rim and to one or two subjacent spots, but often it is more or less extensively dis- tributed over the siphons, sometimes coloring even the whole body surface. Another genus, Cynthia papillosa, is character- istically red all over, and Ascidia atra is entirely black; in Ciona intestinalis, again, the spots are red, while in Ascidia mammillata they are black, and in both they are localized on the margins of the siphons. From ,this diverse distribution and color of the pigment, therefore, one would hardly expect it to be associated with any photic sensitivity of the animal. Ascidia mentula, Ciona and Ascidia atra were tested at Naples while Ascidia atra had previously been tested at Bermuda. The surface of the animal’s body was explored with a pencil of sunlight. For the results on Ascidia atra, see page 323. The three following tables give the results for Ascidia mentula and Ciona; in tables 1 and 2 the diameter of the beam employed was 15 mm. while in table 3 it was cut down to 5 mm. The method of illuminating the animal was to reflect the beam of sunlight with a plane mirror through a cardboard diaphragm of the specified aperture upon the animal as it lay in a depth of 4 to 5 em. of sea-water. Four regions of the body were tested in turn in this manner—incurrent siphon, ganglionic region, excurrent siphon, and the main part of the body. Contraction of the siphons was taken as indicator for the response. If no response 328 EDWARD C. DAY TABLE 1 Reaction of Ascidia mentula to a beam of sunlight 15 mm.? Animals 1 and 3 were red pigmented; 2, 4, and 5 white. The ganglion of each animal was made more accessible to light by cutting away part of the tunic. The tests were made on the second day following the operation GANGLIONIC ANIMAL TRIAL INCURR. SIPHON REGION EXCURR. SIPHON BODY 1 20'’-0 20’’-0 20/’-0 20-0 2 20'’-0 20’’-0 20'’-0 3 20’’-0 20'’-0 20'’-0 1 4 20’’-0 20’’-0 i 5 20’’-0 20’’-0 10/"-+ 20'’-0 15'’-0 ( 90/" se ( 1 20’’-0 20'’-0 20’"-0 20'"-0 | 2 20’’-0 20'’-0 yp 9 9) Ugea os 90’’-0 | 4 20’’-0 20’’-0 20’’-0 5 20’’-0 20’’-0 20’’-0 1 20’’-0 20’’-0 20’’-0 20’’-0 2 20’’-0 20’’-0 20’’-0 3 3 20’’-0 20’’-0 20'’-0 4 10’-+ 20'’-0 20'’-0 5 20’’-0 20'’-0 Caps ( 1 20’’-0 20'’-0 20’’-0 20'’-0 2 20’’-0 20’’-0 20’’-0 4 3 20’’-0 20’’-0 20’’-0 4 15/-+ 20'’-0 20’’-0 5 20’’-0 20’’-0 20’’-0 1 20'’-0 20’’-0 20'’-0 0 2 20’’-0 20'’-0 20’’-0 5 3 10/7-+ 20’’-0 20'’-0 4 20’’-0 2"'-0 20’’-0 5 20'’-0 2"'-0 20'’-0 Total + reactions for 25 trisil sept oe eee 2 eae 4 1 5 0 nee ee NERVOUS SYSTEM OF THE TUNICATE 329 TABLE 2 Reactions of Ciona intestinalis to a beam of sunlight 15mm.? Animal not operated on as in case of Ascidia m. GANGLIONIC ANIMAL TRIAL INCURR. SIPHON REGION EXCURR. SIPHON BODY 1 10’’-0 3/7 30’”-0 15’’-0 2 30’'-0 Baie 15’"-0 1 3 30’'-0 Vie 15-0 4 30/’-0 fae 15’’-0 5 30/”-0 5-4 15’-0 1 10’’-0 Ara 30’”-0 15’’-0 2 30'’-0 6+ 3/4 2 3 30/’-0 a 15’-0 4 30’'-0 rae 15’-0 5 30/’-0 Q’—4 15’-0 1 Hoos oat: 15/0 2 30’’-0 5//—+ 15’’-0 3 3 30’'-0 Hes Pe 15-0 4 30’-0 ModE 15’-0 t 5 12+ - 15’"-0 1 10’’-0 —- os 15-0 2 30’’-0 3/7-+ 15’"-0 4 3 30’’-0 ions y 15’"-0 4 a = 15’’-0 5 30’’-0 v4 15’"-0 1 10’’-0 37 —-+ 30’’-0 15’’-0 D) 30-0 preene 15-0 5 3 30'"-0 a 15’’-0 4 30’-0 ya 15’’-0 | 5 pe oe 15’’"-0 f 1 10-0 4g igen 15"-0 2 30/’-0 3//— 15’"-0 6 3 15’’-+ 33//-+ 15’’-0 4 30’’-0 = 15’-0 5 30’’-0 - 15’’"-0 1 10’’-0 53/7--+ 8//-+ 15’ 2 30/’-0 4/4 15'-0 7 3 30’-0 6”—+ 15’’-0 4 30’’-0 Were 15'-0 5 30’-0 as 15’’-0 Total + reactions....... 5 in 35 trials | 32 in 32 trials} 3 in 6 trials 1 in 35 ae EEE a. PATE FOS *. ee THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 2 330 EDWARD C. DAY TABLE 3 In this table the results are given for a more critical test on Ciona which was made with a beam of sunlight 6 mm.2 in area ANIMAL TRIALS INCURR. SIPHON GANGLIONIC REGION EXCURR. SIPHON 1 10'’-0 eye 10’’-0 D 107=0 ete 10/’-0 1 3 10’’-0 gas 10’’-0 4 10’’-0 ia 10’’-0 5 10’’"-0 10’’-0 10’’-0 ( 1 10’’-0 pee 10’-0 2 10’"-0 ES 10’’-0 2 3 10’"-0 G02 10’"-0 4 10’’-0 iaere 10’"-0 \ 5 10’’-0 5m 10’"-0 1 10’’-0 yee 10’’-0 2 10°=0 Gass 10’’"-0 3 3 10’"-0 Ee 10’’-0 4 10’’-0 10’"-0 5 10’’-0 ea 10’’-0 1 10’’-0 10’’-0 10’’-0 2 10’’-0 10’’-0 10’’"-0 4 3 10’’-0 (ess 10’’-0 4 10’’-0 Quaas 10’"-0 5 10’’-0 10’’-0 10’"-0 1 10’’-0 s7=— 10’’-0 2 10’’-0 10’-0 . 10’"-0 5 3 10’’-0 el 10’"-0 4 10’’-0 ieee 10’’-0 5 10’’-0 RieeE 10-0 1 10’’-0 yee ep lO7=0 2 10’"-0 10’"-0 10’’-0 6 3 10’’-0 (ese 10’’-0 4 10’"-0 ee 10’-0 5 10’’-0 Ri Seu 10’’-0 1 10’’-0 Ala. 10’’-0 2 10’’-0 10’’-0 10’’-0 | 3 Avia 10’’-0 4 4’) 5//-+ 10’’-0 5 4''-0 8/’—-+ 10’’"-0 Total + reactions..... 1 in 35 trials 26 in 33 trials 0 in 35 trials — NERVOUS SYSTEM OF THE TUNICATE 331 Summary of tables 2 and 3 ed GANGLIONIC A. TOTAL REACTIONS INCURR, SIPHON REGION EXCURR. SIPHON BODY For 15-mm.? beam...........| 5 out of 35 | 30 out of 32} 3 out of 6| 1 out of 35 trials. trials. trials. trials. Hor 5-mm.* beam.........-.- 1 out of 35] 26 out of 33) 0 out of 35 trials. trials. trials. Summary for both........... 6 out of 70) 58 out of 65| 3 out of 41) 1 out of 35 trials. trials. trials. trials. B, Average time of response: For 15-mm.? beam..... 4.2 sec. For 5-mm.? beam...... 8.9 sec. was obtained before the lapse of twenty seconds, the reaction was counted as negative and recorded as 20’’-0, as in trial 1, table 1, but if a contraction occurred, as in trial 4, at the end of twelve seconds, it was recorded 12’’—+, indicating a positive response. In table 3 ten seconds was the time limit allowed for the response. It will be seen from table 1 that the body of the animal gave no response out of the five times tested, the ganglionic region but 1 out of 25, the incurrent siphon 4 out of 25 and the ex- current 5 out of 25. Since all the animals were exceedingly sensitive to stimulation by tapping on the jar, the small number of responses here obtained indicate little if any sensitivity to light. Table 2 indicates, in contrast to the results obtained for Ascidia m., that the ganglionic area of Ciona is decidedly sensi- tive to light, yielding 32 responses out of 32 stimulations, while the siphons and body are relatively insensitive. The body responded but once in 35 trials, the incurrent siphon 5 in 35, and the excurrent 3 times in 6 trials. Since the excurrent siphon is much shorter than the incurrent, the 15-mm.2 beam could not be focused upon it without its overlapping the ganglionic region somewhat, it was therefore not tested as frequently as the other parts, but left for a more critical test with a smaller beam. Soe EDWARD C. DAY Table 3 gives the results of this test. Stimulation was made with a beam 5 mm. in cross-section. The pencil of light was small enough to afford a more restricted localization of the stimulus both on the siphons and over the ganglionic area. Under these conditions the ganglion again exhibited its decided sensitivity to light in comparison with the insensitiveness of the siphons, yielding 26 responses out of 33 trials to only one response for the incurrent and none for the excurrent out of 35 trials apiece. From the summarized table it will be noted that when the ganglionic region was stimulated the animal gave 58 responses (closure of siphons) out of 65 trials, while to stimulations of the other regions, the incurrent siphon yielded 6 out of 70, the ex- current 3 out of 41, and the body 1 response out of 35 trails. It was the larger of the two beams of light which was effective in eliciting responses when siphons or body were stimulated. It will also be observed that the latent period of response was longer for the smaller beam (8.9 seconds) than for the larger one (4.2 seconds); i.e., the larger the area stimulated, the shorter is the reaction time. When the body region was illuminated with a beam about 4 cm. in diameter, it squirmed and contracted in that region, while both siphons remained open. Only two of the five Cionae responded in this manner, however. The reaction time in the one case was 18 seconds and in the other 26 seconds. Although the animals lay submerged in a depth of 44 cm. of sea-water, this comparatively slow reaction time may have been due to the thermal effect rather than to the actinic effect. A thermometer placed at that depth and illuminated for twenty seconds showed a rise of temperature of 0.2°C., but no experiment was performed to decide the point in question. NERVOUS SYSTEM OF THE TUNICATE aon SUMMARY OF THE RESULTS OF PREVIOUS INVESTIGATORS Since Hecht (18) has carefully reviewed the literature of the subject, I give below, from my own reading, only a brief summary of the results of both neurohistological and physiological studies made by earlier investigators. A. Neurohistological observations (Van Beneden and Julin, ’84) 1. The ganglion consists of a central fibrillar substance surrounded by a peripheral layer of cells. 2. The largest cells lie to the outside. 3. There are localized groups of cells chiefly at the anterior end and on the ventral side of the ganglion which are suggestive of motor centers. | 4, Large nerve trunks run forward to the oral and backward to the aboral siphon. 5. These break up into branches to form a nerve net about the muscle fibers. 6. There are said to be club, brush, and plate-like motor nerve-endings present. 7. No specialized sensory nerve-endings have been described. 8. A cord of large ganglion cells extends down into the viscera. 9. The ganglion develops from the cerebral vesicle of the larva. 10. The visceral cord develops from the epithelial wall of the central canal of the larval nervous system. 11. The regenerating ganglion develops out of a derivative of the same embryonic tissue from which the original ganglion developed. (Schultze, ’00). B. Physiological observations (for Ciona intestinalis) 12. The siphons are the most sensitive parts. 13. Rapid stimulation (prodding) is more effective than slow continued pressure for the same intensity of stimulus (Poli- manti, 710). 334 EDWARD C. DAY 14. Duration of contraction increases with increase in the strength of stimulation (Kinoshita, ’10). 15. The aboral siphon is the most sensitive, judged by the shortness of the latent period; but judged by the duration of contraction, the oral is the most sensitive (Polimanti, *10). 16. Successive stimulations cause both a decrease in vigor and duration of contractions (Kinoshita, 710). 17. Contraction of the siphon begins at the tip and travels downward, while relaxation begins at the base and travels upward (Po imanti, ’10). 18. There are four types of reflexes: 1) the individual, involv- ing only the local region stimulated; 2) the protective, in which the impulse travels swiftly from siphon to siphon via the gan- glion; 3) the general, which spreads more slowly from siphon to siphon via the base of the animal; 4) the ejection reflex (Jordan, ’08). 19. Narcotizing reagents are: cocain, magnesium sulphate, chloral hydrate, aceto-chloroform, quinine sulphate, nicotine, hydrochloride (Kinoshita, 710), and morphine in weak doses (Polimanti, 710). 20. Excitative reagents are: Strychnine, and morphine in strong doses (Polimanti, ’10). 21. Curare is a partial narcotic according to Kinoshita (’10), but an excitant according to Polimanti (’10). 22. As to the effect of temperature: 30°C. causes frequent opening and closing of siphons; 32° begins to have a benumbing effect, and 35° causes animal to shrink and become unresponsive (Polimanti, ’10). 23. For the effect of light only negative results were obtained by Kinoshita (10). 24, Extirpation of the ganglion produces four chief effects: interruption in the codrdination of the siphons, decrease of sensitivity, increase of latent period and a lengthening of the duration of contraction (Kinoshita, 710; Polimanti, ’10). 25. The ganglion regulates reflexes in a feeble way either by inhibiting or facilitating them (Jordan, ’08; Polimanti, 710). NERVOUS SYSTEM OF THE TUNICATE 335 BIBLIOGRAPHY Froéuuicn, A. 1903 Beitrag zur Frage der Bedeutung des Zentral-ganglions bei Ciona intestinalis. Pfliigers Arch., Bd. 95, S. 609-615. Hecut, Sevicg 1918 The physiology of Ascidia atra Lesueur. I. General physiology. Jour. Exper. Zool., vol. 25, pp. 229-259, 15 fig. II. Sensory physiology. Jour. Exper. Zool., vol. 25, pp. 261-299, 2 fig. Jorpan, H. 1908 Uber reflexarme Tiere. Ein Beitrag zur vergleichende Physiologie des zentralen Nervensystems, vornehmlich auf Grund von Versuchen an Ciona intestinalis und Oktopoden. Zeitschr. f. allg. Physiol., Bd. 7, S. 86-135. Krnosuita, Tooxsaku 1910 Uber den Einfluss mehrerer aufeinanderfolgender wirksamer Reize auf den Ablauf der Reaktionsbewegungen bei Wir- bellosen. I. Mitteilung. Versuche an Tunicaten. Pfliigers Arch., Bd. 148, 8. 501-530. Kowatevsky, A. 1874 Uber die Knospung der Ascidien. Arch. f. mikr. Anat. Bd. 10, S. 441-470, Taf. XXX u. XXXT. : Loes, J. 1891 Untersuchungen zur physiologische Morphologie. II. Organ- biludung und Wachstum, S. 38. Maaenus, R. 1902 Die Bedeutung des Ganglions bei Ciona intestinalis. Mit- theil. a. d. Zool. Sta. Neapel, Bd. 15, S. 483-486. PoLIMANTI, OswaLp 1910 Beitrige zur Physiologie des Nerven-systems und der Bewegung bei den niederen Tieren. IJ. Ciona intestinalis L. Arch. f. Physiol., Jahrg. 1910. Suppl. Bd.S.39-152. (See also Jahrg. 1910, p. 129, same title. 1 Branchiostoma lanceolatum Yarr. Amphioxus.) Scuutrze, L. S. 1900 Die Regeneration des Ganglions von Ciona intestinalis L. und tiber das Verhiltnis der Regeneration und Knospung zur Keim- blatterlehre. Jena Zeitschr. f. Naturw., Bd. 33, S. 263-344, Taf. XII-XIIT. Van BENEDEN, Epovuarp pt CHARLES JULIN 1884 Le Systéme nerveux centrale des Ascidies adultes et ses rapports avec celui des larves urodéles. Arch. de Biol. T. V., pp. 317-3867; pl. XVI-XIX. Recherches sur le développement postembryonnaire d’une Phallusie (Phallusia scabroides Nov. Sp.). Arch. Biol. T. V., pp. 611-638, pl. XXXII. ‘ a ra , . as t? hot ; 4 - i to aie ie ‘ anil g JS : ‘ i, ¥ iY , ¥ « ab J = LNer rl ‘ ; in i Ay Py A. 5 or S Loft es wy { ‘ ie 7 bd 7 ; 5 make | hoo we : i} 4 7 i _ : — * J | ‘ee 7 j eee yee 7 x ie cw tet Si eae a . : 5 iit, a 7? - ee, os bs , 4 = _ i ; j idppneeé “whit iN a a Ay cn bat Ral at 4 | ce i ook tat ’ ; “Jo bp bene iets 0) seat ( sont iJ GP ; y H Caroa ala ae ent hitbe Ns a ey a Mae a Eee ‘nalatin wiialieckhi eeu a mi “uth uel irae" Siemens tol Henh? gale te antalt. da viltgh, proaan sah iae ts RB idate 16 Ramp ee SUT ci ay aa ley HoxAl. Bi Mil sda lio jy. LEAT Opa iad if olaplodgra st open: Lnenedaes sh ah dilsa af oa - i Ns igh en 8 Saeed Sb oN EN trae tube Ar Le dak yh ne ities iad -. ve Birdies an sie ery pieek asm uate , PERL hl a ACR ate ie TAG Ga wees eh cue tee ate ; he elite stk A et af st an tay an git : Na Jeni oir Wr iba Ra aga eae : LE ae cs lana ot airs JUL ! etre 7 ne a ot eine on lll ; : PO? x , wr it 4 7] Ae si «Ct kedeaians we Nant Mgato My yi cy unis ¥ pn Me ER al x sh cae MEP F667 ARS ch a Gy sgn Be Oh iG a (Shit "A i, “sine . Phd Ae i <4 Chole nb! why fa ale eee jr? Be Ss by Marts ae, 2 f, as ‘ao ” : <, oe ef a’ r as | i" apie aly hoe whet a if siete. :) lave abt one gaia Oh Pe ll ae day: - leone ek me} i riers: PY a au, Td | 7 an hea ae ae c an a. a J as . ie Ue a SO ay muri WT i PTL) Vis eons Onin wile. ab. ole ee ~ ‘ag INH OAK bad) anc re a ’ we ‘ ata) i 2 ry } 7) , — * Ayes lie sa " ry vat & - tt ’ — ‘ A ' : ty, ‘ ot pa cc Dob vtientaden Pe albas | MY WA: * Cok ‘int palit sil i ca Likn sage Vhs ~ wid * few vin | Resumen por el autor, Calvin Blackman Bridges Universidad de Columbia Modificadores especificos del color eosina en los ojos de Droso- phila melanogaster El autor ha demostrado la existencia de ocho genes mutantes que producen un pequento efecto o no ejercen influencia alguna sobre el color de los ojos de moscas que son homozigéticas para dichos genes y, en cambio, modifican el color de los ojos del mutante ‘‘eosina,” ligado al sexo. Estas modificaciones “‘espe- cificas’’ y ‘‘desproporcionadas” son ejemplos claros y simples de ‘‘genes multiples.” Cada uno de ellos es el resultado de la co-accion de un gene especifico modificador (crema a, crema II, oscuro, “whiting,” crema III, crema b, ‘‘pinkish,”’ crema ¢) y de un gene particular (eosina) que es necesario como una “‘base”’ o “diferenciador.” La escala de las modificaciones del color eosina producidas por estos diferentes modificadores oscila entre un color rojo mds oscuro que el ‘‘eosina’” y un color blanco puro. En su origen estos modificadores eran por completo independientes uno de otro, y el orden en que aparecen presenta una relacién casual con la seriacién de los colores oscuro y claro. La significacién mas importante de los hechos descritos en el presente trabajo es su relacién con el método mediante el cual la seleccién produce sus resultados. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 1 SPECIFIC MODIFIERS OF EOSIN EYE COLOR IN DROSOPHILA MELANOGASTER! CALVIN B. BRIDGES Columbia University, New York City TWO DIAGRAMS CONTENTS ICR eM RLOVSL It CM Ra Se eee nace snb do poco no ceo aOO Sor no cM ONG Ope o eae 337 (Re er Nl sta ih RA eta i Nh A Ri Re NS RI Ot 338 Grea Le. boa he eee ee LE MPapae eras aialere she leahoteleliete ane ars wiererene:s 341 Werk ands yhibing fe eects acts Ale is eats aoe aio te Faia Re nas tert

