OS, aN aE aia Ne us See siu PEN ret Sat = BAN ‘e pee HERG tN ante Taneguees say 4 Me Ait ( eaee, - pie eee Coan Roatan Ryans i i at H yah rh f a ae tien 4 Hh) Nine < ity A ; 4 i ae \ At eT SE ies ‘ Ba teh i : yy can i Het HI i put i 2 Ales ed he) pate y REN fo ve * NES a hist Pease Hy Nee von aan vt Bn. ti htt a ay Nt Me 4 43474 ee art Net eeeratta SOUR FT| ¥ ue) r} Sy ‘ af a atte (aetal an iM Leap thik bi) bee: j : a ‘agua et Ceri eaeay LAN acd RGR een bed thn } i Gh alanntown ahaa salt a ac faneeRiCa atlas Lette aa 2 ye) Lrasaaayi t i Hats Mia Nemec: vite mis K niles wish ' Sehayh Aiades is ‘ he Sta Ne LN RoR SH Rabe ath RAY tite stats? y ei ha Teeeatow 3.8 R ‘ ‘ vt - IM ee a aay ih =. an MRNA cat, Mt FEAAEL Pel eae Peter rol aioe he satenae he ee Nt a eeie) iaaae the saat hith Nay hs atm < eee Qe Snes PN) acre by re bye HER Oss ase t tts Fee Hite ae ed y see ! ae Hs tie ae ; ae : = eS Kobtentin ms See ESS Ronis a a i i S042 Yui it if i» * a pehieay Mya a eae NG aS EG f Pie Aan : Re PRGA RSTn oe nani fa hh Rip Tear ee aii tt Gini 4 ARR ARE Sn Mi ir Pah : 018 ted ob ae itt Bat tS fat Dag } ORE Ay ALC TEST eR Wa SiR HU A aH \A ary Mes “ San ith) =e, 5 ae a =~ Fosse eer ee Ls Pail i Fastin Hee a : Sa SSS Seaack 2 Se tS SS = = ait 4 4) ih ite ie ( fh aa iy = ft > Peas Pai aed wh a, Seite ’ tee Moree : = Thor, os 2 oo ery ee linge ee. red ate i, de ¥ id | Digitized by the Internet Archive in 2009 with funding from University of Toronto http://www.archive.org/details/journalofexperim16broo en) i ro 1 . ae = : "| (Ne ti / Tey = ie eee em as —) a ‘7 = ha a - fr a iz I 7 5 i] 4 a : , ° ’ ae . 4 we 1 j ct i @ f ; +) — = -— ~ ~ - 7 le : Af ‘a ee , = ihe i ; Pere a we “nh. r a - a —. as en a { a j hi Ce 1 4, a a! aL tee: bade The * ? Sa THE JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY WILLIAM E, CastTLE FrankK R, LILLIE Harvard University University of Chicsgo Epwin G. CoNnELIN Jacques LoEB Princeton University Rockefeller Institute CHARLES B, DAVENPORT THomas H. MorGan Carnegie Institution Columbia University HERBERT S. JENNINGS GrorGce H. PARKER Johns Hopkins University Harvard University EDMUND B. WILSON, Columbia University and Ross G. HARRISON, Yale University Managing Editor VOLUME 16 1 -| i) oie 1914 | $3 ae g | THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA. CONTENTS No. 1 JANUARY E. J. Lunp. The relations of Bursaria to food. I. Selection in feeding and MEL URUSI OMe Bay MAP UPES. he os 8 ccs haha e Nee aoe das A elne So Aeale es fs Auice M. Bortnc anpD RayMOND PEARL. The odd chromosome in the sperma- togenesis of the domestic chicken. Six plates.......................005. CHARLES PacKarD. The effect of radium radiations on the fertilization of NEE Er On tenet ny enns pn pte) me I AE Sh as Od elt ed 8 oh attaket Joun C. Puituips. A further study of size inheritance in ducks, with observa- tions on the sex ratio of hybrid birds. Seven charts...................... GeorGE L. STREETER. Experimental evidence concerning the determination of posture of the membranous labyrinth in amphibian‘embryos. Thirty- 50D EEE S 5 ele 55 ee a Sn he No.2 FEBRUARY E.C. MacDowe tut. Multiple factors in Mendelian inheritance............... Henry Laurens. The reactions of normal and eyeless amphibian larvae to UPRNUES ASISEUITCH AS Oo ente cae © 2 1 Ee ee a ew ROS GEORGE ALFRED BaITsELL. Experiments on the reproduction of the hypo- trichous Infusoria. II. A study of the so-called life cycle in Oxyt icha Jallax and Pleurotricha lanceolata. Sixteen figures (one plate)........... LoranpDE Loss Wooprurr. On so-called conjugating and non-conjugating Weseon armmaectim: One Mewre os .k o.oo. ke. os esses sable eee ArTHUR WILLIAM Meyer. The supposed experimental production of hemo- lymph nodes and accessory spleens. V. Studies on hemal nodes......... ili Cs) 53 85 131 149 AL? 195 211 237 241 iv CONTENTS No. 3 APRIL @ N.E.McInpoo. The olfactory sense of the honey bee. Twenty-four figures 265 CriaupE W. MircHei anv J. H. Powers. Transmission through the resting egg of experimentally induced characters in Asplanchna amphora........ 347 W.C. Auter. Certain relations between rheotaxis and resistance to potassium | cyanidesin Jeopodat 0.0: 20h ssaqunguiaoma een Gere. ay te = ee 397 C.M. Cutip. Studies on the dynamics of morphogenesis and inheritance in experimental reproduction. VII. The stimulation of pieces by section in*Planaria dorotocephala. Four figures. . ...% (0. settsdes ne. ta) ee 413 . G. H. Parker. On the strength and the volume of the water currents pro- duced by. sponges) ...055 2.5... shee ees 22 ae + hae na so ina 443 No. 4 MAY H. H. NEwMAN. Modes of inheritance in teleost hybrids. Thirty-eight figures 447 MarcGaret Morris. The behavior of the chromatin in hybrids between Fun- dulus and Ctenolabrus. Thirty-six figures (five plates)................ 501 Frank R. Lituir. Studies of fertilization. VI. The mechanism of fertili- zationin Arbacia. One figure... 8 sen «salts cisco sols +e) oer eee 523 RaupyH 8. Lititis. Antagonism between salts and anesthetics. IV. In- activation of hypertonic sea-water by anesthetics...................-.-- 591 JONATHAN Risser. Olfactory reactions in ampb‘bians. One figure.......... 617 THE RELATIONS OF BURSARIA TO FOOD I. SELECTION IN FEEDING AND IN EXTRUSION E. J. LUND Zoélogical Laboratory of The Johns Hopkins University EIGHT FIGURES CONTENTS MEMINTRIMI S075 Sree PAL 8 ACE oe ina Nags a ike Shad MASE a aT oes Oe ee eee g 2 Action of the structural mechanism for feeding and the selection of food... 4 i. LEMEIG(CUINDTS ws 3 wih tasers Pl oR n TA ree ICC gine aE een REA a ni AES 5 3) PRPEMECCEMIRDICE oe Cc ok ie, ofarahs eS. a ht ale Ss eS Soe 8 ERO Eee he er 6 ai eee Formation of the vacuole and elimination of residues.................-.-.--- 7 Measurement of the amount of food eaten, and method of experimentation... 8 Internal relations affecting the feeding process...............2...-0002-e00 10 1. The relation of the physiological state of the organism to the feeding TTI EE TEL op ates Ge gts Shee pe Bec Sees ee eR Oe co a eas 10 2. Changes in the physiological state as shown by using the feeding proc- BES 2S, QUST Geo aE Se ee Eee aac eee ee ee a eee 16 BO GherECAUseSs Of iNdLviGUal VATIAtlON. . <2. oe acc cee aie oct eis es eels eo eke 19 ‘Hhevexternal relations of the feeding. process. ....:<.....4.s.0:.202.00+54 re, el) Jeebiiecisiol external factors ON feeding. ..... sac. c.2 ce ees cess eee eee es 20 ROOUCen ration ofthe TOOd SUPPLY sae stash Satine selec eens 20 b. Effect of mechanical stimulation and mechanical injury on feeding. 21 Gaplernnperabube yc cs: es ss 42% ee A eee Begg etre Rex OP tostt an nesy ide oe bo2 |e 23 a. Leer ls Oa oc, Bee eee ae a: a en ee eee 24 @. Winminediielitt! esse setéec oe SEIS Soot dete ORs OI ee 24 fee electric: current: > .. 2.2 ..--2 ss Gero AS BAe he Oe cn eee eae 26 SeEmNeer Hod, and the factors concermeds. as.6. .. 22 .<.h bee we ee ee ees 29 1. Experiments with stained and unstained yolk......................--4- 30 di. SS DuPnBT ravings SUR Se As Aare t= 0) 2 Oogles Ania ales 3 ee 30 RISE CCTI oe ens hee eo). ee ep RE, Fae alec sess 6 aletoetane 34 G. cir isa Tigi 6 os 6 Saeco tn oe Oech Sree ie oid eae eS 37 dL, COGS TOO Gar Sogod ne se COIEAE 6c 5 - 6 SOO Ce EES ts 39 2 2 lpn LU NS Se eer. eS eet) eer 40 i, wlighe walle se? Aas oe Aer acs = 5 Ober eBicrege ate choi itaeaen neice ercich 40 2. The basis for and the nature,of the selection of food in Bursaria....... 4] The relation of Bursaria to digestible and non-digestible substances........ 43 SEP RSRELN AECL APTOS sv.) svg Ae I ra Lad © ais ne sie. 4.0 6 aaa ea 43 2 Loin ion eRe eee 6.oe doen, - Sec cent: open te enemas 6 o ocak 44 SSIITOIST. ach eno Peete oe E55 Stl, 3 ee hs te cen Se cc o.o rota 51 ESL ES ES ee a: gee On ot = cE 52 1 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 JANUARY, 1914 2 Heals “LwNID: INTRODUCTION While studying the phenomena of regeneration and structural regulation in Bursaria, it became desirable to know something of the relations of this organism to its food and of the processes which the solid food material undergoes in its passage into and through the cell and in the elimination of residues. ‘This paper aims to present the first part of these observations and to make a general survey of the relations of this organism to food, leading up to a more detailed study of these and certain other problems of the cell as found in this unicellular animal. Bursaria was found to be much more favorable for the investi- gation of these phenomena than the smaller infusoria such as Paramecium. No study has heretofore been made of the rela- tions of Bursaria to food, so that the facts herein presented are new. The present investigation attempts not only to determine qualitatively whether the relations to food are similar to those more or less known for Paramecium, Stentor, Vorticella and other infusoria, but more particularly to work out and express these relations in a more quantitative way than has been done heretofore. It was found that certain kinds of experimental tests such as those on the rate of digestion, could be made upon this form, which it would have been difficult or impossible to carry out on smaller unicellular organisms. Its very large size offers a a singular opportunity for easy manipulation in many kinds of | work. When in a clear medium it is readily visible to the naked eye at a distance of six or eight feet and individuals may be transferred singly with a pipette without the aid of any magni- fying instruments. Bursaria occurs not infrequently in cultures brought from ponds in the vicinity of Baltimore, though it is less common than many other Protozoa. It can readily be cultivated in large culture dishes in the laboratory. In this way I have had abun- dant material at my disposal for many months. The method of cultivation has been simply the inoculation of an infusion of timothy hay in tap water from the wild culture; by several in- oculations at different times one usually succeeds in obtaining large numbers. Since the food of this organism is not bacteria, RELATION OF BURSARIA TO FOOD 5) so far as I have yet observed, but a variety of other ciliates, flagellates and rhizpods, it is difficult to find a culture medium which can be readily manipulated, and hence pure line cultures can not be obtained so readily as of a form like Paramecium. The problem of pure line cultivation of this organ'sm will not be dealt with in this paper. The material for use in the follow- ing experiments has all been obtained from mixed or ‘wild’ cul- tures, though the reinoculations from the single parent culture brought into the laboratory seven months ago resulted in a small number of pure lines living side by side in the cultures. It is, in fact, preferable in some ways to use material from such wild cultures for the kind of experiments to be considered in this paper. Even without the aid of pure cultures or the application of statistical methods to wild cultures it soon became apparent that there are actually at least two very distinct races of Bur- saria which differ in several diverse characters, physiological as well as morphological.! One form, which under certain food conditions has a tail, has been used exclusively in these experi- ments, since the other form, collected at the same time and lacking a tail, died out early in the experiments. I have observed the following organisms to be eaten and di- gested by Bursaria: Chilomonas, Colpidium colpoda, Vorticella and some of its relatives, Oxytricha, Stylonychia, Arcella, Sten- tor, Paramecium, Stephanodiscus, and some kinds of rotifers. Only once have I observed bacteria to be eaten, and that time in the form of zoogloea. It is, however, certain that bacteria form only a small part, if any, of the usual diet of this organism. The smaller ciliates, flagellates and rhizopods are the favorite article of food. The larger organisms, such as Stentor, are sel- dom successfully captured. Paramecium is quite commonly eaten, though Bursaria does not seem to thrive well on this food. Occasionally rotifers are eaten and it was observed on several occasions that these may remain alive within the vacuole 1 T have been unable to find reference in the literature to more than one form of Bursaria. A consideration of the problems connected with the existence of diverse ‘races’ of Bursaria will be left for a later time. 4 He Js. LUND for as long as five hours before they are killed. It is, however, plainly evident when one follows the development of wild cul- tures from day to day that some forms are eaten in greater num- bers than others and if the smaller forms, such as Colpidium, Vorticella and Arcella, are present in abundance along with such forms as Paramecium, Stentor and Stylonychia, the former kinds serve exclusively as food for Bursaria while the latter are rarely eaten. When the cycle of development of the culture comes to the stage where, for example, Paramecium is in superabundance, then the body of Bursaria may be more or less filled with Para- mecia. In contrast to the above mentioned forms, Spirostomum ambiguum was always rejected. It was often seen to be taken into the oral pouch but invariably was thrown out again, while Paramecia present in the same culture were readily eaten by the same Bursaria individuals at the time of the observations. This is the only case where Bursaria was seen definitely to discrimi- nate between two different forms of Protozoa. By simple methods of observation like the above, it would be impossible to determine just what the principle and the fac- tors are that determine whether Bursaria will feed on only one or several or all of these forms if they be present in all the cul- tures simultaneously, which of course they often are. It is with the object of elucidating these and certain related questions that the following comparatively simple experiments have been per- formed, by limiting and determining to a high degree the condi- tions under which this organism will react to food. ACTION OF THE STRUCTURAL MECHANISM FOR FEEDING AND THE SELECTION OF FOODS An account of the food relations of .Bursaria requires us to examine in some detail the objective processes involved in feed- ing; these are very striking. The highly developed oral appa- ratus with its large cilia, when in operation during the feeding process, may easily be observed. When the organisms are fed on such substances as yolk or starch they usually sooner or later become quiet for a time, and settle to the bottom of the dish or stick to the surface film of the water, then they may be ~ RELATION OF BURSARIA TO FOOD oO observed under a high power of a binocular. Granular sub- stances of different chemical or physical properties may be placed in the medium and the path of each individual particle may be easily observed. Mixtures of these substances may also be made and the paths of the different kinds of particles may be determined. The different paths of particles which come into varying rela- tions with the organism are shown by the arrows in the outline drawings of figure 1. The paths of the arrows are correct repre- sentations of the paths taken by the different kinds of particles. In general the paths taken by particles may be distinguished according to the following outline: 1. Paths of rejection a. Path of total rejection, arrows A 6b. Path of rejection of larger particles taken into the oral apparatus, arrows B c. Paths of rejection’ of smaller particles taken into the oral apparatus, arrows C, and C,. These paths may also be slightly modified by a combination of the avoiding reaction with the different rejection reactions. 2. Path of acceptance of large and small particles (large arrows D) Path A is taken by those particles which under conditions hereafter to be considered (p. 29) never enter the oral apparatus and are only drawn towards the body by the current; for exam- ple, very toxic particles of yolk. Path B is always taken by those particles which are too large to pass out by way of path C, and C, and must be passed back to the exterior by the same way as they were taken in, in order for the organism to get rid of them at all. This may be illustrated by the larger properly treated grains of hard boiled yolk. The path represented by the arrows C, and C, has considerable range of variation in part of its course. It may be illustrated by cornstarch grains; these are of convenient size. The variations in the course of these particles may be divided roughtly into two main divisions; some follow the dotted arrows C, and never directly retrace any (op) E. J. LUND Fig. 1 J, Outline drawing from dorsal side of Bursaria, to show position of the oral pouch in the body, and the paths of variously rejected and accepted particles. A, path of total rejection; B, path of rejection of large particles which are too large to pass out by way of the oral sinus, S.C, and C., paths of rejection of small particles; these pass out by way of the oral sinus, S; D, path of accept- ance. JJ, Outline drawing of sagittal section, in the plane through the body represented by the straight line through 7. A, path of total rejection; D, path of entrance of all particles taken into the oral pouch, same as first part of path D, I. C, path of rejection of small particles, same as C; and C2 of J. E, I the same as # IJ, direction of rejected particles, C, 7, and Ci, C2, IZ. part of their former path; some pass up into the proximal end of the oral pouch but are rejected and returned to the outside by way of the continuous arrows C,. All the paths of rejection under C, and C. converge and lead to the exterior by way of the oral sinus, figure 1, J and JJ S.; they then pass backward under the posterior ventral side, as shown by arrows E. There is but one path of acceptance for both large and small particles. RELATION OF BURSARIA TO FOOD ri Figure 1, //, is a sagittal section through the body in the plane indicated by the straight line through figure 1, 7; it shows the path of total rejection, arrows A, the path of entrance (by heavy arrows D) and the path of rejection of smaller particles (by arrows C). At the point of entrance into the endoplasm the transport of the accepted particles is brought about not alone by the cilia but also if not exclusively, in the case of larger particles, by a peristaltic wave in the wall of the oral cavity behind the par- ticle pushing it into the body. FORMATION OF THE VACUOLE AND THE ELIMINATION OF RESIDUES The vacuoles when formed always contain some liquid, though at times the amount may be very small. The size and shape of the vacuoles varies greatly and depends upon the kind of food eaten, and upon many other conditions, as will be shown. Often the food forms large irregular masses, which in the case of fresh yolk may so completely fill the body that after a half- hour or more of feeding the dorsal side of the body cortex is burst open and the food mass is extruded. The opening then closes and the organism again assumes its usual shape. The rate of formation of the vacuoles is intimately bound up with the same complex conditions which determine their size and shape. The circulation of the vacuoles in Bursaria is not re- ducible to any definite order, such as has been shown to exist more or less definitely in Paramecium, by Metalnikow (’12) and others, and in Carchesium by Greenwood (’94). The vacuoles often become lodged in one place and there digestion is com- pleted. This may often be seen in cases where fat-extracted yolk particles become lodged in the tail. During digestion and resorption the large vacuoles usually become smaller and any residual contents are finally extruded. The residues are always extruded from a small area on the mid-dorsal side of the body of the organism. ‘This may readily be demonstrated by feeding ‘ the animals chinese ink. The changes which take place in the food vacuole from its formation to its disappearance will be considered in detail in a later paper. 8 E. J. LUND MEASUREMENT OF THE AMOUNT OF FOOD EATEN AND METHOD OF EXPERIMENTATION In order to express quantitatively the relations of Bursaria to food, it is necessary to obtain a reliable method for measuring the amount of food taken in a given length of time, under given conditions. The unit of volume employed was that of one grain of fresh hard-boiled yolk of hen’s egg. The eggs were boiled fifteen minutes. These grains are readily eaten by Bur- saria and may be obtained of an approximately uniform size. It is necessary to deal with suspensions of such grains, having a uniform concentration (that is, containing the same number of grains to a given volume). For this purpose stock suspen- sions were made up on suecessive days from yolk of the same egg: these were made uniform by making them up in vials of the same size and comparing each with a standard concentration kept in a vial of the same size. Various known grades of con- centration were then made up by adding a known volume of the stock suspension to 5 ec. of water in a stender dish of 8 ce. capacity. This procedure was found to be sufficiently accurate to avoid the introduction of any observable variation in the © amount of yolk eaten in a measured period of time (page 20.2) Uniformity in the size of the yolk grains was obtained by repeat- edly washing the fresh hard boiled yolk crystals in distilled or tap water and decanting, until the suspension when left to settle leaves a clear supernatant liquid. The smaller grains remain in suspension a little longer than the larger ones and thus may be removed by decantation. Uniformity in size is still further obtained by drawing off the grains from the same level in the clear suspension with a pipette. Some eggs have yolk crystals of more uniform size than others, so that only the eggs best in this respect have been used. 2 In most of the experiments it was necessary to make up only a single stock suspension, since the animals were fed only once and all the feeding was carried out at the same time. In the case of experiments which required the feeding of yolk on more than one day, however, this standard concentration had likewise to be made up anew each day by comparison with that of the day before. RELATION OF BURSARIA TO FOOD 9 The uniformity in size of the yolk grains is of course of para- mount importance in many of the experiments and for some of the conclusions which will be drawn from them. In order that the degree of uniformity might be tested and indicated quanti- tatively, a large number of measurements of grains of the pre- pared yolk suspension were made at different times by means of a stage micrometer. The following shows a typical result of one set of these measurements: 105 grains were measured and the numbers divided at random into three sets of 35 each and the average of the diameters taken. These gave respectively 0.0890, 0.0906, and 0.0837 mm. The range of variation of the diameter was from 0.060 to 0.130mm. The distribution of the variations are shown by the following figures: Diameter of grain mm......... 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.18 BRREHMCN GV e one co). 2 cicis.s se s!e soe os 12 19 16 14 20 13 6 5 By thus being able to obtain a very constant average diameter of a comparatively small number (30 to 50 grains) the errors introduced by the individual variation in size, which in the above example is about as 3 is to 1, is largely eliminated. In order to remove the objection to experimental results based on the vol- ume of granules of varying size, a large number of individuals (20 to 100, depending upon the purpose of the experiment) were used in each experiment and the number of grains eaten was counted; furthermore the experiments were always repeated whenever there could be any doubt as to the validity or signifi- cance of the results. Hence, as will be shown later, the lack of strict individual uniformity of the unit volume is corrected (a) by the fact that the average size of the yolk grain is practically constant, (b) by using a large number of individuals in each experiment, and (c) by repeating the experiment. _ Thus having the form, weight and volume of the units of food eaten made practically constant, we may vary one of their prop- erties—as for example, their chemical nature—by letting them adsorb different kinds of toxic and non-toxic substances which are diffusible or non-diffusible in the native medium, tap or dis- tilled water. We may therefore test the responses to variations 10 E. J. LUND in this one property—namely, the chemical nature of the grain —and its effects. An approximately constant medium was provided by using tap water. This precaution is important, for, as will be shown, the nature of the medium often affects or determines the kind | of results which are obtained. Distilled water was also used but it was found that this extra precaution was not necessary in most of the experiments, and since distilled water is toxic if the organisms are left in it too long or the change is too rapid, it could not have been used in many of the experiments, even if it had been otherwise desirable to do so. The organisms were starved in 400 cc. of tap water for eight- een to twenty-four hours previous to each experiment. At the end of this time they were free from food and residues. Thus an optically clear, active and perfectly normal cell was obtained with which to begin work in all the experiments where uniform- ity in this respect was desired. All the factors with which we are dealing except the ‘physiological states’ of the organisms themselves are known and uniform to within narrow limits, while the one of which we wish to test the effects can be controlled and varied. INTERNA] RELATIONS AFFECTING THE FEEDING PROCESS 1. The relation of the physiological state of the organism to the feeding process | By the words ‘physiological state’ is here meant the condition as a whole, of the equilibria in the physical and chemical reac- tion system, the cell, at a certain time in the duration of its existence.’ This condition or state is to be thought of as being limited to the space which the organism occupies, or is, in other words, internal. However, it is obviously absurd for anyone to attempt ’ This definition is justified because in so far as the facts are at present known, this is the only kind of system with which we have to deal in the cell, and therefore in the present state of knowledge the only logical universal assump- tion for experimental purposes is to define ‘physiological states’ in terms of such known systems, until the universality of the assumption is disproved. RELATION OF BURSARIA TO FOOD ia a definite and strict separation of the internal and external of -any living organism, and especially is this true of the cell. Yet for purposes of presentation, this becomes highly convenient, and it is only for this purpose that the above rough distinction is made here. When all external conditions are made the same in two experiments which nevertheless give different results, the differ- ences must be attributed to different conditions within the organism, and it is, as a rule, only in this way that different physiological states are at present practically perceptible. Differences in physiological state in unicellular animals are made evident most readily in the relations to food, as may be seen from the work of Metalnikow (’12) on Paramecium and by Schaeffer (10) on Stentor. Bursaria affords most excellent material for the elucidation of the relation of these dynamic states to the feeding process and of the fact that this relation changes while the external con- ditions remain constant. These facts are brought out in the following experiments by using both single individuals, and large numbers of individuals collectively, at the same time, and analyz- ing the results. The total quantity eaten and the rate of feeding. Table 1 gives the results of a typical experiment designed to show the difference in the total quantity of food eaten and also the differ- ence in the rate of feeding of Bursaria from different cultures. Material from two different cultures, A and B, was starved twenty-four hours in each of two dishes containing 400 ce. of tap water. 1 cc. of a fresh hard boiled yolk Suspension was placed in each of 16 stender dishes of 8 ec. capacity; 5 ec. of tap water was then added to each. Thirty individuals from culture A were placed in each of 8 of the dishes. Alternately with these 8 sets from culture A were placed 8 sets of thirty individuals each from culture B in the other 8 dishes. At the end of the time intervals noted in the table, in each case, the contents (6.5 ce.) of one dish each of A and of B were transferred into a stender dish with 500 ec. of tap water. This stops the feeding. The individuals were then immediately picked out of these large dishes, placed in 8 cc. dishes and killed in Meves’ fluid. The 13 E. J. LUND counts of the number of grains contained in each individual were taken at the end of the experiment. Table 1 shows (1) that Bursariae living in the two different cultures differ in the total amount of food eaten in the same length of time. In other cases, of course, individuals from di- verse cultures will give identical results so far as feeding is con- cerned, while two or more different cultures may also differ to a greater extent than the above table shows. Moreover, the amount of food eaten by a given culture may vary at different times. The greater the length of time of feeding (within certain limits) the greater the total amount of food eaten. Not only does the total amount of food taken differ in the two cultures, but what is equally important, (2) the rate of feeding varies with organisms from different cultures. This was observed in numer- . ous other experiments. Under some conditions the animals fill their bodies quickly, while at other times this takes place slowly; or only a small number of grains may be eaten. The facts are shown most clearly by the curves A and B, figure 2, representing the number of grains of yolk (ordinates) eaten by the thirty individuals in successive periods of one-half minute (abscissae) throughout the time of the feeding process. Curve A is plotted from the results of culture A and curve B from those of culture B, in table 1. The immediate rapid rise of curve A shows that the rate of feeding of culture A dur- ing the first six successive periods of one-half minute each was about from five to twenty times as great as in any of the subse- quent fifty-seven minute intervals. A similar high initial rate is also shown by curve B (culture B), but here the rise to the maximum was not so steep and the rate during the first six half- minute periods was only about from four to ten times the rate during the subsequent fifty-seven half-minute intervals. In order to show more clearly that the results apply to the individuals taken separately as well as to the averages for all (i. e., to the cultures as a whole) the data may be arranged as in table 2. As this table shows, at the end of sixty minutes all but an insignificant number of animals from each culture had eaten yolk grains: hence, the difference in the amount and 13 OF BURSARIA TO FOOD RELATION = oO rc coc Owinrre re oo mi OD 0° OT 0°CE ONICHA 10 ALVU ADVUAAV = SALONIN % AAISSHOONS NI NGLVaA SNIVUD JO UTAWON COVUTAV i IVOQGIAICGNI udd NG@LVG SNIVUD | 10 Ue | -WON GDDVUAAV OST I8¢ LIT 9TE Vit 6G 16 va TZ GG 8g S6I GG OL OL GE NGOLVa SNIVUD WO vices WON IVLOG | g pun vy 0688 & Zziet9 PTT cr 800 G Z Ozi6 0 O10 911 i 6FEeE 808 le 6 6LFIS STZISTZIGT Le 2 1 \¢ It S 0 F OLFLE 91 0 0 9 zor ee 0 | 2 9 (8 loreto [2 r Ie |r (9 I EO |P E10 (210 |E i (9 0 | 200 01 II 0 0 | eos Fh zl 0 (0 (0 |r jo |r jo 0¢ z 0 1 OT & 8 Fg 6 66ST 0 11/6 cr st 6 6 IL | 6 OLE, 0 G le" | le 8 | G 0 IS OF 9 40) IP iL |G 0€0 0 |t |2 0% 0 IT 1% [2 I z v ~ oo os 219 ¢ 0 FEL 9 IT 10 z 6 Ore it 0 216 3G 0% 00 9F It 0 0 F nN bt I | NN To) I = ~ Yr © _ S = x fo) f 918 0208 Gir % L\S |v 9 § li aa: 20 9 91¢ 7 € % L1Z16 £00 9/8 ¢ Z It (0 900 0 [ 0 0 \¢ % Ol OL 1109 HOVE AM NALVA WIOA HSAUA FO SNIVUD FO UAAWON ‘saunyjna qualaffip on) woLf yopa syonpravpur PLPI 8 GI 00 GI\Ps — ~rY~ oO ooo — = 8 9 0 91 le |9 an Til OLFT O1 rab g 8 0 0 II (0 6 OI 00 Zt 00, 00 ayo -1Nd 0G sana | =NIW } OMe | aot | aM fijsvy) fo sjas qybva fiq wajva y)oh fo suwab fo saqunu 7pj07 ay) pun Burpaaf fo anu ay? fo Uosispdwmos v fo s})nsat ay) Buinoyy LOTav.L 14 Bees. LUD rate of food taken by the two cultures was not due to some sporadic difference caused, for example, by a very high rate of feeding by a few individuals and no food eaten by others, but rather to a uniform difference between the sets of individuals from the two cultures. Therefore the results are typical for the individual as well as for the culture as a whole. Moreover, if we calculated the averages of A and B on the basis of those individuals alone which had one or more grains, the average of A would still be greatly in excess of that of B. - Grains 10 20 30 60 Fig. 2. Showing the rates of feeding by the two cultures A and B, curves A and B, respectively. Plotted from the results of table 1. Minutes 313 5 We may express the variation in the total quantity eaten by the standard deviation of each corresponding group of thirty individuals in A and B, as is done in the last column of table 2. The reciprocal of the standard deviation (c) is a measure of the degree of uniformity among the individuals. It will be noted that there is an increase in the range of variation and the stand- ard deviation with increase in the length of time of feeding; this means that the difference in physiological state among indi- viduals of the same culture finds a fuller and more definite ex- OF BURSARIA TO FOOD RELATION | | | | | | | | She Lith C| PILLS Pe] © ei it| je | Aouenborg | of€ re | ai | ole jel | rece r b) | Aouonborg | 20% eZ T | eels 1) | 1 i) ft | i ee 8 2 Ot} Aouenberg | ofl 680 | be haeleale Pa Les es | IE IT gz) Souenhbosq ‘D001 + =5 nee 7a = GAOPVY FHL HLIM : a al | ise NOSINVdWOD UOT GADNVUUY (3% aDVvd) INGWNINGTdIXA AHONLVATdINAL AO SLTASaU | RA | | | ei| [ie [ a] lie] | | 88°F | Pe eae lhl | Eee ade sel ial o z z ¢€| @ | Aouonboag | g 09 COS Be ico a aa a | | | iE) |e rt ig ie 9 {t | j8| Aouenbory | g | og f° 8 | | ey | ted ste | LP] ©] Br IL 9 9 | souonborq dpe |e 69'S fe aes eae a | | es | TT) @ Tie \2 ¢ Fg) Aouonborg os ie O01 el T | | sles et | || @|Li¢ 9p l9| Sowonbsrg | gg G 88'T | hari ies et nes la IL) |p \¢ |¢ jb ot) Aouonbosrg Elgoaitat.e te I | | | | | | | | ee | iz iz le |e leq Aouonbo.ay q | iT 69°0 | (ea eke tele leelcall nen we | | | 69 Aouonboig i 9F T} | |] EMER] we] | Ree ee) Te) wn |g Aouonbo.ay Virsa ae OO 8L'9 | G I Vi GIL IE IT IL IL] 8) |L IL) jg | Aouenboag V Og ee9 + || rae I ET) | Cee eit 6 LE iE @ iL] | Souenborg |v 0% org | a aria eli ieee dl el ee eimai ele ET | | fe | Aowonboerg | y feu eo F | | LE] le] Bich eer ee ie it) | | jo} souenbeng | iy G 86°F | ey | || IT) | TREC SeeEeeet~ep) Souenbesqy | y € Para | | Nall | | | IT | b| ¢\t i i2 |g | Aouonberg | V T CLT | i | I | | | an Anuonbo.ayy V pe) eae Madd GB FZ |83 <2 ty 0G ots L191) 61 #1 1a u or 6) 8 | F| | 2] pein pestis rapur UD sp pasn si ssavoad Burpaas ay, ‘“g pun y ‘saunqzyn re) 6 ATaVL on) WoLf Didosing fo ajn]8 JooBojorshyd UWL saouasaffyp ay) buanoysy 16 EE, J. LUND pression in the results of the experiment as the length of time of feeding is increased. The final total number of grains eaten when the time is long is then a more accurate index of the rela- tion of the physiological state to the feeding process than if the . time of feeding is short. The greater this difference in the total quantity of food eaten the greater is the difference in the physio- logical state of the different individuals. We, therefore, have in the amount of food eaten, if the length of time of feeding is long enough, a fairly good relative measure of the physiological state of the single individual and the differences in the physio- logical state between different individuals as regards their rela- tion to food at that particular time. 2. Changes in the physiological state as shown by using the feeding process as an index If change in the dynamic conditions of the cell, as regards the food relation, does occur, this should be observable by a change in the feeding process, and such is indeed the fact. This is shown in table 3. Material from cultures C and D was starved in tap water for twenty-four hours. Five active individuals were then picked out from each and tested individually. They were fed twenty minutes and each one was observed continuously during the experiment. The number of grains eaten and re- jected and the time as called off by the observer were noted.! In this way the time record of the relation of acceptance and rejection of food was obtained. The yolk concentration, tem- perature, and so forth were the same in all the tests. As the table shows, yolk grains were at first rapidly eaten. At the end of the first few one-half minute intervals the action of the cilia was frequently reversed, thus rejecting the food after it had been taken into the oral apparatus. There was, therefore, a definite change from eating to the rejection of food by the feeding mechanism. This change was more rapid in general, in the individuals from culture C than in those from culture D. *T am indebted to Mr. K. S. Lashley for kindly aiding me in taking the records of this experiment. RELATION OF BURSARIA TO FOOD LT | Dalene i Pele eeancall i T | | I | | °° poyooloy g | | | | ee ical | | (lial ae beatae I It |‘ paydeooy sf “ 02 ie RWS IE |LE4 |) ty | Oe SHE all a | je ee 9 || | Beanetae aie 8 | | bs ire | Dell GIT “ paydao0y ef 1 | tieehitle! tirl | eect! Fee | iz poqooloy LT | Tepe ede aa Pa el Te alate ale |atie ale la eet I Zz LP | perdesoy TH Lb SLEET) Tee) Cel Tr] €| eeeetrel| elt! « | _ poqoaloy 7 FI | | | I | I Te SP | | |e} | | | BIE] ie @ | pesdeooy J ce eee ES Se ee Ee esl Meee || Per eeee) tl) Tey | | | popooloy i int debe rele alae [eh ef | | | | ea | ai fel a) Be Se ~pardovoy J SINOY FZ POaAIVyG qd wuaLtTAgQ 3g judy | a | any | | | | | | 0G HC tp Set eae tt) | eehtece! thieheeiekee cepa) | | |: -poysotoy rn Bag, ed | eee eel tee al | | € 19 ¢ poydovoy Lg ee) re ze) eroece LIT Fo) @ fen Hevea ee ee eet ieel eh ae I | Paes | IT | | eared 2 phe Ha ele % TIL ¢ | poydaooy 0 |} eel) tl] e| tke eee) eer J} etoeerzctPeier| | | | | |: peqoetoy | aia Iz ai | aie eal | | ee sey ae pasech sale Poydeoan| VL Geer) CELE eC) CoS e|) eserarezt| proe| | / poqooloy U Il | | | | | | | | I | I | | fF ° poydoooy SIL PL] @iE4) OSs) | Lire) Pit) Tee) Eero | '* poqooloy a | [eee Ia ea reali le It lo |. |: poydooow I surpsb | ia | | | | | | | | | | A We J OY aU ed a Pn) 2 was 0% 6 ‘$1 jar 91 I) WI | €I a Il Or 6 8 | L 9 Is p | € | z A T\#| Sa0aNIW awit SPONPLIPUL WO $789], sinoy pz ‘dey “poaarryg QO @NALTAD :F judy € ATAVL 16, No. 1 » VOL. THE JOURNAL OF EXPERIMENTAL ZOOLOGY 18 Beare OND These results from individuals are therefore strictly comparable and in accord with those obtained when a large number are tested at one time (table 1). Now, in order to explain the cause of the change in reaction the suggestion might be offered that Bursaria shows a decrease in the rate of feeding because of the decrease in the amount of space in the body which food can occupy. This is undoubtedly true to some extent in those individuals which do not stop feed- ing until the cell becomes distorted by the comparatively immense mass of food. So far as the volume capacity of a normal indi- vidual of Bursaria is concerned, hundreds of observations have shown me beyond doubt that this may frequently be as much as twenty-five to thirty grains. Nevertheless, reversal of the cilia always takes place sooner or later. But the suggestion evidently ' does not apply to those individuals which show a change in the reaction when only a few grains have been eaten, for it seems impossible to understand how there could be a difference of as much as twenty grains of fresh yolk (table 2) in two normal individuals of equal size, when the cells are under exactly the same conditions, if this result were not due to a difference in the physiological state of the cells. Change in feeding was caused by the periodic reversal of the cilia and the reversal of the cilia in turn in some manner initiated or caused by a stimulus from the food already eaten, for it seems most natural to suppose that the stimulus originated from the change produced by the food mass in the interior of the cytoplasm. The most definite evidence that the change is due to stimulus from the eaten food is found in the radical change in the action of the cilia of the feeding mechanism. If such fed individuals as those in table 3 are left in tap water free from food they may again eat yolk after digestion is par- tially or wholly completed, and again show a similar decrease in the rate of feeding, that is, a reversal of the oral cilia. The total quantity which will be eaten may be greater than that eaten at the previous feeding; but it usually is less, or often none at all. RELATION OF BURSARIA TO FOOD 19 The process of feeding in Bursaria shows it to be a function- ally equilibrating system in its behavior towards food and the condition of its equilibrium at any particular time constitutes the physiological state which the cell is in, so far as its relation to food is concerned. The changes in the increase or decrease in the quantity of food eaten in successive meals and the increase or decrease in the rate of feeding might be discussed in the psychological terms ‘hunger and ‘satiation; but it is evident that the simpler terms quantity and rate express the facts of experiment, while any attempt at definitely determining whether the changes in quantity and rate are the same or different from ‘hunger’ and ‘satiation’ will obviously lead nowhere. Hence it seems better to use the terms, rate and quantity, which have a clear and quantitative meaning. 3. Other causes of individual variation Bursaria at times closes up its oral apparatus. This may take place to such an extent that the opening is smaller than the food particles and then the latter can of course not be eaten. This condition can readily be observed under the binocular and it can always be determined beforehand whether closure has taken place to such an extent that the organisms can not feed. Other minor accidental individual variations are also present to some extent. These may be partly due to the difference in the size of the grains of yolk eaten. Sometimes when an individual is weak, owing to prolonged starving or for some other reason, two or three grains may become stuck in the oral pouch and this prevents feeding until the animal succeeds in throwing them out or by other means they become loosened. The material used was always examined beforehand to make sure that it was in a healthy condition so that these accidental conditions play no part in the final results of the experiments described. Such a series of experiments as the forego'ng do not show us specifically what these complex conditions are which have been cloaked in the phrase ‘physiological states.’ This however is 20 Re 0: LUND not the object of the above experiments: they are only here considered for the purpose of demonstrating the existence of these conditions, the fact of change within them and especially in this connection their rdle in the eaternal phenomena of feeding and food selection in Bursaria, and how they may affect the results which will be given in the following pages. EXTERNAL RELATIONS OF THE FEEDING PROCESS 1. Effects of external factors on feeding a. Concentration of the food supply. The rate of feeding is within comparatively wide limits not dependant upon the con- centration of the yolk suspension, provided it is not too low. This may be illustrated from one out of a series of experiments. The time of feeding was reduced to five minutes for the purpose of bringing out the effect of difference in the concentration more strongly If the animals had been left in the suspensions twenty minutes (the usual time of feeding; cf. table 3) the difference would have been less marked, especially with material which shows a high rate of feeding. Experiment I.° Material from a healthy culture was starved twenty- four hours in tap water. All were perfectly normal and active. The experiment was carried out in 8 cc. stender dishes. The concentration in dish B was 8 times that in dish A. Twenty individuals were placed _mm each dish. The results from trial number 2 represent more nearly the ideal because these two suspensions were kept uniformly distrib- uted during the five minutes feeding, and the individuals were picked out alternately by fives. Both trials, however, express equally well the proportional effect of concentration, namely, 1 to 2, as compared to the proportion of concentration, 1 to 8° (table 4). The concentrations used in the experiment are approximately represented by figure 3. ° The experiments given in this paper are numbered in regular order for the convenience of the reader, and do not represent the actual order. Only a small number of the experiments actually carried out are given. °In all the experiments considered in this paper, where the concentration plays a part, the concentration was intermediate between those used in this experiment (fig. 3). RELATION OF BURSARIA TO FOOD 21 TABLE 4 Experiment I AVG. PER TRIAL DISH NUMBER OF YOLK GRAINS IN EACH INDIVIDUAL TOTAL IND. | Z | ite (Aales | | Grains , {| A i] 31) 1] 5,5, 112 1] 3 1000110 3,9 34) 32 1.6 ite 363744751 20231433138 62 3.1 | SAME YOLK SUSPENSIONS USED . ; _ ae —— Se ete. A a 4 21 1/1) 2) 11 5 5 3.0114 0 4 30,4 a 2 45, | 2.25 ect) 8 18614438415 767241072 81 4.05 A B Fig. 3 Showing the relative concentration of yolk in dishes A and B of Ex- periment 1. b. Effect of mechanical stimulation and of mechanical injury on feeding. e Experiment II. ‘Thirty individuals from the same culture, starved twenty-four hours, were placed in each of six 8 ec. dishes containing 5 ec. tap water. Before feeding, the animals in three of the dishes (Set 1 in the experiment) were mechaniéally stimulated by means of a pipette. The opening of the latter was about ten times the width of Bursaria. The edges of the opening were made smooth by melting. The animals of Set 1 were stimulated by drawing them up into the pipette along with the tap water in the 8 cc. dishes, four times. Equal quantities of yolk suspension were now added to all the dishes. After having fed ten minutes the animals of Set 1 were again stimulated by drawing them along with the yolk suspension into the pipette two times; at the same time the control, Set 2, was stirred by gently shak- ing the dish and not allowing any instrument to touch the animals; hence the distribution of the yolk was the same in the two sets of dishes. All the individuals in Set 1, after having been stimulated, were perfectly normal and not injured. They looked like those of the 22 E. J. LUND TABLE 5 Experiment II Set l. STIMULATED DISH | NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL nee Jit SNe z aie ——- | | fea a | | grains | | | | | | At Jo 1401013 1151120110111 110,212 1 301000) 33 1.1 By case cures 110020000 202000 200 30 01/01 1/0 0 0j01 1) 17 0.56 let Cre sthnnaceee [0/012 122 601111) 3/3 1 111 13,00 403 O10 3 3) 48 | 1.60 | | | icadeeed : a : a - = : Total average. ..........----.-------- pOgEsgeSucceg cue eta deen dbo Josocse dacor See 1.08 Set 2: ContTrRou: Not STIMULATED DISH NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL aes i | | | | a | bia | grains | Ne! | | | | A (357.2363. 5,41419.3\ 121582959 510|377 643.42 157 5.23 Bias henenee 2} 44°36) 8 3} 8} 1) 3) G4) 4) 5) 2 312) 5) A 4) 2) 6) 6) 1) 3) 2) 7) 2 BO ae 3.63 Geet ener 6| 1] 3| 6 5| 6 5| 3| 2 7 4} 4! 4! 3) 3] | 5 4! 3) 2) | 1] 3} 2) 3 5) 6 2 2) 1) 113 | 3.76 | | : * _ | ae _ Total average...... Sieleinr= riot agtehe gee Wiobesdacsc Aedes caacIo OAT OOS. legen gece bebe >: 4.206 control. If a smaller pipette is used or a larger one, and the stimu- lation, by sucking them along with the medium up into the pipette, is more violent, it will strmulate and injure the organisms s9 that they will not eat at all, or at least, not for some time after stimulation. Of course structural injuries are very easily produced, with the result that regulation of the cell must take place before any food can be eaten (table 5). Proof that in this experiment, Set 1, if not in Set 1 of Experi- ment III, the organisms were not injured beyond the capacity « for swallowing, is found in the fact that the great majority did eat, though only a comparatively small number of grains An- other experiment may be given to illustrate the same fact. Experiment III. The animals in Set 1 were not stimulated before feeding, but after they had fed for five minutes they were stimulated by drawing the suspension with the animals in it, up into the pipette only once. Material from a different culture was used in this experi- ment; time of feeding fifteen minutes. The control suspension with the organisms was redistributed once by gently shaking the dish. The animals were all normal in form at the end of the experiment (table 6). In Experiment III the stimulus was only slight as compared to that in Experiment II, yet the effect was marked As stated above, strong stimulation may totally prevent feeding. RELATION OF BURSARIA TO FOOD 23 TABLE 6 Experiment III Ser 1: STrmULATED DISH NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL \TOTAL AVG. PER | YD. she ae WATE a ASAE grains ts ieee (5416869070103 0/1/3/3/401)4 6515000400 96| 3.2 2. ea (00.40 41511 00.901 465.258 41096 6 118 5 60 2 0) 127| 4.0 Ge: “904021 3503101004 2031330800304 59| 1.96 || | | We teste) Total average........2..... 5 brine Leecerenstee pars boaognoodone BORED Ge BE GeCRaeMeD Boh Set 2: Conrrou: Not STIMULATED DISH | NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL) cae | ND: | | | | Me Hey | | | | | grains | | | en | | | | Ve | 69 8 6 6 410 7 65 97717 25 9 914 6 7.6 7 6 O10 7 9 6 201| 6.70 B... | 6 24 66 1/7 9 5/46 4/6 2 7 1 610 0 4) 7) 2 5) 5] 5) 6 7106 3) 152) 5.06 Ch cseeaeeeee | Ot 4) 6 6 9 8) & 6) 9 6 5 5 9 7) 7 9 2) 7] 4) 7] 8 8 712 6 8 6 9) 1 198 6.60 | Seed We La a et ____ Total average eee focopwmece eneDoE eee ee SEGvaRoE bovoogsscusesmatormee Pont The effect of mechanical stimulation must be emphasized be- cause it shows that in any work of this nature it is necessary to handle the organisms gently. This relation must be inferred to apply to work on other Infusoria also, at least to some extent. c. Effect of temperature on feeding. Experiment IV. ‘Thirty individuals starved twenty-four hours, were placed in each of six vials. Each vial contained 5 ce. of tap water. These vials were now placed in large dishes containing water kept at the desired temperatures. The latter were read on a smal! thermometer set inside of each vial. Equal quantities of fresh yolk suspension were added when the temperature had reached the desired point. They were fed fifteen minutes (table 7). - TABLE 7 Experiment IV NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | TOTAL deg. C ae ie Liane lotic | | betel Gi eee 1 0}0.000000000000000d000 00000000 0 3 ea 00000200000 000000090010000000 3 CC ee ce ae 20/131 1/34 210006 4/0 3/5 1102 3110 100404 52 lle Veep 912, 419 6 9/8 2 016 21918 6 5.9 204.24 54.8 7 9675 8 195 10) 2 aie 010| 3 411 6) 913 6151316 4 612 0 915 5/4 813/913 4/2 4 8109 241 39) tor “Soébe Se BBaeeaee All died in from 5 to 10 minutes 24 E. J. LUND The experiment was repeated with closely similar results. At lower temperatures the animals are always unable to eat. As the temperature is raised and the activity of the cell increases, the rate of feeding increases, continuing to increase nearly up to the point where the cell is injured or killed by the heat. At temperatures between 20° and 25°C. (i.°., at about the optimum) the increase in the rate of feeding can be determined only by using a very large number of individuals, since the variations obliterate the effects when a small number is used. All the experiments relating to other conditions were carried on at temperatures ranging between 20° and 27°C. Where neces- sary (as in prolonged experiments on digestion) the temperature was kept constant to within 1° to 1.5°C., throughout the course of the experiment, by keeping the organisms in moist chambers in a constant temperature oven. d. Effect of HCl and NaOH on the feeding reaction. Experiment V. The medium used in this experiment (table 8) was conductivity water.’ Any water less carefully purified is worthless for such experiments, as was shown by experiments carried out with tap water. By comparing the results it was strikingly evident that the acid and base had reacted with the salts and other impurities in the tap water and hence their effect was removed in low concentrations. The animals were washed once in conductivity water before putting them into the solutions. Time of feeding, twenty minutes (table 8). It is seen from table 8 that the base NaOH was much more toxic than the HCl, and that as the concentraton was increased the number of grains eaten became less and less. The chemical relations of the food and medium will be considered in more detail, later on (p. 29). e. Effect of strong white light on the feeding reaction. Bursaria, when kept in dishes with a rather clear medium, often collect in the greatest number on the side of the dish away from fairly strong white light. It therefore became of interest to test what effect continuous light of a high intensity would have upon the rate of feeding. 7 Prepared and used in the Department of Physical Chemistry for conductivity measurements. RELATION OF BURSARIA TO FOOD 25 TABLE 8 Experiment V NaOH é : nbs . TOTAL ‘ MOL. CONC. NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL NUMBER GRAINS 1/400 ne All Ready MPO MUN TUES es ee a aes 0 MEMO REE nice o's. 6 2": A Geadyin: 1Oominubes:1- Lee nosed aed. 0 W/SC066 6-2 eee Many dead at 15 minutes} none eaten at end hi ZU ARTIC DER) Sa eee aes ont Cette Se 0 UA 0 All alive and normal in shape at end of ex- periment; no grains of yolk eaten......... 0 i 0 10004 1) 0 2 1) 2 20 13 20 3100; 19 PRI i. 6 5 8 7| 3513 19 5) 510 9 611) 4 7 5) 8 121 1/0420, re 5 9| 8 8 614 8 aaa 5| 8| 5112 4! sii 6 8! 6 5| 148 CONTROL IN CONDUCTIVITY WATER 7 910 9 9 4) 51010 6 6 612 4 16 5 6178 8 4 155 H Cl OD) All Aas iUal 25} SeTWNANI RESIN oer lg Gd Rei oregus ds meee See | 0 Mea... FUAIVUA0AV4 201 205 4 1/2) - 32 (ol =e 134513 4552605443375 5 6 90 J 6 8 1 8 7 8 6 0 6 3 8 3 1) 5 4! 1) 1) 21} 1 90 J ae 6 5 8 8 7/9 312 3 2 5 814 6 9| 3) 5) 5/075) 128 MAMAN os 2. 8) 7/15| 511/13) 8| 9] 0} 2| 4 113.8 21051 5168; —«130 ae 5101210 9 7 9 314 61059 910107953 162 CONTROL IN CONDUCTIVITY WATER 1212 413 O1114 5 6 2 0) 9} 7 8} 211) 311! 3) 4 137 Experiments VI and VII. White light from the are of an Edinger apparatus was focused upon the stage so that a spot of light 13 inches in diameter, of a very high intensity, was obtained. The light was filtered through a layer of water 1.5 em. in thickness. An 8 cc. stender dish containing thirty normal individuals was placed in the spot of light and the usual quantity of yolk suspension added. A control was kept in weak diffuse daylight. The animals were fed twenty minutes. The following results show that continuous action of intense white light on the animals does not have any effect upon the rate of feeding. Two experiments with controls are given (tables 9 and 10). 26 E. J. LUND TABLE 9 Experiment VI STRONG WHITE LIGHT NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | TOTAL 5) 6 150 ; | | | i i | | | | i i | | | | | | grains 2] 6 5| 61] 8 2 5| 8| 81 2| 3] 5| 6| 6 3| 6 3| 3 7/ 2 7] 6 6 5) 6 1 | | } | | | | | he | | |_| | el tu Koo | | CoNTROL: DIFFUSE DAYLIGHT Sp ca as TS, a = S5643oodoasdiz4s373677 80747925 146 } | Ih il PP > : | wad TABLE 10 Experiment VIT STRONG WHITE LIGHT NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | TOTAL ihe aa ei) ae | | | | | | SA bal } e grains Hosraasssi ce | 4i11] 7| 7] 0 2 W444 ons O5 154 Connon: ‘Dirruse DAYLIGHT 8 82 E 310 0911 8) 9| 2 7| 8 1) 0) 415) 0 8| 2 810 6 7 6 811) 3 4 5 188 f. Effect of the electric current. Weak induction currents have, within a limited time, no noticeable effect upon the feeding, as is shown by the following results from two separate experiments, VIII and IX. The total number of grains eaten by 20 individ- uals in each of two 8 cc. dishes is given in each experiment :° Experiment VIII Dish A—73, total number of grains eaten by 20 individuals Dish B—69, total number of grains eaten by 20 individuals Experiment IX Dish A—65, total number of grains eaten by 20 individuals Dish B—64, total number of grains eaten by 20 individuals Control for Experiments VIII and IX; not stimulated by the current Dish C—89, total number of grains eaten by 20 individuals 8’ The apparatus was arranged in such a way as completely to prevent any effect of substances liberated at the electrodes, by inserting the electrodes in a physiological normal NaCl solution in each of two 8 cc. dishes and from these the circuit was closed through the other two 8 cc. dishes containing the animals, by means, of small N tube connections filled with tap water and plugged loosely with a wad of cotton. RELATION OF BURSARIA TO FOOD 24 When, however, a direct current is used of such strength that the organisms can be made to go to one side or the other by reversal of the current the effect becomes more or less apparent. Feeding can not be prolonged to twenty minutes with a strong direct current, for the organisms are easily injured. To obviate this, time of feeding was limited to five minutes. The animals were made to swim from one side to the other by frequent rever- sals of the current. In Experiment X, they were stimulated by frequent reversal of the direct current during the first minute of feeding and then left to feed four minutes more without stim- ulation. In Experiment XI they were stimulated in the same way during the whole period of feeding. Time of feeding, five minutes (tables 11 and 12). TABLE 11 Experiment X STIMULATED BY DIRECT CURRENT, 1 MINUTE DISH NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | TOTAL a j i | I wile ’ - it “eee PES 2s. - . 3113) 5/11) 310 8 0} 1 611 5/8 4.3. 7)11)8 3 111 | } | ConTROL: NO CURRENT eer So iS< 255 5 7| 6| 4) 6| 3| 810! 7| 7] 6 8 5) O| 4| 3; 3] 9] 4110: 8 118 TABLE 12 Experiment XI STIMULATED BY DIRECT CURRENT, 5 MINUTES DISH NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL grains An 386 oe Coe 1) 3) 4 6 2) 310) 6) 510) 5) 0} O| 2 3} 8 7 1) 3/11 90 ContTROL: No CURRENT Snook) en 5} 5/11) O} 2/18) 3) 111) 011/13)/10 0 5| 7| 0) 9) 5) 6 122 As the results show, feeding was not discontinued under these conditions of strong stimulation by the current, though the or- 28 + Ikea ROUND ganisms show a somewhat smaller total number of grains eaten than in the controls, in the same length of time. The difference is, however, too small to have a clear significance. The strength of current may be increased but usually feeding can never be totally inhibited unless the organisms are injured or killed immediately after the yolk has been added. The preceding experiments show clearly the relation which exists within certain limits, between the feeding reaction of this organism and a simultaneous reaction to certain other types of stimulation. During stimulation with HCl and NaOH, and espe- cially with high temperatures and the electric current, the notable fact is that the reaction to food is strongly persistent under wide ranges of intensity of a second applied stimulus; this is true to such an extent that under some conditions the feeding continues up to the point where the intensity is so high that the stimulus is destructive to the organism. These facts must not be thought to be of general application, for evidently mechanical stimula- tion is quite effective in changing the reaction to food. What the behavior will be under two simultaneous stimuli obviously depends upon the nature of those stimuli. It should be distinctly noted that in all the foregoing experi- ments the chemical as well as physical nature of the food sub- stance has been kept constant while the organism in its particu- lar physiological state has been acted upon by certain external agents; these being of a sufficient variety to indicate clearly what role these different types of factors play in the relation of this animal to food, and to serve as a guide to further inquiry. We now have to see what changes are produced in the feeding reaction by modifying that factor which in the foregoing experi- ments has been kept constant,’ namely, the food. In the fol- lowing series of experiments all the other conditions will be kept constant, or at least arranged in such a way that they may be ° An exception to this might be taken in the experiments with HCl and NaOH for it is a question whether or not these affect the chemical character of the yolk sufficiently under the conditions of these experiments to modify the number of grains eaten. The yolk was not treated previous to the feeding; thus the time was so short and the dilutions so high that any change must have been very slight. RELATION OF BURSARIA TO FOOD 29 properly controlled and accounted for. We shall attempt to determine what the relation of Bursaria is, to specific physical and chemical properties of the food itself. First it will be deter- mined how the external part of the reaction is modified, that is, what is the behavior of the cell in so far as this has to do with the selection of food. SELECTION OF FOOD AND THE FACTORS CONCERNED The object of the experiments described in the present section is to answer the question: Can Bursaria discriminate quantita- tive or qualitative differences between the yolk grains? When fresh hard boiled yolk grains, prepared as described on page 8, are treated with different kinds of water-soluble dyes, the amount of dye which is adsorbed by a grain of yolk varies with the kind of dye used. At first a considerable number of different dyes in aqueous solution were tested in a compara- tively rough way; first, for the relative amount of each dye which would be taken up by the grains of yolk; second, for the rate at which the dyes were adsorbed and the ease with which they could be washed out (reversibility of the adsorption); and third, for the relative toxicity of aqueous solutions of these dyes to the organisms. Among the dyes so tested were fuchsin, lyons blue, methylin blue, eosin, cyanin, gentian violet, saffranin, janus green, congo red, and an aqueous solution of hematoxylin. The results of the following experiments on food selection, in so far as they are related to the dye, depend upon the three factors named: (1) The amount of dye adsorbed (2) The rate of the reversible adsorption reaction, and (3) The relative tox- icity to Bursaria, of the dye in.aqueous solution. It was quickly found that certain dyes were better suited than others, for the particular end in view. Aqueous solutions of saffranin and janus green were found best to fulfil the necessary conditions. Both show a reversible adsorption with yolk, while the velocity of the reversible adsorption is sufficiently low to prevent a too rapid washing out of the stain. By this means one is able to control the amount of adsorbed dye much more 30 Be ge LUND easily than if it could be washed out quickly, and one is also able to control the concentration gradient between pure water and the dye adsorbed by the yolk grain. The toxicity of the different dyes varies greatly, and it was found that saffranin and janus green were best from this point of view also, since both of these are very toxic to Bursaria in higher concentrations but only slightly so in lower concentrations. 1. Experiments with stained and unstained yolk a. Saffranin. Hxperiment XII (a). Object, to test (a) whether or not Bursaria will eat yolk grains which have adsorbed an appreciable amount of the soluble toxic substance saffranin and (b) whether or not the amount of yolk eaten depends, in this experiment, upon the amount of saffranin adsorbed. Equal volumes of a strong suspension of fresh yolk were placed in each of seven stender dishes of 8 cc. capacity. A bright rose-colored solution of saffranin was made up with tap water. To the dishes designated A, B, C, D, E and F was added 5, 4, 3, 2, 1, and 0.5 ce. re- spectively, of this solution, and mixed thoroughly. The seventh dish without stain, was kept as a control. The suspensions were left to settle five minutes, then decanted and 5 ec. tap water added to all the dishes; this was repeated three times. The organisms used were starved twenty-four hours and were in excellent condition. The time of feed- ing was fifteen minutes (table 13) TABLE 13 ‘ iy iments XIT (Oy | SAFFRANIN DISH | pect NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL | || | | 4 | ti | | | | grains | | Halve | | AGE, aa 5 0.00.00000000100110000001 0000010 5 Bee iccts 4 90004803 100001301000 403 40 100 110 34 Cx om 3 30.60.0208 3112.84 249 100 231/109) a) 94 011 61 1D seer 2 2} 0) 0) 3| 3] 1) 0 2 4| 2 1) 21 2 1] 3 3} 2 0 6 5| 3] 0, 4 01 Oo] 0 3) 4 56 JR 1 4/3 1) 27) 6 2 1) 7) 9 0) 3) 7 2) O12) 6 4 0 4) 4 5) 4 1) eS OO 71) 113 pat oh 1 4) 6/11) 6 1) 2) 4! 0} 3] 0} 3| 9| 3) 4] 3) Gl 5| 4! 5 9] 9] 6] 5| 3] 1) 31 8] 2) 71 6| «188 Control SiMe lal Wal ee | | Gree 0 0 5 6 6 7 4 9121 4 9] 9| 1/14) 4| 2! 7| 7 9 Glt2| 1] otaj ol ai ai si 4 gl 7| 188 When stronger solutions of saffranin than that in A were used, no grains were eaten. All the animals at the-end of the experi- ment were normal and-had not been injured. The yolk of dish RELATION OF BURSARIA TO FOOD ou A was now left to soak in its water for fifteen minutes longer, this water was then drawn off and the volk again washed twice with water. Thirty individuals from the same material as used above were now put into the dish. At the end of fifteen minutes the following was the count: Experiment XII (b) ees 4.4.3 9 3, 7, 3, 0, 4, 5, 3, 6, 0, 0, 4, 1, 4,570, 0, 2, 3, 1,3, 1. Total 88 grs- This indicates that we may obtain the same result whether we proceed with a strongly stained yolk and test successively after each washing, or, as in the former experiment, by staining dif- ferent portions of yolk to different degrees to begin with. This was actually done in other experiments not given, and results exactly similar to those in Experiment XII (a) were obtained. To show that even a considerably stronger medium does not injure the animals seriously, a yolk suspension stained with saffranin more strongly than that used in A of Experiment XII (a), was made by leaving the yolk several hours in a very strong solution of the stain. This was washed out several times and then thirty individuals from the same culture material used in the former experiments were put into it and left for fifteen min- utes. They were then picked out and washed once in tap water, and then transferred to an unstained yolk suspension for fifteen minutes. The count gave the following (table 14): TABLE 14 Experiment XIII 5 MIXTURE NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL koe rs Freel joel lost ee (petioles (PST ale c Stained........-..-.; 0) 0} 0} 0} 0! 0} 0} 0} O} O} O} O} 0} 0} O} O| O} 0 OF 0} OO} 2} 0. 0 010 0' 0/0 2 Unstamed..........} 0 2] 0) 3} 0: 0) 0] 0} 0} 1) 0} 3} 1) 0} Oj 2) 1) 3) 2 2) 3} 2) 8) 0} OO} 0 0; 2, 0 30 This shows that they were not injured sufficiently in even this strongly stained suspension totally to prevent them from eating unstained yolk immediately afterward. The cause of most of the 0’s in the count is that, as the toxicity of the solution in- creases, the organisms have a tendency to close up the oral apparatus and do not open it again sufficiently, within the next twenty minutes or so, to be able to take in the yolk grains. Of 32 Ei 3. LUND course if yolk which has been very strongly stained and not washed out sufficiently, is fed, then the concentration of the - medium rises so quickly that they are greatly injured or killed. Now against the conclusions which will be drawn from the re- sults of Experiment XII (a) and (b), as it stands, may still be urged the objection that the reason that so few or no grains are eaten, is because of what one might call a general injury or stimu- lation of the cell by the saffranin which is rapidly being liberated into the water, and that it may not have anything to do with a specific reaction to the chemical character of the food particle as such, that is, to anything like a “‘sense of taste.’’ That this objection does not apply to conditions like those in the above experiment (XII, a and b) where the amount of stain adsorbed even in dishes A and B is very little compared to that in Experi- ment XIII, may be shown by taking the solution of dish A, Experiment XII (a), and placing unstained yolk and Bursariae in it. The result of such an experiment is that the organisms fill up with fresh yolk, showing that the medium in weaker con- centrations does not affect the eating process to any appreciable extent. Further proof of this wil! be given in the experiments immediately following, and also in experiments to be given later. Experiment XIV. To test whether or not Bursaria can select and eat non-toxic yolk grains from among toxic ones, when the two are mixed. Two suspensions were made, one of yolk stained in saffranin twenty-four hours, then washed out repeatedly, the other containing the same kind of yolk washed in the same way but not stained. The two yolk suspensions were mixed immediately before the animals were placed in the mixture. Twenty individuals were used. The time of feeding was fifteen minutes; a control of washed unstained yolk alone, was kept at the same time (table 15). TABLE 15 Experiment XIV MIXTURE NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | TOTAL a: ies | 3 Stained... 6 ememe 000000000 00,0 0 Oxix'x x/x x 0 Unstained........... O 5 2 3 21) 2,429.01) 2.3 ax\x|xx/x/x| ag Control: 1 ail Ly A | | | | | 216 715 4 § 915.8 9 916 325 71512 9) 9 58 8), 218 Unstained yolk..... 1D RELATION OF BURSARIA TO FOOD 33 In the mixture the concentration of the saffranin rose so rapidly that some of the individuals were killed.'!° This is indicated above by X. Yet even in such a strong solution, selection took place, though the number of grains eaten was small compared to the number in the control. Experiment XV. Another sample of yolk less deeply stained than that in Experiment XIV, was washed out many times and mixed with an equal quantity of unstained yolk from the same sample. Thirty individuals of the same material as used in Experiment XIV, were fed for five minutes, instead of fifteen minutes as before (table 16). TABLE 16 Experiment XV MIXTURE | NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL Stained............. | al a] 1 a] of a] 3] of 3} 3 af al of al 0) 3 1] | al al ala} a) al aloldi ala) 43 Unstained.......... 2)6 7 21614 4 11111199 713,906 512/957 210 7) 5/7138 6 6 A repetition of the above experiment with yolk stained a little more deeply gave the following result; twenty individuals used; fed five minutes (table 17). TABLE 17 MIXTURE NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL Sipnines ey a ae 0 0 3 1 7 | 1 4 Unstained........... 400 1 45) 5 * } | | ae i aes The control of Experiment XIV will likewise serve for Experi- ment XV. The results of this experiment are to be explained by the fact that the concentration gradient of the adsorbed toxic saffranin is relatively low with respect to the gradient of the water-soluble yolk substance to which Bursaria reacts in a strongly positive manner. Of course one is not to suppose that it is the relative molecular concentration gradient alone that determines the re- 10 The cytolytic action of saffranin is in some ways more marked than that of janus green. The character of its reversible adsorption reaction also makes it less suited for use in experiments of this kind than janus green, as will appear from results with the latter. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 34 E. J. LUND sult. What is meant in this case by concentration gradient, is the molecular concentration plus the specificity of the substance, that is, in anthropomorphic terms we should say ‘‘the kind of taste’ which the substance has. That the specific nature of the substance is one factor in determining the result, is shown by a comparison of the results of numerous experiments with saffranin, janus green, hematoxylin, and especially other less toxic stains, like congo red (cf. what follows). b. Janus green. A considerable number of experiments have been carried out using this substance, with the same general results as those obtained with saffranin. It is better adapted to bring out the phenomenon of selection than saffranin, causing a sharp discrimination by Bursaria; small quantities adsorbed by the grains are sufficient to bring about rejection. The fol- lowing experiments show some of the relations. Experiment XVI (a) and (6). Yolk was stained in janus green twenty- four hours then soaked in tap water and washed repeatedly. A portion of the same kind of yolk soaked and washed in the same way but not stained, was used as a control and for mixing with the stained yolk. A few minutes before the experiment equal quantities of the stained and unstained yolk suspension were mixed in dish A. A second quan- tity of the unstained yolk suspension of a concentration equal to the sum of those in dish A was placed in dish B. Twenty individuals were placed in each dish and left to feed twenty minutes (table 18 a). TABLE 18 (a) Experiment XVI (a) MIXTURE, DISH A NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL | | | | | Stained eee; 0000000000000 O0 0x xxix x: x eae Unstained...........| 5 3 1] 6 2) 4) 1| 4! 5| 1) 5) 1] 4! 4 1x!)x/x x\x 47 Control, Ve alae lla [eae Mk: | | Dish B: | | | Baba Sa Unstained...........| 91221 son1245to SAT AS51210 114 7| 8|-3) 268 The yolk in both dishes was now washed twice and the experi- ment repeated with control. Time of feeding fifteen minutes. The count is shown in table 18 b. In (a) the solution had become sufficiently strong to affect five of the animals (X), so that they could not be recovered for RELATION OF BURSARIA TO FOOD Bip TABLE 18 (b) Experiment x VI (6) oi MIXTURE, DISH A NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL ed. e's sobs 00000000000000000000 0 Unstained.........../ 0) 0 0 1 0011201224001013 19 Control, | | | ' Dish B: | | Unstained.. 