Please handle with _ EXTREME CARE This volume is Brittle — - and CANNOT be repaired! * Photocapy only if necessary Return to library staff, do not in bookdrop ~ _ “GERSTEIN SCIENCE INFORMATION CENTRE Library staff, please retie with black ribbon and reshelve LINIV. OF ToRONTO LIBRARY ~ Digitized by the Internet Archive in 2009 with funding from University of Toronto http://www.archive.org/details/journalofexperim0S5broo ¥,) ay Oa ie wi tak JOURNAL OF EXPERIMENTAL ZOOLOGY EDITED BY WILLIAM K. BROOKS Johns Hopkins University WILLIAM E. CASTLE Harvard University EDWIN G. CONKLIN University of Pennsylvania CHARLES B. DAVENPORT Carnegie Institution ROSS G. HARRISON Yale University HERBERT S. JENNINGS Johns Hopkins University FRANK R. LILLIE University of Chicago JACQUES LOEB University of California THOMAS H. MORGAN Columbia University GEORGE H. PARKER Harvard University CHARLES O. WHITMAN University of Chicago EDMUND B. WILSON Columbia University ROSS G. HARRISON, Managing Editor 2 HILLHOUSE AVENUE, NEW HAVEN, CONN. VOLUME V PUBLISHED QUARTERLY BY THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY 36th STREET AND WOODLAND AVENUE PHILADELPHIA, PA, CONTENTS No. 1—November, 1907 Davin Day WHITNEY Determination of Sex in Hydatina senta ArTHUR B. Lams A New Explanation of the Mechanics of Mitosis. With Two Figures... . HERBERT EUGENE WALTER The Reactions of Planarians to Light. With Fourteen Figures... . No. 2—December, 1907 HERBERT EUGENE WALTER The Reactions of Planarians to Light. With Fourteen Figures (con- chided) pete Sy lsterie ce wo eta MEMES ora CR eae eee Mary ISABELLE STEELE Regeneration in Compound Eyes of Crustacea. With Sixteen Plates and De Mane WOR endian mace edcae bdeauoheb hor sorgoesse onan H. V. WiLson On Some Phenomena of Coalescence and Regeneration in Sponges. ANS NI SOME MN ateS a hace boo manbeu ss sco un bob aw ou oaU pmaOal dick Hans PrzIpRAM Equilibrium of Animal Form. With) Templo reson rerescaheteeretieeeteha (ter CHARLES ZELENY The Effect of Degree of Injury, Successive Injury and Functional Activity upon Regeneration in the Scyphomedusan, Cassiopea xamachana. \WitchvBourphiipuresssteilt qesicteece tty sa te cletmetstteret teeters Oot acy ett ALEXANDER PETRUNKEVITCH Studies in Adaptation. I. The Sense of Sight in Spiders. With Six INANE, Foc pone ee Ree ah: SE Sk 245 259 No. 3—March, 1908 Gruman A. Drew The Physiology of the Nervous System of the Razor-Shell Clam (Ensis directus, Cons)) sWathi@neiPlate see cen: enecace.o = ote rer 311 FLORENCE PEEBLES The Influence of Grafting on the Polarity of Tubularia. With Twenty- Chay writs choeoss Guid bod O An ou 6 DMbore iD ROME ad oso. vac 32 N. M. STEvENS A Study of the Germ Cells of Certain Diptera, with Reference to the Heterochromosomes and the Phenomena of Synapsis. With Four Pl atesi hire cei uakeee iS cphane ence BRS eea: ADE he kee Sus 9, ed Sie ee Re ere 359 Ratpu S. LILiie Momentary Elevation of Temperature as a Means of Producing Artificial Parthenogenesis in Starfish Eggs and the Condition of its Action....... 375 Tuos. H. Monrcomery, JR. The Sex Ratio and Cocooning Habit of an Aranead and the Genesis of Sex Rationse With aewoll i punes mec canioe aii «6 Meee a Ae 429 No. 4—June, 1908 N. M. STEVENS The Chromosomes in Diabrotica vittata, Diabrotica soror and Diabrotica 12-punctata. A Contribution to the Literature on Heterochromosomes and Sex Determination. With Three Plates...................-. 453 Victor. E. EMMEL The Experimenal Control of Asymmetry at Different Stages in the Devel- opmentofithe lobsters.) ome eee Erte one 471 C. M. CuiLtp Physiological Basis of Form-Regulation. With One Figure .......... 485 H. H. Newman The Process of Heredity as Exhibited by the Development of Fundulus Hybrids. With Five Plates and Sixteen Figures in the Text ..... 503 C. C. GUTHRIE Further Results of Transplantation of Ovaries in Chickens. With Three Bigures acct raer siete rapt esi cterers reels, ee ieiateaer eer eS eee 563 H. S. JENNINGS Heredity, Variation and Evolution in Protozoa. With Twenty-two Figures SiS ce hectare ease eerieiee tee ann. Oe eee 577 DETERMINATION OF SEX IN HYDATINA SENTA BY DAVID DAY WHITNEY Mie oIntro duction: cram c yet eeieveieree vein etctetede ol 1eCotelsfoisislatavers:= nvessioLole/eale/ouclerarearfourtaletere: tae e/e}selel I TE Material and methods: «<..2 2.0 cccceec cence creases deeceneensce cee eeseeuseecteee 3) III Influence of temperature 4 1 Maupas’ experiments......-.00-es cece cece ete cee e rere eee t eee sence r een ers ces 4 2 Author’s experiments.........6 02-6 e secre cere eee e teen eee ener teen e eee ees 5 a Temperature 20° to 22°C. .... 1... e cece eee eee eet e eee t tn t eens 5 b Temperature 25° to 29°C... 1. eee eee eee eee tee teen ete e ene ees 8 c Temperature 14° to 15°C... 2. eee eee eee teen eter e teen eens 9 IV The relative number of eggs which a male-laying female and a female-laying female produc: 10 I Temperature 20° to 22°C... 21. cece cece tenet eet eee e cent ence eens II 2 Temperature 24° to 29° C.... 2.222 see eee erent ee tence ete cnt e erect ees II V_ Early production of male-laying females in a family of daughter-females...............--. 13 Villm lin men celo Mt Oo samc isrovarcitereisveleilalalelsvol-porelaye ofosshaxeseps¥elcVerebesepsltaretceseleieds laidtsiel-(o[olaloriaretiere 15 1 Temperature 20° to 22°C... 2... ee eee ee ee eee ete teen entree cent es eees 16 2 Temperature 14° to 15° C.... 6.00 eee eee eee cent eee eee ee tenet ee eee ees 18 3 Temperature 25° to 26° C.... 1... cece eee eter e eect eee erence tees 18 aVI Male and female'strains........0..20.0ccecccs cer ecce cece cece etsccestesteneceerersce 19 VIII Production of fertilized eggs... ... 0... ce cece cece ec eee cere eet t eee eet e teen eee ene 23 YR S we ary ate ea Yay set saket ole vole aisieke orelalsatcla = ehtetaret=te)sfoin in) tove:wielalete\oL-setehegesonel se eisieieielape=;nheyaks 25 I INTRODUCTION On account of the supposed influence of external factors in determining sex in Hydatina senta, this rotifer has attracted much interest in recent years. As is well known Hydatina produces three kinds of eggs, viz: (1) parthenogenetic eggs which develop into females; (2) smaller parthenogenetic eggs which develop into males; and (3) fertilized eggs which develop into females. Each female produces only one of these three kinds of eggs. ‘Thus three types of females may be distinguished, viz: (1) females which produce females parthenogenetically, or female-laying females, 2 2; (2) females which produce males parthenogeneti- cally, or male-laying females, @ 9; and (3) the sexual females that lay fertilized eggs. Tur Journat or ExreriMENTAL ZOOLOGY, VOL. V, NO. f. 2 David Day Whitney Both female-laying females and male-laying females can be impregnated by males, but on the former, impregnation is sup- posed to have no effect. If the male-laying females are impreg- nated by the male in the first few hours after they leave the egg, such females produce fertilized eggs instead of parthenogenetic male eggs, thus showing that male-laying females can develop into sexual females that lay fertilized eggs. The female-laying female can produce a family of daughter- females, some of which may lay female eggs and others may lay male eggs. With the view of finding out the ratio in which these two classes of daughter-females are produced under various conditions I have carried out the experiments to be described. Maupas found that a temperature of 26° to 28° C. would pro- duce as high as 95 per cent of male-laying females while a tem- perature of about 14°C. would produce as low as 5 per cent of male-laying females. Nussbaum, on the contrary, came to the conclusion that nutri- tion and not temperature is the sex controlling factor. He found that by starving the young females for the first few hours after they emerge from the egg they would produce a high percentage of males, but if they were fed at the time they leave the egg they pro- duce a high percentage of females. Punnett has carried out a few experiments along the lines laid down by Maupas and Nussbaum and finds that neither tempera- ture nor nutrition is influential in determining the sex. He finds, on the contrary, that there are definite “‘sex strains.”’ Some strains produce 40 to 50 per cent of males, others produce a very low percentage, 2 to 5 per cent, while others produce no males at all, although reared through as many as seventy-two generations. The greater part of the work of the present paper was planned and begun in the spring of 1906, under the direction of Prof. T. H. Morgan, before the results of Punnett were published. Not knowing how to obtain proper food cultures the rotifers all died in July and the continuation of the experiments was deferred until October, 1906. Determination of Sex in Hydatina senta 3 Il MATERIAL AND METHODS In the latter part of April, 1906, Hydatina senta was discovered in great numbers in a small pool on the Palisades of New Jersey near Grantwood. The pool was fed by a little stream or ditch which carried away the drainage from several cottages. “The ditch was an extremely favorable place for the growth of Euglena viridis which collected in large patches on the sides and bottom. Immense numbers of Euglena floated down into the pool at the end of the ditch and served as food for the rotifers which abounded there in countless thousands. Sometimes as many as 150 to 250 individuals could be drawn up by a pipette in a few cc. of water. About May 15 the pool dried up completely. The ditch still contained water but no rotifers were found in it after May 20. At this time there were innumerable larve of insects in the ditch and perhaps they exterminated the rotifers by feeding upon them. In all experiments each individual female was isolated in a square or round watch glass which contained about 5 cc. of water and fed with Euglena, other protozoa and bacteria. In order to obtain the Euglena and other protozoa a culture of horse manure and water (one to two ounces to a quart) was made, inoculated with Euglena and allowed to stand for two to three weeks at room temperature. At the end of this time the green coating of alg, Euglena, etc., could be removed from the sides of the glass jar and served as an excellent food for the rotifers. Great care was taken to keep these food cultures uncontami- nated by rotifers. All watch glasses were placed in hot water after each experiment in order to destroy all eggs which adhered to the sides, thus preventing contamination of the following experi- ments by eggs of the preceding ones. The experiments at temperature of 24° to 29° C. were con- ducted in an incubator. ‘Those at a temperature of 20° to 22° C. were conducted on the laboratory tables at room temperature, while those at a temperature of 14° to 15° C. were carried on in an ice chest. These rotifers are exceedingly hardy and can be very easily kept in the laboratory throughout the year. In Mayof 1906 a Euglena culture was prepared in a glass jar containing 2000 cc. of water 4 David Day Whitney and a few rotifers put into it. The jar was covered so as to pre- vent evaporation of the water. Rotifers have lived in it to this time, April, 1907, although no more food material has ever been added. It is absolutely necessary that the surface of the water be free from a scum for the rotifers will die within a few hours if it is present. It is safer, in order to keep the surface free, to tie the horse manure in a muslin cloth and place it in a well covered jar nearly filled with water. III INFLUENCE OF TEMPERATURE I Maupas’ Ex periments The experiments of Maupas were so briefly described that it is very difhcult to understand clearly how he obtained his results. Nussbaum and Punnett are inclined to believe he determined that a female was a male-laying or a female-laying individual by the size of the eggs that she produced. Small eggs being assumed always to give rise to males while larger eggs give rise to females. Nussbaum has measured a series of both male and female eggs and found that in some instances the two kinds of eggs over-lap in size. ‘Thus he points out an error through which Maupas’ results might have been obtained. Isolating and counting the eggs of this rotifer would be exceed- ingly tedious and require almost constant attention. As the sexes can be readily distinguished at any period and as it requires only 36 to 48 hours for a female to mature and produce eggs it seems to me extremely probable that Maupas must have allowed some at least of the eggs to hatch before recording his results. As his experiments are so few and briefly described it may be well to present them here in order that they may be compared with and interpreted by my own. Experiment I. Lot A, tempera- ture 26° to 28°C. Five female-laying female sisters produced 104 eggs; 97 per cent developed into male-laying females. Lot B, temperature 14°C. Five other female-laying females, which were sisters of lot A, produced 260 eggs; 5 per cent developed into male-laying females. Determination of Sex in Hydatina senta 5 Experiment II, temperature 14° C. Five female-laying females, kept from the time of hatching at this temperature, produced 110 eges; 24 per cent developed into male-laying females. The same five female-laying females were then placed at a tem- perature of 26° to 28° C. and produced 81 per cent male-laying females. Experiment III, temperature 14° C. Six female-laying females which had been kept at this temperature from the time of hatch- ing produced 34 eggs, of which 12 per cent developed into male- laying females. The same six female-laying females were then placed at a tem- perature of 26° to 28° C. and allowed to produce 44 eggs, of which 95 per cent developed into male-laying females. These six females were alternately placed at 14° C. and 28° C. several times and always gave a high percentage of male-laying females at the higher temperature. 2 Author's Experiments a ‘Temperature 20° to 22°C. Experiment I, October 24, 1906. A female-laying female was isolated from a jar which was stocked with rotifers collected Octo- ber 2, from the same pool in which the animals were found in the preceding spring. This strain was carried through twelve generations and the percentage of male-laying females determined. Each female was supplied with an abundance of food from the time of hatching, isolated in a separate watch-glass, and kept upon the laboratory table at room-temperature. Table I gives the ratio of the mother individuals producing male and female offspring in the 3264 daughter-females of 95 female-laying females in the twelve generations. This experiment was made in order to obtain the percentage of male-laying females produced at room temperature of 20° to 22° C., in order to be able to have some standard percentage of male-laying females with which to compare the results of the experiments conducted at lower and higher temperatures. 6 David Day Whitney TABLE I Record of the production of male-laying and female-laying females among the 3264 daughter- females of 95 female-laying mothers. Temperature 20° to 22° C. No. | Eggs | Offspring | Per | | No. Eges | Offspring | Per Gen. | 22 ale [> ae || cent -||Gen--) 299 oc | po ercireeceneenn ECCOE | mother fe | oe yc Wh ears: mother laid (sional) Louse eine) SS | eta | Se ape = see | | | | I I ar | ets Vere) | 48a 3 32 8 | 24 25 | | 4 15 3 12 20 I I 28. || 0: |) 28 ° | Ce Ne 223 Sie ess 18+ } 2 | 53 | 8 | 45 | ase | 6 48: 9) rte! Rage eae | 3 27) hn: 23.