)| 32) 12 0 0 } 383 32:).0'-0 1 4996 on ay Oe, AL OS OR OneCare LOn Oman 0 il 4997 ORS ele e253) 0 OF SOV aey Le |p Oreo 0 4998 | 50 47/ 0 8] 1 2 One OM tor 58" 2s 3 0 OF | | | Total.) 123 125} 3 26| 7 7 O40) 118) 130+) 3 is (J) The occurrence of the linkage variation and its testing had diverted me from making tests of the effect of the cream III gene in the absence of eosin as a base. Accordingly, the preceding experiment was repeated, but in such a fashion that half of the flies should be not-eosin. A cream III ebony female from the high value stock was outcrossed to a simple dichaete male. Cynr wy © Oe Hi Daa crossed singly to cream III ebony males of the high stock. The result (table 19, left side) showed that the gene responsible for the dilution of eosin to the cream color gave by itself a color somewhat similar to purple, that is, a ‘magenta.’ But when Several of the dichaete daughters ( ) were back 362 CALVIN B. BRIDGES separations were attempted it was found that this color was not sufficiently distinct from the red to make classification accurate. In the eosin half of the flies on the other hand, the separation was easy and entirely accurate. That is, while the so-called ‘cream II’ gene by itself gave a certain effect, it gave so much more marked an effect in the presence of eosin that it was decided to retain the name ‘cream III’ instead of renaming the mutant ‘magenta.’ The separations in the not-eosin half of table 19 correspond roughly to those of the eosin half, but the large size of the ebony not-magenta class is due to the impossibility of distinguishing the ‘magenta’ character, even though it was un- doubtedly present in most of these twenty-six flies. If this character were accurately classifiable (as magenta) without first laboriously combining eosin with all the flies used in the matings of each experiment, it would be incomparably more useful even if slightly less interesting from our present viewpoint. However, an experiment planned through eight generations had to be abandoned because it was found impractical to distinguish be- tween the cream III and the not-cream III where eosin was not present as a sensitizer. But this aborted experiment revealed that the ebony was a disturbing factor—that in the not-ebony flies the distinction was sharper than in ebony flies. The new possibility arose that the mutation could be used (as magenta) by excluding ebony from all the experiments. While this has not been adequately tested, it seems hopeful that with experience one may be able to use this eye-color without a preliminary eosinization of all the stocks, though it is not to be denied that in the presence of eosin the ease and speed of classification would be greater. CREAM B - An eosin female from a stock of non-disjunction when mated to a bar male gave (culture 82, March 10, 1914) among the eosin sons one whose eye color was as light as that of cream IT or cream III. This male was outcrossed to a wild female and in F, gave creams among the eosin sons, but no disturbance of the color of the not-eosin flies (cultures 183, 184, 185). The EYE COLOR IN DROSOPHILA MELANOGASTER 363 F, ratio was again 12:3:1, as in similar crosses with other re- cessive specific dilutors. But the creams (cream b) which oc- curred in this F, were not as pale as any of the preceding creams. From the circumstances of the appearance of cream b—that it was observed in the F; of an outcross and as a single individual —we should expect it to be a dominant, but as a matter of fact it proved to bearecessive. It seems probable, in explanation, that more creams were actually present in this F, but were over- looked, since attention was distracted by the simultaneous appear- ance in the same culture of still another mutation (lethal 4), and more especially since the effect of cream b is rather slight. Only occasionally was one of the F2 creams as marked as the grand- father, and the mutation might not have been recognized at all were it not that an extreme fluctuant had attracted attention. Since cream b is recessive, we must suppose that the gene was present in both parent stocks. It could have been present in the bar stock and been undetected because of the lack of eosin, without which it has no visible effect. And the character might readily have been present in the eosin non-disjunction stock and have been passed over as an age variation, since as we ordinarily see flies from a stock culture they are of all ages and of corre- sponding densities of pigmentation. A pure breeding stock of cream b was made up for use in back crossing. By this time we were in possession of a good second chromosome dominant ‘star’ and likewise of a perfect third chromosome dominant ‘dichaete,’ which mutants have now be- come the most important in their respective chromosomes. By aid of these two dominants it is very easy to determine in a single experiment whether a given mutant is in the second or third chromosome. Thus, in the case of cream b, a stock of eosin star dichaete was made up and used in making a P; cross to the cream. Then F, eosin males which showed both star and di- chaete and which were heterozygous for the recessive cream were backerossed to cream b females of stock. There is no crossing over in the male of Drosophila, so that if cream b were in the second chromosome, none of the B.C. (back cross) stars should be cream, while half of the dichaete should be cream and half 364 CALVIN B. BRIDGES TABLE 20 The B. C. offspring from the P, mating of an eosin star dichaete male to a cream b female and the backcrossing of the F, eosin star dithaete male to cream b females NON-CROSSOVERS CROSSOVERS (IN MALE) 1916, 9/8 Fiosix a E Cc bas a : E stat bfeteaate Cream b dichaste olen b ee Eosin dichaate 5155 OF fT 5tal| fee 25 26 12 0 0 0) 0 fof ey Meee eee 19 21 21 26 . 0 0 0 0 ORE eee 14 21 18 26 0 0 0 0 HOD ii 14 15 ily 24. 0 0 0 0 Otalaee sek cee 67 82 82 77 0 0 0 0 not. If, on the other hand, the cream were in the third chromo- some, then none of the B.C. dichaetes should be cream, while the star and cream should assort at random. The experiment proved that the gene for cream b is in the second chromosome (table 20). An (eosin) star female and a cream b male selected from the B.C. offspring gave in the next generation the amount of crossing over between star and cream b (table 21). This value of 22.1 includes some double crossing over, and the corrected or ‘map’ distance is probably about 22.5. The chances are in favor of the cream b locus being to the right of star, since star happens to occupy the leftmost of the known loci. TABLE 21 The B.C. offspring given in F3 by an eosin star female and a cream b male from table 20 NON-CROSSOVERS CROSSOVERS (IN FEMALE) 1914, 10/20 Eosin star Cream b Star cream b Eosin Ol Fe eee 36 41 8 8 5593 Saas 22 41 12 12 é OF ehixt ema 47 39 13 7 824 ¥ Dae 28 49 is 15 MLGtal 22. cdie. 133 170 44 42 EYE COLOR IN DROSOPHILA MELANOGASTER 365 PINKISH In the fall of 1913 a stock of eosin black had been made up with which to test the chromosome group of cream II. In the following summer (July 27, 1914) I noticed that a few of the males were somewhat lighter in eye color than the others, but seemed chiefly noticeable because of the weakness of the yellow component of the eosin eye color. The color of the regular eosin male is a pinkish yellow; the color of creams a, II, III, and b is nearly a pure yellow with little of the pinkish tinge, while this new color was somewhat the converse of this and was pale pink with relatively little yellow. One of these males mated to a sister gave all of the sons of this pinkish color and all the daughters of a similar color, which is, however, much harder to distinguish from standard eosin. It seems that this character is somewhat sex-limited in the same direction as eosin. Pure stock of the mutation had been obtained at once through the happy selection of a pure pinkish female which had been taken to be simply an eosin female of somewhat lighter eye color because of being freshly hatched. Since pinkish appeared in a stock of eosin black, material was on hand to test the chromosome group at once. Accordingly, black pinkish females were outcrossed to eosin males and the F, eosin females, standard eosin in color, were backcrossed to black pinkish males. In the B.C. cultures half of the flies were not- black, and the not-black pinkish flies were seen to be less marked- ly ‘pinkish’ in tone than the blacks. In the absence of black the eye color was more nearly like that of the other creams, though the amount of dilution is less than in the case of any of the other creams. In the first of these B.C. cultures (table 22 males and females were both classified together. Some question having been raised as to the accuracy of the separation of pinkish from eosin among females, the cross was repeated, and the more readily classifiable males (last three cultures) gave the same re- sult as before. It was seen that the new or crossover combina- tions were as numerous (51.4 per cent) as the original classes, and this independent inheritance was taken to mean that the 366 CALVIN B. BRIDGES gene for pinkish is not in the second chromosome. While this was a mistaken notion—the true relation being that the gene is so far away from black that in the female there is entirely free crossing over—yet it led to the device of the efficient ‘double mating’ method of ridding a given stock of an undesired recessive. If pinkish were in the third chromosome, then the presence of the black in the pinkish stock could be of no advantage, and might be a very serious handicap, since it would prevent the use of all our third chromosome stocks containing ebony or sooty. The first step in the elimination of black was to mate together some of the not-black pinkish flies of table 22. One-third of the not-black offspring of such pairs should be of the desired TABLE 22 The offspring given by the F, eosin-eyed daughters from the outcross of black pinkish females to eosin males, when back-crossed to black pinkish males NON-CROSSOVERS CROSSOVERS 1914, 9/23 Black pinkieh (Eosin) (Eosin) black Pinkish 525 70 81 7(il 84 526 36 29 32 22, 2424 25 21 24 35 2425 24 27 29 31 2426 28 24 14 29 Motels 183 182 170 201 kind, that is, entirely free from black. Our task was then to pick out from the mixture of pure grays and grays heterozygous for black some pure gray males for outcrossing to eosin females. In this special case we were aided by the fact that black happens to be slightly dominant, that is, the heterozygous blacks are somewhat darker than the pure grays. While this difference is not marked enough to be used regularly in classification, yet it enables us to pick out by inspection a greater proportion of pure grays than we would get by random selection. Four such males were selected as being probably free from black and were mated to eosin females. Into the same bottle with each pair of these flies was put a virgin (red-eyed) black female. The offspring from these two mothers are easily distinguished, since EYE COLOR IN DROSOPHILA MELANOGASTER 367 they are eosin-eyed if from the eosin mother and red-eyed if from the black mother. The offspring from the black mother constitute a test of whether the father were free from black, for in this case none of the red-eyed offspring hatching in the double mating culture should be black, while if the father were hetero- zygous for black half of the red-eyed offspring should be black. Only one of the four cultures gave black offspring, and this culture was then discarded. ‘The eosin-eyed flies of the other three cultures were all heterozygous for pinkish and at the same time free from black. By mating together some of these eosin- eyed flies pure pinkish offspring should be obtained as a quarter of the offspring. A more efficient method, and the one actually followed, was to save the fathers and mate them to their eosin- eyed daughters, since in this case half, rather than a quarter, of the progeny should be pure pinkish. In order to show by an actual test that the gene for pinkish is in the third chromosome, it was decided to take advantage of the fact of no crossing over in the male and to run a back-cross test of a male heterozygous for pinkish and for the dominant third-chromosome character dichaete. It was now realized that the back-cross tests of females heterozygous for pinkish and black had not excluded the possibility of pinkish being in the second chromosome, though they had shown that, if so, it could be only in one or the other end-region and not near black. Ac- cordingly, exactly the same procedure was followed as in the tests for the location of cream b, that is, a pinkish female was outcrossed to a male which had the dominant second-chromo- some star as well as dichaete. The F, eosin star dichaete males were then back-crossed to pinkish females. The result showed (table 23) that the gene for pinkish is in the second and not the third chromosome; for, as well as could be judged, none of the star flies were pinkish, while all the not-stars seemed to be pink- ish, and dichaete was present in half of both the star and the pinkish classes. In the light of this test, and from the fact that there was about 50 per cent of crossing over between black and pinkish, we could place pinkish in either the extreme left or the extreme right end- - THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No. 3 368 CALVIN B. BRIDGES TABLE 23 The B.C. offspring given by F, eosin star dichaete sons, from the outcross of a pinkish female to a star dichaete male, when back-crossed to pinkish females NON-CROSSOVERS CROSSOVERS (IN MALE) 1916, 8/25 : Lid : (Eosin) | é G0 , Pinkish| Star E star | 4,St8%,,| Pinkish | dichaete| pinkish pinkie (Eosin) | dichaete sono (onsets PO Minhitte: piiap rio a On aoe) 0 Tibi tes a 17 soeptABs.(\y a6!’ 02268). Ode, SEROMA 0 9 22 on ig) or]! gt iy are as e 20° Son? air /ivigg? *| \” GRuatict Cone ee Totally a «os a9 6s. | a5 v8’ | oot region of the second chromosome. Fortunately, one advantage of the test just described is that it left us in possession of females heterozygous for star and for pinkish, and a back-cross test showed (table 24) that there is very free crossing over between star and pinkish. Pinkish is known, therefore, to be in the right-hand end of the second chromosome in the neighborhood of arc, speck, balloon, etc. Had the test given almost no crossing over between star and pinkish, we should have known that the gene for pinkish was in the left end, but this was not the case. A test as to whether the pinkish gene would have any visible effect in the absence of eosin showed (table 25) that in a very small per cent of the flies homozygous for pinkish there is a very slight dilution. This dilution is, however, so slight that rarely could one be sure that the effect observed is due to dilution rather than to the slight normal fluctuation of the red. TABLE 24 The B.C. offspring given by a star female from table 23 when back-crossed to a pinkish male : | NON-CROSSOVERS CROSSOVERS 1916, 9/23 a (Eosin) star Pinkish Star pinkish (Eosin) Q 19 30 19 20 saor{ § 26 26 19 16 eC | EYE COLOR IN DROSOPHILA MELANOGASTER 369 TABLE 25 The F2 offspring given by the F, wild-type females and F, eosin males, from the out- cross of pinkish females to wild males (NOT-EOSIN) 1916, 10/28 WILD-TYPE PINKISH) EOSIN PINKISH 5678 121 5) 76 24 5680 52 2 46 17 5703 63 14 59 18 5704 57 6 52 26 5705 8 3 67 18 sR Gtet le woe- sare 373 30 300 103 CREAM C While looking over the eosin stock (July 13, 1916) in search of a virgin female, I noticed that a few of the flies seemed to be unusually light in eye color. If this paleness were of genetic origin, then the gene must be of recent mutation, for on two occasions subsequent to the discovery of creams whose origin might be traced back to the eosin stock, this stock had been bred in such a way as to make it extremely unlikely that any cream or other eye-color mutation then present could fail to be eliminated. A pale eye color due to a mutation within the eosin stock might be an allelomorph of eosin, a specific modifier of eosin, or a non-specific modifier, such as vermilion or pink. The gene for the modifier might be sex-linked or be in any of the other three chromosomes. Only one experiment was carried out with this pale eye color, but this experiment was so planned that it answered all the above points. One of the pale females was mated to a star dichaete male. The F, offspring showed at once that the pale color was due neither to an allelomorph of eosin nor to a sex-linked modifier, for the sons were all standard eosin in eye color. Likewise the color was a strict recessive, both in eosin (the sons) and in red (the daughters). For the production of F:, F; wild-type females and F, eosin star dichaete males were mated together. The first point ob- served in F, was that the gene produced no visible effect by itself, for in that half of the offspring that were not-eosin none of the 370 CALVIN B. BRIDGES eyes were of other than the normal red of the wild-type. On the other hand, among the eosin-eyed offspring about a quarter were of the light color, and this without noticeable sex limitation, that is, the females and males were diluted to the same propor- tionate extent. On analyzing the distribution of the cream with respect to the second chromosome dominant star and the third chromosome dominant dichaete, it was seen that the gene for the cream is contained in the second chromosome, for the cream appeared only in the flies that were not star, all the flies being either star or cream and none being both star and cream or neither, which condition is in accord with expectation from the lack of crossing over in the male. The 2 :2:0:0 ratio observed TABLE 26 The F, offspring given by F, wild-type females and F , eosin star dichaete males from the outcross of a cream c female to a star dichaete male / (we) , ec e) | (we) ve) | ae 1916, 10/20 3! S’ 4 D’ we sy (we) ‘ we we me Ss D S' , Cre Ze 1D Cre D D Cre Db’ 5088! 