16 817 713101112 512 4 512 151516 71013 8 216 the count. The smaller number of grains eaten in (b) by those in dish A is due in part to the shorter time of feeding but more to the fact that the unstained yolk grains had by this time ad- sorbed some of the liberated janus green from the stained yolk grains (see Experiment XVIII, p. 36). Experiment XVII (a) and (6). The results of this experiment are given to show that Bursariae from two different cultures may show different reactions in selection experiments. In part (a) material was used from one wild culture, while in part (b) material from a differ- ent one was used. Both were starved twenty-four hours before using them. All other conditions were alike. Time of feeding fifteen minutes (table 19). TABLE 19 Experiment XV iT MIXTURE, DISH A NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL Sl ales IO a les | erste sg. 00000000000000000000 0 Unstained...........)094000000001111100000 9 Control: dish B, | | | (a) Unstained ........| 5 4) 6| 2) 4| 7, 6 3 9) 3 8 3| 3| 6 4! 3) 2 7| 6 8 99 Biained..... 02. ..0:. 00000000000000000 0 0 Unstained...........| 3) 4| 0| 210 332 7161)660821)05 1) 70 Control: Aaah B. (b) 7| 148 (Wostaimed.....:... 2 413 9 5 7 9) 2) 410 615) 9113) 611) 9) 1) 6 In such experiments as these it was found that occasionally an individual had eaten a stained grain along with the unstained ones, but this happened very seldom in any of the experiments with janus green. 36 Eee uuND. Experiment XVIII. The two dishes A and B in Experiment XVII (b) were left standing for two hours; then they were washed once and tested with the same material used in Experiment XVII (b), in order to show the effect of the adsorption of the liberated stain by the un- stained grains mixed with the stained ones. Time of feeding fifteen minutes (table 20). TABLE 20 Experiment xX VILL, MIXTURE, DISH A NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL Staumed vee Cae, hws JODHOD00000000000000 0 Unstained........... 0 2 072014 : 1000202000 ql 21 Control: dish B, ha a | iain Cnstamed.....<.\.: 125 81513 | 0 e125 4171612151511 0 $10 228 This shows that when the janus green yolk is left with the un- stained yolk for some time, the liberated stain from the janus green yolk is adsorbed by the unstained yolk grains, and as a result the latter are not eaten so readily. If the mixture is left standing too long and then rinsed in tap water, then no grains are eaten. Bursaria can react to such small quantities of ad- sorbed janus green that the amount adsorbed cannot be dis- tinguished by the eye, when the unstained yolk grains mixed with the stained ones are examined. Experiment XIX. To prove that the solution of the janus green which is produced by the liberation of the stain from the stained yolk, does not even in quite strong concentrations prevent the eating of fresh yolk placed in it, the result of one test out of a considerable num- ber made at different times, is given. A solution was drawn off from janus green stained yolk used in an experiment in which no grains af the mixture had been eaten, after standing for some time. To this solution was added unstained yolk. Ten individuals were tested (table 21). TABLE 21 Experiment XIX NUMBER OP GRAINS EATEN BY EACH _USIBUN ADE Fresh yolk in janus green solu- tion, dish Agari ne e. 28 | Se) 6) Sho Oa O comme 27 Control, fresh Tolle in ai Ww Bien. dish) B;... ..., cee eer ee ace 5 | 15] 4 | 5 | OG)! 5: | Sule 44 = RELATION OF BURSARIA TO FOOD 37 It is evident that the solution of janus green drawn from the mixed suspension, and produced by the liberation of the stain from the stained yolk grains of the mixture which was not eaten, did not now prevent the animals from eating unstained yolk grains which were placed in it; hence it was not the stain in solution which prevented the eating of the stained grains of the mixture from which the solution was drawn; but the stain which was adsorbed by the yolk grains of the mixture and diffused from them. Many such similar tests were carried out giving the same result. This does not mean that the solution apart from the yolk grain with its adsorbed dye, may not affect the result of the feeding, for in higher concentrations the solution apart, from the stain upon the grain does affect the feeding proc- ess. In solutions of lower concentrations of the appropriate dye the chemical nature of the grain along with the amount of dye adsorbed, are the essential factors determining the number of grains which will be eaten. c. Hematoxylin. To show further that the specificity of the toxic agent plays a large part in determining whether or not yolk will be eaten, the following experiments are given. It will be noted that in this case we have a substance which has a very different effect upon the cell and its relation to food, from that produced by the substances thus far dealt with. The solution in this case may be made very deep brown while the grains are also stained deeply, and yet the yolk grains are eaten even in solutions which kill the animals if they remain in it more than three or four minutes. Experiment XX: Table 22 (a) and (6). The same quantity of yolk was added to each of nine dishes of 8 cc. capacity, each containing equal amounts of tap water. The dishes were numbered 1, 2, 3, and so forth. To these were added diverse quantities of the 0.5 per cent aqueous solution of hematoxylin by drops, as given in the tables; time of feeding ten minutes. This experiment was repeated with the same suspensions at the end of one hour; time of feeding fifteen minutes (table 22 5). The individuals which died before the count was made are de- noted by X. The tables show that although the solutions, espe- 38 BE. Js LUND cially in higher concentration, are very injurious, the organisms, nevertheless, eat the grains of yolk. After one or more hours the grains become. stained deeply. This was the case in table 22 (b). The increase in the length of time of feeding (i.e., the time the animals were left in the solution) is the cause of the high mortality in table 22 (b). TABLE 22 (a) Experiment XX NUMBER OF DISH | 1 | 2 | 3 | 4 | 5 | 6 7 8 9 Number of drops of | | per cent aq. hema-| | | toxylin...:..:......| 4 | 8 | 42 | 16 | 20 | 24 >| 98 ue (il One Naat) Dig 2 1 4 fh 7 | esl Rh a) ee aa 4 6 0 | 6 | Cage Othe] 7 3 | ae : pete lig cea 4 1 has Ch 11 ee SF a ele Pee ae at at a) testea|| a aaa : ee: oe IAT Ws! Tene he 0 1." ist 1) | Sees Ghyll oye Go x | 4 pele 9 | 9 4 sec ane ese peeceealy rs) yi ape os) PCR ee | <0 3 Oe sy | Pee PAS 6 ae Oral) eared ex Lane Simla < 8 Totaled. ..b al 28 | 28 Rae Beat | 40 | 32+) (7-+)} 89 TABLE 22 (b) NUMBER OF DISH | 1 2 | 3 4 5 oe a 8 | 9 Number of drops of | | | + per cent aq. hema- | | | | Foxy lines coe oi. [yt 8 | 12 | 16 | 20)). 245) 25)) ssn ae (]) 3.) 2) 0 de i.) GAO 14,°6;° 5) 2, 12,2; 45-5; a total of 74 grains. These tests show (a) that although a dye may be toxic to Bursaria, it may nevertheless not affect, to any great extent, the functioning of the feeding mechanism in thé taking in and swal- lowing of the food, though (b) with some dyes total rejection of the food may take place, when the concentration is so low that it has only a comparatively slight cytolytic effect. The former condition is shown to a less marked extent in the experi- ments with saffranin than in the experiments with hematoxylin; while the latter condition is illustrated by the results with janus green. This seems then also to strongly suggest that different substances may affect different parts of the cell differently. Corroborative evidence upon this point, which it would be out of place to consider here, has been obtained from observations showing that the localization of the beginnings of cytolysis of the cell body of Bursaria may differ with the particular nature of the toxic agent employed. — d. Congored. Another stain which is adsorbed readily is congo red. This, however, unlike hematoxylin, can only be washed out in part, that is, its adsorption reaction is not completely reversible. Also since this dye is not as toxic as saffranin or janus green, a large quantity of the stain may be adsorbed and yet not appreciably affect the number of grains eaten, as is shown by the following experiment. Experiment XXI. The yolk was stained twenty minutes in a strong aqueous solution of the dye. Time of feeding twenty minutes. Thirty were used (table 23). 40 Hoo. LUND TABLE 23 LITA oS NUMBER OF GRAINS EATEN BY BACH INDIVIDUAL TOTAL Pan ey Be pe GIRL Ph Congo red, dish A | | | | | | “ik | | isk: Stained): 20.22... 7 36117 6 2 SESS EEE 115 Control dish B "| 3 Weil ] | | | | | | Unstained........ Ws T sags oess7 a4 alae oleae 179 1 } In this case we have a comparatively low concentration gradient of the dye, together with a low toxicity and hence the compara- tively small difference in the readiness with which Bursaria eats the stained and unstained yolk. e. Sudan III. To show that an adsorbed substance which is insoluble in the medium has no determining effect upon the feeding and food selection, Sudan III was employed. This sub- stance is insoluble in water but soluble in ethyl alcohol and fats. Experiment XXII. Fresh yolk was stained in an 80 per cent alco- holic solution of Sudan III for a short time. It was then dried in an oven at 27°C. for twenty-four hours. A control of fresh yolk was also kept. The organisms were fed twenty minutes. The yolk takes on a very deep color with this stain. TABLE 24 Experiment XXIT NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL | a (cer haee oe | Reh Ba Sudan III, stained | | | | NOL Ree roof 4 ,cosccks 131819 O15 3 5 7182515 5 7| 8 9171812 3,26 243 Control, unstained. . 18 3 414 7151 7)22 og 6 9 | eae Ta 274 It is evident that the insoluble Sudan III had no appreciable effect upon the food reaction. Mixtures of these showed no difference in the amounts of the two kinds of yolk eaten. f. Stale yolk. Experiment XXIII. Mixtures of fresh and stale yolk could not be used since the grains of the two kinds of-yolk were visibly indistinguish- able. One experiment is given. The stale yolk was four weeks old while the control was freshly prepared; both were of the same con- centration (table 25). RELATION OF BURSARIA TO FOOD 41 TABLE 25 Experiment XXIII NUMBER OF GRAINS EATEN BY EACH INDIVIDUAL TOTAL FSS Rde TPS AP EM Gaal WRT OD Stale yolk....:......| 0 25, 1] 0, 3, 8 2 1/1 4 0 6 1) 5} 1/3 a) 13 48 Control, fresh yolk. 8 4) 7 aut 4 4 9 31 0 7/9) 5| 2] 2 6 1 01 96 lee | hve eee at eda | | A difference is plainly evident. Other experiments show greater or less difference, depending upon the conditions. 2. The basis for and nature of food selection in Bursaria, as shown by the foregoing and other experiments It must be remembered that in any such experiments as the foregoing the relation to food is in some ways an:entirely new one to the organism. Yet it must be insisted upon that the yolk used in these experiments is assimilable by the organisms (a fact which will be considered at length in a later paper) and especially that whatever the mechanism of feeding and selection in nature is, it must be the same one which is brought into action in these experiments. Hence the criticism imagined above would appear to have no importance for the question under consider- ation here. In fact, it is to be believed that so far as these experiments are concerned, they are only a more strongly empha- sized condition of what we find in nature and that they picture to us, so far as they go, the actual condition of the food relation of Bursaria in its native culture. We may state the results briefly in the following way: First: ‘Yolk grains are rejected if the soluble adsorbed toxic substance makes with the medium a sufficiently steep concentration gra- dient. If this gradient is low relative to that of the yolk-soluble substance, to which Bursaria reacts positively, then the organism may eat the stained yolk, other conditions being equal. Second: (a) Whether Bursaria will eat stained yolk grains or reject them depends also, along with the steepness of the concentration gra- dient, upon the specific chemical properties of the adsorbed sub- 42 Ho) LUND stance in question, and furthermore (b) the substance by virtue of its chemical properties has, at least in some cases, a specific action upon the mechanism of feeding and selection, as is shown by a comparison of the results of the experiments with hema- toxylin, saffranin and janus green. Additional evidence obtained from observations upon the phenomena of cytolysis in Bursaria also points to the correctness of this conclusion. A familar in- stance of a similar nature is the casting off of the peristome by Stentor when stimulated or injured by chemicals. Another in- stance is the fact found by Jennings that the anterior end in Paramecium is more sensitive to mechanical stimulation than are other parts of the body. That in feeding experiments with the Protozoa it is difficult to discriminate closely between the effects of the medium and those of the food substance itself is obvious, since (a) the amount eaten depends upon so many factors other than the nature of food and (b), since the organism selects on a chemical basis, which involves a soluble substance or substances diffusing into the medium from the food particle, hence necessarily involving the external medium to a greater or less extent. It is of course clear that differences in certain physical characters of food may likewise determine whether or not it will be eaten. This is shown most simply by objects which are too large, such as large yolk grains and large individuals of Stentor, which cannot be swal- lowed. From all the facts found from experiments upon food selection by Bursaria, there is no evidence that active selection is based upon either “‘size, weight, form or surface texture’ or any com- bination of these, except in so far as simple mechanical condi- tions would make them effective. All the facts show clearly that the chemical nature of the food is the property upon which the power of discrimination by Bursaria depends. Hence I find no evidence from Bursaria to support Schaeffer’s contention that ‘Stentor selects its food upon a tactual basis and apparently not upon a chemical one” and that ‘‘Stentor reacts in selecting food, to physical properties only or chiefly, and not to chemical properties” (Schaeffer 710, page 131). On the other hand, the RELATION OF BURSARIA TO FOOD 43 facts which have been found in this connection are in agreement with the results and conclusions so far as they have been worked out by Metalnikow (712) for Paramecium. THE RELATION OF BURSARIA TO DIGESTIBLE AND NON-DIGESTIBLE SUBSTANCES 1. The external relations Many substances which are in the ordinary sense chemically indifferent to the organism are likewise eaten, though generally in small quantities. Among these are cinnabar, carbon black, chinese ink, powdered aluminium and the like. The relation of Bursaria to this class of substances is however strikingly differ- ent inside of the cell and to a large extent outside, when com- pared to that relation in the case of digestible and assimilable ones. The fact that some comparatively indifferent substances like the above, are eaten does not affect our conclusion drawn above, as to the paramount importance of the chemical prop- erties of the food in food selection. Chinese ink contains some - mucilaginous matter which as my own observations have shown me, is reacted to positively by Bursaria and hence the ink is quite readily eaten. Carmine isa similar substance which though generally taken to be insoluble in water, is in fact sufficiently soluble clearly to affect the feeding reactions of Bursaria. Fur- thermore, the fact that a substance may be insoluble does not, of course, prove that the stimulus from it is not a chemical one, for it is probable, that with such substances as aluminium, cata- lytic or other specific chemical, or even physical reactions depend- ent upon the chemical properties of the substance, are produced by contact with the plasma membranes. The possible variety of interactions of the cell with different kinds of substances when considered in this order of magnitude may of course be very large. As regards the eating of non-digestible substances, powdered aluminium may serve, in one way, to illustrate the external rela- tions. If a large number of individuals are put into a suspension of aluminium, often few if any will eat any of the particles of 44 HAT UND aluminium and those that do eat it generally take in only a small quantity. This is also true of Sudan III and of carbon black. The quantity eaten varies with the conditions in a similar way, as previously set forth for yolk. Now if fresh yolk grains are added to the suspension of aluminium the animals will often quickly fill up with yolk, but in this case flakes of aluminium become attached to the yolk particles and hence often considerable quantities of the metallic aluminium are passed into the body along with the yolk. Sometimes the quantity of yolk eaten in such a mixed suspension is less than that in the control. This serves to illustrate the sort of equilibrium which exists between the organism and the kinds of substances in suspension, partly determining the amount of food and other substances eaten. 2. The internal relations It was interesting to find that Bursaria possesses what I shall call an internal compensating reaction to those substances which are eaten to some extent, but are not digestible, such as Sudan III, chinese ink, powdered aluminium, and so forth. This com- pensating reaction makes up to some extent, in the “economy of the organism”’ for the lack of a perfect discrimination between indigestible (‘tasteless’) substances and those which can serve as food. It is shown by the fact that indigestible substances are eliminated from the cell usually a long time before the digestion of a similar quantity of food is completed. This may be shown with Sudan III. The results of the experiments are, for the sake of brevity, given by curves. Experiment XXIV: Figure 4. Three sets of twenty-four individuals each were fed Sudan III, cold-ether-extracted yolk, and fresh yolk respectively. They were placed two in each watchglass containing tap water in moist chambers, and examination in this case was made at the end of three, seven, and twenty-two hours. Points on the abscissae - indicate the length of time in hours after feeding, while points on the ordinates show the number of individuals which had extruded Sudan IfI (curve A) in the time intervals between the examinations, or in the case of extracted yolk (curve B) and fresh yolk (curve C) the nu- ber of individuals that had lost all traces of food. In this experiment the observations were not sufficiently frequent to bring out the actual RELATION OF BURSARIA TO FOOD 45 Individuals Hours 3 if 22 Fig. 4 Experiment XXIV. Curve A represents the course of total extrusion of Sudan III; curve B, that of complete digestion of cold-ether extracted yolk; curve C, course of complete disappearance of fresh fat-containing yolk. course of the extrusion of Sudan III or of the disappearance of the yolk from the cytoplasm, but it will be noted that the great difference appears in the observation at the end of seven hours. At the end of twenty-two hours A had no traces of Sudan III, B still had six and C had ten individuals with food. This early extrusion of indigestible substances is considered in more detail in connection with chinese ink in the following." Experiment XXV. Two sets of forty-eight individuals each were fed, one with cold-ether-extracted yolk, the other with chinese ink. Examination of the cell content was made at hour intervals as indicated by numbers on the basal abscissa. Ordinates indicate the number of individuals which had extruded all the ink content in the time indi- cated (curve A). In curve B the ordinates represent the number of individuals fed extracted yolk in which yolk had disappeared at the end of the time indicated by the abscissa. Chinese ink in suspension is much more readily eaten than carbon or aluminium and is therefore more convenient. This is to be explained by the fact that there are present mucilaginous soluble substances in the chinese ink which serve as agents inducing a more positive feeding reaction and are possibly of some slight food value. The ink was not found to be injurious to the animals. The greater part of the ink is thrown out quite early while slight traces may remain for some time longer. The time dur- ing which the ink was retained was taken to end when the last trace Jn order to obtain satisfactory results with such substances as Sudan III and aluminium in aqueous suspension the adsorbed gases should be driven off before feeding. 46 E. J. LUND Individuals Hane 34 5 6 7: 8 1 215 263 34 Fig. 5. Experiment XXV. Curve A represents the course of extrusion of chinese ink by forty-eight individuals; curve B that of complete digestion of a similar quantity of cold-ether extracted yolk by another set of forty-eight in- dividuals from the same culture. had been eliminated; curve A does not therefore represent the actual time at which the greater part of the ink was extruded but should have its maxima farther to the left than shown. This statement applies to all the extrusion curves which are given. In figure 5 curve B, is that of complete digestion; no extrusion of the extracted yolk took place in this experiment. After it had been thus shown that ink fed alone to one set of individuals was extruded long before digestion is completed of a similar amount of extracted yolk fed to another set, experi- ments were carried out to test what the reaction would be if both ink and extracted yolk were fed to the same individuals at the same time. The following two experiments are given to bring out the facts in a quantitative way. Experiment XX VI. Fifty-four individuals were first fed chinese ink and immediately afterwards fed with cold-ether-extracted yolk. Two individuals were placed in each watch-glass containing 5 cc. of tap RELATION OF BURSARIA TO FOOD 47 water and kept in moist chambers. Records were taken noting the presence or total absence of ink and presence or completion of digestion of the extracted yolk at one hour intervals beginning with three and one-half hours up to twelve hours after feeding; three more records were taken at twenty-four, thirty-three and forty-eight hours. The results are expressed in curves in figure 6. Curve A represents the extrusion of ink; curve B, that of complete digestion of yolk. It is seen from the relation of the curves that even in this ease the ink is extruded before digestion of the extracted yolk is complete, provided that a sufficient quantity of yolk has been eaten. It was noted that a short time after the ink had been eaten it became assembled into one or several rather definite lumps. This takes place before extrusion. Closer observation further revealed the fact that when ink particles came to be included in vacuoles containing yolk they were not extruded until the food of those vacuoles had been digested, while those which were not included by the yolk vacuoles were very soon extruded. This fact can readily be made out while one follows such experiments as Experiment X XVI above. Bursaria there- fore has a power of simultaneous selective extrusion of the con- tents of different vacuoles as well as a power of selection in the feeding process. This mechanism obviously compensates for the lack of a perfect discriminative and selective function of the oral apparatus. The results of an experiment (fig. 7) where these facts were taken into account for the purpose of expressing them in a graphic way in curves, is given in the following experiment. A control for comparison was also kept in this case (fig. 7, curve C). Experiment XXVII. Forty-eight individuals were used in each of both the experiment and control. The control (curve C, fig. 7) which was fed ink only, shows a sharp early maximum of extrusion from five and one-half to seven and one-half hours after feeding, with three or four individuals retaining traces of ink as long as ten and one-half to twelve hours. Curve A represents the extrusion of ink in the forty- eight individuals fed both ink and extracted yolk. It shows two max- ima exactly similar to those of curve A in Experiment X XVI. Curve B (fig. 7) represents the course of complete digestion of the yolk in the same individuals as those of curve A. There is only one maximum LUND J. E. “UNWIXBUT 9UO ATUO SBy JAIND ‘Y JAINO JO SB S]eNpIAIpur owes oy} AQ UOT gaSip oj01dulod yo asanod ‘g OAINO {[VUITXBUT OM} SMOYS YI ‘g7enpla -Ipul Inoj-AyJy 04 poj YUL osaurypo Jo WOIsNs}xo 940[du109 Jo os.inod oy4 syuasetdor y PAINT) “TAXX quouodxiy 9 “SIT 8r ee ve zi un or 40 £2 29%G

ey 73 PLATE 2 EXPLANATION OF FIGURES 13 to 18 a First spermatocyte metaphase spindles 13 Extra arm to left toward upper pole. 14 Small round separate chromosome toward lower pole. 15 and16 Round chromosome projecting from the mass toward the lower pole. 17aandb No possible Y-chromosome present. 18a No possible X-chromosome present. 18 b to 21 First spermatocyte anaphases 18b Left pole of spindle more clearly in focus than right; at least one of three projecting arms at the left has a corresponding arm at the right. 19 Two projecting arms at upper pole, and one at the lower. 20 No possible X-chromosome. 21. Projecting arm at right of both upper and lower poles. 22 and 23 Second spermatocytes 22a Equatorial plate, red to left. 22b Metaphase spindle, small round projection toward lower pole. 22. Equatorial plate, no possible X-chromosome. 22d Equatorial plate, separate round chromosome to left. 23 a Anaphase, no possible X-chromosome. 23 b Metaphase spindle, small round projection toward lower pole. 74 SPERMATOGENESIS OF THE CHICKEN PLATE 2 ALICE M. BORING AND RAYMOND PEARL 75 PLATE 3 EXPLANATION OF FIGURES 24 to 34 First spermatocyte divisions 24a Metaphase spindle, projecting arm toward left pole. 24b Equatorial plate, projecting arms. 25a Equatorial plate, no possible Y-chromosome. 25b ande Equatorial plates, mass of chromosomes in two pieces. 26a Equatorial plate, round chromosome to right. 26 b Metaphase spindle, three-parted chromosome toward upper pole. 27 Equatorial plate, two separate round chromosomes. 28 a Metaphase spindle with round chromosome to the right, and an arm to the left. 28 b Metaphase spindle, one small round chromosome to left. 28 ¢ Equatorial plate, three separate round chromosomes. 29 Metaphase spindle, corresponding projecting arms to the right toward each pole. 30 and 31 Metaphase spindle, one round chromosome toward lower pole. 32 Metaphase spindle, projecting arm toward‘upper pole, and round chromo- some toward lower. ; 33 Metaphase spindle, V-shaped chromosome to right. 34 Anaphase, projecting rod-like arm only at lower pole. 36 to 37 Second spermatocyte divisions 35 a Metaphase spindle, no possible Y-chromosome. 35 b Anaphase, no possible X-chromosome. 36aandd Equatorial plates, out of focus. 36¢ ande Equatorial plates, no possible Y-chromosome. 36 b Anaphase, no possible Y-chromosome. 37 aandec Equatorial plates, out of focus. 37 b Metaphase spindle, no possible X-chromosome. 37d Equatorial plate, round chromosome toward upper side. SPERMATOGENESIS OF THE CHICKEN PLATE 3 ALICE M. BORING AND RAYMOND PEARL 46 47 48 49 50 PLATE 4 EXPLANATION OF FIGURES 38 to 53 First spermatocyte equatorial plates Chromosomes massed with projecting arms. Chromosomes loosely massed with space in center. Chromosomes scattered, but still all connected. Chromosomes massed, with a V-shaped body partially separated to the Chromosomes massed, with a three-parted body partially separated to the Three-parted separate body, one-third the size of the whole chromosome. Three-parted separate body, small in proportion to size of whole chromo- plate. Four-parted separate chromatin body to the right. V-shaped separate chromatin body to right. V-shaped separate chromatin body at lower side. Two separate chromosomes, one round and one V-shaped. Two separate chromosomes, one small round one and one rod-shaped. One round chromosome somewhat separated from the rest. The number of chromosomes here might be counted as nine. ol 52 53 One small round separate chromosome to the right. Two rod shaped chromosomes, one separate and one attached. Similar to 52, but rods shorter. ll i a ee, a al SPERMATOGENESIS OF THE CHICKEN PLATE 4 ALICE M. BORING AND RAYMOND PEARL 79 PLATE 5 EXPLANATION OF FIGURES 6410 62 First spermatocyte metaphase spindles 54 No possible X present. 55 and 56 ~Two projecting arms. 57 One projecting rod toward lower pole. 58 and 59 Projecting round chromosome toward lower pole. 60 V-shaped chromosome toward lower pole. 61 V-shaped chromosome toward lower pole and two projecting arms toward upper pole. 62 Four-parted body toward lower pole. 68 to 69 First spermatocyte anaphases 63 No possible X-chromosome present. 64 Projecting arm at upper pole. 65 Projecting rod-like arm at lower pole. 66 Separate four-parted body at upper pole. 67 to 69 Corresponding projecting arms at both poles. 80 SPERMATOGENESIS OF THE CHICKEN PLATE 5 ALICE M. BORING AND RAYMOND PEARL 81 PLATE 6 EXPLANATION OF FIGURES 70 to 84 Second spermatocyte divisions 70to72 Equatorial plates, no separated chromosome. Apparently more than one-half as many chromosomes as in I spermatocyte, although counting is impossible. 73 to 77 Equatorial plates, with separated chromosomes, varying in number and shape. 78 and 79 Metaphase spindles, no possible X-chromosome. 80 and 81 Metaphase spindles, with projecting arms toward lower pole. 82 Metaphase spindle, one round separate chromosome toward lower pole. 83 Anaphase, two separate chromosomes toward lower plate, one round and one three-parted. 84 Anaphase, no X-chromosome present. 85 First spermatocyte equatorial plate. One of the few cells in the smear preparations where it is possible to try to count chromosomes, the number might be nine. 86 to 91 Aceto-carmine preparations 86 First spermatocyte equatorial plate, where there are apparently 6 chro- mosomes. 87 First spermatocyte equatorial plate, where there are apparently 7 chro- mosomes. 88 First spermatocyte equatorial plate, where there are apparently 8 chro- mosomes. 89 First spermatocyte equatorial plate, where there are apparently 9 chro- mosomes. 90 Second spermatocyte equatorial plate, where there are apparently 9 chro- mosomes. 91 First spermatocyte anaphase, where there are apparently 9 chromosomes at each pole. 82 SPERMATOGENESIS OF THE CHICKEN ALICE M. BORING AND RAYMOND PEARL PLATE 6 83 THE EFFECT OF RADIUM RADIATIONS ON THE FERTILIZATION OF NEREIS CHARLES PACKARD THREE PLATES CONTENTS 277 LESS OS Es fe eee ree er ce rene ee: eT ee eg ee 85 meeestire OF the Fagin TAGIAGIONS:..... 2-226. 26oo.ce. 25 socks eee asses 86 Pee WATE OM PRUE. Gk seo) c Sole hte MEME be TI ah. Re Ieee EE 86 STs CUS up ee a RY RI eg ane a ee cet oe, See nae creo ae A eae 91 Mba normal: development, of Nereis «.<...<...-. Si ee AR, REN loerele nea Ry ge me ree es BO et eR eR 93 a. The fertilization of normal eggs by radiated spermatozoa.............. 93 b. The development of eggs radiated before fertilization.................. 98 c. The development of eggs radiated after fertilization................... 100 d. The development of eggs radiated before and after fertilization........ 106 e. The development of eggs normally fertilized, and exposed to radium.. 107 f. Subsequent development of the embryo...........................25-- 109 g. Development of eggs radiated before and after fertilization with ST GTZ IS 07) 1 ee eee an Le ee 109 J EEL ETE a ER Se to oh A it tee ae SU RA ee Me ea Jt CORR ea 112 ppeeaeREPUECETLG! (CURIEC LUSHOT fooye Ue. ais i CS DAML ated chee idee blah Go cies dG ela Sz 119 LLU Cec: (ar rr Ro eat Nee et Tae cod ee 5 fawn 121 sINTRODUCTION The radiations of radium have been used as a stimulus to proto- plasmic activities from the time they were first discovered. The early experiments were made with little knowledge of the nature of the stimulus and with none whatever as to its probable effect on living matter. It was found that the rays affected the tissues in very different ways, acting as a stimulus to growth in some tissues, and as aninhibitor in others. Thus, seedlings have, in some in- stances, been accelerated in growth, and the action of certain enzymes (pepsin, diastase, etc.) favored by exposure to the rays. On the other hand, an exposure to radium produces a marked retardation in the growth of certain tissues and a characteristic degeneration in some of the cell constituents. The latter effects 85 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 S6 CHARLES PACKARD are found chiefly in the rapidly growing structures, such as em- bryonic or regenerating tissue. The nature of the injury has been studied particularly in the former type, and the generaliza- tions made on the effect of radium radiations have been based chiefly on such investigations. A study of the cytological details of the changes thus brought about has been the point at issue in the most recent studies on this subject and is the object of the present investigation. THE NATURE OF THE RADIUM RADIATIONS A discussion of the nature of the radium radiations need not be entered into fully in this paper. It may be mentioned simply that the three types of rays apparently produce rather different effects. The alpha rays, which form the greatest part of the energy given off, are wholly shut out by the glass or mica screens which are usually arranged for the protection of the salt. As the apparatus used in the following experiments and in the in- vestigations of all those who used living tissues as an object for study effectually screened the alpha rays, they may be disre- garded in this discussion. The beta and gamma rays are able to penetrate thin glass, but their activity is greatly diminished in the transit. Their effect on protoplasmic activities is, in general, an injurious one, although Congdon (712) found some evidence of acceleration in the rate of development in the eggs of Drosophila due to the secondary gamma rays. In this case only the eggs which received a very weak stimulus were accelerated. The injurious effects found by practically all observers will be dis- cussed in the next section. HISTORICAL REVIEW It has been found that different tissues show a great diversity in the nature of their response to the radiations of radium. Thus, the spermatogonia and the spermatocytes in the testis of the rat may be killed or greatly injured while the surrounding Sertoli cells and the connective tissue show no effects whatever. The same is true for ova and the follicle cells. K6rnicke (’05) who exposed to radium the growing root tips of Lilium for periods of EFFECT OF RADIUM ON YITERTILIZATION 87 one hour to three days, found that the most noticeable effect was on the nucleus of the cells, the protoplasm apparently undergoing no change. The chromatin was clumped together, and in some eases formed a homogeneous ball. He also found in the tetrad cells a great number of small nuclei, and many extra-nuclear nucleol . Guilleminot (08) found similar changes. With intense ra- diation (200 mg. of the pure bromide) the nucleus of the pollen grain lost all power of division after the fertilization of the ovule. With weaker stimulus the resulting embryo was small and ab- normal in various respects. If the stimulus is applied during fertilization the injurious effect appears early in development, and the plant is unable to repair the injury. If the adult tissue of the plant is exposed, very little effect can be noted. ‘Experiments on the effect of radiations on the germ cells and on the fertilized egg have been carried on by numerous observers. The first of these was Bohn (’03) who used the frog and sea- urchin as the basis of investigations. In general he found that the radiations produce a retarding and injurious effect on the early development of both of these forms. On exposing sea- urchin sperm he found that they very soon lost their motility and failed to fertilize the eggs, a fact not borne out by the experiments of G. Hertwig in the case of Parechinus miliaris, or by my own on Arbacia punctulata. Perthes (’04) has described similar effects of radiation on the fertilized eggs of Ascaris megalocephala. Schaper (’04) and Levy (06) have described the abnormal larvae arising from the radiated, fertilized frog’s egg. The abnormalities consisted in the peculiar shape of the larvae, and the pathological condition of the nervous system. The vascular system also was much injured. Schaper believed that the abnormalities in the embryo were due in a large measure to the destructive action of radium radia- tions on the yolk, the lecithin being destroyed through the ioniz- ing effect of the radiations. This theory, first proposed by Schwarz (’03), will be discussed later. Bardeen (711) has found that X rays (which are similar to the gamma rays of radium) produce a marked effect on both eggs and sperm cells of frogs so that they give rise to abnormalities. The 88 CHARLES PACKARD radiation of the egg produces a greater injurious effect than that of the sperm. In early cleavage susceptibility of the cellsis increased, but in later stages he found, in common with many other observers, that the susceptibility is greatly diminished. A more critical examination of the effect on the embryo of radiating the germ cells has been made by O. and G. Hertwig who also used the frog as a basis of investigation. In four series of experiments they exposed first the fertilized eggs (A series); the sperm (B series); the unfertilized egg (C series); and both eggs and sperm before fertilization (D series). In general they found that the exposure of the egg or sperm alone produced smaller abnormalities than the exposure of the fertilized egg. In the A series and in the D series the development was affected in proportion to the length of exposure and the intensity of the radiation. The same is true, up to a certain point, for the B and C series. But if the sperm or the egg is exposed for a long time the resulting embryo is fairly normal. In such cases, according to O. and G. Hertwig, development is parthenogenetic, since the sperm nucleus néver becomes a part of the cleavage nucleus, but serves merely to initiate development. The same is true when the egg is radiated. The sperm nucleus and not the egg nucleus is active. The embryos developing after the various types of treatment show characteristic abnormalities. Gastrulation may be very abnormal and the older embryos: much bent, as described by Schaper (04). The internal structures also show great. changes. Most seriously affected is the nervous system in which the cells are abnormal and the nuclei broken into granules. The blood system also is injured. Physiological abnormalities are as marked as the morphological. Growth is greatly retarded, due to the slow rate of cell division. Though the muscles develop they are paralized for lack of an adequate nervous system. ‘The nuclear disturbances in the affected cells are similar to those found by earlier investigators, but the plasma of the cells appears to be unchanged. A further study of the exact nature of the injury to the nucleus has been made by P. Hertwig (11) and G. Hertwig (12). The EFFECT OF RADIUM ON FERTILIZATION 89 former found that the fertilized egg of Ascaris megalocephala, after exposure to radium, shows much the same type of abnor- mality that had been observed by others, namely, that the chro- matin alone is affected, being broken up into numerous irregu- lar granules. In division these granules are not exactly equally divided, but part go to one blastomere and part to the other. The astral system and the plasma are not affected at all. Miss Hert- wig concludes that the effect of the radium radiations is a direct one on the chromatin, and not indirect, as Schwartz and Schaper claimed. G. Hertwig (712) has studied the effect of fertilizing the normal sea urchin egg with radiated sperm. The sperm remains active after twelve hours of intense radiation. But such sperm, even though motile and able to penetrate the egg, is incapable, in many instances, of fusing with the egg nucleus. Inside the egg it remains as a compact mass near the latter which divides, being provided with centrosomes derived from the sperm. In some instances it may become involved as a foreign body in the spindle of the dividing egg nucleus, in which case it causes abnormalities in the distribution and form of the egg chromosomes. In divi- sion it usually goes to one blastomere, and in later divisions may operate to produce very abnormal mitoses resulting in an abnormal larva. In other instances it may fuse with the egg nucleus but even then it behaves abnormally and is eliminated during the subsequent divisions. The main point is that the embryos aris- ing from this treatment are parthenogenetic. Thus cytological study has confirmed his earlier hypothesis based on the experi- ments with the frog’s egg. Since there is no visible evidence that the protoplasm of the sperm or fertilized egg has been injured, the Hertwigs conclude that only the chromatin is affected. This conclusion is supported by the fact that the harmful effect of the radiations is much great- er in the fertilized egg, with twice as much chromatin, than in the normal egg fertilized by radiated sperm. Two hypotheses have been advanced to explain the phenomena that have been described. The first, proposed by Schwarz (’04) was based on the fact that egg yolk is decomposed by exposure to the 90 CHARLES PACKARD radium radiations. Although the matter was not chemically determined, it seemed probable that the lecithin was broken up into cholin and tri-methyl-amine and other end products of leci- thin decomposition. Lecithin has been found by many investi- gators in all cells, especially in egg yolk, spermatozoa, pollen cells, plant spores, growing buds, and in all rapidly growing tissue. If then, it is destroyed such cells must necessarily be unfavorably affected. Against this hypothesis can be raised a number of objections. In the spermatozo6n, there can be very little lecithin, and any amount that is destroyed would be so small as to be nearly negli- gible. Yet Hertwig has stated that exposure of frog sperm for one minute is sufficient to produce some abnormality in the em- bryo. It is hard to imagine that this is due to the amount of leci- thin decomposed in so brief a period. On the other hand, the egg contains a relatively enormous quantity, so that the effect should be correspondingly greater, yet such is not the case in the frog. The experiments of Bardeen show only a slightly greater injury when the egg is exposed. It is also found that the fertilized egg, when radiated, develops very abnormally even though the expo- sure is brief. These facts indicate that the lecithin hypothesis cannot explain all of the phenomena although the decomposition of the lecithin under radiation is an actual fact. The second hypothesis, advanced by O. Hertwig is called by its propounder a ‘Biological hypothesis’ although fundamentally chemical in its nature. It is assumed that the radium radiations affect only the chromatin and that all the abnormalities that result from radiations arise in consequence to the injury to that substance. The chromatin, under the influence of the radiations produces a ‘contagium vivo’ which acts like a living ferment in that it increases at each cell division. Thus an originally small amount generated in the sperm will make itself felt after many cleavages, while a large amount generated in the fertilized egg will produce abnormalities almost at once. In cases in which the spermatozoa have been radiated for a long time it is assumed that the ‘contagium vivo’ has increased to such an extent that it has killed itself and at the same time affected the dividing power of EFFECT OF RADIUM ON -FERTILIZATION 91 the chromatin. The destruction of the poison by prolonged ra- diation makes it possible for the egg to develop parthenogeneti- eally. When the fertilized egg is radiated enough poison is gen- erated to prevent cell division entirely. There is no doubt that some substance is stimulated to activity under the influence of radium radiations which acts like a poison in that it produces the apparent abnormalities, and that this sub- stance is intimately connected with the nucleus. But to call this substance a ‘contagium vivo’ and to endow it with the property of producing the observed results does not aid in the solution of the problem. The solution must be fundamentally chemical in nature, even though the precise reactions involved cannot now be discovered. The following investigation was made to find how the early development of the egg is affected by radium radiations when (1) the sperm is exposed; (2) when the egg is exposed; and (3) when the egg is exposed immediately after insemination. The problem was suggested by Prof. T. H. Morgan, to whom I take pleasure in expressing my thanks. METHODS Nereis limbata was used in these experiments. In its sexual phase it can be obtained in great numbers during the summer months at Woods Hole, Massachusetts. This material has a great advantage in that practically 100 per cent of the eggs segment and develop normally under appropriate conditions, showing great regularity in the rate of development. The sex- ually mature individuals swarm in great numbers at the surface of the water on dark evenings, and can be kept, with proper pre- cautions for many hours before they discharge their sexual prod- ucts. For each experiment the eggs of two females were used. It was found necessary to work with small quantities of eggs since the jelly which the eggs extrude at the time of insemination is very mucilaginous and tends to mat together, in which event the eggs fail to develop normally. To prevent this it is necessary to stir them frequently during the first fifteen minutes, at the end of 92 CHARLES PACKARD which time the jelly loses its sticky character. The eggs were killed in Meves’ modification of Flemming’s fluid. The sections were cut 5 uw in thickness and stained with iron hematoxylin. THE NORMAL DEVELOPMENT OF NEREIS The normal fertilization of Nereis has been described by Lil- lie (12) so that an extended account here is unnecessary. My observations are in every respect in accord with his. The follow- ing is a greatly abridged history of the fertilization phenomena in which only the more important events are mentioned. As soon as the successful spermatozo6n is implanted in the egg membrane it loses its motility and remains exterior to the egg for about fifty minutes. Very shortly after its implantation the cortical zone of colloidal material in the egg begins to be extruded, forming a thick jelly which, as it flows out, pushes away all the spermatozoa except the implanted one. The surface of the egg just beneath the perforatorium of the sperm now rises up in the form of a cone (fig. 1). The fertilization cone thus formed soon sinks, drawing with it the sperm which consequently comes to lie in a depression of the vitelline membrane. The cone, in the meantime has taken a deep stain and is a conspicuous object at the periphery of the egg. About fifty minutes after insemination the cone begins to move inwardly toward the center of the egg pulling with it the sperm head, which, as a consequence of being drawn through the membrane, assumes a band-like appearance. (fig. 2). The middle piece and the tail are left behind. The cone, followed by the sperm head, now moves deeper into the egg and revolves through 180° so that the sperm head isin front. The cone is left behind at this stage while the sperm nucleus moves forward. At this time a sperm aster develops (fig. 3). Later this structure divides unequally, the two asters thus arising form- ing the cleavage asters. As the sperm head advances the chrom- atin becomes vacuolated and later breaks up into the haploid number of karyomeres. These gradually fuse, the chromatin spinning out into a delicate spireme. At this time the sperm nucleus is ready to fuse with the egg nucleus (fig. 4). EFFECT OF RADIUM ON FERTILIZATION 93 In the meantime, the polar bodies have been extruded and the egg nucleus, at first very small after the running together of the vesicular chromosomes remaining after the second maturation division, now enlarges. The chromatin forms at first into kary- omeres, just as in the case of the sperm nucleus, and then spins out into a delicate spireme. The two germ nuclei now fuse. The spiremes thicken and break up into chromosomes, which, after the first cleavage, become vesicular and finally run together forming the typical resting nucleus of the blastomeres. The early cleavage needs no description, but a few typical stages that have been chosen arbitrarily for purposes of compari- son with the abnormal larvae will be mentioned. In ten hours (depending somewhat on the temperature) the embryo acquires the protrochal band of cilia and begins to swim about actively. Green pigment at the posterior end develops in about fifteen to eighteen hours. In thirty five hours reddish brown pigment develops just back of the protrochal ciliary ring. A little later the first setae appear. At sixty hours the setae are in two pairs and soon after become jointed. The third pair also develops at about this time. One hundred hours after fertilization the palps are formed, and the whole embryo is markedly segmented. EXPERIMENTAL a. The fertilization of normal eggs by radiated spermatozoa The spermatozoa of Nereis show remarkable vitality under very adverse conditions. If they are taken from a freshly caught, vigorous male and kept without admixture with sea water, they will remain alive for upwards of fifteen hours, and at the end of that time will fertilize fresh eggs in a normal manner, the embryos arising from such a union showing no marked abnormalities. They can live for a short time in practically fresh water, a fact that renders the sterilization of pipettes a troublesome process. If pure sperm be diluted with sea water the spermatozoa die within a few hours. The fresh spermatozoa, as free from sea water as possible, were exposed by putting a small drop of the fluid in a thin walled glass 94 CHARLES PACKARD vial of such a size that it exactly fitted into the depression in the capsule in which the radium salt was held. The radium in this experiment was equal to 4 mg. of the pure bromide. The drop of spermatozoa was so spread out on the bottom of the vial that it formed a thin layer, all parts of which must have received equal radiation. A control experiment was made by putting another small drop into a similar vial which was placed near the other, but screened from it by two sheets of lead. As a rule, the sper- matozoa were radiated for at least twelve hours, for it was found that shorter exposures produced very slight effects. After so long a treatment the motility of the spermatozoa seemed to be un- changed. Indeed they showed no signs whatever by which to distinguish them from the control or from perfectly fresh sper- matozoa. These results are in accord with the statements of G. Hertwig (12) in regard to the spermatozoa of sea urchins. The phenomena of fertilization that can be observed in the living egg are normal. Great numbers of spermatozoa collect at the periphery of the eggs and can be seen to be pushed away by the gradual outflow of jelly which forms as soon as the succes$ful spermatozo6n is implanted in the egg membrane. If india ink is added to the water it becomes evident that no jelly is extruded at the point of attachment, the spermatozo6n appearing to lie at the apex of a cone of ink which marks the region where the jelly is lacking. The time elapsing between the addition of the sperm and the outflow of jelly appears to be normal. Soon after the implantation of the spermatozo6n the fertilization cone develops, reaching up to the surface of the egg membrane just below the point of attachment. The further progress of the fertilization phenomena can be seen best in sections. Eggs killed thirty minutes after insemination show the sper- matozo6n, entirely normal in appearance, still external to the egg, with the perforatorium extending through the egg and into the substance of the fertilization cone. A little later the cone is retracted, pulling with it the sperm head and egg membrane so that the latter forms a small depression in which the former lies. Then the sperm head begins to be drawn out, penetrating further into the cone, and developing about its perforatorium EFFECT OF RADIUM ON FERTILIZATION 95 the usual attachment granules. The cone itself develops normally, and its behavior in drawing in the spermatozoo6n is normal. When the sperm head is in the process of being drawn in by the fertilization cone the first abnormality makes its appearance. The chromatin of the head, instead of forming an unbroken band, as in the normal process may now become broken up into irreg- ular masses (fig. 6). Occasionally, in the normal sperm head as it is drawn through the egg membrane, one finds some evidence of segmentation, but in no case is there so marked a breaking up of the chromatin as is found here. About 10 per cent of the eges show this abnormality. From this stage the subsequent behavior of the sperm heads may differ. The further entrance and development of the sperm nucleus may be normal, or the sperm may entirely fail to gain further entrance. In the first case, the sperm head, preceded by the fertilization cone, penetrates further into the egg, revolving meanwhile. After complete revolution the sperm aster develops. The division of the sperm centrosome was found in only a few instances, but was normal when found. The chromatin becomes vacuolated and later forms into the haploid number of karyomeres. At this stage it may normally fuse with the egg nucleus, or else undergo a curious development which results in its failure to fuse with the latter. When fusion occurs the cleavage nucleus divides normally with twenty-eight chromosomes present. There is, then, either a complete fusion with subsequent normal division, or else no fusion at all, and no division of either germ nucleus. In the latter instances the germ nuclei remain in the karyomere stage. The sperm asters which at first are present now gradually dwindle and finally disappear altogether. Figure 7 illustrates this point. The two asters are hardly as large as the smaller of the original sperm asters. At this time the egg aster has normally disappeared. This case and others to be described are not due to polyspermy, for such eggs are readily recognizable. The number of karyomeres present also shows that only one sperm has entered. Figure 8 shows an egg in which no trace of astral radiations can be found. There are many such eggs in the preparations. The action of the radiation on the sperm, then, has been to prevent 96 CHARLES PACKARD not only its own development, but that of the egg nucleus also. The number of vesicles and karyomeres differ considerably. Normally there should be at most four or five vesicles with twenty-eight karyomeres. Frequently the number of vesicles may be as high as twenty. The appearance of such eggs recalls the condition described by Lillie (’02) in the fertilized egg of Chaetopterus which had been treated with KCI solutions of vary- ing concentrations. He found that the nucleus broke up into many fragments which were distributed irregularly through the protoplasm, and were never gathered up into a single nucleus again. Each particle of chromatin was surrounded by a vacuole of liquified protoplasm. The appearance is similar in many respects to that found in Nereis, although the mode of formation of the vacoules is entirely different. Whether both of the germ nuclei are concerned in this process or only one, cannot be defi- nitely settled. The presence of no extra asters shows that if the sperm nucleus has divided the chromatin only has been con- cerned. In view of G. Hertwig’s experiments on the sea urchin, in which the egg nucleus divided without fusing with the sperm nucleus, it seems reasonable to suppose that in Nereis also the egg nucleus divides by itself. But here the division is very abnormal, being due to something evidently brought in by the sperm. The second class of abnormalities is caused by the failure of the sperm nucleus to enter the egg. There is always a fairly large number of abnormalities of this type in material treated in the way indicated, and each egg presents practically the same appear- ance. It should be mentioned at the outset that this condition is found, though rarely, in the controls. Although the spermatozoa that have been radiated are as active as those in the controls yet some of them fail to effect a complete entrance into the egg. The actual attachment of the spermato- zoon is undoubtedly due to its own activity, but its subsequent entrance into the protoplasm of the egg is due to the activity of the egg itself. It may be inferred, therefore, that the sperm does not, in these cases, call out the proper stimulus for the complete reaction. The egg is able to draw the sperm head in for a short distance only. The details of the entrance of such a sperm are EFFECT OF RADIUM ON FERTILIZATION 97 as follows: The moment the perforatorium becomes implanted in the egg membrane there occurs the usual outflow of jelly. Beneath the perforatorium there forms the usual darkly staining area of protoplasm. The fertilization cone rises to meet the point of attachment, then sinks, pulling the sperm head with it. But as the chromatin is pulled out it is seen to be broken up into irregular masses, a condition not seen in the controls. In some instances the band appears to be continuous, but arranged in a bead like fashion. Thus far, then, the effect of the spermatozoén on the egg has been merely to fail to elicit the full response by which it should be engulfed in the egg. During this period the egg gives off the two polar bodies in a normal manner. But when the egg nucleus begins its recon- struction, it fails to develop in the usual way. The chromosomes increase somewhat in size, and each becomes surrounded by a deeply staining, granular matrix. The chromatin appears nor- mal. The matrix stains about as deeply as the fertilization cone. Each chromosome, surrounded by the matrix, lies in a clear, non- staining vacuole. Occasionally they can be seen to be segmented, a condition not seen in the control. In such eggs there is a total absence of astral radiations. The egg aster disappears very early, and the sperm aster never develops because the sperm is still entirely exterior to the egg. The development of the egg nucleus in such a manner is not due to any poison injected into the egg by the radiated spermato- zoon, but rather to the failure of the sperm head to enter. That this is true may be inferred from the experiments on partial fer- tilization made by Lillie (12). These consisted of centrifuging the fertilized egg at such times that the attached spermatozoén, still external to the egg membrane, was removed wholly or in part. In the former case, when no chromatin was introduced, the eggs extruded the jelly and extruded both polar bodies. But the egg nucleus, instead of forming a typical vesicle with karyo- meres, failed entirely to develop, the chromosomes lying free in the protoplasm. Around each chromosome was a darkly stain- ing matrix, evidently corresponding to the chromosomal vesicle, through no vesicular wall was formed. This condition is an 98 CHARLES PACKARD exact picture of that found in eggs in which the radiated sperma- tozoon has failed to enter. It is evident, then, that the radiation of the spermatozoa is not responsible for the peculiar development of the egg nucleus. Nor is the failure of the sperm to set up an appropriate reaction in the egg which should ensure its entrance an effect of the radi- ations per se. For the same condition is occasionally found in the controls. Probably any treatment which changes the sper- matozoa would bring about the same result. To sum up the results of this experiment, it may be said that the abnormal spermatozoa can be divided into two classes, ac- cording to the reactions they initiate in the egg. Those of the first class are able to activate the egg normally and are conse- quently drawn into the egg protoplasm; those of the second class are able to produce this result only in part. The spermatozoa of the first class induce the normal extrusion of the jelly, and the formation of the fertilization cone, but they fail to develop the normal cleavage asters and fail to fuse with the egg nucleus. The second class of spermatozoa bring about the normal extrusion of the jelly, and the formation of the fertilization cone, but cannot stimulate the egg sufficiently to cause it to draw them in. This defect results in the abnormal development of the egg nucleus. It cannot be said that these spermatozoa are injured in the chrom- atin alone unless it is held that any process that weakens their vitality acts specifically on the chromatin. It is apparent, how- ever, that the chromatin is actually injured. This experiment shows that when failure of the nuclei to fuse occurs, the egg does not develop at all, thus differing radically from the sea urchin egg under similar conditions. b. The subsequent development of eggs fertilized by radiated sperm Cell division in the majority of eggs is normal, although usu- ally somewhat delayed. The chromosomes split in the normal manner and form vesicles at the telophase. Later the vesicles fuse to form the resting nucleus. No irregular mitoses were EFFECT OF RADIUM ON FERTILIZATION 99 found. Those eggs in which the germ nuclei fail to fuse do not divide. The first striking abnormality in the growing embryos appears when they are about twenty hours old. At this time the pro- trochal band of cilia and the green pigment at the posterior end are wholly lacking. Such embryos are of normal shape, but remain motionless, while the others in the dish are swimming actively about. In twenty-eight hours some are still lacking in ciliation, while others have developed the ciliary ring, which shows characteristic abnormalities in the distribution of the cilia. Instead of being in a continuous band, they occur in patches scattered irregularly. Such embryos swim slowly, and with a curious uneven motion, pursuing a devious course. In forty- eight hours, when the control embryos are well provided with green and red pigment and swimming vigorously at the surface, the radium embryos are still unpigmented, or else have the pig- ment broken up into small patches. The ciliation is still abnor- mal. The motion is consequently abnormal, the embryos tum- bling over and over, or taking a spiral course, turning on their long axis. In seventy-eight hours a few of the embryos develop setae and pigment, but both are abnormal. The setae rarely grow out to full length, and are irregularly distributed. The majority of the embryos that have survived as long as this remain in about the same condition that they showed at thirty hours. This mode of development occurs occasionally in the controls, especially if the conditions of growth are unfavorable, but never in as large proportions as among the radium embryos. Unfavor- able conditions obtain if many eggs are allowed to develop in a small amount of water, or if they become matted together after fertilization. If eggs are compressed under a coverglass during early cleavage they develop precisely the same type of abnormalities. In these embryos the effect of the injured spermatozo6n remains latent for many hours, and is displayed for the first time when the embryo begins to differentiate the pigment and the cilia, and begins to grow notably in size. 100 CHARLES PACKARD ~¢. The development of eggs radiated before fertilization with fresh sperm In the second series of experiments the unfertilized eggs were placed in a few drops of water in a watch glass. The glass tube containing the radium was about 1 em. in diameter and 1.5 em. in length. When this tube was set down into the water, the sur- face tension of the water was sufficient to draw the eggs up along the sides of the tube, thus distributing them fairly evenly, and ensuring an equal amount of radiation. Control eggs were treated in the same way except that the glass capsule, similar to the radium capsule was empty. In all of these experiments small quantities of eggs were used. The unfertilized eggs,-after two hours exposure to the radium rays show no change in appearance. The germinal vesicle and the alveolar layer are intact. High magnification shows no alteration in the protoplasm, or in the oil and yolk spheres. When such eggs are inseminated with fresh sperm many immediately throw off the jelly in a normal fashion. As a consequence the alveolar layer quickly disappears. The jelly is apparently normal in consistency, and loses its mucilaginous character in about fifteen minutes, just as with normal eggs. A considerable number of eggs, however, fail to extrude the jelly at once and consequently cannot push away the great quan- tity of spermatozoa that surround them, as can the normal eggs. Polyspermy results in such cases, many eggs being penetrated by a dozen sperm heads. Sections of such eggs show the alveolar layer still intact (fig. 9). This layer presents a fairly normal appearance, with alveoli closely crowded together, but not always radially arranged. Typically it extends entirely around the egg, and is about six or seven uw in diameter. But in these eggs it is very unevenly distributed, being sometimes absent altogether from one side, and massed in a thick layer on the other. There is no definite place at which the layer collects in these cases. The eggs showing this peculiarity are uniformly larger than nor- mal, the increased diameter being due to the greater width of the alveolar layer which varies from eight to fifteen » in thickness. EFFECT OF RADIUM ON FERTILIZATION 101 When massed at one point it exceeds this limit. The explanation for this phenomenon is not known. According to Lillie, the extrusion of the jelly takes place when the’spermatozoén be- comes implanted. As the jelly diffuses outward sea water passes in to take its place, so that the original size of the egg is maintained. If this is true, it may be that the sea water has entered faster than the jelly has been extruded, and consequently, the whole layer increases in width. However, this does not explain the fact that the alveolar layer is unevenly distributed. The increased size of the cortical layer does not persist as measurements show. Within seventy minutes after fertilization it has decreased noticeably, and after eighty minutes has dis- appeared altogether. The following figures are average measure- ments of a large number of eggs. Normal Radiated 45 minutes after insemination................ 106 X 92u 115 X 105p 60 minutes after insemination................ 105 X 87p 111 X 95p 70 minutes after insemination................ 105 X 90u 107 X 95p 80 minutes after insemination................ 