«| «14+ | 7 25 o | 25 ro) | | | 8.938 | 15) 23 tao It I 440 cay | a7 38+ | 9 17 La be 23+ | 2 | 20.0) ara 325) 20h 10 48 15 3300 Stats | | | II 48 | 16 33+ IV Te al ees I so | i+ | 12 43 4 39 | 9+ 2 | 47 15 32 | Bit 13 44 | 10 34. |) 722-5 30 N52 elo Mh iea6 tao 14 49 | +14 35. | 28+ lisa asoonie 71/8 28 ulezo | Sx¢/ 1 Wasa ees 27 40 | 16 16 | 6 10 37+ My I 47 Our 47 ot 17 45 | 4 41 8+ 2 44 2 42 Piaoe Ml 18 50 7 43 14 3 41 40 2+ | 19 | cpa il) A 28 12+ 4 45 ° 45 | | 20 21) |" 0 21 ° 21) 0) 2) See 17 19+ NAD pag ae Mie 12 20 27+ 2220 2 18 10 i) ee 45 4 41 8+ 230+) «925 ° 25 ° Saas o | 36 | © | | eae illsa7 | go. | 14+ ar 49 8 40 16+ 5 | 48 22 26 «| 45+ } | | | XII I 38 3 35 7+ VIE]; 1 30 15 15 | ‘50 2 45 9 36 | 20 | 2 35 I 34 2+ 3 48 4 44 8+ 3 | 27 ° 27 ° 4 31 5 26 16+ 4 24 ° 24 ° 5 shee (aie) 26 | 33+ 5 19 5 14 | 26+ | 6 31 10 21 32+ | latZ 54 To 440) 18+ Vill I 41 16 25 39+ es 45 6 39 13+ Panes 17 I 16 | s+ | 9 35 II 24 31+ 3 47 18 29 38+ | 10 43 5 B85 u) Erte 4 26 ° ye | xr 38 3 35 iia 5 38 | ° 38 ° 12 27 9 18 33+ 13 42 16 | 26 |- 38+ Ix I 24 | I 23 4+ 14 34 ei Og Bae 15 83 15 | 28 | 34+ x I 49 | 13 36 0 | 26+ | 16 37 6 | 31 16 axe 45 10 35 224+ | | 47 31 26 16+ Determination of Sex in Hydatina senta TABLE I—Continued Temperature 20° to 22° C. Now Eeeeet| a Obspaugs |) be Noo ees ||) Otspung, || Se Gen. Se a TERE. || |i ea | ass Gen. 2h) heat Ise — mother | au oon] Lo) {0} ae mother Gs ae? 29 XII 18 26 7 19 | 22+ || XII | 30 16 5 II 19 33s S| 3r 9 oe A 20 4 | 3 41 6+ 32 18 8 | 10 21 20 ° 20 Qi | 33 20 6 14 Bam Aleetsc i ltee 36 | 20 | 34 47 14 33 | | } 31 ut ZO E350) 35 x9) 4 15 24 28 10 i | 35+ 36 26 8 18 25 15 3 2 [20 | 37 II ° II 26 34 I 33 2+ 38 37 II 26 27 15 ° 152) Ce sl 3934 19 15 28 14 3 II 20 40 25 6 19 29 6 I 5 | 16+ | 41 35 4 31 | \ 1 cent 31+ Dict 44+ 3° The nature of the sex-producing power of the daughter-females of each individual mother is given separately in order to show that the ratio between the daughter-females producing male and female offspring varies with different mother-individuals, and also varies as much in the daughter females of sister-mother-individuals. Summary of each generation in Table I and also the final summary of all the generations taken 29 mother TABLE II together. Temperature 20° to 21° C. Offspri Per Werceonl EE eesa | les aR: Sb \inoeher laid 29 29 Sad Gen. Tye 31 15 16 | 48+ | VIZ | I 3 108 12 96 i+ | VOI Ii 2 go 31 59 34+ ee x IV 4 185 39 146 21+ XI Vv 4 177 3 174 I+ xi VAS WG 208 45 163 21+ Eggs Offspring Per aid Ee | ae cent Sa a? _ | a al 135 21) 114 | 15+ | | 169 35 | 134 | 20+ 24 I |} 23 4+ 819 176 643 -| 21+ 49 8 41 16+ 1269 268 | roor | 21+ 3264 654 | 2610 | 204 8 David Day Whitney Table II gives the summary of each generation and the final summary of the twelve generations. In this experiment 95 mother-individuals produced 3264 daugh- ter-females of which 20+ per cent were male-laying females. It will also be noted that the percentage of male-laying females varied in the different generations from I + per cent to 48+ per cent regardless of the number of isolations in each generation. b Temperature 25° to 29° C. Experiment II, October 26. Two female-laying females were TABLE III Record of the production of male-laying and female-laying females among the 208 daughter-females of 26 female-laying mothers. Temperature 25° to 26° C. Temperature 26° to 29° C. | No. Leges| Ofspring| Per No. Eggs Offspring | Per Gen.| 99 | ale __| cent Gen.| 29 laid | — | cent | mother | ] 82/29} oe | mother J 2)9 9 | oS? | | | | | | | StrainIV) I) 1 | 17| o| 17] © Strain TT IX | I | Werey [Naz | ° | | 2 || go" | ral ° | I I (Ts ae Ly | we (or | | | | | | xX 1 |e 31, ONS StrainIII| I I 17 corals 4 | o | 2 Sa sol) 5 ° | {<3} 1h Srh3|) an Fo) IE100 Ul I 13} 4] 9] jot+ AL || |b 48 0] Siena aes ee 3 29 | .25 | 29 No. of records averaged. . 17 | 15 | 12 | 7| 5 2 Rate in mm. per second..) .63 | .625) .565) .55 | -53 | -55 | -375] -39 | -39 mm. per sec. at 39 c.m. intensity, which makes a range of 0.18 mm. per sec. When the same data are rearranged to show the average rate for each individual for all intensities, as in Table V, the extremes are 0.49 mm. per sec. and 0.83 mm. per sec. with a range of 0.34 mm. per sec. In fact the individual behavior of these ten worms, despite their apparent similarity, was sufficiently distinct to allow each one to be thereby identified. Turning. Attention has already been called to the fact that there is less turning in light of various intensities than in the dark. A return to Table I will make plain that there fails to be any TABLE V Average rate of locomotion for each of ten worms (Planaria gonocephala) based on trials with non-directive light of various intensities Identification number of worm........ | I 2 3 4 5 6 7 8 9 10 | | Average ‘rate in eight intensities al | | | pressed in mm. persec.............|0.79 |0-57 0.68 (0.64 0.83 |0.70 |o.72 0.58 [0-49 |0.62 ! = ae Seas definite correlation between the degree of intensity of the light and the amount of turning, although the least turning occurs under the highest intensity. This latter point, however, rests upon a very slight difference and may not be significant. It 1s neverthe- less worth mentioning, since it is in line with the effect of the os The Reactions of Planarians to Light 59 highest intensity upon rate, as well as with certain other evidence to be discussed later. The small excess of clockwise over contra-clockwise turnings is not explainable upon the ground of varying intensities of light. A distribution of the cases under the several intensities of light (Table VI) makes it plain that this peculiarity is due rather to individual causes than to light intensities. Indeed it would be difhcult to conceive theoretically how varying intensities of non- directive light could influence a worm in such a way as to affect the direction in which it turns. ‘The natural expectation accord- ing to chance would be an equal number of turnings in either direction. ‘The excess of clockwise turns seems, therefore, un- doubtedly due to internal causes which render certain worms more liable to go one way than another. In fact, when the records were arranged according to individual behavior it was found that of the ten worms seven averaged a majority of clock- wise turns while only three fell in the contra-clockwise column. TABLE VI Character of turning of Planaria gonocephala in non-directive light of various intensities = 7 = ea =n “Sea se aiihea a = | an . Light in candle meters...............| © | 0.94] II 39 | 78 126 | 155 | 217 | 431 |Total Clockwise turns........... deemuaayie 2g bas. Wier mae | 17) x7) |22 | 24 | 22 | 203 Contra-clockwise turns............--| 23 23) ||) 37; 20 | 17 14 | 166 18 17 | 17 occur in the light than in the dark, but fewer “indefinite” changes. This point requires no further exposition as its corollary has already been given. The behavior of the worm in this respect seems to be more closely correlated with the highest intensity (431 c.m.) than with any other. In the highest intensity employed there are indicated (Table II, p. 52) 40 per cent of definite changes, which is con- siderably in excess of the percentage of such changes made in any other intensity. Onthe other hand indefinite, or S-shaped, changes constitute only g per cent of all records taken at the highest intensity, which is less than half the number of indefinite paths made in any other intensity. 60 Herbert Eugene Walter While the extremes of the series of definite changes indicate a general rise in the percentage of their occurrence with an increase of intensity, and while in the same way the extremes of the series of indefinite changes suggest in general a decrease of frequency with the increase of intensity, it can hardly be maintained that the character of the changes in course is definitely correlated in the majority of cases with changes in intensity. Degree of Wandering. Wandering is not closely correlated with the intensities of light. In Table VII, which deals with the percentage of straight paths made by P. gonocephala under dif- ferent intensities of non-directive light, this fact is expressed nega- tively, since it is held that a straight path is a good indication of the absence of aimlessness or wandering and may thus serve as a negative measure of such behavior. TABLE VII Percentage of straight paths made by P. gonocephala in the dark and also in non-directive light of different intensities Light in candle meters... . Hogeoanmnsite o |0.94| 11 | 39 In this respect again the behavior of the worms under the high- est intensity 1s more pronounced than under any other intensity since the greatest number of straight paths were laid at an inten- sity of 431 c.m. Interval of Response. There seems to be some evidence that the interval of time elapsing between the reception of a light stim- ulus on the part of a worm and its consequent response, may be quite considerable. ‘Three facts were established that may sup- port this conclusion. First, when two-minute records were made under various inten- sities, it was found that the worms averaged a faster rate during the second minute of exposure to the light than during the first, in spite of the facts that the mechanical stimulus due to placing the worm in the light machine had a more quickening influence during the first minute and that the fatigue effects were more likely to appear during the second minute. The actual figures The Reactions of Planarians to Light 61 for the above statement, based upon 240 two-minute trials under various intensities, are 0.645 mm. per sec., the average during the first minute, as against 0.713 mm. per sec., the average during the second minute. Secondly, in these 240 trials, the percentage of turning under all intensities 1s greater during the first minute than during the sec- ond, being 87 per cent and 57 per cent, respectively. “This result may possibly be conceived to be due to a greater steadying influ- ence of the light during the second minute than during the first and to a consequent greater turning than during the first minute. But on the other hand a similar decrease of turning, although not so pronounced, took place during the second minute when the worms were in the dark. It must be admitted, therefore, that the fact of less turning during the second minute may have nothing to do with the interval of response. Thirdly, on several occasions a notable piece of behavior was observed, which may have a bearing on the interval of response. The phenomenon in question AGars occurred in connection with a modification of the experimental held within the light machine to be more fully described later. Briefly this modification con- sisted in making a field of two distinct intensities of light, the latter being projected vertically from above in such a way that a sharp line of demarkation formed a boundary between the two areas. Ordinarily when the worms reached this boundary line as they glided from one intensity to another, they responded promptly to the stimulus caused by the change of intensity. Sev- eral times, however, they were observed to travel indifferently exactly along this dividing line for a distance of several centimeters with half the body in one intensity and half in the other. This curious fact lends itself to various interpretations, one of which is that the response to a new intensity may not be, in all cases, immediate. Manner of Coming to Rest. During the experiments made in the non-directive light apparatus previously described, nor- mal worms could never be induced to come to rest in the light. If allowed to remain in the aquarium they would wander about until they reached the shadow under the diaphragm (Fig. 1, H), 62 Herbert Eugene Walter where they finally stopped, usually in the angle formed by the parafhne wall and the bottom. Loeb’s conclusion (’93b, p. 101) that planarians subjected to directive light come to rest in regions of least intensity, seems therefore to be equally true of planarians in non-directive light. Summary. In non-directive light Planaria gonocephala moves faster, turns less and makes more “‘ definite’ but fewer ‘‘indefinite”’ changes than in the dark. Rate of locomotion; amount of turning; changes in the character of the course, as well as the amount of wandering, do not appear to be correlated with varying light inten- sities, unless in the following instance. Under the highest inten- sity employed, namely, 431 c.m., occurred the slowest rate; the least turning; the greatest number of “definite” and the fewest “indefinite”? responses, together with the straightest paths. The excess of clockwise over contra-clockwise turnings throughout the series of intensities is probably not attributable to light. Continuous exposure to light results in a decreasing rate of loco- motion, although in the second minute of movement as compared with the first an increase in the rate of locomotion takes place, while fewer turnings occur. Rate of locomotion is less influenced by differences in light inten- sity than by certain internal factors which go to make up what may be termed the individuality of different worms. Individ- ual worms may sometimes fail to respond for a considerable inter- val of time to light stimuli that ordinarily produce immediate effects. Finally, planarians subjected to non-directive light come to rest in regions of lessened light intensity the same as they do in directive light. “Gi Abrupt Changes in Intensity Abrupt changes in intensity may be of two kinds: either with reference principally to time or to space. First, those changes are abrupt zm time in which light or dark is suddenly thrown upon the worm, and secondly, those changes are abrupt 1m space in which a moving worm passes immediately from an area of one intensity into a sharply defined area of a different intensity. ‘This topic The Reactions of Planarians to Light 63 will be discussed here only in its relation to non-directive light, the effects of sudden changes in directive light coming more properly in a later section. a Abrupt Changes of Light Intensity in Time Whenever worms were left over night in the experimental aqua- rium completely shut off from light, a large proportion of them would be found at rest in the morning when the light in the hood was again turned on. By removing ie diaphragm (Fig. 1, H), under mie edge of which near the penaiine wall the worms were usually enllcceces it was possible without any mechanical disturb- ance to subject resting worms to sudden non-directive light after a prolonged period of complete darkness. This sudden stim- ulus rarely had an instantaneous effect. ‘The interval of response was often several minutes and frequently non-directive light alone proved insufficient to start the worms into activity. No sudden increase of intensity ever proved powerful enough to throw a gliding worm into the more rapid method of crawling. Pearl (’03, p. 551) stated the same fact after subjecting planarians to much stronger intensities of light than were employed in the present experiments. It was found that P. gonocephala showed a decided response— either some change in course or a wigwag motion of the anterior end—more frequently when suddenly subjected to dark than to light. By inserting a key into the electric circuit it was possible to control the light in the hood to a fraction of a second. Worms in complete darkness were by this means subjected to various intervals of sudden light and worms in light to intervals of sudden dark, the results being at once noted. While the worms were in the dark their behavior could not, of course, be directly observed, but by watching them closely just before the light was turned off and also the instant it was turned on again there was no great difficulty in determining whether a response had occurred during the interval. The results obtained from nearly a thousand trials are indicated in Table VIII. It will be seen from this table that there are more responses than failures to respond and that the responses occur more fre- 64 Herbert Eugene Walter quently when the worms are suddenly subjected to dark than to light. It may be further noted that the excess of the responses in the dark over those in the light increases with the interval of exposure, indicating that the worm’s adjustment to a change in the light stimulus affecting it is not in all cases immediate. The effect of previous exposure, whether to several hours of dark or light, is a factor in these results which will be considered more properly later on. TABLE VIII Percentage of the responses of P. gonocephala in various intervals of time when suddenly subjected to dark and to light of 39 c.m. Number of seconds exposed..............