26 21 32 . 23 Pa Werte age Va Een LE 14., 15 0 56314 40 32 3334 40 42 | 18 16 10 8 0 5654 il Wav 32 8644 AQ Aoial oie ey ey) 21 23 0 Total..... 83 90 97 101 | 103 119 | 50 44 45 46 0 1 Culture 5588 and 5631 produced ‘apterous’as the result of a fresh mutation which took place in the cream c stock (Metz, Am. Nat., 14, pp. 675-692). in the case of star and cream c is in sharp contrast to the 1:1:1:1 ratio in the case of cream ¢ and dichaete in the same experiment. There were as many creams among the dichaetes as among the not-dichaetes, as is expected from the free assortment of genes in different chromosomes. Up to this point there had been no confusion possible among the various modifiers; for the effects produced had been different in one or more of the following respects: in degree, in color tone, in dominance, in specificity, in interaction with other mutants, in the chromosome concerned, or in the locus within that chromo- some. In the case of cream c, there was the chance of confusion, since in tone cream c is not different from cream b, though in EYE COLOR IN DROSOPHILA MELANOGASTER BYE the amount of dilution (degree) there seems to be a very slight difference. The fact that they are both in the same chromosome (Il) made essential a further test of their distinctness. The method of linkage (a comparison of the star cream c value with the star cream b value) would be crucial only in case the values were quite different. An easier method, and one which at the same time disposes of the question of allelomorphism is to make a cross between the two types. When this was done (culture 5721) it was found that the F, flies were standard eosin in color, as is expected from the cross of two distinct recessives. Evi- dently cream c is neither cream b nor an allelomorph of cream b, and this is the only case where confusion was possible. SUMMARY The significant facts with respect to the origin and the inheri- tance of these modifiers are summarized in table 27 in which the primary arrangement is according to the order in which the modifiers were discovered. The first four modifiers appeared in rapid succession in the summer and fall of 1913 and were worked out almost simultaneously. In 1914 three more were found, TABLE 27 Summary of the origin and inheritance of the specific modifiers of eosin 3 ORIGIN \ 2 MUTANT 8 | 22 | Locus a Fin der Date Stock Type Culture| 2 = Cream a.....| Cra | Bridges |1913, 7/15 |Non-disjunc-| w or we |n 43 — | -- tion Cream II....) Grn | Bridges |1913, 9/15 |Lethal 2 W — |II} — Dark........| — | Bridges |1913, 9/23 |Non-disjunc-| w or we jn 100 | — | — | tion experi- ments Whiting......| — | Bridges |1913,11/21|Non-disjunc-| w or we | M67) — | — tion experi- ments Cream III. ..| Crm} Wallace |1914, 2/27 |Lethal la Witon wth 4 | LZ. Creamb.....) Crh | Bridges /1914, 3/10 |Non-disjune-| Wand w*| 82 | II (22.5 tion experi- ments Pinkish......] — | Bridges |1914, 7/27 |Eosin we — | II 106+ Creamc.....| Cro | Bridges |1916, 7 /21 |Eosin we — |II| — 372 CALVIN B. BRIDGES and another in 1916. On various other occasions dilute eosins have been observed whose inheritance has not been followed in detail.? Origin of modifiers. All of the modifiers were first detected in stocks or experiments involving eosin. This fact is not the result of any influence of eosin upon mutation, but has a simple explanation in the fact that the modifiers. produce little or no visible effect except when brought into coaction with eosin, and | hence they pass undetected no matter how numerous, or what their origin, until this condition is satisfied. Only the last two modifiers, ‘pinkish’ and ‘cream c¢,’ were found in stocks pure for eosin, and in these cases the mutation might have occurred previ- ously and have been incorporated with the eosin stock. The other six modifiers were found in experimental cultures in which only half of the flies were eosin; and in these cases it is usually not possible to say whether the modifier originated in or was introduced through the eosin or the not-eosin half of the experi- ment. In the case of cream II, there is good evidence that the mutation had occurred in a wild stock and had lurked there undetected until the cross to eosin brought it out of hiding. Cream b seems to have been present in both parent stocks. Creams a and IJ, dark and whiting, may have arisen in either, though the probability is on the side of the not-eosin parent, since these modifications were likely to have been detected if present in the eosin parent stocks. The scale of modifications. A graphic representation and com- parison of the color differences produced by the modifiers is given in diagram 1. In constructing this diagram two standards were chosen, namely, the colorless eye produced by the interaction of eosin and ‘whiting’, and the unmodified eosin. The modified eosins were then spaced along the line connecting these grades in pro- portion to their intensities of color. That is, homozygous cream II is the lightest of the other modifications and ‘dark’ is the darkest. Pinkish is the weakest of the dilutors of eosin. Creams 7 In 1918 two other creams have been found, and the inheritance of one of these presents features of exceptional interest for the chromosome theory of heredity. EYE COLOR IN DROSOPHILA MELANOGASTER 373 ce, b, and a are quite similar in color, b and ec being especially alike. Cream III is about half-way in color between eosin and white. The grades represent, as near as could be approximated, the mode of each color modification; there is in each case some fluctuation in the direction away from the eosin grade and a greater range in the direction toward eosin. Since the eosin FEMALES MALES Dark Eosin Pinkish eae ace W-W* dar Cream c Dark Cream b w-w? Eosin Cream a Cream II( het) Cream III Pinkish ream Cc Cream b Cream || Cream a Cream III Cream II Whiting Whiting Diagram 1. The scale of modifications of eosin in females and males (modal grades). male is of a lighter eye color than the eosin female, it became necessary to arrange two such series of grades in order to repre- sent the total range. After the separate distributions of color for the female and the male had been plotted, lines were drawn connecting the corresponding types in the male and female series. The surprising fact was then observed that all but one of the lines, if continued, would intersect the base line at about the 374 CALVIN B. BRIDGES same point. This was a striking revelation of a relationship vaguely realized before—that the modification by each of these genes is proportionately as great. in the male as in the female, although the actual modification is much greater in the case of the female. Sex limitation. ‘The above relation does not hold in the case of pinkish, for the connecting line cuts the base line at a point only about three fourths as distant as the others. The dilution of the pinkish male is proportionally greater than that of the pinkish female. It is Just in this respect that eosin itself differs from the other eye colors. The combination of these two ‘sex-limited’ characters, both of which differ in the same direction, increases the sexual dimorphism of the flies to a striking extent. Specificity. The primary characteristic of all these modifiers is that their visible effect upon eye color is dependent entirely upon, or is greatly increased by, the coaction of eosin. Of the eight modifiers, six (creams a, II, b, ce, dark, and whiting) are entirely dependent, giving no visible effect whatever in the absence of eosin, while pinkish gives a very slight dilution and cream III gives a dilution which is strong enough so that by avoiding certain adverse conditions there is the possibility of using it without the complication of eosin. In the case of pinkish the dilution is so slight that only in about a third of the flies is there any detectable difference. With both cream IIT and pink- ish the ease of separation and the sharpness of the difference is vastly greater in the presence of eosin. The term ‘disproportionate modifier’ might perhaps be better for such cases as those of cream III and pinkish. Most of our mutations are what may be called ‘general modifiers,’ since their effects seem to be independent of one another, and the combined effect is cumulative and roughly proportionate. “General modi- fiers’ may be represented by the familiar parallelogram, in’ which the initial condition (wild-type) is one corner, the effect of each gene acting independently is represented by the length and di- rection of the two adjacent sides, and the combined effect (double mutant type) is the opposite corner (diagram 2, a). In the case of the completely specific modifiers, such as cream II, the length EYE COLOR IN DROSOPHILA MELANOGASTER 379 of the side between the wild-type and (not-eosin) cream Il corners is zero, so that the parallelogram becomes reduced to a triangle (diagram 2, b). In the case of pinkish and cream Ill there is an intermediate condition in which the side between eosin and the double type is disproportionately far greater than the distance between the wild-type and the simple modifier (diagram 2, ¢). There is another type of disproportionate modi- fication exemplified by eosin sepia which might be called reversed. a, General modifier b, Specific modifier C, Disproportionate modifier Wild-type Wild-type, cream II Wild-type *s ~Cream III Eosin Pink Eosin / Eosin : ‘Eosin cream II ‘Eosin cream II Eosin pink d, Reversed modifier e, Non-modifier Wild-type Sepia Wild-type Ruby Pink Eosin Ruby pink *\\ Eosin sepia \ Diagram 2. A graphical representation of some types of modification Sepia is considerably darker than the wild-type, becoming a deep blue-black in old flies. One might expect that sepia would cause a proportional darkening of eosin so that eosin sepia would be as much darker than eosin as sepia is darker than the wild- type. This is not the condition that actually obtains. The eosin sepia double form is lighter than eosin, about as represented in diagram 2, d. There are several other curious types of dis- proportionate modification. Thus, for example, ruby (sex- linked) may be described as a ‘non-modifier’ of pink. The effects of these two genes are in the same direction and of like amount, 376 CALVIN B. BRIDGES but they fail to have a cumulative effect, the double type being practically indistinguishable from either single type instead of being as much lighter as each is lighter than the wild-type (dia- gram 2, e). It has been shown for the case of cream II that specificity of a very extreme nature obtains; since cream II is incapable of modifying any of the other mutant eye colors tried in place of eosin. The double recessives (not-eosin) cream II vermilion, and (not-eosin) cream II pink are the same in eye color as simple vermilion and pink, respectively. In the case of whiting the specificity is even more striking; for cherry, an allelomorph of eosin so similar as to be distinguished’ (in the females) only with great difficulty, is entirely undiluted by the whiting gene. Tone. Pinkish is aberrant in still a third respect: the other modifications can all be described as grades of dilution of the yellow-pink color of eosin, the lighter grades being especially dilute in the pink component of the color. Pinkish, on the other hand, gives one the impression that the yellow component has faded more than the pink, so that there remains a slightly greater proportion of pink in pinkish than in eosin or in the other modifications. Dominance. The only one of the modifiers which gives an appreciable effect in heterozygous form is cream II. The amount of dilution of eosin due to heterozygous cream II is of about the same grade that the weakest of the other modifiers, namely, pinkish, gives when homozygous. Fluctuation. The amount of fluctuation in the eye color of the creams due to differences in‘ age, food, etc., was about the same as that observed in the other eye colors and their combi- nations. This question is of some practical importance in the making of classifications and is of theoretical interest in con- nection with the completeness of the seriation obtainable with a relatively few modifiers. Viability. All of these creams are of excellent viability and the observed ratios are very close approximations to expectation. Chromosome. Because of the difficulty of putting the modi- fiers to practical use, it was not at first thought worth while to EYE COLOR IN DROSOPHILA MELANOGASTER 377 determine in which chromosome the gene for a given modifier lies. This information is therefore lacking in the case of cream a, dark, and whiting. Of the five other modifiers, one, cream III, is in the third chromosome, and the other four are in the second. The presence of four modifiers of eosin eye color in the second chromosome is in contrast to, or perhaps: supplementary to, the relative fewness of ordinary eye-color mutations in that chromosome. None of the modifiers were sex-linked, but this is probably not of special significance. Locus. The locus within the second chromosome has been determined for two of the four second chromosome modifiers of eosin, namely, pinkish at about 106, and cream b at 22.5. The locus of cream III has not been found directly because of the presence of C,, in the cream III stock. In heterozygous C,, there has been found to be 4.2 per cent of crossing over between cream III and dichaete, and 5.4 per cent of crossing over between cream III and ebony. This corresponds to a locus on the normal map of about 3.1 units to the left of dichaete or at about 7.9 units to the right of sepia which is the zero-point of the map of the IIl chromoscme. DISCUSSION The facts of the inheritance of these specific modifiers show that each is a definite Mendelian gene on the same footing with the whole body of genes known in Drosophila. Thus, they display clean-cut segregation from their allelomorphs; each is located in and transmitted by a specific chromosome of the Drosophila complex; they give free assortment with genes located in other chromosomes; and with genes located in the same chro- mosome they show linkage, with crossing over corresponding to a “fixed locus. Each of these genes arose by mutation—‘a specific change in the hereditary constitution’”—by the transformation of the materials of a particular locus into a new form having a different effect upon the developmental processes. These specific modifications are clear and simple cases of ‘multiple genes.’ Each is the result of the coaction of a specific 378 CALVIN B. BRIDGES modifying gene (cream a, b, etc.), which by itself produces little or no visible effect, and of a particular gene (eosin) that is necessary as a ‘base’ or ‘differentiator.’ The scale of the modifications of eosin produced by these several modifiers ranges on the one hand to a deep pink darker than eosin, and on the other hand to a pure white. This scale is purely artificial and descriptive, for these modifiers were entirely independent of one another in origin. Furthermore, the order of their occurrence bears only a random relationship to the dark-light seriation. By judicious combination of several such simple modifiers a multiple heterozygous stock could be obtained which would be amenable to selection and which would offer upon analysis a satisfactory parallel to such a case as that shown by Castle’s rats. The first result of selection in such a heterogeneous stock in the direction of lighter forms would be to pick out individuals homozygous for one or more of the modifiers and probably heterozygous for others. These different individuals of course would not necessarily be homozygous for the same factors, and therefore the population might still remain heterogeneous for these factors. Continued selection would result in a greater and greater degree of homozygosity and homogeneity and a consequent slowing down of the speed of the progression of the population in the direction of selection. The grade of the form reached when the population is homozygous and homogeneous would depend on the number and character of the particular modifiers in the initial population. ‘the fact must not be lost sight of'that there is not one lota of evidence to show that either the rate or the direction of the mutation processes that are characteristic of the species are altered by such selection. During the course of any selection _ experiment modifiers and other mutations should arise at the normal rate. In Drosophila roughly 25 per cent of the mutations found were wing or venation characters, 16 per cent body-color characters, and 20 per cent eye colors. Only eye-color mutations would have any effect upon the progress of our selection. The eye-color mutations found in Drosophila have been roughly in EYE COLOR IN DROSOPHILA MELANOGASTER 379 the proportion of 60 per cent of general or non-specific modifiers of eosin, such as vermilion and pink, 22 per cent specific modifiers of eosin, and 18 per cent allelomorphs of eosin. The proportion in which mutations are found should, of course, be distinguished from the proportion in which mutations arise. In this particular example the difference should be considerable, for since all the flies are eosin, the situation is particularly favorable for the detection of specific and disproportionate modifiers of eosin, and the percentage found should be correspondingly higher than in our general work where eosin flies constitute only a small fraction. It is probable that mutation is very much more frequent than appears, since a great many mutations are of very slight somatic effect and would pass undetected except that certain characters, such as eosin eye color, truncate wings, beaded wings, and a few others, are peculiarly sensitive differentiators for eye-color and wing-shape genes, etc. Mutations capable of affecting eye color, wing shape, etc., are presumably not less frequent of origin in ordinary stocks, such as pink eye or rudimentary wings, but these latter are poor reagents for the detection of mutations. Whether or not a specific modifier should arise during the course of selection depends on the length of the selection, that is, on the number of individuals, as well as on the frequency of that type of mutation. The previous. eye colors (all types) may be roughly classified as 90 per cent dilutors and 10 per cent darkeners of eye color which is about the same proportion as that found among the specific modifiers themselves. Cases like that of eosin sepia show that the ratio of darkeners to dilutors considered from the basis of eosin may be far different from that shown by the pri- mary effects of these same genes (differences from the wild-type). Whatever this normal ratio for eosin is, it should obtain in the selection experiment as well. When mutation in the direction of selection occurs, there should be a jump in the speed of pro- gression, and the final grade should be correspondingly more extreme. However, the modifier might be a darkener, but in that case it would not be selected and consequently would have only negligible effect upon the speed and the resultant grade. 380 CALVIN B. BRIDGES No one can deny that progress under selection is theoretically possible by repeated mutation in a single locus. But to accept that as the actual mode in any particular case demands specific proof. Such a hypothesis should not be considered until it has first been demonstrated that the initial constitution of the stock was such as to require the assumption of further acts of mutation. If a fresh act of mutation is required, then adequate proof must be submitted before it can be accepted that this mutation is in the particular locus favored rather than in one or another of the numerous other possible loci. Our experience with many cases of successful selection in Drosophila has been that even in this form, where the work is aided by such special features as a very small number of chromosomes, by absence of crossing over in the male, and by a knowledge of the initial constitution of the stocks that is not paralleled in other forms, it is often a matter of some difficulty to prove that a particular modification arose during the course of an experiment rather than that it was present in and introduced through one of the parents. How much more difficult, then, would it be to prove in a form where the tests are far less precise and on a relatively small scale that all progress observed during the course of selection is due to the occurrence step by step of fresh mutations? And how much more precarious would it be to affirm that these fresh mutations are all of a single locus, when one remembers that there are probably hundreds or even thousands of loci in which a mutation would have effect upon any given character. The evidence in the case of the creams is diametrically opposed to such a type of explanation, since the creams are manifestly of discontinuous origin in as many distinct loci as there are diverse modifications. The similarity between the creams and such a case as that of the rats must be attributed to the presence in the rats of the only con- dition met with in the creams, namely, diverse mutation. In the absence of rigorous tests, no one is justified in assuming that modifiers are not present in any given stock. Not only have the many selections carried out on Drosophila led ‘experi- mentally to such a conclusion, but it follows from a consideration of the facts of mutation. In the carefully pedigreed experi- EYE COLOR IN DROSOPHILA MELANOGASTER 381 ments of Drosophila we continually observe mutations arising, and are often able to locate the time and place of origin to a particular fly and to within one or a few cell generations. Every- one who believes that mutations are now occurring in his stocks in this manner must also accept the probable occurrence of such mutations in the immediate ancestry of his stocks, since it is not likely that either the rate or tendency of mutation have changed within any period with which we deal. The process of mutation thus gives rise in any stock to a complexity and heterogeneity which is only that ‘static condition of diversity’ said by Jennings (Jennings, ’17) to obtain for any species or kind of organism, such, for example, as the hooded rat. When selection is started in such a stock immediate progress should be expected and this should continue until the stock is homozygous for all the genes capable of modifying the selected character. One cannot be certain that observed progress is due to fresh mutation until selection has been continued long enough so that the initial diversity has been removed. The initial diversity means that mutations had previously occurred in various loci, and the presumption is that during the progress of selection the stock will not cease mutating in diverse. loci and henceforth mutate solely in one locus. Muller (’14), MacDowell (716), Sturtevant (18), and others have shown that the results obtained with selection in the hooded rats are, aside from one definite new mutation, exactly those expected from simple se- lection, in a heterogeneous stock of the kind supposed by the Frenchman Jordan, subscribed to by Jennings, and easily ex- plained as the result of previous diverse mutations. It is inter- esting to observe that the list of cases in which unilocal mutations or contamination may be seriously considered is dwindling with the progress of exact knowledge. In our opinion, the attempted distinctions between ‘saltations,’ ‘mutations,’ and ‘variations of slight degree’ have led rather to confusion of thought than to clearer thinking. To us these are all a single class, ‘mutations,’ and the term carries no restrictions of degree, covering the most extreme as well as the slightest detectable inherited variation. Distinctions of degree when ap- 382 CALVIN B. BRIDGES plied to mutations depend largely on circumstances and _per- sonalities, and are correspondingly inexact. Thus, students be- ginning work on Drosophila usually complain that there is almost no difference between pink eye color, for example, and the wild form, so that the classifications are made with considerable un- certainty. With more experience these same workers come to regard this difference as very great and are astonished that they should ever have thought otherwise. Again, some workers are able to distinguish with assurance characters that quite pass the resolving power of other workers seemingly equally experienced. In the Drosophila work great numbers of mutations have been encountered whose somatic effect may be roughly described as ‘shght.’ If degree of effect were marked off as abscissae and number of mutations that actually arise as ordinates, it seems probable that the highest point of the curve would be at the least extreme mutations and the curve would fall rapidly and gradually with the most extreme mutations. For most of the problems in which we have been interested precision of classification is essential, and precision is afforded only by the more extreme mutations and by some few of slight degree but definite character. Accordingly, there has been scant mention of the many ‘slight’ mutations in our accounts, which have dealt in the main with problems in which such characters were of no use as working tools. Likewise, when these mutations are found, they incite little desire to work out the facts concerning their inheritance, chromosome, locus, interactions, etc. However, enough of them have been investigated thoroughly for us to be certain that no departure from normal Mendelian inheritance is involved. Ac- counts of some of these ‘poor’ characters have appeared in Car- negie Publication no. 237, which deals with mutants whose locus is in the first or X chromosome. Thus, for example, dot, bow, depressed, green, chrome, and facet may all be fairly described as of ‘slight’ somatic effect. Accounts of many others will ap- pear in the publications dealing with the second, third, and fourth groups. Even a larger number will be treated.in a section on miscellaneous mutations, while scores will never be referred to at all. EYE COLOR IN DROSOPHILA MELANOGASTER 383 It seems probable that the bulk of the mutations that have been permanent contributions to evolution have been those of slight somatic change. Any organism as it now exists must be regarded as a very complex physicochemical machine with deli- cate adjustments of part to part. Any haphazard change made in this mechanism would almost certainly result in a decrease of efficiency. The greater the extent of the change the more certain the injury, not simply that the particular part is injured more, but also that a disproportionately greater number of ad- justments (morphogenetic, physiological, and ecological) are disturbed or destroyed. Only an extremely small proportion of mutations may be expected to improve a part or the inter- relation of parts in such a way that the fitness of the whole organism for its available environments is increased. ‘Thus, that length of trunk which is most advantageous for elephants in a given environment might be attained by a single mutation; but in this case it is likely that the trunk would be out of balance with the other structures (physiologically or otherwise) so that all the individuals possessing this feature might become extinct before the appearance of changes in these other parts that would make the trunk a real success. On the other hand, a mutation which made only a slight change in the favorable direction would require less extensive supporting changes in the related structures, besides which it might be of immediate advantage. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, No.3 384 CALVIN B. BRIDGES LITERATURE CITED Bripces, C. B. 1916 Non-disjunction as proof of the chromosome theory of heredity. Genetics, vol. 1, pp. 1-52, 107-163. Bripces, C. B., anp T. H. Moraan. 1919 (in press). The second chromosome group of linked genes. Carnegie Inst. Pub. no. 287, part II. Bripces, C. B., anp A. H.Srurtevant. 1914 A new genein the second chromo- some of Drosophila and some considerations on differential viability. Biol. Bull., vol. 26, pp. 205-212. CastLe, W.C., anp J.C. Pinups. 1914 Piebald rats and selection. Carnegie Inst. Pub. no. 195, 56 pp., 3 plates. JENNINGS, H.S. 1917 Observed changes in hereditary characters in relation to evolution. Jour. Wash. Acad. Sci., vol. 7, pp. 281-301. 1917 Modifying factors and multiple allelomorphs in relation to the results of selection. Am. Nat., vol. 51, pp. 301-306. MacDowe tt, E.C. 1916 Piebald rats and multiple factors. Am. Nat., vol. 50, pp. 000-000. Metz,C.W. 1914 Anapterous Drosophila and its genetic behavior. Am. Nat., vol. 48, pp. 674-711. Moraan, T.H. 1912 Further experiments with mutations in eye color of Droso- phila. The loss of the orange factor. Jour. Acad. Nat. Sci. Phil., vol. 15, pp. 321-346. 1915 The réle of the environment in the realization of a sex-linked Mendelian character in Drosophila. Am. Nat., vol. 49, pp. 385-429. 1916 A critique of the theory of evolution. Princeton University Press, 195 pp. Moraan, T. H., anp C. B. Briparss. 1913 Dilution effects and bicolorism in certain eye colors of Drosophila. Jour. Exp. Zool., vol. 15, pp. 429-466. 1916 Sex-linked inheritance in Drosophila. Carnegie Inst. Pub. no. 237, 87 pp. + 2 plates. Morgan, STURTEVANT, MULLER AND Bripeges. 1915 The Mechanism of Men- delian heredity. Henry Holt & Co., 262 pp. Mutter, H. J. 1914 The bearing of the selection experiments of Castle and Phillips on the variability of genes. Am. Nat., vol. 48, pp. 567-476. 1914 The mechanism of crossing over. Am. Nat., vol. 50, pp. 193-221, 284-305, 350-366, 421-434. Sturtevant, A. H. 1913 A third group of linked genes in Drosophila ampelo- phila. Science, vol. 37, p. 990. 1918 An analysis of the effects of selection. Carnegie Inst. Pub. on. 264, 67 pp. + 1 plate. 1919 (in press) Inherited linkage variations in the second chromosome. Carnegie Inst. Pub. no. 287, part III. PPT 93, hed hoi, aren iia y ee v9 vinta ea? 8 ne el tosh nee t. ear ly yi ; ern en “wots j ce Papin oh POT srt rehab? APE: i. am i Resumen por el autor, C. H. Danforth. Pruebas de que las células germinativas estan sujetas a selecci6n, basadas en sus potencialidades genéticas El autor ha emprendido una serie de experimentos para probar la posible existencia de diferencias funcionales en las células germinativas dependientes o relacionadas con la naturaleza de los ‘‘determinantes” transportados por el plasma germinativo. La existencia de tales diferencias se demostr6 por la accién de un agente selectivo actuando sobre las células germinativas. Este agente fué el vapor de alcohol, que al inhalarse por los pulmones se cree que pasa directamente al torrente circulatorio, y desde este a los tejidos germinativos; se administré a gallinas de la constitucién genética escogida. Como indice de cualquier selecci6n que pudiera ocurrir la proporcién relativa de ciertos rasgos, braquidactilia, polidactilia y color blanco, que aparecieron en la prole producida durante los periodos de tratamiento, se compar6é con la proporcién de los mismos rasgos producidos durante periodos en los cuales dicho tratamiento fué suspendido. Los resultados obtenidos indican que por lo menos en el caso de ciertos rasgos la seleccién es posible y que es mas rigurosa cuanto mas severo es el tratamiento. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER JSSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 1 EVIDENCE THAT GERM CELLS ARE SUBJECT TO SELECTION ON THE BASIS OF THEIR GENETIC POTENTIALITIES C. H. DANFORTH Department of Anatomy, Washington University School of Medicine INTRODUCTION A great deal of evidence goes to show that individual dif- ferences among adult animals are commonly due to differences in the germ cells from which the animals have developed. In other words, peculiarities in the adult are, in general, indications of peculiarities in the germplasm. It would seem to follow as an obvious corollary of this that the germ cells of a species must be of almost as many classes as are the adult individuals. This being the case, it is perhaps not unreasonable to expect that germ cells should differ more or less widely in their responses to varying conditions and in their ability to function in the production of zygotes. If such an assumption is shown to be justified, it might be invoked to explain several obscure evolu- tionary tendencies, as well as certain constant departures from normal Mendelian expectations. As a first step in determining whether or not germ cells carry- ing factors for a given unit character differ in their physiological responses from germ cells carrying factors for an allelomorph of that character, it seems desirable to work with the sperm or eggs of a heterozygous animal in which the normal proportion of each class of germ cells produced has already been determined. With such an animal all the germ cells are presumably subjected to the same influences, except as those influenced are varied by the experimenter, and consequently a much better control is possible than could be had by comparing results from different homozygous individuals. 