104 * 9ly 105 X 94p It is thus seen that before the first division, which occurs in about ninety minutes, the eggs are of normal size. Sections show that the alveolar layer by that time has entirely disappeared, except in the few cases in which fertilization did not occur. In the control series the alveolar layer is extruded at once and nothing is seen of it inside the egg forty-five minutes after insem- ination. The germinal vesicle also breaks down soon after the - extrusion of the jelly. I found no exception to this. But in the radiated eggs the vesicle shows some curious modifications in behavior. It may break down at once even though the alveolar layer is still present, or it may remain intact for a considerable period. In the former case, the chromatin, which, during the process of dissolution of the nuclear membrane has collected in the form of chromosomes, is left free in the protoplasm. If the vesicle remains entire, the chromatin, which is now in granules, collects along the periphery, just beneath the membrane. This is the normal condition for it at this period. Occasionally the granules collect in the form of chromosomes which are scattered THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 102 CHARLES PACKARD through the nucleus suspended in a delicate non-staining network. After a time (varying in different eggs from forty to eighty minutes) the vesicle breaks down, and the subsequent history of the chromatin is the same as in the case of eggs in which it disap- peared at once. Its dissolution is apparently independent of the presence of the alveolar layer. Here, then, is a case in which the chromatin is not visibly affected by the radium radiations but the protoplasmic activities are much modified, although the structure of the protoplasm itself remains unchanged in appearance. This fact is in opposition to Hertwig’s view that only the chromatin elements are affected by the rays. The mode of entrance of the spermatozo6n into eggs which re- tain their alveolar layer is considerably modified from the normal. The fertilization cone rises up under the point of attachment of the perforatorium as usual, and later sinks into the egg, pulling the sperm head with it. But it takes no stain at all. Figure 12 shows an egg killed forty-five minutes after fertilization. The spermatozoén has become implanted in the egg membrane, and has begun to penetrate a short distance into the alveolar layer. No cone is seen. A later stage is shown in figure 10 in which the alveolar layer is still intact and is wider than normal. -The sperm heads have penetrated nearly their entire length. They never completely enter the protoplasmic area of the egg, but remain in the alveolar layer until it is extruded, at which time some of them continue their course inward. Intermediate stages in this process have not been found. It may be inferred that the sperms enter entirely since at the time that the cortical layer is extruded (in about eighty minutes) none can be found exterior to the egg, while within many may be found in various stages of development. The process of maturation in those eggs which fail to extrude the alveolar layer at once is much modified. The germinal vesi- cle disappears normally in the great majority of cases. .The chromatin, now free in the protoplasm, may move toward the animal pole where the spindle normally should form, or it may remain in the interior of the egg. Figure 11 shows a case in point. The chromosomes are ill-formed, show no particular arrangement s EFFECT OF RADIUM ON FERTILIZATION 103 and are deep in the egg. Careful search has not revealed any asters or spindle fibers. In figure 12 the spindles have formed, as well as the chromosomes, but the asters are lacking. Here, too, the spindle forms in the center of the egg. There are other cases in which very small asters are present, with fairly normal chromosomes. Occasionally figures like figure 13 occur, in which the achromatic portion of the figure is normal, but the chromo- somes are segmented. A more normal type is seen in figure 14. The normal condition seems to be due to the fact that the alveolar layer has been extruded at the periphery nearest to the spindle, allowing the usual relation to obtain between the protoplasm and the sea water in the egg protoplasm. In a majority of cases, however, the alveolar layer is extruded normally. In all instances the spermatozoa may enter normally or may show a characteristic abnormality, which, curiously enough, is similar to a condition found in normal eggs fertilized by radiated sperm. It consists in the failure of the fertilization cone to stain the usual dark color. Figure 15 shows such a con- dition. It will be noticed that the sperm is normally implanted and has already become anchored in the substance of the cone by minute attachment granules. The cone itself is transparent, and traversed by a delicate achromatic network. Immediately around it the protoplasm is of normal appearance. Later the cone sinks into the protoplasm in the normal manner, pulling with it the sperm head. The non-staining quality apparently has no effect on its function. Lillie suggests that the normal deep stain of the cone is due to a coagulating fluid injected by the sperm into the egg through the perforatorium. If this is true we have another evidence that the chemical reactions of the egg protoplasm, have been changed by the radiation. The formation of the first polar body may be very abnormal, but the abnormalities concern chiefly the achromatic portions of the spindle. As a rule the chromatin is but slightly affected. Figure 16 shows a well marked tri-aster with large centrosomes which are abnormal in size. It will be noticed that the chromo- somes, even in the minutest details of splitting are perfectly nor- 104 CHARLES PACKARD mal. This condition can hardly be interpreted as a first and second polar division combined, since it occurs much earlier than the normal time for the latter. The distribution of the chromo- somes shows also that the two lower asters do not represent the second spindle. The sperm is still exterior. Figure 17 is more normal, having but two asters. The spindle has not revolved, as it should have done at this time. The normal figures are at the telophase of the first polar spindle. The asters are of normal size and appearance, and the chromosomes show no abnormality of any sort. It is curious that they have taken up a position outside of the spindle, more or less in the place which they should occupy if the spindle had revolved. There are many other abnormalities in the formation of the spindle, but the examples given are typical. In practically no case in which the spindle forms do the chromosomes show any abnormality. The second polar spindle may be entirely suppressed or may develop abnormally in several respects. When the first polar body is not extruded, the second also may fail to be given off. Figure 18 shows what is evidently a second maturation. The egg was killed eighty minutes after insemination, at a time when the control eggs had already extruded the second polar body, and possessed a well developed egg nucleus. The spindles are normal in general appearance, but abnormally small, and placed in thé center of the egg. If is difficult to say what would be the future history of such an egg. Evidently the polar bodies will not be extruded. Figure 19 shows a similar case except that the spindles are larger. It will be noticed that one of the sperm asters has formed before the revolution of the sperm head and cone, an unusual occurrence. It is evident, then, that the injury done by the radium has affected both the chromatin and the protoplasm, particularly the latter. When the alveolar layer is not extruded the astral systems are abnormal and consequently the chromosomes. In such cases it is difficult to say how far the chromatin has been injured. EFFECT OF RADIUM ON FERTILIZATION 105 After the maturation period the remaining chromatin collects in one or more large vesicles of normal appearance. Occasionally the number of such vesicles may be very large and the karyomeres very numerous. The extra number may be due to the fact that the polar bodies have not been extruded, so that there is much extra chromatin. The subsequent behavior of these vesicles differs accordingly to whether they fuse with the sperm nucleus or not. In the former case the fusion may be entirely normal, and the subsequent division perfectly regular. Or the cleavage nucleus may never divide, but, on the contrary, increase consider- ably in size until it is nearly half as large as the germinal vesicle. The chromatin, in the meantime, breaks up into minute granules and passes out from the nucleus into the surrounding protoplasm. Figure 20 shows such an instance. The nuclear wall displays no breaks through which the granules might have passed out bodily. Such a condition might be explained on the basis that the egg chromatin had been injured by the radiations and had formed, as a result, a poison, which affects the sperm chromatin. The general effect of the radiation is to cause the chromatin to break up into granules. This is the explanation suggested by G. Hertwig. If the sperm nucleus does not fuse with the egg nucleus it fails to develop past the spireme stage, and in many cases does not develop even as far as that. Up to this time it has behaved normally. Even when many sperms enter (fig. 19) each one pursues the normal course in the egg. In figure 21 are shown two sperm nuclei in several separate vesicles, in which the karyomeres have fused and commenced to draw out into spiremes. Which vesicles belong to the egg and which to the sperm nuclei is diffi- cult to determine since they have become exactly the same in ap- pearance. In such a case, if a poison has been generated, it has failed to produce any marked effect on the egg chromatin itself. The only apparent abnormality is the failure of the nuclei to fuse. Division of the egg may occur without the appearance of asters and without division of the chromatin. Figure 22 shows an egg in which all the chromatin is in one blastomere. Figure 23 shows the protoplasm dividing without any division of the chrom- 106 CHARLES PACKARD atin. The vesicles, which have not developed into the spireme stage, are scattered irregularly about in the protoplasm. The large number of vesicles, with forty-two karyomeres, and two asters, indicates that two sperm nuclei are present. But they have ceased to develop and their astral systems have failed to develop normally. This failure is evidently not due to polyspermy for polyspermic eggs develop in a different and characteristic manner. To sum up: The effect of radium on the unfertilized eggs is seen after fertilization with normal sperm in both the protoplasm and the chromatin. The failure of the surface layer to be given off is due to a change in its constituents under the influence of radium. This failure leads to abnormalities in the formation of the polar bodies. The chromatin at first displays very little evidence of injury. Later, however, the injury shows in the fail- ure of the egg nucleus to develop properly. The sperm nucleus also fails to develop in these cases. d. The development of embryos from radiated eggs fertilized by normal sperm The later development of these eggs is not different from that of the normal eggs fertilized by radiated sperm, but the propor- tion of unfertilized eggs is greater, and the rate of development is slower. This may be inferred from the fact that the sperm, in many cases, cannot enter the egg for more than an hour after insemination because the cortical layer is still present. Occa- sionally some eggs divide into three or four parts at once, or, if into two blastomeres, the relative sizes of the two are abnormal. The subsequent development is also greatly retarded. When the control embryos are already in the early trochophore stage, these have not yet acquired the protrochal ring of cilia. A large proportion die before they reach the trochophore stage. Whether the proportion of abnormalities is greater in this series then in the first is hard to say, since always in the controls, a few embryos are abnormal. The type of abnormality that develops is precisely the same as that described in the preceding section. EFFECT OF RADIUM ON FERTILIZATION 107 e. The development of eggs normally fertilized, and exposed to radium In this series of experiments, the eggs were inseminated and stirred frequently for five minutes, then put over the radium cap- sule, which had an activity equal to 4 mg. of the pure bromide. At intervals a small quantity was taken out, half of the eggs being killed, and half put into sea water and allowed to develop. The early stages of fertilization are normal in all respects. The jelly is extruded at once (this began before the eggs were exposed to the radium) and consequently only one sperm enters. The external phenomena of implantation of the sperm and the rising up of the entrance cone are normal. A study of the sections shows that in many cases the fertilization cone does not take the usual dark stain, but remains as a very light, nearly transparent vesi- cle, similar to that described in the second series of experiments. (fig. 15). The sperm enters normally, however, as in the former case. In the meantime the polar bodies are normally extruded. The first indication of abnormal development is seen in the behavior of the germ nuclei. In many instances these do not fuse, but remain separated, at some distance from each other and inde- pendently develop chromosomes and normal spindles. Figure 26 shows an early stage of the process. The nuclear walls are still intact. The sperm nucleus has not developed as far as the egg nucleus, being still in the closely wound spireme stage. The egg nucleus with its aster has moved from its position under the polar bodies, away from the approaching sperm nucleus. Figure 24 is a later development. Probably the structure nearer the polar body is the sperm nucleus which has failed to develop much further than the stage shown in the preceding figure. The egg nucleus, however, has lost its nuclear membrane and has formed distinct chromosomes which are not as clear cut as normal, but show some evidence of becoming broken up into granules. Fig- ure 25 shows a still later development. The nuclei have migrated as far from each other as appears possible and liebetween complete and normal asters. In each nucleus the haploid number of chromosomes can be determined. In figure 27 the division has 108 CHARLES PACKARD gone still further. Judging from the size of the chromosomes, it seems probable that they have split, fourteen daughter chromo- somes going to each pole; but the number could not be exactly determined. The chromatin in this case is perfectly normal. In the instances mentioned, and in many others that have been observed, it is apparent that the plane of one spindle does not coincide with that of the other or with the first normal cleavage plane. Occasionally, when the nuclei fail to fuse, the astral systems do not develop, or, if they do, reman only a short time. The chrom- atin in such cases degenerates, forming irregular, granular masses which lie free in the protoplasm (fig. 28). One germ nucleus has evidently degenerated faster than the other and its chromatin has become much scattered and takes the hemetoxylin stain very lightly. When the fusion of the nuclei occurs the process appears to be normal. Abnormalities appear when division begins. As a rule the chromatin shows few signs of injury, but the astral systems are abnormal. Very commonly a tri-polar figure arises (fig. 29). The asters-are normal in appearance and the centrosomes are normal. The chromosomes are very irregularly arranged, as would be expected. But each one is well formed, showing no evidence of degeneration. Frequently, however, the chromatin displays undoubted signs of injury, as shown in figure 30. Not only are the chromosomes indistinguishable but some of the strands are now broken up into minute granules. There are all gradations between such a con- dition and a normal figure of which there are very few in the material. The effect, then, of the radium on fertilized eggs is much more marked than that on the unfertilized eggs. This accords with the results of O. and P. Hertwig, who found that the embryo frogs arising from radiated fertilized eggs showed more marked abnor- malities than those in the other series. The effect on the chrom- atin are similar in some respects to those described by P. Hertwig in the egg of Ascaris. It is evident, however, that the protoplas- ric activities of the eggs of Nereis are almost equally affected EFFECT OF RADIUM ON FERTILIZATION 109 with the chromatin, and that the fertilized eggs in the two cases respond in very different ways. f. Subsequent development of the embryo The great majority of the eggs do not divide. In some cases they may divide into three or four blastomeres at once, as would be expected from the division figures already described. Very few of the eggs continued to divide after ten hours of development. A small number of early trochophores developed but no pigment formed in them, and the cilia were abnormal. No embryos sur- vived after twenty-four hours. g. Development of eggs radiated before and after fertilization with normal sperm The early phenomena of fertilization in eggs radiated for two hours before fertilization and for varying periods after fertilizaon, do not differ to any extent from those described in the second series. In some lots of eggs the alveolar layer is extruded nor- mally in almost every instance, while in other lots it 1s retained for some time after the implantation of the sperm. In the latter case polyspermy invariably occurs, as many as ten sperms entering at practically the same time and developing at the same rate. The mode of entrance of the sperm in either case is the same as that described in the second series. In many instances the fer- tilization cone is an almost transparent vesicle. But, as in the previous cases, this change in the staining reaction does not indi- cate any change in function, for the spermatozo6n is drawn in normally, and the whole process is without any other indication of abnormality. The further development of the sperm head is normal. The head and cone revolve, and an aster forms at the base of the head. The chromatin, as the head enlarges, collects into afew karyomeres. But the number of vesicles which contain these structures may be very large. In some eggs there may be as many as forty, indicating that each sperm has broken up into at least four. Their subsequent development will be described after the behavior of the egg nucleus has been traced. 110 CHARLES PACKARD The germinal vesicle breaks down as a rule, within a few minutes after insemination, even though the alveolar layer is still present. The egg centrosome divides normally, the daughter centrosomes migrating to the opposite sides of the group of chrom- osomes which lie free in the protoplasm without any definite ar- rangement. If the alveolar layer has been extruded the spindle usually moves to the surface where it develops normally in many cases. In such instances the chromosomes are not at all affected, but split into the normal crosses and rings with great precision. The achromatic figure also appears to be normal. When the alveolar layer is not extruded at once, and, in some cases, even when it is, the whole first polar figure remains near the center of the egg. Judging from the scarcity of extruded polar bodies, it is probable that the polar body is not often extruded. There are, however, several marked departures from the course outlines above. Figure 31 shows a condition found in several eggs. There is but one centrosome and no real aster. The spindle fibers are normal and attached to the chromatin masses. No normal chromosomes are present, the chromatin remaining either as a long narrow thread, or as irregular, homogeneous masses. In each instance it is in two distinct groups separated by a varying distance. The size of the chromatin thread indi- cates that splitting has not occurred in the right hand group. In the other group the chromatin is in very irregular masses of unequal size. Figure 32 is a later stage of the same condition. In this egg there is a normal sperm head and cone already revolved, — but the sperm aster is not apparent. Some of the chromosomes are fairly normal, while others are mere masses of chromatin. The separation of the groups and of the spindle fibers attached to them is a curious phenomenon. In some instances both polar bodies are given off, but in many cases none at all are extruded. During this time the sperm heads have developed up to the vesicular stage, as has been mentioned above. Their further development may proceed along different paths. Jn some in- stances the karyomeres apparently break down and fuse together, taking the form of fine spireme, or even breaking up into chrom- osomes. If several sperm heads enter there are at least as many EFFECT OF RADIUM ON FERTILIZATION Itt asters, and in many cases, a much greater number. But such asters are not connected with each other by spindle fibers nor are they centers for chromatin masses. In some instances a great number of asters is present but their origin is uncertain. If we assume that each original sperm aster has divided normally, it would be necessary that fifteen sperms entered the egg about at the same time. This number is greatly in excess of any observed ease. The extra asters may have arisen de novo. Perhaps the most common abnormality is the presence of a great number of vesicles without any aster. In view of the facts just presented, namely, that sperm asters develop when poly- spermy occurs, it is evident that the extra vesicles represent a single sperm nucleus that has divided, or else the egg nucleus, or finally, sperm nuclei that have lost their asters. Figure 33 is a case in point. The only aster present is extremely small. The vesicles, of which there are a great number in the egg, are still intact although the karyomeres have about disappeared, having gone into the spireme form. At this stage we should expect the germ nuclei to fuse but in this case, one hundred minutes after insemination, when the control eggs have already divided, we find no evidence of fusion. Such eggs do not divide but show the very common phenomenon of budding. By many observers this phenomenon has been described in eggs that are rapidly degenerating. In the eggs of Nereis, at about the time when cleavage should take place, the protoplasm at the periphery separates off into buds of irregular size and shape. They contain no chromatin unless by chance some of the vesicles are in the région where the budding occurs. The protoplasm of the buds and the region directly contiguous shows marked signs of degeneration, appearing to be more fluid and possessing only a few protoplasmic granules, instead of the closely crowded quanti- ties present in the more normal regions. The general effect of the radium radiations on the eggs treated in this manner seems to be more on the protoplasm than on the nucleus. The surface layer is greatly affected and as a conse- quence may fail to extrude the alveolar layer. When division would normally occur, only a budding takes place. If the sperm 112 CHARLES PACKARD asters are developed from the egg protoplasm, it is evident from their peculiar behavior that the latter substance has been injured. The chromatin also is affected, as is shown in the cases of irregu- lar formation of the maturation chromosomes, and in its be- havior in the nuclear vesicles. But the many cases of perfect formation of the chromosomes and their normal splitting, even in the minutest details, indicate that the effect of the radiations on the chromatin has not been profound. DISCUSSION The work of G. Hertwig and others points to the conclusion that the chromatic elements of the nucleus are the only parts of the cell which are disturbed by the radium radiations. The con- clusion that no other constituents are affected has been drawn from the fact that the usual methods of technique do not reveal any chemical changes in the protoplasm. Such a criterion is, in many cases the best we have, but it is evident that staining will not reveal all of the cell constituents which are known to be present, which, like enzymes, operate on the protoplasm of the cell, changing its chemical constitution; nor will it give evidence of the nature of such changes. There are no visible changes in the radiated sperm—a fact that led Hertwig to believe that the protoplasm was uninjured. But I believe that in the sperm of Nereis there is evidence that a proto- plasmic change has taken place. This change becomes apparent through the activities of the egg in the early stages of fertiliza- tion—activities which are incited by the stimulus of the sperm itself. In the first place, when the sperm becomes implanted in the egg membrane it injects into the egg a coagulating substance which, according to Lillie, changes the protoplasm so that it takes a fairly deep stain with hematoxylin. This staining reaction is frequently lacking. It cannot be said that the egg protoplasm has been changed since all of the eggs used in the experiment were normal. Therefore it may be inferred that the protoplasmic elements in the sperm have been modified. EFFECT OF RADIUM ON FERTILIZATION iS Secondly, the sperm in many instances fails to gain entrance into the egg. The engulfing of the sperm is undoubtedly due to the activity of the egg, since the sperm is motionless from the moment it becomes implanted in the egg membrane. Such ac- tivities are incited by some substance in the sperm itself, and are not due merely to the pricking of the membrane by the perfora- torium, for in many instances the implanted sperm does not enter. Theseat of such a stimulating substance is not certainly known. Probably it does not reside in the nucleus, since all of activities of the egg, such as the throwing off of the cortical layer, the formation of the fertilization cone, and its subsequent movements, occur while the nucleus is still intact and entirely outside of the egg. It is very likely that the stimulus resides in the head cap of the spermatozo6n, and is injected into the egg with the coagulating fluid. The failure of the sperm to enter is due to the injury to this substance and a consequent failure properly to activate the egg. Such an injury is not due specifi- cally to the radium treatment since old spermatozoa, kept fourteen hours in a vial will produce the same results. Probably any treatment which weakens them in any way has this effect. These two instances, I believe, indicate that the sperm proto- plasm has been injured by the radium radiations. That the chromatin of the sperm has been injured has already been pointed out. The difference in behavior between the Nereis sperm and that of the sea urchin, as described by Hert- wig, should be mentioned. In the former case, the sperm nucleus fails to develop past a certain point, and fails to fuse with the egg nucleus. The latter structure also fails fo develop due apparently to the failure of the sperm centrosome. In the sea urchin, on the other hand, the sperm centrosome develops perfectly, becoming the dynamic center of division for the egg nucleus, which divides by itself as a rule in a normal manner. If the sperm nucleus has migrated into a position close to the spindle it acts as a foreign body in disturbing the normal be- havior of the chromosomes. In some cases the two nuclei fuse, but the sperm nucleus is soon eliminated in subsequent divi- 114 CHARLES PACKARD sions. Development is consequently parthenogenetic, a condition which does not obtain in the Nereis egg. It should be noted that the sea urchin egg is an especially favorable subject for par- thenogenetic development, while Nereis is not. Consequently the two cases are not exactly parallel and the behavior of the’ one should hardly be expected to be repeated in the other. The generally accepted theory that the cleavage centrosomes are derived from the divided centrosomes brought in with the sperm has recently been questioned by Lillie who believes that in Nereis the centrosome appearing at the base of the sperm head is derived through the interaction of the sperm chrom- atin and the egg protoplasm. The fact that the middle piece is always left behind on the outside of the egg membrane seems to support this view. A discussion of the theory is not in place here, but it may be pointed out in passing that this fact alone is not conclusive evidence that no formed centrosome has been introduced by the sperm, since in many insect sperma- tozoa the centrosome is contained in the sperm head. If, however, ih Nereis, the centrosome is a product of the sperm chromatin and the egg protoplasm we have evidence that the sperm nucleus has been injured. If, on the other hand, the older view as to its origin is true we have evidence that the radium radiations have affected a protoplasmic structure. A discussion of the effect of radiations on the unfertilized egg raises some interesting questions concerning certdin protoplas- mic activities of the egg. It has been shown that in many eases the cortical zone, composed largely of colloidal material is not extruded at the time of implantation of the sperm, but remains for a considerable period within the egg, and is given off slowly during subsequent development before cleavage. As a result polyspermy occurs. The extrusion of the surface layer is due, according to R. 8. Lillie, to the sudden increase of per- meability of the plasma membrane, resulting in an outflow of substances previously impermeable to it. That the membrane is rendered somewhat permeable at the time of implantation of the sperm is evident from the fact that water is taken up, result- ing in a swelling of the alveoli. But the normal permeability is EFFECT OF RADIUM ON FERTILIZATION 115 not fully attained. The failure of the cortical layer to be given off may be due to a change in the plasma membrane, or to a change in the colloids of that layer. The possibility of such a change will be discussed later. The formation of the polar bodies has been shown to be ab- normal both in respect to the chromatic and achromatic elements of the spindle. It might be concluded from this that both parts were equally affected by the radium radiations but this is not necessarily true, for it has been shown that the presence of the cortical layer has an inhibiting effect on mitotic phenomena. This suggestion has been made by Bataillon who holds that in the cortical layer are certain katabolic products of metabolism which hold the egg in check until they are given off. The in- creased permeability at the time of fertilization allows them to be extruded. The resulting egg membrane is therefore a result of such an action and is incidental, indicating merely that certain more fundamantal phenomena have been taking place. In Nereis, when the cortical layer persists, the maturation divisions are always abnormal, remajning permanently in the interior of the egg or else showing other defects which have been described above. If, however, the cortical layer is extruded in part, the spindle migrates to that region and the figure is normal. It should be mentioned, however, that occasionally in eggs from which the cortical layer has been extruded, the same type of ab- normalities may be found. The radium radiations therefore, have affected not only the surface layer, but also the deeper lying protoplasm. That the abnormalities in the chromatin appear during the formation of the chromosomes, when they are being built up from the protoplasmic constituents indicates that the latter themselves may have been modified, or that the agents through which they are so built are affected, as well as the chromatin itself. These abnormalities are even more strikingly manifested in the division of the fertilized egg, as described above, and also by P. Hertwig. If we attempt to explain the phenomena already described by the hypotheses of Schwarz or of Hertwig we are at once involved 116 CHARLES PACKARD in difficulties. The latter involves a number of difficult assump- tions. First, when the sperm is radiated, the poison generated in the nucleus must remain altogether in that structure, since the protoplasmic activity concerned in movement is not affected, or else it must be assumed that the poison is specific for the chro- | matin itself, not affecting the protoplasm. And secondly, it must exert no influence on the egg protoplasm during division, since a normal haploid division can occur provided the sperm head does not mechanically interfere in the spindle. If the sperm head fuses with the nucleus its poison exerts no influence on the egg chrom- atin, other than to cause it to eliminate the sperm chromatin. But it has been shown that in Nereis the protoplasm of the sperm has evidently been changed. Hence it would be necessary to postulate that the poison had extended into the protoplasm and had affected the sperm centrosome—a condition contrary to the first supposition. In the egg it is very evident that the proto- plasm has been affected. Hence the hypothesis that poison is generated solely in the nucleus is not tenable. A modification of the poison theory will, I believe, explain more of the facts, and such a modification will be discussed later. The lecithin hypothesis is hardly as satisfactory as the other. The fact that the sperm nucleus is undoubtedly affected by the radiations makes it evident that they are not specific for the lecithin in the protoplasm. The behavior of the radiated egg is more easily explained for it has been shown that the cortical layer, which contains considerable lecithin is affected. If that substance which is distributed throughout the egg is utilized in the upbuilding of the chromatin, the phenomenon could be explained. These two hypotheses, therefore, fail at certain points when they are used to explain the facts observed in the fertilization of the egg of Nereis. I believe that I have given good evidence that both hypotheses may be true in part, since both protoplasm and chromatin are affected by the radium radiations. The question now arises as to the means by which such changes can be brought about. EFFECT OF RADIUM ON. FERTILIZATION ay The solution of the problem lies, I believe, in the fact that the protoplasmic and nuclear elements are not directly affected by the radiations but only indirectly by means of enzymes which are activated by the treatment. In support of this view may be mentioned a number of observations. The experiments of Schwarz (’03) and of Wohlgemuth (’04) prove that lecithin is readily decomposed by treatment with radium. But the latter observer, who investigated the action of radiations on pure organic substances such as olive oil, asparagin, peptone, and finally lecithin, found entirely negative results. In no case were there any observable changes in the substances. From this he concluded that the breakdown of the cells and tissues which con- tain these substances was due, not to the direct effect on them of the radiations, but to the increased autolytic activity of enzymes in the cells. The question whether all cell constituents (nucleo proteids, simple proteids, lipins, carbohydrates, and salts) are affected can- not be determined, but an experiment by Wohlgemuth, the re- sults of which are significant in this connection, throws some hight on the subject. To test further his hypotheses that the autolytic enzymes are activated by the radium radiations, he exposed portions of tuberculous lung for varying periods, and found that the total amount of nitrogen given off was at first increased fourfold. In the end, however, the control material showed the same total amount as the other portion. But the experiment indicated that the enzymes which bring about a decomposition of the nitrogenous material in the cell are greatly activated. We should look therefore, to the proteins and to lecithin for evidences of decomposition. That enzymes are activated by radium radiations has been shown by numerous observers. They exert an accelerating effect on peptic digestion, on the diastatic enzymes of the blood, liver, saliva, and pancreas. ‘‘This favorable action is not always ob- servable immediately; very often retardation occurs during the first twenty-four hours, this being gradually neutralized, and then replaced, if the experiment is sufficiently prolonged, by an acceleration”? (Euler). The promotion of plant growth is also 118 CHARLES PACKARD attributed by Falta and Schwarz (11) to enzyme activation. Gager (’08) and Congdon (’12) have also found some evidence of acceleration in growth, if the intensity of the radiation is not too great. When Drosophila eggs are placed close to a strong prep- aration of radium, their development is retarded. From these experiments it may be concluded that many kinds of enzymes may be activated but that the lytic enzymes are more stimulated than those that play a synthesizing role. The decomposition of the nitrogenous compounds of the cell is in part normal and in part abnormal. Lecithin breaks down into cholin and finally into trimethyl amine, a reaction which does not take place in life. But the decomposition of the nucleo- proteids occurs constantly under normal conditions. Under the influence of enzymes these substances are oxidized to nucleinic acid and finally into some of the purin bases (guanin, hypoxanthin, etec.). The reverse reaction also occurs, by which the complex nucleo-proteid is built up again from the simpler materials. The seat of this process is at the nuclear wall, where the nucleo-pro- teids and the protoplasmic proteins adjoin. If, then, we assume that under the stimulus of the radiation the katabolic changes in the nucleo-proteids takes place at a more rapid rate than the synthesizing reactions (an assumption that is warranted on the basis of Wohlgemuth’s results), we have an explanation of the behavior of the radiated chromatin in breaking up into granules, and failing to divide normally. The acceleration in growth, found by Gager and Congdon when the material was exposed to slight radiation, may be explained on the ground that under such conditions the synthetic processes are stimulated, while under a more intense radiation the opposite reaction obtains. The behavior of the radiated cells must, I believe, be inter- preted in the light of the increased activity of these autolytic en- zymes which act both on the chromatin and on the protoplasm. A radiated spermatozo6n is affected chiefly in its nuclear constitu- ents which are broken down into simpler compounds. That these are still acid is indicated by their staining reactions. When such a cell enters the egg, the nucleus is unable to develop normally, and to divide with the egg nucleus, because it cannot build up EFFECT OF RADIUM ON- FERTILIZATION 119 nucleo-proteids at a normal rate. Hence it is either eliminated, as Hertwig has shown, or remains stationary, after developing up to a certain point. In like manner, the egg nucleus, after the egg has been radiated, fails to develop. That the normal sperm which fertilizes such an egg may develop, as suggested by Hertwig, indicates that the egg protoplasm, made up of “‘simple” proteins, has not been greatly changed. In Nereis evidently such proteins have been affected, as shown by the abnormal behavior of the sperm nucleus. When the fertilized eggs are radiated, the radia- tion is applied at a period when the synthetic process in the formation of the chromosomes is most active. Hence, any acceleration of the autolytic enzymes results in a speedy deterior- ation of the chromatin. The latent period which is found in many instances when the radiation is neither prolonged or in- intense may be explained by this hypothesis. Only a slght acceleration is brought about by the treatment, but the effect of autolysis is cumulative, and becomes manifest after a period, depending on the rate of acceleration. The nuclear material has been broken down to such an extent that, after division, the chromosomes cannot be built up again. In this connection it might be suggested that the normal nuclear division takes place when the ratio between the amount of nucleoprotein and simple proteins of the protoplasm has reached a certain limit. After the division the ratio is changed and constructive action again takes place. SUMMARY AND. CONCLUSION 1. Radiated spermatozoa of Nereis react in two ways to the normal egg. They may normally stimulate it, and be drawn in, but subsequently fail to develop, or they may fail to stimulate the egg sufficiently and so remain external. In the first case, the sperm nucleus and aster may fail to develop and to fuse with the ege nucleus. In the second case, the egg nucleus develops with- out an aster. 2. The radiated egg at the time of fertilization may or may not extrude the cortical layer. In either case, the maturation phe- nomena are more or less abnormal. The germ nuclei develop 120 CHARLES PACKARD abnormally and mitosis does not occur, although the protoplasm may divide. 3. Radiation of the fertilized egg results either in a failure of the fully developed germ nuclei to fuse, or in abnormal division of the cleavage nucleus. 4. Eggs radiated before and after fertilization show very marked evidences of protoplasmic degeneration. 5. In general, both protoplasm and chromatin are affected. 6. The previous hypotheses do not satisfactorily explain these facts. 7. It is suggested that the radium radiations act indirectly on the chromatin and protoplasm by activating autolytic enzymes which bring about a degeneration of the complex proteids, and probably by affecting other protoplasmic substances in the same manner. EFFECT OF RADIUM ON FERTILIZATION 121 LITERATURE CITED BarDEEN, C. R. 1909 Variations in susceptibility in amphibian ova. Anat. Kec:, vol. 3: 1911 Susceptibility of amphibian ova to X rays. Am. Jour. Anat., vol. 2. BaTalLuon, E. 1910 Le probléme de la fécondation circonscrit par l’impregnation sans amphimixie, et la parthénogénése traumatique. Arch. de Zool. Exper. et. Gén. 5 ser. Tom. 6. Boun, G. 1903 I. Influence des rayons du radium sur les animaux en voie crois- sance. II. Influence des rayons du radium sur les oeufs vierges et fécondés et sur les premiers stade du développment. Compt. rendus. BonnevikE, K. 1909 Chromosomenstudien No. 2. Arch. f. Zellforschung, Bd. 2, Heft. 2. 1910 Uber die Rolle der Centralspindel wihrend der indirecten Zell- teilung, ibid Bd. 5. CoHNHEIM, O. 1912 Enzymes. New York. Wiley. Conepon, E. D. 1912 Effects of radium on living substance. I. The influence of radiations of radium upon the embryonic growth of Drosophila ampelo- phila, and upon the regeneration of Tubularia crocea. II. Comparison of the sensitiveness of different tissues in Allolobophora feotida, and in Cambarus affinis to the beta rays of radium. Bull. Mus. Comp. Zodl., Harvard, vol. 53, nos. 7 and 8. Conkuin, E.G. 1912 Cell size and nuclear size. Jour. Exp. Zodl., vol. 12. Euter, H.1912 General chemistry of the Enzymes. New York. Wiley. Fiscuer, M. H. 1902 Artificial parthenogenesis in Nereis. Am. Journ. Physiol., Vol. 4. GoptewskI, E. 1908 Plasma und Kernsubstanz in normalen und der durch aus- sere Factoren verinderten Entwicklung der Echiniden. Arch. f. Ent- wickl-Mech., Bd. 26. 1910 Plasma and Kernsubstantz bei der Regeneration der Amphibien. Arch. f. Entwickl-Mech., Bd. 28. GUILLEMINOT, M.H. 1908 Effets des rayons X sur le cellule végétal. Journ. de Physiol. et de la Pathol. génerale. HamMarsTEN, ©. 1909 Textbook of physiological chemistry. New York. Wiley. Hertwic, G. 1911 Radiumsbestrahlung unbefruchteter Froscheier and ihre Ent- wicklung nach Befruchtung mit normalem Samen. Arch. f. Mikro. Anat., Bd. 77: 1912 Das Schicksal des mit Radiumbestrahlten Spermachromatins im Seeigelei. Arch. f. Mikro. Anat., Bd. 81. 122, CHARLES PACKARD Hertwic, O. 1911 Die Radiumkrankheit tierischer Keimzellen. Arch. f. Mikro. Anat., Bd. 77. Hertwic, P. 1911 Durch Radiumstrahlung hervorgerufene Veriinderungen in den Kernteilungsfiguren der Eier Ascaris megalocephala. Arch. f. Mikro. Anat., Bd. 77. Just, E. 1912 The relation of the first cleavage to the entrance point of the sperm. Biol. Bull., vol. 22. Korosy, K. v. 1911 Radioactivitét und Fermentwirkung. Arch. f. Gesammt. Physiol., Bd. 137. Kostaneckl, K. 1904 Cytologische Studien an kiinstlich parthenogenetisch sich entwickelnden Eier von Mactra. Arch. f. Mikro. Anat., Bd. 64. Levy, O. 1906 Mikroscopische Untersuchungen zu Experimenten tiber den Ein- fluss der Radiumstrahlen auf embryonale and regenerative Entwick- lung. Arch. f. Entwickl.-Mech., Bd. 21. Litun, F. R. 1902 Differentiation without cleavage in the egg of the annelid Chaetopterus pergamentaceous. Arch. f. Entwickl.-Mech., Bd. 14. 1911 Studies of Fertilization. I. The cortical changes in the egg. II. Partial fertilization. Jour. Morph., vol. 22. 1912 III. The morphology of the normal fertilization of Nereis. Iv. The fertilizing power of portions of the spermatozoon. Jour. Exp. Zool., vol. 12. Lituiez, R.S. 1909a Oxidative properties of the cell nucleus. Am. Journ. Physiol., vol. 7. 1909b General biological significance of changes in the permeability of the surface layer or plasma membrane of living cells. Biol. Bull., vol. 17. 1911 The physiology of cell division: The action of salt solutions fol- lowed by hypertonic sea water on unfertilized sea urchin eggs, and the role of membranes in mitosis. Am. Journ. Physiol., vol. 27. Mastne, E. 1910 Uber das Verhalten der Nucleinsaure bei der Furchung des Seeigelei. Hoppe Seylers Zeitshr. Bd. 67. MENDEL AND LEAVENWORTH, 1908 Chemical studies on growth. VII Am. Journ. Physiol., vol. 21. Prertues, G. 1904 Versuche iiber den Einfluss der Roentgen— und Radium- strahlen auf der Zellteiling. Deutsche med. Wochenschr. Jahrg., Bd. 30. RUTHERFORD, EH. 1905 Radioactivity. SHakE.L, L. F. 1911 Phosphorus metabolism during early cleavage of the echino- derm egg. Science, vol. 34, no. 878. EFFECT OF RADIUM ON FERTILIZATION 123 ScuHaper, A. 1904 Experimentelle Untersuchungen iiber den Einfluss der Radiumstrahlen auf embryonale und regenerative Vorgiinge. Anat. Anzeig., Bd. 25. Scumipt, H. E. 1907 Uber den Einfluss der Roentgenstrahlen auf die Entwick- lung von Amphibieneier. Arch. f. Mikro. Anat., Bd. 71. Scuwartz, G. 1903 Uber die Wirkung den Radiumstrahlen; eine physiolo- gischechemische Studie am Huhnerei. Arch. f. Physiol., Bd. 100. Tur, J. 1906 Sur l’influence des rayons du radium sur le développment de la rousette Seyllium canalicula. Arch. de Zool. expér. et gén., tom 5. Werner, W. 1905 Zur Kenntniss und Werwertung der Rolle Lecithin bei der biologische Wirkung der Radium- and Roentgenstrahlen. Deutsche med. Wochenschr. WoxuucemouTH, J. 1904 Zur Kenntniss von der physiologischen Wirkung des Radium. Berliner klinisch. Wochenschr. Witson, E. B. 1901 Experimental studies in cytology. II and III. Arch. f. Entwickl. mech., Bd. 13. 1906 The Cell. New York. Macmillan. ZvuEuzeR, K. 1905 Uber die Wirkung der Radiumstrahlen auf Protozoa. Arch. f. Protisten., Bd. 5. All the drawings were made with a camera, with a 2 mm. Zeiss apochromatic objective, and a No. 6 compensating ocular, unless otherwise stated. PLATE 1 EXPLANATION OF FIGURES 1to4 Normal Fertilization. 1 Attachment of the sperm; 55 minutes after insemination. 2 Penetration of the sperm head; 60 minutes after insemination. 3 Telophase of the second maturation division, and approach of the sperm nucleus; 65 minutes after insemination. 4 Just before the fusion of the germ nuclei; 90 minutes after insemination. 5to9 Fertilization of normal eggs by radiated sperm. 5 Penetration of the sperm head, showing absence of the fertilization cone; 60 minutes after insemination. 6 Development of the egg nucleus, and failure of the sperm nucleus to enter; 115 minutes after insemination. 7 Gradual disappearance of the sperm aster; 105 minutes after insemination. 8 Total disappearance of the sperm aster. The sperm nucleus has divided into many vesicles; 115 minutes after insemination. 9to29 Fertilization of the radiated eggs by normal sperm. 9 Polyspermy. This egg has failed to extrude the alveolar layer. 2 mm. apochr. and No. 0 eyepiece. 10 Detail of the same; 70 minutes after insemination. 11 Formation after the first maturation figure without astral system. 12 Same, with spindle but no asters; 45 minutes after insemination. 13. Same. The chromatin is segmented, but is splitting longitudinally 45; minutes after insemination. 124 EFFECT OF RADIUM ON FERTILIZATI( IN CHARLES PACKARD PLATE 1 2 ‘s & Re Re Ses 2 A ars OH I uae” ig 5 r = J ® 4 ta & .e 5 x o%@e® e @e e,. @ .: ® 5 os: Py! ge es e ®%e Py, a” 9 e. BS i @ @ *e@ 6 e 2«@° ®. 8 a s°? 3 2 - ————— ie: @ \ 7 } ve 8 & e O° © vA ZO e { vs © es ¢ > x e 2 & oe. Je” x ge ®e 4) J ) @ e oe E: ~ Pe TI es ae vee a . ye CL ae “oe ®e ¥ AS te 4 pate 12 ray go : PLATE 2 EXPLANATION OF FIGURES 14 First maturation division occurring at a point where the alveolar layer has been extruded; 60 minutes after insemination. 15 Entrance of the sperm, showing the non-staining vacuole. 16 Tripolar spindle of the first maturation; 55 minutes after insemination. 17 First maturation spindle, showing an abnormal distribution of the chromo- somes; 60 minutes after insemination. 18 The second maturation. The first polar body has not been extruded; 80 minutes after insemination. 19 The same; two sperms are entering in one, the aster has appeared before the revolution of the cone. 2mm. X No.0. 20 The cleavage nucleus. The chromatin is passing out through the nuclear wall; 120 minutes after insemination. 21 Polyspermy; 70 minutes after insemination. 22 Division of the protoplasm without a distribution of the chromatin. 2mm. apochr. X No. 0 eyepiece. 23 The same; the same magnification. 24 to 30 Normally fertilized eggs radiated. 24 Approach of the germ nuclei; 105 minutes after insemination. 25 The nuclei have migrated to oppositie sides of the egg; 105 minutes after insemination. PLATE 2 EFFECT OF RADIUM ON FERTILIZATION CHARLES PACKARD @e® ° @oc®. 22: 8. @ ° @®. ee0 @ e: @ %6 de e o°. 9 ~~ Gy = = %e = @e eon x4 ®e@ v °° @ %e ® e ° e Se @ ©.; o@ ee. ae . TQ oe ay 2 ; f ae aon . ° 6 oe © @ e joe fe e ~ e@ e ¢ 3 e e oS, ) : 18 ae - ©, 86 ma) to ©, e. ~ -@ e@ ax gi 7 e @*ee ®. te . Ss yes @ ® @. 2 + . 3 oe @ fe \ @ as o. : —_ fe on > J Ret} ars —E [ Pe Py EAT, 23 ~ a “es 22 cu we =) a a 5 Jz ioe 7 ad) S< a =." /y 24 25 30 PLATE 3 EXPLANATION OF FIGURES The same. Each germ nucleus is dividing by itself; 115 minutes after insemination. Disintegration of the germ nuclei; 115 minutes after insemination. Tripolar spindle; 120 minutes after insemination. Degeneration of the germ nuclei; 120 minutes after insemination. 31to34 Eggs radiated both before and after fertilization. 31 32 33 34 Abnormal polar body formation; 60 minutes after insemination. The same. The egg showing the phenomenon of budding. The same; both 120 minutes after insemination. EFFECT OF RADIUM ON FERTILIZATION CHARLES PACKARD —_ — — i - —S ~ — Cg —— > ~ 4} 9 } a0 S 30 Ae VA oN > Ane : UN wr ¢ 28 27 So Me... e . i ~o) atc0o “eo @-. ~ \ és een? e A eee,e%e > /@ 0.2) @e® S\ e"ee@ °°? 31 32 e*/ @°5 © x on ~ @°%e = if Ps e corey {% @ \See “@ a ON °*. pate ye @ IS e \ & & e 8 Se a : ag: aoe foe ee = ee? ° e° e ee PLATE 3 A FURTHER STUDY OF SIZE INHERITANCE IN DUCKS WITH OBSERVATIONS ON THE SEX RATIO OF HYBRID BIRDS JOHN C. PHILLIPS Laboratory of Genetics of the Bussey Institution SEVEN CHARTS In a previous number of this journal the writer reported the result of a cross between two races of ducks differing in size. The number of offspring produced was small, but tended to show a marked increase of variability in size in F,. In the breeding season of 1912 more data were obtained from the same stocks, 57 F, and 31 F, ducks being raised to maturity. Growth curves were made for all these ducks from an early age until such time as they were considered adult, or nearly so. The average length of this period is 142 days. The growth curves give a good check on the adult or autumn weight of the individ- ual, and enable one to assign the proper weight figure to each bird with very fair precision. The ducks were weighed at intervals of one week to two weeks. The curves will be considered later on, together with the growth rate of a race of pure wild mallards reared in 1912. As a further test of the parent stocks, which it may be recalled consisted respectively of Rouen ducks and domesticated mallards of a particular strain, 9 pure Rouen and 20 mallards were reared in 1912. The weight of these animals at five months of age is recorded in table 8. To produce another lot of F; ducks, one of the original Rouen males was mated with three female mallards taken at random from among the same birds used in the previous work. For the F, generation, the same three F, females as had been used the previous year, Nos. 83, 101, and 106, were mated with F,; male 151 PHILLIPS JOHN C. 132 6 OL vy GI PLI 696 G6S1 L Ol G6 CI Soe |= Con ¢ OL 00 ST COL 90 | OFT COL Gr OL PST ELT | 6LF I 20 IT LG GTI ISI VGG PEor SPV € G8 G PS 16 L8SI 9ST 91° 6 Sil | Zit | PZ0T Silas vl IT OS We 6s 4) 86 9g 2 € 6 TL] GCG L966 OSES 6OVIG 96ES GEGG ries 1 eae SHIVNGA wo saivw ao | SXVUS NI ae fe ea Oe ae acne’ Re OILVIA ae | esi ead ‘ GQUuvanyvis | GdUVANV.LS NVGAW OPLI L99T OOLT O99T ISLT SOOT SPLIT Ss90T OL9G OSGG GESG OGL SN VUD Sa IVW 40 LHDIGM NVGW OF VE ine OP ial val 1G 9¢ 9] Mil (5 Ol GIL Ms €V 6S ce {noqge OF FHoGe 61 Gl ¢ i 6 9 ¢ G P G STIVNAA | Sav AO ne) ua aWwoON UqAaWOAN SAID. Asay) Waanjaq splighy WL pup ‘syonp pppoe pup uanoy we pybram T WIdaViL “SOUL G “SOUL G sAep 6CT 94 SET SARP OOT OF FET CVp [G1 4 PEL ‘soul G “soul ¢ ‘SOUL G Yen “soul UT ‘soul Gg G “SOUL G siBok & GOV UO SUOLPDALISGO JO ‘Uesd Fy ‘uos |] ‘Uosd Fy] ‘uod Ty ‘Uo 35] ‘uos Ny ‘Uosd pe ‘los pz YOOYS [BUISLIO | pourq -Wl0d f pus e ‘Z| ‘Udd VF | ‘Uos pe | ‘Us pz yoO4S [BUTSLIO dnoup ““spriqAy jo sjeqoy, “spuqay ZI6T tones SspriqAy L161 pue OL6I 7+") DIBITRIN WoNOYy,? sarwas alyjua fo hanuiungy SIZE INHERITANCE IN DUCKS 133 TABLE 2 Summary of measurements of 1912 ducks in centimeters BILL TARSUS NECK Bison! mean... 5.5 5.24 | 35.6 68 .2 Bree: ro D Aaa 0.27 0.31 1.69 2.53 Niele CAV cs: 4.90 5.91 4.74 3.71 ule! ages 5.5 5.12 35.5 68 .0 Bets cee fool D oeeen 0.29 0.29 2.22 3.67 | GEV re: 5.27 5.66 6.26 5.39 mean... 5.21 5.0 32.9 61 .2 Beh. Seeks | $.D 0.22 0.19 1.31 3.54 aranales C.V 4.22 3.80 3.98 5.78 mean 5.35 5.0 32.3 63 .1 evecce | S.D 0.24 0.19 1.98 2.51 | C.V 4.48 3.80 6.13 3.97 103, for the first three clutches of eggs, with F; male 105 for the last two clutches (see former paper). Female 83 died after the first few eggs had been taken from the pen, so that nearly the entire F, lot was raised from only two mothers. Is is thus evident that as regards the number of the parent individuals in each lot there is fully as great a chance for variability among the F, ducks as among the F, ducks. Both generations ran together in the rearing yard and in the maturing yard in exactly the same way as in 1910-1911. The weights of the birds reared in 1912 are given in tables 4 to 7. In table 1 are given the means and certain other statistical constants for each lot of birds together with the results obtained in the two previous years, thus summarizing the entire experiment. The standard deviation in weight of the 1912 birds is greater than that of the 1910 and 1911 birds. The mean weight of the 1912 birds is somewhat less than that of the 1910-1911 birds, especially the F, birds. It is plain that, in spite of the greater variability of the 1912 F, birds, induced no doubt mostly by their greater number (57 animals instead of 13), there is still a small excess of variability of the F, over that of the F; generation (at least among the males). Finally, the combination of the three years’ work gives a very satisfactory basis for comparison, the F: lot comprising 70 anima!s THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 134 JOHN C. PHILLIPS and the F, lot 64. The results for 1912 are not, however, so striking as those for 1911, but point in the same direction—to a partial segregation of multiple size factors, or to a modification of the gametes brought about by their association in F, zygotes. The increased variability in F; is far greater in the male sex than in the female. As a further index of size differences, external measurements were taken on all the adult 1912 ducks. These were as follows: Bill: exposed culmen. Tarsus: the tibio-tarsal measurement as used by the orni- thologists. Neck: from the anterior point of the breast bone to the tip of the bill. The front end of the breast bone is a point which can readily be felt. This measurement is easier after rigor mortis has passed. Length: tip of bill to end of tail feathers. A summary of these measurements is given in table 2, for each sex and generation. Among the eight pairs of coefficients of variability obtained the F, coefficient is greater than the F; co- efficient in five cases, the same in one case, and less than the F, in two cases. Accordingly greater variability of the F, birds is not very clearly shown in the measurements taken. GROWTH OF DUCKS In a recent paper Goldschmidt gives the result of some size work in ducks. He does not deal with adult weights, but with a coefficient of growth determined by dividing the weight at ten weeks by the first or hatching weight. In this way he investigated the early growth rate of several breeds. He obtained growth figures for one lot of F, Peking female mated with wild mallard male, and records a great divergence, which he attributes partly to segregation. It is to be remarked that Goldschmidt’s wild mallard race reached a weight of 1470 grams, which would sug- gest that it might have been contaminated, for this weight is very high. The average for my original stock of wild mallards is 1073 grams for the males, and 922 grams for the females, while the autumn, or five month’s weights for the offspring were 949 and SIZE INHERITANCE IN DUCKS ao 892 grams respectively. The difficulty of obtaining pure wild mallards in parts of Europe is well brought out by Rogeron in his work ‘‘ Les canards.”’ Goldschmidt lays stress on the importance of environmental factors in the growth rate of ducks. To show the influence of earliness or lateness of hatch on body weight, I have prepared table 3. In this table each of the four groups of 1912 ducks (both sexes and both generations) were split into early and late hatches, and the average age and average adult weight found for each of the eight groups thus obtained. The result is very strik- ing, and shows the advantage of lateness over earliness in every one of the four cases. The growth curves show the same phenomenon. For example in chart 4, curve F,¢1 represents the weights of early hatched birds and shows the greatest weight at 60 days of age but a very low maturity weight at 105 days. On the other hand, curve F,¢25 (late hatched birds) shows a smaller weight at 60 days of age but much greater weight at maturity. The early hatched birds tend to mature quicker and remain smaller. This result only emphasizes the importance of outside conditions as pointed out by Goldschmidt. In charts 2 to 5 growth curves are given for both sexes and both generations of the cross-bred birds. Each curve represents the average growth of one sex of each clutch of eggs. At the point marked P, at 50 to 70 days of age, there is always a depression which corresponds with the removal of the ducklings from the rearing yard to the maturing yard, where the free access to water and greater space tended for a short time to keep their weight down. There is a slight tendency to a depression at the time of moult from the first juvenile to the first adult plumage, which takes place between the ages of 110 and 140 days. In all cases the birds were in full plumage when killed, and the completion of the full plumage is a very good criterion of the time when the adult or autumn weight has been attained. Increase in weight beyond this point seems to be due mostly to an increase of fat, and is generally much more marked among the females than among the PHILLIPS JOHN C. 136 SUBIS /QOT ‘SUIBIS UL JYSIOM oSvIDAV sAep OFT ‘pet uoym ose oyvurrxoidde 61 eung ‘surgoyey jo oyep oyeutxoidde SUIVIS GOP] ‘SUIBIS UI JYSIOM OSvIOAG shep SFI ‘pet ueyM ose oyeurxoidde VG Avy ‘Suryoyey jo oyep oyeurxoidde SUIBIS G/RT ‘SUIBIS UT 4YSIOM VSB1VAB sfep CFL ‘pal[h], wey ose oyvurrxoidde st oung ‘surqoyey jo ayep oyeurxoidde SUIBIS GOQT ‘SUIBIS UT JYSIOM 9SBIOAR shep OFL ‘peTTt{ ueyM ose oyevutxoidde x6 Avy ‘Suryoyey jo oyep oyeurxoidde ‘SoTBU poyoyey 94R] 2 ee cccee vo ole SOTRUR So[BUul peyqoyey Aypareo TT SUIVIDS ZOGT ‘SUIBIS UL JYSIOM VSVIOAT sfep OFT ‘pey[o{ ueyM ode oyeurxoidde 8I aung ‘suryoyey jo oyep oyeurxoidde SUIVID FOP] ‘SUIBIS UT YYSIOM VSvIIAG shep Cpl ‘pelt ueya ode oyeurrxoidde 66 Avy ‘Suryoyey jo oy¥p oyeunxoidde SUIVIS EFT ‘SUIBIS UI YYSIOM OBBIOAB shep OFT ‘peyTry ueyM ode oyeurtxoidde 0c oung ‘Surgozey jo o4ep oyeurxoidde SUIBIS ZEGT ‘SUIBIS UI YYSIOM oSvIOAB sXep OFL ‘pe, wey ose oyevunxoidde LG Avy ‘Suryoyey jo oyep oyeurtxoidde "+ "*s97RUIAF 948] OT “** gayeurey ApIva TT * SoTBUI 948] JT sajeur AyIBa GT NOILVEGNGD 27 NOUWVHENAD LT Buryayoy 970) Lo fizuva pun syonp GIG] fo az1s waamjaq UoYnjay € HTaVL 137 DUCKS SIZE INHERITANCE IN qYySIOM JO SUIBIS PUL 95¥ JO SA¥p UI pozyoTd ‘syoIyo UonoY o[vuUr Jo 9AIND YRMOAY T JABYD ° N 2 ° S fo) a ° o ogt osi ori O€l OL 09 OS Ov oe ord ol Bae ee [| 00z Sees ae. i 00 bE ® /| 009 p uanoy es Griselda 008 Bee asec / 000! BeBe 0021 (ee ae eee = oor! Cee ee ee ae: 7 0091 Esl ose Vecsey eee emer dee of ost Bee eee Ae eee 0002 Or ie ae ea ser et ee ee 00zz co aka lO a al oovz il lat la lla 009% PHILLIPS JOHN C. 138 suryoyey JO op1O oY} OYBOIPUT SOAIND OY} JO WYSII oy 4e s[eBsoUNU OY, “SYONp o[VUI ly JO Spooaq XIs Jo SoAIND YYMOIH Z Avy OBL OLt 091 ost Ort O€1 erat Ol ool 06 OL 09 os Ov oe ex4 ol 0 i = == oot —<— 002 4+ O00€ OOov ——— 00S 009 002 008 = Ba 006 oooL 0021 ooel oovt ——|00G1 zé'y oo9gt === lofovel oo#8l ~—1| 0061 SIZE INHERITANCE IN DUCKS 139 150 160 140 120 No 100 130 90 i : baa eee 70 LL WZ i ie wins 40 Chart 83 Growth curves of six broods of F, female ducks il i LZ ay 2S 0 2 oo o o o o os ge °o (ern Me) sae (oy (2) (oe, [o} (Si te) o (oy te) fo} (over >y 8 & oR te Ma Oh = CO AO eee On iD. ee = N = vn = - - — — - - - a JOHN C. PHILLIPS 2000 1900 1800 1700 1500 1400 1300 1200 1100 1000 3 Chart 4 Growth curves of five broods of female F, ducks SIZE INHERITANCE IN DUCKS 100 SSRs a Ses SSR __SEaSSNSReE SESS NRE SRR EC See ee AE ee a DN. ee | ee NN et eh Jae ae SR aRae i 90 80 70 60 50 40 30 20 10 SOO OR COMO HO FOO) O'S ey! Stet (kad ped ee eeh ve. Col Gol fe) fe) (oy Gla e woh (os bile) ee Ot ans oe Sr oepemon en 2) 8 81'S ss 38 88 3 BR FB ‘ 141 Chart 5 Growth curves of five broods of F2 male ducks "spat [[eur eyeurey pyr oand Jo spoorq 90.14} Jo SoAINO YMOIH 7 WLRYD Spre]yeu oyeur pr oand jo spoosq sory} JO SAAIND YYMOIH 9 JIvYD PHILLIPS JOHN C. | 002 = : } —| ss 008 5 es lar sH WM 2 006 WM © rm _|| { : — ee eee — 0001 0011 — +d ———00z) SIZE INHERITANCE IN DUCKS 143 males. The small figures at the end of the curves refer to the sequence in date of hatching time of the different clutches of eggs. The curve for the Rouen ducks shows the remarkable growth of this race between the ages of 30 and 70 days. The curves for the small race are not available, but I have added two curves for 26 individuals of a pure wild race of mallards in charts 6 and 7. These birds are slightly smaller and much more uniform in size than the ‘English’ mallards. They show a standard deviation of 43.5 for the males and 23.93 for the females, and a coefficient of variability of 4.5 for the males and 2.6 for the females. The males, as among the other races, show a much greater tendency to vary. This fact may be a universal one for ducks; it is in agree- ment with the more uniform c. v. for the females in both F, and F, generations. INHERITANCE OF COLOR IN ROUENS This is mentioned because Goldschmidt (p. 189) expresses some doubt as to the homozygous nature of the color of the Rouens. In my strain at least there is no question about this. In crosses with mallards, 46 F; males were all true to the mallard type of pattern, while 34 F; males gave only two birds which were slightly ‘off color.’ These two birds had large white collars with some white on the breasts and primaries. Goodale showed in crosses between Pekins and Rouens that a very complicated assemblage of color types resulted in the first hybrids. He attributes this to heterozygosis among the Pekin males, which view would appear to be the correct one. THE SEX RATIO OF F, SIZE DUCKS Owing to the present general interest in disturbed sex ratios of hybrid animals it is worth while to call attention to theresults obtained in ducks (table 1).. In both seasons there was a great preponderance of males among the F; offspring though among the F, birds, where the parents were similar in size, the sexes are near- ly equal. Adding the results for both years together, we obtain 144 JOHN C. PHILLIPS the following proportions of the sexes: In F;, males 46, females 24; in F,, males 34, females 30. In the first case is found practically a two i one ratio, which, in view of the large number of observations, is significant. Guyer called attention to an excess of males in various hybrid birds, especially when the parents were from widely separated species or different genera. His results are amply confirmed in a recent paper by Goeffrey Smith and Mrs. Haig-Thomas on pheasants. An excess of males is apparent in fertile hybrids, but seems to be greater in sterile hybrids. I myself have found that in hybrids between Reeves and Torquatus pheasants, Jungle fowl crossed with pheasants, and Swinhoe pheasants crossed with Silver, a great excess of males occurs. _ It is therefore very interesting to find the same result obtaining in a cross between two domestic races of ducks, very dissimilar in size, but derived from the same wild species, and producing fer- tile offspring when intercrossed. It is possible to explain the in- equality of the sexes as due to a selective death rate on the part of the two sexes, a larger proportion of male zygotes surviving, but the supposed selective death rate, if it occurs at all, must be effective very early in embryonic life, for out of 76 eggs set in 1912, 60 hatched, and 57 were reared to maturity, producing 36 males and 21 females. Most of the eggs, which failed to produce young were sterile or at least discarded as such after ten days incubation (exact number not noted) so that a selective fertiliza- tion seems to be a more probable explanation in this case, but an early embryonic selective mortality cannot be ruled out. In conclusion I wish to express my thanks to Professor Castle for help in working out details. SUMMARY 1. The high coefficient of variability of both the parent races, especially the mallards, may perhaps increase the variability in F4. Possibly the mallard race may contain an admixture of the small, so called ‘toy’ race, which would account for some of the small individuals which occasionally appear. In spite of this, however, the male ducks show a very considerably increased variability SIZE INHERITANCE IN DUCKS 145 of weight in F, over their variability in F; and the females show a very small increase. The numbers are now thought large enough to have a significance. 2. If a stock like the pure wild race mentioned above, could be used as a small parent, it is very likely that a much more striking segregation of size might be observed. 3. The present experiment does not throw any further light on the theoretical side of the question except to diminish the possi- bility of the existence of any large and clear cut size units in birds, which would result in an easily recognized numerical ratio. 4. The male sex is much more variable in size than the female. 5. The Rouen race used in this work was homozygous for color (mallard color) when crossed with a mallard race. 6. The growth curves show that a very satisfactory age for studying size of ducks is 140 to 150 days, at the assumption of the first adult plumage, but also point to a marked effect of earli- ness and lateness of hatch on the rate of growth and ultimate size. This difference probably varies greatly from year to year and cannot be regarded as uniform. 7. A disturbed sex ratio occurs among the F hybrids, a result apparently of the difference in size of the parents, for it is not seenin F;. This ratio has resulted in the preponderence of males over females in the proportions of nearly two to one. BIBLIOGRAPHY Goopate, H.D. 1911 Studies on hybrid ducks. Jour. Exp. Zodl., vol. 10, p. 241. GoutpscumipT, R. 1913 Zuchtversuche mit Enten. Zeit. fur Induct. Abstamm- ungs- und Vererbungslehre. Bd. 9, p. 161. Guyer, M. F. 1909 On the sex of hybrid birds. Biol. Bul., vol. 16, p. 193. Puituirs, J.C. 1911 Inheritance of sizein ducks. Jour. Exp. Zodl., v.12, p. 369. Smitu, GEorrry, AND Mrs. Hatc-THomas. 1913 Onsterile and hybrid pheasants. Jour. Genetics, vol. 3, p. 39. 146 JOHN C. PHILLIPS TABLE 4 Weights and measurements of F, male ducks reared in 1912 NO. Ss fae es WEIGHT BILL TARSUS NECK ante | grams cm. : cm. cm. cm. 305 May 23| Sept. 28) 1740 | 5.9 5.5 35.5 | 68.0 306 May 23 | Sept. 28) 1290 5.3 Baik Bis) 61.2 307 May 23 | Sept. 28) 1470 5.6 5.3 Boo 67.3 308 May 23} Sept. 28) 1830 5.8 Deo Syl 3} 67 .1 310 May 23} Sept. 28) 1100 4.9 4.6 28 .7 58.5 312 May 23! Sept. 26) 1760 6.1 6.5 38.5 72.0 313 May 30 Nov. 6 1690 5.9 4.5 ay 70.5 314 May 30| Nov. 6 1700 Dn5 R23) 37 1 69.9 318 May 30! Nov. 6 1640 5.6 5.3 36.0 69.8 319 May 30; Oct. 14 1600 5.3 5 (0) 35.8 68 .2 320 May 30) Oct. 26 1720 ye! 5.4 36.3 69.0 321 May 30} Oct. 26 1500 5.9 5.4 36.0 66.6 323 May 30 | Oct. 26 1590 5.6 OES 35.0 67 .0 325 May 30) Oct. 14 1500 OES 5.15 35.0 66.5 326 May 30| Oct.14| 1600 | 5.5 5.15 | 36.8 | 69.0 344 May 30| Oct. 14 1600 5.8 5.2 36.8 70.0 346 May 30/ Oct. 26 1460 oe 4.9 30d 66.8 347 May 30| Nov. 6 1770 5.4 5.1 35.5 68.5 348 May 30} Nov. 6 1550 IY G51 Biel 66 .2 351 May 30} Nov. 6 1680 5.4 5.3 36 .2 69.0 354 June 15 Oct. 26 1655 3 573 36.1 69 .2 358 June 15) Nov. 6 1860 6.1 yey 35.0 69.0 360 June 15) Nov. 6 1890 5.8 ez yf 70.4 361 June 15} Oct. 26 1570 5.4 5io, |) soto 70.5 365 June 15) Oct. 26 2130 5.6 5.3 36.5 70.8 370 June 15! Nov. 15 1700 eres 5.4 33 .6 66.8 371 June 15| Nov. 6 1560 5.4 520 37.8 69.7 373 June 15) Nov. 6 1750 5.4 oe 35.0 68 .0 375 June 15| Nov. 15 1700 5.4 5.2 35.5 68 .6 378 June 25| Nov. 15 1750 ee Dee 34.5 68.3 379 June 25| Nov. 15 1630 5.9 5.4 36.2 69.8 381 June 25| Nov. 15 1800 Dien yet 35.4 68.8 383 June 25} Oct. 14 1700 oe ea 35).5 67 .0 385 June 25 | Nov. 15 1800 5.8 5.5 36.2 69.9 387 June 25 | Nov. 6 1700 5) 8° onl 35.0 67.5 390 June 25 | Nov. 6 ileérA 5.9 5.2 36.4 69.7 TABLE 5 Weights and measurements of F, female ducks reared in 1912 No. capricorn | aac | WEIGHT BILL TARSUS NECK ae grams | cm. cm | cm. | cm 309 | May 23 | Sept. 28 1200 | 5.4 5.0 33.0 62.9 315 May 30| Oct.14| 1500 | 5.1 5.0 31.6 64.0 317 May 30 | Oct. 26 1560 5.1 5.2 35.8 65.0 324 May 30| Oct. 26) 1385 5.3 5.1 32.2 63.9 327 May 30| Nov. 6 1580 Sez 4.9 32.3 63 .4 330 | May 30| Nov.6/| 1420 56. |. 50 34.2 66.0 335 | May 30] Nov. 26; 1450 | 53 | 4.8 33.1 63.8 338 May 30| Nov. 26] 1230 | 4.9 4.7 31.5 61.8 340 May 30| Oct.14) 1500 | 5.15 4.9 34.0 62.0 342 May 30) Oct.14) 1330 5100) pares 32.0 63.0 343 May 30/ Oct.14) 1290 A828 Wi AGS Soe 63.0 350 June 15 | Oct.14| 1390 5.1 4.9 32.4 63.6 359 June 15) Oct.15} 1700 5.6 5.4 33.5 66.0 364 June 15| Oct.15| 1600 5.3 5.1 32.5 63.8 367 June 15| Oct. 26) 1530 4.9 4.9 30.8 62.3 369 June 15| Oct. 26; 1380 FB} 5.0 34.2 64.4 372 June 15} Nov. 6 | — 1900 5.7 55 3Y/ 70.7 374 June 15! Nov.6 | 1540 5.2 5.1 32.5 64.0 386 June 25 | Nov. 6 1450 Sell 5.0 33 .6 65.0 388 June 25! Oct. 26! 