-- 5 | 10 15 20 25 30 | Average Percentage of responses in light............ 51 | 59 54 54 48 46 52 Percentage of responses in dark............. 63 66 73 75 71 71 7° Excess of responses in dark....... ROSA OS 12 7 19 21 23 25 18 It should be added that Bdelloura gives a remarkable response when enveloped in sudden darkness. It will frequently forsake its attachment under these circumstances and unattached in the water go through violent contortions. ‘This striking response can be called forth by an exceedingly brief interval of dark, namely, the shortest time required to turn the electric light off and on. Nagel (’94, p. 387) speaks of animals thus affected by sudden shadow as “skioptic.”’ The relation of Bdelloura to light falls into a somewhat different category, however, than that of the fresh-water planarians, since Bdelloura is positive to light, while fresh-water flat-worms are negative. b Abrupt Changes of Light Intensity in Space Several devices were employed to test the behavior of planarians passing abruptly from an area of one intensity of non-directive light into another. ‘The most successful device tried was that in which two lights of different intensities were mounted overhead iy The Reactions of Planarians to Light 65 in the hood of the apparatus already described in Fig. 1, the mingling of their rays being prevented by the insertion of a ver- tical diaphragm (Fig. 3, C), which extended from the region between the lights down to the surface of the aquarium. In order to place the diaphragm in position it was, of course, necessary to remove the heat screen (Fig. 1, C), the presence or absence of which, however, would not have affected the results sought since the water in the aquarium itself was nearly 2 cm. deep and thus Fig. 3. A, stationary light; B, sheet iron walls of hood; C, vertical diaphragm separating the two lights; D, horizontal diaphragm; £, sheet iron roof of hood; F, plate glass aquarium floor; G, paraffine wall of aquarium; HH, diaphragm to shut off reflections from wall of aquarium; J, wall of reflector box; F, open side of box; K, mirror; L, black tunnel; M, black draperies cutting off side light; N, table sup- porting reflector box and end of tunnel; 0, movable light; P, track for movable light; 9, narrow, hori- zontal diaphragm attached at right angles tothe lower side of the diaphragm C, in order to prevent the light rays from the two sources of light, 4 and O, from overlapping. constituted an efhcient heat screen. By keeping the hood sta- tionary and causing one of the lights (Fig. 3, O) to slide up and down at will, it was possible to bring about various contrasts of 66 Herbert Eugene Walter intensity in the field below. The complete plan of the appa- ratus Is given in Fig. 3. The principal variations in the behavior of Dendroccelum and Phagocata upon reaching the critical line separating the areas of two intensities are indicated diagrammatically in Fig. 4. The dotted line represents the boundary separating two areas of different light intensities. The arrows represent the types of paths made by Dendroccelum and Phagocata. For the sake of simplicity the worms are represented as going in one direction; that is, into one of the two contrasting intensities, but the same types of paths resulted as well when the opposite direction was taken. The angles made in crossing the critical line were also more varied than those represented in the diagram. Type 4 represents a passage without re- sponse; B,an angular change of course made at the critical line; C and F, aloop-like return effected after a short excursion into the new intensity, and G,a sharp turning aside, while A indicates a halt at the critical line, as if a barrier had been encountered. Finally D and E represent a temporary pause on the part of the worm accompanied by wigwag move- ments of the anterior end of the body. In the case of D the wigwagging is immediate, but E typifies a case when there occurred in the response an interval of such a nature that the ' significant movements were not made until Fig. 4 the worm had advanced at least its own length into the new area. Of these types all, with the exception of 4, are to be regarded as reactions to differences in intensity encountered. The most questionable are the infrequent types C and F, which may be otherwise explained as arcs in a curving course which might have occurred in a field of uniform intensity. By far the commonest type was D, plainly the least doubtful of the series. The Reactions of Planarians to Light 67 As a result of over 3000 observations on the manner in which the critical line separating the two intensities was passed, three facts become evident. First, responses were considerably more TABLE IX Kind and percentage of responses of Dendrocalum and Phagocata in passing from one intensity of non- directive light to another Turn-backs| No Wigwags and full Loops Angular Total Character of course responses) (Types | stops | (Types courses reece (Type 4) Dand E) | (Types | C and F)| (Type B) G and H) Going into greater intensity, PETCEN Greys eteressfulshessreleat ests 79 II 6 2 2 2I Going into lesser intensity, PEM CODG gararaintatelstetar lererstaroks 5° 36 5 8 I 50 Average responses, per cent.| 64.5 tals 5-5 | 5.0 Aly 35°5 frequent when the worms were passing into the lesser intensity than they were when entering the greater intensity. Secondly, lack of response is more frequent than a visible response of any kind since 64.5 per cent of the crossings made over the critical line were of the type 4. Thirdly, the responses at the critical line were more frequent when the worm was upside down, 1. e.. moving on the surface film, than when it was on the floor of the aquarium. This latter point was illustrated most fully by Phagocata, which, being an active worm, takes quite readily to the surface film, so that it was possible with this species to get a series of observations in which the behavior when crossing the critical line on the bottom of the aquarium could be compared with that when the same line was encountered at the surface film. Table X contains the results of these observations. The doubling of responses when the worm is on the surface film is probably not due to an unequal receptivity of light stimulus by the dorsal and ventral surfaces of the planarian as might at first thought seem possible. As will be shown further on, the worm’s rate of locomotion on the bottom of the aquarium is nearly the same whether the light comes from below or from above, pro- 68 Herbert Eugene Walter vided the amount of light in both cases is equal. Planarians, as Pearl has emphasized, are strongly thigmotactic. Naturally, then, their response to contact is much greater when they are on the glass bottom of the aquarium than when they are suspended on the less resistant surface film. In other words, the less the worm is influenced by the stimulus of contact the freer it is to respond to the stimulus of light. TABLE X Percentage of the responses made by Phagocata at the critical line separating two intensities of non-direc- tive light either on the bottom of the aquarium or on the surface film | Number of No response Response | observations per cent per cent Ounithe surfacedilia!acones chs apeanastosuiss cess 740 45% 544 On: thetbottome. oancrsosen ee eee aneiacg cee ai 1664 76 24 MMotalisiepsannc te oxelateernvctsy se corers/o\sralsinyo: Se co-fiassiecto 2404 60} 39% Finally, a series of experiments was tried in which the contrast between two intensities was varied by raising or lowering one of the lights in the hood. It was found that the responses made by Phagocata under these circumstances increased with the increase in contrast between the two intensities as shown on the bottom line of Table XI, where these contrasting intensities are expressed in a ratio between the constant light taken as unity and the movy- able light. The fact that responses by no means invariably occur when bright light and complete darkness are suddenly substituted for each other (see Table VIII) rendered a further extension of this series unnecessary. The contrasts here used form probably a much greater range of intensity contrasts than the worms ever encounter in nature. Attention to the details presented in Table XI brings to light the fact that, although the number of responses is correlated in a general way with an increase in the contrast between the two illuminated areas, as shown in the bottom line of the table, yet the percentage of the responses is further influenced by the actual degree of the intensities employed. For example, when the two ze fs 7 ot o§ | SE-fF | zh | SLE | S-ob gf sasuodsas yo | . asejuaoiad “ay Lotz rio, Siz 66 ZOz 66 foz 66 goz lof 116 : =| Ss We a oe == | suonearasqo zor | €11 66 ZOz 66 | 101 | zor 66 for | Sor} ztr | €or | zér | 26 | Sar | zex | che | €xx | Ler | oor if jo saquinyy ssdsaz jo'atjog 19 7S ob oS Ly | | eb a4 6S $$$ 9S \ stan 6 5% ISI] apquaoyy +h: g6z of zgz gz biz oz g61 Lg" Lor 69°ztr 61° 9g LS*€9 tL: Lt gi tt If 9 ssdsaa JoaSqg gt | $68 S-1+ | Sz 5A bz S-o1! $*€€| S-cz| S-gz \ aEaionsee WyTty apquaoyyy Lg'9gS | joo of E loz'r1z| Lg: Ltr $L- Lor oS zg SI gg Og°zS |Lg° Lb Lg“ LE if 2 1431) Kapuonvig Ws yuvsuog or: | 1: ap ||) aa ee Wl seb spe | rep || arse sees || sae Ts LOO Gtem | OC lithe |esets Insta |b sag I: ae) WS aquaoyy, 1f-gr 69°21 60°91 | Sb: £1 | Forze Sack) $9°6 | t1'g | LE-9 [es ‘ch-b | O6'£ Sz-€ |€6-z ghz forz g6-1 6-1 zf-1 or -40> Aysuaynt “a . | at} am joonr 1: 61-L1 meL1-Srjr:St—Erin:€r-11| ors 1r-6 |r: 6-L/ rR ES «Il res etter Sect Pasvadut Kyonposd s1 sassuajur asoyr Ma2msag 18P43U0) ayt uayor 14311 aa1oeup-uoU fo sainsuaqut om Surwsvdas aur poosps9 ays yw spor Eobe us pwr0dnyg fo sasuodsas fo aSviuariag IX ATAVL 70 Herbert Eugene Walter areas of light were respectively 33.16 and 68.18 c.m. the ratio was practically the same as when the two intensities were 16.3 and 33-16 c.m., yet the percentage of responses in the two cases is decidedly different, being 10.5 per cent in the former, and 56 per cent in the latter case W hen ve lesser of the two lights was 33.16 c.m. there were invariably fewer responses than aa the lesser light was 16.3 c.m. The latter intensity is undoubtedly nearer the planarian’s optimum intensity, and the apparently inhibitive action of the higher intensities agrees perfectly with certain facts already detailed, as, for instance, that the activities of Planaria gonocephala were less pronounced at 431 c.m. than in lesser inten- sities; and, again, that all planarians show more responses on going into a lesser than when going into a greater intensity. Attempts were made in some other ways to subject planarians to areas of contrasted intensities and, although the results were less satisfactory in general than those obtained by the method of using two buerhead lights of different intensities just described, yet certain facts were ibeatie out that may be worth recording. In the first of these attempts two concentric rings of white paper, each about two centimeters wide and having between them a space of a couple of centimeters, were fastened to the under sur- face of the aquarium floor. The white paper thus arranged reflected the light upward and made areas of gradually increased intensity as compared with the remainder of the aquarium floor through which the light passed without reflection. Worms placed in the center of these circles would consequently be obliged to pass from one intensity of light directly to another, ee er the direction of the radius they might be taking. When worms were actually tested, it was found that they exhibited considerable modification in their movements, particularly when approaching the edge of the paper backgrounds. Owing to the considerable thickness of the plate-glass floor of the aquarium as well as to the fact that white paper is a surface which scatters the light falling upon it, it was found that there was formed, not a sharp line of demarkation between two intensities, but rather a penumbra-like margin of intermediate light. This apparatus was therefore abandoned as unsatisfactory. The Reactions of Planarians to Light 71 The difficulties presented by paper as a reflecting surface largely disappeared by the substitution of a plain mirror in its place, since the surface of a mirror is such that all the light striking it at right angles is reflected at right angles. When, therefore, an unmounted mirror was brought into contact with half of the under surface of the aquarium floor the whole field was thereby divided into two regions sharply separated from each other. Of these one was supplied with light from above only, while the other received the same light plus Aeanl: an equal amount reflected from the mirror below. With the aid of this device an increase of II per cent was gained over the responses obtained when white paper instead of a mirror was used as a reflector. Both Phagocata and Dendro- ccelum were tried by this method. In 76 per cent of the trials made, 7. ¢., in 125 cases out of 165, the worms showed a visible modification in their behavior on reaching the boundary of the two areas of light. It was nevertheless decided that this method was an uncertain test of behavior, since the body of the worm, although fairly translucent, would by no means allow all the light that fell upon it to pass through and be reflected, and consequently the difference of the two intensities to which it was being sub- jected could not be easily estimated. Summary. When sudden light or dark envelops planarians (Dendroccelum, Phagocata and Planaria) the response, if any occurs, is often not immediate. No one of the intensities of light which were employed in these experiments when introduced suddenly was sufficient to make the worms forsake gliding for crawling. Sudden dark calls out more responses than sudden light, while the number of responses increases with an increasing interval of exposure to the stimulus. Bdelloura is decidedly “‘skioptic.”’ Worms encountering the edge of a reflecting area which increases the intensity of the light without introducing any other barrier, show a marked degree of response. ‘The percentage of response is considerably larger when a mirror instead of white paper is used to produce the reflecting surface. If worms are allowed to pass from one intensity to another sharply separated from it, their responses are more frequent upon passing into the lesser intensity 72 Herbeit Eugene Walter than when going into the greater. The average number of fail- ures to respond to these contrasts of intensity reaches about two out of three. Phagocata, at the critical line separating two contrasting inten- sities, responds oftener when on the surface film than when glid- ing over the bottom of the aquarium. The number of responses increases with the increase in the con- trast between the two intensities employ ed, but the percentage of response is greater, regardless of ratio, when one of the lights is of low intensity (13.6 c.m.) than when both are of higher intensity (33 + c.m.) 2 PHOTOTAXIS The term “phototaxis” was introduced by Strasburger (’78) in a study of certain swarm-spores, to indicate movements which were parallel with incident light rays. The term has since been extended by several authors to include similar movements on the part of animals. Any organism is said to be positively phototac- tic when it moves toward the source of light in the direction of the rays and negatively phototactic when it goes in the opposite direc- tion. The purpose of this section is to consider the phototactic move- ments of planarians, as distinct from their photokinetic behavior, (A) when the light remains constant, (B) when the light is changed either (a) in intensity or (b) in direction, and (C) when phototaxis is combined with responses of a different kind. A In Constant Directive Light Orientation. With the exception of Bdelloura all the planarians studied are, under normal conditions, negatively phototactic so far as their first movements in directive light are concerned. To obtain quantitative data for this statement it was necessary to construct an apparatus in which the worms to be tested could be placed quickly and with as little mechanical disturbance as possible in the center of a unit circle with the long axis at right angles to the direction of incident light. The circle was marked off into degrees so that by noting the place at which a worm made its exit a quan- The Reactions of Planarians to Light 73 titative measure of the amount of turning toward or away from the source of the light under the given conditions was obtained. The apparatus finally utilized for this experiment was based upon a device employed by Parker and Burnett (’oo) in testing the relative behavior of normal and eyeless planarians when sub- jected to directive light. Its arrangement is shown in Fig. 5. On the top of a table (4) in the dark room was placed a rectan- gular aquarium (BCDE), the ends of which (BE and CD) were Fig. 5 A, Top of table; BCDE, rectangular aquarium; BE, glass end; FG, round swinging aquarium; H, copper wire attached to ceiling and supporting the swinging aquarium FG; J, movable light; 7, diaphragm; K, surface of water in outer aquarium; L, surface of water in inner aquarium; M, lens. made of glass while the floor and sides were of wood painted with camera-black. Within this aquarium a second cylindrical one (FG), made entirely of thin glass and measuring 20 cm. in diam- eter, was suspended from the ceiling by means of a fine wire (/7) attached to a swivel to allow turning. On the floor of the outer aquarium and directly beneath the inner one was drawn a circle 74 Herbert Eugene Walter 10 cm. in diameter and marked off plainly into arcs measuring 5 degrees each. An incandescent lamp (/), placed on the table at approximately the height of the inner aquarium floor, could be manipulated at any desired distance, while a diaphragm (Ff) prevented much of the light from reaching either the upper surface of the water contained in the two vessels or the floor of the outer aquarium whence it would be reflected. A biconvex lens was then so interposed as to make the light rays practically parallel upon theiremergence fromit. ‘Their course through the inner aqua- rium was kept parallel by means of the medium of water on both its inner and outer sides. A nearly uniform intensity over the entire floor of the swinging aquarium was thus obtained and the objection arising when the inner aquarium is used in air, viz: that it acts as a converging lens, was obviated. Side reflections were eliminated by enclosing the light (/), together with the interven- ing space between it and the diaphragm, with black screens. When a worm introduced into the inner aquarium began to glide, it could with slight mechanical disturbance be quickly cored) by means of moving cen aquarium, into any desired posi- tion with reference to the light, and then swung so as to bring its posterior end exactly over the center of the stationary circle below. Various species of planarians were started in this manner at right angles to the light. Out of 386 cases, 371, or 96 per cent, emerged from the 10 cm. circle at a point farther away from the light than that toward which they were originally directed. ‘This is taken to mean that 96 times out of a hundred the worms were negatively phototactic. If, however, the method of reckoning negativeness employed by Parker and Arkin (’o1) on the earth- worm is used, the foregoing per cent would be somewhat less. These authors assume (’o1, p. 28) that the apparently positive responses of a normally negative animal, such as the earthworm, may be due to causes other than light, in which case an equal number of responses of like nature might be expected to occur on the negative side as well as on the positive. A number equal to the sum of these apparently positive responses should therefore be subtracted from the total of the apparently negative responses The Reactions of Planarians to Light 75 in order to obtain approximately the amount of unquestionable negativeness. By following this method in the case just given, the per cent of negativeness would be g2 instead of 96, but since this method assumes that normally negative worms are never posi- tive, which is contrary to the evidence to be given later, the most accurate estimate of negativeness would probably fall somewhere between these two percentages. Bdelloura, on the other hand, behaves in the same way only three times out of ten, therefore showing itself to be positively phototactic. This difference in orientation becomes more marked if the total number of degrees, that is, the amount of positiveness and nega- tiveness of emergence from the circle is used as the basis of reckon- ing, instead of only the number of times of emergence. Such a quantitative computation is shown in Table XII. TABLE XII Amount and kind of orientation to directive light exhibited by various species of planarians in 396 trials | Number of Total de- | Total de- |Percentage of Percentage of trials grees positive |grees negative) degrees neg. | degrees pos. Negative worms (Dendroce- lum, Planaria, Phagocata). 386 | 566 10157 94.7 ces Positive worms (Bdelloura) .. .| 10 397 50 11.2 88.8 Although the actual number of trials for Bdelloura in this table is small, they are characteristic of what was observed in a large number of unrecorded instances. The amount a planarian may deviate from the direction in which it is pointed, depends upon the direction of the light imping- ing upon it. A negative species deviates from a straight course least when headed away from the source of the light and most when headed toward it, while an intermediate degree of deviation occurs when the direction of the light is at right angles to the long axis of the worm. In the case of Bdelloura the converse is true, as shown in Table XIII. Rate of Locomotion. In obtaining the rate of locomotion of worms subjected to directive light, the double aquarium apparatus 76 Herbert Eugene Walter just described was used. After the worm to be tested had been placed in the inner aquarium and had begun gliding, it was so oriented that the tip of its posterior end came precisely over the center of the subjacent circle 10 cm. in diameter. The exact time of its departure from the center of the circle was then noted and the instant thereafter that the tip of the posterior end passed over the circumference of the circle was again taken and the worm’s course plotted at once on a duplicate circle sheet. Each worm was given four trials in this manner, being started in four different directions, toward the light, away from the light, and with the long axis of the body at right angles to the light, first with one side to the light and then with the other. TABLE XIII Amount of average deviation tn 2400 trials expressed in degrees of a circle, exhibited by negative planarians, (Dendrocelum, Planaria and Phagocata), and a positive one (Bdelloura) when pointed toward, away from, and at right angles to the source of light Direction in which the worm was pointed with regard) tOsthesl ser hiti tess /alejararacstcvatevaveracchey shia. ’eps/s a wierd ontynusith one At right angles Toward Away from Negative planarians, degrees.........-.-+++++++s0-- 48.1 128.7 2759 Positive planarians, degrees..... : eveuarsvele Waxege 49. 39-3 132.1 The time of the worm’s emergence from the circle was not taken with a stop-watch because the observer’s hands were otherwise occupied. Instead a small clock, ticking half-seconds, was placed conveniently near. By counting the number of ticks during the interval of the worm’s transit from the center to the circumference of the circle the time consumed could be determined within less than a half-second. After tracing the worm’s course on a dupli- cate circle sheet and measuring the same by means of a map measurer, a unit of distance was obtained, which together with the known unit of time consumed in covering this distance, fur- nished all the data necessary for computing the rate of locomotion. Ten representatives of Dendroccelum lacteum, Planaria macu- lata, Phagocata gracilis and Planaria gonocephala respectively were given four trials apiece by the method just explained. ‘The results are presented in Table XIV. From the 160 records thus obtained it becomes evident that the average rate of locomotion The Reactions of Planarians to Light fh is greatest when the worms are pointed toward the light, and least when they are pointed in the opposite direction, while an inter- mediate rate occurs when they are started at right angles to the light. This result is at variance with the findings of Parker and Bur- nett (’00, p. 381), who incidentally reported that Planaria gono- cephala when started away from the light traveled faster than when started toward the light. TABLE XIV Average rate of locomotion, expressed in mm. per sec., of various species of planarians when started torward, away from, and at right angles to the source of directive light of 27 c.m. intensity. : Dendroceelum | Planaria |Phagocata Planaria Total Species | lacteum | maculata | gracilis | gonocephala| average Direction in which the worm was pointed with reference to the light At right angles'.....5....2.% 0.855 1.475 1.445 0.980 1.19 LGR Eid logoseoopagnaouneder 0.g10 1.505 1.430 1.205 1.26 wma ye ir OM cra )axerereesey~layate/2) b s s Le Leo eH [pet eee =| os vo eed ae s by 5 Sgilze|sa > a S S| & le as|A =< los Duration of movement in a typ- ical set of experiments, minutes) 47 60+ | 60+ 56+ 18 15 | 6s Percentage of orientation to light upon coming torest.......... 41 41 30 30 Wigwag responses at the critical line separating two intensities of non-directive light, per cent 39 39 8.5 | 8.5 Average number of seconds re- quired to escape from a circle | 10 cm. in diameter ........... 48.8 | 59.6 | 39-6 | 49.3 | 62.2 27.6 | 44.9 Precision of response Deviation in degrees upon emerg- ing from a circle 10 cm. in dia- meter when headed away from | the ightsrcyarcievaa iterate s(cte(=teieiets= ay Hae D ey 29.1 26.3 10.1 11.4 | 10.7 Rate of locomotion in mm. per SECanaciteccsee el tteretersiace 1.47 | 1.075| 1.395| 1.28 | 0.85 0.85 pigment in its eyes only, may be brought to a standstill very read- ily by means of light stimulation. With the exception of the eyes it may be possible that the photoreceptive apparatus is not differ- entiated to such an extent that it could fairly be said that any part of the translucent planarian body is entirely free from the direct stimulation of light. The relation of pigment to light reactions The Reactions of Planarians to Light 127 is, however, by no means settled in the foregoing observations. This matter should be finally tested by comparisons in the behay- ior of different individuals of the same species showing variation in pigmentation or of identical individuals at different times when their phases of pigmentation are unlike, rather than upon indi- viduals of different species. It has proven impossible to include such a consideration in the present paper, but the preliminary steps toward attempting an analysis of the function of pigment with reference to light reactions have already been made and it is expected that a discussion of this phase of planarian behavior will be presented later. It may be stated here that when Planaria maculata is fed with a drop of human blood, a decided increase in pigmentation makes its appearance within a few days, due probably to the oxidation of the hemoglobin in the blood corpuscles with which the planarians have Become gorged. ‘This single observation suggests an experi- mental means oe controlling che amount of pigment in a single individual and it may reasonably be supposed that tests of eae ior before and after excessive pigmentation will contribute direct evidence upon the part played by pigment in reactions to light. Summary. Mutilated planarians in general respond to light with less accuracy than normal individuals. Small worms move more slowly than large ones in the case of those species whose photoreceptive apparatus 1s not solely confined to the anterior end of the body. In the case of Dendroccelum lacteum, whose photoreceptive apparatus is relatively greater in small individuals than in large, the rate of incomouonet faster among the smaller than among the larger. Small worms orient with less accuracy than large ones. Pla- narians in the relaxed, resting position are less responsive to light than when they are stretched out in the act of gliding, a result probably of the greater exposure of the photoreceptors to light in the latter instance. The anterior end of the body 1s the chief photoreceptive region and in certain worms, such as Dendroccelum lacteum and Bdel- loura candida, the anterior end is apparently the exclusive seat of this function. 128 Herbert Eugene Walter No marked difference in response to light is shown between worms stimulated on the ventral surface and those equally stim- ulated on the dorsal surface. Aside from the eyes, which form at least a part of the photo- receptive apparatus, no definite light-receiving organs were recog- nized. Planarians possessing dark-colored pigment distributed over the body show in general greater activity when subjected to light than forms in which there is no dark pigment except in the eyes. The central nervous system, as well as the more exterior parts of the planarian, may possibly be stimulated directly by such light as passes through the translucent body. B The Phystological Basis of Behavior The continually changing adjustment in any organism between the incoming and the outgoing energy gives rise to varying phases of metabolic balance, which may be designated as different “ physi- ological states.” Such physiological states form a noticeable fac- tor in the behavior of any animal, a fact towhich Jennings (’o4b, p- 109) in particular has called attention. That the difference between such states is great may be readily demonstrated. A planarian’s response to directive light when it is in a relaxed, quiescent condition is plainly different from that exhibited after it has been vigorously disturbed by a brush. In fact, it is extremely difficult to get two animals that are in precisely the same physiological condition, or the same animal in precisely the same state at two different times, since the exact adjustment of physiological states is too. delicate a matter to be controlled by the present gross experimental methods. The attempt is ordinarily made to eliminate from experiments, so far as possible, the disturbing element of changing physiological conditions, that is, to keep constant all the factors except the one which is being subjected to test, and those results are counted as most successful in which such disturbance is reduced to a minimum. It is the purpose of this section first, to give a possible classifica- tion of the different physiological states in which a planarian may The Reactions of Planarians to Light 129 be, and, secondly, to pass briefly in review some of the many ways in which light may change the physiological state of such a worm. a Classification of Physiological States It is by no means easy to define even a simple physiological state, since the subtle changes form a continuous series of conditions which pass imperceptibly into each other. An arbitrary classification for convenience may, however, be made as follows: 1 Relaxation, or rest. 2 Slight activity, without locomotion. 3. Normal activity. 4 Violent activity. 5 Rigor. 