385 386 Cc. H. DANFORTH The choice of a method of procedure presents a number of difficulties. Attempts to bring special influences to bear on cells within the living body are attended by many uncertainties and, on the other hand, the treatment of suitable germ cells outside the body in most cases requires the development of a special and difficult technique. The desideratum would seem to be the possession of a means of administering definite amounts of reagents to the germ cells without otherwise disturbing their normal environment. Despite the fact that it does not lend itself readily to quantitative investigations, the inhalation method devised by Stockard seems to be the most satisfactory thus far employed. A considerable number of reagents are suitable for use by this method, but ethyl alcohol is the one that has been most completely tested (Stockard, ’13; Cole and Davis,’14; Stockard and Papanicolaou, ’16; Nice, ’17; Pearl, 7171). For these reasons ethyl alcohol administered by the inhalation method was employed in the experiments reported in this paper. Pearl (’17) has interpreted results obtained by himself and others as indicating that inhaled alcohol affects germ cells dif- ferently according to their vitality, destroying some, injur- ing some permanently, and producing no more than a tran- sient effect on others. In the case of the fowl, he believes that practically maximum treatment with alcohol and certain other vapors results in an elimination of weaker germs with a con- sequent rise in the average vigor of the chicks actually pro- duced. In other words, treating the parent with alcohol vapor results in a selection of germs in favor of those which produce the most vigorous young. It was not shown in his experiments that any characteristics other than those dependent on general vigor were affected by the treatment, the Mendelian ratios for the few characters involved being essentially the same in both experiment and control. At the time when Pearl’s papers appeared the writer was Just concluding a breeding experiment with poultry in which several 1 An extended list of titles relating to the whole subject may be found in the paper by Pearl, ’17. EVIDENCE OF GERM CELL SELECTION 387 Mendelian characters were being investigated. Since some of these characters were not represented in Pearl’s material, and the writer was at the time casting about for a method to test for germ-cell selection, it seemed worth while to continue breed- ing from the same birds with former conditions maintained except for the introduction of alcohol treatment. The work was planned in no sense as a repetition of Pearl’s experiments, but rather as an attempt to extend his methods to a few characters not known to be directly dependent upon inherent vitality. The writer takes this opportunity of acknowledging his in- debtedness to the valuable and suggestive paper mentioned above. The first experiment, 1, was made in the spring of 1917 with results that seemed to point to a selective action of the alcohol vapor. The matter is one of such importance, however, that a repetition of the work seemed desirable before publication of the data. Consequently three further experiments, 2, 3, and 4, were carried out in the spring of 1918. As will appear below, the results obtained in the latter work proved to be essentially consistent with the findings in the first experiment, which fact seems to justify their publication. GENERAL NATURE OF THE EXPERIMENTS In each experiment a cross was made between normal, pure- bred stock on the one hand and hybrid, heterozygous birds on the other. The hybrid individuals were subjected for certain periods to two daily treatments with alcohol vapor, during which time as far as possible all eggs laid were incubated (table 2). Since the original purpose of the work was to test the pessi- bility of selecting germ cells rather than of modifying them through influences brought to bear on the soma, each experl- ment is brief and intensive. A control was obtained in each case by saving the egos from the same flock, kept under as nearly identical conditions as possible, during a period before or after the aleohol experiment. All eggs that did not hatch were opened and the character of 388 Cc. H. DANFORTH the contents recorded. Since the eggs were candled frequently many embryos were obtained only a few days after death and consequently in a good state of preservation. Very few of the dead embryos were found to be too macerated to afford the desired data. Some living embryos were purposely taken. For the sake of brevity, data obtained from eggs laid during the periods of alcohol treatment will be referred to as A, those from eggs laid in the control periods as C. Each experiment, therefore, has an A and a C subdivision. In experiments 1, 2, and 3 the heterozygous individuals were males, in 4 they were females, The characteristics tested for their possible response to al- cohol vapor were brachydactyly, polydactyly, and eolor. The presence or absence of booting (feathers on the tarsi) and the apparent form of the comb were also recorded, but the former has been found to be only another manifestation of the factor which produces brachydactyly, while chicks that have been allowed to develop have shown that comb form cannot, in the material used, be accurately determined in embryos and young specimens. Sex was not recorded. The nature of the pe- culiarities under investigation may be briefly indicated. Brachydactyly is a condition described by Danforth (’19) which manifests itself in a more or less pronounced shorten- ing of digit IV of the foot. This shortening involves the length, and in extreme cases the number, of bones in the digit. Nearly all brachydactyl birds are booted, but there is a small percent- age that has unfeathered tarsi. Conversely, there is an occa- sional individual with feather tarsi which is not brachydactyl. Breeding experiments have seemed to establish the fact that these conditions are interchangeable from the point of view of heredity or, in other words, that they merely represent different expressions of one and the same underlying cause. Conse- quently, a single term may be used to cover all of these phae- notypic manifestations, although it must be borne in mind that in a certain small number of cases the word ‘brachydactyl’ used in this sense is not strictly literal. Brachydactyly may be recognized after about the tenth day of incubation. The EVIDENCE OF GERM CELL SELECTION 389 index of brachydactyly? was computed for all chicks that hatched and each booted individual was assigned to one of three grades based on the number of feathers present on the shanks and toes. Polydactyly, a well-known condition, seems to be the final manifestation of an early disturbance in the developing rudi- ment of digit I of the foot. The character is quite variable, ranging, in the present material, from a single enlarged hallux to a condition in which the hallux is replaced by two digits with even the occasional indication of a third. The cases may be arbitrarily grouped into three grades, dependent upon the degree to which the peculiarity manifests itself. Rarely poly- dactyly was found to extend to the wings. The term as here used is a misnomer to the extent that a certain number of in- dividuals which clearly manifest the fundamental character in question really have only the normal number of toes. Poly- dactyly can be recognized with certainty and probably in all cases after about the seventh day of incubation. Color in the present paper refers to only two grades, ‘black’ and ‘white’. While the pure-bred birds used were either clear snowy white (Leghorns) or uniform glossy black (Minorcas), a large percentage of the young of mixed ancestry showed some color indication of their hybrid origin. This was particularly true of the white chicks, many of which had one or more small spots of black down. By ‘white’ may be understood pure white down or feathers or white plumage with only a little black pigment in the form of a few dark spots. ‘Black’ covers all other shades, even though some individuals classified as black subsequently turned out to be barred or mottled. From the twelfth day of incubation (or even earlier) every chick could be put unhesitatingly in one or the other of these categories. * The index of brachydactyly is an arbitrary value obtained by dividing the sum of the lengths of the two fourth toes by the sum of the lengths of the two second toes. With brachydactyl chicks this gives a value equal to 1 or less, while with normal chicks the value obtained is more than 1. The division is carried to the second decimal place and the quotient then multiplied by 100 to eliminate fractions. 390 Cc. H. DANFORTH METHOD OF GIVING THE ALCOHOL TREATMENT Aleohol was administered by the inhalation method devised by Stockard (713) in his work with guinea-pigs and subsequently employed by Pearl (717) in the experiments with poultry. Pearl’s technique was followed in its essentials, except that owing to the conditions of the experiments it was possible to use a some- what simpler inhalation chamber. For experiments 1, 2 and 3 a glazed earthenware crock 20 inches high and 15 inches in diameter, with a measured capacity of 1.98 cubic feet, or about 56 liters, was fitted with a tight cover and a galvanized-iron ‘ false bottom. The false bottom was perforated by seventy- three 23-inch round holes and raised on legs 214 inches high. The cover was made of thick matched boards subsequently soaked in paraffin and padded along the surface of contact with the crock. A window 8 inches square was cut out of the center and covered by a piece of glass set in paraffin. When a treatment was to be given, the false bottom was tipped on edge and pieces of cotton soaked in 95 per cent ethyl alcohol were placed in the crock. A little additional alcohol was poured in, the false bottom replaced and the cover fitted over the top. At the end of from fifteen to twenty minutes, when the atmosphere was found to be saturated with alcohol vapor, the bird was quickly placed inside on the false bottom, the cover being raised as little, and for as short a time, as possible. Although frequent tests failed to reveal any odor of alcohol on the out- side, the whole chamber was generally covered by heavy cloths except when observations were being made through the window. Between each two treatments the crock was cleaned and aired. From 60 ec. to 70 ec. of 95 per cent aleohol were used for each treatment. The amount that vaporized was roughly determined in the following manner. Fresh cotton was soaked in alcohol and then squeezed as dry as possible; 66 cc. of alcohol were then poured over the cotton and on the bottom of the crock. At the end of an hour and thirty minutes, during which time a treatment was given, the alcohol on the bottom was sopped up and the pieces of cotton squeezed as dry as before. EVIDENCE OF GERM CELL SELECTION 391 In this way it was possible to recover 42 ec., leaving a balance of 24 cc., as the approximate amount that had evaporated. The absolute alcohol vaporized into the tank therefore prob- ably ranged between 20 cc., and 25 cc. for treatments last- ing about an hour. There was always a rise in temperature amounting to about 3.5°C. during an hour treatment (e.g., on May 21, from 26° to 29.5”), and this of course facilitated evapora- tion of the alcohol. The treatments were usually begun about 9 a.m. and 5 p.m. Their duration varied considerably, as will be noted below. With the hens used in experiment 4 a slightly different con- tainer was employed for administering the alcohol. Enameled specimen jars about 15 inches square and approximately 12 inches high, with close-fitting covers, were found to have a capacity of 1.6 cubic feet. These jars were used as inhalation chambers, and one or two hens placed in each after an excess of alcohol had been allowed to evaporate in them for from twenty to thirty minutes. REACTIONS OF THE ALCOHOLIZED FOWLS The degree of alcoholization obtained is perhaps best indi- cated by the observed responses of the birds to this treatment. While a certain amount of resistance or accommodation was acquired as the treatment progressed, it was apparent that the males were throughout less affected than were the females. The alcohol absorbed on each occasion was sufficient to cause at least a mild degree of intoxication. Since the reactions were very constant, the details of a single treatment may be given as illustrative of the typical behavior of one of the males fol- lowing the inhalation of aleohol vapor. On May 29, 1917, the inhalation chamber was prepared as usual and male no. 8 treated from 10 a.m. to 11.15 a.m. During the whole time he sat on the false bottom turning his head only when disturbed by objects seen through the window. When undisturbed the front of the head rested against the side of the jar. The nictitating membranes passed over the eyes at the rate of thirty times per minute. Notes on respiration were made as follows: 392 Cc. H. DANFORTH Time Rate Remarks 10.15 40 per minute Shallow. 10.25 25 per minute Deeper. 10.50 22 per minute Deep, regular. 115 22 per minute Deep, but irregular. At the end of the treatment the bird was breathing noisily and with difficulty. The mouth was open. 11.15. Removed from the tank. Quite passive. Did not struggle. 11.17. Returned to the breeding pen. Stepped very high, but not unsteadily. Showed indications of hyperexcitability. Shied at a feather and jumped out of the way of a ‘chick that approached from behind. The labored breathing improved rapidly. 11.20. Picked up pieces of grass and called hens. Very gallant. 11.22. Crowed repeatedly but abnormally. 11.25. Ate heartily when the flock was fed. This behavior was quite characteristic of the males follow- ing both morning and afternoon treatments. The ability to crow was always affected. Occasionally one was unsteady on his feet and they frequently misjudged distances in attempt- ing to jump or fly. Repeated rubbing of the eyes on the back was characteristic, but despite the severity of the treatment, no whitening of the conjunctiva was noted. As a control for the above-described behavior, on the after- noon of the same day the inhalation chamber was again thor- oughly cleaned and aired and the same male treated exactly as before except that no alcohol was used. He was kept in the chamber from 3.10 to 3.45 p.m. During the first half of this period he sat quietly on the bottom of the container, but during the latter half he stood, possibly owing to the gradual exhaustion of oxygen. Respiration fell from 42 to 30 per minute. The moment the cover was removed from the tank he made violent attempts to escape, struggling and squawking. (Nothing of this sort ever happened following an actual treatment.) Returned to the breeding pen he displayed none of the customary excit- ability, high-stepping and gallantry towards the hens. After a regular alcohol treatment given an hour later he again showed the usual reactions indicative of intoxication. It will be seen from the foregoing that the immediate phys- iological effect of the alcohol upon the males, while unmis- EVIDENCE OF GERM CELL SELECTION 393 takable, was nevertheless remarkably slight considering the strength of the treatment. It should be noted in passing that there is the possibility of a mild degree of asphyxiation, but this probably is of no importance in the present connection. The five hens used in experiment 4 were affected much more severely. They were commonly completely overcome and entirely unable to stand for several minutes after being released from the inhalation chamber. Staggering, which was rarely noticeable in the males, was of regular occurrence with these females. Despite precautions taken to the contrary, all but one of them ultimately died from the effects of the treatment. The one remaining individual developed, or had inherently, a considerable power of resistance. . Aside from the immediate but transient effects of the alcohol, nothing was noted except a slight loss in weight. For example, in the fifty-seven days from February 9 to April 7 no. 28 (treated), with an initial weight of 1803 grams lost 133 grams, while in the next fifty-seven days no. 28 (untreated) gained 160 grams, and no. 27 (treated) fell from 1825 grams to 1695 grams, a loss of 130 grams. DESCRIPTION OF THE EXPERIMENTS The method of administering the alcohol and the general character of the obvious physiological responses have already been indicated. It remains only to outline the special features of the individual experiments, of which a condensed summary is given in tables 1 and 2. The control period preceded the périod of treatment in experiments 1, 2 and 4, but in experiment 3 this relation was reversed. The surroundings and care of all the flocks during both A and C periods were made as constant as pos- sible. Eggs were put in the incubator in all cases, except 1-C, twice weekly and on definite days, so that no egg had been laid more than four days when incubation commenced. Experiment 1. (February 15, 1917 to June 2, 1917). The hetero- zygous parent (male no. 8) was raised at the laboratory from mongrel ancestors. His mother (no. 6), was a monster of the 394 Cc. H. DANFORTH Pygopagus parasiticus type, which, however, produced only normal young. She was a small hen with a rose comb and the plumage of a barred Plymouth Rock. She was not brachy- dactyl, polydactyl, nor booted. The father (no. 5) was white with a pea comb and unusually heavy feathering on the neck, suggesting Asiatic blood. He was of moderate size, polydactyl, and booted. This is all that can be said of the ancestry of no. 8, the individual which supplied the traits studied in this paper, but the nature of the experiments is such that a more com- plete pedigree would be of little additional value. Male no. 8, was hatched in 1914. His coloring was approximately that of a barred Plymouth Rock, but with the breast somewhat spotted and with more or less white in the hackle, saddle feathers, TABLE 1 Comparative statement of the conditions of each experiment DURATION IN DAYS DAILY DOSAGE (PART A) EXPERIMENT NUMBER Part A Part C Maximum Minimum Average 1 36 63 2h 25 me | lhe elm. leh Abas 2 65 34 PAN, Yfke Wyn Oa ln, Wes i. 3 58 63 Ang Oia, Wn, Ose || Pl, By sen, 4 43 64 Pin, 22m, pb lay ants || Pela) re 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 Bare uh tec ge” [incubated tga foepeee | ine chal |. Laraeed ; x 150 150 5 39 51 55 C 362 300 341 | 108 72 86 : A 191 180 2 is 153 14 C 106 104 1 19 69 15 4 i 151 150 35 14 60 41 C 166 155 49 12 60 34 , rn 39 39 0 6 13 20 C 151 150 2 45 76 O7 Totals: A and C { A 531 519 42 70 277 130 separately C 785 709 86! 