1610 5.3 5.2 3207 65.7 391 June 25} Nov. 15) 1520 5.1 5.1 32.2 64.8 * TABLE 6 Weights and measurements of F. male ducks reared in 1912 DATE OF DATE OF | | | TOTAL NO. HATCHING | KILLING WEIGHT | BILL TARSUS NECE LENGTH | } grams cm cm. cm. cm. 247 May17/ Oct.14| 1330 5.25 4.7 35.0 66.0 288 May 17| Sept.28| 1360 5.85 5.1 35.5 66.3 311 May 30 Oct.26 | 2180 5.6 5.0 38.0 VORS 316 May 30| Oct.13 | 1700°| 5.65 Bray i ae.0 69.0 328 May 30 Oct.26 | 1680 5.4 5 oeerhn 235-0 69.8 329 May30)/ Oct.13 1490 5.1 5.1 34.6 65.0 331 May30| Nov.6 | 1645 5.5 5.2 36.0 69.0 332 May30| Nov.15| 1945 6.0 5.6 38.0 72.0 336 May 30| Oct.13 | 1670 5.5 4.9 36.0 67.0 341 May 30| Oct.26 1100 4.8 4.5 29.7 58 .4 349 May 30} Nov.15| 1600 5.8 4.5 36.7 69.5 352 Junel5|} Nov.15} 1480 5.2 5.2 33.0 65.0 356 June15 |} Nov.6 | 1870 5.5 5.5 35.5 69.0 362 June15 | Oct. 26 1890 5.4 oa! 35.3 69 .2 368 Junel5 | Nov.15| 1850 5.2 5.2 32.5 66.5 377 June25| Nov.15| 1750 5.6 5.3 34.8 68 .0° 389 June 25 | Noy.15} 2400 5.8 5%. | 938.0 75.0 147 148 JOHN C. PHILLIPS TABLE 7 Weights and measurements of F. female ducks reared in 1912 NO. S ea pa ed WEIGHT BILL TARSUS NECK ee grams cm. cm. cm. cm. 283 May 17 | Sept.28 1330 4.7 4.9 29 .4 58 .0 289 May 17 | Sept.28 1390 ‘ial 4.7 29 .6 59.5 296 May 17 | Sept.28 1450 5.25 ome 32.0 62.0 333 May 30 Oct.26 1520 yar 4.9 34.4 66.0 337 May 30 | Nov.15 14380 oo Ry 34.0 65 .0 339 May 30 | Oct.26 1415 5.6 5.0 32.5 64.0 345 May 30| Nov.15 1300 yall —48 30.0 60.0 353 June 15| Nov.15 1600 5.4 4.9 30.8 62.5 357 June 15) Oct.26 1600 5.8 Sell 34.2 66 .4 363 June 15| Nov.6 1600 Divo 4.9 31.0 63.8 366 June 15} Oct.15 1640 5.4 3} 31.2 65 .0 376 June 25} Oct.15 1800 5.4 Deo 30.0 62.5 380 June 25} Oct.15 1770 5.0 5.0 29 .9 62.8 382 June 25} Oct.15 1800 5.6 5.4 33.8 66.5 . TABLE 8 Weights in grams at five months of age of Rouen and mallard ducks reared in 1912 ROUEN MALES ROUEN FEMALES MALLARD MALES MALLARD FEMALES 2440 2290 950 830 2570 2320 1080 940 2680 2340 1120 970 2750 2340 1150 980 2360 1150 1020 1290 1045 1300 1060 1100 1110 1160 1200 1240 1310 EXPERIMENTAL EVIDENCE CONCERNING THE DETERMINATION OF POSTURE OF THE MEMBRANOUS LABYRINTH IN AMPHIBIAN EMBRYOS GEORGE L. STREETER Department of Anatomy, University of Michigan THIRTY-EIGHT FIGURES The purpose of the present paper is to report a series of experi- ments that substantiate the conclusion that the posture of the membranous labyrinth and the position of its canals is determined by some force or influence that interacts between the labyrinth and its environment. The exact nature of this control or in- fluence has not yet been determined, but from the experiments that are to be described it is evident that it possesses a force capa- ble of producing a complete rotation of an embryonic vesicle that has been displaced in a reversed position. The existence of a postural influence of this kind introduces a new factor in organogenesis. It means that organs do not develop inertly in the position that they happen to find themselves; on the contrary, there is a certain amount of adjustment of position through forces interacting between them. According to the conception of what we might call the theory of passive organ development, all organs develop passively in the position in which they are located at the outset. Perfect form according to this theory is eventually obtained because originally the anlages of the various organs are so perfectly placed, and all the stresses so care- fully calculated, that the subsequent increase in size, and mutual pressure against each other, produce a final normal disposition of all of them. In contrast to this we may now speak of self-place- ment of organs or individuality of organ development, according to 149 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 150 GEORGE L. STREETER which there is, in addition to the passive pressures of adjacent structures, a motive element, the force of placement, which helps to maintain, and corrects if necessary, the relative position of an individual organ. This force of placement may be defined as the product of the combined motive forces interacting between an organ and its environment. We are already familiar with the striking phenomenon of muscie migration and the movement of nerve-ganglion masses. We can now add to this the potential movement and adjustment of entire organs during their course of development, and it is not necessary to add that this autostatic tendency is doubtless much more pronounced in some organs than in others. The writer’s attention was first directed to the possibility of experimentally altering the position of the developing labyrinth, during some investigations on equilibration in amphibian larvae (Streeter 06). In these experiments it was found that the cells constituting the ear vesicle are specialised very early, and though transplanted to an abnormal environment they continue to differentiate themselves in the usual way into a recognizable labyrinth. Lewis (’07) showed in a subsequent paper that this was true while they were still in the stage of an uninvaginated plate, quite in contrast to the surrounding cells forming the car- tilaginous capsule, which Lewis clearly proved were not prede- termined in this way. In a later paper (Streeter ’07) additional evidence was given of the high degree of developmental independence possessed by the early labyrinth cells. It was shown that even fragments of the vesicle may develop independently of the rest of the vesicle, and any individual part, for example the endolymphatic append- age, may be quite normal in cases where the remainder of the labyrinth is very abnormal. It was also shown (Streeter ’09) that when two primitive vesicles are crowded together into the same pocket they do not fuse and form together one large laby- rinth, but remain as two distinct labyrinths. Moreover, it was shown that this developmental independence of the vesicles extends to the difference existing between a right and left-sided organ. The dextral or sinistral character, or laterality of the ear POSTURE OF MEMBRANOUS LABYRINTH ph vesicle, is not controlled by the environment but is determined by the intrinsic character of its own constituent cells. A left-sided ear vesicle when transplanted to the right side develops into a labyrinth having all the characteristics of a left-sided organ; the anterior canal is formed on the caudal border of the labyrinth, the posterior canal on the oral border, and the lagena, which normally is directed caudalward, is found extending forward toward the eye. It was found, however, that the ear vesicle, though capable of this marked power of self-differentiation, apparently was not in all respects independent of the surrounding structures. The posture of the developed labyrinth, the situation of its canals and various chambers, seemed to be controlled by the environment. Deliberate rotation of the ear vesicle into abnormal positions and even transplantation to the opposite side of the body, always resulted in a labyrinth possessing a normal attitude with reference to the brain, ganglion masses, and the surface of the body. In. all of seventeen experiments performed for this purpose the results were positive, and the writer consequently came to the conclusion that there is some influence interacting between the ear vesicle and its environment that constitutes the chief factor in the deter- mination of its placement. At the same time an experienced foreign investigator had been working independently upon a similar problem (Spemann ’06 a and ’06 b). His experiments consisted in removing the ear vesi- cle in young Rana esculenta larvae and then replacing it in an in- verted position for the purpose of studying the consequent abnor- mal body movements and their correlation with the anatomical results of the operation. He also studied the influence of the epi- thelial labyrinth upon the surrounding connective tissue, and its relation to the formation of the cartilaginous and bony labyrinth. Concerning this latter problem he did not come to a definite con- clusion. The former problem, however, gave positive results. He found that the inverted vesicle continued in its development and formed a more or less complete labyrinth, but, unlike my speci- mens, the orientation of the developed labyrinth corresponded to the operative displacement of the vesicle. There had been no subsequent readjustment of position. After the appearance of 152 GEORGE L. STREETER my papers (Streeter ’06 and ’07) in which I reported the experi- ments in which, after various displacements, the labyrinths all recovered their normal postural relations, Spemann (’10) made further experiments and reported more at length with regard to postural development. His results confirmed his earlier obser- vations, and his paper contained illustrations showing sections of larvae of different ages in which the inverted ear vesicle has remained inverted though otherwise practically normal. Out of twelve operations nine labyrinths remained inverted, two re- covered partly, and one recovered perfectly the normal position. Regarding the latter specimen he surmised that it had not been sufficiently turned, and had slipped back into its normal position. The results, therefore, as obtained by Spemann are directly con- tradictory to those obtained by me. On account of the great embryological significance of the question of environmental pos- tural control, further experiments were plainly warranted both for the purpose of testing my own previous results and if pos- sible for finding some explanation of the discrepancy existing be- tween the observations of Spemann and myself. From the study of these new experiments and from an examination of the accom- panying photographs it will be plainly apparent that they confirm my previous results. Before entering into a further discussion of them we may proceed with a description of the details of the experiments themselves. METHOD OF OPERATION Obviously the conditions of the operation should be planned so that the environmental control would be put to an extreme test. In most of my previous experiments, and in all of Spemann’s, the vesicle operated upon, after being taken out and rotated one way or another, was replaced in its own pocket. Recovery of the original posture after such an operation might be explained by the possibility of the vesicle not having been completely detached; some small strand of tissue might suffice to draw it back into the original position. Also it is conceivable that, since the vesicle and any adherent mesodermal fragment exactly fitted the pocket LS es Se POSTURE OF MEMBRANOUS LABYRINTH 153 from which they were taken, they might on that account slip back into their original position. Both of these objections were avoided by transplanting a vesicle from one specimen into the emptied ear pocket of another specimen. This certainly involved complete detachment, and it is obvious that the transplanted ear vesicle would not exactly fit in a pocket in which it did not originally belong. All of the following experiments were done in that way. The experiments were carried out on larvae of Rana pipiens, and the operating stage was the same as that used in previous experi- ments (Streeter ’06, fig. 3, p. 547). This is just at the end of the non-motile stage, and the outer form of the larvae shows a distinct tail bud, and the eminences on the head caused by the optic cup and head ganglia. The ear consists of an invaginated saucer- shaped mass of cells Just in the process of being pinched off from the deeper layer of the skin, with the edges inverted and in the process of completing the closure of the vesicle; that is, transition from auditory cup to auditory vesicle. In this, as in previous papers, the attempt is not made to distinguish between these two stages, and the term ‘ear vesicle’ is used to cover both. The technique of the operation is as follows: Two larvae are removed from their gelatinous capsules and placed side by side in distilled water, under a binocular microscope. With two no. 12 embroid- ery needles, a linear tranverse incision is made through the ecto- derm of one of the specimens over the site of the ear vesicle, the incision being about three times as long as the vesicle, and the right side of the animal always being used for the operation. The lips of the wound are than gently everted, forward and backward respectively, which discloses the thin lateral wall of the vesicle, or if this is torn in the removal of the skin then one sees the pigmented concave interior of the vesicle. With the needles the vesicle is now loosened from its pocket and cast away. This leaves an empty pocket, free for the transplantation. The right vesicle of the second specimen is uncovered in a similar way, loosened from its pocket and then slipped into the empty right auditory pocket of the first specimen. In tucking in the vesicle under the edges of the wound care was taken in all cases to place the vesicle so that the lateral or concave open surface of 154 GEORGE L. STREETER the vesicle should lie against the brain, and the median or convex surface toward the opening of the wound. The lateral and median surfaces of the vesicle are easily recognized and therefore this transplantation in a reversed posture can be done with great accuracy. This method of placing the vesicle has a double ad- vantage. It furnishes a complete rotation of 180 degrees and it is a secure way to transplant the vesicle which is very easily wedged in the pocket in that posture. The vesicle shows no tendency to escape, and as the lips of the wound are not very wide apart, it is not necessary to take any precautions about holding them to- gether. They take care of themselves and need no superimposed weights or further attention. The specimen is now set aside and in the course of three or four hours all traces of the wound have disappeared. The specimens are allowed to go on with their development for fourteen days, the period usually necessary for the formation of the canals, at the end of which time they are preserved in a chrome-acetic mixture and are ready for examina- tion. In most of the operations two specimens were utilized, one always being discarded after transplanting its right vesicle. In four experiments the writer succeeded in making a complete exchange of the right vesicle between the two specimens and pre- serving both of them. This involves great care and the addition- al effort required does not justify the procedure. The preserved specimens were imbedded in paraffin, cut in series and stained with hematoxylin and congo-red. Examination of them showed that nineteen of the experiments were successful, in that the transplanted vesicle had continued in its development far enough to recognize its different parts and their posture. Wax-plate models after the Born method were made of them all, and draw- ings of these, together with a photograph of a selected section from each series, are shown in figures 1 to 38. I may here state that these photographs were made with the apparatus belonging to my colleague, Professor Novy, and it gives me pleasure to take advantage of this opportunity to acknowledge his courtesy and his assistance in their preparation. Other operations were performed with a view to the determina- tion of characterand degree of the environmental control, but these POSTURE OF MEMBRANOUS LABYRINTH 155 will not be reported on in this paper; we limit ourselves to the question of whether environmental control does or does not exist. It may be pointed out that the above operative procedure offers a complete and severe test for the existence of such control. It is so arranged that we have for the experiment a transplanted foreign ear vesicle, which does not naturally fit in its new pocket and which is intentionally placed in a posture as abnormal as possible. RESULTS OF THE EXPERIMENTS At the outset it may be stated that in all cases where the laby- rinth had developed into a structure with sufficient completeness for the identification of its relations, it was found that the position of its canals and various chambers as regards the surrounding structures was practically normal, in spite of the manipulation it had undergone at the time of the operation. As has previously been shown (Streeter ’07, p. 483) the posture of the labyrinth can be determined both by the histological structure of its walls and by its outer form, as determined by wax-plate reproductions. The description of these morphological characteristics will not be repeated here; it will be sufficient to state that they are so definite that the various parts of the labyrinth can be recognized without difficulty, even though they happen to be incomplete or unequally developed. For the sake of compactness table 1 is annexed in which are detailed the separate features of each of the nineteen transplanted labyrinths as found on examination. Each feature is marked ‘normal’, ‘imperfect,’ or ‘absent,’ indicated by the signs N, I and 0. These are used in a liberal sense so, that ‘normal’ signi- fies practically normal, and includes structures abnormally large or small; ‘imperfect’ signifies quite abnormal, the characteris- tics, however, being sufficiently well defined for identification of the structure; and ‘absent’ means either completely absent or unrecognizable. The features as listed will for the most part explain themselves. It may be mentioned, however, that by canal planes is meant the relative position of the canals as regards each other. Planes 156 GEORGE L. STREETER drawn through the anterior and posterior canals would meet at an angle of about 95 degrees at the crus commune. This is quite constant in all of the labyrinths. The lateral canal is about perpendicular to the other two. At the end of two weeks the planes are not so well defined as they are later, but when the canals are otherwise normal the planes can be determined, even as early as that. From a study of this table it will be seen that we can arrange our experiments in three groups, based on the completeness of the TABLE 1 LABYRINTH hte [08 Woe ges Ste BS. | 2 /| :onesiem NO. g o|° = ey it) aS =a o ° of : Be) 8 = Silvia =. | BB) CE) ge ie | a a < n 2 a oe D pe | ey Sele |e] Sage Beloe |S. | 230) seem & Dn < a =< ° o | > =| a < < ee Normal group T|ININININ]) Ne] N | NUN | N | NO) Ne eee IEININININ| N | WON |} N. | N | NN She Ii |NIN|NIN| N | Ne] NN | NS] NN ONS IVIN|ININ|IN|) N| N |] I N N Ona N N V1) NUN OE ON aN ON | ON > Ns I VIININ|IN|IN| N | N | N N N N N | oN N VII|N|N|N|N/ I N | N | N-| 0 | N JON] VIIL|N|N|N|I| N | N TN | N 1) NSN ee Abnormal group IX |N|IN|N/N| I N Neu st O N N N N X | AUN AN NA Mey NN | oi O)| No) eas XI|N/I|N|N| I NN} Neal"sO | ay I I N XII/N|N|N|N/ I IN) l= aN I Oo ON Ne ia N XIE |N{}T|N|T| N | N°] N |°N | OF -05) N “Sieg SOV SANG SIIB Ney by IN | aL oe 0. ]°O 1 Ree XV EN EN N I |-N |.0 |) N | NS SVT [NGL NG NG ol! os, I Nl) ONG | al 1 N | N Doubtful group xvit |nNin|N|[r| 1 | N r{ i/o o | 1 rae VI ea eee ie 2? | @ | ON =) hea 1}0|0| 2 AP ee te NOOM Cm) oT S| 2 POSTURE OF MEMBRANOUS LABYRINTH 157 resultant development of the transplanted labyrinth. These are: firstly, those in which the labyrinth is nearly normal; secondly, those in which the labyrinth is abnormal though the main parts of it are easily recognizable; and thirdly, those in which there might be some question as to the identification of the differ- ent parts of the labyrinth. There are eight in the first group, eight in the second group, and three in the last group. Based on their order in these groups we may now proceed to the exam- ination of the individual labyrinths. | Labyrinth I: (figures 1 and 2). The photograph shown in figure 1 is taken from Series X 2 C, slide 5, row 3, section 7. The study of this labyrinth shows only one slight departure from the normal, and that concerns the Saccus endolymphaticus which lies some- what cephalad to the crus commune. It is on this account that it does not show in figure 1, which is directly through the crus. What looks as though it might be the edge of the sac, is the pig- ment layer of the skin. Aside from its slight anterior displace- ment it bears the usual relation to the chorioidal roof of the fourth ventricle and is normal in size. Labyrinth IT: (figures3 and 4). Figure 3 is taken from Series X 2 A, slide 4, row 1, section 8. Here as in Labyrinth I the endo- lymphatic sac is shghtly in front of the crus commune. The crus is a little wider than in the former, so both it and the sae can be seen in figure 3. This labyrinth shows no noteworthy departure from the normal in either its histology or general form, aside from the slight displacement of the endolymphatic sac. In the draw- ing of figure 4 the groove between the ampullae of the anterior and lateral canals has been exaggerated a little more than is warranted by the reconstruction. In figure 3 the thickening of the ventro-lateral labyrinth wall is partly due to the crista of the lateral canal which it represents, and partly is due to the obliquity of the section of the ampulla. The characteristic relation of the endolymphatic sac to the chorioidal roof of the ventricle can be seen. The large ganglion on the ventro-median wall of the laby- rinth is the proétic ganglion; the acoustic ganglion lies in a simi- lar position but is found in the more caudal sections. Some of 158 GEORGE L. STREETER the acoustic root fibers, however, can be seen entering the side wall of the brain near the endolymphatic duct. Labyrinth IIT: (figures 5 and 6). Figure 5 is taken from Series X 7 A, slide 4, row 2, section 4. In the reconstruction of this labyrinth there is apparently nothing abnormal. And on exam- ination of the section the only departure from the normal is a slight. deficiency in the cartilaginous capsule in the region of the lateral center. The parabasal plate seems normal. The gan- glion shown in figure 5 is the proétic near its junction with the acoustic ganglion. The section passes through the endolym- phatic sac, the crus commune and the lateral semicircular canal. It shows the intimate relation existing between the endolym- phatic sae with the chorioidal membrane. Labyrinth IV: (figures 7 and 8). Figure 7 is taken from Series X 11 B, slide 3, row 1, section 5. This section was selected be- cause it shows the characteristic protrusion of the lagena on the ventromedian wall. The section is toward the caudal end of the labyrinth and passes through, besides the lagena and large vesti- bule, the posterior canal and the bulging caudal edge of the lateral canal. Sections a little in front of this show the lateral canal opening into the vestibule. Evidently in the process of trans- planting this vesicle some injury was inflicted on the cells that were to form the middle part of the median surface. The result- ant defect includes the absence of the endolymphatic appendage and a marked imperfection in the crus commune. Otherwise the general form of the labyrinth is quite perfect, as can be seen in figure 8. Labyrinth V: (figures 9 and 10). Figure 9 is taken from Series X 1B, slide 2, row 3, section 5. The section passes through the endolymphatic sac, the short crus commune, and the lateral eanal. The lateral vestibular wall is thin and distended and represents the dropsical type, which is a common deformity. There is a large acoustic ganglion connected centrally with the brain. This section shows well the relation between the endo- lymphatie sac and the chorioidal membrane. In figure 10 the shading of the lateral vestibular wall makes it look thinner than | shown in the model and does not give the swollen appearance it POSTURE OF MEMBRANOUS LABYRINTH 159 Sac. endol. ' C. sc. post. \ ‘ ' ’ A ‘ Lagena -* Sac. C. sc. post. Lagena / EwiGeiscai lat 6 Sac. endol., C. sc. post. C. sc. ant. 1 i 1 1 1 “.C. se. lat. 8 -C. sc. ant, Crus commune (imperfect) eae ' wnF C. sc. post. C. sc. lat. 160 GEORGE L. STREETER ought to have. The lateral wall in reality presses directly against the lateral canal. Aside from this dropsical tendency the general form of the labyrinth is quite normal. There is some deficiency in the cartilaginous capsule, particularly in the lateral center. Labyrinth VI: (figures 11 and 12). Figure 11 is taken from Series X 4 A, slide 7, row 1, section 4. This section passes through the endolymphatic sac, crus commune, vestibule with its thickened macula acustica, and the lateral canal. The endolymphatic sac comes in contact with the chorioidal membrane in sections oral to this. Both histologically and in its general form this labyrinth is quite perfect. Figure 11 is the only photograph in the series that was retouched in any way. Here the negative was scratched to show the outlines of the lateral canal and lateral wall of the vestibule more distinctly. Labyrinth VII: (figures 13 and 14). Figure 13 is taken from Series X 7 B, slide 6, row 3, section 3. The section passes through the endolymphatic sac, the crus commune, and the combined vestibule and lateral canal. The lateral canal exists only as a lateral pouch from the general vestibular cavity. A partial separation is indicated by an indentation on its dorsal surface, as can be seen in figure 14. The indentation does not completely perforate the pouch, so the canal is incomplete. There seems to be a general defect of the caudal portion of the labyrinth. So that, in addition to the imperfect lateral canal, the lagena is absent and the posterior canal small though otherwise complete. The acoustic ganglion is large (figure 13) and has a well developed root connecting it with the brain. The endolymphatic sac pre- sents the usual relation to the chorioidal membrane. Labyrinth VIII: (figures 15 and 16). The photograph shown in figure 15 is taken from Series X 2 B, slide 5, row 1, section 2. In the process of embedding and mounting this series, some of the sections were injured, though not enough to interfere with the identification of the different parts. Thus it can be plainly seen that figure 15 passes through the endolymphatic appendage, the crus commune, the vestibular pouch, and the lateral canal. The anterior half of the labyrinth is quite perfect, as can be seen in POSTURE OF MEMBRANOUS LABYRINTH 161 Sac. endol. C. sc. ant. Sac. endol. \ C. sc. post. ae Sac. endol. Ps C. sc. post. a (imperfect) ~—~““="C. sc. lat? 162 GEORGE L. STREETER figure 16. The posterior region is defective in that the posterior canal and crus commune consist of a common pouch, taking how- ever the usual shape of the posterior canal. ‘The lagena devel- oped as a pair of short pockets, one extending medially and the other caudally. Apparently the anlage was divided. The endolymphatic sae (figure 15) presents the usual relation to the chorioidal membrane. Labyrinth IX: (figures 17 and 18). The photograph shown in figure 17 is taken from Series X 8 A, slide 2, row 4, section 9. It passes through the endolymphatic sac, the crus commune, and the vestibule with the lateral canal opening into it. In thesections oral to this the lateral canal becomes completely separated, though it is small and is deficient in the region of the ampulla. The anterior canal is small but well formed. The posterior canal is completely formed but les closely against the vestibular wall. The vestibule is small throughout and lacks the lagena. The acoustic ganglion is well developed and is connected with the brain in the normal way. The endolymphatic sac (fig. 17) bears the usual relation to the chorioidal membrane. This is the first of what we have grouped as abnormal labyrinths. Labyrinth X: (figures 19 and 20). The photograph shown in figure 19 is taken from Series X 11 C, slide 1, row 2, section 8. In this labyrinth the injury involves the endolymphatic appendage and the anterior canal. The latter exists as a blind pouch ex- tending orally from the crus commune. This should be compared with figures 26 and 28, where a similar defect involves the posterior canal. The remainder of the labyrinth is quite perfect. Figure 19 shows the crus commune, the vestibule, with a portion of the acoustic ganglion, and the lateral canal. The endolymphatic appendage is entirely absent. Labyrinth XT: (figures 21 and 22). The photograph shown in figure 21 is taken from Series X 8 B, slide 2, row 2, section 8. Here there is considerable abnormality in the ventral half of the labyrinth, and the labyrinth is correspondingly reduced in size. The anterior and posterior canals and the endolymphatic sac are fairly normal. In figure 21 can be seen the endolymphatic sac with its usual relation to the chorioidal membrane. The POSTURE OF MEMBRANOUS LABYRINTH 163 Sac. endol, C. sc. post. (imperfect) \ Lagena-’ Sac. endol. ' ' au ae 4 C. sc. post. \ ee \C. se. ant. C. sce. lat. (imperfect) 24 Sac. endol. , C. sc. post., \C. se. lat. (imperfect) 164 GEORGE L. STREETER section also passes through the anterior canal just oral to the erus commune and through the vestibular pouch. The latter is much reduced in size. It possesses a macula with a diminutive nerve and ganglion, but thé lagena is absent. The lateral canal exists as a lateral pouch projecting from the vestibule (fig. 22). In spite of the abnormal shape due to the deficient vestibule the planes of the anterior and posterior canals intersect at the usual angle. Labyrinth XII: (figures 23 and 24). The photograph shown in figure 23 is taken from Series X 7 C, slide 2, row 3, section 8. The section passes through the endolymphatic sac, the crus commune, the vestibule with the large pouch extending laterally from it and representing the lateral canal. The endolymphatic sac bears the usual relation to the chorioidal membrane. The anterior and posterior canals are fairly normal. The vestibule is enlarged and irregular, and illustrates the dropsical type. It possesses the usual macula with nerve and ganglion (fig. 23). The lagena is absent. Labyrinth XIII: (figures 25 and 26). The photograph shown in figure 25 is taken from Series X 4 C, slide 1, row 3, section 5. It passes through the crus commune, the vestibule with its thickened floor, and through the lateral canal just being pinched off from the vestibule. Slightly oral to this the separation becomes com- plete. On examination of the reconstruction of this labyrinth (fig. 26) it can be seen that the chief injury involves the dorsal part of the labyrinth, and as a result the whole labyrinth is under- sized. The greatest defect is in the posterior canal, which is represented by a bud extending caudally from the crus commune. A similar one is seen in figure 28 and the same kind of defect in an anterior canal has already been seen in figure 20. The crus commune and the anterior canal are small though complete. The involvment of the median surface is shown by the absence of the endolymphatic appendage and the lagena. Labyrinth XIV: (figures 27 and 28). The photograph shown in figure 27 is taken from Series X 6, slide 1, row 4, section 2. This labyrinth is of the dropsical or vesicular type. It is possi- ble to recognize its general position but aside from the fairly POSTURE OF MEMBRANOUS LABYRINTH 165 26 Crus commune - C. x. post. \ (imperfect) ~>~ | <= 28 Crus commune C. sc. post. ! €. sc. ant. (imperfect)! ---m x . sc. lat. 30, C. se. ant. : (imperfect) Sac. endol. , C. sc. post. , i Gease-mlars Sac. endol 32 / Cs sc. ant. C. sc. -post Ate << (imperfect) (imperfect) », THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 1 166 GEORGE L. STREETER normal lateral canal it is poorly differentiated (fig. 28). In this kind of a labyrinth after undergoing differentiation for a short time the process ceases. After that the only further sign of development is an inciease in size, which thus produces the dropsical appearance, such as is seen in figure 27. The section is cut obliquely through the oral end of the lateral canal and the anterior canal. This section was chosen because it shows the macula and its nerve ganglion connection with the brain. The anterior canal, like the crus commune, is only slightly separated from the general vestibular cavity. The posterior canal is rep- resented by a bud extending caudally from the region of the crus commune, such as we have already seen in figures 20 and 26. There is no definite sign of a lagena or endolymphatic appendage. Labyrinth XV: (figures 29 and 30). The photograph shown in figure 29 is taken from Series X 1 A, slide 6, row 3, section 8. It passes through the oral end of the labyrinth, showing the endo- lymphatic sac, the small anterior canal, and the oral end of the lateral canal opening widely into the vestibule. As can be seen in figure 30, the caudal part of this labyrinth is fairly normal. The oral end is quite imperfect, the parts being small and not properly differentiated. The anterior canal is only a pouch not completely separated from the vestibule. The lateral canal has a deep impression on its dorsal surface marking the beginning of a separation. This labyrinth has not completely recovered its normal posture. It is displaced forwards and is tilted partially forward so that the endolymphatic sac, instead of bearing the usual relation to the chorioidal membrane, is opposite the mid- brain and lies embedded in the prodtie ganglion. Though con- sidering its abnormal shape it is not far from having the normal posture. Labyrinth XVI: (figures 31 and 32). The photograph shown in figure 31 is taken from Series X 4 B, slide 2, row 2, section 4. The section passes through the endolymphatic sac, which in the sections a little more caudal touches the chorioidal membrane in the characteristic way. The section also passes through the crus commune and the vestibule with the lateral canal opening into it. The lateral canal is quite perfect, but the anterior and poste- POSTURE OF MEMBRANOUS LABYRINTH 167 rior canals are only represented by pouches opening out of the ves- tibule in common with the crus commune. The shape of the pouches is relatively normal, and they are only deficient in not being pinched off from the vestibule. The lagena is quite normal. The acoustic nerve and ganglion are diminutive. Labyrinth XVIT: (figures 33 and 34). The photograph shown in figure 33 is taken from Series X 3 A, slide 2, row 3, section 13. Due to some mechanical or other injury the last three laby- rinths in our series are quite imperfect, and from examination of the sections alone (figs. 33, 35 and 37) one would be quite unable to orient them. The reconstructions are of great assis- tance in this respect. The section taken from Labyrinth XVIT was selected because it shows the ‘labyrinth fragment’ which evidently was a portion of the main vesicle that became detached and developed as a separate little sac (fig. 34, 7). In figure 33 it can be seen outside the cartilaginous capsule, partially covered by a cartilaginous coat of its own. It is not connected either by nerve or duct with the main labyrinth. It is possible that in the course of the operation a fragment from the native vesicle was left in the operation-pocket, and eventually formed this structure. Examination of the reconstruction shows that there is an inter- ruption in the lateral canal, its two ends being sealed off. One never finds an open end in these imperfect labyrinths, they are always sealed off. The anterior canal is small but well formed. The posterior canal with the crus commune from a common pouch communicating with the vestibule. The vestibule has a normal macula with diminutive nerve and ganglion. There is no lagena and no endolymphatic appendage. Labyrinth XVITT: (figures 35 and 36). The photograph shown in figure 35 was taken from Series X 3 B, slide 2, row 1, section 11. The whole oral region through which this section passes is of the dropsical type. The caudal region is not so pathological. The posterior canal, though incompletely separated from the vestibule and the crus commune, is otherwise well formed. The caudal half of the lateral canal is fairly normal, but the oral portion is distended like the rest of this portion of the labyrinth. Ventrally in one region this large labyrinth bulges through the parabasal 168 GEORGE L. STREETER C. sc. ant, ! Crus commune C. sc. post, ' ' ' ! ‘ i 5 1 r) oa ' L- eee. Rear C. sc. post. “ost (imperfect) i Crscrlats 38 C. sc. lat. 7 \ wf----—-— . | 7] 7 -- “in ay } ee } c plate and protrudes in the pharynx. On the lateral surface of the anterior canal there is a deep pit indicating the usual site at which it is pinched off from the vestibule and crus commune, otherwise the separation is lacking. The presence of a lagena is questionable, and the endolymphatic appendage is entirely lack- ing. This, like Labyrinth XVII, through quite abnormal, can be fairly definitely oriented. Labyrinth XIX: (figures 37 and 38). The photograph shown in figure 37 was taken from Series X 11 A, slide 2, row 2, section 2. POSTURE OF MEMBRANOUS LABYRINTH 169 This is the last and most imperfect labyrinth in our series. It must be regarded as only a fragment. Apparently there is a lagena with the characteristic finer structure, shape and position. There is also a canal that corresponds in form and position to the lateral canal. Otherwise the different parts cannot be very well differentiated. This labyrinth is included in spite of its fragmen- tary character because it shows that even such an imperfect labyrinth is apparently affected in its posture by the influences interacting between it and its environments. In a summary of the first eighteen of these experiments the result that concerns us most is that in all instances where the transplanted labyrinth had developed with sufficient complete- ness for identification of its various parts it was found that it had almost perfectly regained its normal posture. It was this fact that formed the chief object of our investigation, but we may add the following concerning the resultant abnormalities produced by the manipulation of the vesicle at the time of the operation. a. In 8 out of 18 experiments there developed a practically normal labyrinth; b. Where a defect occurs following the operation it is usually localized in some particular region of the labyrinth; it may be confined to the anterior, posterior, or median portion, while the remainder of the labyrinth is quite perfect; c. The three different canals are defective with about equal frequency (30 per cent), though in an individual labyrinth where the canals are involved the imperfection is not distributed among them equally; and the planes of the canals, whether they are defective or not, are usually normal; d. Defects of the endolymphatic appendage occur with about the same frequency as those of the canals; e. The lagena is imperfect more often than any other part (in 8 out of 18 cases) ; f. The acoustic nerve and ganglion were always present, though in a few cases quite diminutive; the connection of the nerve with the brain wall can almost always be recognized. 170 GEORGE L. STREETER DISCUSSION The results that have just been recorded seem to show conclu- sively that there exists some influence between the transplanted ear vesicle and its environment that tends to control its posture, and that an inverted vesicle is thereby rotated back into the nor- mal position. At the same time we must recognize the fact that Spemann (710), from investigations of the same problem with experiments of much the same character, came to quite different conclusions. On examining and comparing our results, however, it will be seen that they are not necessarily contradictory, as thought by Spemann, but may perhaps be better described as differing in degree. Including those in the present paper I have now reported 36 experiments in which the posture of the ear vesicle was especially studied. In 12 of these the ear vesicle had been rotated 180 degrees from its normal position; in 5 of them the ear vesicle had been transplanted without special placement; and in 19 of them the ear vesicle was transplanted and at the same time placed in a definite abnormal posture. In all of these 36 cases the labyrinth regained its normal posture. In Spemann’s twelve reported cases In which there had been simple inversion of the ear vesicle, nine of the ear vesicles remained inverted, while two partially and one completely regained their normal posture. The latter Spe- mann regarded as an unsuccessful operation, surmising that it had slipped back into position directly after the operation. The fact of ‘slhpping’ into the right position is the point of the whole matter. It is the remarkable fact that an ear vesicle, though rotated or turned or transplanted in an extremely abnormal posi- tion, nevertheless ‘slips’ into the correct position, that I am trying to establish in this paper. It is of interest to examine Spemann’s methods in search of some explanation for the difference in our results. For his ex- periments he used Rana esculanta larvae while I used Rana pipiens. It is not likely, however, that this would account for the difference in our results. His operations were performed at the time the ear vesicle is in the process of detaching itself from POSTURE OF MEMBRANOUS LABYRINTH Al the deeper layer of the epidermis. This is the same stage used by me. However, in the technique of the operation our methods are different. Spemann raises a relatively large quadrilateral flap of skin, thereby exposing the ear vesicle, which he loosens and inverts so that the anlage of the ductus endolymphaticus points downward. He then replaces the skin flap and secures it in position by a weight, consisting of a slightly curved strip of cover glass. There are two factors here that may account for the differ- ence in our results. In the first place Spemann, by reflecting a large skin flap, exposes a larger area of the deeper structures and perhaps thereby injures the environment in a way that lessens the postural interaction between it and the ear vesicle. It will be remembered in my operations there was only a linear slit opened, which by the spreading of its edges was sufficient for the manipulation of the vesicle. In the second place the use of the weight, as is done by Spemann, to hold the skin flap in apposition may retard the movement of the vesicle and prevent its rotation. From the nature of my wound no weight was necessary. In addition it may be mentioned that Spemann always placed his vesicles so that the median or convex side remained toward the brain, while in most of my experiments it was made to point lateralward toward the skin. In this respect the displacement of the ear vesicle in my cases was more extreme than in Spemann’s, making it all the more difficult for my vesicles to obtain a normal posture. Furthermore, in Spemann’s experiments the ear vesicles were replaced in their own original pocket, while in all of my last series they were transplanted to another specimen, here again adding to the difficulty of their postural adjustment. In all other respects there seems to be no essential difference in our methods. Judging from my own experiments there certainly exists a de- cided tendency between the ear vesicle and its environment that serves to control the posture, though we know from Spemann’s experience that under certain circumstances this tendency is interfered with and the necessary corrective rotation is not accomplished. 172 GEORGE L. STREETER The development of a normally placed labyrinth from an in- verted vesicle can only occur in one of two ways. Either the ear vesicle at the time of operation consists of indifferent cells which are capable of forming various parts of the labyrinth in accord- ance with how they chance to le (‘harmonisch-aiquipoten- tielles System’); or the ear vesicle itself rotates as a whole after the operation, so that the cells originally intended for the differ- ent parts are brought to lie in their correct relation, where they continue in the fulfillment of their destined development. As was clearly argued by Spemann ’10, it cannot be explained by the former. All our evidence points to a high degree of differ- entiation of the cells of the vesicle, and it is conspicuously proven by their possession of laterality, which has been described in the earlier part of this paper. We could not otherwise have a left- sided labyrinth on the right side of the head. This leaves us with the alternative that the displaced ear vesi- cle does not stay in the position in which it is placed, but rotates into a normal posture. Regarding the nature of the force that produces the rotation there is yet little information. One must take into consideration at least three possibilities which either separately or in combination may explain its action. In the first place, it may be an active phenomenon on the part of the ear vesicle, that is, intrinsic motility of the vesicle itself; secondly, it may be based upon some attraction existing between some por- tion of the vesicle and the brain or other structure; and thirdly, there may be some purely mechanical basis for it. In the first place, regarding the intrinsic motility of the ear vesicle itself, we are familiar with the flowing motion of protoplasm in the case of amoebal pseudopods, and Harrison (’10) has de- scribed the remarkable movements of the protoplasmic processes of nerve cells. In these instances there is a movement of one part of a single cell in relation to the rest of the cell. In the case of the ear vesicle, however, we should have to consider a mass- movement of a group of cells. Such movements have already been described for small masses of cells, such, for example, as the lateral line rudiment, which, as has been shown by Harrison (’03) migrates in the course of a few days, all the way from the head POSTURE OF MEMBRANOUS LABYRINTH 173 region to the tip of the tail. Perhaps of a similar character is the shifting of groups of ganglion cells within the central nervous system as described by Kappers (’08) and the author (Streeter 08). These, however, are movements of small clusters of cells. An example of intrinsic motility of a larger epithelial mass, which will be more analogous to the ear vesicle that we are dealing with, is afforded in the healing of skin wounds of young larvae. The epithelial coat, in such cases, spreads like an elastic sheet from the surrounding area over the wound. This mass-like move- ment of the epithelial coat is very characteristic. On watching the healing of a denuded area one can see the pigmented epithe- lial layer gradually spread in from all sides of the wound, covering in, as it does so, the exposed mesodermal elements, until even- tually its edges tightly pucker together at the center. This is accomplished not by the formation of new epithelial cells but by the stretching out of the cells already there, as can be plainly seen by the alteration of the pigment pattern. The marked tension on the adjacent epithelium resulting from this movement in the region of the wound is strikingly shown by the way in which the pigment line existing along the dorsal crest of the larva becomes deviated toward the operated side. In such cases the epithelium belonging, we will say, to the left side of the body, is drawn well over to the right side, thus indirectly aiding in the covering of the seat of operation. For an excellent description of the behavior of the epidermis in the healing of wounds in larval Necturus, the reader is referred to the paper of Eycleshymer (07). If the epithelial ear vesicle can move and adjust itself to varying conditions in the same way that the skin epithelium moves and adjusts itself we could then understand the rotation of the vesicles recorded in our experiments. A second possibility is that the nerve and ganglion mass may serve to draw the vesicle into its proper position. At the time of the operation some of the ganglion element is usually trans- planted with the vesicle, and as the latter develops we find the ganglion closely attached to the thickened part of its floor which is to form the macula. At the same time fibers have grown out from the central end of the ganglion to attach themselves to the 174 GEORGE L. STREETER side of the brain wall. This tendency to an early nerve-ganglion connection between the ear vesicle and some portion of the brain wall is evident even when the vesicle is transplanted in a strange environment such as in front of the eye (Streeter 06). The nerve seeks the brain wall, and is it conceivable that after it is securely attached it would act as a check or guy rope on the vesicle. With the subsequent growth and the change in the relative positions of the different structures, the nerve attachment would pull on it and, depending on its previous position, would tend to rotate it one way or another. When the vesicle is in front of the eye there are so many abnormal factors present that one could not expect a successful correction of posture through this means. But in the auditory region where other things are favorable it is con- ceivable that the nerve attachment might at least help in the adjustment of the position of the vesicle. As a modification of this idea it is conceivable that there exists some mutual attraction between the macula, ganglion, and brain wall which tends to draw them together by some psysico-chemical process. This would result in the same effect on the labyrinth as a whole as a tension of the acoustic nerve, and would serve in the same way as an assisting force in the adjustment of the position of the vesicle. In an analogous way it is possible that the endolymphatic append- age may be attracted to the side of the brain wall and thus tend to draw the vesicle into position. Wherever the endolymphatic duct and sac are well formed there exists a constant relation between them and the brain which I have never heretofore seen mentioned. The endolymphatic sac always lies closely applied to the membranous roof of the fourth ventricle near the rhombic lip. Some force may bring these two structures together, and thus we would have a secondary correcting influence on the posture of the whole vesicle. This, it must be admitted, cannot be the whole explanation, as we have vesicles with correct posture in cases where the endolymphatic sac is entirely absent. As a third possibility there is to be considered, from a purely mechanical standpoint, the shape of the vesicle and the bed or pocket in which it fits. The auditory pocket is bounded on the median side by the relatively firm brain wall, ventrally by the = POSTURE OF MEMBRANOUS LABYRINTH LAS developing cartilaginous skeleton, laterally and dorsally by the non-resisting auditory capsule and skin. In front of the prodtic ganglion mass and the optic vesicle, and caudally the vago-glosso- pharyngeal complex. These different structures present differ- ent degrees of resisting pressure, and thus from different directions there are these compression forces acting upon the ear vesicle, to which, when normally placed, it is properly adjusted by its shape and the firmness of its different parts. Normally there is an equilibrium between the compression forces of the environment and the resisting forces of the vesicle. When the vesicle is ab- normally placed there is a disturbance in this equilibrium which continues until the vesicle regains its normal position. Thus we would have the mechanical tendency for the disturbed vesicle to fit itself into the right position. An objection to this explanation immediately suggests itself, and that is the fact that vesicles having an abnormal form, and that could not possibly fit well in the usual pocket, notwithstanding, right themselves almost as well as the normally shaped ones. As another mechanical factor one might think of gravity. It is well known that gravity con- stitutes a decisive factor in certain embrylogical processes. Hert- wig (99) and Wetzel (04) experimentally produced deviation from the normal development of the egg by alterations in gravity through the use of a centrifugal machine. It is true that gravity controls the position within the egg membranes of the amphian larvae in the early stages. The ventral side, due to the yolk mass, is heavier and is always down. In the case of the ear vesicle the large macula in its floor is thicker and presumably heavier than the other portions of the vesicle wall, and we might assume this fact as the reason that we always find the macula towards the ventral side and thereby a factor in the posture of the vesicle as a whole. However, when the larvae are removed from the mem- branes, as is necessary for purposes of operation, the conditions are quite altered. Being no longer supported by a gelatinous sphere which is easily kept’ properly erect by gravity, the larvae fall to the bottom of the dish and rest on their side, and we imme- diately have an abnormal direction of gravity, which persists throughout the critical period in the development of the ear ves- 176 GEORGE L. STREETER icle. If gravity were the controlling factor all of the ear vesicles in larvae removed from their membranes would be obliquely placed, whether operated, upon or not. This we know does not occur, and therefore we may safely assume that gravity does not exercise any great influence on the posture of the ear vesicle. LITERATURE CITED EycLesHyMER, A. C. 1907 The closing of wounds in larval Necturus. Amer. Jour. Anat., vol. 7. Harrison, R. G. 1903 Experimentelle Untersuchungen tiber die Entwicklung der Sinnesorgane der Seitenlinie bei den Amphibien. Arch. f. mikr. Anat., Bd. 63. 1910 The outgrowth of the nerve fiber as a mode of protoplasmic move- ment. Jour. Exp. Zodl., vol. 9. Hertwia, O. 1899 Ueber einige durch Centrifugalkraft in der Entwicklung des Froscheies hervorgerufene Verinderungen. Arch. f.Mikr. Anat., Bd. 53 Kapprers, C. U. A. 1908 Weitere Mitteilungen ueber Neurobiotaxis. Folia Neuro-Biologica, Bd. 1. Lewis, W. H. 1907 On the origin and differentiation of the otic vesicle in am- phibian embryos. Anat. Rec., vol.1. (Amer. Jour. Anat., vol. 7). Spremann, H. 1906a Uber eine neue Methode der embryonalen Transplantation. Verh. d. Deutsch. Zo6l. Gesell., 1906 b Uber embryonale Transplantation. Verh. d. Gesell. Deutsch. Naturf. u. Arzte, 78 Vers., Stuttgart. 1910. Die entwicklung des invertierten Hérgriibchens zum Labyrinth. Arch. f. Entwick.-Mech., Bd. 30. SrrEETER, G. L. 1906 Some experiments in the developing ear vesicle of the tadpole with relation to equilibration. Jour. Exp. Zodl., vol. 3. 1907 Some factors in the development of the amphibian ear vesicle, and further experiments on equilibration. Jour. Exp. Zodl., vol. 4. 1908 Nuclei of origin of the cranial nerves in the 10mm. human embryo. Anat. Rec., vol. 2. 1909 Experimental observations on the development of the amphibian ear vesicle. Anat. Rec., vol. 3. WeTzEL,G. 1904 Zentrifugierversuche an unbefruchteten Eiern von Rana fusca. Arch. f. mikr. Anat. u. Entwick., Bd. 63. MULTIPLE FACTORS IN MENDELIAN INHERITANCE! E. C. MacDOWELL Sheffield Scientific School From Osborn Zoélogical Laboratory, Yale University HISTORICAL Since cases of simple Mendelian phenomena have been very frequently described, the interest of the student of genetics has shifted from the attempt to prove, or disprove, Mendelism to- wards the investigation of the extent of the application of Men- del’s fundamental principles. Perhaps this is why the appar- ent exclusion of anything Mendelian from size inheritance by Castle’s work on the ear length of rabbits, aroused so much in- terest. Castle (’09) found that the ear lengths of rabbit offspring were in general intermediate in relation to the parents. This seemed to indicate a simple blending inheritance in which the size-controlling elements from either parent were permanently joined in the offspring, never to segregate. Yet in the same rab- bits Mendelian ratios were being given and segregation was taking place in the color of the hair. The appearance of a paper by H. Nilsson-Ehle (’09) brought a new possible interpretation of the ear length crosses. In this paper evidence was presented to show that there may be two or more Mendelian factors for the same character, which factors develop the character whether they appear alone or in any combination with the others. Nilsson-Ehle found in certain crosses of oats and wheat in which the colors of the glumes and seeds were considered, that simple Mendelian ratios (3:1) were generally given, but in cer- 1 THis paper is based on an investigation carried on in the Laboratory of Ge- netics of the Bussey Institution with assistance from the Carnegie Institution of Washington. A full report of the investigation has been submitted to the Car- negie Institution for publication. This statement of some of its main features is published by permission. For the conclusions drawn from the facts recorded the author is alone responsible. UTZ THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 2 FEBRUARY, 1914 Ue: 178 E. C. MACDOWELL tain strains he obtained consistent ratios which indicated di- and tri-hybrid crosses (15:1, 63:1); in other words, there were two or three factors involved, each of which alone or in combi- nation with the others, produced the same color, and only in the absence of all such factors could the recessive color be seen. The application of this theory to the case of the rabbit ear lengths was immediately made, East (710), Castle (11), Lang (11), by supposing that ear length depended upon several fac- tors each of which behaved as a simple Mendelian unit, but lacked dominance. This would mean that we could see the difference between an ear whose length had been based upon four factors, and one whose length depended upon five or six factors. It would be supposed that the more simplex factors, or doses, there were present, the longer the ear. If the long-eared race bore four duplex factors (8 doses) for length, and the short-eared race bore two similar but not allelomorphic factors (4 doses), the hy- brids obtained by crossing these two races would each have six factors in a simplex condition (6 doses). The relative sizes of the two parents and the hybrids would be expressed by the num- ber of doses in each case. The second generation would give rab- bits with various numbers of doses and their ears would range all the way from the length of the long-eared race to that of the short-eared race. Unless the two factors in the small race were allelomorphic to two of the four in the large race, there could be found rabbits with ears longer than the long-eared race, and others with ears shorter than the short-eared race, for in the sec- ond generation would be formed combinations involving as many as 12 doses and others with none. There would, however, be a much greater number of rabbits with intermediate than with extreme ear-lengths. Just as in a mono-hybrid cross without dominance there are two heterozygous individuals to one of each of the pure types, so in a multo-hybrid cross, where each factor, independent of the others, is giving a 1 : 2:1 ratio, the individ- uals with all the factors in a simplex or heterozygous condition would be most numerous. The higher and lower grades would appear in decreasing frequency as the extremes were approached (Tammes 711). Now if only a few animals were raised in F, MULTIPLE FACTORS IN MENDELIAN INHERITANCE 179 it would be expected that they would have intermediate ear lengths. The numbers obtained by Castle were small enough not to show the new combinations expected on this theory. On account of the many external conditions that may influence the development of the hereditary size factors, one could not ex- pect to find the various grades in F, clearly enough separated to discover the number of factors that might be involved. How- ever, evidence favoring a multiple factor interpretation of size would be found in the appearance in F, of extremes not found in F, or the increased variability of F, over F,, especially if ex- tremes beyond the parental races appear. In plants numerous experiments favoring a multiple factor hypothesis have been described. Since definite ratios may be obtained for qualitative differences, such characters afford the most convincing evidence. Supporting Nilsson-Ehle’s results, East (710 and ’11), and East and Hayes (’11) in endosperm and pericarp colors of maize, found ratios which are apparently to be explained only by multiple factors. Tammes (11) crossed two different sorts of blue Linum; an intermediate first gen- eration was followed by a second generation with a wide range of variability, the intermediate shades having the greatest frequencies. From quantitative differences in color one is led to quantitative differences in shapes and forms. The work of Kajanus (711 and 712) on the roots of Brassica and Beta affords evidence to sup- port the hypothesis that various form types of these roots are due to combinations of multiple factors for length, or length and roundness. Emerson (’10) reports increase in the variabil- ity of the shapes of the second generation from certain crosses with squashes and gourds. These cases of form differences do not show dominance. Shull (’11 b) described a case involving a double gene for the same character, that is dominant. He found that the flat and triangular seed capsule of Bursa bursa-pastoris is dominant to the oval and, in section, round seed capsule of its recent mutant heegeri. The ratios of the second and third gen- erations strongly indicate that there are two genes involved, each of which can produce the pastoris type of capsule. 180 E. C. MACDOWELL From quantitative differences in form we come to quantitative differences involving size alone. The possible application to size inheritance of this theory of multiple factors, so strongly indicated by color characters, was soon realized by Nilsson-Ehle (07). He found that crosses between races of wheat or oats with long and short stalks gave intermediate hybrids whose off- spring had stalks ranging all the way from the length of the long- stalked race to that of the short-stalked race. The length of the head of a certain wheat acted in crosses as though it were dependent upon a strong dominant shortening factor and two weaker factors for length, that lacked dominance. Spillman (’02) in crosses involving the length of wheat heads, and Emerson (11), in certain crosses involving the height of corn plants, re- port intermediate first generations followed by second genera- tions in which the ranges of variability included, or even ex- ceeded, the extremes of the parental races. East and Hayes (’11) present crosses of maize involving ear length, number of rows per ear, size of grain; Emerson (’10) sizes of beans; Tammes (’11) sizes of Linum seeds and petals; in all these cases the same char- acteristic results were found, namely, a constant and intermedi- ate first generation and a second generation with wide variabil- ity. Philips (12) gives a preliminary report on size crosses in ducks which indicates a like increase in the variability of the second generation birds. Certain structural characters have been found that seem to depend upon two or more similar units of inheritance. In oats the presence and absence of ligulae, and the arrangement of the spikelets in the head—either on all sides of the rachis or only on one side—give strong evidence of multiple factors, definite ratio being obtained (Nilsson-Ehle ’09). Such physiological characters as winter hardiness, rust re- sistance and flowering times of cereals have been found by Nils- son-Ehle to show increased variabilty in crosses. Tammes (711) found similar results in the opening or remaining closed of ripe Linum capsules. They partly open in the first generation; in the second generation some remain closed, others fully open, while most are half opened. Tschermak (’11) and Leake (11) pre- MULTIPLE FACTORS IN MENDELIAN INHERITANCE 181 sented studies on the blossoming times of peas and cotton. In both plants crosses between late and early blooming varieties gave plants with intermediate blooming time. In the second genera- tion some plants began to bloom as early as the early blooming variety and others as late as the late blooming variety. Perhaps as strong evidence for the assumption of different fac- tors influencing the same character as is afforded by ratios in the second and third generations, is found in crosses between apparently similar races which give marked variability in the second generation following a first generation that was like the two parents, and no more variable than the parental races. In “such cases the characters in the two lines that have been crossed are supposed to be caused by different factors which produce the same effect. In the second generation, where recombinations first have a chance to appear, new grades are found. Cases have been mentioned that show extremes beyond the parental types. In these cases there must have been a different number of factors in each parental race, if we suppose for the moment that the factors are all of equal power. In the following cases each parental race may be supposed to have the same number of factors (although not identical ones) and so the races appear to be alike. The occurrence of two white races of sweet peas which give colored offspring when crossed is undoubted evidence that two different factors are involved in producing color. In the follow- ing cases the factors are supposed to be similar to each other so that their presence cannot be determined until the second genera- tion. Nilsson-Ehle (711) found two strains of red wheat, which, when crossed together gave whites in the second generation. Whites also appeared when certain lines of black oats were crossed, in the ratio of fifteen blacks to one white. To lines of oats (Nilsson-Ehle ’09) whose spikelets were arranged on all sides of the rachis gave, in the second generation, some plants with spikelets on one side of the rachis. Kajanus (’11) reported three crosses between apparently similar races of long beets which gave the same form in the first generation, while in the second genera- tion some very long individuals were obtained. In crossing dif- 182 E. C. MACDOWELL ferent strains of cotton of equal height Balls (07) found a great variability in the heights of the second generation. The first generation was unifornly close to 100 cm., the second generation ranged from 30 cm. to 100 cm. Hayes (’12) has crossed varieties of tobacco with similar numbers of leaves which produced plants with the same number of leaves in the first generation. The leaf number for the plants of the second generation was very variable. Keeble and Pellew (’10) have worked out a very clear explanation of an increase in variability found in a second generation follow- ing the crossing of two semi-dwarf races of peas. The parental races were very constantly between 3 and 4 feet tall. The first generation showed increased vigor, but a constant height of 7 to 8 feet. The second generation ranged from 13 to 8 feet. This case differs from those preceding in that the factors, to whose recombination the new grades in the second generation were due, did not produce the same effects. One factor increased the dis- tance between the nodes, the other increased the thickness of the stem, and this thickening enabled the internode to attain greater length. One parental race had the factors for thick stems and short internodes, the other had the factors for thin stems and long internodes. The recombinations in the second generation produced plants with long internodes and thick stems (8 feet as well as plants with short internodes and thin stems (1} feet). The combinations found in the parents also occurred. More- over, the ratios in the four classes closely approximated expec- tation. In this case height is dependent upon two Mendelizing factors. In other cases size may depend upon a single factor. Some of these cases are mentioned: tall and dwarf peas, Mendel; tall un- branched habit vs. dwarf branched sweet peas, Bateson and Pun- nett (’08); axial vs. terminal position of bean pods, Emerson (04); tall vs. dwarf Antirrhinum, Baur (’11); dwarf vs. normal tomatoes, Drinkard (’08); long vs. short styles of Oenothera, de Vries (01, p. 435); long vs. short wings of Drosophila, Morgan (11); long vs. short hair in various mammals, Castle (?05) ; brach- ydactylous digits in man, Farabee (’05) and Drinkwater (708); MULTIPLE FACTORS IN MENDELIAN INHERITANCE 183 short stocky vs. long slender legs of Dexter-Kerry cattle, J. Wilson (’09). The evidence presented by plant breeders seems to carry con- viction. It has shown that the assumption of multiple factors is the most simple theory to explain certain ratios in crosses in- volving color and its absence. Similar ratios are also found in crosses involving certain structural characters (ligulae and the arrangement of spikelets in wheat). Now if qualitative char- acters may depend upon multiple factors in certain cases, there is Just as good evidence for saying that certain morphological characters depend upon multiple factors. Size crosses differ from crosses involving multiple factors for color and structural char- acters in that in the latter cases the color or the structure may be absent, and, so definite ratios can be found; whereas in the former size as such cannot be absent, so in most cases no definite ratios can be found. Otherwise, these two types of characters are strikingly similar. Both show wide variations in the second generation not found in the parental or first generations. Both types show similar wide variations in the second generation in crosses between certain strains that appear to be alike and in these cases the first generation is like the parents. Both types may show mono-hybrid ratios in the same sorts of characteristics. Any theory to account for the wide variations in the second gen- erations of crosses involving characters which cannot afford defi- nite ratios must also account for this similarity with cases where definite ratios may be found, and again, with cases where sim- ple Mendelian ratios and complex ones are found for the same character differences. Size and Mendelian inheritance are not incompatible, as is shown by many mono-hybrid ratios from crosses involving size. In one case (semi-dwarf peas) it has been shown even by ratios that size may depend upon two distinct Mendelian factors. Surely the most simple theory that can be given to explain this phenomenon of increased variability, which in itself cannot be doubted, being reported by many investiga- tors for so many characters, is that there exist units of inheri- tance, introduced by the parents, to the segregation of which units in the germ cells of the hybrids, the new combinations are due. 184 E. C. MACDOWELL EXPERIMENTAL In 1908 Castle started a second experiment that should test this hypothesis by obtaining larger numbers of animals so that the variabilities of the first and second generation could be com- pared. As the size of the whole animal was to be considered, crosses were made between a small male rabbit from a small race, the Himalayan, and a series of large females that had been used in the ear length crosses. The pedigrees of the females are known and for several generations there were only slight differ- ences in the weights of the pairs of ancestors. In 1910 the ex- periment was put into the hands of the writer. By that time the original crosses had been made and weights for growth curves had been recorded for the animals raised up to that time. I wish to acknowledge here a keen appreciation of my indebtedness to Professor Castle for the privilege of completing this work which he had planned and already started, and to express gratitude for the assistance and advice that he has contributed towards the completion of this work. In most cases the second generation consisted of a back cross of the first generation females to their male parent. A few back crosses of first generation males to their female parent were made and a few crosses between first generation males and fe- males. Measurements were made of the skull and long bones of the rabbits after the bones had attained full size. Fifteen months was set as the age for killing a rabbit, As it was found that the bones were fully grown at that age. Sixteen measurements for each set of bones were recorded. a. Coefficients of size It was found that there was enough lack of correlation between the various measurements to give different results when different characters were considered, so it became needful to obtain a number for each animal in which the various measurements would be equally represented to express the size of each animal. As the measurements varied from 2 to 10 em. no absolute average could be used; for a small deviation in a short measurement MULTIPLE FACTORS IN MENDELIAN INHERITANCE 185 would have far greater significance than an equal deviation in a long measurement, and so in any average the large deviations in the long measurements would entirely overbalance the equally important small deviations in the short measurements. The fol- lowing method was used to obtain a number for each animal, which may be called the coefficient of size (C. 8.), a number based on relative deviations. The average of a character for one fraternity was used as a dividend into which were divided the individual measurements of the animals in the same fraternity. The quotients so formed gave a series of ratios expressing the relative sizes of the various sibs above or below theirmean. By this method the ratios of the other characters of one animal to the corresponding fraternal means were obtained. The average of these ratios gave the coefficient of size (C. 8.) for that animal. These coefficients range from 0.930 for the largest animals to 1.070 for the smallest ones. In classifying them, classes 0.005 in extent were used. From these distributions, standard devia- tions were calculated. Since the means of all fraternities lie always in the column whose value is 1, it will be realized that the sums of all the individuals in each column will give a correct de- scription of the variability of one generation of a whole family; in other words, different fraternities from the same family may be classified together. The standard deviations of these family distributions will be based on deviations from the various fra- ternal means. This will give a more accurate result than if the actual deviations had been calculated from the means of all the individuals averaged together, as would be done in applying the formula for standard deviation to the values themselves. It must be supposed that the means of all the fraternities do not coincide. The mean of all the values would give the extreme variates even greater deviations and the standard deviations would be higher than those obtained by this method. It will be seen now that the coefficients of size of all the animals in one generation may also be classified together. In the tabulation of these coefficients of size it is at once ob- served that all the most extreme coefficients belong to individuals 186 E. C. MACDOWELL from the back cross. Naturally both the F, and the back cross offspring occur in greatest numbers about the middle class and decrease in either direction away from this class. But the back cross extends much further than the first generation. It is un- fortunate that the numbers of the Fi; are very much smaller than those of the back cross. It may be claimed that with larger numbers more extremes would have been found in the F;. This is undoubtedly true, but it is highly improbable that any indi- viduals that approached the back cross extremes would have TABLE 1 Standard deviations of coefficient of size of the first (F1) and back cross (B.C.) gen- erations in the various families. The numbers of individuals are given for each fraternity (No.). FAMILY F, NO. | B.C. NO. 647 0.0145 4 | 0.028 = 0.002 | 25 1443 | | 0.026 + 0.003 17 1471 0.015 7 1491 0.015 4 0.019 + 0.001 24 1493 0.009 13 0.021 + 0.001 60 1531 | 0.027 = 0.001 45 1532 | 0.027 = 0.003 | 15 1537 / 0.020 = 0.001 46 2011 0.019 5 0.025 = 0.003 | 11 1493! x 2379 0.019 14 | 0.017 = 0.001 | 31 Total..........| 0.014 + 0.001 33. * ~=«0.023 = 0.0007 243 1 This family is not from a size cross. Is is used as a check, so is not included in the totals. been found. The standard deviations show that there is a real difference in the variability of the two generations. In table 1 the standard deviations in terms of the coefficients of size are given for the two generations, one family at a time as well as for the totals of each generation. In every family resulting from a size cross the standard deviation of the back cross exceeds that of the first generation. Family 1493 x 2379 was carried along as a control as the parents were nearly the same size. It will be noted that the standard deviation of the back cross, in this case, MULTIPLE FACTORS IN MENDELIAN INHERITANCE 187 is slightly smaller than that of the first generation. In families 1443, 1531, 1532 and 1537 there were not enough F, individ- uals to use in determining coefficients of size. The back crosses of these families are given for comparison with the first genera- tions and back crosses of other families. In family 1491 the standard deviation of the back cross is the same as the highest first generation standard deviation, namely, that in family 2011. In 1493 and 1537 the back cross standard deviations are very nearly as low as the highest first generation standard deviations, but the standard deviation of the first generation of family 1493 is much lower than that of the back cross of the same family. This is the largest family obtained. b. Classification in relation to the parents The data were next treated in a way to show graphically the size relations between the parents, the first, and back cross gen- erations, considering one character at a time. This method of treatment also shows the relative variability of the first and back cross generations and offers another method for combining differ- ent families. This method may be called classification in rela- tion to the parents. For each family 15 classes were formed be- tween the parental classes. ach parental measurement is taken as the middle of a parental class; the extent, or range of each class equals one-sixteenth the difference between the parental measure- ments. Into classes so formed the first and back cross genera- tions were separated. This was done for all characters in all families. From these classifications the number of animals of all families that fall half way between their parents in regard to any measurement is readily found by adding the frequencies in the middle column; likewise the sums in other columns show the distribution of animals in other relations to their parents. The male parent in all families was the same animal. The female parents differed in size to some extent. This means that in the various families and in the totals, individuals are put together that are not exactly similar as to size. They are, however, sim- ilar as to their relative positions between their respective parents. 188 E. C. MACDOWELL In general the size of all the measurements of the first genera- tion is plainly intermediate, although the modes and means of the various frequencies are markedly above the midparental. This high distribution may readily be interpreted as the effect of increased vigor from the cross. Similar plant crosses show a like increase in size in the first generation (see East ’09, Shull 09, Darwin ’76). A marked difference in variability is seen in these first generation frequencies when different characters are compared. When the range is wide it is as much extended in one direction as the other, showing that the extreme lows are not to be accounted for by underdevelopment, as an asymmetrical variation might suggest. Such a difference is found when the distribution for the ulna, which has a wide range, is compared with the distribution of the skull length, which has a very limited range (table 2). Four of the remaining measurements are sim- ilar to the skull length; two are intermediate; six are wide, slightly less than the ulna. Yet these frequencies include the same in- dividuals. Pearson (’02) found similar inequalities in different parts of the human body. Hatai (07) found greater variability in the length of the nasal bone and in the zygomatic width in skulls of albino rats than in any other skull characters. The variations in the ranges of distributions of the different characters, described for the first generation, are also found in the back cross. The means of the back cross fall near the class half way between the midparental (middle class) and the small parent to which the first generation females were back crossed. A very marked feature of these back cross frequencies is the number of measurements that are as low and lower than the small parent, and as high as, and higher than the mode of the first generation frequencies in which their mothers are included. Since the father and grandfather in all the families was the same rabbit, the actual values of the male (small) parental class in all families is the same, and the values of the adjacent classes differ very slightly in the different families. This means that all animals falling in classes near his class have very nearly the same sizes in the different families. This makes the occurrence of these short measurements the more convincing evidence of segregation. GS PUD SjUudWaINSpaU om) 61 616 Walks G ST] 40} 91 | St ay} fo Ajquyrqvispa aarnja 6 ATAVL 4 9y) Moys 07 ‘syzbua) Duyn pun syjbua) 7Nnys UO pasng ¢ | jet fer jet for fox or zt oz oc tt er\e|elzltielz| |r] |_| One Geeta ete hy WT % I he | W I gi 8 |9T J91 [re |se [ze \ch 81 ler & 9 |e | | poke: €|4 01s js ie ¢z | | at cr mjor] o6|s|zi9|¢ | | @ | %} 4] o jr—le-le—|s—-|e—-l9—|2—|9—|6— Or |11—let—le1—| va Me Hh | t = ~ U is 2 WS (O°) sso49 yong ay? pun ('y) Uworvsowab jsf ay) fo ; i J Bun jo Yysuory [ys jo Y}suo'T = 18 UMLOVUV AD sjuaiod 07 U0NDjaL Ur UoYynorfissy7) 190 E. C. MACDOWELL With assistance from Mr. 8. Wright standard deviations were calculated for the total first generation and back cross frequencies, as classified in relation to the parents, for each measurement (table 3). These constants do not give actual deviations because the totals are classified.in a purely relative manner. However for purposes of comparing the two generations these standard deviations are permissible because the frequencies in both gen- erations are classified in relation to the same parental classes. TABLE 3 Means and standard deviations of the first generation (F,) and back cross (B.C.) based on classifications in relation to the parents, similar to those given in table 2 and expressed in terms of those classes. MEASUREMENT | MEAN F\ | mpan B.C. of | o Fi o B.C. skull length, total..... 10.80 D200 1.94 + 0.15 2.60 + 0.08 skull length, partial, 1. 10.56 5 .24 2.76 = 0.21 2.69 = 0.08 skull length, partial, 2. 10.02 5.28 2.50 = 0.19 2.80 = 0.09 skull width, anterior... 8.39 2.01 3.17 = 0.24 3.52 + 0.11 skull width, posterior... 