6 Exhaustion. In the first of these states there is a minimum expenditure of energy caused by the ebb of the katabolic processes. The second and fourth states indicate what are often referred to as conditions of low and high “tonus,” but as this term has a technical significance with reference to muscle reactions, it will not be used in this classification. Thethird state, thatof normal activ- ity, is the average condition; it is the most desirable one to main- tain in testing the animal’s responses to different stimuli. By rigor is understood a state wherein there may be an excessive outgo of energy, but unaccompanied by movement, while under exhaus- tion is cneleded the condition when energy is not being released because there is none to release. That excessive or continuous light stimulation may go beyond the point producing rigor or exhaustion and may actually cause death, has been repeatedly proven in the case of bacteria by a long line of observers. The inhibitive effect of excessive light upon other organisms than bacteria has been pointed out by Berger (00) with reference to Cubomeduse; by Pearl and Cole ® Tyndall (’78), Downes and Blunt (’77, ’78), Jamieson (82), Duclaux (’85a,’85b, ’g0), Arloing (’87a,’87b), Roux (’87), Dandrieu (’88), Raum (’89), Pansini (’89), Janowski (’90), Buchner (’92) , and Ward (94). 130 Herbert Eugene Walter (02) in the case of various infusoria as well as Hydra, Hyallela, Clepsine, Stichostema and Physa; by Yerkes (’03) for Goni- onemus and by Carpenter (’05) for Drosophila. b Changes in Physiological States Induced by Light A variety of stimuli besides light may cause an animal to pass from one physiological state to another. For: example, the sense of phototaxis was reversed through mechanical stimulation by Towle (’00) in Cypridopsis and by Holmes (’or, ’o5b) in Orches- tia and Ranatra. The following typical illustration of the manner in which changes from one physiological state to another succeed each other is offered as a basisof comparisonwith the responses tolight itself, which are about to be described. In the absence of mechanical stimuli a planarian may be in a state of relaxation. Very gentle mechanical stimulation causes the worm to lift its anterior end and move it cautiously about, bringing the animal into a state of slight activity without locomotion. If, now, the mechanical stim- ulus is prolonged or increased in intensity, enough energy is released to put the animal into gliding locomotion, when it may be fairly said to have passed into the state of normal activity. Pro- vided the stimulation is made still more pronounced, the worm can next be forced to forsake gliding for crawling or humping, so passing into the state of violent activity. Further, it is possible by vigorous shaking to throw the worm, temporarily at least, into a condition of inactivity through excessive stimulation, during which the animal would remain quiet, not because it is failing to release any energy, but because it is unable for the time to set free its energy in the form of locomotion. In other words, it is in the state of rigor. Last of all, if mechanical stimulation is repeatedly applied a condition of exhaustion will appear when the worm has no more available energy and so is unable to move at all. Effect of Different Intensities. As already pointed out, no inten- sity either of directive or non-directive light was found sufficient to change the condition of normal gliding into crawling. The Reactions of Planartans to Light 131 Moreover, light of any intensity or direction frequently proved ineffective in arousing a quiescent worm into any state of apparent activity, particularly if the worm had but recently passed into the state of rest after a prolonged period of exercise. Effect of Excessive Light. In the experiments with non-direc- tive light it appeared that Planaria gonocephala, when subjected to an intensity of 431 c.m., showed somewhat less activity than at lower intensities, both with respect to rate of locomotion (Table III, p. 57) and to the number of turnings made (Table VI, p. 59); yet, so high a degree of intensity of the light stimulus was appar- ently not sufficient to cause a change into the physiological state of light rigor. It was comparatively easy, on the other hand, to transform Bdelloura candida by means of excessive light from the state of normal activity into that of light rigor. Effect of Sudden Change in Light Conditions. A sudden change in light intensity either by increase or decrease is more effec- tive in producing a new physiological state than an equal grad- ual change. ‘The sudden withdrawal of the lamp to a consider- able distance, for example, is usually sufficient to throw a worm from a normal state into violent activity, that is, from a gliding movement into a disturbed state in which the anterior end is waved actively about. But if the light is gradually withdrawn the same distance the worm will usually not pass into a different physiological condition. The sudden introduction of complete darkness was never found sufficient to reduce an active worm more than temporarily to the resting position. Sudden dark might temporarily halt a moving worm, but it would not cause it to come to rest and assume the relaxed contour. In Bdelloura candida sudden dark, instead of checking the animal’s movements, threw it into violent activity. Effect of Continued Exposure to Light. Continuous exposure to light results in fatigue, which finally causes planarians to change from the state of normal activity to that of relaxation. ‘The tend- ency toward such a change is shown in Table XXXV, where the responses of a number of worms newly subjected to light stimu- lation are contrasted with the responses made by the same worms after they had been moving about for several hours in the light. 132 Herbert Eugene Walter The fresh worms show more activity than the fatigued worms do. Otherwise expressed, the worms have a tendency to change into a lowered physiological state upon continued exposure to light. TABLE XXXV Fatigue effects due to continuous exposure to non-directive lights forming adjacent fields of different intensities, as shown in the behavior of Phagocata gracilis Ratio oF THE TWo -96: 13-4501 Average INTENSITIES iene ae 3-45 3 Percentage of the re- Going Going Going | Going Going | Going sponses at the crit-| into into FF into into 4 into into a ical line separating greater lesser & | greater | lesser Pe greater lesser is = iS ' 0 f 2 & the two intensities intensity intensity 2 intensity intensity 2% intensity intensity oa Fresh worms........ | 10.5 | 21 16 45-5 47-5 | 46.5 | 28 34-4 | 31+ Fatigued worms.... 2.5 9-5 32-5 33-5 | 33 17.05 21.5 | 19.5 It may be incidentally noted in Table XXXYV that, as has already been pointed out in another connection, the percentage of responses is greater when the contrast between the light intensities is greater, and that both fresh and fatigued worms respond oftener upon going into the lesser intensity than when going into the greater intensity. The time required for a worm placed in directive light to come to rest; that is, to run the gamut from the state of normal activity to that of rest, becomes gradually shorter with continuous expo- sure. As fatigue increases the worm shifts down the scale of physi- ological states in less time than when freshly subjected to directive liohe A specific case of this kind has already been described in the paragraph on “duration of movement”’ (p. 105); where in 39 consecutive trials the change from normal activity to relaxation was first made in 18 minutes, but the thirty-ninth time in Ir} minutes, while the fortieth time even mechanical stimulus failed to arouse the exhausted worm from the resting position. Effect of Previous Exposure to Dark. Worms kept several hours in complete darkness make a larger percentage of re- sponses to changes in their light environment than those which previous to experimentation have been several hours in light. Individuals removed from the stimulus of light for any consider- The Reactions of Planarians to Light 133 able time are more responsive when subjected to it, for the reason that they are in a physiological state farther removed from fatigue than those worms which have remained a long period in the light. This point is brought out in Table XXXVI. TABLE XXXVI Percentage of reactions of two worms, Planaria gonocephala,to a sudden change in light intensity both when previously kept several hours in the dark and also when previously exposed for several hours to light Percentage of responses Number of observations After several hours in the light........... 54 100 After 48 hours in the dark............... 66 100 Summary. Physiological states grade imperceptibly into each other, but may be tentatively divided into: 1, relaxation; 2, slight activity; 3, normal activity; 4, violent activity; 5, rigor; 6, exhaus- tion. Various stimuli besides light may induce a change from one physiological state to another. No light intensity lower than 431 c.m. is sufficient to throw a worm into a higher state than that expressing normal activity, nor is the absence of light sufficient to bring a planarian to rest. Excessive light intensity shows a tendency to carry Planaria gono- cephala from a state of normal activity to one of rigor. Bdelloura candida is easily changed into a condition of rigor by light. A sudden change of light intensity acts more immediately than a gradual change in causing planarians to pass from one physio- logical state to another. Continuous exposure to light induces fatigue, finally resulting in the passage of the worm into a state of continuous relaxation, in which condition it becomes practically indifferent to light. Repeated trials of the time required in constant light to come to rest show that a progressively shorter interval occurs between the state of normal activity and that of relaxation until a point of complete inactivity is reached, the worm finally remaining in the latter state for a prolonged period. 134 Herbert Eugene Walter Planarians kept for some time in darkness pass into a state in which they are more’responsive to light than individuals exposed for a similar length of time to light. G Psychological Basis of Behavior Among the first questions that naturally arise concerning the behavior of planarians in light are those which approach the mat- ter from a psychological point of view. How much can planarians actually see, and can they, by repeated experience, “learn” to adapt themselves to changes in the light surrounding them ¢ To this kind of inquiry it is most difficult to give a satisfactory answer, for the reason that it is impossible to go beyond conjecture and inference in judging what any animal, aside from man, can see or know or experience. It is only possible to state, in more or less definite terms, the responses which animals make to light, since it is beyond man’s power ever to experience how animals “feel” under any circumstances. a. How Much Can Planarians See? Broadly speaking it may be said that planarians can distinguish light from darkness. The experiments described on pp. 84, et seq., relating to planarians placed ijn aquaria so surrounded by backgrounds as to produce regions of different light intensity, point to this conclusion, since when subjected to such differential environments the worms come to rest in the darkened areas. Again, the numerous responses made at the critical line separa- ting two light intensities may be regarded as evidence of some power of discrimination on the part of the worm between dif- erent intensities of light. It is probable, furthermore, that planarians can distinguish a moving object when that object is of sufficient size and contrasts with its surroundings in its degree of illumination, for the reason that a moving object from which light is reflected, means the same to a worm coming into the vic:nity of the object as any other change in the direction of light, such as might be caused by moving a The Reactions of Planartans to Light 135 lamp from one position to another. ‘To changes in directive light planarians are known to respond very definitely, and consequently they may be said to distinguish the motions of objects. With regard to true seeing, however, in the sense of distinguish- ing the forms of objects, it is safe to assume that planarians have almost no power whatever, since their eyes are optically unable to form images even if the central nervous system were highly enough developed to interpret images when formed. In the case, there- fore, of Planaria alpina, which, according to Collin (91, p. 180), “shuns” Planaria gonocephala when the latter has been put into the same aquarium with it, seeking “‘strenuously to escape” from its larger relative, the conclusion does not necessarily follow that P. alpina sees an enemy and experiences the sensation of fear. As previously pointed out (p.95), the whole matter is probably explainable on the basis of negative chemotaxis alone. ‘To attri- bute fear, therefore, or any other similar complex sensation, to an organism whose responses are so plainly of a simple reflex nature, is to go quite beyond the evidence. In the performance of the two great life processes of nutrition and reproduction, light is apparently in no way a direct aid to pla- narians, since they thrive in situations from which light is entirely excluded, as in caves, and since they habitually frequent places where this factor is reduced to a minimum. Light cannot, then, be regarded as a directly essential factor in the life of planarians. That light is not essential to the activity of protoplasm has more than once been demonstrated. Engelmann (’79), for example, showed that the streaming protoplasm of plant cells occurs nor- mally in darkness, while Maupas (’87) found ciliates multiplying as rapidly in the dark as in the light. b~ Are Planarians Able to “Learn” ? With regard to the ability of these worms to acquire upon repe- tition an abbreviated form of response; that is, to “learn,” a few suggestions may be drawn from experiments already described in other connections. It will be remembered (p. 93) that when a small aquarium 136 Herbert Eugene Walter was delicately mounted upon a turntable, such as is used in “‘ring- ing’’ microscopic slides, a very slight rotation was sufficient to bring to a halt momentarily a gliding worm in this aquarium. It was possible to control this momentary response to such a point of nicety that the anterior end of the worm could be made to halt for an instant without interfering with the onward loco- motion of the posterior end. If this slight rotation was repeated at intervals of a second it was found that the worm under obser- vation halted with less and less certainty, until after a dozen or more trials it continued to glide on without halting at all. In ordinary phraseology the worm had learned by experience not to be alarmed by a sudden mechanical shock. ‘The lesson, how- ever, was always very soon forgotten, for after an interval of less than a minute, during which the aquarium remained stationary, the worm responded exactly as it did at first, whenever a slight rotation was made. In a similar way the skioptic response of Bdelloura candida became less pronounced upon repetition, until it was possible to throw a shadow upon the animal without obtaining any response at all. Again, when worms were placed in a field of non-directive light, parts of which were of two different intensities, the number of wigwag responses made at the critical line separating the two intensities grew less after the animals had repeatedly crossed the line. At first the new condition of sharply contrasted light intensities in the worm’s field of locomotion called out a large percentage of wigwag responses. Later, however, by repeated experiences the worm became familiar with this feature of its environment and made fewer wigwag motions. A definite instance of such a decrease in response is given in Table XX XVII. TABLE XXXVII Responses of Planaria gonocephala on crossing the line separating two intensities of non-directive light Wigwag movements No response Percentage of response First 25 crossings....... 21 4 84 Second 25 crossings..... 19 6 76 Third 25 crossings...... 