184 277 162 Totals: A and C = dod \ A+C} 1816 1228 1281 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, ete. A. The aleohol 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, Le., brachydactyly (including booting), polydactyly, and white color. The homozygous parents were six single-combed black Mi- norea 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, Viz": 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. ©. 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 i1. 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 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. Alcohol 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. Alcohol 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 C. 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 * 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 sean Present ete 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 Cc 26 48 30 47 38 35 3 Av 52 45 52 48 53 37 C 38 51 37 53 48 7 4 A 18 14 (2) (31) 12 16 ” Cc 48 77 (28) (105) 63 55 Totals: A and C Tetales seen aa A 178 191 139 247 120! 108 separately Wigan ©) 181 283 168 352 1491 127 Totals: A and C Ginn AAS } A+C 359 474 307 599 269 235 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 463 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- eidence of brachydactyly under conditions such as obtain im 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 ‘brachydactyt TABLE 4 Percentage distribution of characteristics. Compare table 3 EXPERIMENT BRACHYDACTYL POLYDACTYL WHITE Number Part} Observed |: Expected Observed Expected Observed Expected 7 A | 46.7+3.1| 39+3.0 | 29.4=2.7| 3642.9 100 100 § 39.2=2.5| 39=2.5 | 33.2=2.1) 3642.2 100 100 9 { AG | A32a--o2 1) 09-5) Oat. 92,9). 302.9) | 5010-322) 50-=3.2 COirsbell slip 39a 8) | 09. 0-201) o0--oef | o2o1-—4, Ol 50-=o. 9 3 { A | 53.6-£3.4!| 39==3.3) |) 52/03) 4) 36323 || 59:9=3.5| 50=3:6 CaM 23 ooo or) | 40eU==35_5) 36-3. | 5605-=3,-0), 00--ont , { A) 5622-5, 9! 0395.1 42.8+6.3] 50+6.4 C | 38.4+2.9] 39+2.9 53.4+3.1) 50+3.0 Totals { A | 48.2+=1.8) 391.7 | 36.0=1.7| 361.7 | 52.7+2.2) 50+2.2 C | 39.0+1.5| 391.5 | 36.0=1.6) 361.6 | 53.9+2.0] 50+2.1 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 C. 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 aleohol 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 brachydacty], 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 C. H. DANFORTH an interpretation here would imply that the aleohol 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 alechol 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 cireumstanced 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 AruittT, ADA Hart, AND WELLS, H. Gipron 1917 The effect of alcohol on the reproductive tissue. Jour. Exp. Med., vol. 26, pp. 769-778. Coe, L. J., anp Bacnuuper, 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. Cote, 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. DanrortH, 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 aleohol on white mice. Am. Nat., vol. 51, pp. 596-607. PearRL, Raymond 1917 The experimental modification of germ cells. Jour. Exp. Zoél., vol. 22, part I, pp. 125-164; part II, pp. 165-186; part ITI, 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. StrockarD, C. R., AND PapanicoLtaou, G. 1916 A further analysis of the hered- itary transmission of degeneracy and deformities by the descendants of alecoholized 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 aleohol on treated guinea-pigs and their descendants. Jour. Exp. Zoél., vol. 26, pp. 119-226. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 3 Resumen por el autor, P. W. Whiting. Estudios genéticos sobre la mariposa de la harina, Ephestia kithniella Zeller Las nvestigaciones genéticas efectuadas con la mariposa de la harina, procedente del Mediterrdneo, han demostrado la herencia de diversas variaciones de forma y color. Asi, por ejemplo, los defectos de los palpos labiales se heredan irregular- mente. La lengua hendida se hereda como recesivo pero las mariposas que genéticamente contienen este factor aparecen normales somaticamente a causa de las condiciones del medio ambiente. El factor “oscuro,” como indica su nombre, oscurece el drea media de las alas anteriores y aclara la base y mdrgen externo. No se ha estudiado cuidadosamente pero depende probablemente de una diferencia de un solo factor con un hetero- aigoto intermedio y variable. El factor “‘sooty,’’ S, ennegrece la base y margen externo y aclara el drea media, produciendo de este modo un efecto algo contrario al de “oscuro.” Acttia como un dominante casi completo sobre el tipo, s; pero en cul- tivos que contienen individuos con el factor oscuro y otras vari- aciones menores, heredadas independientemente, presenta una inter-graduacién completa con el tipo. El factor ‘‘negro,’’ b, (cuerpo y alas) acttia como un recesivo simple y completo sobre el color gris tipico, B. En la mariposa negra homozigd6tica, el color ‘‘sooty” presenta dominancia invertida actiando como un recesivo. Para denotar esta particularidad los simbolos del locus para sooty se han invertido, sS.bb. En las mariposas sooty-negras homozigéticas, SS.bb, el efecto dilutivo del factor sooty aparece en el drea media. El indice diheterozigético es: 9 sooty, 3 tipo, 1 sooty negra y 3 negras. El autor propone una hipétesis para explicar la distribucién del pigmento en las alas, inversién de la dominancia, inter-graduacién de los caracteres por los efectos invertidos de factores que complican los resul- tados, ete., la cual supone la existencia de una accién ondula- toria de un agente productor de color y otro inhibidor del mismo. Los caracteres estudiados no demuestran la existencia de liga- miento u otras complicaciones en el mecanismo hereditario. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIVGRAPHIC 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 fe Untroduic hlonwerets es iets cis Sa ilae donee ets mecha. « spy. sistas yas © 414 MEDistribUiionsandstaxOnOMiypre erected: ceicies cele ests seals 414 18, Sowa OF tilae WOW ls lotsa cole nida eandoeus Shoba0odop0eoOdEeoDoS 416 GeRochMIGC sees Pat as Mee mean Senne tas ea uale cheigee wena 417 II. Observations, experimental data, and conclusions..................-.. 419 A. Description and origin of variations noted...................+-. 419 ee DescripuionsorvariablOUster well eyoes ote ole ee)elaye ale viele ate 419 ET COOTER oO en ATE Ce ee hate eh Tacit ale Mercere ae 419 DE SIZO sae ARS PR Lerch ia okeiaei a ae fessional hbase 3 419 Gepibe RES DIMES Erte chia he = = at | 25 ger at | 5055 at | 50 55> at 50 i daily 50 as *iaily 50 kv. daily | 50 kv. daily | 50 ae Husky 10 17 17 14 11 12 20 20 34 29 25 21 28 69 30 46 35 30 28 39 79 40 51 42 36 34 55 90 50 54 47 40 39 67 96 60 58 53 44 44 77 99 70 63 59 48 52 88 100 80 67 65 56 63 96 90 74. 74 69 79 98 100 84 83 S4 91 99 450 WHEELER P. DAVEY These readings, taken from the smooth curves of the graphs, do not differ from the actual experimental data by more than 1 per cent. Except while being x-rayed or counted, the beetles were kept in an incubator at 34 to 35°C. In order to make sure that the results were not affected by some possible ‘temperature co- efficient of life,’* the controls were taken out of the incubator while group JA was being rayed, and were kept out during the whole raying. Since group JA was rayed the longest time each day, this meant that the controls were cooled off for a longer time than groups IW, IX, IY, IZ. Therefore, if cooling off for a few minutes each day happened to tend to increase the length of life, then the controls were made to live longer than they otherwise would. The actual increase in length of life observed in groups IW, IX, and IY is, therefore, not due to any possible effect of temperature, but occurs in spite of it. After so many boxes of beetles in JA were dead that the time of raying group IZ was greater than the time of raying JA, the controls were kept out of the incubator while group IZ was being rayed. Some data not given in the graphs may be of additional in- terest. Each group was divided into two subgroups of about the same number of individuals each. It was found that the idiosyncrasy was great enough so that the curves of the corre- sponding subgroups could not be exactly superimposed. How- ever, it was found that this idiosyncrasy was always less than the changes in death rate caused by x-rays. By way of illustra- tion, table 2 shows the percentage of beetles dead in each sub- group, a) on the day when 50 per cent of the controls were dead; (b) on the day when 50 per cent of the x-rayed group were dead. This table shows that the lowest death rate among the controls (group TV) was higher than the highest death rate among the beetles of groups IW, IX, IY. It is interesting to note in this connection that the total dose received by these beetles was greatly in excess of that minimum dose which, when given all at once, would have caused pre- mature death. ? Loeb and Northrup, Proc. Nat. Acad. Sci., Aug., 1916. PROLONGATION OF LIFE DUE TO X-RAYS 451 TABLE 2 PER CENT OF TRIBOLIUM CONFUSUM DEAD GROUP Approximately 50 per cent controls Approximately 50 per cent x-rayed dead tribolium confusum dead Subgroup (1) | Subgroup (2) Subgroup (1) | Subgroup (2) 39th day 56th day iV AT 54.2 52.8 60.0 IW 41.6 42.1 48.1 D250 39th day 74th day IV 47.7 54.2 59.9 70.3 IX 32.4 38.3 44.7 54.1 39th day 67th day 1V 47.7 54.2 58.6 68 .2 JING olen 36.7 48.5 DOnZ 39th day 38th day LV ATE 54.2 46.8 53.4 IZ 52.9 54.5 50.1 52.4 39th day 14th day IV LN 7 54.2 epg 24.9 JA 88.5 91.7 A3 .7 46°8 A further analysis of the data of groups IV to JA will be of interest. The curves shown in figure 1 were each replotted on probability paper‘ (fig. 2). It was found that each curve was composed of portions of three accurate probability curves, joined end to end. It is as though there were three causes of death, or perhaps three definite groups of ages. These three portions of the death-rate curve will be termed A, B, and C. Portion C represents those beetles which lived the longest in their group. 4 The ordinates of probability paper are so spaced that the ordinary curve of the probability integral is represented by a straight line. 452 WHEELER P. DAVEY 80 FAESENT IN THE | 70 | ca v Fig. 2 PROLONGATION OF LIFE DUE TO X-RAYS 453 TABLE 3 ———— Se PER CENT DIED GROUP DAILY DOSE A B Cc IV Control 44 26 30 IW 4 32 36 32 IX 123 26 26 48 SY 25 21 35 44 IZ 50 23 61 16 JA 100 64 17 19 Table 3 gives the death rate per 100 in each group for A, B and C. It is evident that the smallest daily dose (group IW) decreases the death rate of ‘A’ and that those beetles which are kept from dying of ‘A’, die of ‘B.’ Deaths from cause ‘C’ are practically unaltered. A larger daily dose (group IX) causes about half of those which would normally die of ‘A’ to die of ‘CG.’ A still larger daily dose (group TY) causes half of those which would have died of ‘A’ to die of ‘B’ and ‘C.’ A still larger daily dose (group IZ) acts much like the previous dose in causing almost half of those which would have died of ‘A’ to die of "By but it differs from it in that some of those which would have died of ‘C’ are prematurely killed. The largest daily dose employed (group JA) caused about a third of those which would have died of ‘B’ and ‘C’ to die of ‘A.’ It is hard to interpret all this. It may be that life cannot exist except in the presence of a small amount of radio-activity. The radio-activity of the earth may not have been of the optimum value, so that some benefit was derived from the X-Tays received each day. The following is an effort at an alternative explanation. The evidence given by group JA shows that the lethal action of x-rays is tied up in some way with cause of death ‘A.’ It is well known that the lethal action of X-rays is more marked on cells in the process of division than on those in the resting state. Therefore, small daily doses (larger than a certain minimal value) can kill off those few cells which happen to be in a state of division at the time of raying. The death of 454 WHEELER P. DAVEY these few cells stimulates the production of more to take their places between the periods of raying. Therefore, small daily doses, instead of increasing the death rate from cause ‘A,’ actually decrease it by stimulating the processes of repair. The whole individual beetle, therefore, has a smaller chance of dy- ing from ‘A’ and is compelled to die of either ‘B’ or ‘C.’ When the daily dose is increased to such a value that the daily de- struction of cells is equal to or greater than the production of new cells, premature death occurs, from causes ‘B’ or ‘A’ (see groups IZ and JA). B. PROLONGATION OF LIFE DUE TO SMALL SINGLE DOSES OF X-RAYS Five groups of approximately 850 individuals each were taken. ‘These were known as groups JB, JC, JD, JE and JF. Group JB was the control. = AM Group JC was given 100 at 50 KV. — 50 M.A. MAI Group JD was given 200 Iv at 50 KV. — 50 M.A. Group JE was nee at 50 KV. — 50 M.A. Group JF was given 400 = at 50 KV. — 50 M.A. 252 The beetles were rather old, so that the controls were all dead on the fortieth day of the experiment. There were so few bee- tles still alive after the thirty-fifth day that the results of the last five days are not of the same order of accuracy as those of the first thirty-five days. MAM 2p? at 50 KV.) had the same death rate as the controls. After the tenth day the death rate was considerably less than that of the controls. The two groups were divided into two equal sub- groups, and although it was found that the idiosyncrasy was such that the subgroups were not exactly alike, still, after the The first ten days of the experiment, group JC (100 PROLONGATION OF LIFE DUE TO X-RAYS 455 tenth day, the highest death rate of JC was lower than the lowest death rate of the controls. During the first seventeen days of the experiment, group JD (200 MAM 25? After the seventeenth day the death rate of group JD was less than that of the controls. After the twentieth day the death rate of JD was identical with that of JC. When divided into two equal subgroups, as described above, it was found that after the twenty-second day the highest death rate of group JD was less than the lowest death rate of the controls. During the first twenty-nine days of the experiment the MAN death rate of group JE (800 pee 2 at 50 KV.) had a higher death rate than the controls. at 50 KV.) was greater than that of the controls. After the twenty-ninth day the death rate of JE was less than that of the controls. A AN The death rate of group JF (400 : at 50 KV.) was at all times greater than that of the controls. These results are shown graphically in figure 3. Figure 4 contains an analysis of these same curves by means of prob- ability paper, showing that, as in the case of experiment A, the curves are composed of accurate portions of probability curves placed end to end. All of the above results seem to be a direct confirmation of the curves given in the previous paper (loc. cit.). The effect of concentrated single doses is not nearly so marked as the effect of a series of small ‘homeopathic’ doses. This seems to be much the same law as is already well known in serum therapy and in the action of certain drugs. In the case of serum therapy, this law has been shown to be identical with the law of adsorption. If it could be rigorously shown that the effects of exposure to x-rays follow the same general law, we should conclude that the x-rays are responsible for the production of some substance, perhaps in the blood, which is later adsorbed. 