11.00 Deley 3.68 + 0.27 | 4.58 = 0.14 GEGt ie hay Fg ch ae 10.12 5.69 2.56 = 0.19 2.88 = 0.09 AUR Yee CAC te Vicente 9.20 4.35 3.47 + 0.26 2.87 + 0.09 MENON OUE Ms. dos aes on: 10.27 4.75 2.04 = 0.15 2.99 = 0.09 MMOH AKG MOLY Pack Gagne eon e 8.95 3.16 | 2.90 = 0.22 3.29 = 0.10 uM ens eee 12.20 5.86 3.65 = 0.27 3.55 = 0.11 UNS ee. eis 4 eee 12.24 5.27 5.43 + 0.42 4.61 + 0.15 REMMI Se Sey. ee Ge i os 11.83 5.54 3.85 + 0.30 3.38 = 0.10 tibia il .90 6.16 4.07 = 0.31 3.98 = 0.12 Six measurements show significant increase in the back cross over the first generation, three show decrease in the back cross, and four lie within the probable error, meaning practically no change. This is a mathematical demonstration that there is enough lack of correlation between the bone measurements for different characters to give different results. Since twice as many characters show increase as decrease, it can be concluded that there is greater variability in the back cross offspring than in the offspring of the first generation. MULTIPLE FACTORS IN MENDELIAN INHERITANCE 191 c. Body weights When the work was taken over by the writer it was believed that body weight could be taken as a measure of size. Ac- cordingly weights of the rabbits were recorded weekly. From these records growth curves were plotted by which the adult weight for each animal was to be determined. That this could not be done with any nicety, was clearly demonstrated by the study of some 300 curves. Rabbits’ weights are very sensitive to changes in conditions; and to obtain curves that would be smooth enough to determine adult weights with accuracy would require more perfect experimental conditions than it has been possible to obtain in raising large numbers of animals. In many cases fat is deposited in such a way that there is no flattening of the curves at about 150 days, as in most cases; instead it may con- tinue to rise for a year without flattening. Pregnancies and nursing disturb the curves of the F, females. The most important information the growth curves afford is their vouching for the recovery of animals that have been sick. In spite of fluctuations, one can see in nearly every one the trace of a regular curve. Through a recognition of the normal type of curve it soon became possible to determine whether at a cer- tain point in a curve an animal was above or below its normal. Based on this element of regularity, which, when shown by a part of a litter, gives the type of curve the other would be expected to have followed, adult weights were estimated. Approxima- tions were made within 100 grams; in a few cases within 50 grams. Coefficients of variability were calculated for the F, and back cross fraternities (table 4). Since the means were absolute, standard deviations could not be used in making comparisons. In ten fraternities the C. V. of the back cross animals are higher than any of those of the first generation fraternities. In five fra- ternities the coefficients of variability are lower than the highest coefficients of the first generation. The weights show, then, that greater variability is found in many of the back cross fraternities than in any first generation fraternity. Whereas by themselves 192 E. C. MACDOWELL TABLE 4 Coefficients of variability of weights, arranged by generations. Each fraternity is considered separately in the back cross (B.C.) FAMILY C. V. of Fi | No. C. V. of B.C. | No. 647 5.16 [ere Salad | 0 Settee | 85 | 5.58 6 10.27 7 1471 4.09 | i | 1491 6.22 | 4 10.46 10 6.32 7 1493 8.75 bY Sts 13 .56 8 | 8 .30 | 8 9.29 12 9.94 6 1531 13 .46 16 10.45 | 14 1532 10 .27 ly ae 1537 7.35 | a6 8.23 13 2037 | 10.83 | = a8 these weights might bear little or no conviction, on account of the roughness of their estimation, as cumulative evidence in connection with the results based on bone measurements, they certainly may be considered. CONCLUSION The conclusion to be drawn from these observations seems clear. The back cross is more variable than the first filial gen- eration. This appears in the relative distributions of the co- efficients of size of the two generations, whether compared by observation or by standard deviations: it is found when the two generations are classified one character at a time, in relation to the original parents, whether single families or whole genera- tions are considered: and finally the coefficients of variability of the estimated body weights support the same conclusion. There occur characters among the back cross offspring that are smaller than the corresponding characters in the small parent and others that are larger than the modes of the first generation large parents. MULTIPLE FACTORS IN MENDELIAN INHERITANCE 193 This conclusion is very similar to many of the cases cited above. All of these deal with heritable characters quantitatively different. They are subject, to a greater or less degree, to fluctu- ations that, not being heritable, may be roughly assigned to environment. Offspring from crosses between extremes are gen- erally of an intermediate nature. In the following generation new forms appear that are similar to the original parents or even more extreme. The greater number of individuals are inter- mediate. In certain cases crosses between similar lines, after a first generation like the parents, give a second generation in which a wide range of grades appear. These are the facts that can be definitely ascertained from the work that has been done. On the probability that these same phenomena will always be found, a law may be stated: the second generation of a size cross will show greater diversity than does the first generation or the parental lines. All practical application will come from this. The interpretation of multiple factors can be applied to all the facts. It goes hand in hand with the mutation, and pure line doctrines of de Vries and Johannsen, and in its breadth of application, and its comprehensive simplicity, this theory, based on the assumption of the segregation of distinct units, is very attractive; by its use as a working hypothesis important facts have been discovered; its acceptance and further development will help to establish a broad and unified system of heredity. BIBLIOGRAPHY Batis, W. L. 1907 Mendelian studies of Egyptian cotton. Journ. Agri. Sci., vol. 2. Bateson, W. anD PunnEtTT, R.C. 1908 Experimental studies in the physiology of heredity. Rep. Evol. Com., No. 4. Baur, E. 1911 Einfuhrung in die experimentelle Vererbungslehre. Berlin. CastLe, W. E. 1909 Studies of inheritance in rabbits. Pub. Carnegie Inst. of Wash., No. 114. Darwin, C. 1876 The effects of cross and self fertilization in the vegetable kingdom. London. DrinKarp, A. W. 1908 Inheritance in tomato hybrids. Virg. Agric. Exp. Sta. Bull. 177, p. 18. Drinkwater, H.+ 1908 An account of a brachydactylous family. Proc. Roy. Soc., Edinb. 28. East, E. M. 1910 A Mendelian interpretation of variation that is apparently continuous. Amer. Nat., vol. 44. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, NO. 2 194. E. C. MACDOWELL East, E. M. 1911 The genotype hypothesis and hybridization. Amer. Nat., vol. 45, pp. 160-174. East, E. M., anp Hayes, H. K. 1911 Inheritance in maize. Conn. Agri. Exp. Sta. Bull. No. 167. Emerson, R. C. 1904 Heredity in bean hybrids. 17th Am. Rept. Agri. Exp. Sta. of Nebraska. 1910 The inheritance of sizes and shapes in plants. Amer. Nat., vol. 44. 1911 Genetic correlation and spurious allelomorphism in maize. 24th Ann. Rept. Agri. Exp. Sta. of Nebraska, note, p. 88. FaraBeg, W.C. 1905 Inheritance of digital malformations in man. Papers of Peabody Mus. of Amer. Arch. and Ethn., Harvard University, vol. 3. Hayes, H. K. 1912 Correlation and inheritance in Nicotiana tabacum. Conn. Agric. Exp. Sta. Bull. 171. Kasanus, B. 1911 Genetische Studien an Beta. Zt. f. indukt. Abstammungs- u. Vererbungslehre. Bd. 4, pp. 137-179. 1912 Genetische Studien an Brassica. Zt. f. indukt. Abstammungs- u. Vererbungslehre. Bd. 4, pp. 217-237. KEEBLE, F., AND PeLtew, C. 1910 The mode of inheritance of stature and time of flowering in peas (Pisum sativum). Journ. Genet., vol. 1, p. 47. Lane, A. 1911 Die Erblichkeitsveshaltnisse der Ohren der Kaninchen nach Castle und das Problem der intermediaren Vererbung und Bildung kon- stanter Bastardrassen. Zt. f. indukt. Abstammungs-u. Vererbungs- lehre. Bd. 4, pp. 1-23. LEAKE, H. M. 1911 Studies in Indian cotton. Journ. Genet., vol. 1, pp. 205- 272. Morean, T. H. 1911 An attempt to analyze the constitution of the chromo- somes on the basis of sex-limited inheritance, in Drosophila. Jour. Exp. Zool., vol. 2, pp. 365-411. Nitsson-Hute, H. 1907 Om liftyper och individuell variation. Botan. notiser, Pp. 113-140. 1909 Kreuzungsuntersuchungen an Hafer und Weizen, I. Lunds Uni- versitets Arsskrift, ING ee Atay ssa. onminae2 1911 Kreuzungsuntersuchungen an Hafer und Weizen, II. Lunds Universitets Arsskrift N. F. Afd. 2, Bd. 7, nr. 6. Puities, J. C. 1912 Size inheritance in ducks. Jour. Exp. Zool., vol. 12, pp. 369-380. Suuuui, G.H. 1911 Defectiveinheritance ratiosin Bursa hybrids. Verh. Naturf. Ver. Brunn. Bad. 49, p. 157. Spruypman, W. G. 1902 Exceptions to Mendel’s law. Science, N. S., vol. 16, p. 794. Tames, T. 1911 Das Verhalten fluktuierend variierenden Merkmale bei der Bastardierung. Extrait du Recueil des Travaux botanique Néer- landals. Bd. 8, Livr. 3. TscHERMAK, E. v. 1911 Uber die Vererbung der Bliitezeit bei Erbsen. Verh. Naturf. Verein. Brunn. Bd. 49. DE Vrigs, H. 1903 Die Mutationstheorie. Leipzig. Witson, J 1909 The origin of the Dexter-Kerry breed of icattle! Sci. Proc. Roy. Soc., Dublin, N.S. 12, p. 147. THE REACTIONS OF NORMAL AND EYELESS AMPHIBIAN LARVAE TO LIGHT HENRY LAURENS Osborn Zoélogical Laboratory, Yale University TWO FIGURES A great deal has been written about the photic reactions of amphibians, but this has been concerned chiefly with the adults, the larval forms having received but scant notice. In the spring of 1913 a series of experiments were carried out upon the larvae of Rana pipiens, R. sylvatica, and Amblystoma punctatum to determine whether they were sensitive to light, and if so, whether they were sensitive to light received through the skin as well as to light received through the eyes. I take this opportunity to express to Dr. R. G. Harrison my thanks for suggesting this piece of work to me, as well as for his continued interest and criticism. Banta and McAtee (’06, p. 71) have noted that the larvae of the cave salamander are much more responsive to light than are the adults, and that the younger larvae are more responsive than the older. Both larvae and adults are negatively phototactic. Eycleshymer (’08) found that Necturus larvae were negatively phototactic, both in their natural environment and in the aqua- rium; furthermore, that they orient in a definite way, such that the light falls with equal intensity upon the two sides of the body. The only other mention made of the reactions to light of am- phibian larvae is that by Franz (10 and 713). Franz describes tadpoles as being indifferent to light—non-phototactic—except when they are crowded into,a small space, under which conditions they will all orient themselves to the rays of light, so that their heads are directed toward the source of light. Franz cites this case as one of the numerous examples of so-called phototaxis, 195 196 HENRY LAURENS which he has called a ‘Fluchtbewegung,’ caused by the abnor- mal conditions under which the animals find themselves. My experiments were carried on in a basement dark-room, the temperature of which varied between 15° and 19°C. A 60-watt Mazda lamp on a 110-volt circuit was used as the source of light. This was placed in a wooden box, in one end of which there was an opening 8 cm. square, through which the light was projected. The larvae whose reactions were to be tested were placed singly in a flat cylindrical glass dish, of 22 cm. diameter, and 8 cm. deep, holding 2500 ec. of water. The sides and bottom of the dish were covered with black paper, except on the side toward the light where a small window, 3 cm. wide, and extending from near the top to the bottom of the dish was cut. The dish was placed so that its middle point was at a distance of 50 cm. from the lamp, at which point the light had an intensity of about 192 candle- meters. Between the lamp and the dish, and 15 cm. from the former, a screen with an opening 3 cm. high, and 1.5 em. wide, through which the light passed, was set up. The larvae used varied somewhat in size, but at the beginning of the experiments the tadpoles were about 12 mm., the Ambly- stoma larvae about 18 mm. long, none of them being shorter than 15 mm. During the course of the experiments, which were be- gun early in April, and continued through May and part of June, the larvae grew in size, until the tadpoles were about 20 mm., the Amblystoma 30 mm. long. In order to find out whether the larvae were sensitive to light received through the skin, it was necessary to remove their eyes. The method used for removing the optic vesicles was that described by Lewis (04, 705 and ’07) and by LeCron (07). The instruments used were a pair of very finely ground Noyes iridectomy scissors, and a fine pointed pair of forceps for hold- ing the embryos. The stage at which the tadpoles were operated is that figured by Harrison (’04, p. 201) when the tail bud is just beginning to be perceptible (fig. 1). The Amblystoma embryos were operated at the corresponding stage (fig. 2). All the em- bryos were operated under a binocular microscope. They were placed in watch glasses in a 0.2 per cent normal salt solution, in REACTIONS OF AMPHIBIAN LARVAE TO LIGHT 197 which they remained until the wounds had healed, after which tap-water was gradually added. When the larvae had attained a certain size they were transferred to battery jars, each larva being placed in a single jar, which was numbered. All the larvae tested were kept separate in small battery jars about half full of water and in whieh a few water plants of vari- ous kinds were placed. The frog tadpoles will feed upon the Fig. 1. Embryo of R. sylvatica, to show the stage of development used in the beginning of the experiments (after Harrison). X9}. Fig. 2. Embryo of Amblystoma punctatum, to show the stage of development used in the beginning of the experiments. X 93. leaves of these plants but the Amblystoma larvae will not. The Amblystoma larvae were fed regularly during the course of the experiments on small crustaceans which they devoured eagerly, the blinded larvae seeming to have very little difficulty in seiz- ing their quickly-moving prey. The jars were numbered, so that the reactions of the individual larvae could be followed. It was found necessary to isolate the Amblystoma larvae early for 198 HENRY LAURENS the reason that when several are kept together in a single dish they will nip each other’s gills, legs and tails, rendering them- selves unfit for experimentation. The larvae were tested in groups of ten, and each individual in a set of tests was given in order a single trial, until each one of the ten had been tested once. This was then repeated until each had had ten trials. Reaction time 2 to 5 seconds, average 3.14 seconds Workers tested, 25 Only once were all the bees in a case excited Bee stings: 13 vibrated antennae and acted as if noticing odor 6 moved an inch or two until they were directly over the vial 3 stroked antennae and acted as if noticing odor 2 moved slightly and vibrated antennae 1 moved away quickly 1 turned around quickly as if noticing odor é 1 jumped quickly and vibrated antennae as if noticing odor 1 moved quickly, vibrated antennae, and followed odor when vial was moved beneath the case Reaction time 2 to 3 seconds, average 2.16 seconds Workers tested, 28 Smoke: During all these experiments it was absolutely necessary to refrain from smok- ing tobacco because the least smoke in the laboratory invariably excited the bees. They produced an uproar by rapidly vibrating their wings, became very restless, and remained in this unquiet state for some time after being irritated. The least amount of smoke from a bee keeper’s smoker, whether inside or near the outside of the house, also usually excited them. Queens Oil of peppermint: 3 moved away quickly 3 raised antennae and moved away slowly 2 moved away quickly and vibrated antennae 2 moved away slowly 1 vibrated antennae 1 arose and moved away Reaction time 2 to 7 seconds, average 3.8 seconds Queens tested, 12 Once all the workers in a case were excited Oil of thyme: 3 stroked antennae and moved to one side 2 arose quickly and moved to one side slowly 2 moved away slowly and vibrated antennae 2 vibrated antennae 1 moved away slowly and vibrated antennae 1 lifted antennae and arose slowly Reaction time 3 to 5 seconds, average 3.9 seconds Queens tested, 11 286 N. E. McINDOO Oil of wintergreen: 4 raised antennae and moved away slowly 2 moved away slowly and stroked antennae 2 moved away slowly and vibrated antennae 1 moved away slowly 1 vibrated antennae 1 arose and moved slightly Reaction time 3 to 7 seconds, average 4.2 seconds Queens tested, 11 Honey and comb: 4 vibrated antennae and moved away slowly 2 vibrated antennae 1 arose and moved slightly Reaction time 5 to 10 seconds, average 6 seconds Queens tested, 7 Pollen: 4 vibrated antennae and moved slightly 2 vibrated antennae 1 moved only slightly Reaction time 4 to 10 seconds, average 6.3 seconds Queens tested, 7 Leaves and stems of pennyroyal: 3 vibrated antennae 2 vibrated antennae and moved away slowly 1 vibrated antennae and turned around slowly 1 vibrated antennae and arose Reaction time 3 to 8 seconds, average 5.1 seconds Queens tested, 7 Drones Oil of peppermint: 14 moved away quickly 3 raised antennae and moved away quickly 3 arose and moved away quickly 2 moved away slowly 2 moved away quickly and vibrated antennae 1 raised antennae quickly Reaction time 2 to 3 seconds, average 2.3 seconds Drones tested, 25 Only once were all the workers in a case excited Oil of thyme: 14 moved away quickly 4 moved away slowly 3 moved slightly and vibrated antennae 3 arose slowly and moved away quickly 1 raised antennae OLFACTORY SENSE OF THE HONEY BEE 287 Reaction time 2 to 4 seconds, average 2.16 seconds Drones tested, 25 All workers in the case were excited only once Oil of wintergreen: 12 moved away quickly 6 arose quickly and moved away slowly 3 arose, stroked antennae, and moved away slowly 2 raised antennae quickly and moved away slowly 2 moved antennae slightly Reaction time 2 to 4 seconds, average 2.48 seconds Drones tested, 25 Honey and comb: 5 moved antennae slightly 4 vibrated antennae vigorously 2 raised antennae and turned around as if noticing odor 2 vibrated antennae vigorously and tried to reach the honey through the floor of the case vibrated antennae as if noticing odor arose slowly and vibrated antennae 1 placed antennae on floor of case and acted as if noticing odor 1 stroked antennae and turned around as if searching for honey Reaction time 2 to 10 seconds, average 3.8 seconds Drones tested, 19 Pollen: 5 vibrated antennae and moved slightly 2 vibrated antennae vigorously 2 vibrated antennae and acted as if noticing odor 2 raised antennae and moved away slowly - 2 arose quickly and stroked antennae 2 arose and moved away slowly 1 tried to get through the floor of the case 1 moved only slightly Reaction time 2 to 6 seconds, average 3.56 seconds Drones tested, 17 2 2 _Flowers of honeysuckle: 5 moved away quickly 4 moved away slowly 3 vibrated and stroked antennae 3 vibrated antennae as if noticing odor . 3 arose quickly and moved away slowly 1 raised antennae quickly and moved away quickly 1 arose quickly and turned around as if noticing odor Reaction time 2 to 5 seconds, average 2.8 seconds Drones tested, 20 288 N. E. McINDOO Leaves and stems of pennyroyal: 8 moved away quickly 6 raised antennae and moved away slowly 3 moved away slowly 3 vibrated antennae vigorously 2 arose and moved away slowly 2 arose slowly 1 arose quickly and stroked antennae Reaction time 2 to 5 seconds, average 2.74 seconds Drones tested, 25 Leaves and stems of spearmint: 6 moved away quickly 5 vibrated antennae and moved slightly 4 arose and moved away slowly 3 moved away slowly 1 raised antennae and moved slightly Reaction time 2 to 4 seconds, average 2.55 seconds Drones tested, 19 Once all the workers in a case were excited Leaves and stems of scarlet sage: 4 arose quickly and moved away slowly 4 vibrated antennae slightly 3 moved away slowly 3 raised antennae and moved away slowly 2 vibrated and stroked antennae and moved away slowly Reaction time 2 to 10 seconds, average 3.37 seconds Drones tested, 16 Summary To summarize the preceding data, it is found that 87 per cent of the individuals tested moved away from the peppermint, 80 per cent from thyme, 89 per cent from wintergreen, 6 per cent from honey, 29 per cent from pollen, 45 per cent from honeysuckle, 78 per cent from pennyroyal, 73 per cent from spearmint, 69_ per cent from scarlet sage, and 4 per cent from the bee stings. Thus, to all of these odors, except those of honey, pollen, honey- suckle, and the bee stings, bees react negatively. For the odors of honey, pollen, and honeysuckle the most characteristic re- sponse is in searching for the source of the odor by turning around over the vial with the head almost on the floor of the case and by vibrating the antennae vigorously. In response to the bee- .} OLFACTORY SENSE OF THE HONEY BEE 289 sting odor the bees usually moved quickly toward the source of the odor when the vial was moved 2 or 3 inches fromthem. By including the data from the use of all of these ten odors, except that of bee stings, the average reaction time for the very first responses of all the workers is 3.29 seconds, whereas the average reaction time for all the drones is 2.86 seconds. By using the data for the first responses to peppermint, thyme, wintergreen, honey, pollen, and pennyroyal, the average reaction time for all workers is 3.4 seconds, for all drones 2.9 seconds, and for all queens 4.9 seconds. Hence, in spite of the fact that drones are less active in the observation cases, they responded somewhat more quickly than did the workers and much more quickly than did the queens. It is evident, therefore, from these data that the olfactory sense in the honey bee is acute and that the sensitiveness to various odors is most highly developed in the drones, and least highly developed in the queens. EXPERIMENTS ON MUTILATED BEES IN OBSERVATION CASES In the preceding pages the behavior of unmutilated bees in ob- servation cases and their reactions to odors are discussed. The behavior of bees that have been injured in various ways will now be discussed, together with an account of the experiments upon them with various odors. In these experiments all the bees in the same observation case were mutilated and they were given the same kinds of food as those on which the uninjured bees survived. : ANTENNAE Entomologists now generally agree in the belief that the organs of smell in insects are located on the antennae. Bees with either the right or left antenna pulled off are much less pugnacious than are those with the antennae intact, and they “pay less attention”’ to each other. They appear otherwise normal, except that their ability to communicate is considerably decreased. They feed the queen and each other, eat normally, and often stroke each 290 N. E. McINDOO other as do those with both antennae intact. While those with both antennae intact lived on an average of 9 days and 3 hours in observation cases, the bees with one antenna removed lived under the same conditions on an average only 6 days and 18 hours. an Bees with one antenna pulled off and with 2 to 8 joints of the other one cut off never ‘“‘pay any attention’’ to each other and very seldom are seen fighting, but are just as apt to fight a hive- mate as a stranger. All of these eat more or less, but the more joints cut off the second antenna the less normal they appear. The greater the number of joints severed, the less number of days they live. Such bees, on an average, lived 5 days and 11 hours in the observation cases. Bees with both antennae pulled off at first wander about in- side the case more or less aimlessly. They seem to have no means of communicating with other bees. They run against each other and then try either to stroke or to fight one another, al- though they fail to do either. Only occasionally do they feed one another and only a few are seen eating. After a few hours they become inactive and usually remain so until they die. In their activity they are similar to young bees. They ‘“‘pay little or no attention” to other bees and in most cases do not move unless another bee runs against them. ‘The ones studied lived on an average of only 19 hours in the observation cases. Bees with both antennae cut off at their bases behave very similarly to those just described. Those observed lived only 2 hours on an average in the observation cases. When the antennae were varnished with shellac or celloidin the bees were not at all normal. They failed to eat; many ran around and acted ‘‘crazy”’ and tried to clean their antennae, but in this they failed. Those with shellac on their antennae lived only a few hours, while those with celloidin lived, on an average, 16 hours in the observation cases. The antennae of some were covered with vaseline. These bees at once cleaned their antennae and were soon normal. The left antennae of 15 workers were pulled off at the base with a small pair of forceps. The responses of these bees to the OLFACTORY SENSE OF THE HONEY BEE 291 odors of the three essential oils are similar to those described for bees with both antennae intact. The average reaction times are as follows: Oil of peppermint 5.2 seconds, oil of thyme 4.3 sec- onds, oil of wintergreen 4.8 seconds. The average time for 15 specimens with their right antennaggpulled off gave similar re- sults, as follows: Oil of peppermint 4.8 seconds, oil of thyme 4.1 seconds, oil of wintergreen 4.6 seconds. The average time for workers with one antenna pulled off is: Oil of peppermint 5 seconds, oil of thyme 4.2 seconds, oil of wintergreen 4.7 seconds. This gives a general average of 4.6 seconds for the three oils, while the general average of the same oils for bees with both antennae intact is only half as much, that is, 2.3 seconds. Ac- cording to these results, and on the theory that the antennae contain the organs of smell, it would at first glance seem that each antenna plays an equal part in receiving odor stimuli and that when either antenna is absent the average time for the responses is consequently doubled. Bees with either the right or left antenna pulled off and with from 2 to 8 joints of the remaining flagellum cut off gave the following average reaction times when using the three essential oils: seconds “2 Gaui BNUSfS) CEES ST Ye ae es ch eS cea BA AS 15 2k (CTRGIS) TTS) 0 a Rene oS Sota es i rR Pea eae ey Gi dt £5) TOMUaNS} Tat ace een PES cls.t. ic Gitaeta caetaeeelis erases & Aes ner see eae ere eae 56 DR OLMUSEINISSUMN OMe .ch.1 Ree E rachis eae creeeut- tye yoigicius Sears: 2 Saks 27 FmOMMiS MISSIMPen on. +l lene eee FN ee ce Ke Nao «sist 98 SUMMERS SIME Pas cas So ree oes Sie s eat dee ate OO Thus, in general, the greater the number of joints of the sec- ond antenna removed the longer it takes the bees to respond to odor stimuli. Workers thus mutilated are not at all normal and live on an average only 5 days and 11 hours. Bees with both antennae either pulled off or covered with cel- loidin entirely fail to respond to the essential oils. They also are quite abnormal in their behavior. Two drones, one of which had 4 and the other 5 joints of one flagellum missing, were taken in this condition from a_ hive. 292 N. E. McINDOO These were apparently normal in other respects, but when tested with 6 of the 10 odors used in these experiments they gave an average reaction time of 3.16 seconds, whereas the average for the same odors with unmutilated drones was 2.9 seconds. Ninety middle-aged bees,with both antennae pulled off were placed on the combs of the observation hive at 4 o’clock in the afternoon. To prevent them from flying from the hive, their right wings were clipped. These bees were more restless than normal bees. They wandered about and continually crawled out of the hive. Whenever one was found outside the hive it was put back, although all of them eventually escaped in this way, because two days later not one of them, either alive or dead, was found in the hive. The behavior of these mutilated bees was similar to that of the antenna-less bees in the observation eases. They took no part in the activities of the hive and ap- peared to have lost all means of communicating with the other bees. The next day at various hours 16 of them were found, dead in frontof the hive. The third day at 8 o’clock.3 more were found, and on the fourth day at 4 o’clock one more. In three days 20 were found dead, while the other 70 certainly crawled out of the hive and escaped. Counting the time until their dead bodies were found, these 20 lived on an average only 21 hours, while in the observation cases the average length of life of such bees was 19 hours. From these combined results it is evident that bees with their antennae pulled off are not nor- mal and therefore whatever results are obtained by experiment- ing with them must always be discounted. Immediately after pulling off both antennae of a worker, it was placed on a comb with the other bees in the observation hive. A small piece of cotton wet with oil of peppermint was held 3 inch in front of this antenna-less bee, and afterward smoke from a bee keeper’s smoker was gently blown on it. In performing these experiments with 10 different workers, not a single mutilated bee reacted in the least, whereas all the other bees soon fled from the oil of peppermint and caused an uproar by rapidly vibrating their wings when smoke was gently blown on them. Similar experi- ments were performed, by using oil of thyme, clove, wintergreen, OLFACTORY SENSE OF THE HONEY BEE 293 and cedar and the same bees were afterwards tested with smoke. Not one of 50 mutilated bees so treated showed any reaction, while the normal bees never failed to react. The mutilated bees in these experiments were restless and wandered about considerably. They often crawled into cells. Occasionally one cleaned itself and sometimes other workers cleaned them; very often a mutilated worker fed one to three other bees before it crawled off the comb and disappeared. How long these mutilated bees lived can not be stated because they were soon lost among the other bees. Three or four hours afterward two or three of them were found lifeless in front of the hive. Miss Fielde (’03 a) believes that the olfactory organs which an ant uses in recognizing enemies lie in the fifth and sixth an- tennal segments. To ascertain if one or more particular anten- nal joints of bees bear olfactory organs which are used in recog- nizing strange bees, 3 to 8 joints were cut from both antennae of 3 lots of 9 middle-aged bees each. These mutilated workers were introduced into an observation case and their behavior was studied. No abnormality in behavior was noticed among them except that they were slightly less active. Occasionally when first introduced one attacked another, although not seriously, and no injury was ever done to any of them. They lived scarcely 9 days on an average. Nine bees with 2 to 8 joints of both antennae amputated were introduced into an observation case. The following day several strange bees, one at a time, were put into the same case. Most of the mutilated bees at once noticed a stranger and in a few seconds one of the strange bees was attacked. Both fighting bees were immediately removed from the case, the stranger was dis- carded, while the mutilated bee was killed and its remaining an- tennal joints counted. This experiment was repeated seven times. In all, 72 mutilated bees and as many uninjured ones were used. In only a few instances did the mutilated bees at- tack each other. This can always be prevented by keeping the bees isolated for a short time after the operation. When strange bees were put among the mutilated ones almost one-half of the 294 N. E. McINDOO former were attacked. When the remaining antennal joints were counted it was found that in the pugnacious bees from 2 to 6 antennal joints were missing and in those that did not attack the strangers from 2 to 8 joints were absent. Only the last 8 joints of the antennae of a worker contain the supposed olfactory organs and when these 8 joints were removed all of these organs are eliminated. According to Miss Fielde’s theory the fifth and sixth antennal segments of the worker bee would be the ones which carry the olfactory organs by which strange bees are rec- ognized. Furthermore, Miss Fielde claims that the tenth an- tennal joint of an ant contains the organs through which the colony odor is recognized and if this jot is removed colony- mates will attack each other. Judging from the foregoing ex- periments, no particular antennal joint of a worker bee contains the organs by which the odor of sister bees is received, because the mutilated bees never fought each other regardless of the number of antennal segments amputated, when they were kept out of the cases for a few minutes after the antennae were injured. To determine whether bees with mutilated antennae are nor- mal, 11 joints of both antennae of 12 middle-aged bees were cut off. When introduced into an observation case these bees did not fight, ‘‘paid no attention” to each other, were quiet, and only one ate candy. When strange bees were put among them the mutilated bees always gave the strangers the right-of-way and did not attack them. They lived only 6 hours on an average. Both antennae of 95 middle-aged bees were burnt off with a red-hot needle. These workers were also abnormal and lived only 17 hours on an average. The antennae of 30 drones were similarly burnt off and they too lived only 17 hours on an average. Thirty middle-aged bees were immersed in water for 15 min- utes. When removed they appeared entirely lifeless and their antennae were pulled off at once. They revived and lived only 19 hours on an average. Since bees whose antennae are mutilated after they become adults are abnormal, the following experiments were performed with immature bees. A frame of brood was removed from a hive, the cells of sealed brood were gently uncapped and both OLFACTORY SENSE OF THE HONEY BEE 295 antennae of 100 pupae ranging in age from 14 to 18 days from the laying of the egg were carefully cut off near the head. Great care was exercised not to injure the immature bee in any way other than by the amputation of these appendages. The cells were again closed with a very thin layer of beeswax, a wire-cage screen was placed over these uncapped cells so that the adult workers could not pull the mutilated bees out of their cells, and the frame was put into the hive. In a few days these bees began to emerge from their cells. They were removed from the hive and placed in observation cases with intact middle-aged bees. All bees thus mutilated were abnormal and lived as adults about 5 days on an average. In like manner both antennae were cut from 300 pupae 13 and 14 days old from the laying of the egg. In this instance a wire screen was placed over the entire frame so that the mutilated bees could not escape when they emerged. Seven days later they began to emerge and most of them in a short time crawled from their cells. They soon mixed on this comb with other young bees that were unmutilated and they had plenty of food but they were all dead 5 days later. Seven of the workers with their antennae burned off recovered from the operation sufficiently to eat candy and to move about freely, but they were far from being normal. These lived from 1 to 11 days. When tested with the three essential oils, pep- permint, thyme, and wintergreen, they responded readily. Their general response was to move slightly and vibrate the stubs of their antennae; one rubbed a leg against the abdomen and 3 moved their heads quickly. Average reaction time for oil of peppermint 3.5 seconds, for oil of thyme 4.5 seconds, for oil of wintergreen 4.3 seconds, and for all three oils 4 seconds. Fre- quently, when the antennae were cut or burnt off, the insects were placed on the table and tested with these odors. They often moved away from the odors but generally did not react to odors or anything else and often scarcely moved even when touched with a pencil. In previous experiments bees cleaned off any substance put on their antennae. To prevent this the tarsi of their front legs 296 N. E. McINDOO were burned off with a red-hot needle, removing their antennal cleaners. About one-fourth of the bees so mutilated died with- in 12 hours, but the remainder appeared quite normal in every other way. On the second day eleven joints or the entire fla- gellum of each antenna was coated with liquid glue. Since the antennal cleaners were removed, these bees could not remove the thick coating of glue. They were quite abnormal and most of them did not live long. However, after gluing the flagella of many bees, 21 were finally obtained that were fairly normal and they responded to the odors of the above oils without failure. The general response was either to arise and move away quickly or to vibrate their antennae. Often when one had fallen down apparently lifeless the odor was placed under it; it arose almost instantaneously and moved away quickly. After 3 or 4 days all the surviving bees had succeeded in cleaning their antennae but most of them died before this time. However, some lived 12 to 14 days. The average reaction time to oil of peppermint was 3.1 seconds, to oil of thyme 3 seconds, to oil of wintergreen 2.8 seconds, and for all three oils 2.9 seconds. To ascertain if the odor of the glue itself affected these re- sults in any way, glue was placed on the top of the thorax. This dried in a few minutes and the bees were not able to get it off. They were entirely normal and were tested with the foregoing odors. The average reaction time to oil of pepermint was 2.4 seconds, to oil of thyme 2.8 seconds, to oil of wintergreen 3.1 seconds, and for all three oils 2.7 seconds. The workers tested were 19 in number. From all the experiments performed on the antennae of bees it is evident that when these appendages are mutilated in the slightest degree, the bees are never entirely normal, even though they apparently recover from the effect of the shock. ‘The greater the number of joints removed or covered, the greater | the abnormality. Furthermore, there is no.reason to assume the presence of the olfactory organs on the antennae, because when these appendages were burned off, the general average of the reaction time for the three oils was 4 seconds; when glued, 2.9 seconds; and when unmutilated, 2.6 seconds. This slight in- OLFACTORY SENSE OF THE HONEY BEE 297 crease of time can certainly be attributed to the abnormal con- dition of the bees, and these results indicate that the olfactory organs are located elsewhere. At most it can be claimed only that the antennae may assist in the receiving of odor stimuli. MAXILLAE AND LABIAL PALPI Various observers claim that the palpi of insects are the seat of the olfactory organs. In the bee the maxillary palpi are al- most wanting, but the maxillae and labial palpi possess pore-like sense organs. These appendages were cut off at the base. When introduced into observation cases, bees so treated appeared quite normal in all other respects and when strange bees were put among them they lost no time in attacking the strangers. Never- theless, they certainly were not completely normal, for they lived only from 3 hours to 4 days, with 24 hours as an average. When tested with odors, they gave responses similar to those of unin- jured bees. The reaction time was as follows: Oil of pepper- mint 2.2 seconds, oil of thyme 3.7 seconds, oil of wintergreen 4 seconds, honey and comb 4.4 seconds, pollen 5.8 seconds, and leaves and stems of pennyroyal 3.9 seconds. This gives a gen- eral average of 4 seconds, whereas for the same odors with un- mutilated bees the general average was 3.4 seconds. PROBOSCIS Several writers have described sense organs on the proboscis of insects. To determine whether these have an olfactory use, the proboscides of 36 workers were cut off close to the base. These bees seemed quite normal in most respects and many of them even tried to eat candy, but of course they could not accom- plish much without this appendage. They lived only 7 hours on an average. The average reaction time to oil of peppermint was 2.6 seconds, to oil of thyme 3 seconds, and to oil of winter- green 3.2 seconds. This gives a general average of 2.9 seconds, while for the same odors with unmutilated bees the average was 2.6 seconds. Twenty-two workers were tested. We can probably attribute this difference of 0.3 seconds to the abnor- mality of the mutilated bees. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 3 298 N. E. McINDOO MANDIBLES Janet (’11b) describes a sense organ in the mandible of the honey bee which he thinks may have an olfactory function. To ascertain this experimentally, the mandibles of 30 workers were amputated close to the base. These bees appeared completely normal and did not fight each other. They lived from 5 hours to 15 days, with 7 days as an average. The average reaction time to oil of peppermint was 2.6 seconds, to oil of thyme 3.3 seconds, to oil of wintergreen 4.6 seconds, to honey and comb 7.7 seconds, to pollen 6.2 seconds and to leaves and stems of penny- royal 4.7 seconds. These give a general average of 4.8 seconds whereas the average for the same odors with unmutilated bees was 3.4 seconds. We may attribute this slight difference in time either to the injury caused by the amputation, or to the fact that the mandibles help to perceive odors or to both. The bees tested were 20 in number. BUCCAL CAVITY Huber (1814) states that the seat of the organs of smell in the honey bee is the buccal cavity, whereas Wolff (1875) discovered some sense organs on the epipharynx of the same insect. The epipharynx lies in the mouth cavity and forms a portion of its roof. ‘To determine whether this cavity has anything to do with olfaction, Huber’s experiment of filling the mouth cavity with flour paste was repeated. With the aid of a small pencil brush the mouth cavities of 25 workers were thus filled. When the paste had become perfectly dry, the bees were put into observa- tion cases. They seemed otherwise entirely normal and all tried to eat candy although some were unable to move their probosces on account of the hard paste. They lived from 2 hours to 16 days, with 74 days as an average. However, by the end of the fourth day the paste had come out of their mouths. The aver- age reaction time to oil of peppermint was 2.55 seconds, to oil of thyme 2.75 seconds, and:to oil of wintergreen 2.75 seconds. This gives a general average 2.68 seconds. The average for OLFACTORY SENSE OF THE HONEY BEE 299 the same odors with normal bees was 2.64 seconds. Twenty workers were tested. It would seem that neither the buccal cavity nor the epipharynx has anything to do with olfaction. MORPHOLOGY OF OLFACTORY PORES In 1857 to 1860 Hicks described for the first time some pecul- iar structures found on the bases of the wings and legs of insects. He called them ‘vesicles’ and suggested that they have an olfac- tory function. From the following description of the morphology of these structures it will be seen that the word ‘vesicle’ is less appropriate than the word ‘pore.’ Since they have an olfactory function in the honey bee (pp. 333-341) they may be called ‘olfactory pores.’ DISPOSITION In studying the distribution and number of the olfactory pores of the honey bee, workers and drones just emerged from the cells were selected on account of their lighter color, but since queens at this stage were not available at the right time, old dark-pig- mented ones were used. The three legs from one side of 6 individuals each of workers, queens, and drones and the three legs from the other side of 9 of these 18 bees were examined under a high-power lens. The wings from both sides of 8 workers, 8 drones, and 4 queens and the stings of 15 workers and of 9 queens were likewise examined. In all, 81 legs, 80 wings, and 24 stings were searched for pores. The wings have upper (dorsal) and lower (ventral) surfaces and the legs have inner and outer surfaces. The inner side of the leg faces the bee’s body, and the outer side is directed from the body except in the case of the front legs, which are directed forward and the sides are therefore reversed. These pores are found in groups, and for convenience in studying and describing them the groups on each side may be numbered from 1 to 21. The first five of these groups are found at the bases of the wings, groups 6 to 18 inclusive on the legs, and the remaining three groups on the sting (figs. 1, 2, and 14, A). The numbers of the 300 N. E. McINDOO ABBREVIATIONS 1A, 3A, first anal vein, third anal vein AgCyt, deeply staining cytoplasm Art, articulation ArtCh, articular chitin ArtM, articulation membrane 1Az, 2Ax, 3Ax, first axillary, second axillary, third axillary AxC, axillary cord BC, body cavity BGI, alkaline gland of sting BlCor, blood corpuscle BlSin, blood sinus Brb, barb BrbHr, barbed hair C, costa Ch, chitin ChL, chitinous layer ChM, chitinous marking ChTh, chitin of thorax Con, cone _ConnT, connective tissue Cu, cubitus Cyt, cytoplasm FFI, Forel’s flask or sensilla ampullacea Gv, groove Hr, hair Hyp, hypodermis HypCW, hypodermal cell wall HypNuc, hypodermal nucleus HypNuc', hypodermal nucleus that has formed a hair IXS, thick membranous lobe that over- laps sting Let, lancet M, muscle MB, muscle bundle MBNuc, nucleus of muscle bundle Md, median plate MFI, mouth of flask N, nerve NB, nerve branch Neu, neurilemma NeuNuc, nucleus of neurilemma NeurNuc, neuroglia nucleus NF, nerve fiber NkFI, neck of flask Ob, oblong plate of sting Pg, peg, club, or sensilla basiconica Por, pore PorAp, pore aperture PorApHr, pore aperture of hair PorB, pore border PorBHr, pore border of hair PorCl, pore canal PorHr, pore of hair PorPl, pore plate, pore canal, orsen- silla placodea PorW, pore wall PorWHr, pore wall of hair PPgq, pit pegs, champagne-cork organs or sensilla coeloconica PsnCl, poison canal of sting PsnSc, acid poison sac of sting Qd, quadrate plate of sting R, radius R-+M, radius and media united Sar, sarcolemma SarNuc, nucleus of sarcolemma SC, sense cell SCG, sense cell ganglion Scl, scutellum of mesotergum SCNuc, nucleus of sense cell SCNuc', nucleus of sense cell without cell wall SCNucl, nucleolus of sense cell SF, sense fiber Sh, shaft of sting ShB, bulb of shaft of sting SpHr', 2, 3, three varieties of spinelike hairs StnPlp, palpus of sting 1Tar, 2Tar, 3Tar, 4Tar, 5Tar, firstto fifth tarsal joints Tg, tegula THr, tactile hair or sensilla trichodea Tn, taenidia of trachea Tra, trachea TraNuc, nucleus of trachea Tri, triangular plate of sting 1 to 21, groups 1 to 21 of the olfactory pores OLFACTORY SENSE OF THE HONEY BEE 301 AUENTA 42 Clearer 4 Trocharnsrer’ // Fig. 1 Diagram of ventral view of a worker bee, showing the location of the different groups of olfactory pores as indicated by the numbers. groups on the legs apply’ to similarly placed groups on all the legs, but the groups on the front and hind wings are seemingly not homodynamous and are given different numbers. If a group has been found to be always similarly placed and always 302 N. E. McINDOO BOE: oy iy e el Bie: 28 ( ee Cee (E Ca \ Fig. 2 Diagram of dorsal view of a worker bee, showing the location of the different groups of olfactory pores as indicated by the numbers. to have the same general shape, it may be regarded as constant and if the pores are close together it may be regarded as definite; otherwise it may be considered as inconstant and scattered. OLFACTORY SENSE OF THE HONEY BEE 303 As the worker bee is most suitable for study, the pores on the wings, the third leg, and the sting of specimen No. 15, which has been drawn (figs. 1 and 2), will be described in detail. The differences on the other legs of this specimen and the variations found in the three castes will then be discussed. Fig. 3 Ventral view of base of front wing of a worker bee, showing groups 1 and 2 of olfactory pores. X 45. Fig. 4 External view of group 1 of olfactory pores of a worker bee. X 465. Fig. 5 External view of group 2 of olfactory pores of a worker bee. X 465. Groups 1 and 2 (figs. 1 and 3) lie on the lower surface of the front wing. Group 1 occupies almost a central position on the subcosta, while group 2 lies near the anterior and distal margin (fig. 3, Sc.). These groups are somewhat triangular in shape, with their apices facing each other, and the concave edge of group 1 always faces the posterior margin of the subcosta. A light band (represented by a line in fig. 5), surrounds the distal 304 N. E. McINDOO end of group 2. In groups 1 and 2 there are 281 and 50 pores respectively; in the former group (fig. 4) the pores are close to- gether while in the latter (fig. 5) they are more scattered. The pores vary considerably in size; in group 1 most of them are small, with the diameter of the largest three times that of the smallest, while in group 2 almost all of them are comparatively large, with the diameter of the smallest one-half that of the largest. In group 1 scarcely any rows are discernible, but in group 2 the rows are more sharply defined. The openings (PorAp) in the pores are usually round, although several are oblong, and the long diameter of the oblong pores is more or less parallel with that of the group. Group 3 is found on the upper surface of the front wing (fig. 2) and occupies nearly the entire surface of the median plate (fig. 6, Md), leaving only a narrow margin on all sides. This group is long and slender, with its tapering end pointing toward the proximal and posterior margin of the subcosta (Sc). Its pores are considerably scattered, although they le in more or less definite rows (fig. 7). The pores are about equal in diameter and the diameter of the largest is never more than twice that of the smallest. In this group there are 174 pores, nearly all of which have round apertures. Group 4 is present on the lower surface of the hind wing (fig. 1) and covers most of the anterior half of the union of the radius and media (fig. 8, R + M). Its distal end is the more pointed and its pores are arranged in irregular rows (fig. 9). There are 83 pores of about equal size. Group 5 lies on the upper surface of the hind wing (fig. 2) and occupies nearly all of the surface where the radius and media unite (fig. 10, R + M) and extends slightly over the subcosta (Sc). The proximal and narrower end of the group points di- rectly toward the proximal end of the subcosta. There are 209 pores, which can scarcely be said to lie in rows (fig. 11). Most of the pores are small, with the diameter of the largest three times that of the smallest. Group 6 is a double group having the shape of a figure 8, located at the proximal end of the femur (fig. 1) on the outer OLFACTORY SENSE OF THE HONEY BEE 305 Fig. 6 Dorsal view of base of front wing of a worker bee, showing group 3 of olfactory pores. X 45. Fig. 7 External view of group 3 of olfactory pores of a worker bee. The dis- tal end with two hairs is placed beneath the remainder of the group. X 465. Fig. 8 Ventral view of base of hind wing of a worker bee, showing group 4 of olfactory pores. X 45. surface near the posterior margin. The smaller portion of group 6 lies nearer the posterior edge of the leg (fig. 12, A) and has 9 pores, none of whose openings lies with its long diameter exactly transverse to the long axis of the leg. The other portion of the group has thirteen pores, and the long diameter of most of its openings are exactly transverse to the long axis of the leg. The diameter of the largest pore is about twice that of the smallest. 306 N. E. McINDOO G ¢ ©) ao® : Por Str o. 5 ©) Cr : =) me a Tee, ag. 08,6 2° © 2 OS CO OH O eo @ eS) O ©) oS ‘ORCMEC) a = Be. 3? oo® @ 68, o Coo ©6 © ood © @ Sec L ‘ 36 @® G §68 660,6° 6 66° {+ %o 56 0 6,8qe See iT] Fig. 9 External view of group 4 of olfactory pores of a worker bee. X 465. Fig. 10 Dorsal view of base of hind wing of a worker bee, showing group 5 of olfactory pores. X 45. Fig. 11 External view of group 5 of olfactory pores of a worker bee. X 465. The majority of the pores on the legs are slightly oblong and have openings which as a rule have their long diameters more or less transverse to the long axis of the leg. Groups 7, 8, and 9 lie on the outer side of the trochanter (fig. 1). Group 7 has 16 pores and lies at the distal end near the ‘anterior margin, group 8 extends along the anterior margin from group 7 almost to the articulation of the trochanter with the coxa, and group 9 extends from the proximal end along the medi- an line two-thirds the distance to the femur. In all three groups the diameter of the largest pore may be three times that of the smallest. The pores of groups 7 and 9 (figs. 12, B and D) have their long diameters directed obliquely across the leg, while in OLFACTORY SENSE OF THE HONEY BEE 307 Fig. 12 External view of alee pores as they appear on the third leg of a worker bee, X 465. A is group 6; B, group 7; C, three olfactory pores and a hair pore (PorApHr) from group 8; D, three olfactory pores and a hair pore from group 9; E and F, a few olfactory pores and hairs from groups 10 and 11; G, six olfactory pores and a hair pore from group 12. group 8 (fig. 12, C) most of the long diameters are parallel to the long axis of the trochanter. Groups 10 and 11 are found at the proximal end of the nas (fig. 1) on the outer side. Group 10 is near the anterior edge and group 11 near the posterior. The pores of these groups are 308 N. E. McINDOO Fig. 13 External view of olfactory pores and some spinelike hairs as they appear on the third leg of a worker bee, X 465. A, from group 13; B, from group 14; C, from group 15; D and E from groups 16 and 17 respectively; F, from group 18, the two pores are drawn twice too close together. the largest of all and they do not vary greatly in size. They are oblong and the long diameters are in some cases oblique and in others transverse to the long axis of the tibia (fig. 12, E and F). They lie in rather straight rows, along the long axis of the leg. Groups 12 to 15 lie on the inner side of the trochanter (fig. 2). Groups 12 and 13 are located at the distal end, with group 12 near the posterior edge and group 13 near the anterior. Group 14 lies on the inner side in a position near the anterior edge simi- lar to that of group 8 on the outer side. Group 15 lies near .the posterior margin extending the full length of the trochanter (figs. 12, G and 13, C). In these preparations occasionally one end of the slits extends outside the pore wall (PorW). ‘This may be due to the treatment with caustic potash. a, @ 42 a OLFACTORY SENSE OF THE HONEY BEE 309 Groups 16 and 17 (figs. 2, 18, D and E) are similar to groups 10 and 11 except that they are on the inner side of the proximal end of the tibia. Group 18 consists of two or three pores as large as those on the tibia, which are found on the ventral side of the second and third tarsal joints (fig. 2). These two pores and three hairs are represented in figure 13, F, although in the drawing the relative distance between the pores is reduced one-half. Group 19 is located on the shaft of the sting (fig. 14, A) and its pores are greatly scattered, but most of them occur about midway between the two ends. | 2.76 19 Oi ian Vaseline on abdomen as control.............) 2.78 18 97 is Bagel armament eee eet oie einer) te - |) =420 7 Oo ran Plheellavelmed inet cee gees 5 a haln eee): 2.9 21 Pay Maxillae and labial palpi cut off............ 3.3| 4.0 | 19 LO IPEQHDOSCISICUUKOLLAE Rep omieey- pier dente rene fans = 2.9 22 Ones Memndiblesteubroiire naam rete ere ocr tor | 3-9 |, 4.85) 20 Fs al ae Flour paste in mouth............. Pees RC cian: | 2.68 20 TG Teles Wings cut off beyond pores...............-. 320 17 9 | 23 Stingsrexthacued SYS ~~ N= 1d NA ei ri ri -— Cre Ig uve G16 POWERS H. MITCHELL AND J. Cc. 364 SYUVNAY OOT OOT 66 Ri) ING) udd 6 Ri) LINGO udd oon OS Oo eS fx} To) (=) cI es Mi fT | oooxn on ot SHLVad TO YaaWoNn OL OL I 8 | 3) 8 CT je 39 il ¢ G I 8 g 6 6 0) i 8 V | I ¢ Or | G 6 Ms G ¢ OT | I IT 14 | ‘S 9 0 I i g € CT 0 i € G | I 9 i i FI G | € - 61 I I SI G € ia! 0 I I | ¢ | I GT G I 9 | OT I OL 9 I € g I | xtmva 66 6L | NI LN@uvd | Ri) ao 4o STva ugdaWwon ugdqaWwon penuya0op—t ATaVL UGAHWON |-dIAIGNI JO} | “ICY I ; q Ay | HG MBAR ZI16I ‘NGD ONI | -MOTION wo | | LNGuvd a0 | | HLYIA AO ALva | 0G LI 8T Lal 1G 61 8 qh! &1 6 Ol LZ Gl g Sh 4148 LT | [ “Ie 1G 8z 9T GG 8 &G Ol 0G Cie ei Se er Ol ial SI él &6 Or 0G 8 Si | @ te 4 1G Te Geel 91 O& 91 8Z | 8 LG Us Cl6T DNOOXK DNOOA 10 asuld 40 uadWwaNn HLYId JO ALVa Saluas | NI ‘NAD fo uaqdnWon — ee 365 IN ASPLANCHNA AMPHORA INDUCED CHARACTERS | | | | | | SoCal | 518 «= ica 0 ie) ) |= UO ieam Se) al) aaa 9 9 H DOs | GROW Y DieeiZs elerse ase 30 I g Or G H Ci Me eRe Ci sll gee iu I b IZ 7 H gg Sy, lee (Oh ae Cl ae eee I g ZG g “OSIp FS0Y_ | =H €F Lg 0 € ) iz p Z Avy ! T Avy H 19 GES al FI 6 el es 0G 08 H cg G9 0 L | &I Pe be 02 =| 66 H GZ cL 0 ¢ GT ie ie OL 0z 9% H € 19 I ¢ Ol lie oc 91 G H 8L ru 0 p | °% lb Stolen ee 6 re H'I'S 001 0 0 ¢ 0) | ¥ IZ ¢ 61 Ieee. “68 19 OF) 2 fui I LT le sea LT H (050) Se) hae ORs | OT cele) I 91 Or 91 Hae wieOoy Pesce hice Tr =) FOr. KOT I FI 1% FI H 0) ie aed hm 0 ae facet gil Vetus I T SI aT H oh 8¢ | z 8 II I ZI IZ ‘a H Pg OF ome Cle fe le =e eT O CZ OL H gI GS alia, ne Piers 9 6 we 8 H SF og 0 (is anit La lee 13 L H 68 1 to 8 ail Cee vlpeg 6 p H oT Gg ee taal) ela Ie Palone H cI G8 GUE eae ee]! ane eatdly 0G Po Tidy H ee Bes sOm aes Ol g 08 Se (0s H SP Gh == 10 Ol II I 6G 1% | 66 H OF Uo Oh a aN ee) linea I 8Z 0G | 8 H OSes ic OR a) =. 0 L L LE 92 val 9% [gl 2 2 egal | SAR Sees igs 9 ese fees Soa NMI fai | Qt | Fe | Z8 a, | mice | z eat red rea rad H 89 a A a I Gn ees 055° |. 02 i 09 Ee i) % 6 9 age GL “AVIN Lh | Sie a8 366 C. W. MITCHELL AND J. H. POWERS The histories of J and K, on the other hand, are very different and were a great surprise to us at the time the experiment was conducted. The transition from the saccate to the humped type which was so tardy and transient in H and J, appeared in K promptly in the second generation and in J in the third. More- over, when once produced the humped type became singularly constant throughout almost. the entire series. But once in J and three times in K did single generations occur in which a part of the individuals only reverted to the saccate type; yet, though no preference was shown in the selection of humped individuals to continue the series at these times of reversion and although no change of food was made, the humped type was immediately resumed in the next generation. This constancy of the higher type was in striking contrast to our entire previous experience in the rearing of pedigree series of this rotifer. While in mass culture the humped type is readily main- tained yet in no other instance of a long continued series of iso- lation cultures had it been possible to avoid frequent reversions to the smaller saccate type. The only instance where this was done for a score or more generations? was where a special food stimulus was applied. That our success with J and K was not due to food supply employed is proven by the fact, which we have already suggested, that the copious food supply was taken from the same source as was that fed to H and J and was given to all in excess of consumption. In size J and K were of course larger than H and I, being of the humped type; the average number of young pro- duced per individual approximated closely to fourteen in J and to fifteen in K. A study of the metabolic rhythm of the four series, which may be made out by noting the number of young present in successive generations, shows the relative independence of the four series in regard to each other, and also in regard to external conditions. Their crescendos and diminuendos are not coincident. Never- theless, a closer approach to parallelism obtains between J and K, on the one hand, and between H and J, on the other, than obtains between any two of the contrasted series. ? See article I, table 10, history of Series D. INDUCED CHARACTERS IN ASPLANCHNA AMPHORA 367 In order to test further the nature of H and J in regard to their latent potentiality to produce the higher type, derivative mass cultures were started from time to time, it having been proven by all our previous experience that mass cultures favor the pro- gressive mutation to the larger types. Irregular food supplies were purposely resorted to in these cultures as favorable to the transition. Transitions did indeed occur to the humped type, and in a few instances cannibalistic campanulates also made their appearance. This latter transition, however, was far less frequent than in parallel mass cultures derived from J and K. The natural conclusion suggested by the conduct of the above series is that we do have the transmission of experimentally induced factors through the sexual gametes and the resulting resting egg. While this result is not shown morphologically in the generation immediately derived from the zygote yet it would seem to plainly exist as an inherent factor, the factor which we have termed the physiological potential. In order to accumulate further evidence for the fact of this transmission and also to ascertain whether such transmission is, as it were, a uniform process or whether it may perhaps be variable or cumulative, or regressive, we continued the rearing of further series derived from resting eggs produced by the same stocks. As the labor of rearing these stocks is very considerable and we wished to mass abundant proof upon the points tested we choose to limit the work to inbred series only. Series derived by inbreeding individuals of I and J It may be first explained that owing to the diverse inherent tendencies of reproduction it was easy to obtain resting eggs from humped series like J while it was a matter of much difficulty to obtain them from saccate series like 7. It was thus not feasible to begin a new parallel series of these two types simultaneously. Adequate centrols however were always conducted. ‘To test the inheritance of the high potential of J new series were derived as follows: On January 15, during the twenty-first generation of J, a mass culture was obtained from the fourteenth daughter of the 368 C. W. MITCHELL AND J. H. POWERS mother in the pedigree line. This culture was treated according to our regular method for inducing male production.*? The effect was quite successful. Resting eggs followed. On January 20 sister eggs from one individual were isolated under conditions favorable for hatching, which began by February I.— Ten of these young were isolated in cultures normal as to food supply and to all other known conditions. These ten individuals became the parents of ten series which, for brevity, we designate collec- tively as J 2. The effort was made to parallel J 2 with series derived from the series H or J, but as already mentioned, difficulty is always expe- 3JIn an article appearing in Science, vol. 38, pp. 786-788, A. Franklin Shull, in discussing this method of determining sex or male production, offers certain objections based upon his study of Hydatina senta. In regard to the relation of physiological rhythm, nutrition and male production, we should state that since the last two may be directly controlled or modified by experimental conditions, and since the first may also be modified, though not by direct methods, we have a means whereby we may govern this relation at will. That Shull has healthy lines which pass through long periods of parthenogenetic female production, is not surprising. We have had parallel cases; for instance, one of our healthy lines had given rise to a continuous parthenogenetic female production for over sixty generations. However, conditions were clearly not such that male production was possible, but later when conditions were altered to those favorable to male production, this hitherto total female-producing line suddenly threw a large percentage of male producers. That this was due en- tirely to some unknown internal factor would seem all but improbable. But even accepting this, would it he!p in explaining results which have been obtained? In our study of sex determination we have attempted to emphasize that the individual is the “‘point of action’ rather than a number of generations. A propos to this we would call attention to those cytological facts obtained by Erlanger and Lauterborn (R. Zool. Anz., Bd. 20), Lensen (Zool. Anz., Bd. 21), and Jennings (Bull. Mus. Comp. Zool., Harvard Coll., No. 30), upon Hydatina and Asplancha which tend to prove that during the period of maturation sex is determined by the casting off of the polar bodies. The reduction in the num- ber of chromosomes occurring when the second polar body is cast off, results in male production, while in ova which cast off only one polar body there results female production. The time at which this maturation occurs cor- responds to the time at which the factors we advocate are active in the deter- mination of sex. It would be rather difficult to correlate this evidence, it seems to us, with definite internal causes or to chemicals which act only upon the preceding generation, as is offered in explanation of sex determination by Shull. INDUCED CHARACTERS IN ASPLANCHNA AMPHORA 369 rienced in obtaining resting eggs from the saccate type. However, a mass culture started on January 29, from the fifty-third genera- tion of J and remaining purely saccate, produced by February 6 resting eggs. These began to hatch by February 12. Unfor- tunately, at this time the labor of caring for the enormous pro- geny of the ten lines of J 2, as well as of the original series, ren- dered it impossible to establish more than a single additional series. This, started from one of the before-mentioned resting eggs drawn from the saccate series, J, we term series J 2. It thus. forms a parallel or contrast series to the ten lines in J 2. Other individuals derived from sister eggs, as well as from other resting eggs produced at the same time in the same culture, were set aside to form mass cultures. useful in a general way as checks or con- trols. The history of J 2 as well as of the contrasted series J 2 are given in table 2. It will be seen from table 2 that seven of the series in J 2 were bred for ten generations, the other three succumbing to total male production after six to eight generations. J 2 was bred to thirty generations. It will be noted that again sexual reproduction fails to wipe out the induced but hereditary tendencies of the contrasted series. Although conditions were in every way parallel the transition from the saccate to the humped type occurred early in each of the ten lines of J 2, while in J 2 no transition whatever occurred. In reality the contrast is even stronger than indicated by this general statement, for not only is there no tendency toward mutation displayed during the entire thirty generations of J 2, but in not one of the above mentioned mass cultures derived from the resting eggs of the same source did the humped type appear. The exact generation, on the other hand, in which the humped type appeared in the lines of J 2 were, respectively, the ninth, fifth second, third, third, third, fourth, sixth, fourth and third. It is noteworthy that even in the one instance where the transition in the direct line was delayed till the ninth generation sister indi- viduals to the one in the direct line made the transition as early as the third generation. POWERS EL? C. 370 MITCHELL AND J. Jal c& 89 I & ] 6 61 Or H G6 8 i Gl I G 8I ial 81 | 26 H 98 ial I heat) I J oT 8 91 8 H 001 0 Onreiners 0 I ial 8 ia! L H vor | 0 | 0 6 0 G &I 6 Gl 9 H OMe «he 0) Sri= CO ata c 9 I Or gT OL ¢ as eh LOO Sy re ee ecu a 7 8 8 L v $ OOT 0 0 te i 20 i! ¢ € ‘] fe ts 8 OOT 0 I om | O i Sed G & La | I ee Sees | eae | | | | 2161 Z16T 6 ON H 001 os |> 0 OE aii 20 | 9 1Z Ol H Lg &P I i & & 61 8 6 6 Ss 48 él 0 L i I oT 8 9T 8 >) Lg &F | 0 de & G ial L ial L 8 06 Ol | I 6 i! Us Gl i! IT 9 ) OOT 0 0 IT (eo G Or Il 6 g 8 oor 0 | 0 & 20 I 8 & 8 i Hepetd< 18 8 Oot 0 G 9 0 G i 8 i 6 8 00r 0 0 1-30 Pe asiene HI g z ete eur I [oe Eel i | | Z161 Z16T = 66 6p | SHLVaa 66 6 NieaNava AnypN DNOOX DNOOX saluas SUUVNaH WAXL 40 BOG || aio 40 10 ROwavos|eeeewe ees | ato rsuld to | NI‘N@D so ING) udd INGO udd | UaaWwoNn uadnoNn uaenWon Re ae HLUIa JO ALVA uUGaWON | HLYIA AO ALVa uadaWwoNn ON sjuaind paduny “pf fo woynsauab [Tg fo ponpraypur FI] fo aanjjno sspw woul padisap “% [ sailagy fo hsojsiyy 6 ATAVL INDUCED CHARACTERS IN ASPLANCHNA AMPHORA 13 ee 0 iy 01, ataneeg L ll S|. 001 0 i | @ I iy OR ll 47 S161 2 : phe ry se i : ae es He MC See 19 € @ 9 | ral |HS | ‘oor 0 0 Or 0 | : SI Or Nae | xOOT 0 z er ame 91 CT Moseley SA 0G 6 0 Ol I SF eee I et \ L00r 0 , 0) FI 0 Lo Or | $1 12 eee ae IZ 1 Il g 1s tad Galen | &y H 06 Ol 0 6 I | ay SG lms O1 S‘H 00T 0 0 ¢ OF ol, Sa wg € H ‘pold ‘sig H‘1‘S Z9 88 fd ¢ g | z ay eu OL I | rate ae = ss E a — = Sem y ‘ON aa eal O01 0 I 9 0 | At ed | OE Ce cc" ole. OT 9 |. % I iz eG eile ARs ZI z L I I rad Ol Mm | OP 0¢ I 8 8 at 0z LI a a eae 5 aeG Open Or I I 8I II | oo eee eco. Oy I | OT i au in ae $e | ST 0 L I I a lFrg iat ie rae 0 ¢ l l ial Wert H “pold ea ee 0 Or g P G “qo | g1 I ZIT | & “ON POWERS H. Wi. Cc. 372 MITCHELL AND J. panuryuUojg—G ‘ON penuyu0p—s WIAV.L 4 H 00T 0 | I | Z 0 aT II g II p S OT 0 at ee 0 | P 6 9 8 g Ss 001 OF oes Fe Ce <= i 9 “dea g 9 z | | I G qaq I | | GI6I | CI61 : ~ a = asi 95 Geist ccil| Pe ae H SI 18 0 I L | 8 SI l H 8Z ZL 0 Z G ¢ 91 l [oi 9 (ea 2 19 gE 0 P | 2 I eI 9 ete ¢ H 001 0 I iH 0 Z 61 fl 61 P H cL cZ I g I come 8 aie sg 8 g H ‘poig sig} g 18 ZI 0 8 | = z ¢ ‘qoq | 6 P z | | leat Z 4d I | S161 | ZI6L ys (oie i Fa oo > s: SOI OINy a . . ee ara O01 0 ear p 0 8 Wenge. Or | H 06 Ors) | Weer | “6 I I &Z Or &Z 6 Ie 9g HOb TNE, Mee i l a a8 1Z ral 8I 8 > ei 18 Oi) =p 9 Or Z 91 91 Cl L eo et 16 6 0 ia I I ial ral FI 9 H O01 0 0 Ol 0 Z ral Ol ral c H | oF Beat AO | ee 9 9 Ol “eq 1 See: r | | ZI6I | 2161 Le ea SAMS ae eS . Sn [Oa Al Oe ‘ ee 56 6p | sHivaa | rane 5p Maeaeee | Z One aaa eNnox | onnox sara SHMUVNAY AdAL JO tO fc ce) IO AO qostivo 40 LSula Ao NI‘NG@S JO | Lago udd LINGO wad | UdaWon | GaWwon Yaanon Eyes pg ips Ss dee | UWHeWoON HLUIA AO ALVA UaaqWwon 373 INDUCED CHARACTERS IN ASPLANCHNA AMPHORA —_ 0 OOT 0) Nooo mrAoooconnon OO MO19 & ON 10 -— 1 ~ ~ eo) 2g) mo oooonno & “ON xs OoOoOoTrT q ~ N Nar 1 ~ CS TD oo O rt of Ns os ST CT tea &G OG SI 9 oo Oo © NANN GC “qo | o NOD — aoa n nm mn o o o om Sc rM RNRNNNANRNRNNMNM NRNNNNWHNMNM MR M S| eel Gel Si (=) Ign! Genea(=) S) Se (Se) Sr sSu fay (=>) Us) SoS onononcdr- &> ®O & oS © 30 Oo © & CO SS are al _ a4 Daoroontwttono ooo eo 6 O& © 29 SoS = N N wd OD SH rd eS (SS (eS eat oon on oo n =! S&S mr HOON 19 29 CO CO 1190 6 SH HOD OO ht >) Sooo iia! N _ - - Ne - X Se) ™~ =< = S S | | Re NR RON Soe te I OOO. eK en Oe on | oa N Hi ~~ oS Se oD OD i oD oD oD OD an nN | a a 3 : Bo : Ss =e = | A mormon moN re © © © A139 OOO HID ANN rig = NS Qs rt N NN = aSOonrntwooonn Sonn Hi190o DO OOon N oO OD | N Oo oD | N OD OD a an 1a = 3 edieie = Peete om ou Oo hm a | oO fe _ | = = < = eS Br ma Ba = o VES * RHEOTACTIC RESPONSES mie) (ise Bos = a ne STOCE COMPARED ote D Zao s: n BO |#| zs ¢ | 6 | 888 oM 86] BY peHO x <. | Soe Baa = 70 eet 2 OCma OE < < a rn, a3 ard 3 E hee ca in | | | e | | ’ =. ‘ Ell. |2:8ormore 50) ease 0.0002 27,7126 3) 381) 96 6.2L Prien ieee ....| 0.0002 | 13) 43 18 36 3.1.7 | 114) 7:14 Meroe enarcl ema 0.0002 5 60 40 29 | 8.9| 14:46 | } | Wt }2.0 or less........:-..| 0.0002 | 10] 32] 21| 39} 8 1-7 |) Gos) aioe 2 sb ior mone. Asch - 0.001 | 10 63 34 3 28 | 7.0) 4:69 12.0 or less...... aan 0.001 | 19) 35) 38) 22) 511.78) 5.9) B11 The results exhibited in table 3 are too few to be more than indicative; however, since they agree with all the other observa- tions I think they approximate the truth of the situation. This is that while high efficiency in the current usually accompanies a larger percentage of positive responses and a higher rate of metab- olism than low efficiency, yet considered alone the differences are neither so marked nor so sure as when the sign of the rheo- tactic reaction is made the index of the metabolic condition. One pertinent comparison remains in the relation between survival-time and rheotaxis. What are the rheotactic reactions RHEOTAXIS, RESISTANCE TO POTASSIUM CYANIDE 407 which accompany relatively long and relatively short resistance to potassium cyanide? A synopsis of these results is presented in table 4. Upon making this comparison one is struck, first by the relative closeness of the rheotactic responses given under the two condi- tions and second by the fact that the longer survival-times have always the lower percentage of positive responses. One of the factors causing the relatively low percentage of positive rheo- TABLE 4 Showing the relation between long and short survival-time and rheotaxis. Tempera- ture 17 to 21° C. al] 4a 12am | 2 Y g/22 | sess |e], | 83 = Sn wos a | cAI A Be Apel S| 9 2b 2 | 2° ate E| @ a5 STOCK SURVIVAL-TIMES COMPARED ) BC SEAS i) 4 2? Blades ella, |) 42 Ss 826 ae a a Be, (ce Voenno teas ||. = I-II Hy HOUTSEOTE ESseanares sneer e 29 4:03 65 25) 10 \2.5| 10.5 ‘0.0002 7 hours or more............| 20} 9:16] 60) 32, 18, [2.5 11.1 (0.0002 18 hours or more............| 21} 9:09) 56) 32) 10) 22.3) 7.3 (0.0002 utr 28 hours or more............| 18} 39:10 47) 41, 9} 32.1) 6.5 poop |) 4 hours 15 minutes or less... 20 3:01 51 38 8 32.1) 6.1 0.001 52.1) 6.1 (0.001 | 6 hours 15 minutes or more 16 8:58 34 43 18 tactic reactions with the shorter survival-periods is that isopods sometimes give a high degree of negativeness with a relatively short survival-time (cf. Allee 713), although on the average such isopods live longer in the cyanide than do highly positive cnes (see also p. 405). The indirect method Only isopods from Stock IT were tested by this method. ‘These were placed in 0.00001 molecular postassium cyanide solution in two-liter flasks immediately after having been tested for rheo- taxis. Three easily distinguished individuals were placed in each flask. In order to secure more even temperature the flasks were placed in running water. The temperature averaged 18° C., but 408 W. C. ALLEE varied + 4°. The isopods were examined each morning and night. As soon as one died it was removed with a long pipette. In general the method is, I believe, a trifle more reliable than the direct method since the exact determination of the death point is of less significance. However, since the method is so much slower it was used only to check the results obtained with the direct method. The findings are summarized in table 5. They support, as far as is possible with a limited number of cases, the findings with the direct method. TABLE 5 Showing the relation between survival-time in 0.00001 molecular potassium cyanide solution and the rheotactic reaction in Stock II. Temperature range 14 to 22° C.; average 18° | | . a a Zz 4 > a 'S5 il =i B 4 feo] > 3] a D = I I a AVERAGE RHEOTACTIC REAC- S o Zn = TION IN PERCENTAGE ee = De RHEOTATIC RESPONSES COMPARED 9 OF TOTAL TRIALS S +) A 2 a Az ms o oH Q = <5 <8 =) a as | aS 2 ——_-, e | BE | ge a ts ae =, |e ew . < | ea qu Soh GSE, = ae See ke | 4 sk 60 per cent or more +...... 19 86r 4) 5 9 | 207 | sO26a oat 2 a. | 61 | 20 3/24 |10.0| 26 40 per cent or less +...... 