12 13 48 Fourth 25 crossings ..... | 8 17 32 The Reactions of Planarians to Light 137 It will be seen that when Planaria gonocephala was first intro- duced into a field of contrasted intensities, it made the wigwag response at the critical line marking a change of light intensity, in 84 per cent of the first 25 crossings, while during the second, third and fourth sets of 25 crossings, the per cents uniformly decreased until at the fourth 25 crossings the number of wigwag responses fell to 32 per cent. It may be objected that the instances thus far cited in this section find a more reasonable explanation upon the hypothesis of fatigue, but the same surely cannot be said of the fol- lowing case. It was found that Planaria maculata oriented itself to directive light at successively shorter intervals when the position of the light was suddenly changed. To produce such a series of re- sponses there was placed in the dark room a shallow aquarium with an electric lamp at either end, under the control of the right and left hand, respectively, of the experimenter. A planarian was placed in the middle of the aquarium and the right-hand light turned on. As soon as the worm was fairly oriented to this light and gliding away from it, the right-hand light was turned off and at the same instant the left-hand light turned on. The time in seconds required for the worm to orient to the new light; that is, to turn 180° and begin to glide away, was recorded. On p. 89 a typical series of records of such responses is given, in which the number of seconds required for re-orientation when the source of light was reversed, varied from 260 seconds, at first, irregularly down to 35 upon the sixteenth trial. It will be seen from this series that the worm acquired by experience some degree of facility in adapting itself to certain variations in its environment which it would never be liable to encounter in nature, and that this adapta- tion cannot be explained as due to fatigue. Davenport and Cannon (97, p- 32) found similarly that “‘ Daphnias respond more quickly and accurately to light after having made several trips to it.” It is quite certain, however, that any educative attainment which a planarian may experience, or which a planarian may acquire, is exceedingly evanescent and also that there is no evi- dence that the worm emerges from reflex behavior into responses connected with consciousness. 138 Herbert Eugene Walter Summary. The existence of feeling or consciousness among planarians is a matter of pure conjecture. From their responses it may be inferred that they are able to distinguish dark from light, as well as objects in motion, but it is not clear that they can distinguish the forms of objects. The knowledge which planarians have of objects in their imme- diate environment, such as food, enemies, etc., depends largely upon chemical and tactile means. ‘They are, therefore, as well able to go through the entire range of their activities in the dark as in the light. Upon repetition planarians may in some instances become accus- tomed to, or acquire greater facility in, responding to stimuli, but this result of experience is almost instantly lost, so that it is doubt- ful whether these animals possess more than the merest rudi- ments of the primary criterion of consciousness, namely, the ability to learn. VI GENERAL CONCLUSIONS Probably the questions which have occupied the greatest share of attention throughout the literature dealing with the reactions of organisms to light, are the following: 1 Is the direction or the intensity of light of more importance in orientation ? 2 Which theory best explains orientation and phototaxis, the theory of trial and error or that of the tropisms ? 3. How far is behavior with respect to light, adaptive ? I DIRECTION OR INTENSITY Before the part played in the behavior of planarians by either the direction or the intensity of light can properly be discussed, it will be necessary to present a brief historical résumé of certain general conclusions reached by investigators along this line. A Historical Cohn (’53), Strasburger (’78) and Loeb (’g0, ’93a) attributed the directive effect of light to the action of the rays. In a later The Reactions of Planarians to Light 139 paper Cohn (64) abandoned his first position and came to regard intensity as the important element in light, a position also main- tained by Famintzin (67), Engelmann (’83), Oltmanns (’92), Verworn (’or) and even by Loeb (’93b) in the case of Planaria torva, which he found came to rest in accordance with the intensity, and regardless of the direction, of the light. Davenport and Can- non (’97) modified this point of view by attempting to show that direction and intensity may each operate independently, producing, respectively, “phototaxis”’ and “photopathy.”’ Holt and Lee (’or) followed with an excellent summary of the whole controversy, emphatically maintaining, in opposition to Davenport and Can- non, that intensity alone is the only possible operative factor in light stimulation and that direction of the rays has no effect what- soever except in determining a greater intensity of light with refer- ence to one part of an organism as compared with other parts. Among more recent investigations Holmes (’03), experimenting with the same organism that fled Oltmanns to ascribe the greater importance to intensity, namely, Volvox, declares himself in favor of direction, while Zeleny (05), on the other hand, gives an instance of Serpulid larve going both toward the source of the light and away from it; that is, moving regardless of direction, in order to arrive in regions of increased intensity. Carpenter (’05) found that the pomace fly, Ampelophila droso- phila, will orient to the direction of light after it has first been sufhciently aroused by the intensity of the light, while both Yerkes (99) and Towle (’oo) maintain that direction and intensity are by no means mutually exclusive, and that each may play a part simul- taneously in determining the behavior of an organism. Lastly, it has been made clear by Parker (’03) that, besides direction and intensity of light, the size of the source of illumina- tion may determine the orientation. ‘This theory explains why butterflies alight upon a patch of reflected sunlight which produces a large but faint retinal image instead of flying toward the sun itself, which forms only a small but intense retinal image. In the case of planarians, however, this phase of light stimulation is not operative, since the eyes of these animals are incapable of forming retinal images. 140 Herbert Eugene Walter B Conclusions with Reference to Planarians The behavior of planarians may in general be more satisfac- torily explained by regarding, with Loeb, the intensity rather than the direction of the light as the principal operative factor in light reactions. At the same time there is much evidence that the inten- sity utilized by the organism, is intimately associated with, and powerfully modified by the direction of the light. As a basis for these conclusions the following points will be considered. First, the distinction between direction and intensity; secondly, the way in which directive light modifies the intensity with reference to planarians; thirdly, the action of intensity without the modifying effect of direction, and finally, modifying effects of factors other than light. a. The Distinction Between Direction and Intensity Theoretically it is plain that light per se with respect to any fixed point, may be regarded in two distinct aspects, namely, that of intensity and that of direction. ‘The intensity of light under ordi- nary circumstances varies inversely as the square of the distance and is independent of the position of the source of light. That is to say, at any points equidistant from its source, light has the same intensity, but the more remote the less is the intensity at any given point. ‘The direction of light, on the contrary, is depend- ent solely upon the position of the source of the light and in no way upon the distance. When intensity and direction are con- sidered with reference not to a fixed point but to an organism presenting three dimensions and made up of differentiated pro- toplasm, the basis of light relations becomes more complex. Light cannot here be treated as a phenomenon per se but must be considered in relation to a differentiating organism. It is true that intensity in the case of the organism, as in the case of a fixed point, varies with the distance from the source of the light. A decided difference, however, appears in the case of the organism inasmuch as, owing to its structure, the intensity received by it varies also in accordance with the position of the light. ‘This The Reactions of Planarians to Light I4I second form of variation in intensity is directly due to the fact that the organism has a solid form and is not homogeneously photo- receptive. The direction of light with reference to the organism, presenting as the latter does a structurally diversified form, is influential only as regards the position of the source of light, just as in the case of a fixed point. Any change in the position of the source results, then, in a redistribution of the intensities falling upon the organism, so that again the intensity received varies in accordance with the position of the light. It is this factor of posztion in light that has been termed the direc- tive influence of light and it is seen to be due to variations in the intensity of light with reference to the organism, and not to any peculiar property of light iself. By “‘non-directive light,” on the other hand, is understood those conditions which secure for the organism equalized or symmetrical intensity with respect to the parts stimulated. If this interpretation is correct there can be no response, strictly speaking, to the direction of light exclu- sive of intensity although the factor of intensity may be continually modified by that of direction in the light relations of organisms. b The Modifying Influence of Direction It is undeniable that the planarians experimented upon exhib- ited without exception a definite characteristic phototaxis, that is to say, they habitually go either toward or away from the source of light according as they are respectively positive or negative. In analyzing this phototaxis it seemed desirable to eliminate so far as possible the factor of intensity, but the attempt to do this was only partially successful owing to physical limitations. A step was made, however, toward subjecting worms to directive light without at the same time exposing them to a variation in intensity by inserting a biconvex lens between the source of the illumination and the aquarium, thus making the diverging rays of light parallel throughout their course in the aquarium. By this “means was formed a field equal in its amount of illumination at 142 Herbert Eugene Walter the two ends of the aquarium, the one opposite and the one next to the source of light, with the exception that there was a slight difference at the two ends due to the fact that light in its passage through water is partially absorbed. But modification of light in any degree results in producing less intensity at the farther end of the aquarium, though this difference is less pronounced when a lens is employed. ‘Therefore, although worms placed in this apparatus went with considerable precision in the direction of the propagation of the light, there is no certainty that their behavior was not due simply to differences in intensity. Worms which are thus apparently traveling directly in accordance with the direc- tion of the light, are meantime being subjected to different inten- sities at the anterior and posterior ends of the body, for the reason that the anterior end is more or less shadowed by the rest of the body, since the latter cuts out a certain portion of the light received at the posterior end. That direction of light is a factor by no means to be disregarded, even if it cannot be proven to be the immediate cause of phototaxis, is apparent when it is recalled that slight changes in direction call out corresponding changes in the course of the gliding pla- narian, whereas considerable changes in intensity when the direc- tion remains constant and particularly when such changes are gradually made, may fail entirely to produce corresponding changes in the worm’s behavior. ‘This is due to the fact that slight changes in direction may cause considerable changes in the asym- metry of illumination. When a worm, for example, is receiving horizontal light from behind, its head is more or less in shadow, the sides of its body being at the same time equally illuminated. The moment the light is shifted in even a small degree to one side, one entire side of the animal may receive an increase of illu- mination and the opposite side be thrown into shadow. Thus a slight change in position initiates a fundamental change in the distribution of intensity over the planarian’s body. c Instances of Behavior Due to Intensity Alone The effect of intensity as a separate factor from the directive influence of light is clearly demonstrable in certain phases of The Reactions of Planarians to Light 143 light reactions. Yo isolate intensity by excluding the possibility of directive light; that is, to secure equalized intensity with refer- ence to the organism, is not difficult and the manner in which this was done, with non-directive light falling upon a horizontal field from above, has been sufficiently detailed in the body of the paper. It may be briefly recalled that planarians experimented upon by this method showed a certain unmistakable degree of response which could be referable only to differences in equalized intensity. For example, the rate of locomotion was found to be faster in any non-directive intensity up to 431 c.m. than in darkness, although light in itself was not always sufficient to start a worm into ac- tivity, nor was its absence sufficient to check an animal already in motion. Again, though no close correlation between behavior and the degree of intensity was found to exist, there appeared cer- tain general results which were plainly referable to intensity differences only. Instances of such results are the behavior of Planaria gonocephala (which was modified in several particulars at 431 c.m. as compared with its behavior at. lower intensities) ; the coming to rest in regions of diminished intensity of individuals of all species except Bdelloura; and the increase of wigwag re- sponses corresponding to an increase of intensity differences when a field of contrasted intensities was used. It is interesting to observe that increase in the intensity of non- directive light, and continued exposure to non-directive light of constant intensity, both tend to produce the same behavior that would result in directive light. Under any of the three conditions just mentioned there resulted by actual experiment fewer turnings, fewer ‘indefinite changes” and more nearly straight paths on the part of planarians than occurred when the worms were (1) placed in non-directive light of lower intensity, (2) subjected a short time to non-directive light of constant intensity, or (3) left in darkness. Now, fewer turnings, fewer “indefinite changes,” and more nearly straight paths are ordinarily characteristic results of directive light, so that here is a case of reactions, which if resulting from the employment of directive light would be termed phototaxis, occur- ring in non-directive light as the result of intensity alone. Mast (03) experimenting upon the reactions of planarians to thermal 144 Herbert Eugene Walter stimuli obtained a similar result. He observed that apparently “negative” as well as “positive” responses resulted when the animals were subjected to non-directive thermal stimult. Another noticeable phenomenon with reference to responses to intensity is, that more wigwag responses occurred at the critical line separating two different intensities when the lesser of the two intensities was 16 c.m. than when it was 33 c.m. (Table XI, p. 69). Similarly responses were more frequent when planarians were subjected suddenly to dark than when they were flooded suddenly by light, and, throughout a large number of series, responses were invariably more frequent ben the worms were passing into a region of diminished intensity than when they were entering an area of increased intensity. It is to be