456 WHEELER P. DAVEY 00 eee ot ee 60. |__| LL KOUP 40 |__| 1, |CONIRULS FOR GROUPS JF 29 | 4 874 WVU §. Jog sete wile Halt py a as Ec 60 aa SC 40 / | 00 A @S0KV-$0/18 2. 4 96 3 IWDIVIDUMLS eh LL _|__feraup Jo $0 Sp oe oe GO SONV-SOLA 20 | 2_|838 OMIM S 100 — yee GROUP JE go ies) 300 EEE @S0NV-S0/148 me | | e462. NOU s me yt i GROUP LF ae | 4002 @s0nY-50/ 837 |NOVIDULS 20 0 10 4 30 40 LYS Fig. 3 PROLONGATION OF LIFE DUE TO X-RAYS 457 end Z | ee aire GROUP JB ee oth een am “anriels rar 6 am may sl alieali ig Wah ~ We ah : | | laa pinlaits | B63 20 se are wes DWV - 508 a > S —|— 3b | GROUP JF | ar MNIOUES | (PSOKY- LON | BPO 7) 20 WHDOD 99 929 9999 *LLA0 Fig. 4 458 WHEELER P. DAVEY . SUMMARY 1. It has been shown that the life of Tribolium confusum may be prolonged by the use of a purely physical agent, i.e., x-rays. 2. The prolongation of life due to a series of small daily doses is greater than that of larger doses given all at once. 3. The lethal effect of an x-ray dose is less if it is split up into a series of small daily doses than if it is given all at once. 4. A method of graphical analysis of results has been de- scribed by which the number of causes of death may be de- termined from the death rate, and by which the effect of an external agent upon each of these causes may be studied. 5. Using the same kind of organism throughout the whole experiment, the work reported in this and the previous paper has shown that, by merely varying the size of the dose, a purely physical agent (x-ray) may be made to produce at will, 1) a stimulation; 2) a destructive effect which occurs only after a latent interval, and 3) an instant destructive effect. Research Laboratory, Genera] Electric Co., Schenectady, N. Y. vl, eben os ' A ae | Resumen por el autor, Carl R. Moore. Sobre las propiedades de las gonadas como reguladores de los caracteres somaticos y psiquicos II. Crecimiento de las ratas machos y hembras gonadectomizados El autor da una descripcién del crecimiento de ratas después de separar las glaindulas sexuales en los individuos jévenes de ambos sexos. Estos experimentos se emprendieron para deter- minar si existe una diferencia constante en el peso de los indi- viduos de ambos sexos (como sucede en el crecimiento normal) aparte de una influencia secundaria de las glindulas sexuales. Los resultados obtenidos demuestran de una manera definitiva que hay una diferencia potencial de peso en los dos sexos, la cual no esta relacionada con la influencia que ejercen las glandulas sexuales; los machos pesan invariablemente mas que las hembras. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 2 ON THE PHYSIOLOGICAL PROPERTIES OF THE GONADS AS CONTROLLERS OF SOMATIC AND PSYCHICAL CHARACTERISTICS II. GROWTH OF GONADECTOMIZED MALE AND FEMALE RATS CARL R. MOORE Hull Zoological Laboratories, The University of Chicago ONE FIGURE During a study of the effects of transplanted gonads in modi- fying somatic and psychical characteristics in mammals, it was highly desirable to know the relations of the growth curve of gonadectomized male and female rats before the effects of homoplastic transplantation of gonads could be correctly interpreted. Many investigators have shown that the growth curve of normal male rats is consistently and considerably above that of normal female rats.2 Steinach,’ after gonadectomy of young male and female rats and subsequent transplantation to each of the gonad of the opposite sex, noticed that the growth (weight) of the females had a tendency to be above the normal for females and that of the males to be lowered from the normal for un- operated males, provided the transplantations were successful; and he has not only associated these changes with a supposed modification of sex, but has used the changes as a criterion of sexual changes. Stotsenburg, however, in a careful series of experiments, has shown that the elimination of the ovaries of young female rats, without the subsequent transplantation of testis, causes the growth curve of the spayed females to be increased 17 per cent 1 Moore (19). 2 See Donaldson (15). 3 See Steinach (’10, ‘11, 12, 713). 459 460 CARL R. MOORE to 30 per cent above that of normal females;* he has proved also that the removal of the testis from young males does not in- fluence the subsequent growth curve as compared with normal males. As the presence of the gonads, at least in case of the female rat, is a modifying element of the growth curve, it is highly desirable to know whether there is a difference in the growth of the two sexes aside from any influence exerted by the gonads. Is there a latent, potential, sex difference between the male and female growth aside from any modification brought about by the presence or absence of the sex glands? Stotsenburg’s data as presented are not adequate for giving an answer to this question. This paper contains the results of a comparison of the growth of completely castrated males and spayed females of the same litters and shows conclusively that there is constantly a differ- ence in weight in the two gonadectomized sexes. The spayed females do increase in weight relative to the normal females and approach more nearly the weight of the males, but in every case the males are heavier than the females. MATERIAL Seven litters of the common white rat (Mus norvegicus albinus), composed of fifty-four rats, forty of which were castrated or spayed, the remaining fourteen serving as controls, constitutes the material used. Each litter was kept in a sepa- rate cage, but in the same room during their growth, and the entire experiment was confined to the period from November 29, 1917, the birth of the oldest litter used, to September 1, 1918, when the last weighing was made; and since the weight of each litter was recorded for 180 days, it will be realized that corre- sponding weights for each litter were made at comparatively the same time of year. The diet used was a constant one and consisted of milk and bread daily, a small amount of meat twice each week mixed 4 Stotsenburg (713). GONADS AS CONTROLLERS OF CHARACTERISTICS 461 with the bread and milk and a small amount of grain (corn or oats) placed in the cage once each week. The rats were allowed to eat all they would at one time each day and weighings were made seven to ten hours after the meal. Gonadectomy was performed under anesthesia at ages of twenty-three to thirty days, and the gonads were removed from both sexes by a midventral incision. Table 1 is a tabulation of the material used; the normal male and female served as controls, though the relation of operated to unoperated animals is not considered in this paper. TABLE 1 LITTER AGE AT OPERATION ae Taz eee ceri CONTROLS NORMAL naan e eae oo lors, | ARE Vg Ia aig 11 A’B3 23 2 6 Io, 192 30 A'B} 26 33 3 lc; Fe 48 A'B} 30 3 3 17,19 24 A1B1 51 2 2 lot, ES 17 A'B! 29 3 3 Wohi 25 A2B? 26 3 3 197,19 38 A'1B1 31 2 2 eta 9 Motsiisiseven litters)... 0)... 18 22 Stotsenburg has already stated that, at about 200 days old, many times rats are affected by certain ailments that interrupt the ascent of the normal growth curve, and for this reason, as well as on account of the death of several animals, the weight records only include growth up to 180 days. But considerably before this time the rats would have been sexually mature and entirely adult. Table 2 is a tabulated record of the average weight of indi- viduals of each sex in each litter from the time of operation up to the 180th day, as well as the number of individuals of each sex in the litter at that age; the last column is the percentage of the weight of the males above that of the females at each weigh- ing, the percentages having been calculated from the average weight of the individuals of each sex as given in the table. An THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 28, NO. 3 462 CARL R. MOORE TABLE 2 ss sw SSS CASTRATED MALES SPAYED FEMALES WEIGHT OF MALES Number of Number of 3 ABOVE FEMALES animals animals Average weight ee i ee ee ee Litter 11 A’B3 AGE IN DAYS Average weight days grams grams per cent 23 2 25.0 6 24.6 6 46 2 58.0 6 53.3 orl 59 2 91.0 6 79.5 14.4 90 i 122.0 6 106.6 15.0 120 1 170.0 6 140.1 21.4 150 1 > 208.0 6 165.0 26.0 180 1 207 .0 6 162.4 27.0 ' Litter 30 AB} 3 26.0 3 26.3 —1.1 3 58.0 3 56.0 3.5 3 93 .6 3 85.0 10.1 90 3 124.0 3 109.6 13.1 120 3 167.0 3 142.0 17.6 150 3 199.0 3 174.0 14.3 180 3 188.3 3 164.0 14.8 Litter 48 A1B 30 3 22.0 3 19.6 1193 572 43 3 30.6 3 26.6 15.0 60 3 59.0 3 50.0 18.0 90 3 110.6 3 93.6 18.1 120 3 136.0 3 122.6 10.9 150 3 142.6 3 130.0 9.6 180 3 161.6 3 148.0 9.1 Litter 24 A1B1 51 2 55.0 2 49.0 12.2 97 2 128.0 2 IW) 8.8 120 2 152.5 2 136.0 IPA 150 2 178.0 2, 169.0 5.3 180 2, 185.0 Py 172.5 US Litter 17 A‘B1 29 3 27.0 3 27.0 45 He 48.5 3 49.6 —2.2 60 2 73.0 3 74.3 —1.3 90 2 106.0 3 100.0 6.0 120 2 149.5 3 134.6 ih) 150 2 163.5 3 139.3 Ul7f583 180 2 176.5 3 155.0 153574 GONADS AS CONTROLLERS OF CHARACTERISTICS 463 TABLE 2—Continued OEE EM OS ere ene a ee CASTRATED MALES SPAYED FEMALES WEIGHT OF MALES ABOVE FEMALES Average weight AGE IN DAYS Number of animals Number of wnimals Average weight Litter 25 A?B? days grams grams per cent 26 3 28.6 3 27.3 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 15fe3 2 125.0 25.8 150 3 190.3 2 148.5 28.1 180 3 195.0 2 156.5 24.6 Litter 38 A‘B! 31 2 30.0 2 31.0 12.9 44 2 59.0 2 51.0 15.6 60 2 89.0 2 85.5 4.0 90 2 117.5 2 111.5 5.3 120 2 161.5 2 148.5 8.7 150 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. "499.1100 Ajoyeurxordde ATUO st vAand oy} Jo yaed sty} fose owes oY} A]JOVXO 7B poysIoM Jou 919M SI9z4I] [[B “BuruUIseq oy} 4e yey} JOR oY OF onp st Skep AyxIs 03 AZAIYY WOIJ SOAIND oY} Jo yavd UOYyoIg oy, ‘SoAmnd oy} Aq pozuosoidos ore pu poyndwod 010M SoBV UBATS 4V 810}}1] [|B 10} XOS TORO JO FYSIOM OFRIOAB OY} SITY WOIJ '(Z o[qGBy}) SABp Ul osu poyeUSISOp OY} JB 1o}}I] YORI JO XOS YOU JO JYSIOM OBIOAG OY} SuIsn AQ pojoNA}SUOD O1OM SOAIND OY, ‘SoTBuley poAvds Jo ouo IAMOT OY} ‘SoTBUT po}BIYSBO JO Y} MOIS OY} S}Uosetdod oAINO Joddn oy, “SABP UT sTBUIIUG OY} Jo 99v JY} OBSSTOSqE oY PUL qySiom Apoq Jo surviSs yUdSeidod SoyVUIPIO OY], ‘S}RA O[BWOJ PUL O[BVUT POZIWIOJOOpLUOS Jo YYMOIT JO OAIND T “SI Ost 02) og! ost Ovi oeT oz! ol OoL 06 08 Or 09 os Ov of oe oT Ov ‘\ \ N ze N ‘ \ ra sS SN \ \ ‘ » s 91 09 08 BEE SCoS ee te sot ed) Oe a a bbe ost BER MeBe iS bed a i beg rth ts od ea wld ete st fee rae 002 pais te a 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 ina 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 cases, 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 Donatpson, H.H. 1915 Therat. Memoirs of the Wistar Institute of Anatomy and Biology, no. 6, Philadelphia. LAWRENCE, J. V., AND Rippiez, 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. Zodl. vol. 28, no. 2. } Rippie, Oscar 1917 The theory of sex as stated in terms of results of studies on pigeons. Science, vol. 46, no. 1175, pp. 19-24. Sreinacn, 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 Saéugetieren durch Austausch der Pubertatsdriisen. Zentribl. f. Physiol., Bd. 25, S. 723- 726. 1912 Willkiirliche Umwandlung von Siugetier-Mannchen in Tiere mit ausgeprigt weiblichen Geschlechtscharakteren und weibliches Psyche. Pfliigers 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. Stotsensurc, 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. Rece., 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 colocan a la luz solar, hasta tal punto que el agua del cultivo puede contener hasta + 16 cc. de oxfgeno libre por litro, mientras que en la oscuridad no hay desprendimiento de dicho gas. El autor colocé 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 ce. por litro) y en la que contenia una can- tidad mayor de este gas (2 a 8 ce. 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 Department of Zoology, University of Nebraska, Lincoln, Nebraska Recent papers by Shull and Ladoff (’16) and Shull (718) 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 ce. of old stable-tea culture water and poured into stender dishes, 1 inch in diameter, OXYGEN AND MALE PRODUCTION 471 and placed 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 in it. 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 in old stable culture water that is devoid of all food substances ! La g : |e | eel | LOTS LIGHT CONDITIONS TIME, 1918 Fe S = & Ze = ea | 43/88/88 py =) = rs 5 & B Z B & i 4 ce. ce. ce. ce. 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 42] 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. e.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 a a LOT TIME, 1918 ie 1 { A 4 p.m., Nov. 9 B 4 p.m., Nov. 9 f A | 4p.m., Nov. 7 B | 4:15 p.m., Nov. 7 9 C | 5 p.m., Nov. 8 D | 4 Pp.m., Nov. 9 E | 11 a.m., Nov. 10 F | 5 p.m., Nov. 11 [ A | 9.a.M., Dec. 17 3 B | 10 a.m., Dec. 18 | C | 10 a.m., Dec. 19 D | 9a.Mm., Dec. 20 ‘| A | 3p.m., Dec. 23 4 B | 3 Pp.m., Dec. 24 C | 3 p.m., Dec. 25 D | 3 v.m., Dec. 26 (| A | 3 p.m, Dec. 25 B 3 P.M., Dec. 26 5 C | 3 p.m., Dec. 27 Dy | 3 pm. Dee. 28 || E | 3 p.m, Dee. 29 CULTURE WATER TESTED OXYGEN PER LITER cc. 7.97 7.02 WATER Tap-water Tank rain-water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Old culture water Unfiltered Filtered Filtered Filtered Filtered Filtered Unfiltered Filtered Filtered Filtered Filtered Filtered Filtered Filtered Filtered Filtered Filtered Filtered Filtered 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 ce. 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. to a 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 14 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 ee. 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 is 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 2cc. OF CHLAMYDOMONAS AND A FEW DROPS OF POLYTOMA In 50 cc. CULTURE WATER IN EACH EXPERIMENT hase TIME, 1918 Sunlight Darkness Number | Number | Per cent | Number | Number | Per cent of 9 of S| of fo) of fe) 1 April 9-15 47 8 14+ Poi 25 48+ 2, April 15-21 266 28 9+ 284 78 21+ 3 April 17-22 171 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 BM 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 4 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+ Motals 0s. Jy. ee 1590 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 table 3 ee ee ee vite, 1918 UE 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 ec., were put into 10 ec. of the culture water and well stirred. Then 3 cc. of this mixture was added to°50 cc. of the filtered old stable-tea culture water. To the lot that was placed in sunlight 0.2 to 1 ce. 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 TABLE 5 MALE PRODUCTION 477 Showing that in sunlight where 2 to 15 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 EXPERIMENTS PERIOD TIME, 1918 DIRECT SUNLIGHT water Oxygen per li- tested ter Culture Number females treated Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End Beginning End 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 4 p.m., Oct. 10 a.m., Oct. 11 a.m., Oct. 11 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. 