1 DISCUSSION It was shown in an earlier paper (Allee 12) that isopods of the pond mores are less positive in their rheotactic reaction than those of the stream mores. This is again illustrated by these studies and light is thrown on the metabolic condition of the isopods giving this difference in response. Thus the rheotactic reactions of 72 isopods of the stream mores (Stocks I and I), taken partially during the lowered responses of the breeding season, averaged 61 per cent positive, 28 per cent negative, and 11 per cent indefinite, with an average efficiency in the current of 2.5. These gave a mean survival-time in 0.0002 molecular potas- sium cyanide of five hours and fifty-nine minutes. Of the pond mores 51 were killed in 0.0002 molecular cyanide solution with an average survival-time of twenty-three hours and RHEOTAXIS, RESISTANCE TO POTASSIUM CYANIDE 409 ten minutes. These pond isopods had given an average rheo- tactic response of 51 per cent positive, 37 per cent negative, 10 per cent indefinite and 2 per cent no reaction. Their average efficiency in the current was 2.2. Another lot of 55 pond isopods which gave a rheotactic average similar to the above, were killed off from normal conditions in 0.001 molecular potassium cyanide; these gave a mean resistance of five hours and fourteen minutes, just forty-five minutes less than that given by the stream isopods to a cyanide solution one-fifth as strong. All this means that the stream isopods, the ones that as a rule are more highly positive to and more efficient in their reac- tions in a water current than pond isopods, have a shorter survival- time in equimolecular solutions of potassium cyanide and therefore have a higher rate of metabolism than do the pond isopods. If only the highly positive isopods of the two mores are chosen for a comparison of their resistance to potassium cyanide, another _ situation is revealed. The individuals of the two groups selected on the basis of the similarity of their rheotactic response show a great difference in their survival-time in equimolecular solutions of the cyanide and hence in their metabolic rate. In table 2 (p. 404) the stream isopods with a rheotactic reaction 60 per cent or more positive averaged 88 per cent positive and gave a mean survival-time in 0.0002 molecular potassium cyanide of five hours and twenty-four minutes. Isopods from the pond mores selected on the same basis averaged 83 per cent positive and gave a sur- vival-time of twenty-one hours and sixteen minutes, in the same strength of cyanide solution. Here is a great difference in the metabolic activity of the two groups of isopods but a close simi- larity in their rheotactic reaction. It can also be seen in table 2 that the pond isopods that give a low positive response have a longer survival-time in potassium cyanide than those that are more highly positive. This must mean that the rate of metab- olism which is accompanied by a high degree of positiveness is relatively high when compared with the mean metabolic rate given under the conditions to which the isopods are acclimated, but that it may not be an absolutely high rate when compared THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 3 410 WwW. C. ALLEE with a fixed standard. This topic will immediately be discussed more fully in dealing with the reactions of individual isopods. Thus far I have been dealing with averaged results. For generalizations these are necessary and in order to save space the individual records have not been given. How far does the work with individuals support the averaged results? Before taking up this point it is necessary to point out that each individual has in all probability a rate of metabolic processes slightly different from that of any other isopod and further, that it is not a fixed standard rate of metabolism that finds its expression in a high rate of positiveness but rather a relative rate. That is, when the metab- olism of an isopod is rapid for that individual it tends to go posi- tive to the water current, when less rapid, negative, when still less rapid, indefinite, and when least rapid no reaction at all is given. But a rate of metabolism that is rapid for one isopod may be slow for another, and intermediate for a third. Now all three would show the same survival-time in potassium cyanide and yet , their rheotactic reaction would be highly different. Thus forexam- ple, isopods 96 and 101 from Stock III had a survival-time of three hours and forty-five minutes in 0.001 molecular cyanide solution. The former gave 50 per cent positive, the latter 100 per cent positive reactions. Isopod No. 97 was killed in the same solution with the other two with a survival-time of three hours and fifteen minutes, yet its rheotactic reaction was only 60 per cent positive. Since it is obviously impossible to test the survival-time of an isopod in its varying rates of metabolism, one is forced to do the next best thing, that is, take average results. There is still another complication in this regard. An individ- ual isopod may give identical rheotactic reactions when its rate of metabolism varies greatly. Thus an isopod kept at 20°C. comes to have a normal mean metabolic rate, also a normal mean rheotactic response. When the metabolic rates goes above this normal mean the isopod tends to become more positive, when below, less positive to water currents. Put the same isopod at 5°C. The metabolic rate is depressed and the positiveness also decreases. But in time the isopod becomes acclimated; a new metabolic mean is established and the rheotactic reaction goes up RHEOTAXIS, RESISTANCE TO POTASSIUM CYANIDE 411 to about its old average and plays up and down as the metabolic rate varies about its new mean. So the rheotactic reaction is an expression, pot of the absolute metabolic rate of the animal, but of the relative metabolic conditions under which the isopod is acclimated for the time being. Further experimental evidence on this point will be published elsewhere in conjunction with Dr. Shiro Tashiro. But even with this relative scale of metabolism determining the rheotactic reaction of the isopod in place of an absolute scale, the fact remains that individuals having a short survival-time usually have a high rate of positiveness to a water current. This was determined for individuals as follows: An assistant read me the rheotactic reaction of isopods in groups of two from the same stock, giving individuals that had at least 10 per cent of differ- ence in their positive rheotactic response. Basing judgment on the rheotactic reaction alone I was able to predict the relative survival-time in 106 out of the 160 cases tried. This shows that the rheotactic reaction in 66 per cent of the cases was a correct indicator of the absolute metabolic conditions of the isopods compared. Isopods with a highly negative reaction gave the most difficulty, while it was impossible to predict the survival- time of isopods that did not move at all in the current. The explanation of this difficulty has already been given (pp. 405-406). Jennings (’06) clearly outlined the problems in this particular field of animal behavior when he said, ‘‘We are compelled to assume the existence of changing physiological states. This assumption besides being logically necessary, is of course sup- ported by much positive evidence drawn from diverse fields, and there is reason to believe that in time we shall be able to study these states directly.” With the application of the cyanide method to problems of animal behavior this prophecy is a step nearer its final fulfilment and it is now possible to demonstrate directly as regards the rheotactic reaction of isopods that high positiveness is the expres- sion of a relatively high rate of metabolism, and low positive- ness, of a low metabolic rate. December, 1913 412 W. C. ALLEE LITERATURE CITED ALLEE, W. C. 1912 An experimental analysis of the relation between physio- logical states and rheotaxis in Isopoda. Jour. Exp. Zoél., vol. 13, pp. 269-344. 1913 Further studies in physiological states and rheotaxis in Isopoda. Ibid., vol. 15, pp. 257-295. AuLEE, W. C., AND TasHrRoO, SuHrro 1914 Some relations between rheotaxis and the rate of carbon dioxide production of isopods. Journ..An. Beh., vol. 4. (In press) Cuitp, C.M. 1918 Studies on the dynamics of morphogenesis and inheritance in experimental reproduction. V. The relation between resistance to depressing agents and rate of metabolism in Planaria dorotocephala and its value as a method of investigation. Jour. Exp. Zoél., vol. 14, pp. 153-206. 1913a Certain dynamic factors in experimental reproduction and their significance for problems of reproduction and development. Arch. Entw. Mech., Bd. 35, pp. 598-641. GepPeRT, J. 1899 Uber das Wesen der Blausiure-vergiftung. Zeitschr. klin. Med., Bd. 15, pp. 208-307. # JENNINGS, H.S. 1906 Behavior of the lower organisms. 366 pages, New York. SHELFORD, V. KE. 1911 Ecological succession. I. Stream fishes and the method of physiographic analysis. Biol. Bull., vol. 21, pp. 9-36. TasHIRO, SHiRO 1913 A new method and apparatus for the estimation of ex- ceedingly minute quantities of carbon dioxide. Am. Journ. Physiol., vol. 32, pp. 136-145. STUDIES ON THE DYNAMICS OF MORPHOGENESIS AND INHERITANCE IN EXPERIMENTAL REPRODUCTION VII. THE STIMULATION OF PIECES BY SECTION IN PLANARIA DOROTOCEPHALA C. M. CHILD From the Hull Zoélogical Laboratory, University of Chicago FOUR FIGURES I. METHODS By means of the direct susceptibility method it is possible to follow the changes in rate of metabolism resulting from section in pieces. This method, which has been described in detail else- where (Child ’13 a), consists in determining the susceptibility of individuals or pieces to concentrations of KCN, alcohol, and so forth, which kill within a few hours. In such concentrations the higher the general rate of metabolism the greater the suscepti- bility and the shorter the life of the animal or piece in the solution. Tn this way the susceptibility serves as a means of distinguishing and comparing the general rate of metabolism in different indi- viduals, regions, or pieces. Since the changes in susceptibility following section depend upon size of piece and region of the body from which it is taken it has been found most satisfactory to determine the susceptibility of series of pieces, representing both different fractions of the body length and different regions of the body, at various intervals-after section. Of course a different series of pieces must be used for each susceptibility determination but by standardizing the material used and the external conditions a remarkable uniformity of results is possible. The method of experiment is as follows: The bodies, exclusive of the heads, of ten worms of equal length and similar physiologi- 413 414 C. M. CHILD cal condition and always from the same stock, are cut into fourths, sixths, eighths, or twelfths; each lot of ten pieces representing approximately the same region of the body is kept together and isolated from other lots and its susceptibility tested at the desired time after section. In all the susceptibility determinations of pieces presented below animals from the same stock were used and great care was exercised in selecting those of equal size, as experience has shown that in a particular stock size is the best criterion of physiological ‘condition, and also in cutting the pieces so that they should rep- resent as nearly as possible the same regions of the bodies of different individuals. It is of course impossible to avoid some differences in size and level of pieces but the results will show ‘how unimportant these differences are. Jn all susceptibility deter- minations KCN 0.001 m was used in filtered water from the same source and at the same temperature and the same quantity of KCN solution was used in each case. In this manner the susceptibility of series of ten each of fourths, sixths, and eights, of wellfed animals 18 mm. in length has been determined, immediately, twelve, twenty-four, forty-eight, seventy- two, ninety-six and one hundred and twenty hours after section, and for some series of pieces three and six hours after section also. For twelfths the susceptibility has been determined, immediately, twenty-four and forty-eight hours after section, but it is impossible to avoid considerable relative differences in size in pieces as small as these and such differences complicate the results. With pieces smaller than twelfths no general susceptibility determinations have been attempted. “Tmmediately after section” in these series means that as soon as the cutting of the series of pieces was completed they were placed in KCN. Since the preparation of the series requires from ten minutes to an hour according to the size and number of pieces the first pieces cut remain in water for some time before being placed in KCN. It was ascertained, however, that this made no difference in the results since the susceptibility does not change appreciably during several hours after section. In nearly ee ath 14 DYNAMICS OF MORPHOGENESIS 415 all cases the determination for each size and each time has been made repeatedly so that conclusions do not rest on single series and besides this the results of these experiments are amply con- firmed by large numbers of others in which only certain regions of the body instead of the whole were used in susceptibility determinations in relation to size and region. In the susceptibility determinations, as stated the whole body ex- cept the head was cut into fourths, sixths, ete., and the susceptibility of all pieces was determined and recorded, but in the following tables only the pieces from the first zodid (see Child ’11 e) are included. ‘The pieces of the posterior zodids show nothing funda- mentally different from those of the first zodids and double the length of the tables. Only a sufficient number of susceptibility records are given to show the general course of the changes in susceptibility. The tables of these records given below are sup- plemented and confirmed by many other similar series in my notes. The tables are in the form of the susceptibility tables in an earlier paper (Child 13 a). The column headed “‘length of time”’ gives in hours and minutes the length of time that the animals or pieces have been in the solution at each observation. The column headed ‘“‘lots’”’ gives the serial numbers for each time of the different lots of ten each, beginning with the most anterior pieces as Lot 1. The lot-numbers are the same as the numbers of the pieces in the corresponding figures. The five columns headed ‘‘I to V” under “stages of disinte- gration’’ distinguish more or less arbitrarily five stages in the visible changes which the animals undergo in dying. These five stages are briefly as follows: I. Still intact, showing no appreciable disintegration. Such animals or pieces are always alive and moving about. Il. In whole animals, from the first traces of disintegration to the beginning of disintegration along the margins of the body, which usually follows disintegration of the head. In pieces from the beginning of disintegration at one or both ends to beginning of disintegration along the lateral margins. Considerable motor activity may still be present. 416 C. M. CHILD III. In both whole animals and pieces disintegrations of lateral margins of body until it is completed and swelling of body begins. Movement may still occur to some extent in the intact parts. IV. Swelling of the body to complete loss of epithelium and loss of shape. Motor activity has ceased. V. All further changes after swelling and loss of shape are completed. The recognition of these different stages makes it possible to distinguish more clearly slight differences in susceptibility which would otherwise not be evident. The numbers in each of the five columns for each time and each lot are the actual numbers of animals or pieces of that lot which are in that stage at that time. Lots of ten each were used and condition was recorded every half-hour in all cases. From these tables it is possible to determine at a glance the regional differences in susceptibility and by comparison of the tables for pieces of different size the effect of size and region upon the changes in susceptibility following section appears and com- parison of the susceptibilities of pieces of the same size at different times shows the course of the susceptibility changes. Il. STIMULATION 1. The susceptibility of whole animals In order to show how the susceptibility of the pieces changes following section it is necessary to determine the susceptibility of whole animals as a basis for comparison. ‘Table 1 gives the susceptibility of a lot of ten worms 18 mm. in length, that is, of the same size and physiological condition as the animals used for the preparation of pieces. 2. The one-fourth pieces The regions of the body represented are indicated in figure 1. Pieces 7 and 2 represent the body of the first zodid and only the records of these pieces are included in the tables. DYNAMICS OF MORPHOGENESIS 417 Comparison of table 2, the one-fourth pieces, immediately after section, with table 1, the whole worms, shows that the susceptibility of the pieces is somewhat greater than that of the whole animals. Susceptibility determinations twelve and twenty- four hours after section show a lower susceptibility than table 2 but no great change occurs during this period. Table 3, forty- eight hours after section, shows a susceptibility lower than that of whole animals. 418 C. M. CHILD TABLE 1 Susceptibility of ten worms 18 mm. in length to KCN 0.001 m STAGES OF DISINTEGRATION LENGTH OF TIME IN KCN I II Ill IV Vv 4.00 4 6 4.30 4 5 1 5.00 a 3 5.30 3 a 6 .00 10 6.30 10 7.00 7 3 7.30 4 4 2 8 .00 4 6 8 .30 |: 2 8 9 .00 10 TABLE 2 Susceptibility to KCN 0.001 m of one-fourth pieces immediately after section STE ee STAGES OF DISINTEGRATION IN | = | I | II Ill Iv Vv LD ated te Ales 500. Sa ae ed ne 1 9 il ios ie a2 . 1 6 3 1 so «| 6 | 3 1 7 3 1 7 3 ant { 2 9 1 1 2 7 1 6.00 { 3 ; , 1 ri 3 6.30 { 5 ‘ : 1 4 6 7.00 { : : 6 7.30 { Pe ii iW cannes et a 8.00 9 Ste Sle 2s Ae S16 — TABLE 3 ile lee Susceptibility to KCN 0.001 m of one-fourth pieces forty-eight hours after section LENGTH OF TIME STAGES OF DISINTEGRATION KCN } LOTS a I | re ur IV Vv 1 9 1 5.30 { ? " 1 4 6 | 6.00 { 9 8 62 | 1 1 9 | 6.30 { 3 § 5 | “A 1 1 5 Z 5 en { 2 6 2 2 Z 1 1 5 2 2 ee { 2 2 3 3 2 1 1 2 2 5 age { 2 2 3 3 1 1 1 if 2 i | ou { 2 2 3 3 2 Be 1 1 1 8 a0 { 2 2 1 3 4 : f if 1 9 9.30 \ ; é eae 10.00 1-- |} ----}---- |--------- fr 10 10.30 2 2 8 11.00 2--|}----------|}----------- 10 TABLE 4 Susceptibility to KCN 0.001 m of one-fourth pieces one hundred and twenty hours after section : STAGES OF DISINTEGRATION LENGTH OF TIME Ki LOTS -: i Ir Oi Lye Wi i! 1 7 3 3.00 \ ; - : 3.30 { ; ; ; 1 5 4 1 fe {| 2 4 4 2 1 4 4 2 4.30 { : j : i 1 4 1 3 2 a { 2 4 ew TA 5.30 { : 2 : aS la 1 2 Lt slg ©. { J==--------—--} ---- ee IE ih) 6.30 1 A? Sena 7.00 (ee 2 ee en re | - 10 420 C. M. CHILD is 2 From forty-eight hours on, the susceptibility begins again to increase as the pieces undergo reconstitution. At seventy-two hours it is somewhat greater, at ninety-six hours still greater and table 4 for one hundred and twenty hours shows that it has now again become greater than that of the whole worms (ef. table 1). | As regards the susceptibility of the first and second fourths, ,_ tables 2 to 4 show that the first fourth is at all times more sus- .» ceptible than the second, that is, the original axial gradient (Child 12 a, 13 b, 13 ce) appears in the difference in susceptibility between the first and second fourths. DYNAMICS OF MORPHOGENESIS 471 ‘ 3. The one-sixth pieces Figure 2 shows the region of body included in each of the one- sixth pieces. In table 5 to 8 only records of the pieces 1, 2 and 3 which represent the body of the first zodid are included. Comparison of tables 5 to 8 with each other and with table 1 shows that the susceptibility of the one-sixth pieces immediately after section (table 5) is greater, but after twelve hours (table 6) is less than that of whole worms. TABLE 5 Susceptibility to KCN 0.001 m of one-sixth pieces immediately after section | STAGES OF DISINTEGRATION akc LoTs I 5a III EVs } Wire 1 10 3.30 2 ii 3 3 7 3 | 1 i 3 4.00 2 4 6 3 4 6 | 1 2 8 4.30 2 1 6 3 3 2 6 2 | i 2 6 1 1 5.00 | 2 | 2 6 2 | 3 | 4 5 1 ih an ee 7 1 5.30 2 6 3 1 3 1 7 1 1 1 2 4 1 3 6.00 2 2 6 2 3 3 2 5 1 2 1 2 5 6.30 2 5 5 3 1 3 6 1 2 1 rf 7.00 2 4 6 3 1 9 1 1 9 7.30 Se eee eee St = 0 pe Se Se ee ee i () 8.00 1 ee eee eee 10 422 C. M. CHILD Susceptibility determinations twenty-four and _ forty-eight hours (table 7) after section show a slight further decrease in susceptibility. From this time on the susceptibility begins to increase as reconstitution proceeds. At seventy-two hours it has increased slightly, at ninety-six still more and at one hundred and twenty hours it is once more higher than that of the whole worm. As reconstitution goes on it continues. to increase. These sixths show another interesting feature. Comparison of Lots 7, 2 and 3 in table 5 shows that 2 and 3 are considerably TABLE 6 Susceptibility to KCN 0.001 m of one-sixth pieces twelve hours after section LENGTH OF TIME STAGES OF DISINTEGRATION in KCN AS = Bo aD 2 s it Il a aa Sie ok ‘ Hee Vv [) 1 9 1 5.00 | 2 10 3 8 2 1 7 3 | 5.30 | 2 10 | 3 Uf 2 1 | | 1 5 4 1 | 6.00 | 2 8 2 | 3 6 2 2 ie ai 4 3 2 ees | 6.30 | 2 5 3 2 | 3 3 1 3 2 1 | 1 4 3 2 1 7.00 | 2 5 3 2 | Se ik 2 4 3 | i | | 4 1 5 7.30 2 Weeder) 1 2 2 3 2 1 1 6 1 1 9 8 00 2 ie | 2 2 5 3 1 Z 7 1 } 1 9 § .30 9 1 2 1 6 | 3 1 9 9.00 [t= BE Sree BS eek es ==-- - Sap | 2 2 bere | a = 1 J----~- |. --~--|----- ==A10 9.30 | 2 | 2 8 10.00 9 athena _ eG): Ae oe 22 ea DYNAMICS OF MORPHOGENESIS 423 more susceptible than 7; in other words, the susceptibility of pos- terior is greater than of anterior pieces. This change constitutes, so far as the different pieces are concerned, a reversal of the axial gradient of the first zodid of uninjured worms. It is a char- acteristic feature not only of sixths but, as will appear below, of all shorter pieces and is due to the fact that posterior pieces are more stimulated by the act of section than anterior. After twelve hours (table 6) the effect of this stimulation has - disappeared in Lot 2 so that its susceptibility is lower than that S TABLE 7 Susceptibility to KCN 0.001 m. of one-sixth pieces forty-eight hours after section y. oS ISIN’ Z N LENGTH OF TIME anes STAGES OF DISINTEGRATION mv KCN | poe FA) ar I - Il ir LV: } Vv (| 1 9 1 | 6.00 2 10 | 3 10 1 7 3 6.30 2 fF 10 | 3 | 10 f SS ae 4 2 7.00 2 10 | | 3 10 | | 1 2 2 3 = | 7.30 2 9 1 3 9 1 1 2 1 2 5 8.00 om) 5 3 1 1 3 4 3 2 1 | i! : 2 8 8 .30 2 6 1 1 1 3 2 4 1 o 1 1 2 8 9.00 2 1 3 1 5 3 (gilli. ho je. 3 3 ters se | 2 = 10 9.30 4 2 6 3 [ 3 1 3 6 2 ] 9 10.00 5 : 2 ; 95 ae = |= 40 10.30 z ; i & ‘ ae 424 Cc. M. CHILD of Lot 1 while Lot 3 has not yet fully recovered and its suscepti- bility is about the same as that of Lot 1. After twenty-four hours further recovery has occurred and after forty-eight hours (table 7) the susceptibility of Lot 2 is lower than that of Lot 1 and that of Lot 3 slightly lower than that of Lot 2, that is, the original axial gradient of the whole worm has now reappeared in the relative susceptibilities of the pieces. As reconstitution proceeds, the susceptibility of all the pieces rises and since the more posterior regions of the first zodid undergo more reorganization than the anterior regions in giving rise to a new individual, the susceptibility of the posterior pieces rises TABLE 8 Susceptibility to KCN 0.001 m. of one-sixth pieces one hundred and twenty hours after section LENGTH OF TIME STAGES OF DISINTEGRATION In KCN | | LOTS : < | I I I IV | v | 1 4 6 3.00 | 2 6 4 | \ 3 9 1 f 1 10 3.30 | 2 10 | 3 3 i (| 1 9 1 | 4.00 | 2 10 | 3 3 ff | 1 | 6 | 4.30 | 204 | ee 3 i | She ex 3 iu (| 1 |} 4 3 2 1 5.00 2 | 5 3 2 | 3 | 4 5 1 | 1 | 5 |. 3 5.80 2 | 6 ic 3 3 | 6 3 1 [ 1 2 Wiles 6.00 | 2 4 6 3 2 2a) 6 f ibe ~---}----}-10 6.30 2 See SY ( | 3 2 8 7.00 3 = : = Say ap DYNAMICS OF MORPHOGENESIS 425 more than that of the anterior. In table 8, one hundred and twenty hours after section, this has occurred and the suscepti- bility of all three lots is almost the same. Still later the sus- ceptibility of the posterior pieces provided they undergo complete reconstitution, may become even greater than that of the anterior pieces. 4. The one-eighth pieces The regions represented by the pieces are shown in figure 3. In tables 9 to 12 only the records of pieces 1 to 4 are given. Table 9 compared with tables 5,2 and 1 shows that in the first four one-eighth pieces the susceptibility is higher immediately after section than in the one-sixth pieces or one-fourth pieces or in the whole worms. Not only do these pieces complete disinte- gration earlier than the others, but the first traces of disintegration appear earlier. After twelve hours, however, as shown in table 10, the suscepti- bility has decreased so that it is about the same as that of the whole animals. At twenty-four hours and also at forty-eight hours (table 11) it is still about the same. Seventy-two hours after section the susceptibility is again higher, at ninety-six hours still higher and at one hundred and twenty hours (table 12) it is higher still. Many of these short pieces in the more posterior regions do not undergo complete reconstitution but remain head- less and in these the susceptibility does not undergo as great an increase as in those which produce wholes. But the one-eighth like the one-sixth pieces show regional dif- ferences in the increase in susceptibility immediately following section. In table 9 it is evident that the anterior pieces (Lot 1) are in general least susceptible, the second eighths (Lot 2) con- siderably more susceptible, the third eighths (Lot 3) slightly more susceptible than the second, and the fourth eighth (Lot 4) about like the third. Here, as in the one-sixth pieces, the regions of the body which in the uninjured animal have lower susceptibility show a higher susceptibility after section, that is, the stimulation resulting from section is least in the anterior piece and increases THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 3 426 ¢. M. CHILD from piece to piece in the posterior direction except in the last two pieces where there is little difference. After twelve hours (table 10) this stimulation has disappeared and now the susceptibility decreases from piece to piece in the posterior direction, i e., the original axial gradient has reappeared. At twenty-four hours the relative susceptibilities of the one- eighth pieces remain about the same as at twelve hours. In table 11 (forty-eight hours) some slight irregularities in the susceptibility of different pieces appear. Lot 1 is most susceptible, Lot 2 less — es = DYNAMICS OF MORPHOGENESIS 427 TABLE 9 Susceptibility to KCN 0.001 m. of one-eighth pieces immediately after section LENGTH OF TIME IN KCN 2.30 3.00 3.30 4.00 4.30 5 .00 5.30 a — STAGES OF DISINTEGRATION LOTS af - I cae Ir IV Vv 1 10 2 6 4 3 7 3 4 10 1 10 2 3 oh 3 4 5 1 4 | 6 4 1 | a 3 2 2 5 3 3 | 4 5 1 4 4 5 1 1 6 4 2 8 2 3 5 4 1 4 6 4 1 3 i 2 6 3 1 3 4 2 2 2 4 4 2 2 @- 2 1 6 4 2 1 4 4 ji 3 2 3 2 3 4 3 5 2 1 g- 1 2 3 2 5 3 1 + 5 4 2 1 7 1 6 4 2, 3 7 3 2 8 4 2 8 1 1 9 2-- ----- ----- -----}----- -- 10 3 - —- —-- -~-----~ ----- -- 10 4--|----- |--~-- |---~-~ -~---}|-- 10 z 1 5 5 1 --’----- Im ---- ----- ----- - 10 Cc. M. CHILD 428 TABLE 10 Susceptibility to KCN 9.001 m. of one-eighth pieces twelve hours after section ao ANMOMWINMOChHAHr~MADAARAARACS > S| re wal ir! | | ee | sal | | ln eT | ] I i) : | 1 | ok q Aaa AMNANHHHN AMI |, FOI | ao | Lan] a <4 | | ea fe x | | ee | | ena z | 3 eal | o ANMAMMANMIAMDNA Nar Mae ae A ao een 3 (4 ° 7 = ior ky | =o <> [el a ! q | | | BOSE || sard nN SeHNM MONAHAN AN 4 al ee | | | | ee ee ee | ea | | HH MODSWAMDAAM MONA OMAHA YW OD A | 4 ten | = | | : = | | | | | | g aN OHA NH HE NMDA NO HAN Ow dH NOMAD AINOM HAIN ODA HN SAH NY HH 4 sie | : ~ ~ —— St ia — | a a Be S S S =) S S i) 2 GS oF S % S % e@ °* S % S- ae = ae H bal ve) Ve) =) © is id [e4) wH jon) [=P) | OR Z | 12] = TABLE 11 Susceptibility to KCN 0.001 m. of one-eighth pieces forty-eight hours after section | 2 STAGES OF DISINTEGRATION LENGTH OF TIME IN CN LOTS I Il it TV Wi 1 9 1 | Z 10 | | | 3.30 > wed l 4 10 [ 1 fi 3 2 10 4.00 ; | 4 10 1 6 4 | 2 10 | 4.30 | 2 : ; | 4 10 | I 1 4 5 1 5 | 2 10 | | 5.00 | 3 9 1 | 4 10 | | | 1 | 7 3 | | 2 eae! 3 | 5.30 | 3 5 A 1 | 4 9 1 | 1 a eee: 3 2 4 6 p00 | 3 4 2H iw 3 il | 4 8 2 1 | 20) ONES 4 | 1 2 her! 3 eS pe Set Wet ok 5 3 1 | 4 ho nee 8 f 1 | 3 2 5 2 ie Rae ea 3 G00 | 3 | 1 HERD 2 5 | 4 He, hele eG 1 2 | 1 ta 1 3 6 Z | 2 jaa a get) 3 3 7.30 3 9 3 5 | 4 | 3 AD 49 1 [ === —— | J-----[; - 10 2 | 1 3 6 8.00 : | ; ‘ { 4 | 1 1 2 6 2 1 1 8 $8.30 3 1 9 4 | 1 9 2 | 1 9 9.00 3 1 9 1 LEE SS 22 ee a =10 Dee See es | jae i) 9.30 ic aN Ea ane ae 430 C. M. CHILD susceptible, Lot 3 more susceptible than Lot 2, and Lot 4 again less susceptible than Lot 3. These irregularities are undoubtedly incidental and due either to differences in size of pieces, too large pieces showing too low and too small too high a susceptibility, or to stimulation of some of the pieces which cannot always be avoided and which distinctly increases susceptibility. TABLE 12 Susceptibility to KCN 0.001 m of one-eighth pieces one hundred and twenty hours ; after section ae OF TIME Ge | Mee Leet eee Oe Se I II | IIL IV Vv 1 | 7 oe 6) 2 8 2 | 3.30 : i 4 Camere: * | Ne ae ie a OF ha 4.00 5 arse oh 4 5 5 1 5 5 2 2 4 4 4.30 5 F ; : ; 4 hpi 4 1 1 1 | | “4r6 2 2 2 ges 2) 2 4 5.00 5 | 3 : ; : 4 | 3 | 4 1 1 1 | 1 | | 6 4 2 | | | emer ee | 8 5.30 ) 3 isp : ; l 4 | 2 3 5 ce 4 en ee ag 6.00 5 : . ; 2 ale ton ----- -----|-----|-- 10 6.30 3 | | 9 8 ( 4 3 7 3 a ----|-- 10 7.00 t Oo Gee hie eas ae DYNAMICS OF MORPHOGENESIS 431 At seventy-two and ninety-six hours the gradient in suscepti- bility from piece to piece is essentially that of the uninjured ani- mal. At one hundred and twenty hours the same is true, as table 12 shows. Here Lot 1 has the greatest susceptibility, Lot 2 is next, Lot 3 next, and Lot 4 about like Lot 3. One-eighth pieces from the posterior region of the first zodid undergo but little re- constitution and almost never produce whole animals while in the more anterior pieces the frequency of complete reconstitution is greater, consequently the differences in susceptibility in table 12 one hundred and twenty hours after section are due in part to a greater degree of reconstitution in anterior than in posterior pieces. 5. The one-twelfth pieces The difficulty of cutting even large individuals into as many as twelve pieces of anything like equal size sets a limit to the study of consecutive series of pieces. It is of course possible to cut single pieces much smaller than this with a considerable degree of accuracy but in consecutive series the difficulties are much greater. Figure 4 indicates the regions represented by the first six pieces which make up the first zodid. Tables 13 and 14 give the sus- ceptibilities immediately and forty-eight hours after section. The susceptibility of all the pieces is very high immediately after section, being higher than that of any of the larger pieces (tables 2, 5, 9) or of the whole worms (table 1). The survival- time of even the least susceptible pieces is only half that of the whole worms and larger pieces. These short pieces are evidently more stimulated by the act of section than are larger pieces. After forty-eight hours, however, (table 14) the susceptibility of all pieces has decreased greatly and is only slightly higher than that of uninjured animals. The effect of section has at least largely disappeared. Among the one-twelfth pieces only the more anterior produce whole animals and these do not in all cases. The more posterior pieces usually remain headless. Consequently the later increase 432 Cc. M. CHILD in rate associated with reconstitution is much greater in the anterior than in the posterior pieces. It is in general similar to that in the one-eighth pieces and requires no special comment. DYNAMICS OF MORPHOGENESIS 433 TABLE 13 Susceptibility to KCN 0.001 m of one-twe!fth pieces immediately after section ) | | STAGES OF DISINTEGRATION RON LoTs | e us I l ul V V ( l a) | 2 10 | 3 8 2 5 hea 3 | f 6 CaN 8 | | 1 10 | | 3 Go ee 4 | pot 4 Gar Ae pu) 5 Dele TI eG 2 6 bers” Ghee 2 1 2 Sh PA NS | 2 5 3 2 ] 3 5 3 Tae 1 HALL 4 3 3 2 2 5 1 3 DR Bell 4 6 1 1 4 4 | 1 1 1 4 3 1 | 2 3 4 3 3 Sut ili ee 5 2.30 | ; ee | ow ig 5 i's Paes | 7 [ 6 1 | 9 ( 1 ny ial Te | lei, SS 2 | iL Aes eae 8 3 1 1 8 3.00 ; ; 4 5 . 1 9 [ 6 1 9 | 1 1 9 a Be ave = t 32 = ee == i 4.00 4 Ve = |p | Se = 10 {| ee Sir | @=—=-=—-=—-=-= SS == |= = =-— |>— 16 [ = =| S=es5) 52225) -= 10 434 C. M. CHILD TABLE 14 Susceptibility to KCN 0.001 m of one-twelfth pieces forty-eight hours after section LENGTH OF TIME | STAGES OF DISINTEGRATION ae LOTS |. _- 22 Ji II Ill IV Ne (| 1 OF 4 1 2 9 1 3 9 1 3.00 ji ‘a 5 10 6 1s | 1 9 1 y) ii 2 1 3 9 | 1 3.30 ji a 5 10 6 10 1 9 1 2 i 1 1 1 3 9 1 = 00 | 4 | 10 | Sper coy .| 1 | 6 10 | 1 6 3 1 | 2 ee. |. 3 ia 3 Boo 8 & 1 4.30 | i ; 4 ! 5 8 1 1 | 6 10 | 1 2 4 3 1 2 2 1 1 6 | 3 Shae |, ard | 1 5-00 4 ee ic oe 1 1 1 a 25 8 | 2 6 10 ( 1 2 2 2 4 2 | 1 1 8 3 2 4 3 1 5.30 | i mae | = [.| 5 ie 1 | 2 6 | ris 3 | 1 4 6 2 19M 9 3 | feniete 2 7 6.00 5 : 5 6 1 3 Cale 6 ith 2 DYNAMICS OF MORPHOGENESIS 435 TABLE 14—Continued ~ STAGES OF DISINTEGRATION Se te ae LOTS ain oon See eee ia, 7 | I II lr IV | Vv ie ign | R¥G ee 1 | Seg 3 Aig *=| 9 6.30 A 2 e 5 5 1 4 | 6 3 3 ee a | jee ea | ee eee FS: er Ses ae I ----|-- 10 ee ee ee ——— 10) 7.00 i Fades | ’ 7 5 2 If) RANE * Ores! 5 6 2 Aeoays,| 4 4 | hare 8 7.30 | 5 | 1 | arm 7 | 6 | 2 8 ( i lose es 522s 22525 = 10 8.00 = = ba ee SS eee ee 2210) G Ss 22252 SSoss 45555 555== == 10 III. CONCLUSIONS A number of conclusions which will be shown later to be of the greatest importance for an understanding of the process of reconstitution are to be drawn from the data presented in the tables above. These are briefly stated and discussed in the fol- lowing paragraphs. 1. A temporary increase in rate of metabolism, a ‘stimulation’ lasting at least several hours, results from the act of section. Such a stimulation is to be expected for the act of section severs nerve cords and of course produces extensive injury to other tissue. Susceptibility determinations of pieces three hours after section, of which the records are not given, show that at that time the susceptibility is as high as immediately after section. From this time on it gradually decreases and twelve hours after section the susceptibility of the pieces is about the same as that of the unin- jured animals. 436 C. M. CHILD This also is to be expected, for the piece is no longer in con- nection with its chief sources of stimulation, that is, first the head region and secondly other parts of the body. As regards move- ment it is much less active than the intact animal and if it were not for the wounds at its ends and the growth processes beginning there its rate of metabolism would undoubtedly fall far below that of the uninjured animal. 2. The temporary increase in rate of metabolism following section varies in amount inversely as the length of the piece. A comparison of tables 2, 5, 9 and 13 will show that in general the smaller the pieces the higher the susceptibility. The one- fourth pieces show in the tables practically no increase in sus- ceptibility following section. As a matter of fact the regions adjoining the cut ends do show an increased susceptibility but it does not involve the piece as a whole. That the shorter piece should be more stimulated than the longer by section is also in accord with expectation. The shorter the piece the more direct the injury and resulting stimulation of its conducting paths and tissues. In very short pieces almost the whole length becomes involved in the wound reaction at the two ends. Tables 10 and 14 show that the susceptibility of the one-eighth and one-twelfth pieces does not fall during the forty- eight hours following section to so low a level as that of the longer pieces. This difference is undoubtedly due to the fact that in the shorter pieces the wound reaction affects the whole rather than merely the two terminal regions as in the longer pieces. In the long pieces the susceptibility of the two terminal regions is distinctly higher than that of the rest of the piece at this stage. 3. The temporary increase in rate of metabolism following section varies in amount according to the region of the original body which the piece represents, being least in the most anterior and increasing in successively more posterior pieces. This regional difference in stimulation is associated with length of piece and is not apparent in the long one-fourth pieces, but in the one-sixth, one-eighth and one-twelfth pieces it is distinct. The rate of metabolism in the first zodid of uninjured animals is highest in the anterior region and decreases more or less regularly DYNAMICS OF MORPHOGENESIS 437 in the posterior direction (Child ’*13.c). In the one-fourth pieces of table 2 this gradient continues to exist after section, the rate of the second fourth being lower than that of the first. In some other series the rates of first and second fourths immediately after section are about the same, that is, the original gradient has disappeared. In the one-sixth, one-eighth and _ one-twelfth pieces (tables 5, 9 and 13) we see that in general the susceptibility increases from piece to piece posteriorly, i. e., the original axial gradient is reversed. This can be due only to the fact that more posterior pieces are more strongly stimulated by section and this is actually the case. A comparison of the susceptibilities of Lot 1, which represents the most anterior fourth, sixth, eighth or twelfth in tables 2, 5, 9 and 13, shows that the susceptibility of this anterior region is but little affected by length of piece and undergoes but little increase after section until we reach the one- twelfth pieces, where the increase is greater than in the longer pieces. The more posterior pieces are much more affected by section and the tables show that in general the more posterior the level of a piece the greater the degree of stimulation following section. As regards the regional differences in rate immediately after section, the results of the susceptibility method have been con- firmed by estimations of CO, production very kindly made at my request by Dr. Tashiro with the apparatus devised by him (Tashiro 713). In one-fourth pieces or larger the second piece shows a lower rate of COs production than the anterior. In one-sixth pieces the anterior piece has the lowest rate of CO, production and the rate increases posteriorly from piece to piece. This regional difference in degree of stimulation agrees with the wellknown fact of observation that when such animals as earthworms and planarians are cut in two the posterior piece usually reacts much more strongly than the anterior and it is also in full accord with the theory of antero-posterior dominance developed in preceding papers (Child ’11 d, 713 b, ’13 ¢), in fact it constitutes valuable experimental evidence for this theory. The greater stimulation of posterior as compared with anterior regions in consequence of section can only mean that the rate of 438 Cc. M. CHILD metabolism in the anterior region is to a greater degree independ- ent of conditions in other parts of the body and so is but little altered when conducting paths are cut. The rate of metabolism in the posterior region on the other hand, must be in high degree dependent upon its connection with other regions, for when this connection is severed the stimulation of conducting paths increases its rate very greatly. The more anterior levels of the body are largely independent of the posterior while the more posterior levels are largely dependent upon the anterior. The degree of independence or subordination of a body-region then depends upon its level in the body. There is thus an axial gradient in degree of stimulation result- ing from section with its lowest point at the anterior and its highest point at the posterior end and this is evidently the expres- sion of a gradient in degree of subordination. That these gradients are expressions of the axial gradient in rate of metabolism described in the preceding paper (Child ’13 a) cannot be doubted. It was pointed out in that paper that the region of highest metabolic rate in a living cell or cell aggregate must be much more independent of other regions than they are of it and so must inevitably dominate other parts to a greater or less extent and within a greater or less distance from it. In Planaria where a simple axial gradient in rate from the head posteriorly exists the head is unquestionably the dominant region. As regards functional relations this is self-evident but it is also true at least for certain features of morphogenesis (Child ’11 d). As pointed out above, the stimulation gradient in pieces is merely an expression under special conditions of the axial gradient and of antero-posterior dominance. The increase in stimulation from section with increasing distance from the anterior end means simply increasing dependence upon stimuli coming from more anterior regions. 4. A remarkable relation between frequency of head formation and degree of stimulation after section exists. The greater the in- crease in rate of metabolism after section the less the frequency of head-formation. DYNAMICS OF MORPHOGENESIS 439 In the first paper of this series (Child ‘11 ¢) it was shown that the frequency of head-formation decreases as the length of the piece decreases and also from anterior to posterior regions. In the present paper we have seen that the degree of stimulation increases as the length of the piece decreases and also that the degree of stimulation is greater in posterior than in anterior pieces. It would seem that these two features, degree of stimulation after section and frequency of head formation, must be in some way associated or have a common foundation. But before the re- lation between them can be made clear it is necessary to find when in the history of a piece it is determined whether the piece shall give rise to a head or not. And consideration of this prob- lem, the time of head determination, must be postponed to an- other paper. The chief purpose of the present paper, however, is to point out this inverse relation between head frequency and degree of stimulation after section. The following paper will afford an insight into the nature of the relation. 5. After the temporary increase in rate of metabolism following section has disappeared a second slow and (relatively) permanent increase in rate occurs, in connection with reconstitution and this is a process of rejuvenescence resulting from the reorganization and reduction of the pieces. The writer has called attention elsewhere (Child ’11 a, 713 d) to the occurrence of rejuvenescence in connection with experi- mental and asexual reproduction in planarians. The suscepti- bility tables in the present paper show merely the earliest stages of this process at one hundred: and twenty hours after section (tables 4, 8, 12). This slow increase in rate in the pieces, which begins after three or four days and which indicates the beginning of rejuvenescence is relatively permanent, i. e., it disappears only gradually as the animal undergoes growth and becomes older in consequence of feeding. The degree of rejuvenescence is in general inversely proportional to the size of the piece and di- rectly proportional to the degree of reconstitutional change. In this connection another expression of the axial gradient in rate of metabolism may be mentioned, although it is not clearly shown in the tables. In general, the degree of rejuvenescence in recon- 440 Cc. M. CHILD stitution is greater in posterior than in anterior pieces of the same size, provided of course that both produce new whole animals. The amount of reorganization in the reconstitution of a piece undoubtedly increases as the level of the piece becomes more posterior. The amount of new tissue behind the head and the degree of change in the alimentary tract both increase from anterior to posterior pieces. In other words, the farther from the head a given region of the body is, the greater the alteration it must undergo when brought into direct relation with a dominant head region as every piece is when a head forms at its anterior end. IV. SUMMARY 1. A temporary increase in rate of metabolism, a ‘stimulation’ lasting at least several hours, results from section. 2. The temporary increase in rate of metabolism following section varies in amount inversely as the length of the piece, long pieces being very slightly or not at all stimulated and short pieces strongly stimulated. 3. The temporary increase in rate of metabolism following section also varies in amount according to the region of the body which the piece represents, being least in most anterior and in- creasing in successively more posterior pieces. In the shorter pieces this difference is sufficient to increase the rate of posterior above that of anterior pieces. 4. The differences along the axis in the stimulation from section are merely special expressions of the axial gradient and the antero-posterior dominance which results from it. 5. An inverse relation exists between the frequency of head formation and degree of stimulation of pieces following section. The shorter or the more posterior a piece the greater the degree of stimulation and the less the frequency of head formation. The nature of this relation will be considered after presentation of further facts. 6. The temporary increase in rate of metabolism in the piece after section is followed by a gradual fall. Twelve hours after section the rate of metabolism in the pieces may be as low as, or DYNAMICS OF MORPHOGENESIS 44] lower than, that in corresponding regions of intact animals. During the following thirty-six to forty-eight hours the rate of metabolism remains about the same in the pieces. 7. After three or four days the rate of metabolism in the pieces begins to rise slowly as reconstitution proceeds. This rise in rate is relatively permanent and constitutes physiological re- juvenescence. It disappears only as the animal increases in size and becomes older with feeding. Its amount depends upon the degree of reconstitutional change and the size and region of the piece. Provided complete reconstitution occurs the degree of rejuvenescence is greater in posterior than in anterior and in small ‘than in large pieces. LITERATURE CITED Cuitp, C. M. 191la A study of senescence and rejuvenescence based on ex- periments with Planaria dorotocephala. Arch. f. Entwickelungsmech., Bd> 31, Ho 1911 b Experimental control of morphogenesis in the regulation of Planaria.. Biol. Bull., vol. 20, no. 6. 1911 ¢ Studies on the dynamics of morphogenesis and inheritance in experimental reproduction. I. The axial gradient in Planaria doroto- cephala as a limiting factor in regulation. Jour. Exp. Zodél., vol. 10, no. 3. 1911 d Studies, ete. II. Physiological dominance of anterior over posterior regions in the regulation of Planaria dorotocephala. Jour. Exp. Zo6l., vol. 11, no. 3. 191le Studies, ete. III. The formation of new zodids in Planaria and other forms. Jour. Exp. Zodl., vol. 11, no. 3. 1912a Studies, ete. IV. Certain dynamic factors in the regulatory morphogenesis of Planaria dorotocephala in relation to the axial gradient Jour. Exp. Zodl., vol. 13, no. 1. , 1913 a Studies, etc. V. The relation between resistance to depress- ing agents and rate of metabolism in Planaria dorotocephala and its value as a method of investigation. Jour. Exp. Zodél., vol. 14, no. 2. 1913 b Certain dynamic factors in experimental reproduction and their significance for the problems of reproduction and development. Arch. f. Entwickelungsmech., Bd. 35, H. 4. 1913 ¢ Studies, etc. VI. The nature of the axial gradients in Pla- naria and their relation to antero-posterior dominance, polarity and sym- metry. Arch. f. Entwickelungsmech., Bd. 37, H. 1. 1913 d The asexual cycle of Planaria velata in relation to senescence and rejuvenescence. Biol. Bull., vol. 25, no. 3. Tashiro, S. 1913 A new method and apparatus for the estimation of exceedingly minute quantities of carbon dioxide. Amer. Jour. Physiol., vol. 32, no. 2. -_ i q ‘ iv - A i } 4 : —~ Py i a C3 dad a x fr . \ , S \ ; ie “ap \ i i r ; : f . 1 b | a be : , r. 4 j ma / a } f of CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, No. 247. ON THE STRENGTH AND THE VOLUME OF THE WATER CURRENTS PRODUCED BY SPONGES! G. H. PARKER Ever since Grant discovered the currents produced in water by living sponges, this activity has been a matter of interest to the naturalist. The arrangement of the canals within the body of the sponge, the disposition of the motor elements or choanocytes, and the control of currents by means of sphincters have all received much consideration. Very little attention, however, has been given to the currents themselves, their strength and volume. In a paper published in 1910 in the eighth volume of The Journal of Experimental Zodlogy, I recorded the fact that in Stylotella, a sponge from the coast of North Carolina, the pressure of its currents was equal to that of a column of water from 3.5 to 4 mm. in height. This sponge is of small size and generates only a slight disturbance in the surrounding water as compared with what may be seen in the vicinity of large sponges in the tropical and subtropical seas. Here the water often wells up so abundantly from the sponge as to deform the surface of the sea much as a vigorous spring deforms that of a pool into which it issues. Sponges of this description are frequently met with in the waters about the Bermuda Islands, and, during a recent sojourn at the Bermuda Biological Station, I took the opportunity of measuring the currents from some of these organisms. The method that I used in making these measurements was the same as that which I had previously employed for Stylotella. A whole sponge or a large portion of one was transported in a tub or bucket to the laboratory and a glass tube of appropriate size was tied securely into one of its oscula. Care was taken to select for this purpose an osculum whose canal system had been unin- 1 Contributions from the Bermuda Biological Station for Research, No. 32. 443 444 G. H. PARKER jured in detaching the sponge from its natural base or, if it had been necessary to break open the canal system anywhere, such breaks were closed by ligatures of soft string. The tube with the sponge attached was then held rigidly in place in the bucket of water and, after five minutes or so, the height of the water in the tube over that in the bucket was measured by means of a milli- meter rule attached to the side of the tube. Without disturbing the tube and its attached measure, the canal system of the sponge was then cut open so that the water in this system was in free communication with that in the bucket. The water in the tube immediately fell and, after it had come to a constant level, TABLE 1 ° Showing pressures in millimeters of seawater exhibited by the currents of seven species of sponges from the Bermuda Islands | HEIGHT OF COLUMN OF SEAWATER IN MILLI- AVERAGE NAR IO SEONG METERS FOR FIVE DETERMINATIONS HEIGHT Tethya sp. Close to T. seychellen- sis (HP. Wright) 2.2. Seen | Spirastrella sp. Close to S. vaga- | bunda, var. tubulodigitata) 2.5 2.0 2.0 2.2 bo or bo oO Dendy =. saa ees ke. eee 2.5 2.0 2.0 2.5 2.0 DD) Pachychaling sp-:.-.---<---eeee- | re 2.0 1.5 1.0 120 1.3 Spinosella sororia (Duch.et Mich), 3.0 Deby 3.0 3.0 3.0 2.9 Medanta spies eee [het ea. Se eeee | 2.0 2.0 2.5 2.0 2.0 2a Stelospongia sp................5- | 25. 2.0 2.0: =e 2.1 3.0 3.0 2.0 2:5 3.0 ar | Hircinia variabilis F. E. Schulze.. | a reading of its height was again taken. The difference between this reading and that made before the sponge was cut open was assumed to be a measure of the pressure of the water current produced by the sponge. Determinations of the current pressures by the method just described were made in seven species of sponges, and the details of of these measurments are given in table 1. At least five deter- minations were made for each species. I am under obligations to my friend, Prof. H. V. Wilson for the identification of the sponges upon which these determinations were made. The surprising feature of the determinations in the table is that they all indicate a very low current pressure, lower even than that WATER CURRENTS PRODUCED BY SPONGES 445 of Stylotella. Notwithstanding this general lowness of pressure one of these species, Spinosella sororia, the common Bermuda finger-sponge, produces in the shallow sea a conspicuous deforma- tion of the natural surface of the water, and this deformation can also be observed in specimens that have been transported to the laboratory. As Spinosella was a very satisfactory species to work with, a determination of the volume of water that passed through a single finger of this sponge was attempted. This was made by sinking the tube that had been fastened into the osculum of a finger of Spinosella under the surface of the seawater and thus allowing the water driven by the sponge to escape. ‘The escaping current deformed the surface of the water in the bucket in much the same way as is often seen over a vigorous sponge finger in its natural position in the sea. The rate of the flow of water from the finger was then determined by measuring the velocity of floating particles, such as carmine and so forth, that were carried up the tube by the water current. This proved to be very close to 20 mm. in five seconds. As the diameter of the tube was 17 mm. the finger must have discharged a little over 4.5 cc. of water in five seconds. At this rate the discharge would amount to some 78 liters a day. This finger measured about 10 cm. in length and averaged 4 cm. in diameter; its osculum had a diameter of about 2 ecm. Spinosella colonies often consist of as many as twenty such fingers and, assuming that all the fingers work at the same rate as the one upon which the measurements were made, such a colony as a whole would strain in a day about 1575 liters of sea- water or over 415 gallons. : Since all the measurements recorded in the table were made upon sponges that had been removed from their natural attach- ments and transported to the laboratory and since the effect of such treatment is known to retard rather than accelerate the cur- rents, the results therein recorded are believed to be below rather than above the actual working capacity of the several sponges. This opinion is borne out by an observation made on Spirastrella for which I am indebted to Mr. W. J. Crozier and Mr. D. H. Wenrich. A specimen of this sponge was found partly exposed at low water. Although the osculum was out of water, there was 446 G. H. PARKER still a free flow from it. A cut was made on the side of the sponge at the level of the outer seawater and the sponge was taken to the laboratory. Close measurements showed that the edge of the osculum over which the water lifted by the sponge had been flowing was 4 mm. above that of the sea level, demonstrating that this sponge in the undisturbed condition in which it was found had been overcoming a pressure of 4 mm. of seawater. As the highest pressure that had been obtained from Spirastrella in the laboratory was 2.5 mm., it seems probable that the laboratory determinations for all the species tested are a little below the actual maxima for this form of activity, though, for reasons already given in my former paper on Stylotella, they cannot be much below this maximum. From these observations it is clear that the currents produced by sponges consist of relatively large volumes of water flowing at low pressure and that currents of this kind are capable of producing such deformations of the surface of the sea as can be observed above large sponges. It would be interesting to ascertain what proportion of the suspended material in the seawater is screened out as the current passes through the sponge, but no observations on this point were made. MODES OF INHERITANCE IN TELEOST HYBRIDS H. H. NEWMAN From the Hull Zoélogical Laboratory, University of Chicago, and the Marine Biological Laboratory, Woods Hole THIRTY-EIGHT FIGURES CONTENTS PPROLOGUCTLON «2. <7 cee 5 + EI cio = one oes Sapa ciighg Shines Sey 447 Historical review of the work of Appellof, Moenkhaus, Kammerer, Bancroft and Newman on teleost hybridiMameat .. 6... -.- 0.5 eee e cece een ceees 448 New experiments: Reciprocal crosses between species of the genus Fundulus and the genus CyprimodOn@ie ame... 0... cc ccc ee ce cae cee eee ee ees 457 a Material and methods: Ae... cee ice ge oe we eee eo ee eee 457 . Tabulation of data.. a. De he Rha ne hs RRC: Es, AEAOO 3. Summary and discus of Bia. adi en as a ORS Me Ted en 467 Inheritance of pigment charaeters in Fundulus hybrids..............--.--.-- 475 1. A statement concerning pigment characters in Fundulus species. ...... 475 Pee TADUIATIONS Of CataMMMMNP : 2260606 f. oe'os ce eed ees a aacle neg ye diag: 478 3. Summary and discussion of data on pigment inheritance............... 482 General summary of conelusiOMs.....................--+--- Sree he 2) he 486 1. The nature of the influence of foreign spermatozo6n on cleavage and CauEty CLEVCl O[) EMMYS ¥5.2 och. oy sac caus an dioleids oqabteawvelssneue ole,éc2 sewers 486 2. Harmful versus beneficial effects of foreign spermatozoon.............. 487 3. Phylogenetic ir @ of parent species and success in development of hybrids. . RN as, access ae hea ar oo ee Ete ee eae ee 487 4. Modes of inheaie ance in Pieleast hy aly aad ie Mendelian hypotheses OlAG Orin ANP TEE AULON 971.) leis coc aclo cae: Se Okie crete silere es 488 5, Inheritance immreeipmpcal crosses...)......6...2... 00002 e eee ee eee es 489 Ean ona iyi > ae Soo en ocd sel atheyin ss OSAl Gale one se Sees Pee ee 489 INTRODUCTION The present contribution has a twofold purpose, to review and discuss the literature on teleost hybrids and to present a new body of experimental data, which seem to throw added light on various mooted questions’and to lead to rather definite conclu- sions concern?ng certain problems of inheritance. As material especially available for the study of the early stages of hybrid development the teleosts are unexcelled by any \ 447 THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 4, way, 1914 448 H. H. NEWMAN other group of animals known to the writer, although a much larger number of investigators have chosen to work with echino- derms. ‘Teleosts possess decided advantages over the echino- derms in that they intercross in practically any combination and without the aid of the reagents used in artificial parthenogenesis, which are necessary ir order to cross echinoderms, and which may affect the character of development and thus interfere with the study of heredity. It is, moreover, a simple matter to rear the young of teleosts to a stage in which many definitive charac- ters can be studied, while hybrid echinoderms have never to my knowledge been reared beyond a comparatively early larval stage. In addition to these two fundamental advantages tele- osts afford especially good material for this type of investigation because the large number of species in any neighborhood give a very wide range of possible combinations; because the eggs and young are large and easily handled; and because the con- trasting characters are more numerous and more clearly defined than are those of echinoderms. HISTORICAL REVIEW OF THE WORK OF APPELLOF, MOENKHAUS, KAMMERER, BANCROFT AND NEWMAN ON TELEOST HYBRIDS The earliest published work on teleost hybrids that has come to my attention is an important paper by Appelléf that appeared early in 1894. I regret that this paper escaped my notice until after the publication of my two former papers on Fundulus hybrids, for it would have been highly suggestive for certain phases of my own work. The paper, although rather generally overlooked by investigators, should be’ of considerable interest to all those who are engaged in the study of the developmental mechanics of heterogenic hybrids. Appelléf was probably the first to attempt crosses between very distantly related species. His combinations include such ill-assorted matings as the stickle- back and the flounder, the cod and the cunner, the cod and the flounder and others equally bizarre. In each case he found that, even though a mature hybrid between the two parent types be unthinkable, development in the cross fertilized eggs proceeds MODES OF INHERITANCE IN HYBRIDS . 449 for a considerable period, in most cases through the cleavage period, in a normal manner. In all cross bred eggs the cleavage seemed to be identical with that of pure bred eggs except that in rare cases the time rate of the cleavage process was markedly retarded. This was especially noticeable in the hybrids from Spinachia eggs fertilized with Gasterosteus sperm. An observation of Appelléf’s, which to my mind is of consid- erable morphological significance, is to the effect that the most critical period in the development of heterogenic hybrids is the period of transition from cleavage to germ layer formation—the period of ‘gastrulation.’ In comment upon this finding I may say that my own results have for a number of years forced me to similar conclusions and that I have on hand an investigation dealing at some length with this point. It is of more than pass- ing interest to note how this idea lines up with the rather prev- alent opinion that the process of cleavage is purely a function of the egg, and that the spermatozo6n, though codperating with the egg nucleus in the mitoses of cleavage, does not begin to influence the process of heredity until during the formation of the gastrula. It would appear that even a very distantly related sperm may stimulate development in an egg and coéperate har- moniously with the latter through the period of cell multiplica- tion, but that when the hereditary functions of the germinal elements come into play discordant chemical interactions take place that effectually disrupt developmental harmony. ‘This barrier to further development I have found to be less sharply drawn than Appellof supposed, as will become evident when the data subsequently presented in this paper claim the readet’s attention. Next in chronological order comes the work of Moenkhaus (’94) in which he relates the results of his crosses between Fundulus heteroclitus and Menidia notata. This work had as its object the study of the behavior of maternal and paternal chromatin in hybrid eggs and embryos in which the chromosomes of the parental species were visibly different in form. The observa- tions show that the paternal chromosomes retain their identity in the zygote and for a considerable period remain in a close 450 H. H. NEWMAN group sharply marked off from the maternal group. Subse- quently, however, the two types intermingle, though still retain- ing their specific characters. The rate of cleavage was, accord- ing to Moenkhaus, the same as in pure bred eggs of the same species. It is a noteworthy fact that when Fundulus eggs are fertilized with Menidia sperm, the period of gastrulation is safely passed and a well-defined, though short and defective embryo is formed although the two species represent two distinct orders of teleosts. Kammerer (’07) published a rather extensive account of hy- bridizing two species of perch, Perca fluviatilis and Acerina cernua. In brief his report is as follows. In the spring of 1905 a fisher- man brought to the laboratory two specimens of an unknown type of perch, which, at the suggestion of Doctor Przibram, were adjudged to be hybrids between Perca fluviatilis and Acerina cernua, because in a number of respects they appeared to be intermediate between these two species. In order to test the correctness of this diagnosis it was decided to intercross the hypothetical parent species in order to determine whether the hybrids resembled the questionable wild individuals. It was found that the two species did cross successfully and that a number of young fish hatched and grew well for about six months, at which time they showed some characters which slightly re- sembled those of the supposed adult wild hybrids. The latter were successfully crossed, since they were both fertile females, with males of both Perea and Acerina, the young resembling closely the paternal species in each case. This is the only case on record, so far as I am aware, of fertile hybrids being produced by crossing two genera. It is rare, in fact, to find undiminished fertility in hybrids between two closely related species of the same genus, and even inter-varietal hybrids frequently show incomplete fertility. Had Kammerer reared to maturity and successfully bred the unquestioned hybrids which he obtained he would have made good his case, but as the matter stands there is considerable cause to doubt the validity of his conclusions. The chief value of the paper in the present discussion lies in the fact that there is presented a considerable array of data, MODES OF INHERITANCE IN HYBRIDS 451 tending to show that in practically all characters examined the hybrids reared in the laboratory are strictly intermediate between those of the parent species. This is true for such matters as numbers of scales in a row, numbers of fin-rays, numbers of teeth, and other integral variates. Characters involving dimen- sions show a similar blending between the two species. No men- tion is made of the discovery of any type of inheritance sugges- tive of Mendelian dominance. In addition to the crosses mentioned above Kammerer made a number of heterogenic crosses between teleost species,: the results of which led him to conclude that the degree of success in the development of hybrids depends not upon the closeness of relationship of the parent species but upon their similarity of habitat. The fallacy of this conclusion is evident in view of the success of Moenkhaus and the writer in cross fertilizing prac- tically any two species of teleosts. The first attempt to follow in detail the development of teleost hybrids. was made by the writer (Newman ’08). This paper dealt exclusively with the reciprocal hybrids between Fundulus heteroclitus and Fundulus majalis. The former species develops nearly twice as rapidly as the latter and the reciprocal hybrids both have a developmental time rate intermediate between those of the parent species, though more nearly like that of the egg than that of the sperm species. AIl characters involving the time factor show blended inheritance, but many characters not dependent on the time factor were shown to exhibit more or less typical Mendelian dominance. Some of the dominant characters noted were as follows: F. heteroclitus type of chromatophore dominant over that of F. majalis; size of young on hatching or at the time of maximum development that of F. heterocli- tus in both reciprocal crosses; the rate of cleavage nearly pure maternal; the viability of young hybrid larvae, when capable of hatching, equal to or greater than that of the more viable species, F. heteroclitus; resistance to lack of oxygen or presence of carbon dioxid of F. heteroclitus egg hybrid as great or greater than that of the more resistant species; susceptibility to these reagents greater in F. majalis egg hybrids than in the more susceptible 452 H. H. NEWMAN species, F’. majalis; the depth of pigmentation of young F. heter- oclitus egg hybrids as pronounced or more so than the darker species, F. heteroclitus. These cases in which the hybrids seemed to show an exaggeration of dominance exemplify the phenome- non of ‘hyperdominance.’ In view of the fact that at the time when this paper was written the new Mendelism was sweeping all before it, I took pains to emphasize especially all types of inheritance which appeared to me to be essentially non-Mende- lian. This stand was taken advisedly in the hope of checking the growing impression that all inheritance was on analysis Mendelian, and that cases of apparent blending, even in the F, generation of hybrids, was the result of incomplete analysis of the factors involved. Chief stress was laid upon the importance of studying heredity as a process, of observing the genesis of characters rather than limiting observation to the definitive conditions as they appear in adults. In this connection it was noted that maternal and paternal influences showed alternating periods of predominance, so that a character might appear as a dominant at one time only to be superseded by the character that had apparently been recessive. The definitive condition represents only the end re- sult of a struggle between opposed parental tendencies whose ups and downs may be observed from day to day in developing hybrid embryos. The question as to whether the foreign sperm exercises any influence upon the rate and character of early cleavage was in this paper answered in the negative for the reason that in all early experiments the differences were very small and not always in the same direction. This paper was followed nearly two years later by another dealing with the same species of fish (Newman ’10), in which the question as to the influence of the spermatozo6n on early devel- opment was reéxamined. Very searching statistical methods were applied to the relative rates of early cleavage in pure and hybrid F. majalis eggs. It was found as the result of five experi- ments involving very large numbers of eggs, that there was a slight but constant acceleration of cleavage as the result of the MODES OF INHERITANCE IN HYBRIDS 453 use of foreign sperm. On this score I took occasion to criticise the position of Conklin, Godlewski and others, who, on the basis of the work done on echinoderm hybrids, had declared that the early development of hybrids is purely maternal and that it is only in stages later than the gastrula that the influence of the spermatozo6n begins to make itself felt. There was in my ex- periments a real effect of foreign sperm on the rate of early devel- opment and this effect was not due to a more speedy impregna- tion of the egg membrane by the foreign sperm head, as was shown by the experiment in which all spermatozoa in both hybrid and control cultures were destroyed with distilled water within less than five minutes after insemination. By this time all spermatozoa had entered the eggs, those of the foreign and native species having entered with equal promptness. These results were criticised by Godlewski (’11) because they were not confirmed by cytological studies. He did not explain, however, just what he meant by a cytological confirmation of a statistical treatment involving many thousands of eggs, no small sample of which could be expected to show anything of value. On the basis of what seemed to be a demonstration that foreign sperm actually effects a slight change in the rate of early cleavage, the writer entered upon a general discussion of the role of the spermatozo6n in early development and came to the conclusion that, in addition to initiating development, the pater- nal germ plasm reacts almost immediately with that of the egg so as to change the rate and character of metabolism and thus accelerate or retard cleavage. Naturally the cytoplasmic organ- ization of the egg imposes upon the young embryo certain mor- phological restrictions, but in all those matters that have to do with rate of chemical reaction the spermatozoén begins to exert an influence as soon as fertilization occurs, as soon as the two ontogenies are engrafted one upon the other. The work of Godlewski, Baltzer and others, who had experimented with echi- noderm hybrids, was cited by the writer in confirmation of his conclusions and it is to be regretted that, in citing Godlewski’s results certain misstatements were made, for it was in response to these inadvertent inaccuracies that Godlewski took the whole — 454 H. H. NEWMAN paper rather severely to task. The fundamental contention was, however, scarcely touched upon by this critic, as was shown by the writer (Newman ’11) in a reply to Godlewski, in which the minor inaccuracies in reporting the latter’s results were admitted and differences of interpretation were brought clearly upon a controversial plane. It may then be eonsidered as fairly well established that, at least in the case of hybrids between closely related species, the sper- matozoon exercises at first a slight and then a progressively more pronounced influence upon the rate of early development, but fails to exercise any really hereditary function until the embryo begins to differentiate tissues and organs. This point of view will, I believe, be justified by the new data herewith presented. This somewhat radical change of front on my part is the result of sev- eral years of study and has been forced upon me especially by the experiments about to be reported, the results of which are herewith to some extent anticipated. In the Proceedings of the Indiana Academy of Science for 1910, published a year later, Moenkhaus reports upon an exten- sive series of experiments dealing with ‘‘Cross fertilization among fishes.”’ These range from extreme heterogenic crosses between different orders of teleosts to those between closely related spe- cies of the same genus. A useful summary of the chief facts and conclusions is furnished and is here quoted: 1. The eggs of any of the species of teleosts tried may be impregnated by the sperm of any other species tried. 2. The number of eggs fertilized is usually great, i.e., 75 per cent or more. This bears no relation to the nearness of relationship of the two species concerned. 3. Normal impregnation is the rule, di- and polyspermy being the exception. 4. Development in its early stages proceeds normally, the deleteri- ous effects of the two strange sex products upon each other showing only at later cleavage or subsequently. | 5. The rate of development in the early cleavage stages is always that of the egg species. Any effect of the strange sperm upon the rate of development shows itself by slowing the process regardless of whether the rate of the sperm species is faster or slower than the egg species. A period of great mortality in the developing hybrids is gastrulation. MODES OF INHERITANCE IN HYBRIDS 455 6. If the heart is formed, although it pumps no blood, the embryo may remain alive for a considerable period, yolk absorption taking place to a varying degree. If the heart handles blood and the blood vessels are differentiated, the embryo is likely to develop to the point of hatching. 7. The numerous abnormalities appearing in the hybrid embryos are due to a deterioration in the developmental processes, resulting prob- ably from the poisonous action of the sex products upon each other. 8. The success of the hybrids, i.e., the stage to which a given hybrid will develop, is correlated with the nearness of relationship of the two species used. 9. The mixing of unrelated sex products is looked upon as analogous to the transfusion of unrelated bloods, the more distantly related the two species concerned the greater their toxicity. For the most part the results of Moenkhaus confirm the earlier results of Appell6f and extend the horizon of our knowledge of the field in question. I am unable, however, to accept any of the last six conclusions listed above, and will present my reason for thus dissenting after presenting my new results. Bancroft (12) has made the most recent contribution to the literature on the hybridization of teleosts. He repeated the ex- periments published by the present writer in 1908, intercrossing Fundulus heteroclitus and F. majalis, paying especial attention to the heredity of pigmentation. The avowed object of the experiments was to reéxamine the hybrids “from the Mendelian and physiological points of view.’”’ Bancroft confirms my obser- vations and in most points agrees with my conclusions. In his search for evidences of Mendelian dominance he overlooks many intermediate conditions. In one instance he attempts to explain what had seemed to me a perfect case of blended inheritance as the result of the interaction of two separate unit factors. He confirms my results and conclusions very exactly when he con- cludes that ‘‘in general, characters connected with rate of devel- opment show blended heredity and it may be that such charac- ters are so intimately associated with extra-nuclear substances such as the yolk that complete dominance is not obtained.” It is noteworthy, in view of Moenkhaus’s statement ‘“‘that any effect of the strange sperm upon the rate of development shows eitself by slowing the process of development regardless of whether 456 H. H. NEWMAN the rate of the sperm species is faster or slower than the egg species,’ that Bancroft takes the position that: As regards the rate of development of the embryos my observations confirm those of Newman on most points. The development of the F. heteroclitus egg hybrid was shown to be slower than that of its maternal parent; and the development of the F. majalis egg hybrid, during the early stages, was faster than that of the pure F. majalis. After hatching the F. heteroclitus egg hybrid seemed more vigorous and grew faster under like conditions than either of the pure forms. It is somewhat strange that Moenkhaus failed to note these facts for he must have performed the experiment a number of times. Bancroft’s paper adds materially to our knowledge of the heredity of pigment in teleost hybrids. His analyses of these processes is more nearly adequate than that of any other writer. In the general discussion of new data I shall have occasion to come back to this paper for more detailed treatment. This then is the history of the development of our knowledge of heredity in teleost hybrids. From this array of facts and theories it is possible to extract certain well-defined problems and unsettled questions, which it is my intention to put to the test of further experiment. A list of some of the more important questions is herewith introduced in order to clarify the issue and to hia the attention of the reader upon -the problems involved. . At what point in the development of an embryo does the sperm begin to exercise an hereditary as opposed.to a me chemical effect upon the character of development? 2. Is the rate of cleavage ‘inherited,’ in a strict sense, or is it merely a function of egg size and of yolk content? 3. Is the effect of foreign sperm to be looked upon as neces- sarily deleterious, and hence toxic, or is the effect bad only when there arises so pronounced a disharmony of the two engrafted ontogenies that no blending or compromise can be reached? 4. Is it true that all characters concerned with or resulting from the rate of development are inherited in the blended fash- ion? Is rate of development strictly an hereditary character? 5. Is the degree of success in the development of hybrids in any way closely correlated with the nearness of relationship of MODES OF INHERITANCE IN HYBRIDS 457 the two species used? If not, upon what factors does success in crossbreeding depend? 6. Are structural or physiological characters inherited in the exclusive or in the blended fashion in the first generation of hybrids? 7. What is the function of the spermatozo6n in early develop- ment? Is its initial effect merely equivalent to that exercised by mechanical and chemical agents used in artificial partheno- genesis, or is this only part of its function? 8. Is there any truth in the analogy suggested by Moenkhaus between the effects resulting from mixing strange sex products and those resulting from the transfusion of unrelated bloods? Each of these questions will receive its proper share of atten- tion elsewhere in the paper. NEW EXPERIMENTS: RECIPROCAL CROSSES BETWEEN SPECIES OF THE GENUS FUNDULUS AND THE GENUS CYPRINODON 1. MATERIALS AND METHODS The four species of fish used in the present experiments belong to two genera of the family Poeciliidae and are common in the waters about Woods Hole. They are Fundulus heteroclitus (figs. 1 and 2), F. majalis (figs. 3 and 4), F. diaphanus (figs. 5 and 6), and Cyprinodon variegatus (figs. 7 and 8). F. heteroclitus, F. majalis and Cyprinodon are marine or brackish water species, while F. diaphanus is strictly a fresh water form though tolerant of slightly brackish water. In some regions F’. heteroclitus is found with F. diaphanus in fresh water ponds into which the sea flows at high tide. The eggs of F. majalis are the largest, averaging about 2.7 mm. in diameter, those of F. diaphanus come next with an average diameter of 2.3 mm., those of F. heteroclitus average 2 mm., and those of Cyprinodon are much the smallest, averaging scarcely 1.5mm. The yolk of F. majalis eggs is denser and of deeper yellow color than that of the other species; that of F. heteroclitus is denser and yellower than in F. diaphanus but less so than in F. majalis; that of Cyprinodon is almost colorless. 458 H. H. NEWMAN During the months of June and July all of these species are spawning in the Woods Hole region and little difficulty is experi- enced in obtaining abundant material for crossbreeding. The most successful experiments were those in which insemination was accomplished by the dry method. In this way it was possi- ble to avoid exposing either of the sexual products to waters of harmful concentration, for eggs were stripped from well dried females into bowls containing no water and they were mixed with fresh milk or macerated testis. After allowing a few min- utes for the sperm to impregnate the eggs, water of the proper kind was added and excess sperm washed out. It was found best to rear the eggs of F. heteroclitus, F. majalis and Cypri- nodon in sea-water and those of F. diaphanus in fresh water. All thrive well in about 25 per cent sea-water, but do no better than when the two kinds of natural water are used. Likewise all four kinds of eggs thrive in fresh water, but the eggs of F. diaphanus do not do well in natural sea-water, in which they undergo various degrees of plasmolysis. No less than three crosses of each type were performed and careful comparisons were made possible through the frequent camera drawings of individuals and details. The illustrations herewith published are taken from camera drawings which seem to represent the most typical condition. The method of pre- senting the data is as follows: A chronology is given for each cross and beside it in an adjacent column are given the details of synchronous stages in the development of the control or pure bred eggs of the maternal species. Only such details are men- tioned as seem of particular significance as diagnostic of relative rates of development or types of inheritance involved. In order that repetition may be avoided the chronologies for the recip- rocal hybrids between F. heteroclitus and F. majalis are omitted, and the reader is referred for the facts to the previous paper in which these crosses were first dealt with (Newman ’08). The references to details of pigmentation are intended to be merely suggestive, since the study of pigment inheritance forms a spe- cial topic for subsequent discussion. 2. TABULATION. OF DATA - TABLE 1 a. Crosses between species of the same genus Comparing the hybrids from F. diaphanus 2 X F. heteroclitus ¢ with pure bred 2 hours 6 hours 20 hours 2 days 3 days 4 days 5 days PURE F. DIAPHANUS (CONTROL) First cleavage Advanced cleavage Blastoderm with germ ring faintly defined, embryonic shield barely visible (fig. 10) Germ ring halfway around the yolk, embryonic axis well de- fined, no neural tube (fig. 13) Blastopore closed, embryo with short tail, lenses of eyes len- ticular in form, mid-brain broadly open, no heart-beat or circulation, first chromato- phores under hind-brain and on adjacent yolk Heart-beat beginning to be established in nearly all em- bryos, no circulation, scatter- ing stellate dark chromato- phores on top and sides of brain, black yolk chromato- phores scattered over whole yolk, a few small red yolk chromatophores of dull orange color (fig. 16) Embryos about like those of pure F. heteroclitus at 4 days, | heart-beat strong, circulation | well established, head chro- | matophores all brownish and | small, large black yolk chro- matophores arranged along the principal vitelline vessels, red yolk chromatophores large, stellate, and dull red- dish brown 459 F. diaphanus F. DIAPHANUS 9 X F.HETEROCLITUS * Same A larger proportion in more ad- vanced stages than in control Germ ring well defined and well established embryonic shield (fig. 11) Germ ring nearly around yolk, embryo long and with brain and optic vesicles well defined (fig. 14) Blastopore closed, tail long, lenses spherical and partially enclosed in optic cup, entire brain closed, heart beating in majority of embryos, much more numerous and better de- fined chromatophores on brain and yolk Heart beating strongly, yolk and body circulation well estab- lished, pigmentation much more advanced in every partic- ular than in control, red yolk chromatophores redder than in control (fig. 17) | Embryos less markedly in ad- vance of control than at 4 days, a few large black head chroma- tophores of the F. heteroclitus type have appeared along with numerous small brownish head chromatophores of the F. dia- phanus type, red yolk chro- matophores intermediate in color between the red-brown of the maternal and the brick red of the paternal species 450 H. H. NEWMAN TABLE 1—Continued F. DIAPHANUS 9 XF. HETERGCLITUS co 1 week | Embryo as in fig. 19, almost Embryos as in fig. 20, tail shorter even with the hybrid strain in than in control, other details stage of development, other | as before details about as before TIME PURE F. DIAPHANUS (CONTROL) 9 days | Embryos comparatively light | Embryos much darker than con- | colored, tail long and slender | trol, decidedly larger and more advanced 12 days | None hatched | Nearly all hatched 13 days A few hatched | All hatched and very active 14 days About half hatched 15 days All hatched 3. weeks Larvae vigorous and would Larvae larger and more active doubtless live indefinitely than control, evidently grow- ing faster under like conditions MODES OF INHERITANCE IN HYBRIDS TABLE 2 461 Comparing the hybrids from F. heteroclitus 2 xX F. diaphanus & with pure bred 2 hours 4 hours 20 hours 52 hours 3 days 4 days 5 days 1 week 13 days 14 days 3 weeks F. heteroclhitus PURE F. HETEROCLITUS (CONTROL) | First cleavage | 32 to 64-cell stages _ Advanced cleavage but no germ ring visible Blastopore closed or nearly so in all embryos, pale chroma- tophores appearing under hind-brain, a few red yolk chromatophores on yolk near head, no black yolk chroma- tophores Heart beginning to beat feebly | in nearly all embryos, no vi- telline circulation. A few black yolk chromatophores, and fewer head chromato- phores Vitelline and body circulation established and all types of chromatophores abundant Embryos nearly as advanced as those of F. diaphanus at 6 days _ Body shorter and stouter than in hybrid, otherwise little dif- ferent except in details of pig- . mentation A few hatched _ All hatched | Larvae vigorous F. HETEROCLITUS 2 XF. DIAPHANUS oO Same Like control except that a small percentage of eggs are less ad- vanced than any of the pure bred eggs, having 16 cells or less Some embryos show the early steps in germ ring formation, others are less advanced than any of the control The normally developing embry- os noticeably more advanced and more deeply pigmented than controls, F. diaphanus type of head chromatophores appearing on sides of mid- brain, afew black yolk chroma- tophores Heart beating quite strongly in a few embryos, the majority about as in control, a consid- erable percentage of abnormal, retarded embryos derived from the eggs that showed lagging cleavage at 4 hours Almost identical with control in stage of development except the abnormal specimens, which from here on will be ignored as they never hatch; the normal embryos more heavily pig- mented than control Very much like control; pigment characters given in detail later Body rather more slender than control, a few more deeply pig- mented than any of the con- trols The majority hatched All hatched except abnormal specimens Larvae indistinguishable from control except in details of ig- mentation 462 H. H. NEWMAN TABLE 3 Showing development of hybrids from F. diaphanus 2 X F. majalis 3; for control 2 hours 6 hours 20 hours 2 days 3 days 4 days 5 days 1 week 9 days 12 days 13 days 14 days 15 days 16 days 3 weeks see first column of table 1 | First cleavage like control | A smaller proportion of the more advanced stages of cleavage than in control _ Blastoderm much smaller than in control, germ ring scarcely dis- tinguishable, no embryonic shield (fig. 9) Germ ring one-third around the yolk, embryonic shield small and only slightly elevated from germ ring, embryonic axis not estab- lished (fig. 12) Blastopore still open, in some cases germ ring is only three-fourths around yolk, optic vesicles flat with no optic cup, no lenses, no enlargement of hind-brain, no chromatophores Embryo much less advanced than in control, no heart-beat, no body or yolk pigment (fig. 15) Embryo about as advanced as control was at 4 days, chromato- phores on head and yolk few as compared with control and all more delicately branched as in paternal species, no vitelline cir- culation but the majority of embryos have heart-beat established. Nearly half of the embryos appear at this time as if they were abnormal The normal embryos have developed very rapidly since the estab- lishment of vitelline circulation and are now nearly on a par with the control; many specimens, however. have failed to estab- lish vitelline circulation and appear anemic and deformed _ Even the more advanced embryos have lagged behind control; a number of abnormal embryos have died from anemia, the most pronounced differences between hybrids and controls have to do with pigmentation ; None hatched None hatched A very few hatched About a fourth hatched, larvae not very active _ Nearly all of the normally developing embryos hatched; a few have died without hatching Nearly all larvae dead, the few that have survived are not as large or as vigorous as the pure bred larvae of the same age, living under identical conditions TIME MODES OF INHERITANCE IN HYBRIDS TABLE 4 PURE F. MAJALIS (CONTROL) 3 hours 6 hours 22 hours 2 days 3 days » 4 days 5 days 1 week THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, NO. In 2- and 4-cell stages 16 to 32 cells Blastoderm beginning to form faint germ ring Germ ring about halfway around yolk, triangular embryonic shield, embryonic axis a thin | line Germ ring nearly around yolk but blastopore still widely open, optic vesicles flat with- out optic cups, no lenses, no mesoblastic somites Embryos with many somites, well differentiated brain ves- icles, no heart-beat, no circu- lation, no pigment Embryo large, heart-beat and circulation well established, no chromatophores Embryos very much larger than hybrids, a few slender black chromatophores on sides of mid- and hind-brains, slen- der black chromatophores on yolk, no red yolk chromato- phores 463 F. MAJALIS 9 XF. DIAPHANUS o Same Mostly 32-cell stages Blastoderm much broader than in control, germ ring well de- fined, embryonic shield begin- ning to appear as a thickening of germ ring Germ ring in more advanced em- bryos as much as two-thirds around yolk, embryonic axis more sharply defined than in control Blastopore closed, optic cups formed, lens flat, 4 to 7 somites Embryos smaller but more ad- vanced than control, no heart- beat or circulation, scattering slender black chromatophores on yolk, none on body Embryos about as advanced as control in general develop- ment, but much smaller in size, all show the paternal influence in pigmentation, since there are both black and red chro- matophores on yolk and black chromatophores of the F. dia- phanus type, or nearly so, on the sides of the hind-brain Hybrids much smaller, eyes paler, head chromatophores almost pure paternal in type, black yolk chromatophores almost pure paternal, red yolk chroma- tophores intermediate in form and color between two parental types. Red chromatophores in three rows on side of body as in F. diaphanus 464 H. H. NEWMAN TABLE 4—Continued PURE F. MAJALIS (CONTROL) F. MAJALIS 9 XF. DIAPHANUS 18 days 20 days Hybrids between F 5 hours 18 hours 24 hours | | Head and body chromatophores | On the head the F. diaphanus all very finely branched and type of chromatophore pre- scattering, yolk chromato- dominate, but there are always phores both red and black some of them with a tendency very slender to branch like those of the ma- ternal species; red yolk chro- / matophores much larger than in control and of an interme- diate form and color. Body of embryos beginning to appear anemic on account of the inter- ference with circulation pro- duced by stretching of the heart, as in other F. majalis egg hybrids described else- | where i Several hatched Very little change since 9 days except that a number of the _ weaker embryos have died , All hatched Many dead, others very feeble. There has been very little ad- vance since the first week of development except in the case of the chromatophores. None hatched in any experiment b. Crosses between two genera TABLE 5 . diaphanus Q X Cyprinodon variegalus About half of the eggs fertilized and in 2- and 4-cell stages, a few obviously polyspermic, others indistinguishable from control Advanced cleavage of regular character in the majority of develop- ing eggs, but polyspermic eggs show great irregularity and con- siderable cytolysis All monospermic eggs normal in appearance but less advanced than control, the beginnings of cytolysis in a few blastodiscs All eggs show cytolysis of marginal cells of blastodise, indicating that development has ceased and disintegration has set in. The limit of development in this cross seems to be the end ofthe cleavage period although a few embryos have begun to show the first indications of a germ ring, indicating that gastrulation or germ layer differentiation had begun but could progress no further MODES OF INHERITANCE IN HYBRIDS 465 TABLE 6 Hybrids between Cyprinodon variegatus 2 X F. diaphanus 3 2hours About 75 per cent of eggs cleaving and 4- to 8-cell stages. Cleavage regular without any indications of polyspermy 6 hours _ All fertilized eggs in advanced cleavage stages, like control but less advanced 20 hours | A few of the eggs that had cleaved are dead and part ally disinte- grated, but the majority have reached the end of the cleavage period and are beginning to show evidences of gastrulation. All are decidedly less advanced than control, in which all eggs are in germ ring stages 26 hours | Nearly half of the embryos that had reached the stage described above now show cytolysis, the others are still alive and in early germ ring stages 2days | Only four embryos still alive. These have the germ ring one-third to one-half around the yolk, a triangular embryonic shield and a well-defined embryonic axis. Many embryos have died in early germ ring stages 3 days Two embryos alive. These have blastopore nearly closed but show very little embryonic differentiation 4 days | Both embryos dead without having advanced beyond the condition noted on the previous day. It is possible for a few individuals of this cross to live practically through the period of gastrula- tion, but they are unable to enter upon embryonic differentia- tion. It is to be noted, however, that this cross succeeds much better than the reciprocal cross described in table 5 TABLE 7 Hybrids between F. heteroclitys 9 * Cyprinodon & 2 hours About 20 per cent of eggs cleaving. Polyspermy evident in nearly half of these eggs if one is to judge by the irregularity of blasto- meres. The commonest rregularity consists of the production of two cells of very unequal size 5 hours Nearly all that had cleaved are in advanced cleavage stages and appear to be quite normal. Evidently early irregularities have had little effect on the cell arrangements of this period 12 hours Many eggs in advanced cleavage stages disintegrating. These are doubtless the polyspermic eggs. Others still normal in appear- , | ance 20 hours | All embryos dead, some evidently have but recently ceased to develop. These had succeeded in completing cleavage but show no signs of having entered upon gastrulation 466 H. H. NEWMAN TABLE 8 Hybrids between Cyprinodon 2 X F. heteroclitus 2hours A large proportion of eggs cleaving and in 4- to 8-cell stages; no signs of polyspermy 5 hours A great variety of cleavage stages, from 8 to more than 32 cells, for the most part regular or nearly so 20 hours | A few embryos have reached the end of the cleavage period and have started to form germ rings. The majority have begun to dis- integrate at the end of the cleavage period 2 days All those that were developing at 20 hou:s now dead, several hav- ing reached a condition in which the germ ring was about half around the yolk and a flat embryonic shield with no visible em- bryonic axis had been differentiated. This cross is more success- ful than the reciprocal cross, but less so than when the same eggs are fertilized by the sperm of F. diaphanus TABLE 9 : Hybrids between F. majalis 9 & Cyprinodon o& 3 hours About 25 per cent of eggs cleaving and in 2- and 4-cell stages, many irregular and some evidently polyspermic 6 hours Only a very few eggs alive, others show various degrees of cyto- | lysis 10 hours | All dead and disintegrating after reaching advanced cleavage stages _ Note: Another cross of these two species gave even less success than that here indicated TABLE 10 Hybrids between Cyprinodon 2 X F. majalis 3 2 hours | Nearly all eggs cleaving in an apparently perfectly normal manner. 4hours | A wide range of cleavage stages ranging from 8 to 32 cells. Those that are much retarded look unhealthy 20 hours | All but two embryos dead in advanced cleavage stages. The two that are alive are in early germ ring stages, showing flat embry- | onic shield. Both are much less advanced than control 2days | The two embryos that were alive at 20 hours are now dead without having made much progress since the last observation. This cross is more successful than the reciprocal, but less successful than either of the other Cyprinodon egg hybrids MODES OF INHERITANCE IN HYBRIDS 467 3. SUMMARY AND DISCUSSION OF DATA SHOWN IN TABLES a. Successful crosses Out of twelve crosses possible between the four species dealt with in this paper four have been successful in the sense that development up to hatching took place and more or less viable larvae were produced. The most successful crosses are those between FI’. diaphanus and F. heteroclitus, in spite of the fact that the former is a fresh water species and the latter a marine species. These two species cross reciprocally with equal facility, producing in each case some larvae that hatch earlier than those from pure bred eggs of the maternal species. These early larvae are also, in both reciprocal crosses, unusually vigorous and seem to excel the larvae of either parent species in viability and rapidity of growth. It is important to note here that F. heteroclitus devel- ops somewhat more rapidly than F. diaphanus, but that the developmental ‘rates of both reciprocal crosses are more rapid than those of the respective egg species, and that in the case of the F. diaphanus egg hybrids this is very marked. ‘These facts are entirely out of accord with the observations and state- ments of Moenkhaus, who claims that the invariable effect of the foreign sperm in fish hybrids is to retard development, never to accelerate it. He considers this universal retardation to be the result of some injury to either the egg or sperm substances of the zygote, akin to the well known hemolytic effects observed “in experiments dealing with the transfusion of foreign blood. The hemolysis parallel fails to apply, however, in these cases and in others where no retardation occurs but where a pronounced acceleration is evident to any one who takes the trouble to make carefully controlled studies of comparative rates of development of pure and hybrid strains. In these crosses in which early accel- eration has been observed it is conceivable that we are dealing with effects akin to those rejuvenating or stimulating effects so often noted when diverse strains of the same stock are crossed. Certain new combinations of morphological and physiological characters are more readily produced and occur more rapidly 468 H. H. NEWMAN, than those normal to either strain or species. I am inclined to interpret the speeding up of the developmental process in these hybrids as the result of the introduction by the sperm of a for- eign enzyme, which produces abnormally rapid dissociations in the egg materials, and in this way hastens the processes of metab- olism and development. Whatever be the chemical explanation of the acceleration the fact remains that at a very early period, certainly long before the end of cleavage, the hybrid eggs are developing more rapidly than the pure bred ones. This can be detected without the use of any refinement of method and should be obvious at a very much earlier period if the methods pre- viously used by the writer (Newman 710) were employed. If anyone has abundant time and patience he might readily demon- strate an acceleration in the first cleavages, but the writer has foresworn any further attempts of this character as involving labor incommensurate with the reward involved. Next in point of developmental success are the crosses between the eggs either of F. diaphanus or of F. heteroclitus and the sperm of F. majalis. Although the hybrids herewith treated develop successfully and produce a fair percentage of vigorous larvae they go much more slowly than the controls. It is also to be noted that the hybrids between F. heteroclitus 9 xX F. majalis ~ are much more viable than those between F. diaphanus 9 X F. majalis ¢. An interesting commentary upon the generalization of Moenk- haus that foreign sperm always has a deleterious influence upon the development of the egg, resulting in retardation of develop- ment, is to be had from a study of the chart including figures 9 to 20. Here the developmental rates of eggs of F. diaphanus, fertilized by three species of Fundulus sperm are shown in three parallel vertical columns, the middle column showing the control (pure F. diaphanus) that on the left F. diaphanus eggs fertilized with F. majalis sperm, and that on the right eggs of F. diaphanus fertilized with F. heteroclitus sperm. It will be readily noted that the effects of the two types of foreign sperm are of an exactly opposite character, the one producing marked retardation and the other equally marked acceleration. No theory depending MODES OF INHERITANCE IN HYBRIDS 469 on the idea that the effects of foreign sperm are always injuri- ous, resulting in retardation, can explain both of these equally obvious results. Another point of rather general import is also illustrated by this chart. It often happens in cases where hybrid strains go faster or slower for the first few days of development, that, in so far as the general external evidences of degrees of development may be relied on as criteria, hybrid and pure strains reach a point at which they are for a short time on a par. ‘This is well illustrated by the figures showing the three strains of F. diaphanus eggs at the end of one week (figs. 18, 19, 20). After this time acceleration or retardation is again evident, usually up to the period of hatching. I am unable at present to offer any suggestion as to the conditions underlying this apparent interruption in the developmental rhythms of these strains. In all cases in which hybrids successfully weather the period of gastrulation and enter upon the period of embryo formation, further success seems to be conditioned by the ability or in- ability to establish nutritive relations with the yolk. In cases where hybrids differentiate a circulatory system but fail to com- plete the assimilation of yolk, the difficulty is evidently due to lack of balance between the rate of embryonic differentiation and that of yolk digestion, as will be shown later. Although in the hybrids under discussion (F. diaphanus ¢ X F. majalis ¢ and F. diaphanus ¢ X F. majalis ¢) a great many of the em- bryos become abnormal for these reasons, it must not be for- gotten that in every experiment a large proportion of them sur- mount the difficulties of yolk assimilation and produce normal vigorous larvae, some of which, at least in the F. heteroclitus egg hybrids, are more viable and grow faster than either of the pure bred strains. It is difficult to imagine what factors under- lie this wide range of success of individual hybrids of the same parentage. It has been suggested that the best results are ob- tained when both germinal products are at the optimum state of maturity, and that if either one or the other germ cell be at all under-ripe or stale the result is sub-optimal development of varying degrees depending on the degree of over- or under-ripe- ness of egg or sperm. A more nearly probable explanation, how- 470 H. H. NEWMAN ever, is based on an almost diametrically opposite assumption and is suggested by a series of experiments which I performed with the idea of testing the effects upon development of freshness and staleness of the germinal products. Similar results were obtained several times, but sometimes what seemed to be the same treatment gave different results. Eggs of F. heteroclitus were fertilized with the sperm of F. majalis which had been kept alive for about twenty minutes in 45 per cent sea-water, a con- centration that greatly prolongs the life of these spermatozoa. A very small percentage of eggs cleaved, but these-developed normally and showed scarcely any trace of the paternal influ- ence either in rate of development or in details of inheritance; while the control strain, from the same batch of eggs fertilized at the same time by more sperm obtained fresh from the same male, showed the usual retardation in development and the typical hybrid characters described here and subsequently. It might then be concluded that, when the sperm is so stale that it is barely able to initiate development in the egg, it plays a role equivalent to that of the reagents that produce artificial parthenogenesis, but is unable to take part in the differentiation of hereditary characters. According to this view, the most active sperms might have the most deleterious effects upon the egg mate- rials of another species and give rise to serious incompatibilities whose result is more or less pronounced abnormality, and cessation of development; while the sperms that have lost some of their vigor are less likely to disturb the developmental rhythm of the egg+and thus more likely to give normal embryos. b. Unsuccessful crosses It will have been noted that, although the reciprocal crosses produce viable larvae, the hybrids between F. majalis ¢ and F. diaphanus ~ and those between F. majalis ¢ x F. heteroclitus # never produce larvae, since they never truly hatch. They may live for weeks within the egg membrane but gradually suc- cumb to anemia. The difficulty is almost certainly one involv- ing retarded or partial yolk assimilation, since in the final stages MODES OF INHERITANCE IN HYBRIDS 471 of development there is a large external yolk sac full of undigested yolk. In fish the development of the pericardium, heart and vitelline circulation are intimately associated, and under normal conditions the yolk sac diminishes in size at such a rate that it, together with the pericardium and heart, are simultaneously drawn into the body cavity of the embryo shortly after hatch- ing. In these abnormal hybrids, however, the differentiation of the heart and pericardium is at first accelerated, while yolk di- gestion goes more slowly than in pure bred embryos. Since the three structures are intimately associated, this failure of the yolk sac to diminish causes the pericardium to enlarge and stretches the heart into a long straight tube which beats feebly but carries, in later stages at least, no blood. Thus the supply of nutriment is cut off and growth ceases. Various organs continue to differ- entiate even without an external food supply, so that we may have an embryo developed which, though much smaller than the normal, has reached a stage of advancement equivalent to that seen in young larvae of pure bred strains. In last analysis the stoppage of development seems to be conditioned by a lack of coordination between two processes, that of the differentiation of the protoplasmic materials, which is accelerated by certain agents brought by the foreign sperm, and that of yolk assimila- tion which fails to progress as rapidly as in normal eggs. ‘The result is that the embryo gets to the point when heart pericar- dium and yolk sac should be taken into the body cavity and is prevented from so doing by the large mass of undigested yolk. The yolk of F. majalis is optically denser and of different color from that of the other species and it may well be that a specific enzyme carried by the sperm of the same species is necessary for its complete dissociation and assimilation. When I say that these abnormal embryos, burdened as they are with a permanent yolk sac, never hatch, I do not mean that they may not be shaken or dissected out of their membranes. When Bancroft claims that these embryos occasionally hatch I judge that he means that they may sometimes lose the egg mem- brane. Real hatching, however, is brought about by well marked violent struggling of the larva, and no such hatching struggles 472 H. H. NEWMAN have been observed in any of the numerous experiments dealing with F. majalis egg hybrids that have been under my observa- tion during the last eight years. Moenkhaus, moreover, agrees with me that these hybrids never hatch. c. Inter-generic crosses It is possible to produce six crosses between Cyprinodon and . the three species of Fundulus, as shown in tables 5 to 10. In no case is the cross successful in the sense that a larva or even an advanced embryo is produced. None of the hybrids go far enough to show specific characters, but all go through the cleay- age period more or less normally and some develop through the period of gastrulation and begin to show embryonic differentiation. The six crosses may be significantly arranged in the order of their success in development, as follows: 1. Cyprinodon ¢ X F. diaphanus 7. A small percentage of the embryos go practically through the period of gastrulation and begin to show the rudiments of embryonic differentiation. 2. Cyprinodon ¢ xX F. heteroclitus 7. A large proportion of the eggs go through to the end of the period of cleavage in a nearly normal manner and a few advance to a stage in which the germ ring has covered about one-third of the yolk and a flat embryonic shield is developed. Here the stoppage occurs in the midst of the process of gastrulation and before the embryonic axis is established. 3. Cyprinodon ¢ X F. majalis 7. A few embryos go through the cleavage period and show the beginnings and early steps in gastrulation, a well defined germ ring with flat embryonic shield being formed, as in figure 11. 4. F. diaphanus ¢ X Cyprinodon ¢~. Many embryos go through the period of cleavage and a few of them forma germ ring but get no further, the most advanced condition being like that shown in figure 9. 5. F. heteroclitus ¢ rie - ; 4 5 yA AL, rz 67 2 3 gd Be = ss e = ie = o: = heey ' \ es x ; atts vee ne & 11} ot fea (sy) PLATE 1 > > o fo} PLATE 2 EXPLANATION OF FIGURES '; anaphase, first cleavage. '; anaphase, first cleavage. ol4 PLATE 2 BEHAVIOR OF CHROMATIN IN HYBRIDS MARGARET MORRIS 515 ———— aa—<— -— ~ 14 15 16 17 1s 19 EP ya 40 40 +0 +O +40 +40 KOS OS OS GOK PLATE 3 EXPLANATION OF C 1; anaphase, first cleavage I 9; anaphase, first cleavage C o; anaphase, first cleavage C o; anaphase, first cleavage. F o; anaphase, first cleavage. C 1; anaphase, first cleavage. 516 FIGURES (X 2500). (X 2500). (X 2500). BEHAVIOR OF CHROMATIN IN HYBRIDS MARGARET MORRIS \ | } M Dw ww bw Ww aor WN HF © bt bo | ye fff 40 40 40 40 40 40 40 40 40 xXxxKXXKXKXXKXXKXKX bo ie) PLATE 4 EXPLANATION OF FIGURES F &; telophase, first cleavage. C &; telophase, first cleavage. C o’; nucleus, 2-cell stage. F &; nucleus, 2-cell stage. C &; prophase, second cleavage. C &; anaphase, second cleavage. C &; prophase, fourth cleavage. C &#; anaphase, fourth cleavage. C &; anaphase, sixth cleavage. 518 BEHAVIOR OF CHROMATIN IN HYBRIDS PLATE 4 MARGARET MORRIS PLATE 5 EXPLANATION OF FIGURES 29, 30, 31 and 32. F 9 & C o@; normally dividing cells from 12-hour stages. 33 FQ & C @&; large cells with irregular nuclei from 12-hour stage. ) qouqur Mog | OF-OF ‘D9 fyZ Wolf 9OUdIOIp popiooq | [Ardvo % GIO'O + SNOM ut Ge'o (2) pozAjoyso | 4801 oY} ‘oulnjsvjq wI0oy % oz wo | =. Joyooys '(% ¢ *p9) youjur uorsodoid yews y | (% 06-08 *y9) {y0RB,UI UTeUIET SBF0 qSO | [Aure-1 % “JOA #0 + SNOW ure ce’ (9) | Joyoo|s OB[NysETG % 0% “V9! 4OVIUT MOF BS POY] | pozAjoyAd ysoa oy, “gov jut % Op-ge “yp JAynq-u % "Joa T + GNOM uw 4gg'9 (Gc) oB[NYSB[ UII0; % YZ JOyoolR ‘Do SpazkjoyAo sour {yoRqut s#¥0 May pozAjoyAd ysor oy} {yorqur % ez-0z °9 jAdoad-u % ‘sjoa zZ + SNOM Ww cco (F) ‘s onpnysvyq | | JOIN UW CeO UNIO} 9UOU {49BJUT ULBUIOA 910UT 10 % 26 powyur sddo []e AT[BoOOVlT “SOA Of + GNOM Ul GeO “soa OOT (£) UMOP Uoyorq OOS SyoRQUT “TB Ur CE'D) sddo0 ysour {(% ¢-z) ov[nyseiq Moy y | qgoujur sddo [je fosuvyo ou AT[worpoerg | “sjOA OT + GNOY uw eGo “S[OA OOT (2) V SoLlug sv ourrg V Sollog sv ouleg SNOM Ut gg’ (1) { saisagy penuryu0h)—T WIAV.L 600 RALPH 8S. LILLIE Nevertheless, the after-treatment with hypertonic sea-water initiates development in a certain small proportion of these eggs, some of which form blastulae, though most stop short of this stage and undergo cytolysis. It is clear from the altered character of the response to hypertonic sea-water that some persistent modification has been produced by the previous treat- ment with salt-solution. What the nature of this modification is can only be surmised at present. The most probable general hypothesis seems to be that the plasma-membrane has been al- tered in some definite way—possibly rendered more permeable or more susceptible to alterations of permeability under changed external conditions. A changed state of electrical polarization would presumably accompany such a modified condition. On this view the increased responsiveness to hypertonic sea-water is analogous to the increased responsiveness of frogs’ voluntary muscle which has been sensitized by brief exposure to isotonic solutions of (e.g.) sodium citrate or other sodium salt.!° This form of sensitization is almost undoubtedly dependent on a surface-alteration, since it is produced within a few seconds by salts which either do not penetrate the normal plasma-mem- brane or do so with extreme slowness.1! The degree of protection afforded by the above alcohols may be decidedly greater than that shown in the above series. The following experiments (table 2) are especially favorable in this respect. These experiments also bring out clearly the effective- ness of anesthetics in inhibiting the characteristic action of hypertonic sea-water. Ethyl alcohol and phenyl urethane were used in addition to the anesthetics of table 1. It will be seen from an examination of table 2 that the pro- tective action of the alcohols is well-marked, while that of the urethanes, especially phenyl urethane, is comparatively slight. Amyl aleohol is distinctly superior to the others. These results are typical, as an examination of table 5 will show. A small proportion of the protected eggs remain unaffected by the after- 10 Cf. J. Loeb, Am. Jour. Physiol., 1901, vol. 5, p. 362. ‘1 Cf. my paper in the Proceedings of the Society for Experimental Biology and Medicine, 1910, vol. 7, p. 170. 601 ANTAGONISM BETWEEN SALTS AND ANESTHETICS S339 40BjUI OU | ‘% Z-[ “po ‘evpnyseyq uso; M2} WY “pBep ole [je Apivan owyar % Z-[ po feBinyseyq ou ‘yueurdojoa -ap yNoyyIM pap [eV ATION (% €-Z) qovqut ureurod Moy B fpazAjoyAD s#30 ysSo]Y avin} -SB]q ON ‘pazAjo}Ad 4soq % OF v9 {4youqut ureuror Auvyy qoequt % Ty) ‘avynyseyq ou {peep are qe ATpwotoVlg avyny GL (% uMOp ueyoig 4Sa1 oy} fo1our 10 % 09 ‘pa {sno1sumnu sB[Nyse]g | “v9 UMOP UeyOIG 4sad 0g <) oB[Nyselq WII0J S599 Jo UOTZI0doAd osuieT uMOp Uaxo.1q a emt} {joejurl = ulBured 6G "Dds aBTNySBTq % OE-0Z "PD q0BYUI UIBUT -31 % OI-G ‘pQ ‘pezAjoyA0 sBHo soy (% OI-G v9) aR[nysv[q Moy AJIATZBIBdUIO;: uMOp UayoIq qsor oy} £(% | ‘v9) qovqutr MO} Bf aB[NySB]q WIOJ % YE “PD UMOP Udyog ySod oy} £40R7 UI UIvUIOI 9, C ySol oy} 'ssd0 youqur % 7-T ul | D9) JOBJUL MO} B fpvoap ySOP | ‘v9 fovTNysvTq uIOF % OF | JOBYUI UIBUIOI 97, C-F ‘DQ ‘aB[nysey[q uo; (% [ *v2) Mo} ‘yuesudo[eAep jo susts qNOYJIM pvop a1B SBS9 yso] | OY} SoB[Nysvlq uTIOJ % OG “YOY UMOP UIYOId 91 4YSo1 UMOP UdYO1g 91v YS OY} /OB[NASB[G UO; 01994 }80UB SUTUIBIWOO (gq) d}JOYISOUB UIOIT aary (y) UALYM-VHS OINOLUGdAH HIM aaivauL s))a dO NOILIGNOO i % 09-06 “PO | pvop o1e ssdo [][@ AT[BOTQOVIg pezAjoyfo SUOT}OR sanjoniorde a4VT 8359 yORqut % OT “va Qysts ATOATPBIeL 0190301 | pozAjojAd 4Sor 043 yoeyur uteulel 9% Q9-0G 0 | pozAjojyAo 4ysor oyg £308 UT (°% 06-08 °29) Aqtsofeur os1eT | pozApoyA0 ysaa oy) £ (% 02-09 ‘D)) JOBYUL UTBUIOI SSGO YSO]\ ie ae qsot oy} fjyouqul % Og-Gz “POD pazAjoyAd 4Sor oY} :joeqjurl uleumlel 94 Q&-0% “PO Ricans pu ouvyyjeim [Aueyd % 90°0 + SNOM Ww Eg"0 (8) oueyyain [Aqye Yo 1 + SNOM WW g¢°0 (2) Joyooys [Ardea % c10'0 + SNOM W ¢¢°0 (9) JoyooTs [Aure 2%, 194 F'0 + SNOM UES" (¢) Joyooye JAyng % | TOA T + SNOM W G¢'0 (F) SoA F + SNOM W joyoo,s [Adoud 9% | S]OA4 @ + SNOM U g9°0 (€) Joyooye [Ayyo % e¢'0 (2) SNOM @ ¢¢°0 (7) QNO1V SNOILOATIOS HIIM GALVaAUL S9P)X FO NOILIGNOD SNOTLOIOS “9]qQD} AY} Ud waarb $v (sunoy 81 jmoqn layfp) B burnopof hop ay} UO ‘sbbo ay) {0 WO1IPUOI YT, alan sbhba ay) Laypm-pas ovu0jsadhy ur sajnunu og afy “LIJDN-DIS JDWLLOU OF PIULNIAL ‘WoYnNjos-7)08 Burpuodsaslod ay} UL SD WOI)DL]WAIUWOD BUDS AY) U2 O1AYISIUD BUDS AY) BUuIuIDZUOI LayDM-Das DIUO,LAd fy UL (gq) J4od sayjoUD ‘(JOD N Ut Gg SawUNjOAa QT snjd waJDM-vas sauNjoa QOT) Lajvn-bDas avuojadhiy ur paonjd avan sbhba ay) fo juod sayy) saynuru gy noqy "10729 Waar WoL WAjUaoUOd ay} ULIYOYISAUD ay) BuruwjwWoI YF NOM WU gg 0—:uoynjos-jos Burpuodsassoa sv 07 fijyoa41p *SUOLINIOS=]]DS 9A2Q09d SAL BY] UL SD SUO1]DL]UIIUOI DUDS AY) UL 892994} aUL paisafsuvi, spn shba fo 10) yona -saup Burmopjof ay) bururjU09 saynm-vas 0] saNunU gg jnogn sof pasodxa a.an sbBa nrangQsy pozy.4afuy JAJDN-DAS SVY] ULOL . 6 WIdVL "SIINUIW F SDM UW01]NIOS-7]D8 ay} 0} aLnsodxa "SI6l ‘91 nbn 602 RALPH S. LILLIE treatment with hypertonic sea-water and remain intact the next morning; but the majority either undergo cytolysis or develop to a blastula stage. The proportion of eggs forming blastulae is in general smaller the greater the protective action, indicating that when the membrane-forming action of the salt is prevented the eggs respond less readily to the hypertonic sea- water; some entirely fail to respond. It will be noted also that the presence of the anesthetic in the hypertonic sea-water greatly diminishes or entirely annuls the effectiveness of the after-treatment. Such hypertonic sea- water is not, however, entirely indifferent in its action, as is shown by the fact that relatively few eggs so treated remain intact; most undergo cytolysis and a few may develop. The same result appeared in six other series of experiments with hypertonic sea-water containing anesthetics. The favorable effect of the treatment is, however, almost entirely prevented by the presence of anesthetics in the protective or anesthetizing concentrations. Removal of oxygen from hypertonic sea-water or the addition of cyanide also prevents its characteristic action, as Loeb found for Strongylocentrotus.2 Cyanide has the same effect with Arbacia eggs (cf. table 3). Loeb has interpreted these facts as indicating that some chemical process involving oxidations underlies the favorable action of the hypertonic sea-water. That this agent acts by modifying chemical processes in the egg is also indicated by the high temperature-coefficient of the times of exposure.8 The fact that anesthetics have the same effect on the action of hypertonic sea-water as suppression of oxidations seems highly significant. Evidently the hypertonic sea-water induces some process of an oxidative nature; this process is checked or prevented by anesthetics. Now the anesthetics appear to act by altering the state of the lipoid components of the plasma-membrane, thus rendering this structure more resist- ant to change than normally: it thus appears probable that the 12 J. Loeb, Biochemische Zeitschrift, 1906, Bd. 1, p. 183. 13 J. Loeb, loc. cit.; also University of California Publications, Physiology, 1906, vol. 3, p. 39. ANTAGONISM BETWEEN SALTS AND ANESTHETICS 603 hypertonic sea-water exerts its characteristic action primarily by changing the state of the plasma-membrane, and that the oxidative processes underlying the favorable effect of this after- treatment are a function of certain membrane-processes. We have thus further though somewhat indirect evidence that the plasma-membrane is a controlling factor in the intracellular oxidations." ; Chloral hydrate has little effect in preventing the action of 0.55 m KCNS. In five experiments with this anesthetic, in concentrations of 0.2 per cent to 0.1 per cent, the highest pro- portion of eggs remaining unaltered next day was ca. 10 per cent (table 5). Potassium cyanide showed no signs of protective action in any experiment. The series shown in table 3 illustrates the results obtained with these compounds. It will be noted that although chloral hydrate and cyanide are almost without. influence on the cleavage-initiating action of 0.55 m KCNS, both prevent entirely the favorable effects of after- treatment with hypertonic sea-water. This action was highly striking in the above series, since in every experiment the great majority of eggs formed blastulae after treatment with the pure hypertonic sea-water. Experiments with sodium vodide In experiments with this salt all the above alcohols showed - well-marked protective action. The series summarized in table 4 will illustrate. The eggs were exposed to the freshly prepared 0.55 m Nal for four minutes; this is too brief an exposure to cause membrane-formation in all eggs, and about 20 per cent remained intact next morning; the proportion remaining intact was, how- ever, much greater in the anesthetized lot, and fewer of these eggs formed blastulae. A second similar series with five minutes’ exposure to 0.55 m Nal, and containing in ‘addition to the above alcohols ethyl urethane and chloral hydrate, gave similar results, although the protective effect was on the whole less pronounced. Urethane 14 For more direct evidence of a relation of membranes to oxidations, cf. my recent paper in the Journal of Biological Chemistry, 1913, vol. 15, p. 237. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 4 RALPH 8S. LILLIE 604: q08jur oB[NJSBIG WI; MOF B fOV[Nysv[q UOJ BUON | % O8-GL “V9 aAOGe SB VUIBG yorqur (4GL<) | MO} B fav[Nysvlq Woy oUON qovyur Moy Bf pez]OzA9 [[B Ajrvou [oVv[Nysv[q WO} VUON | | aelnyse[q waoy Aywoleyy pezAjoyAo {eB ATS | avjnysviq w10y 2% Og-OL “2, -orpovad fovTNysv]q WAOJ UON | oVTNySBIq UIO, % ER-CL “V) aB[Nysv[q WIoF % GL VO oy}oy4seUB Surarezuoo (q) opjouqsous wory aol} (V) MALVM-WAS OINOLUAGAH HOIM GELVAUL SONA JO NOILIGNOO pozA, -0jf9 s3S0 je Alyworjoeid oyeipAy (B10; yo ‘yooyo oatjyoozoid ou 10 97}9V'T | % 10°0 + SNOM “E90 (¢) pozAjoyAo [[B oyerpAy [Bs1o[yo Ajreou { yooyo 9aTz90,0I1d FYBITS | 2% STO + SNOM | ESO (F) s330 4o84UI | oyeipAY [BIo[yo Moy @ SpazhoyAo JIB ysouTy | % Z'0 + SNOM ™ G¢'0 (€) NOW pozhjoq4o s83o Te Aywoyovig | OOOT/U + SNOM "4 E¢"0 (Z) pozAjoyAd Sado [Be A[[VoTPOVL HLIM GALVaAUL SODA AO NOTLIGNOO GNOTV NOIWOTOS SNOM ¢¢°0 (T) SNOILOIOS "SaINULUL OG ‘La}DN-DAs 9YU0} -sad hy 07 fsanunu 4 spm suor1njos-]]0s 07 ainsodxgy *8W0717N]OS-]]DS 24} OF Jafsun.} asofag sanuvu og Lof apruvfia pun aynaphy yp.107yo Buruwmuoo sayom-nas w2 7fa) aiam sbba ay. °@ 2190) fo sarias ay] U2 sD aus ay? Spa ainpavoid aYf, “SI6T ‘9g ysnbny § ATAV.L 605 ANTAGONISM BETWEEN SALTS AND ANESTHETICS paezAjoyAo qsor ay} £(% §-Z “D9) 4ORy -UI Moy faB[Nyse[q WO; oUON pozAjoy4o Jopulvulal £4084UI UleUIeI 9% OI-G "po faB[nyse]q UIoJ OUON pozsyT -OJ40 4Sor 04} Syouqut (% ¢ >) Moj fovlnysv[q WIo0y oUON ael[nyseyq ou f40eqUI °4CT-OT !pezAjoyAd S380 YSOyL pazAjozA0 ysor 043 % T > ‘e[NysBiq Maj Bf% QG< ‘4084 -Ul uUlvmel uoljiodoid o51e'7 o}jeyysoue Buty} TOD (q) qoVqJUL UIBUIOL 9% 0% ‘Dd $(% ¢ po) aByNysByTq Maq | qOB4UI UIBUTOL 2% G ‘Dd '(% G *p9) oB[NysBlq MoT pozAjoy£0 4sar oy} fjyowqjur ureulol % 0Z ‘po faB[nyse[q UI0J % OL “2D qoeyur uorz10d -o1d {jews {pozAjoyho sade ysour {(% G ‘p9) Moy ov[Nyse[g | qovqut ureules % 0Z “9 5 (% ST-OL “99) T “dxq Ul UBY} JOMOy ov oB[NYSVI_ qowjul ureuled % CT-OT ‘po Saepnyse{q Wo % Og “PD | }OBJUT UTBUIO %/, O9-OG “D,) | qovqUI UIBUIOL % O8-OL “PO | + . qOB4UI UIBUIAL 9% QS “9 (% 04-09 DI) 4OBYUI UIBUIAI SS50 SOT (% OL?) ABP JXOU JOVJUL OIG SAO YSOP | 4084 | -UI UIvUel % (QZ “po !ABp opjayysout wOsy 901} (y) HALYM-VAS OINOLYAdAH HLIM GALVaUL SOOa AO NOILIGNOD aLan sbBa ay) LajpM-ngs Ur sajnurmu GT “nd waif @NOTVY SNOILO'TIOS HLIIM GaLVaudL §s99a AO NOILIGNOD | qxou pozdjoyAo ov S330 4SOy } “SOINUVU OF pauwuas fay) asayn ‘soYyay)saUD ynoypINn pUud YIN ‘LajpM-vas ‘saynuru FY SpM SUOLNIOS-]]D8 ay) 07 ainsodxa ay], joyooys [Ardvo <10'0 + I®N Bt GeO (9) joyoo,s [Aure-1 % 104 F'0 + IBN WGG'0 (¢) joyooye [Aynq 0A T + IBN W G¢'0 (F) joyooys [Adoad ‘JOA Z + IBN WCCO (8) Joyooye JAY} % JOA 7 + IBN W GeO (Z) I@N WGG'O (1) SNOILO'IOS oua)iodhy 0) postafsrn.) “SUOLYINIOS-])D8 AY} 0) 4afsuna) asofag sajnunu og sof soyayjsaun Gurinonjof ay) Burumjuod sajpm-vas 0} pasodxa auan sbba ayt, i> IT ysnsny | 9g ysnsny | pysnsny | € z | I Givd dNVY UWHWNON Deiat poe Saaner ouvyyoin TAuoYyd 2% 9070 | euceae iohcas tose Rol ouvyyoin [Ayo of CZ'T with 6 Weel, 0. als 6) 6lve) t/a lee ouvyjoin [Aye % Gar stata taliel a) elrae) CIRGES: 0 4. wrieheuen 6 9uo0ja1o[ yo % 70 Deca ot ier Aa joyooye [Aides % eT0'0 Fie ES NS ‘oyooye [Aure-1 % +A F'0 Vou Oi oon .cIrRe coc) cc Joyooye [Aynq-u 7 “AT Jerre eee eee " [OYoo]B [Aynq-u % "AST lod Se ht oy ood Joyooys [Adoud-u % “AZ Psaneieres peso joyooys [Adoad-u % °a Fz sat OO Gy TORE Oo . JOyoos pAqye % cy, t u |ishisiie) 'e)Venieie be teice) ©) (sh e)i5) ce ayearpAy B10; 49 % T'0 | OO Ce eu, CMCHORC tic chin t ayeipAy [eroyyo % cL‘0 | PUSMwls, eho ieksieiisvisila eile, 6 oyeipAy [eroyyo % Z'0 SNOM U g¢"0 NI NOILOIOS OLLAHLSHNY ro -pdas p fo s}nsas ay) saavb Uwmnjo0d yooYy0aa Yoo "$91U9S OIDL "(% & uy}? ssa) shonjp) fivp 7xau yonjur sbba hun fr maf 1f2) SNOM WU 6¢°0 DAL f-INJOYISOUD ALN AY] YIN JWAUWIHAL) BY) SASDI YD UT *(UOYN,OS AY? 07 aunsodaxa sazfD sunoy O@-ST “a°d) 7ybvw saao sayom -pas U2 Burana) aj fp JoDJU1 paurpuas yoy) sbba fo uwoysodoud pajwurysa ay) aah saunby ay J, alNnsodxa ay], “SaynUuu F SOM YNOM Ul Gg'0 0} "SIINUWUL OG NOD LOf (JUAWILIELA JDY] UL PASN UOYNIOS-])DS 9Y} UL SD WOYHDLJUIIUOI BUDS BY) UL IYaYISOUM AULDS BY}) DYaYyISaUD ay) BuLiuimzU0d saypM-nas YN ‘GNOM Ue ogo ur buropjd 07 fiysnorasd ‘pajnas, avan shba ay) sasvo yo UZ ¢ ATAV.L 608 RALPH S. LILLIE a given anesthetic may interfere with one set of processes and not with another has long been known. It would appear rather, as suggested above, that while chloral hydrate and the urethanes impart to the plasma-membranes a degree of resistance sufficient to prevent the normal changes of cleavage, this resistance is insufficient to prevent the relatively violent action of the pure salt-solution. The alcohols, for some reason as yet obscure, protect the egg more effectively against the latter action. They are, however, quite ineffective in preventing the cleavage-initiat- ing and cytolytic action of fatty acids. The ability of an anes- thetic to suppress or prevent a given process thus depends both on the nature of the anesthetic and of the process itself. The experiments with fatty acid about to be described will illustrate this. The cleavage-initiating action of fatty acids in the presence of anesthetics The above anesthetics entirely fail to interfere with the for- mation of fertilization-membranes or the initiation of cleavage by fatty acids. In a number of experiments, conducted similarly to those already described, in which the parthenogenetic agent was sea-water containing acetic or butyric acid (2 to 3 cc. fatty acid plus 50 cc. sea-water) the results were entirely negative. The addition of the anesthetic seems indeed to increase the injurious action of the acid. Chloral hydrate is equally ineffec- tive. The urethanes were not tried. The following series with the alcohols will illustrate (Table 6). From these experiments it is clear that the above anesthetics are completely unable to counteract the action of the fatty acid. Their antagonistic influence seems to be confined to the salts. Experiments with sea-water containing additional calcium and magnesium chloride (without altering the osmotic pressure) gave an analogous result. It was thought that possibly by increasing the proportion of these salts in the medium the eggs might be rendered more resistant to fatty acid. The following solutions were used: 100 volumes sea-water plus respectively 609 ANTAGONISM BETWEEN SALTS AND ANESTHETICS uees Ov[N4Se]q 0M} IO 9U0 {pazAjoyAd [[e ATIVAN pezdjoz49 [TV paz4jo449 TLV, pozAjoyA9 ITV %1 > :eBpny “SBIq Moy B SpazATo4Ad []e A[IVON oTJOYISoUB Bulurezyuod (_q) pezAjoyA0 Ysor O44 :% OG ‘vo favpnyseyq wo; Auvyy pozAjoyAo [ev fdojaaep ou0N pezs] -0j40 ][B fovpnyseyq wu0J sUON aB[nyseyq unos % G “pa fpazh[o}Ad S830 YSOPY % 01-09 “v9 |] todxg Ul uBy} Jomoj ovpnyselg aVB[NySB[q WI0F % 06-08 pozAJoyA0 [TV pezd]o449 ITV pez4]o449 [TV pozA[o}A0 [TV paezAjo4A0 ITV pezd]oq49 TV jou ~ooye [Aidvo % ¢TO'O + oureg (9) joyoo -[8 [Aure-1 % “a $0 + oureg (g) Joyooys [A4nq-u % *A T + oureg (fp) JOyoo|s jAdoid-u % ‘joa g + oureg (¢) loyooye [A190 % "Joa F + ouBG (¢) prow o1Agng *y% + aoqyeM-vag (1) O1oyjsouB ynoyyIM (Vy) UALVM-VUS OINOLUAdAH HLIM CALVAUL-YALIV SDD AO NOILIGNOO NOTY SNOILOTOS HLIM GGiVaGuNL S994 AO NOILIGNOO SNOILaTOS 9 ATAVL Ssmoj0f sp spm sinoy 6] 4a, fo sbBa ayy fo UOYIPUod ay T, “sanunu Og Lof ‘sayjoy;sauD ynoyyN puv YR ‘LaDM-Das o.U0,LedhYy 0) pasodxa auam avd saynuru Gy fo qpa ~L9qUL UD Lif @ *(pron arahing “E *00 & snjd saypm-nas *99 YG) prov druhyng 92% BuruwyUuo0d saym-nas YUN sajnunu g Hof paynany way) aLan fay], “samnuru gg sof 4avM-vas Ut sjoyooyp Burmonpof ay) fo suoynjos ay) 07 pasodxa auan sbBa "UL “SI6I ‘0g Isnbny 610 RALPH S. LILLIE 10, 20 and 50 ce. 0.35 m MgCl, similar mixtures of 0.35 m CaCl, and sea-water, and mixtures containing both salts; 2 cc. % butyric acid was added to 50 cc. of each solution, and the action on unfertilized eggs was tested as above. The results again were entirely negative as regards protective action: in all cases two minutes’ exposure to these solutions was followed by cytoly- sis. After-treatment with hypertonic sea-water proved ineffec- tive in all cases, none of the eggs so treated forming blastulae (with the exception of a few from 100 volumes sea-water plus 10 volumes 0.35 m MgCl.). The presence of an excess of calcium or magnesium in the sea-water thus prevents the eggs from devel- oping favorably later, although it does not hinder the cytolytic action of the fatty acid.'7 17 }xperiments, with unfertilized eggs conducted in the summer of 1911 on the antitoxic action of CaCl, on isotonic NaCl solutions containing acetic acid gave entirely negative results. The following series will illustrate. Eggs were left for two hours in the solutions, then returned to normal sea-water and fertilized. The results were as follows (condition of the eggs next day): Solution Result (t) Pure 0.55 m NaCl........ J GSE Se OE Te es cae vnennan ca. 40-50 % form blastulae (2) 95 vols. 0.85 m NaCl+5 vols. — CaCh.....-..0.ccceccccecees all form blastulae (3) 0.55 m NaCl+— 20 CHICOO He ce eee Cen Sic oesiont eden all eggs dead (4) 0.55 m NaCl+— 20 CH3;COOH+ | (CCL sak eee ne a ee all dead (5) 0.55 m NaCl+yy CHECOOH. he ee eee apne eee eines all dead (6) 0.55 m NaCl 79 CH:;COOH+ > CaCl... ccc veces all dead (7) 0.55 m NaCl+—— io CHsCOO HS yale cars Os sos is veneer eee all dead (8) 0.55 m NaCl+ G99 CH;:COOH+ | a OAC IS: 31.1. dc Saeeee eee all dead (9) 0.55 m NACH aon CHSC OO Hira sn cts: jee ae ee eee all dead (10) “0.55 m NaCl+— se CH;COOH+ CAG: 28 cde eer all dead Thus CaCl, in concentrations which completely prevent the toxic action of the NaCl solution has no effect on the toxic action of acetic acid. CaCl, also showed no antitoxic action in 0.55m NaCl containing NH,OH in cone “on #, to s%,. It also failed to anlage’ HCl in concentrations from ,4, to os ine with weaker solutions (,4,, to ;.%,), HC!) some antitoxic effect was seen. Arbacia eggs thus differ from Fundulus eggs, which ae protected by salts to a considerable degree against injury by acetic acid in ;¥, concentration (cf. J. Loeb, Biochemische Zeitschrift, 1912, vol. 47, p. 151). Fundulus eggs are, however, surrounded by a resistant chorionic membrane. ANTAGONISM BETWEEN SALTS AND ANESTHETICS 611 CONCLUSIONS We reach thus the general result that the formation of fertiliza- tion-membranes and the initiation of cleavage may be prevented by anesthetics when the parthenogenetic agent is a neutral salt, but not when it;is a fatty acid. This contrast is what would be expected on the assumption that the essential action of the anesthetic is superficial, and consists in rendering the plasma- membrane more resistant to alterations of permeability. Hence the salt, which does not readily penetrate the unaltered egg and produces its effect by increasing the permeability of the plasma- membrane, is rendered less effective when the membrane has been rendered relatively resistant or stabilized by the anesthetic. The fatty acid, on the other hand, which penetrates the plasma- membrane readily under all conditions, by virtue of its lipoid- solubility, is not prevented in its action by anesthetics. Fatty acids and neutral salts represent two classes of agents, one of which penetrates the egg-surface readily, the other with difficulty and apparently only after increasing the permeability of the plasma-membrane. Both induce parthenogenesis in a typical manner. It is significant that of these two parthenoge- netic agents one should be influenced in its action by anesthetics, the other not. The fact that it is the more penetrating of the two which is uninfluenced seems to indicate that the agent pro- duces its essential effect by acting on some portion of the egg- cytoplasm situated within the most external surface-layer, and that calcium and anesthetics inhibit the action of salt-solutions because they prevent the access of the salt to this critical region of the egg. Otherwise it is difficult to understand why the anes- thetic, which apparently stabilizes the surface-layer and hinders alteration of permeability, is without influence on the action of the fatty acid, although it inhibits the action of the salt. The en- trance of the salt but not of the lipoid-soluble fatty acid would be hindered by agents which act (like anesthetics and Ca salts) by preserving semi-permeability unaltered, since semi-permeability relates to lipoid-insoluble substances'’ only, as Overton first showed. 18J.e., non-colloidal substances. Ruhland has shown that various dyes which form colloidal solutions are exceptions to Overton’s general rule. Cf. Jahrb. wiss. Botanik, 1908, Bd. 46, p. 1. 612 RALPH S. LILLIE It is of course also possible that the actual entrance of the salt into the egg is unnecessary, and that a purely superficial action sufficient to increase permeability to a critical degree and thus cause a .definite depolarization-effect is all that is necessary. There is at present no certain means of deciding between these two alternatives. If the salt mcreases permeability to a suffi- cient degree, it will naturally enter the egg and produce certain effects in its interior. It is, however, clear that there is nothing specific about the salt-action; all that is needed is that it should be sufficiently energetic. The entrance of special substances from outside into the egg is not necessary for parthengenesis: the effects of temporary warming and mechanical agitation upon starfish eggs are a sufficient proof of this. On the other hand, certain substances as fatty acids and lipoid-soluble alkalis, do undoubtedly produce their effects by penetrating the egg.!® The comparative ineffectiveness of the lipoid-insoluble and non- penetrating alkalis and acids indicates this clearly. The most probable conclusion seems to be that the same effect can be produced by a purely superficial action, like that of a salt or the electric current, as by one operating at some region within the interior of the egg. There are many indications that the primary effect in the acti- vation of the egg—whether by the spermatozo6n or a partheno- genetic agent—is superficial and consists in an alteration of the surface-layer of protoplasm—the region somewhat vaguely designated as plasma-membrane. The term ‘plasma-membrane’ in its application to the surface-film seems at present to require more precise definition. The conception of this structure as a thin homogeneous haptogen membrane exercising passive me- chanical and osmotic functions is clearly inadequate. It must rather be regarded as essentially a superficial portion of the living protoplasm, characteristically modified in its composition and physical properties by surface forces. We must thus ascribe to it a characteristic chemical organization and metabolism as well as the characteristic physical and other properties, such as 1° Cf. Loeb’s recent paper on ‘“‘The comparative efficiency of weak and strong bases in artificial parthenogenesis,’ Jour. Exp. Zodl., 1912, vol. 18, p. 577. ANTAGONISM BETWEEN SALTS AND ANESTHETICS 613 selective semi-permeability, that have led to its distinction from the more internal protoplasm. Conceived in this way, it corre- sponds closely to what morphologists designate as the cortical region of the egg, or at least to the most external layer of this region. Changes in this region form a highly characteristic accompani- ment of fertilization in many if not in all eggs; associated with these changes is a marked temporary increase in the general permeability of the surface-layer. The relation of these surface- ‘changes or cortical processes to the initiation of cell division and development is evidently a critical one.2° Once they are accomplished the developmental mechanism, hitherto held in check, resumes operation and—provided external and other conditions are favorable—continues its course automatically to the adult stage. It is clear, from the diversity of the con- ditions that may initiate development, that some process specific . to the egg and quite independent of the nature of the activating condition forms the primary event in fertilization. The sper- matozo6n or the parthenogenetic agent in some way removes the hindrance to this process. What the nature of the latter is may be partly inferred from the results of recent experiments on the physiology of fertilization. The observations described in this paper support the view that some change in the cortical region of the egg-protoplasm, beneath the most external semi- permeable layer of plasma-membrane proper, forms the initial stage of the fertilization-process. The immediate surface of the egg has semi-permeable properties relatively to most water- soluble lipoid-insoluble substances, and apparently must under- go increase of permeability in order that such a salt as Nal or KCNS may produce its characteristic effect. As already said, it is uncertain whether the salt acts by entering and then affecting directly the state of the colloids in the cortical region, or whether it acts without entrance, possibly by altering the electrical polar- ization of the plasma-membrane. It seems probable, however, from the general effectiveness of lipoid-alterants, that some 20 Cf. F. R. Lillie, on The cortical changes in the egg of Nereis. Jour. Morph., 1911, vol. 22, p. 361. 614 : RALPH S. LILLIE change in the condition of the lipoids—possibly in the inter- relations between lipoids and proteins—is the primary effect produced and that this change then initiates some specific chemi- eal reaction which determines directly or indirectly the char- acteristic surface-changes of fertilization, namely, secretion of cortical material, formation of fertilization-membrane, tempor- ary change in osmotic properties of the plasma-membrane with accompanying electrical depolarization. This view emphasizes the analogy between the activation of the resting egg and the general process of stimulation.*! In - stimulation the primary event is a depolarization of the limiting membrane; similarly in the fertilization-process the electrical variation accompanying the above surface changes forms most probably the critical or ‘releasing’ event on which the rest of the process automatically follows. One gains the impression that in the resting egg-cell, as well as in the resting muscle or nerve, certain substances are hindered from interacting by the . electrical polarization at the cell-surface; just as in a battery with open circuit the chemical reactions on which its opera- tion depends are held in check by the polarization at the sur- face of the plates: that is, the passage of ions into or out of solution is thus prevented and with it all effects, chemical and other, dependent on the flow of electricity through the circuit. Under analogous conditions in the living cell a brief depolariza-’ tion might suffice to release the impediment to the chemical interaction forming the primary event in the response—whether to stimulation or (in the case of the ‘egg-cell) to fertilization. It’ is noteworthy that in many if not in most irritable cells the response is specific and constant and independent of the char- acter and intensity of the stimulus; that is, recent research indicates that the ‘‘all or none”’ law applies to irritable elements in general,2? and not only to heart-muscle, and it may be said 1 For a fuller discussion of this analogy, cf. my recent paper in the Journal of Exp. Zo6l., 1918, vol. 15, p. 23. 2 The solution-tension of the ions being compensated by the electrostatic at- traction between the plate and the oppositely charged adjoining layer of solution. °3 Cf. especially the recent articles from the Cambridge Physiological Labora- tory by Lucas and Adrian in the Journal of Physiology. ANTAGONISM BETWEEN SALTS AND ANESTHETICS 615 with some qualification to apply also to the resting egg-cell. The removal of an inhibiting condition, whatever the means employed, is the essential requirement for fertilization. What follows is determined entirely by the nature of the egg itself. The recent Work of F. R. Lillie reinforces still further this general point of view. His results indicate that in fertilization a union of a specific amboceptor-like substance, contained in the egg-cortex, with some other specific substance, also furnished by the egg, forms the primary or determinative event. The spermatozo6n acts by removing the hindrance to this interaction, but other agents may act similarly—hence the possibility of parthenogenetic fertilization. This conception makes it clear why the presence of the sperm is unnecessary to the activation of the egg, and suggests that in its activating capacity this struc- ture plays essentially the part of a specific releasing mechanism, in a manner which is thus closely analogous to that of a stimu- lus, as already indicated. The ensuing developmental processes are specific to the egg under investigation and require special analysis in each case. SUMMARY The chief experimental results and general conclusions of this paper may be briefly summarized as follows: 1. The action of pure isotonic solutions of neutral salts (0.55 m KCNS, Nal) in inducing formation of fertilization-membranes and cleavage in the unfertilized eggs of Arbacia may be pre- vented by anesthetics as well as by calcium and magnesium salts. The effective concentrations are those which just suffice to pre- vent cleavage in fertilized eggs. 2. The anesthetics are less effective than calcium or mag- nesium, and vary characteristically in effectiveness. The mono- hydric alcohols of the aliphatic series are the most favorable of those tried; the order of relative favorability runs: amyl >n- butyl > n-propyl and capryl > ethyl. Phenyl and ethyl ure- thanes have comparatively slight action, and chloral hydrate still less. Cyanide is ineffective. 24 Cf. Science, N. S., 1913, vol. 38, p. 524. 616 RALPH S. LILLIE 3. The anesthetics have no inhibiting influence on the cleav- age-initiating action of fatty acids. Since fatty acids readily penetrate the unaltered plasma-membrane, while salts do not, and since both agents are equally effective in inducing parthe- nogenesis, this difference indicates that the parthenogenetic agent acts at some point within the most external layer of the egg. 4. The favorable effects of after-treatment with hypertonic sea-water are prevented by anaesthetics as well as by cyanide. This result indicates that hypertonic sea-water, as well as anes- thetics, acts by modifying the condition of the plasma-membrane. CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. NO. 248. « OLFACTORY REACTIONS IN AMPHIBIANS JONATHAN RISSER ONE FIGURE CONTENTS ATR IAC LUT UT OTIAE Wt: ou py aM ow cod 2 ack) cyte fs SS SEMPRE St state hc God aialalan slots 618 “Pinpcl qinval CyevenRce eed 5.35 oo o.ce t SeRe See eee ere chai cla 6 aan ce en 618 lemNelatton Of OG OTS CONGO so... 5.5 sos cinta Be watts ee od ais Sirius 8 ssa wince 618 Pea Merbental sean dermenh OG Spr i sx.) s.5) avasesscc aie se ayo re ake ore hie os oeennece 619 ESTEE ISIS 8 2 oe co UR Mee ica nei ac Gris acre ei Iie ic 620 Peraperiments swith) DOU-Mving LOO... 52.5. :05,i2ise ges leeds tee ie ple tie oe 624 cog 7 Gens Ciera Saray TR 80) 26 Se iA RO ng Ac 625 Fie IMREHIETE ES as Bias etawals Seconda ae Oe Are Rea eI 5 ie Se A 625 & [SSO RITES tes 6 6 low 6 bia 5 RoIR OD EERIE oni HES SOE SC Se OLEAR Earn c 626 peetimardial onldrayandsdarkness)... <2. 1.< see i Muchacsedl cles deere 627 RR ass MPITAIME TGS oh ecco o's ashes sae os nace nis on td ROR EA meats 628 ile, JAOPOR METAR 5 eb a Gite i ea oe nee ee ea eer eed acre he 628 PMmRATeIOLNOWNOL OUONSStROAM. < .2ir wc .uyl os Maced: aie obo ae eee tonne 630 3, IMIGGINGOID, 56 deen be os cee Oo aR See ene ee tee i Sener cE eR chee tee 630 SUD SbANCESP USEC MIMy GHETLEStS sce. o5- + «salsa Sree Sa es 645 rmtan eRe orm hoy is MN ONS ie he AA ey SAE a Aca elas Mates s EE 647 2 STIG TST TL pala RRB en ae ST 648 Sn er ade ti gl Me lg i bi SRR ANS ot ng RE mR nee 649 1 Sere SS eee Eo a Cea ea a a Ir 651 618 JONATHAN RISSER INTRODUCTION It is the purpose of this paper to record some observations and experiments on the olfactory reactions in certain amphib- ians, more especially in the toad (Bufo americanus LeConte). Initial experiments were tried on frogs (Rana virescens and R. catesbiana), both larval and adult, but the work did not prove promising and was therefore given up. These investigations were undertaken at the suggestion of Dr. G. H. Parker of the Zodlogical Laboratory of Harvard College and to him for his kindly interest and helpful advice I wish to express my deepest appreciation. FOOD AND ODORS 1. Relation of odors to food There is little evidence that the relation of odors to food has been taken into account in any investigation of the habits of the different amphibians. In the quantitative studies of the stomach contents of frogs and toads no evidence has been found to show that certain foods are preferred by these animals. Fischer-Sigwart (’97) believed that frogs and toads were indis- — criminate feeders. Needham (’05) has shown the food of the bullfrog (R. catesbiana) to be extremely varied. Lockwood (’83) in speaking of the toad, says ‘I do not believe it can smell. It catches insects, but only when such probable food is in motion.” Knauer (’75), however, mentions cases where decomposing animal food was rejected by toads after having been taken into the mouth. Schaeffer (’11) found that certain caterpillars were refused by frogs in a similar manner. In both cases other factors may have been of disturbing influence. Hartman (06) found no special preponderance of one species over another in the insects taken from the stomach of toads collected at ran- dom. Garman (’92) has records of the food of toads from which one might conclude that ants were more sought after than other insects. Such a condition, however, may have been due to the fact that younger newly metamorphosed toads being close to the ground met ants more frequently than they did OLFACTORY REACTIONS IN AMPHIBIANS 619 other insects. Hodge (’98) refers to the large numbers of house- flies eaten by the toad. Schaeffer (’11) mentions the variety of insect food taken by frogs in confinement. Quaintance and Brues (’05) show how toads make use of very diverse forms of insects as food. * Slonaker (’00) and others have fed meat to toads, by simulating the motion of small insects to attract attention. The behavior of toads and frogs in confinement leads to the inference that food ordinarily must be in a living condition and in motion to be attractive. Among food materials taken by the toad under normal cir- cumstances there are many insects with characteristic odors. Conradi (’01) infers that the odor of the cucumber beetle (Anasa tristis) is inimical to the toad, but Hill (73) observed no disastrous results to toads that had fed on this insect. Neither frogs nor toads hesitate to make use of other vertebrates as food, when occasion offers, or to devour members of their own species. Can it be shown that these animals are stimulated to seek for food or refuse it because of odors? With this question “in mind the following feeding experiments were carried on with the toad (Bufo americanus LeConte). 