inferred that all these phases of behavior are due to the probable fact that the lower inten- sities compared are nearer the worm’s optimum as regards light than the higher ones, since the latter apparently have a tendency to inhibit activity. Lastly, the relative part played by intensity of light varies decidedly in different species of planarians. ‘The relative inten- sity in different parts of an aquarium, when no lens is used to les- sen the contrast, has comparatively little influence upon Phagocata gracilis, as its extensive wanderings (typically reproduced in Fig. 11) toward and away from the source of light, indicate. Planaria maculata and Planaria gonocephala, on the contrary (Figs. 13 and 14), notwithstanding their ability to come toward the light in the direction of the “rays” throughout the farther half of the dish, seemed invariably to encounter an impassable barrier as soon as they approached within a certain intensity, thereby showing a more delicate responsiveness to intensity differences. d The Modifying Effect of Other Factors In attempting to analyze the relative bearing of the intensity and of the direction of light upon the behavior of planarians there must be constantly kept in mind two general sources of error which are always present when these factors of light are in operation. ‘These The Reactions of Planarians to Light 145 are (1) the physiological state of the organism at the time of obser- vation, and( 2) the simultaneous effect of other stimuli. A physiological state may be directly traceable to known causes, such as previous exposure to other stimuli or the condition of meta- bolic balance in which the animal chances to be at the time of observation, or, again, it may be the result of factors at present unknown, which consequently, although in active operation, are not susceptible of analysis. In any case it is certain that the uncontrolled factors comprehended under the term “ physio- logical state” prove individual planarians to be not identical Sigeha nists, but organisms possessing a more or less definite individuality. Moreovens it has been shown that differences in physiological state play a greater part in the determination of behavior than do intensity differences in the light stimulus. When a planarian is approaching a state of fatigue, for example, it becomes indifferent to differences of intensity. With regard to the simultaneous effect of other stimuli acting in conjunction with light, it has already been pointed out that behav- ior is the resultant of all the factors, external as well as internal, which may be acting upon an organism at a given time, and that consequently the effect of any one of the operating factors, such as that of light, for example, cannot be determined unless the value of the other factors involved is also taken into account. In support of this view, which is so self-evident, it will be recalled that some of the ways in which the responses of planarians to light may be modified by geotaxis, thigmotaxis, goniotaxis and chem- otaxis, were touched upon. Summary. Direction and intensity are separable qualities of light. Direction is dependent upon the relative positions of the light and the organism, whereas intensity depends upon the dis- tance between the light and the organism as well as the initial intensity of the light. When applied to living organisms intensity may act independ- ently of direction, or in conjunction with it. Direction cannot act independently of intensity upon organisms, since the latter possess definite form and consequently cannot receive the light at a single point. 146 Herbert Eugene Walter With reference to an organism, directive light is resolvable into unequalized intensity and non-directive light into equalized intensity. Asymmetrical intensity in directive light is largely due to the partial shadowing of that part of the body farthest away from the source of the light. Slight changes in the position or direction of the light may cause considerable changes in the symmetry and the degree of the shadow effects and consequently in the relative intensity of the light on different regions of the body of an organ- ism. To different degrees of equalized or symmetrical intensity pla- narians show considerable response, but the correlation between their behavior and the degree of intensity is not so close as it is in the case of asymmetrical intensity. Increase in intensity of non-directive light, continued exposure to non-directive light of constant intensity, and change from dark- ness to non-directive light, all tend to bring about apparent photo- taxis similar to that occurring in directive light. Responses are more frequent on the part of planarians in inten- sities approaching the optimum than in higher intensities, where there is a tendency to inhibition. . Relative differences in responses to various intensities are due to specific differences between planarians. The physiological state of an organism together with the influ- ence of known stimuli other than light are constant sources of error in estimating reactions to light. ‘These factors taken to- gether play a more important part in planarian behavior than light stimulus. Finally, the action of light upon planarians is a function of its intensity, which, under certain conditions, is emphasized by the direction of the light. 2 TRIAL AND ERROR OR TROPISM? It is apparent from the preceding section that light may have two effects upon organisms. Of these, one is a kinetic effect, arising from the intensity of the stimulus and resulting in a gen- The Reactions of Planartans to Light 147 eral activity termed photokinesis, while the other, connected indi- rectly at least with the direction from which light impinges upon an organism, is called phototaxis. In the case of planarians these two phases of light stimulation have been shown to be intimately associated and both operative. Carpenter (’05) pointed out in the case of the pomace fly that phototaxis occurs only when pre- ceded by photokinesis or some other reaction, and such an inter- relation of the two is undoubtedly of wide occurrence. The object of this section is to inquire into the causes underlying pho- totaxis. Loeb (’93b) has shown that phototaxis is the result of orientation. It does not necessarily follow, however, that orien- tation invariably results in phototaxis. In fact Dearborn (’oo) found that crayfishes would orient to an electric light introduced into the water near them without making any considerable loco- motor movements in consequence.’ To the question of how orientation of organisms to light is caused, three possible explanations may be presented: 1, Chance result of photokinesis; 2, reflex response to directive stimuli; 3, voluntary action. Since the first hypothesis seems entirely inadequate to account for the uniformity of orientation in pla- narians, and the third alternative is out of the question with refer- ence to these animals, a consideration of the reflex responses to directive stimuli may be taken up at once. There are two general theories which attempt to explain the way in which orientation occurs through reflex responses to stimull. ‘These theories are first, the trial and error theory of Jennings and Holmes, and secondly, the tropism theory of Verworn and Loeb. By the trial and error theory orientation, with its consequent pho- totaxis, 1s interpreted as the result of repeated attempts on the part of an organism to become adjusted to any given stimulus. Those attempts which fail to result in adjustment to the stimulus are “errors,” and as such are followed by other attempts until finally some one secures the necessary adjustment. ‘Trials of this kind may be made in different ways according to the organism 7 Throughout the following discussion orientation will be understood as a position assumed with refer- ence to the light while phototaxis will be made to include motion toward or away from the source of the light. 148 Herbert Eugene Walter in question. Among the infusoria and rotifera, as Jennings has shown in a masterly series of papers,® such attempts at orientation are made by means of a “ motor reflex,” consisting in (1) a sudden withdrawal from the stimulus, (2) a rotation toward a structurally defined side of the asymmetrical organism, and (3), lastly, an advance in a new direction. In the case of organisms which do not possess marked asym- metry the trial and error method, as pointed out by Holmes (’o5a), resolves itself into a series of “‘random movements;”’ that is, a number of apparently experimental movements are made, which finally result in the best adjustment to the stimulus. In both of these methods the organism acts as a unit and not in response to localized stimulation received asymmetrically. The tropism theory, on the contrary, 1s based upon asymmetrical action as the result of asymmetrical stimulation. If an organism receives a stronger stimulus on one side of its body than on the other, the result, whether direct or indirect, is that it moves in such a way that this asymmetrical stimulation becomes symmetrical. In other words, orientation occurs. It is unfortunate that the tropism theory was made to apply to the behavior of the infusoria, since it has been shown beyond doubt by Jennings that exact observation of the behavior of these organisms and an analysis of its details does not admit of the tropic interpretation, but is, on the other hand, explained by the trial and error theory of motor reflexes. It is also to be regretted that the unquestionable rout of the tropism theory, as applied to cer- tain protozoa and other asymmetrical forms, should have led to an attempt to exclude it from the remainder of the animal king- dom. In a paper on the tropism theory Jennings (’o4a) names as an essential criterion of tropism the direct unilateral stimulation of the motor organs. After showing how inadequate such an assump- tion is to explain the orientation of animals, particularly that of Infusoria, he continues (’o4a, p. 104), “We should perhaps con- ® See bibliography in Contributions to the Study of the Behavior of Lower Organisms. Carnegie Inst. of Washington. Publication No. 16. 256 pp. 1904. The Reactions of Planarians to Light 149 sider here a modification of the original form of the tropism theory that has been proposed by some authors. ‘This is in regard to the assumption that the stimulating agent acts directly on the motor organs upon which it impinges. For this it is sometimes proposed to substitute the view that the action of the stimulating agent is directly on the sense organs of the side on which the stim- ulus impinges and only indirectly on the motor organs through their nervous connection with the sense organs. When thus modi- fied the theory of course loses its simplicity and its direct explain- ing power, which made it so attractive. In order to retain any of its value for explaining the movements of organisms, it would have to hold at least that the connections between the sense organs and the motor organs are of a perfectly definite character so that when a certain sense organ is stimulated a certain motor organ moves in a certain way. When we find, as we do in the flatworm (see the following paper), that to the same stimulus on the same part of the body, under the same external conditions the animal reacts sometimes in one way, sometimes in another, the tropism theory, of course, fails to supply a determining factor for the behavior.” It seems to me that the mechanism by means of which the asymmetrical response is brought about is immaterial, so long as that response can be shown to be the result of asymmetrical stimu- lation. Asymmetrical response might occur either from direct stimulation of the motor organs as was implied in the earlier papers on the infusoria, or by means of a more complex method, consisting of stimulation of the sense organ, transmission to the central nervous system and thence to the motor organs. The outcome in either case would fulfill the demands of the trop- ism theory, if asymmetrical response to asymmetrical stimulation be taken as its criterion. In the quotation just cited, the objection that such transmission compels stereotyped behavior is hardly valid, since stereotyped reaction is by no means the only alterna- tive of asymmetrical stimulation. ‘That flatworms do not respond uniformly to directive stimuli cannot be disputed, but that fact does not exclude the possibility of all tropic reaction on their part. The imperfection of response may be simply the result of imper- 150 Herbert Eugene Walter fections in the worm’s nervous circuit, assuming that planarian reactions are due to indirect rather than to direct stimulation of the motor organs. In fact, repeated evidence of the failure of a constant and perfectly invariable orientation on the part of pla- narians has been given in the preceding pages. Such failure, moreover, is quite as likely to occur in the application of the trop- ism theory to behavior as it is in the case of the trial and error theory, since stereotyped reactions and forced movements, as Holmes (05a, p. 112) has emphasized, are no more characteristic of tropisms, which depend upon a differentiated stimulation and response, than they are of trial and error movements, resulting from a single motor reflex given in response to all kinds of stimu- lation. Furthermore, it has been urged that tropism indicates a simpler form of reaction than trial and error for the reason that it in- volves only a local part of an organism while the motor reflex of trial and error requires that the organism act asa whole. Conse- quently, since motor reflex has been indisputably demonstrated as the method of infusorian phototaxis, Jennings (’04a, p. 95) asks, “Should we conclude that the reactions in the higher metazoa are simpler and less unified than in the protozoa f” That the motor reflex, which occurs with machine-like uni- formity, regardless of the point where the stimulation is received, is more complex in character than the stimulation of an asym- metrical part of an organism which may depend for its response upon sense-organ, nervous transmission and motor apparatus is an assumption difficult to sustain. It seems more reasonable to agree with Harper (05) in placing tropism higher in the evolu- tionary scale than trial and error. The fallacy that “tropism leads nowhere; it is a fixed final thing like a crystal’ (Jennings, ’o4c, p. 251), while trial and error alone offers possibilities of the higher evolution of phototaxis, has already been answered by Holmes, who points out that trial and error, at least that phase of trial and error depending upon motor reflex, is even more fixed and stereotyped than the reactions occurring in accordance with the tropism schema. To quote: “The end result of both methods is the same, 7. ¢., to get the organism away The Reactions of Planarians to Light 151 from the stimulus. In the one case it is accomplished by direct reflex without more ado; in the other, only after a considerable waste of energy in inconsequential vermiculations” (Holmes, "05a, p- 110). It is at least conceivable that under the tropism schema, as the nervous differentiation of an animal becomes more complete, the ability of the organism to interfere with and modify its machine- like responses to external stimuli might also increase, resulting in a flexibility of behavior which would present quite as much variation for natural selection to act upon as that evolved by the trial and error method. ‘This point of view by no means denies that trial and error is the usual “method of intelligence” (C. L. Morgan ’o0, p. 139). It is simply an attempt to recognize in the method of tropism also one of the possibilities of evolutionary progress in behavior and as such holding a higher position in the scale of evolutionary methods than trial and error by motor reflexes, It has been shown (p. 143) that planarian responses of an appar- ently asymmetrical character may occur as a result of symmetri- cal stimulation. Similar instances in the case of