10 a.m., Oct. M., Oct. 28 10 A.M., 10 a.o., 4 P.M., 10 A.M., Oct. Nov. 2 Oct. 26 28 3l Nov. 1 29 30 5 p.m., Nov. 3 10 A.M., 5 P.M., Oct. Nov. 4} 42] 7.81 bl 42| 5.97 42/12 .11 42}11.72 12 DARENESS Daughters 8 = s Daughters e ile. |s 36 7 eee 3 a. ese! ies g 99/79] 8% 128) &. aslo o|ae 3a cc. | Cc 12 9} 1155 4) 16/80 3 29| 15/34-+4+ 12) 25|67+- 4 27| 13)382+ 4| 19/82+- 4 28} 8)/22-+ 15} 21/58+- 5 28} 12|30 7| 33\82+ 4 9} 1/10 1) 787+ 4 10} 8/42+ 3} 17/85 16 17| 33/66 | 42/5.64 15} 35\70 12 25} 25/50 | 42/4.81 19} 31/62 12 35| 15/30 | 42/5.33 14] 36/72 12 10) 38/79-+-| 42/5.86 9) 34/79+ 24) 26/52 12 42/6 .25 + ea i THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL, 28, No.3 478 DAVID D. WHITNEY TABLE 5—Concluded DIRECT SUNLIGHT DARKNESS | & a = Daughters 8 a S Daughters Dn 3 ee my 3 c 5 PERIOD TIME, 1918 S 3 5 a Es a FI = 5 | iiss! 2 Ss ) Slade s £ol/e3}8 |s3 z 2 2s] &. [eelee|r9| 8% (3818. l2sloolng| o% é aol pe 188 5/38] 28 55 a a Oy [Foe Ve A NO lio: Te & cc. cc. cc. cc. | Beginning | 9 a.m., Nov. 1] 42] 2.15} 12 42'2.15] 12 "|| End 3 p.m., Nov. 5] 42] 5.86] | 30] 20/40 | 42/5.33 2| 48/96 14 Beginning} 9 a.m., Nov. 2) 42} 0.97] 12 42/0.97) 12 ‘End 3 p.M., Nov. 6] 42] 7.81 33] 17/84 | 42/5.93 6) 44/88 i ‘Beginning | 10 a.m., Nov. 3} 42] 2.65} 12 42)\2 .65) 12 \| End 3 p.m., Nov. 7| 42| 7.44| | 14] 36172 | 42l6.45| | 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} 18/86 | 42/7.81 10} 40/80 ae Beginning | 11 a.m., Nov. 12) 42] 3.59] 13 42/3 59] 13 \ End 4p.m., Nov. 16} 42] 7.81 32} 18/36 | 42/7.81 9) 41/82 19 Beginning | 10 a.m., Nov. 13} 42] 3.50} 13 42/3 .50| 13 End M., Nov. 17 | 42| 8.21 27| 23/46 42|8 .21 11} 39/78 oa Beginning | 9 a.m., Nov. 14] 42] 3.12] 13 42)3 12] 13 “|| End 4 p.m., Noy. 18} 42/15.63 36| 14/28 | 42/7.81 18] 32\64 SSCUINATTL INV? oe, Ses. peer et ee tee 9+* 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 & TIME, 1918 aaah d eae ee TIME, 1918 Gee a ae 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 {| a.m., sun November 15 loudy ORIG | | 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 13-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 ce., 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 cases 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. Piz is |+o¢ os's] 2b | Ol] a& | 09 | 08 | 02 Is 2] 2 | ystp wadg Zo AON “W'd G pug \y SI |0S'€| cb | OT a €1 jog €) Z| ystp uedQ | 6T “AON “Wa g | BuruuIdog | | C6\9F |b |+0S OES Chal ee Lp ae | $6 e2hal Ss tp | Sb | ysip uedo 1% “AON “Wd pu by c €I |€6'3| ch | SI] a GI |€6'3| Ze | Usip ued | gt AON “Wag | Suruudeg Zsitr |6 |+0¢ vO Ol Cre) 22 Et | PZ Lee St 182} a | ystp uodo 0% “AON “Wd G ae 7 SI \08'F| 2h | oT] a SI 08h SF | ysip ued | ZT “AON “WV OT aeienaae PS|SF |S |+0S 80'S] ch | O88) Gd | 96) 8FIZ, £02) &k ysip wodoO 61 “AON “Wa F pug ¢ s €I |2¢'s| ce | O's) a EL |2o°¢) cb | ystipuedQ |) gf ‘AON “WV 6 | SurUUIdeg ca) a gojge {41 |+-0¢ 98'S} cb | OF G2 | 88] WH) 9 18° 2} oP ystp ued 81 “AON “Wd § pul i z - &I |29°s) ch | O08) a ST |2G'§| GF ystp wod¢O GT ‘AON “Nd Suruuisog | | = < 824/68 |IT |+0S Ly S| cb | O8€ A | 2] IF) 6 18°24) Zh | ystp uoddO LI ‘AON “N'Y OT pus || Vv I e €1 |80'S| ae | O's] a €I [80'S] GP} Yysip wedQ | FI “AON “Wa pH | Suruursoeg | / a 99 99 “09 99 “09 Ss pa | as ed alls || ee | | eae = Seti aes 3 Sol kelee & |e 8 Bo| & | & cS es of] Be) FP |e) g 5 A Pia iS 2, es es |= e, 8 | é z= q, su] of =| 38 Q gfe| $s 40 Sar g ee ia S + SIO ‘WIL aomaa SLOT Be 3 a ce, 5 A slog qaneqg Soe S siogysne a AIlddOS UlVY GAHSINIWIG SSANYUVA NI IOULNOD are wayyy Jad Uabhixo aasf fo +99 4 07 B Burummjuos sajpm aanqyna ur paonpoud ssafyou sp ‘uaa sad FZ ‘saypmas a buronpoid-ajou huni sv paonpoud sayy tad uablixo aasf fo +99 ¢ 0} [ fo unui oy) Burvurmjuos sajpn asnyynd Ur sdafyos ay. Bunoys — 4 QTaVvL MALE PRODUCTION OXYGEN AND +89/TE SPC +E8/9Z OLITE GLE v1 I OL 9Z &1 iat 0% +0¢ el reat &I reat eT €1 ST é1 &1 €1 19° SOP OL T OVS 0G 90°& ZO 1 So's LV T 8o°s IGG OS F GP oP = S aa RA ARQ AA oo S S ae AF AA 98 OF og 98 +£6 tS 08 9g OL i 0G §1 SF cI 1@ GP OF Ge. OL 6G cI rea é1 I &I €I ~ ystp uodgQ udd() ystp ysip ysip ysip ued( uod¢ ysip uwodg ysip uedg uedg usd Tere ystp uedg ystp ued¢g u9dQ usdgQ ysip ystp uodg uadgo YStp Md ystp uedg ysip ued uadg uadg ystp ystp uedg uodaQ ysip ysip OF GS GG FG IG &@ 0G 00] “O9@] ‘VIC “AON ‘09(] “AON 0a] “AON "AON "AON "AON "AON “AON “AON "AON “AON “AON "AON "AON "AON “i'd & CW'V 6 “Wd ¢ CNW Il pug sumursog pug Suruuisog pug SULUUIDIGT pad SUI BIg pu suruuis9g pug suruuisoq pud Suruuiseg pug surluuisag pug Suldulsog pad Suruuisegq \ V J by \y \y cT iat él él Il Ol WHITNEY DAVID D. 484 TFL|ST9ZIZ GLIOE |FL | 8% +Esice |Z | 0 CsiIv 6 | GT “wlalol ez g Be o | 0] 70 | &E 5 25 40 3B BS BS siojyaneq | § iq A1dd08 Go € 18° Teme 19°C| GP SL |88°%| oP 90°) oF €1 |29°Z| oP OL T| GF €1 |Z9°Z| CF ‘90 | ‘90 se| # | 2 A beedalheas 8 | s e E. 3 ulv GaHSININIa eee e ele ee ele ee ele ee ele ee ele eee 30 Test (a) Ajddns ny TPL! 099} 12e 94) 8& | GI +62} 62 | IT 08) OF | OF v7) a} © a wo | 40 i) & g soy 40 S10} q3neq SSANYUVA NI IOULNOD sopeuley jo = 19}1] Jed uasAXG CePOrOmnDeO Crh Critic) has Get) 0 CM INO. Uo th oad eth Glo Din cron puo 4u OSVIDAY ystp wodo ysip u0edO ystp uedo ystp wodo ystp uedo ysip uedo Ajddns ity pm a ————————— athe cle, efemetete) 6)-e ter dts (ce © (0.6 cle tprale's (ele.8)¢ Suruursaq 4% eSvIvAy Areuiuing Hehe nalie octets we pieho ude ee: lo fo] po Melanin) istidietien ep igiiprw! 0.6 8/ese a 8h Bip OBvIOAY “TOT, ; 09 OO SE f P pu 700g “WV [I | Suruurdeg Me 61 g 00q “Nd T pug |) e-o0q “NV TI | Suruusog J Af st 09 be . U es Soar all LI Z 09q “N'Y 6 suruuiseg SI6T ‘AWG aormad SLOT pozso} 19} 8M 91ng[ND papnpuog—L WIAVL 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 ec. 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 -OpAWe[ yy) JO uoljtIppe sinoy ZZ rel | Aep dutinp ung | $2°¢ | GO] cP | Jo}JW SoINULOL YT potozpIy | Ur Sarpuvys 10}RM-ureyy URL Wid Fale Ora | sanoy ZZ rel by Avp Sutimp ung | e¢-G | GQ | GF pete} guy) | Ul Surpurys J0yeM-urey Lure “Wd F | g sanoy ZZ, rel | CoE OOL| Ge peiey[guy) | Ul durpurys 1aojpem-urey Lure “wap | Vv |i sAep [BioAas rel Aep Surinp ung | 9f'¢ | OT Cr poroy]yuy) | ul Surpurys s9yVM-urTeyy g user “Wad fF] sXep [eioAes rel e OI ON = ae pesdoy[guy) | ul Burpuvys s97zeM-urey gQ:uee “Wd Zp | V yystu Surpoo sanoy 9¢ ref -o1d jo ssouyieq | 91°¢ | OL] ZF pele}[Juy) | Ui Surlpuvys soyeM-urey | g-uee “WV O@g | g smmoy gg rel ( G $99 1) CP po19}[guy) | UL Surpuvys soyea-urey | gcuee “wv ogg | y J sedid Avp dutinp ung | e290 | ¢@:0| ZF porte} guy | ur surpueys s903eM-urey pouee “Wap | g sodid T €2°0 0 CV pesiey[guy) | ur BurIpueys 1078M-uUIeYy Pp uve “Wa P| V “09 *29 *99. eo YaLIT | , 2 z 5 4 Phe Seana ae bah e 2 S ee an SNOILIGNOO HALYM AGLY M 6161 ‘AW SLOT zZ g | -AXO iS) a8 | at PISDILIUL JOU $i Sda}DM asay} fo Juazuod vabhxo ay) sayon asnqno PIO 40 19D UWA OFUL Ind au S.inoy [D.41aAa8 LOf UNS oY} UL BurpuD{s Uaag aaDY YorYNn sDUOWOphiuD)]Y;) fo sassDLU uayn yoy) Buinoyg 8 HIAVL 487 MALE PRODUCTION AND OXYGEN Aep Surimp ung Aevp sutinp ung Avp Suiinp ung Avp surmmp ung ite 61 Ol Ww 96° 9g © Ven) usd o ¢'0 00 00 00 00 0-0 0°0 0°0 0-0 00 seu -owopAmR[ yO Jo uoryip -pB doyye o0u0 4B pdso4 [ly] poroy,yuy) pos194 LT poroy yu seuoul -opAure[y9 Jo uorrppe Joye soynurut G pIso}[ ly] por0y,yuy) por guy poroy [hl pero} ft peroyyguy polo} Gul) pero poroyyL peroy {yup persq[gua/) sodid IOVEM-ULRY sodid 10} VM-UIBYT sodid IOPVM-ULCY sodid 19JBM-ULBY ul Surpurys ul Surpueys ul Surpurys ul Suipurys sanoy 96 ref Ul SUIPURYS 10}BM-UTeYy sanoy 96 ael UL SUIPURIS 10} BM-ULRY sanoy 96 ae UL SUIPUBIS 10}BM-UIeYy sanoy ZZ tel Ul SUIpuRys 10} BM-ULRY sinoy ZZ ae UI SUIPURYS 10}eM-UIRYT sinoy ZL avl SUIPUvIS Ul 10}8M-UTRYY sanoy ZZ ae UI SUIPUBYS 10}VM-UlBY sodid Ul SUpIURyS 10}8M-UTBY sodid 19} @M-ULBY sedid UL SUIPUBIS 10} BM-UTRYT sodid 19} VM-ULBY ur Surpueys ul Surpurys 8 “UBS ue “uRe’ une ‘uBe ‘uBe “uel ‘uel ‘uBe ‘uve ‘uBe ‘uBe ‘uBe ‘uRe ‘uel te ia ‘ ‘ te ia te W WwW "Ww WwW Ww “Ww ‘Ww ‘dae ‘a9 ‘19 Pare rae Pare Pare ta “iA OO 8 —— —, ————— —— ~— xD WHITNEY DAVID D. 488 ‘WV Sulinp ung ‘Wy Sulinp ung ‘Wy Sulinp ung ‘WV Sulinp ung Aep surimp ung Aep sutinp ung 6g ¢ oT ¢ 09°2 12% 98°0 96 °S Lg°0 69'T 9€°0 29 T 96°0 "09 SVNOWOCANYVTHO ZO SNOILIGNOO LHDIT UALIT ud Nao -AXO g°0 ¢°0 0°0 ¢°0 agaay SYNOK -OdAWNVIHO GV GP GP GY GP GP GP GP GP GP “99 SBuoul -opAue[yO Jo WoryIppe I0jJB sunoy fF podtog [ly] SCUuOUL -opAure[yO Jo WOryIppe Joyye sanoy Ff pojyuBooqy pote}, Gul) SBuoul “opAwE[yD JO WOrIIppe 1ojje saunoy Ff posoz[ ly sBvuoul ~opAuue[YyD Jo uorzIppe Joyje smnoy fF pojyuvood por0} [NT peroyyun svuoUl “opAWB[yD Jo UOIZIppe IojJe SoyNuIUL YT Pelo4[T poyuRoog pered tT poyuvooq, ur ul ur ur agsn. WaLVM SNOILIGNOD YALVM surpueys Surpueys Ssurpueys surpueys surpueys Surpueys Surpueys sodid 10]BM-ULEYT sodid 10}BM-UIB YY sodid 10} BM-ULB YY sedid 104BM-ULBY sodid 194BM-UIE YY sodid 10}'BM-UIB YY sodid 104 BM-ULB YT 10}VM OIN{TND PIO 10}BM OINATND PIO 19}BM 9IN4[NI PIO JOVBM 9IN4[NI P[O ULV M IT ‘use “Wa yz | O | IT ‘usp “W'd 7 IT ‘uve “Wd % Il “ave “Wd F Il ‘uve “Wd F Il ‘uve OWad p IL ‘uve Onwd p OL uve “Wa g OL wer “Wd ¢ Or ‘uve “Nd e OL ‘uBe “Wd ¢ 616] WIL | panuyuog—8 WTAVL 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+ to 4+ 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 + ee. to 1+ ce. 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 + ec. 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+ ce. 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, Oey domonas. =. a: Vis Ee Say’ exh 2 ee y= be (6) en ERRATUM 7 to The Journal of Experimental Zoélogy, volume 28, number 3 (July, 1919), David D. Whitney, author, page 491, third line from bottom, the word not has bs been omitted between the words oxygen is and a factor. Paragraph 7 should é read as follows: 7. The general conclusion is that oxygen is not a factor in causing a production of males except inasmuch as it is necessary for all life l processes and activities of the rotifers. e- n- en SS he 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 ec. 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 ec. of oxygen per liter. a 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. 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 pres per LOR ¢ Sige with with the was of fr This Shul male of 03 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 + ee. to 1+ ce. 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 + ec. 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 + ce. 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 ee. 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 ee. 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 Snuuu, 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. Zodél., vol. 24, no. 1, October, pp. 101-145. SUBJECT AND AUTHOR INDEX apres multiplication and regenera- tion in Sagartia luciae Verrill........... BAvkeerces: J. Percy. A nutritional study of insects with special reference to micro-organisms and theirsubstrata.. 1 Bripees, Carvin B. Specific modifiers of eosin eye color in Drosophila melano- PANY Fae o> deaeeber Sonne doo ceeanoCriaC Goon an 337 Brinces, Cavin B. The genetics of purple eye color in Drosophila..........+++++++++ C= are subject to selection on the basis of their genetic potentialities. Evi- Gence that Revi cee ald steels eins = ance _. 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........ ---+-+++++: 337 Color in Drosophila. The genetics of purple Gir 3nd Goes OGED onra dO RUOESC ONDE GODOOCDDOUS 265 ANFORTH, 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 SmMall'GOses Of X-TAYS. vices oni e secs sc anae Davis, DonaLtp Watton. Asexual multipli- cation and regeneration in Sagartia luciae Werrill irocnreeic nana teria tices too eicieeets 161 Day, Epwarp C. The physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to sensory PESPONEES Sena ielae cr Teen aek ce eile eiciale Drosophila melanogaster. Specific modifiers GHEOBINIE VE COLOL AM seers ele ere nies eaeie eistareie 337 Drosophila. The genetics of purple eye color UW ee tote areal ohare aie lslow-siv.cloteieie css oo ninlew/sinvacatsle Nsjate 265 1 be production. The bearing of ratios on theories of the inheritance of winter.. 83 Eosin eye color in Drosophila melanogaster. Specific modifiers of................--+.-+ 337 Ephestia kihniella Zeller. Genetic studies on the Mediterranean flour-moth......... 413 Eye color in Drosophila melanogaster. Spe- cific modifiers of eosin................-..- 337 Eye color in Drosophila. The genetics of POIEDE Oa eee emia nica Catton aivis s tuculesiooy cars 265 ACTOR. in causing male production in Hydatina senta. The ineffectiveness Of OXY REWGS A iedacs caine nocd se none sean 469 Flour-moth, Ephestia kihniella Zeller. Genetic studies on the Mediterranean..... 413 ANGLION to sensory responses. The physiology of the nervous system of the tunicate. I. The relation of the LIST): SOU COHAN eo cock Sec OU 307 Genetics of purple eye color in Drosophila. MG RPMMIEM IG Gs oslo ovina. coeescdunssusieeieen 265 Genetic potentialities. Evidence that germ cells are subject to selection on the basis Nign sl oe aga onan pemote CoP aeOmencs 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- dence hatin). ts lao cee hee een ome we 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., anD MacMuLiEen, GRACE. The bearing of ratios on theories of the inheritance of winter egg production...... 83 OMOZYGOUS yellow mice. The fate of 125 Hydatina senta. The ineffectiveness of oxygen as a factor in causing male pro- GuUuCHOMA ance soenoeeice ee cae asanen sn 469 NHERITANCE of winter egg produc- tion. The bearing of ratios on theories Of thet tees ee a ridofra ee tomate woe 83 Insects with special reference to micro-organ- isms and their substrata. A nutritional Bud y/OF sos. eie xs oa cmt clenclowie sigan = 2 ea IRKHAM, Wrii11am B. The fate of homozygous yellow mice.........-.-- on kee | ares of Tribolium confusum apparently due to small qoses of x-rays. Prolonga- TAO TMOLN: aos ee socism in teleiee inertness ==. 447 Lucran VerrmL. A sexual multiplication and regeneration in Sagartia............-- 161 AY Geet tence Grace. Goodale, H. D., and. The bearing of ratios on theories of the inheritance of winter egg pro- Quction! o: ssn eee areca ce 83 Male production in Hydatina senta. ineffectiveness of oxygen as a factor in CHURN cr clea woe eae Seiten ein alte en ses SI 469 Mice. The fate of homozygous yellow.... . . 125 Micro-organisms and their substrata. A nutritional study of insects with special reference to..... Be ene See copes Selene Modifiers of eosin eye color in Drosophila melanogaster. Specific.... 337 Moore, Cart R. On the physiological properties of the gonads as controllers of somatie and psychical characteristics. Fe The rater enact ees. eaten 137 Moore, Cart R. On the physiological properties of the gonads as controllers of somatic and psychical characteristics. II. Growth of gonadectomized male and Ferdblerates jester nana cn wate cnet, = «inte 459 Multiplication and regeneration in Sagartia KoGiney Vermiile | ABCTURE acs ce cscs coun tec 161 493 494 INDEX ERVE ganglion to sensory responses. N The physiology of the nervous system of the tunicate. I. The relation of aT Re ETE e oS ers an oomeciedos Seaee 307 Nervous system of the tunicate. I. The 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 BUBStraters Ate ierc tee or olets icuetstotinttetene talons XYGEN as a factor in causing male production in Hydatina senta. The ineffectiveness Of..............- abe: pore 469 HYSIOLOGICAL properties of the gonads as controllers of somatic and psychical characteristics. I. The rat. Onithe 422 4ssiadasl eA ds ae eed: eee 137 Physiological properties of the gonads as con- trollers of somatic and psychical char- acteristics. II. Growth of gonadecto- mized maleand femalerats. Onthe. . . 459 Physiology of the nervous system of the tunicate. I. The relation of the nerve ganglion to sensory responses. The..... . 307 Potentialities. Evidence that germ cells are subject to selection on the basis of their MONCH Cats. Soe. ee ee se sare ates: Saphaete 385 Production in Hydatina senta. The in- effectiveness of oxygen as a factor in GsUsing: Male. fs SFn Ve arcu. 4 bs. 84 = tote 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............. eo Psychical characteristics. II. Growth of gonadectomized’ male and female rats. On the physiological properties of the gonads as controllers of somaticand . . . 459 Porple eye color in Drosophila. The genetics eet a. pilser he ORE tee anaventa-a Poor Ass 265 | Pee ee on theories of the inheritance of winter egg production. The bearingof. 83 Rat. On the physiological properties of the gonads as controllers of somatic and psychical characteristics. I. The......... 137 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 and................ 161 Responses. The physiology of the nervous system of the tunicate. I. The relation cf the nerve ganglion to sensory........... 307 AGARTIA luciae Verrill, Asexual multi- plication and regeneration in............. 161 Selection on the basis of their genetic poten- tialities. Evidence that germ cells are BUDJECU LOD Secs eee ee ecto eiaiane > eres 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 properties of the gonads as controllers of...... . ._.. 137 Somatic and psychical characteristics. II. 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 (heir i es eer 8 eee ae ay System of the tunicate. I. The relation of the nerve ganglion to sensory responses. The physiology of the nervous............. 307 RIBOLIUM confusum apparently due to small doses of x-rays. Prolonga- tion of life of ..... IS DAIS es PSA eee 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.............. 41 Wuitney, Davin D. ‘The ineffectiveness of oxygen as a factor in causing male pro- duction in Hydatina senta...............- 469 Y-RAYS. Prolongation of life of Tri- bolium confusum apparently due to smnall dosesio fe 27-2ea At Re Beene 447 ; ie Roy e ‘ea TU sh ee Oe ‘abe yaa Meer || ei. on Pa erat ee ees au. fy « ia 7 hia le et re ee Price J y ; he P PAbT ry a 47 : wa" + ms U ok ATE i up, § it a, Le ifs ay | ; y eee ar a BV Arut we Brpt? Me a rear v oe a eter al Paine eteapeon tnd age M } ae aay Ried Md, be My " vel, Coeer. "=. eo . ae abd, tat a ; — eed r ; : b QL The Journal of experimental 1 zoology cop.2 Biological & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY . cep byw “ ee Oe " Fa pe BDI PPh ity A PF