2. Materials and methods The toads used in these experiments were obtained in the vicinity of Cambridge, Massachusetts, quite late in the year. They were kept in a large box containing soil and leafmold. in a moderately cool basement room. The soil was kept damp and the box dark. Some of the animals buried themselves in the soil, others took shelter under bits of wood. The animals reacted normally in all respects, taking food when offered. The food consisted mainly of mealworm larvae (Tenebrio molitor), of earthworms and dungworms (Allolobophora foe- tida), flies and other insects. Some of the toads were removed to large jars for greater convenience. Not all the azxmals were alike, some showing a greater tendency to hibernate than others. Food was generally given the toads at intervals of several days, but this procedure was modified as necessary. For convenience the experiments were carried on with the THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, NO. 4 e 620 JONATHAN RISSER toads confined in a large metal pan. In this way the animals were kept within bounds but were still allowed much freedom of movement. 3. Hxpervments The first experiment was made to determine whether choice was made between two forms of food with specific odors. Pre- vious to the trials earthworms had been fed freely to the toads, and it might be supposed that because of the peculiar odor of the dungworm, the latter would be refused as food. Toads No. 1 and No. 2 were in the box in which the experi- ments were conducted before the dungworms were introduced. Toad No. 1 saw the worms soon after they had begun to crawl. It quickly hopped toward the worms and attempted to snap up the nearest worm. The attempt was unsuccessful and the lips were wiped with the forefeet. Sandgrains on the lips evidently produced a mechanical stimulus which was followed by the wiping. The dungworms were then freed of all the adhering sand and again placed in the center of the box. Although the move- ment of the worms was sufficient to attract the toads, no further attempt was made to take the worms. When a mealworm larva was put in with the dungworms, it was quickly snapped up by the nearer toad, No. 1. Dividing a dungworm then into a number of pieces, these were placed in a shallow vessel before the toads. Into the vessel there were also dropped two mealworm larvae. From the mass of wriggling pieces one of the mealworms was immediately »elected and swallowed. Upon the dorsal surface of the other mealworm a drop of oil of pennyroyal was now placed; the meal- worm with the oil was also quickly taken. Toad No. 2 made several unsuccessful attempts to take the remaining dungworm, but «bandoned the dungworm, taking later some mealworms that had been put in the box. Similar trials with two other toads gave like results, leading to the conclusion that the dung- worm is not used as food under normal condition, a conclusion that was subsequently shown to be false. OLFACTORY REACTIONS IN AMPHIBIANS 621 Some days later, the same toads, Nos. 1 and 2, were again tested with the dungworms. Three worms of medium size placed in the center of the box were immediately noticed by Toad No. 1. _ No attempts to take any worms followed. The odor character- istic of the worfns was evident and the cutaneous exudate was visible. .A mealworm dropped into the vessel among the dung- worms quickly became covered with the slime of the worms. That the mealworm was recognized as a new object by the toad, is proven by the actions of the toad. The attitude peculiar to the animals when watching some object, was immediately as- sumed. As soon as the mealworm had crept away from the dungworms, the toad snapped it up. Another mealworm dropped into a dish containing some fragments of dungworm was watched by the toad as long as it moved. After it ceased moving it was of no‘more interest to the toad. No attention was paid to the fragments of dungworm although they were wriggling. A second mealworm smeared with the exudate from the earthworm was quickly snapped up by No. 1. This reaction was followed by the wiping of the lips with the forefeet; whether this was done to remove the slime adhering to the lips was not determined. The wiping action again took place when the same toad took another mealworm similarly prepared. Toad No. 2 exhibited the same reactions under similar conditions. A series of trials with the different toads proved that frag- mented dungworms were never taken, even though they ex- hibited decided movements. The trials also showed that the toads preferred the mealworms, and that the odor of the dung- worms when applied to the mealworms was not deterrent to the toads. . Trials carried out with the same toads, to determine their preference for mealworms over earthworms, indicated that the mealworms were the more desired form of food. These were selected from among the earthworms. It is not clear why this should have been so; previously and later in the course of the experiments both earthworms and mealworms were taken promiscuously. 622 JONATHAN RISSER In a subsequent trial with Toad No. 1 mealworms placed among dungworms were taken after they had moved away from the dungworms. Broken pieces of dungworm taken into the mouth were followed by the wiping action. The exudate alone did not affect the toad in a similar manner. Although the odor was evident on mealworms smeared with the exudate, this did not hinder them from being taken, the only effect being a pecu- liar ‘gaping’ action after the mealworm had been swallowed. To determine whether the color of the mealworm was a deciding factor, attempts to approximate the color of earth- worms and dungworms were made. Mealworms tinted to re- semble the color of the earthworms were taken without hesitancy. - Such mealworms were always taken from among fragments of dungworms and earthworms. In some trials for determining the effect of distinctly ab- normal odors and natural foods, the following observations were recorded. The toads experimented upon were No. 3, which was very responsive, and No. 4, which was sluggish. Dungworms were put in the experimental cage in a Petri dish. They attracted the attention of Toad No. 3, but otherwise it did not respond. The dungworms were then cut up and again placed in the cage, but again the toads did not react. Mealworms were now sub- stituted for the dungworms and No. 3 immediately took two of these. Oil of pennyroyal was put on certain mealworms and when these were introduced, two were taken by Toad No. 3. No discomfort was shown by the toad. Fragments of dungworms were placed in the cage, but were not taken by the toads. Mealworms covered with dungworm juice were quickly taken, and accidentally one piece of a dungworm was taken and swallowed; another piece was brushed from the mouth. The mouth was opened several times in succession as though the fragment of worm had been unpalatable. Four days later the same toads took mealworms with the juice of dungworms, and the exudate smeared upon them. Both toads took mealworms smeared with oil of pennyroyal unhesi- tatingly. Again later, the same toad, No. 3, did not discrimi- OLFACTORY REACTIONS IN AMPHIBIANS 623 nate between earthworms and dungworms. The latter were irritated in such a manner as to make the integumentary exudate distinctly noticeable. This was not«deterrent in any case. Whatever may have been the cause of the refusal of the dungworms in the earlier trials has not been made clear, for later trials showed no aversion to the dungworms on the part of the toads. A series of trials with the toads, in which the conditions of feeding were carried to a degree far beyond what might be expected to occur at any time normally, showed that many odors were not repellent when associated with food. To hungry toads was given the choice of taking mealworms with unknown odorous substances, and others without such substances. Ether, chloroform and alcohol could not be made use of, because of their fatal action on the mealworms. Clove oil, oil of cedar, oil of pennyroyal, oil of bergamot, oil of cit- ronelle, aniline oil, carbon bisulphide and iodine in saturated solutions were made use of. In the experiments there often arose the question of a fatal dose. But none of the toads died during the trials. Several of the trial records are transcribed: Toad No. 1. One mealworm eaten, after which two meal- worms with oil of pennyroyal taken within one minute; only effect was ‘gaping’ several times. Rest for ten minutes, then mealworms put in a dish too high for toad to get into. Toad ‘in attentive attitude, and again when offered took clean worms and also one with oil of cloves. Odor was not repellent, nor was any after-effect noticed. Four days later the same toad ate three mealworms, two with clove oil and one with oil of citronelle. Toad No. 3. Took in two successive trials, two clean meal- worms, two with oil of pennyroyal and two with carbon bi- sulphide; and later two clean mealworms, two with aniline oil, and two with oil of rose geranium. Toad No. 5. At one trial, took two mealworms with oil of pennyroyal and one earthworm with the same oil applied. The oil of pennyroyal appeared to be more irritating to the lips of the toads than the other substances used; the wiping 624 JONATHAN RISSER of the lips was the only reaction seen after the toad swallowed worms treated with this oil. Iodine, when taken into» the mouth with the larvae seemed to be more productive of discomfort than the oils did. The larvae were not refused at any time during the trials, nor were the substances deterrent because of the attendant odors. 4. Experiments with non-living food Some experiments were made to determine whether animal food that gave no evidence of being alive would be taken by the toad. In the feeding trials previous, only mealworms in motion were used for food; no attempts to use inert worms were made. Pupae of the mealworms were offered the toads. When the pupae were first dropped into the cage, they often made spasmodic motions, but these soon ceased. Sometimes the toads took the pupae when thus in motion. Mealworm larvae, freshly killed and motionless, were never taken. By means of a delicate strand of silk threaded into a larvae, these could be dragged over the bottom of the cage in a manner to imitate partially the living condition. The toads could be induced to snap up such larvae, and when once in the mouth they were not rejected. Knauer (’75) gives instances of the rejection of decomposing earthworm. My experience was to the contrary. Bits of meat fashioned into semblance of meal- - worms when thus drawn past the toad, and taken into the mouth, were not rejected. Dead flies also could be suspended and moved before the toads. These were attractive to the toad even if abnormally ‘odorous. Artificial larvae were fed in a few instances to the toads. These were made from absorbent cotton and paraffine, and could be substituted for the living worms if set in motion by means of the thread. Toad No. 1 swallowed two such false larvae in the course of one trial. These false larvar were smeared with oil of clove and iodine solution in different trials. OLFACTORY REACTIONS IN AMPHIBIANS 625 In the presence of the living larvae, no attention was paid to motionless artificial larvae, but if these were set in motion they were sometimes taken. Effects of the substances used in these experiments may be summed up’as follows: The odor of the iodine solution was not deterrent in any manner; the after-effect seemed to be more disturbing in the mouth region than did that of other substances. Oil of pennyroyal and oil of rose geranium when they touched the lips led to wiping the mouth-parts with the forefeet. No similar effect was noticed from clove oil, cedar oil, and bergamot oil. When bisulphide of carbon was applied to the mealworms, the toads gave some evidence of discomfort, gulping several times in succession after having taken larvae prepared with this substance. Aniline oil did not cause any re- actions that seemed to result from stimulation of the mouth- parts, nor was it repellent because of its odor. FOOD AND DARKNESS The diurnal retreat of the toad into dimly illuminated places and its habit of feeding in the night are well-known. Frogs on the contrary are most active during the daytime. In con- nection with the feeding experiments on the toad it was desir- able to know what degree of darkness was prohibitive to its finding food. Are other stimuli present, do smell, hearing and touch share in the reactions of seeking food? 1. Materials The same individuals that were used in the previous. experi- ments were used in these trials. The toads appeared to be in normal condition, giving no evidence of having undergone any untoward experiences. No changes were made in the care given the toads. The experimentation was done in a photographie dark-room with controllable illumination. The experimentation chamber was a box, 30 inches long, 20 inches wide, and 8 inches high. Loam and leaves had been allowed to remain in the box for some days, thus giving the cham- ber some semblance to the toad’s natural habitat. 626 JONATHAN RISSER 2. Experiments To ascertain whether any stimulus other than the optic is called into play in the finding and taking of food the following experiments were performed. Toad No. 1 was placed in the box, which was empty except for a covering of moist filter paper on the floor, and allowed to remain undisturbed for some time. There was no excited hopping about during this period of illumination. The hght having been cut off, the toad crept along the wall, and finally came to rest in a corner where it attempted to burrow as though it were on soil. After two hours of darkness some mealworm larvae were introduced into the box. These were placed centrally on the floor and as these changes were made in darkness the larvae were not seen by the toad. The toad was then lifted from its position in the corner of the box and placed imme- diately over the glass plate on which the larvae were and facing them. In this position there was no thigmotactic stimulation, as there would have been in the corner position. By listening to the sounds within the box the movements of the toad could be followed. The position of the larvae and the toad were noted at the end of two minutes of darkness. The toad had resumed the corner position, the larvae had moved out and away from the plate, one of them being within 2 inches of the toad. Two minutes of darkness brought no further movement on the part of the toad. The larvae were distributed around the sides of the box. After another minute of darkness, the light was turned on and at the first visible movement of the mealworms, the toad turned and jumped toward them. There followed no movements on the part of the toad in successive periods of darkness; but each time the box was illuminated the least movement on the part of the mealworms called forth char- acteristic movements from the toad. Toad No. 2 was tried with dungworms in a manner similar to that just described for Toad No. 1. No notice was paid to the worms in the dark but in the light the toad followed the: 7 OLFACTORY REACTIONS IN AMPHIBIANS 627 worm some distance and finally picked it up, using both the tongue and lips to do so. In another trial, the toad was placed in the center of the floor and the worms were distributed in the portions of the box near the Walls. The illumination was lateral, thus leaving one side of the box quite dark. Movement of worms in the darker regions of the box were not seen by the toad, even when it was turned toward the worms. When a worm moved in the light part of the box, the movement soon attracted the attention of the toad and the characteristic reaction followed immediately. Repetition of the experiment several times with other toads gave the same results. When the dungworms were replaced by pillbugs (Oniscus), these were immediately caught when in the light, but not when in the dark. Other trials with mealworms, earthworms and dungworms gave like results; only during the periods of illumination was the food sought and then with no discrimination in favor of any particular form. Other trials, with very dim daylight as a means of illumi- nating the chamber gave identical results. If the mealworms or the dungworms were in the darker portions of the chamber, they were unnoticed by the toad. When the box was covered so as to cut off all illumination, there were no movements on the part of the toad; darkness prohibiting entirely any food seeking activities. b* In the lighted portions of the chamber the worms were only seen if in the direct visual field; if somewhat to the rear or to the side of the toads they apparently could not be seen, except when ~ in very vigorous motions. 3. Abnormal odors and darkness In connection with the darkness experiments, trials were made with the same odorous substances that had been used in the preceding experiments. The same individual toads were used as before. Mealworms were treated with appli- cations of oil of clove, carbon bisulphide, and oil of rose gera- 628 JONATHAN RISSER nium and were allowed to remain in the experimental chamber with the toads under the same light conditions as described in the preceding trials. The presence of the odors did not stimulate the toads to seek or avoid the worms in the dark. Again, as observed in the previous trials, the presence of the odorous substance did not deter the toads from taking food in the light. 7 ODOR-STREAM EXPERIMENTS 1. Apparatus From the experiments previously recorded there was no positive evidence that odors were concerned with the taking of food by the toads. Odorous substances when taken into ~ the mouth by accident do not stimulate the receptor organs to the degree of inhibiting the act of deglutition. A possible explanation for this might be sought in the relation of the ex- ternal nostrils to the mouth. Even if the olfactory function were present to but a slight degree, it might be possible to dem- onstrate its presence by leading the odors directly to the nasal opening. If the vapors could be brought directly to the epithe- lial surfaces in appropriate manner, reactions might occur. It seemed desirable to devise some method for doing this. After some trials, an apparatus of a satisfactory kind was finally devised. An outline of it is shown in figure 1. The arrangement of the parts is as follows: From the reservoir, A, water displaced the air contained in B. By means of screw clamps at x and y the rate of flow of the water was controlled. The clamp, x, was adjusted for the flow, while clamp y is used for starting and stopping the stream as a whole. C is a small reser- voir in which was suspended a vial, D, containing the odorous substance used. The distal end of the outlet tube, 6, was sub- merged in the material contained in D; the air stream carried over and out of C would consequently be impregnated with the odor to be tested. By means of appropriate connections and the nozzle tube, d, three millimeters in diameter and bent in proper form, the air stream was led into the experiment chamber, £. Elastic suspension of the tube, d, automatically raised the tube OLFACTORY REACTIONS IN AMPHIBIANS 629 and allowed freedom of movement. For control tests there was provided a duplicate of the tube, d, which could be connected directly with B; thus avoiding any possible errors by contami- nation from the material in D. The chamber, /, was an open eylinder, the ower end resting in a shallow vessel containing soil easily changed and moistened. To eliminate other dis- 7 | Fig. 1 A, reservoir; B, air chamber; C, odor chamber; D, vial; H, experiment chamber; F, lamp; a, water tube A to B; b, airtube B to C; c, outlet tube for odor stream; d, nozzle tube for chamber £; e, vessel containing sand; x, y, clamps. turbing factors the cylinder was covered with black cloth. The illumination of the chamber, FL, was by an electric lamp of 8 c¢.p. This was of advantage in orientation, as the toads are positively phototropic (Pearse ’10). The apparatus was arranged near the water supply in a base- ment room from which daylight could be excluded. Several other chambers similar to H were provided in order to adjust ~ 630 JONATHAN RISSER the toads to the new conditions. Toads not accustomed to the chamber were at first much disturbed, reacting to various other influences. Attempts to crawl out of the chamber, to bite the end of the tube, d, or to be restless within the chamber were some of the reactions noted. After adjustment normal re- actions were then exhibited, such as feeding or burrowing into the soil. The apparatus was kept well ventilated between periods of experimentation as well as during the testing periods. 2. Rate of flow of odor-stream The rate of flow of the air stream was made as uniform as possible. If allowed to flow uninterruptedly the 3000 cubic centimeters of air in B were displaced by the water from A in from 80 to 90 minutes. At this rate the flow is approximately 35 cubic centimeters per minute. That such a stream is not sufficiently strong to be perceptible can be demonstrated by plac- ing the nozzle tube near the moistened lips or the tongue. - The flow of the stream could also be judged from the frequency of escaping bubbles when the end of the tube was immersed in water. The tests as carried out followed a uniform plan consisting of, first, a control test with air, and secondly, a trial proper with the odor stream. 3. Methods In the control test the toads were first subjected to the air stream coming from B. The nozzle tube, d, was directed toward various body regions, namely, flank, axillary region, anal region, and surface of the eye (as near as possible without contact). After this test for possible stimulation of integumentary sense organs, the air stream was next directed to the nostril to ascer- tain if there was any stimulation of the nasal epithelium. The test with the odor stream followed next and the same parts of the body were tested as before. The toads used in the feeding experiments were used for the tests with the odor stream and were normal in all their activities. OLFACTORY REACTIONS IN AMPHIBIANS 631° The one exception was Toad No. 4, which seemed to be in a semi-hibernating condition during the entire period. Two other toads taken in the fall of 1912 were likewise tested with this apparatus. Although no two individuals were identical, the results of the trials agree in all essentials. 4. Substances used in the tests Oil of cloves, oil of pennyroyal, oil of rose geranium, cedar oil, bergamot oil, aniline oil, carbolic acid, olive oil, castor oil, and cod liver oil were used, though some of the substances gave entirely negative results. : Trials were also made with the odor stream from food ma- terials used by the toad. For these tests mealworm larvae, earthworms, dungworms, and cockroaches and other insects were put directly into the chamber, C and the air stream was allowed to carry over any odors that were present. In no case was there evidence of olfactory stimulation from these bodies. 5. Experiments The attempt of Graber (’85) to determine the olfactory reactions in some of the amphibians proved unsatisfactory. Aronsohn (’86) and Gourewitsch (’83) noted the effect of odors on the rate of respiration. The animals, while confined in covered beakers, were exposed to odors of turpentine and eau de cologne with disturbing results. Early in my own tests it was seen that chloroform, ammonia, ether, turpentine, formol or alcohol when introduced into the chamber were disturbing. Even though the air stream with these ‘substances was not near the nostril, reactions followed very quickly and with the proximity of the stream to the nose the disturbance was increased. Ammonia, chloroform and turpentine when directed upon the anterior head region induced the toad to jump away from the tube to one side or even out of the chamber. The odor of turpentine resulted in a very characteristic attitude, the head > 632 JONATHAN RISSER being bent down between the forelegs, the body raised from the ground and best described as humpbacked. Ether or alcohol did not produce reactions so decidedly vigorous, though the toads retreated quickly from the tube. The respiratory movement was retarded by all the substances that stimulated the olfactory end-organ. ‘Tests made with the odor stream when the act of inspiration was suppressed were devoid of any resultant motor activity. After the earlier tests with ammonia, ether, etc., the essen- tial oils were used and some were found to be more effective in calling forth. characteristic reactions than others. Some were entirely negative in effect, as, for instance, castor oil, cod liver oil, and olive oil. What reactions could be considered as directly called out by the stimulation of the olfactory organ? When the odor stream from oil of cloves or pennyroyal was allowed to spread over the anterior region of the head and pre- sumably enter into the nasal cavity with the inspired air, the first motor act consisted of a slight bending down of the head, away from the nozzle of the tube, and a cessation of the respira- tory movements. If the tube were removed quickly, the toads soon resumed the normal position and the respiration move- ments went on again with little interruption. If the odor stream was allowed to flow continuously upon the nostrils, respiration was entirely suppressed for some time. Often the toad made motions with the forelegs resembling wiping. When very much stimulated by the stream the animals moved away from the tube. The bending of the head was most noticeable when the toads were partly hidden in the sand. At such times respiration ceased, the head was bent more anteriorly and the. animals endeavored to burrow down into the soil. The act of burrowing could be hastened by directing the tube repeatedly at the nostrils during such a test. This position of being partly buried in the soil was most favorable for observation. When ~ thus buried the animals did not move about and the stream could be directed more accurately against the nostrils than at other times when the animals were free to move about. OLFACTORY REACTIONS IN AMPHIBIANS 633 The tests were carried out in periods of approximately five- minute durations, followed by intervals of rest varying in length of time. It sometimes happened that the animals were too rest- less to allow a fair interpretation of their reactions; in such cases the tests were discontinued. Other individuals again, were amenable to the trials for several hours without showing them- selves to be disturbed. Precaution was taken that no odors entered the chamber otherwise than by means of the appropriate tube. Records of the trials indicated what reaction followed the stimulus applied to the nostrils. The records shown in table 1 will serve as samples. TABLE 1 Toad No. 1 TIME . STIMULUS REACTION 3.00 | air _ moved to one side 3.05 | repeated burrowing normally | toad was allowed to rest fifteen minutes 3.20 odor of oil of clove head bent downward | repeated head bent downward repeated wiped with right foot repeated moved away to left = repeated immediately moved away to left repeated wiped; attempted to climb 3.25 | repeated wiped again Toad No. 2 10.30 air at rest repeated at rest repeated at rest 10.35 repeated at rest Toad allowed to rest five minutes 10.40 | odor of oil of pennyroyal wiped repeated head bent downward repeated immediately head bent downward repeated moved toward right repeated — head bent downward | repeated immediately head bent downward 10.45 | repeated * head bent downward 634 JONATHAN RISSER Bergamot and cedar oil resulted in fewer reactions than oil of clove and oil of pennyroyal, showing them to be less stimulating than the other oils. As already indicated, the stimulus was effective only while inspiration*was in progress, for pressure of the stream was so slight that it could not force the odorous particles into the nos- trils. The reactions were not characteristic of the different odors. It might be said that the reactions differed in degree, only, since there were only differences of vigor with which the reactions were executed. Even individuals differed in this respect at different times during the trial periods. Other factors such as tempera- ture or tendency to hibernate may have had some influence in this aspect of the reactions. Absolute parity of the tests could of course not be obtained. A The relative effectiveness of the odors .in stimulating the receptors was as follows: Most effective and approximately equal were oil of cloves, and oil of pennyroyal; less effective were oil of rose geranium and cedar oil; and least effective, with reactions infrequent, were bergamot oil and carbolic acid. : Odors from castor oil, olive oil, cod liver oil, living meal- worms, earthworms, dungworms, cockroaches, decaying meat and decaying leaves in soil were without effect. 6. Controls and operations The controls used in the tests seemed to be confirmatory of the presence of the olfactory function; yet a possibility existed that other receptors had been stimulated at the same time. The ophthalmic branch of the trigeminal nerve supplies the region of the head anterior of the nostrils, with some fibers possibly present within the nasal capsule. This being so, stimuli coming to this region of the head could call forth reactions easily through this branch as well as through the olfactory organ. To ascer- tain whether this branch had been stimulated several sets of experiments were tried. The nostrils were closed by suturing with, silk thread. Closing the nostril in this manner was of more serious consequence to the OLFACTORY REACTIONS IN AMPHIBIANS 635 toads than to fishes. This established an effective hindrance to the respiratory current, therefore it was found impracticable. If the stitches were not placed deeply, the muscular movements of the nostril soon caused the thread to cut through the margins, Furthermore, complete closure of the nostrils was apparently fatal in several preliminary trials. Filling the nasal aperture with melted vaseline was only partly successful, entrance to the capsule being prevented by the nostril valves. The vaseline could be “troweled” into the opening. It was evident from the actions of the animals that the vaseline was discomforting” The toads showed decided restlessness and attempted to wipe their heads. Some tests were made with toads so treated. When both nostrils had been successfully closed, no reactions traceable to odorous substances could be observed. If only one of the nostrils was closed, the odorstream, when directed upon the open nostril, was effective in causing a reaction. Clove oil and pennyroyal were the substances used-in the tests with the nostrils partly or wholly closed, these having been most efficient in calling forth reactions in the normal toads. To be certain of the assumption that the reactions noticed with the odorstream were called forth by the stimulation of the olfactory receptor, operations of two kinds were performed on the toads. In one of the operations, the olfactory tract was: severed. In the other, it was necessary to section the ophthalmic branch of the fifth nerve. After etherization, the olfactory tracts were severed at the anterior border of the eyeball. The ophthalmic branch of the trigeminal was cut by piercing the integument of the optic capsule at the anterior inner angle of the orbit and cutting across the floor of the orbit. After recovery, only the specimens reacting normally to other stimuli were used for the tests. Toads in which the olfactory tracts had been severed did not react to the odorstream. If partly buried in the sand, they were undisturbed by odors coming through the tube. The head was not drawn down, nor did they move away from the tube. One THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 15, No. 4 636 JONATHAN RISSER peculiar action was noted in the toads so tested; when the stream was first directed to the nostrils, they sometimes would gasp, - often repeatedly, but without interrupting the respiratory act. The reactions of the toads in which the olfactory tracts had been severed are well exemplified in the record of Toad No. 1 (table 2). The reactions of toads in which the ophthalmic branch of the trigeminal nerve has been cut is shown in table 3. | Since oil of cloves and oil of pennyroyal had given the best reactions in the normal toad these substances were used after the operations. As may be seen from the record above, the opera- tion on the ophthalmic branch of the fifth nerve did not disturb the olfactory function. Ether and chloroform did not affect the operated toads in any manner different from the normal animals. After cutting the olfactory tract the animals were apparently less active than toads with the ophthalmic branch severed. TABLE 2 Toad No. 1 TIME | STIMULUS i | REACTION 1.10 odor of oil of clove resting repeated often moved forward | repeated often wiped once _ 115 with continuous odorstream | respiration normal Toad allowed to rest for ten minutes anid chamber well ventilated 1.25 odor of oil of clove gasped at first 1.30 continued for five minutes | wiped once Allowed to rest ten minutes 1.40 | odor of oil of clove continuous for four minutes wiped once 1.45 further stimulation gasped once | Allowed to rest ten minutes 1.55 odor of oil of clove quiet; respiration sus- | : pended | continuous odorstream respiration resumed 2.00 | continuous odorstream | no reactions | Allowed to rest ten minutes 2.10 | odor of oil of clove | 2.15 | continuous stream changed position once and respiration normal The latter were more responsive to the optic stimulations. toads with the olfactory tract cut preferred to lie quietly in the sand. The feeding reactions were not inhibited by either operation. Some toads collected in the late summer of 1912, tested simi- larly, gave evidence coinciding with the data of the previous tests. The results of these experiments show that odors stimulate the olfactory receptors, whereby certain motor activities of OLFACTORY REACTIONS IN AMPHIBIANS unquestionable value to the animals are called forth. TIME 9.35 9.55 TABLE 3 STIMULUS REACTION odor of oil of clove repeated repeated repeated repeated Toad allowed to rest fi odor of oil of clove repeated repeated repeated Allowed to rest five odor of oil of clove repeated | repeated _ repeated; applied to left nostril repeated; applied to right nostril Allowed to rest twent odor of oil of clove repeated repeated repeated repeated Toad allowed to rest five odor of oil of clove repeated repeated repeated repeated head down wiped with right forefoot and moved backward moved away from tube head down and moved backward moved away from tube ve minutes head down withdrew from tube . head down wiped head twice minutes head down, drew back head down head down wiped left side of head wiped right side of head y minutes head down head down drew back head down head down minutes drew back, drew back jumped away drew back, drew back head down, head down jumped away, drew back The 638 JONATHAN RISSER EXPERIMENTS WITH TADPOLES 1. General Knauer (’75) emphasizes the fact that some of the larval anurans are less phytophagous than is commonly supposed and instances the fondness of toad tadpoles for animal food. Holmes (07) also states that decomposing insects, earthworms, etc., are acceptable as food to frog tadpoles. That certain European anurans feed indiscriminately like the earthworm is a well- known fact. Whether vegetable material must not be in the early stages of decomposition before becoming available as food for tadpoles has not yet been determined. It is certain that many unicellular and filamentous algae may be ingested repeatedly before digestion is complete. Nagel (’94) denied the olfactory function to aquatic verte- brates. His theoretical objections have been refuted for the fishes by the work of Parker (’10; 11), Sheldon (11), and Cope- land (712). The question still remains open as regards am- phibians and its solution depends upon suitable material and methods. This was reason sufficient to attempt an investigation of the olfactory function in the tadpole of the toad. Preliminary trials gave evidence at once that the tadpoles were responsive to the presence of decaying animal matter in their immediate vicinity. Tadpoles placed in a vessel containing filtered water for a day or two, the feces being removed frequently, soon come to be in a state of hunger. Water not filtered contained at all times organic material enough to form a delicate film upon the walls of the vessel. This film of partly decomposing organic matter is a source of food for the tadpoles and must be removed. For like reasons the feces must be taken away. In the preliminary trials particles of dead earthworms, dead fish or bits of meat undergoing decomposition were placed in the vessel containing the hungry tadpoles Such particles of food were quickly found by the tadpoles. The feeding trials were performed on tadpoles of three suc- cessive years. The first set was obtained at Woods Hole, Massachusetts, on August 10, when the tadpoles were about to OLFACTORY REACTIONS IN AMPHIBIANS 639 metamorphose. Some of the larvae had one, others both pairs of appendages avell developed, and a few specimens passed through the final phases after having been brought into the laboratory. They were kept in shallow vessels with some of the bottom detritus from their original habitat. Experimentation and natural causes gradually diminished the numbers available for the trials, but the tadpoles appeared to be normal in their activities at alltimes. The tadpoles surviving through the period of experimentation were killed by accident in the month of January following. The second set were tadpoles taken in May of the following year in the vicinity of Cambridge. They did not lend them- selves well to experimentation. Conditions appeared to be unfavorable for them in the laboratory. Experiments therefore were carried on with but small numbers, but so far as these went they were corroborative of those from the first set. The tadpoles included in the third set, were taken in June, 1913, from two different localities: namely, Cambridge and Woods Hole, Massachusetts. The tadpoles were experimented with particularly to show that the olfactory reactions might be completely checked by a certain procedure and then revived. 2. Methods The method of procedure in the trials was as follows: The tadpoles were placed in filtered water for twenty-four to forty- eight hours before experimentation began. Then the food was introduced into the vessel; the reactions being noted in accord- ance with the shifting of the food placed in the vessel. Food used in the tests consisted of particles of fish, earthworm, or meatin decomposition. At first the food without any envelope was placed free in the vessel, but later it was wrapped in cloth. In other trials again, two packets were placed in the water; one containing the food, the other without it. Very little difficulty was experienced in noting the reactions of the tadpoles. When the two packets were presented, the tadpoles distinguished quickly the one containing the food. 640 JONATHAN RISSER When confined to shallow vessels, the actions of the tadpoles of the toad and of the frog of approximately the same stage, were somewhat different. In general, the frog tadpoles swim along the vessel walls. The toad tadpoles, on the other hand, move in all directions over the bottom or near the upper sur- face of the water. When toad tadpoles came into the vicinity of the food, either covered or open, the reactions seemed to indicate stiniulation of some kind. They swam from side to side, and often when near the food would turn directly to it. When the food was found the tadpoles attacked it very eagerly. When two packets had been introduced the tadpoles did not remain upon the packet without food nor did they nibble it as they did the one containing food. 3. Experiments a. First set of tadpoles. The records of several of the trials are here given in detail: TIME 8.20 Food packet placed in shallow dish containing eight tadpoles 8.22 Food found by first tadpole 8.23 Food found by second tadpole 8.39 Food found by six tadpoles S$ 8.50 The position of the food was changed 8.55 The food found by three tadpoles 8.57 The food found by five tadpoles ' 8.05 The food found by seven tadpoles In another trial a small piece of fish not covered was placed in the dish. 10.30 Food placed in dish containing eight tadpoles 10.40 Food found by six tadpoles In another trial two packets were placed in the dish, one with food, the other without. 2.85 Food placed in dish with six tadpoles 2.43 All the tadpoles had found the food and were nibbling None of the tadpoles were at the false packet. 2.55 Reversed position of the two packets 3.20 Food found by five tadpoles In a similar trial with two packets, one containing fresh meat, the other without. 10.42 Packets placed in dish with fifteen tadpoles 10.50 Food found by twelve tadpoles 10.56 Reversed position of packets “11.05 Food packet found by ten tadpoles OLFACTORY REACTIONS IN AMPHIBIANS 641 The totals of the trials carried on with the larva of the first set gave these results: In 160 trials, food packet was found 120 times. ‘Trials carried on from time to time up to the accidental loss of the tadpoles resulted in similar ratios. b. Second set of tadpoles. The conditions were unfavorable for keeping these tadpoles and because of this circumstance only a small number of individuals were available. The trials were carried out in a manner similar to those for the first set. It needs only be said that the results were similar to those obtained from the experiments of the first set. Precautions were taken with this set to have the two packets identical in appearance, and to transpose them in position. To avoid the accidental finding of the packets as much as possible, they were placed some distance from the sides of the vessel. The toad tadpoles swam more rapidly than the frog tadpoles, and were also more erratic in their movement in the water. It could be easily determined whether they were influenced by the presence of the food mass near them. In a few cases only did there seem to be a visual stimu'us influencing the tadpoles to react; such being occasioned by the very lively actions of other tadpoles already at work on the food mass. c. Third set of tadpoles. For the purpose of verifying the results of the previous experiments, similar experiments were performed with the tadpoles of the third set in June and July of the third season. As stated, these tadpoles were obtained from two different localities and kept successfully in the labora- tory, metamorphosis being deferred for the time being. The tad- poles were unquestionably those of Bufo americanus, having: been obtained earlier than the time when the mating eall of Fowler’s toad was heard. : The food used in these experiments was dead and partly decomposed frog tadpoles, earthworms, and beef liver. These materials were either placed freely in the vessel or covered with cheesecloth one or two layers in thickness. In a similar way, as previously described, two packets were used in some of the trials; one containing the food, the other being identical in appear=- ance but without food. 642 JONATHAN RISSER A number of trials in which the food was placed in the vessel with the tadpoles showed that the reactions of the tadpoles depended somewhat on the size of the vessel. In a small vessel the time necessary for the food to be found by a certain number of the tadpoles was usually less than in a larger vessel. The same is true if the food is placed near the side of the vessel, say within an inch of the side wall; in such a case the animal swimming through the zone impregnated by the odor is influenced more quickly than when the food is centrally placed. In the trials with two packets discrimination was very evi- dent in the actions of the tadpoles. In the beginning of the trial the reactions were at all times apparently without choice. After the packets had been allowed to remain in the water for a short time, the tadpoles always endeavored to feed from the packet containing the food. Even if the tadpoles rested on the ‘dummy’ packet very little attempt was made to nibble. Trans- position of the two packets was accompanied by a corresponding redistribution of the tadpoles. This last method was modified several times, by exchanging the envelopes of the packets, or by substituting an envelope saturated with the odor of the food material for the food itself. When this was done, the tadpoles congregated upon the food- saturated envelope, finding it as they did the food open in the water, or when contained in the envelope and allowed to remain for some time. An attempt was made to determine whether the tadpoles would orient themselves to an odor-saturated waterstream which was allowed to flow into a vessel containing them; but this experiment was fruitless of results, although there was some slight evidence that the current was especially stimulating. Corroborative of the earlier findings of the experiments on the two previous sets, experiments carried out with the third set were of greater value in this study in what may be called operative tests. To determine whether the reactions of the toad tadpoles as already described are due solely to the stimulus received OLFACTORY REACTIONS IN AMPHIBIANS 643 by the olfactory organ, attempts were made to inhibit such stimulation in ‘Various ways. This was found to be rather difficult and in the light of the earlier attempts the evidence was not very convincing. The diminutive size of the tadpoles is the principal factor militating against successful operations. Cutting the olfactory tract is the method by which inhibition may be made absolutely certain. In this method the chief difficulty lies in determining the proper degree of anesthetization for the operation and subsequent revival of the animal. Chlore- tone of 0.1 per cent was used for this purpose, as the tadpoles could be brought into a vessel with fresh water and revived. After being anesthetized the tadpoles were bedded in a bit of absorbent cotton held in the hand, and with a needle the cranial case was pierced in the median line anterior to eyes. The shock attendant upon the operation or the manipulation was dis- astrous in most cases, and only afew specimens survived. Within a few days these tadpoles became less vigorous, and finally all succumbed. The few individuals so operated upon and tested for reactions re- sulting from stimulation to the olfactory sense organ did not give sufficient evidence from which satisfactory conclusions could be drawn. In another manner the inhibition of the stimulation and the reactions was also attempted. The external nares of the toad tadpole are comparatively large and it was possible to fill the nares with white vaseline, the tadpoles being bedded in moist absorbent cotton. Tadpoles so treated showed the presence of the vaseline to be disturbing in effect. The plugs prevent the respiratory stream from entering the nasal openings, the swim- ming movements are not so vigorous as under normal conditions, the chief endeavor of the tadpole being to free itself from the disturbing material. The temperature of the water is to be taken into account here, as the vaseline could be removed quite readily by placing the tadpoles in water slightly warmed. As soon as the vaseline has been dissolved the tadpoles again act in normal manner. This method of treatment was made use of in the first 644 JONATHAN RISSER and second sets, but on account of the comparatively small num- bers at hand, the results could not be considered as conclusive. With the material of the third set the chief aim was to deter- mine whether the presence of the vaseline plugs was beyond doubt inhibitive of the apparent olfactory reactions. Some operations of cutting the tract were carried out, but the after effects were in most cases disastrous to the tadpoles, and there- fore special stress was placed on plugging the nasal openings. Making use of the same individual tadpoles in the successive trials, these groups were isolated in filtered water for some time previous to each trial, and tested repeatedly for reactions to the food as presented (without an envelope) while the nasal openings were in normal condition, or when filled with the vaseline plug. The trials with the tadpoles in this series were carried out in the following manner: The tadpoles were isolated in filtered water for at least twenty-four hours, in a few cases forty-eight hours, but not beyond this, as the tadpoles showed that a longer period without food was disastrous to them. After isolation for the designated length of time, the tadpoles were tested for the presence of food in the water; immediately after this the nasal plugs were put in, and the tadpoles transferred to a vessel free of food or vaseline. The food was then introduced. Unless the food mass was placed immediately in the proximity to the tadpoles they exhibited no such activities as they pre- viously had. When swimming and moving about in the vessel, there were no such positive attempts to find the food as when normal. The trials were repeated several times with different groups of individuals, in each case allowing several days to in- tervene between the trials, the tadpoles being provided with food and water from the stock. The tadpoles showed no bad effects from the plugs of vase- | line in the nazal openings, reacting freely in the later trials as well as in the earlier experiments. Tests made with the tadpoles having the plugs in the nares showed them unable to discriminate between two packets one of which contained food, while previous to the plugging definite choice had been made by them in favor of the food packet. OLFACTORY REACTIONS IN AMPHIBIANS 645 The conclusion’ reached from these experiments as described for the different broods of tadpoles is that the presence of the nasal plug is effect ve in inhibiting any stimulus coming to the sense organ and consequently no corresponding reaction follows. DISCUSSION 1. Toads In its method of obtaining food the toad seems to respond to the visual stimulus entirely. This stimulus is apparently effective only when it involves motion. It is not always fol- lowed by perfect reaction, for substances inappropriate as food are often taken accidentally. Rejection of such material occurs in compliance with mechanical or tactile stimulation. Nor does the gustatory function appear to be of any im- portance in feeding. Gaupp (’04) does not consider the function of the epithelial endplates of the mouth cavity to be established. He refers to Bethe as bringing forward the best evidence favoring them as tactile organs. It has been shown that strong solutions of picric or acetic acid applied to the epithelium of the mouth cavity will cause appropriate motor activity. Such reactions take place without reference to the point of application of the acid, whether this be in regions supplied with endplates or de- void of them. The latent period between stimulus and reaction is of appreciable duration. These considerations speak against the belief that these organs are gustatory in function. Until the contrary is proven, the similarity of food materials of frogs and toads argue for similarity in structure of the epi- thelium of the mouth. Observations and feeding experiments tend to confirm this view. Food does not remain in the mouth cavity any great length of time. Deglutition follows almost instantly, and therefore.the sense of taste would be of minor importance. Under certain conditions regurgitation may occur. If substances disagreeable to taste are taken into the mouth the animals might be expected to resort to this expediency more often. 646 JONATHAN RISSER In no ease in the course of the feeding experiments was cog- nizance taken of the unusual substances coming into the mouth to the degree that food was ejected or regurgitated. The ex- perience cited by Knauer (’75) of toads refusing decomposing earthworms is probably referable to tactile stimulation. The nature of the organisms serving as food, for the toads under natural conditions is such, that materials differing in texture markedly from the normal might be sufficient to cause refusal. | To establish any connection between food used by toads and the possible odors inherent to the food seems difficult at present. Although evidence is negative, this is qualified by the fact that the data are really not sufficient to establish any conclusion on this question. Natural foods are apparently taken indiscriminately. Stimulation of the receptors may take place; the presence of such stimulation and the effect are not yet demonstrable. Our inability to recognize the quality of particular motor reactions following certain stimulations does not argue against the absence or refinement of reactions. The experiments of Graber (’85) were such as to allow no great value to be attached to them. The unmodified methods used for forms differing so greatly in phylogenetic position and in habit with the very doubtful reactions as recorded, speak against the acceptance of his data as important. To a similar degree the experiments of Aronsohn (’86) on odors and respiration are of little value in indicating the use of the olfactory organ in anurans. His experiment does not pre- clude stimulation of the trigeminal nerve. His choice of sub- stances and the manner of experimentation favor the possibility that the fifth nerve is involved. The instance mentioned by Conradi (01) has not had confirmation of any kind. In the experiments described, the presence of abnormal odorbearing substances has not given origin to stimulation sufficiently strong to inhibit the desire for food. In all experiments carried out with the lower animals and their reactions toward solutions or vapors there exists the pos- sibility that solutions or vapors are more dilute than was in- OLFACTORY REACTIONS IN AMPHIBIANS 647 ’ tended. Whether stimulation is to take place in air or water, it is extremely difficult to work with solutions or gases abso- lutely standardized. The results of the experiments as con- ducted show that odors when in relation with food are not sufficiently deterrent in action to compel the toad to refuse such food. Odors of natural surroundings may stimulate the toads to certain reactions. At present there is no evidence that odors of soil or water are effective in any degree on the olfactory organ of the toad. 2. Tadpoles In contrast with the tadpoles of the frog, the toad tadpole may be claimed to possess an olfactory sense, and possibly to a much greater degree than might be supposed. Anatomically considered the two species appear similar. Differences make themselves evident in a closer study of certain parts. The nasal openings in the toad tadpole are relatively larger than in the frog tadpole. The water stream into the nasal passage of the toad tadpole is therefore of greater magnitude than in the frog tadpole. Contrary to Exner (’78), as quoted by Gaupp (’04), the nasal openings in both species serve for the incurrent water stream. If to this stream are added other substances: dilute solutions of methylene blue or particles of carmine, the toad tadpoles are very quick to respond to the stimulus. Such substances added to the water current flowing into the nasal chamber of the tadpoles of the frog (Rana virescens and R. catesbiana) produce similar reactions. Toad tadpoles react at the immediate entrance of the first substances into the nostril; tadpoles of the frog will permit the stream to flow into the nose for a long period, reacting much more slowly to the stimulus. Probably there is a me- chanical stimulus from the carmine which sets free the response. It is more than probable that toad tadpoles recognize certain foods and their odors. When given the choice as between decay- ing animal matter and decaying plant substances the former is preferred. 648 JONATHAN RISSER In the series of tests there was always positive reaction in the presence of the animal food when opportunity for such choice was given. Organs of taste have not been demonstrated in the mouth of the tadpole. On the other hand, the differentiation of the nasal epithelium into the characteristic olfactory organs and the supporting structures takes place very early and we may safely assume the reactions noted to be the result of stimulation of the olfactory receptors. 3. Conclusion The anurans have been called microsmatic animals; such dis- tinction being based on histological and anatomical comparison with the animals classed as macrosmatic. The nasal organ of the Anura is a common respiratory and olfactory organ, and in this respect conforms to the organ pos- sessed by the higher vertebrates. A chambered nasal cavity of considerable magnitude supplied extensively with olfactory epithelium and adequate connections with the central nervous system, predicate functional activity of the sense organ. The life habits and the phylogenetic position of the anurans suggest that such functional activity not only be present but adequately developed. The presence of receptor organs identical in structure with those found in the higher vertebrates further postulates that functions peculiar to these structures be identical. The receptor peculiar to the olfactory organ of all verte- brates consists of a neurone whose cell body is peripheral in posi- : tion. The distal portion of the neurone is characterized by the protoplasmic processes projecting above the level of the sur- rounding cells, while the proximal end is attenuated and gives rise to one of the fibers of the olfactory nerve. This type of. receptor is directly comparable structurally with neurone cells found in the epidermis of many invertebrates (Parker ’12). Neurones such as these, found in some of the invertebrates are distinct portions of the receptor-effector system and have been demonstrated as extremely sensitive to chemical stimuli. That EE — OLFACTORY REACTIONS IN AMPHIBIANS 649 such receptors, olfactory in function and responding only to stimuli when the cell surfaces are dry were peculiar only to air breathing animals was the earlier assumption. The work of Aronsohn (’86) and Veress (’03) has shown that in man and the higher vertebrates the olfactory epithelium is bathed by glandular secretions and whatever stimulus reaches the receptor must do so in the form of solutions. More recently Baglioni (’09), Parker (10; 11), Sheldon (711) and Copeland (712) have demonstrated in aquatic animals the stimuli inducing certain reactions to be identical with those noted in the air-breathing animals. If this assumption is valid, as it appears to be, that the olfac- tory receptor is the simplest and least differentiated of receptor neurones and stimulated by extremely dilute solutions, we should find the olfactory organ in Anura readily stimulated. _ That well defined and characteristic motor reactions have not yet been recognized as results due to stimuli varying in quality may be due to lack of observational data and methods of experiment. SUMMARY 1. There is no evidence that toads react to olfactory stimuli pertaining to soil, water, etc. 2. The character of the food is not differentiated by attendant odors to the degree that the adult toad thus distinguishes it. 3. Substances of unusual character and odors, when associated with food, do not stimulate the olfactory organs in such a man- ‘ner as to bring the toad to refuse the food. 4. The presence of such substances in close proximity to the toad, and invisible because of darkness are not repellent in effect on the toad. 5. Odorstreams specific in character, made to flow over and into the nasal openings stimulate the olfactory sense-organ; such stimulation causing definite motor activities to follow. 6. Appropriate operations are confirmatory that the stimulation by such odorstream is olfactory. Section of the olfactory tract inhibits the reactions. Olfactory stimulation and reactions are 650 JONATHAN RISSER not affected by section of the ophthalmic branch of the trigeminal nerve. : 7. Under circumstances allowing discrimination, the tadpoles of the toad prefer animal foods. 8. Such discrimination appears torest upon the appropriate stimulation of the olfactory receptor. 9. Tadpoles of the toad show by proper reactions that animal food is recognized, although not visually perceptible. 10. The receptor organ so stimulated must be a distance receptor and thus is olfactory in function. 11. In the metamorphosed toad the visual stimulus is the principal and guiding factor in procuring food. Therefore, it is inhibitory in relation to other stimuli and their resultant reactions. Postscript. Since the preparation of this paper, Copeland has published in The Journal of Animal Behavior, vol. 3, pp. 260 to 273, an account of the olfactory reactions of the newt Diemyctylus and has shown that this amphibian can scent food under water as a fish does. OLFACTORY REACTIONS IN AMPHIBIANS 651 BIBLIOGRAPHY AronsoHN, E. 1886 Experimentelle Untersuchungen zur Physiologie des Geruchs. Arch. fiir Anat. und Physiol., Physiol. Abth. Jahrg. 1886, pp. 321-357. Baationi, 8S. 1909 Contributions expérimentales 4 la’ physiologie du sens olfactif et du sens tactile des animaux marins. Arch. Ital. Biol., tome 52, pp. 225-230. Conrapi, A. F. 1901 Toads killed by squash-bugs. Science, new ser., vol. 14, p. 816-817. CopreLanp, M. 1912 The olfactory reactions of the puffer or swellfish, Spheroides maculatus (Bloch and Schneider). Jour. Exp. Zool., vol. 12, pp. 363-368. Exner, S. 1878 Fortgesetzte Studien ueber die Endigungsweise des Geruchs- nerven. Dritte Abhandlung. Sitzungsber. d. k. Ak. d. Wissensch. Math.-nat. Cl., Bd. 76, Abth. 3, pp. 171-221, Taf. 1-2. FIscHER-SIGWART, H. 1897 Biologische Beobachtungen an unsern Amphibien. Vierteljahrschr. d. Naturf. Gesellsch. Ziirich, Jahrg., 42, pp. 238-316. GarMANn, H. 1892 A synopsis of the reptiles and amphibians of Illinois. Bull. Ill. State Lab. Nat. Hist., vol. 3, pp. 215-389. Gaupp, E. 1904 A. Ecker’s und R. Wiedersheim’s Anatomie des Frosches. Dritte Abt., ix+961 pp. GourewitscH, A. 1883 Ueber die Beziehung des Nervus Olfactorius zu den Athembewegungen. Dissertation. Bern, 8vo, 18 pp. GraBer, V. 1885 Vergleichende Grundversuche ueber die Wirkung und die Aufnahmestellen chemischer Reize bei den Tieren. Biolog. Centralbl., Bd. 5, pp. 385-398, 449-459, 483-489. HartMan, F. A. 1906 Food habits of Kansas lizards and batrachians. Trans. Kansas Acad. Science, vol. 20, part 2, pp. 225-229. Hitt, T. 1873 Note on Bufo americanus. Proceed. Amer. Assoc. Adv. Sci., 22nd meeting, part 2, pp. 23-24. Hopae, C. F. 1898 The common toad. Nature Study Leaflet. Biology Series, No. 1. Worcester, 8vo, 15 pp. Homes, S. J. 1907 The biology of the frog. 2nd Ed. New York, 8vo, ix+ 370 pp. Kwnaver, F. K. 1875 Beobachtungen an Reptilien und Amphibien in der Ge- fangenschaft. Wien, 8vo, 57 pp. Lockwoop, 8. 1883 Bufo americanus at play. Amer. Nat., vol. 17, pp. 683-684. NaGet, W. A. 1894 Vergleichend physiologische und anatomische Unter- suchungen ueber den Geruchs-und Geschmackssinn und ihre Organe. Bibliotheca Zoolog., Bd. 7, Heft 18, viii+207 pp, 7 Taf. THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 16, No. 4 652 JONATHAN RISSER NEEDHAM, J. 1905 May flies and midges of New York. Bull. N. Y. State Mus., No. 86, 352 pp., 37 pls. Parker, G. H. 1910 Olfactory reactions in fishes. Jour. Exp. Zool., vol. 8, pp. 535-542. 1911 The olfactory reactions of the common killifish, Fundulus heteroclitus (Linn.). Jour. Exp. Zool., vol. 10, pp. 1-5. 1912 The relations of smell, taste, and the common chemical sense in vertebrates. Jour. Acad. Nat. Sci., Philadelphia, vol. 15, Second Series, pp. 221-234. Pearse, A. 8. 1910 The reactions of amphibians to light. Proceed. Amer. Acad. Arts and Sci., vol. 45, pp. 161-208. QuAINTANCE, A. L., AND Bruzgs, C. T. 1905 The cotton bollworm. U.S. Dept. Agriculture, Bureau of Entomology, Bull. No. 50, 155 pp. ScHarEFFER, A. A. 1911 Habit formation in frogs. Jour. Animal Behavior, vol. 1, pp. 309-335. SHELDON, R. E. 1911 The sense of smell in selachians. Jour. Exp. Zool., vol. 10, pp. 51-62. StonakeR, J. R. 1900 Some observations on the daily habits of the toad (Bufo lentiginosus). Indiana Acad. Sci., pp. 167-170. Vergss, E. 1903 Ueber die Reizung des Riechorgans durch directe Einwirkung riechender Fliissigkeiten. Arch. ges. Physiol., Bd. 95, pp. 363-408. SUBJECT AND AUTHOR INDEX LLEE, W. C. Certain relations between rheotaxis and resistance to potassium cyanide in Isopoda.................---- 397 Amphibian embryos. Experimental evidence concerning the determination of posture of the membranous labyrinth in............ 149 Amphibian larvae to light. The reactions of normal and eyeless................ 195 Amphibians. Olfactory reactions in......... 617 Anesthetics. Antagonism between salts and IV. Inactivation of hypertonic sea-water [ont ar Cele alc, (6c een aie a Oy 4 ea ee 591 Antagonism between salts and anesthetics. IV. Inactivation of hypertonic sea-water yesnesthotics! ome Ne fcc reece sai: 591 Arbacia. Studies of fertilization. VI. The mechanism of fertilization in..... 523 Asplanchna amphora. Transmission ‘through the resting egg of experimentally induced HEACLELS IT cas octane erence cians Desc 347 AITSELL, Groras Atrrep. Experi- ments on the reproduction of the hypo- trichous Infusoria. II. A study of the so-called life cycle in Oxyticha fallax and Pleurotricha lanceolata. ...........:....- 211 Bee. The olfactory sense of the honey....... 265 Birds. A further study of size inheritance in ducks, with observations on the sex ratio CAT (ONG Spats Le RR Pee ye hg oa eee 131 Borine, AvicE M. and PEARL, RAYMOND. The odd chromosome in the spermato- genesis of the domestic chicken...... SaOee 53 Bursaria to food. The relation of I. Selec- tion in feeding and in extrusion.......... 1 HARACTERS in Asplanchna am phora. Transmission through the resting egg of _ experimentally induced ................. 347 Chicken. The odd chromosome in the sper- matogensis of the domestic............... 53 CuiLp, C. M. Studies on the dynamics of morphogenesis and inheritance in experi- mental reproduction. VII. The stimula- tion of pieces by section in Planaria doro- ECT a Ee ee ares so ote oer oer 413 Chromatin in hybrids between Fundulus and Ctenolabrus. The behavior of the........ 501 Chromosome in the spermatogenesis of the domestic chicken. The odd Conjugating and non-conjugating races of Paramaecium. On so-called.............. 237 Ctenolabrus. The behavior of the chromatin in hybrids between Fundulus and........ 501 Currents produced by sponges. On the strength and the volume of the water.... 443 Cyanide in Isopoda. Certain relations be- tween rheotaxis and resistance to potas- SGanRB Rp An ond Sccbo tor sacoc DOD OCR eee 397 UCKS, with observations on the sex ratio of hybrid birds. A further study of SIZE WHETIGANCE LO) 525-\5 locks 6 eee ss «3's 131 ES of experimentally induced characters in Asplanchna amphora. Transmission through the resting..............--...-- 347 653 Embryos. Experimental evidence concerning the determination of posture of the mem- branous labyrinth in amphibian......... 149 Extrusion. The relations of Bursaria to food. I. Selection in feeding andin............. 1 EEDING and in extrusion. The rela- tions of Busaria to food. I. Selectionin. 1 Fertilization of Nereis. The effect of radium TACIAGIONS | ONTENO sar h-jcntarsuaee ire ei reer 85 Fertilization. Studies of VI. The mechan- ism of fertilization in Arbacia............ 523 Fundulus and Ctenolabrus. The behavior of the chromatin in hybrids between....... 501 EMAL nodes. The supposed experi- mental production of hemolymph nodes and _ accessory spleens. Studiestome te oc ence ses ee epee 241 Hemolymph nodes and accessory spleens. The supposed experimental production of V. Studies on hemal nodes.............-. 241 Honey bee. The olfactory sense of the...... 265 Hybrid birds. A further study of size inherit- ance in ducks, with observations on the SOX THLIO OL Renate saan. Crooner telat 131 Hybrids between Fundulus and Ctenolabrus. The behavior of the chromatin in........ 501 Hybrids. Modes of inheritance in teleost.... 447 Hypotrichous Infusoria. Experiments on the reproduction of the II. A study of the so-called life cycle in Oxyticha fallax and Pleurotricha lanceolata...............+.+- NFUSORIA. Experiments on the repro- duction of the hypotrichous II. A study of the so-called lifes cycle in Oxyticha fallax and Pleurotricha lanceolata........ 211 ORISUZON Ose soe oa. als tees eteorens eae hetero eters 131 Inheritance in experimental reproduction. Studies on the dynamics of morphogenesis and VII. The stimulation of pieces by section in Planaria dorotocephala........ 413 Inheritance in teleost hybrids. Modes of..... 447 Inheritance. Multiple factors in Mendelian.. 177 Isopoda. Certain relations between rheotaxis and resistance to potassium cyanide in... 397 | tae RINTHinamphibianembryos. Ex- perimental evidence concerning the de- termination of posture of the mem- PLranous: 22) ..c cere oe ets . 149 Larvae to light. The reactions of normal and eyelessianip hi biaie sep eee eae tetera 195 Laurens, Henry. The reactions of normal and eyeless amphibian larvae to light.... 195 Light. The reactions of normal and eyeless amphibian larvae t0...-..--- 0.00. +e 195 Linuiz, FRANK R. Studies of fertilization. VI. The mechanism of fertilization in VAT ACTA. coo elle Dteetine eae be rebsoteetovalel=s= 523 Linuie, RatpH 8. Antagonism between salts and anesthetics. IV. Inactivation of hy- pertonic sea-water by anesthetics........ 591 Lunp, E.J. The relations of Bursaria to food. I. Selection in feeding and in extrusion... 654 ACDOWELL, E. C. Multiple ? factors in Mendelian inheritance.. PS 177 McInpoo, N. E. The olfactory. sense 5 Of Be i ONY DOs. 5 ck eee a ieee eye eiae elaine 265 Membranous labyrinth in amphibian embryos. Experimental evidence concerning the de- termination of posture of the............. 149 Mendelian inheritance. Multiple factors in... 177 Meyer, ARTHUR WrLLIAM. The supposed experimental production of hemolymph . nodes and accessory spleens. V. Studies on! hemalimodessecreeretaice cosas scale 241 MircHELL, CuaupE W., and Powers, J. H. Transmission through the resting egg of experimentally induced characters in PAF} hal ouoey evant 0) eos cz) 5 og agee Rano dec 347 Morphogenesis and inheritance in experi- mental reproduction. Studies on the dynamics of VII. The stimulation of pieces by section in Planaria dorotocephala 413 Morris, MArGaret. The behavior of the chromatin in hybrids between Fundulus and: Crenolabnustenes.ahs cs coh ess ees 501 EREIS. The effect of radium radiations on’ the fertilization (Of. 0 550...6.6.050 5. 85 Newman, H. H. Modes of inheritance in teleost hybrids) .n aos seecy soa cee 447 Nodes and accessory spleens. The supposed experimental production of hemolymph V. Studies on hemal nodes............... 241 Nodes. The supposed experimental produc- tion of hemolymph nodes and accessory spleens. V. Studies on hemal........... 241 @ lesincune reactions in amphibians.... 617 Olfactory sense of the honey bee. The...... 265 Oxyticha fallax and Pleurotricha lanceolata. Experiments on the reproduction of the hypotrichous Infusoria. II. A study of the so-called life cycle in................. 211 ACKARD, Caries. The effect of radium radiations on the fertilization of Nereis...... Sach Gnunsineis soascee manos clon 85 Paramaecium. On so-called conjugating and non-conjugating races of................. 237 ParKER, G. H. On the strength and the volume of the water currents produced by BPONLES Hs na areite ere reat nateieys sive aieivie able 443 PEARL, Raymonp, Borine, AuicE M., and The odd chromosome in the spermato- genesis of the domestic chicken............ 53 PurILuips, JouN C. A further study of size in- heritance in ducks, with observations on the sex ratio of hybrid birds.............. 131 Planaria dorotocephala. Studies on the dy- namics of morphogenesis and inheritance in experimental reproduction. VII. The stimulation of pieces by sectionin........ 413 INDEX Pleurotricha lanceolata. Experiments on the reproduction of the hypotrichous Infu- soria. II. A study of the so-called life cycle in Oxyticha fallaxiand | evens. sen 211 Potassium cyanide in Isopoda. Certain rela- tions between rheotaxis and resistance to. 397 Powers, J. H., MitcHeELL, CLAupDE W. and Transmission through the resting egg of experimentally induced characters in Asplanchnajamphora.: .. c0.2-ste snes 347 ADIATIONS on the fertilization of Nereis. The effect of radium.......... Radium radiations on the fertilization of Nereis. he effect. Of ).)..2 sas nce meee 85 Ratio of hybrid birds. A further study of size inheritance in ducks, with observa- tions!on! the: Bex... «'. ...4. us aaceseeee eee eee 131 Reactions in amphibians. Olfactory......... 617 Reactions of normal and eyeless amphibian larvae*to light: “Che .s\..2, ccm smile 195 Reproduction. Studies on the dynamics of morphogenesis and inheritance in experi- mental VII. The stimulation of pieces by section in Planaria dorotocephala..... 413 Resistance to potassium cyanide in Isopoda. Certain relations between rheotaxis and.. 397 Rheotaxis and resistance to potassium cyanide in Isopoda. Certain relations between... 397 * Risser, JONATHAN. Olfactory reactions in amphibians, «5 0-!iss0% 6 Sekt salsrale ee eee ee and anesthetics. Antagonism be- tween IV. Inactivation of hypertonic sea-water by anesthetics................. Sense of the honey bee. The olfactory...... 265 Sex ratio of hybrid birds. A further study of sex inheritance in ducks, with observa- tions: OD. the: « .ij.ci stele orale erent 131 Size inheritance in ducks, with observations on the oer ratio of hybrid birds. A further Btudy Of. . ..: os doeceoclvcle oc ceen ee eee 131 Spermatogenesis of the domestic chicken. The odd chromosome in the.............. Spleens. The supposed experimental produc- tion of hemolymph nodes and accessory. V. Studies on hemal nodes............... 241 Sponges. On the strength and the volume of the water currents produced by....... 443 STREETER, GEORGE L. Experimental evi- . dence concerning the determination of posture of the membranous labyrinth in amphibian embryOs. sp seseeeere eee sent 149 ELEOST hybrids. Modes of inheritance fe OODRUFF, LoranpE Loss. On so- called conjugating and non-conjugat- ing races of Paramaecium............ 237 i iy Ny, ie / ‘ : : - oA F : etme oF iS etn an QL The Journal of experimental 1 zoology J68 v.16 COped Biological & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY ae He a be y a cite f Hage i this aie t Pe aint eee # inte ii aaah asi ay ean FA es 6 aU ath nist CF DO OUT seh y etal at hh with it} ee Bie re a ee riNd iad fein id oy fet aah ‘ ae vehnap g: uy eae bal rete Natt a aie ae ao faviybed beds, Of eas raraak wath sii heed ons a ii it bs Abs vie ni he i if Pa fen pa is ba ate + ; ny « : / OY tae ty Pentan deri Gas fs : ‘ Beg Veg ates rs 4 TAN i re end pe ripe eau be dy Pedestal AS Re ii he 5 en Rea VAIS mane Sphe eB RY? * iy + Pinata satay Ue Toatte PIE Ladi tat) ds cm hs an ia a , i igtaet SUS ke aN ns Me ne ‘ Sp eT