planarians have also been demonstrated by Mast (’03) with reference to thermal stimuli. This, however, is no exception to the validity of the tro- pism theory, in which asymmetrical responses result from asym- metrical stimulation. Because a planarian may make an appar- ently phototropic response when subjected to symmetrical stim- ulation, is not evidence against the supposition that the usual phototropic response is due to asymmetrical stimulation. The “wigwag’’ movements of planarians, to which repeated reference has been made in the preceding pages, resemble super- ficially the “random movements” of the earthworm as described by Holmes. They do not, however, ordinarily appear to be the basis of trial and error selection resulting in orientation, since in a majority of cases, after a worm Hales and makes wigwag movements it continues on its way without a change of direction. The movements of Bdelloura candida, as shown in Fig. 12 form an exception to ordinary planarian behavior in this respect As a rule wigwag movements are probably occasioned by a general Eeambence arising from some stimulation which thowrs 152 Herbert Eugene Walter the worm into a different physiological state. Exploring move- ments, such as these seem to be, may bring about asymmetrical stimulation, in which case the worm makes a tropic response. It was particularly noticed that when planarians received light from below, the anterior end of the body was frequently tilted back and forth as if to make it possible for the light when coming from such an unusual direction to enter the pigment cups of the eyes. The phenomenon suggested the craning of necks and bob- bing of heads among a crowd of people who are all trying to see the same object at once. Wigwag movements seem to be oftener connected with changes in the intensity of light than with changes in its direction. When the latter occur, tropic response is immediately the result. In the course of the experiments previously described wherein the worms glided from an area of one intensity of non-directive light into another it was noticed that in a majority of cases when the critical line was not crossed at right angles, no change in course occurred, even when the worm halted and made wigwag move- ments. Of course at a certain instant of any diagonal crossing of the critical line one eye must receive more stimulus than the other, in which case according to an inflexible tropism theory asymmetrical response ought to occur. But such a response does not frequently appear aa the reason for this becomes clear when it is remembered that a considerable number of responses were shown to occur which were called “latent wigwags”’ (Fig. 4, £), be- cause they failed to make their appearance until in some instances the worm had passed more than the length of its body beyond the critical line. Since, therefore, latency of response to intensity is by no means uncommon, it is evident that the brief interval of asymmetrical stimulation occurring when a worm glides diagonally into an area of different intensity is not sufficient to result in an asymmetrical response. Two conclusions, then, seem reasonable, namely, that phototaxis as related to planarians is primarily due to asymmetrical response resulting from asymmetrical stimulation, and that wigwag move- ments, together with similar apparent trial and error forms of behavior, contribute chiefly to this end, 7. ¢., to phototaxis. The Reactions of Planarians to Light 153 Summary. Orientation may occur without phototaxis. Two theories have been advanced to explain orientation and phototaxis in lower organisms, namely, the trial and error theory and that of the tropisms. “The former may be based upon “‘ motor reflexes” or upon “random movements” according to the sym- metry of the animal. The tropism theory rests upon asymmetrical response to asym- metrical stimulation. It does not necessarily depend upon the direct stimulation of the motor organs, nor is it essentially stereo- typed in its character any more than are trial and error responses by motor reflex or random movements. The tropic form of response may, and probably does, require a more complex mechanism than that which causes the motor reflex, consequently it is the form of response to be logically expected among planarians, since the motor reflex has been proven to be the form utilized by the protozoa. Tropisms, as well as trial and error movements, provide, through the modifying control of an evolving central nervous system, sufh- cient latitude of variation for natural selection to work upon in the evolution of higher forms of behavior. Asymmetrical response may, in certain cases, result from sym- metrical stimulation, but ordinarily its cause is asymmetrical stimulation. Wigwag movements are occasioned most frequently by changes in intensity, and they may result in orientation and phototaxis by assisting an organism to secure asymmetrical stimulation. Latency of reaction accounts for some of the failures in orien- tation which often occur even when asymmetrical stimulation is acting upon an organism. Finally, the orientation and phototaxis of planarians is more consistently explained by the theory of tropisms than by the theory of trial and error. 3. ADAPTATION. It remains, finally, to inquire how far the reactions of planarians to light are adaptive; that is, how far the response to light is “of 154 Herbert Eugene Walter such a kind that it better insures the existence of the individual, or of the race” (T. H. Morgan ’o03, p. 1). It is evident that the generally negative character of the reac- tions of planarians to light indicates a tendency on the part of these worms to reduce as much as possible the amount of light stimulation received or to avoid it altogether. The rigor effects of excessive stimulation furnish evidence also that light is a factor in a planarian’s environment which it finds unavoidable and unwelcome and to which it is adapted only in a negative fashion. In fact the vague distinction separating “lower” from “higher” animals consists largely in the ability of higher animals to assume an active aggressive rather than a passive defensive relation toward the factors making up their environment. For example, the evolution in animals of the visual organs, which in the planarians is only inceptive, enlarges the possible range of photic responses until light becomes an essential factor in an animal’s environment, contributing largely to its welfare by enabling it to see its food, to avoid its enemies and to select its mates. It is plain that light plays no such important part in the activities of planarians, for, as has already been pointed out, light per se is not essential to pla- narians, since they are known to ihe successfully in dark caves. Moreover, so far as known, light does not influence the regenera- tive or reproductive processes of planarians in any way whatso- ever. The formation of pigment may perhaps be regarded as an adaptation to light conditions, inasmuch as anna possessing pigment are thereby shielded to a certain degree from excessive stimulation. With reference to activities connected with nutrition and repro- duction, planarians are not dependent upon light stimulation. They are otherwise equipped, since they doubtless find their food by chemotactic means and avoid whatever enemies they may have, not aggressiv ely nor activ ely by retreating from visible foes but malen in a passive way by remaining Concealed from enemies that might see them. They have no organs of defense but survive by escap- ing attention. In this sense their negative phototaxis may be regarded as of protective value and consequently adaptive. Furthermore, the geographical distribution of fresh water pla- The Reactions of Planarians to Light 155 narians has been shown by Borelli (’93) and Wilhelmi (’04) to be chiefly dependent upon temperature and almost not at all upon the amount of illumination to which they are subjected. Voigt (’04) noticed that worms when hungry may be seen wandering about even in patches of bright sunlight with apparent disregard of light. This seems to be a case of the light reactions becoming over- balanced by other responses. Summary. Light is not an essential factor’ in planarian activities, since the behavior necessary to the welfare of the individual and the race is mainly referable to other factors. A planarian’s response to light is of a passive character, which may have an adaptive significance only in so far as its phototaxis tends to conceal the worm from its enemies. The presence of pigment may also be regarded as an adaptive condition induced by the animal’s relation to light. The evolution of the photoreceptive apparatus of the planarian has not reached the degree of differentiation necessary to enable it to secure for itself such adaptations to the factor of light in its environment as would make aggressive activity possible to it in a manner characteristic of higher animals. VII BripLlioGRAPHY ARLOING, S., ’87a—| Letter on Mechanism of Destruction of Microbes by Light. | Ann. Inst. Pasteur., vol. i, pp. 594-596. *87b—Les spores du Bacillus anthracis sont réellement tuées par la lumiére solaire. Comp. Rend. Acad. Sci., Paris, tom. 104, pp. 701-703. BarDEEN, C. R., ’ora—On the Physiology of the Planaria maculata with especial Reference to the Phenomena of Regeneration. Amer. Jour. Physiol., vol. v, no. 1, pp. 1-55. ‘otrb—The Function of the Brain in Planaria maculata. Amer. Jour. Physiol., vol. v, no. 3, pp. 175-179. BartHELeémy, A., °84—Sur la physiologie d’une Planaire verte (Convoluta Schultzii). Comp. Rend. Acad. Sci., Paris,tom. 99, pp- 197-200. Beer, T., ’01—Ueber primitive Sehorgane. Wiener klin. Wochenschr., 1gor, Nr. 11-13. 156 Herbert Eugene Walter Bercer, F. W., ’oo—Physiology and Histology of the Cubomeduse. Mem. Biol. Lab., Johns Hopkins Univ., vol. iv, pt. 4, pp. 1-84, pl. 1-3. Birce, E. A., ’97—The Vertical Distribution of the Limnetic Crustacea of Lake Mendota. Biol. Centralbl., Bd. 17, Nr. 10, pp. 371-374. Boun, G., ’03a—Sur les mouvements oscillatoires des Convoluta roscoffensis. Comp. Rend. Acad. Sci., Paris, tom. 137, pp. 576-578. ’o3b—Sur le phototropisme des Artiozoaires supérieurs. Comp. Rend. Acad. Sci., Paris, tom. 137, pp. 1292-1294. ’03c—Les Convoluta roscoffensis et la théorie des causes actuelles. Bull. Mus. Hist. Nat., Paris., tom. 9., pp. 352-364. Bore ut, A., ’93—Osservazioni sulla Planaria alpina (Dana) e catalogo dei Den- droceli d’aqua dolce trovati nell’ Italia del Nord. Boll. Mus. Zool. ed Anat. comp. R. Univ. Torino, vol. viii, no. 137, pp. T—03. Bucuner, H., ’92—Ueber den Einfluss des Lichtes auf Bakterien. II. Mittheilung. Centralbl. f. Bakteriol. u. Parasitenk., Bd. 12, No. 7, pp. 217-219. CarpENTER, F. W.,’05—The Reactions of the Pomace Fly (Drosophila ampelophila Loew) to Light, Gravity and Mechanical Stimulation. Amer. Nat., vol. xxxix, no. 459, pp. I157—-I7I. CarriERE, J., ’82—Die Augen von Planaria polychroa Schmidt und Polycelis nigra Ehrb. Arch. f. mikr. Anat., Bd. 20, pp. 160-174, Taf. 9. CurcukorF, G. D.,’92—Recherches sur les Dendrocoeles d’eau douce (Triclades). Arch. de Biol., tom. 12, pp. 435-560, pl. 15-20. Coun, F., ’53—Ueber eine Gattung aus der Familie der Volvocinen. Zeit. f. wiss. Zool., Bd. 4, pp. 77-116. ’64—Ueber die Gesetze der Bewegung mikroskopischer Thiere und Pflanzen unter Einfluss des Lichtes. Jahresber. Schles. Gesell- sch. f. vaterl. Cultur, Bd. 42, pp. 35-36. Cottiyn, A., ’91—Mittheilung uber Planaria alpina (Dana). Sitzungsber. Gesellsch. naturf. Freunde, Berlin, Jahrg. 1891, Nr. 9, pp. 177-180. Curtis, W. C., ’°02—Life History and Reproduction of Planaria maculata. : Proc. Boston Soc. Nat. Hist., vol. xxv, pp. 515-559. DatyeELL, J. C., ’14—Observations on Some Interesting Phenomena in Animal Physiology, Exhibited by Several Species of Planaria, illustrated by colored figures of living animals. Edinburgh, 8°, xi + 151 Pp:s tpl. ’*53—The Powers of the Creator Displayed inthe Creation; or, Observa- tions on Life Amidst the Various Forms of the Humbler Tribes of Animated Nature. London, 4°, vol. 11, 359 pp., 46 pl. The Reactions of Planarians to Light 157 DanpriEu (DE Carcassonne), ’88—Influence de la Lumiére dans la Destructions des Bactéries; pour servir a Etude du “Tout a l’Egout.” Ann. d’Hyg. pub. et de Méd. lég., Paris, sér. 3, tom. 20, no. 5, pp. 448-451. Darwin, C., ’44—Brief Descriptionsof Several Terrestrial Planariz and of Some Marine Species, with an Account of their Habits. Ann. and Mag. Nat. Hist., vol. xiv, pp. 241-251, 1 pl. *81—The Formation of Vegetable Mold through the Action of Worms, with Observations on their Habits. New York, 12°, 326 pp. Davenport, C. B., anp Cannon, W. B., ’97—On the Determination of the Direction and Rate of Movement of Organisms by Light. Jour. of Physiol., vol. xxi, no. I, pp. 22-32. Dearsorn, G. V. N., ’o0o—Notes on the Individual Psycho-physiology of the Crayfish. Amer. Jour. Physiol., vol. iii, no. 9, pp. 404-433. Dexace, Y., ’86—Etudes histologique sur les Planaires rhabdocceles acceles (Convoluta Schultzii O.Schm.). Arch. Zool. Expér. et Gén., sér. 2, vol. iv, pp. 109-144, pl. 5-6. Downes, A., AND Brunt, T. P., ’77—Researches on the Effect of Light upon Bacteria and other Organisms. Proc. Roy. Soc., London, vol. xxvi, pp. 488-500. *78—On the Influence of Light upon Protoplasm. Proc. Roy. Soc., London, vol. xxvii, pp. 199-212. Ducraux, E., ’85a—Influence de la lumiére du soleil sur la vitalité des germes de microbes. Comp. Rend. Acad. Sci., Paris, tom. 100, pp. 119-121. ’85b—Influence de la lumiére du soleil sur vitalité des micrococcus. Comp. Rend. Acad. Sci., Paris, tom. 101, pp. 395-3098. *90—Action de la Lumiere sur les Microbes. ‘ Revue critique. Ann. Inst, Pasteur, vol. iv, pp. 792-800. Duces, A., ’28—Recherches sur l’organization et les moeurs des Planariées. Ann. Sci. Nat., tom. 15, pp. 139-183. Enceitmann, T. W., ’79—Ueber Reizung contraktilen Protoplasmas durch plotzliche Beleuchtung. Arch. f. ges. Physiol., Bd. 19, pp. 1-7. *83—Bacterium photometricum. Ein Beitrag zur vergleichenden Physi- ologie des Licht-und Farbensinnes. Arch. f. ges. Physiol., Bd. 30, pp. 95-124, Taf. 1. Famintzin, A., ’67—Die Wirkung des Lichtes auf Algen und einige andere ihnen nahe verwandte Organismen. Jahrb. f. wiss. Bot. (Pringsheim), Bd. 6, pp. 1-48, Taf. 1-3. ; FranpsEn, P., ’01—Studies on the Reaction of Limax maximus to Directive Stim- uli. Proc. Amer. Acad. Arts and Sci., vol. xxxvii, pp. 185-227. 158 Herbert Eugene Walter Fiuner, N., ’o6—Notizen zur Biologie von Convoluta roscoffensis Graff. Biol. Centralbl., Bd. 26, Nr. 1, pp. 24-26. GamB LE, F. M., anp KEEBLE, F., ’03—The Bionomics of Convoluta roscoffensis, with Special Reference to its Green Cells. Quart. Jour. Micr. Sci., vol. Ixvii, pp. 363-431, pl. 30-31. Geppes, P., ’79—Observations on the Physiology and Histology of Convoluta Schultzii. Proc. Roy. Soc., London, vol. xxvill, pp. 449-457- Giss_Ler, C. F., ’°82—A Marine Planarian and its Habitation. Amer. Nat., vol- XVI, 00: Wy pp: §2=53;, gs 1-2. Graber, V., °83—Fundamentalversuche iiber die Helligkeits- und Farbenemp- findlichkeit augenloser und geblendeter Thiere. Sitzungsb. d. Wien, Akad., Bd. 87, pp. 201-236. GrarF, L. von, ’84—Zur Kenntnis der physiologischen Function des Chloro- phylls im Thierreich. Zool. Anz., Bd. 6, Nr. 177, pp. 520-527- Groom, T. T., unp Logs, J.,’g0—Der Heliotropismus der Nauplien von Balanus perforatus und die periodischen Tiefenwanderungen pelagischer Tiere. Biol. Centralbl., Bd. 10, Nr. 5, pp. 160-177. Haper.anpt, G., ’91—Ueber den Bau und die Bedentung der Chlorophyllzellen von Convoluta roscoffensis. Appendix (pp. 75-90) to L. von Graff, Die Organization der Turbellaria accela. Leipzig. 4°, Holzsch. u. go Taf. Harper, E. H., ’05—Reactions to Light and Mechanical Stimuli in the Earth- worm Pericheta bermudensis (Beddard). Biol. Bull., vol. x, no. 1, pp. 17-34. Hesse, R., ’96—Untersuchungen iiber die Organe der Lichtempfindung bei niederen Thieren. I. Die Organe der Lichtempfindung bei Lum- briciden. Zeit. f. wiss. Zool., Bd. 61, pp. 393-419, Taf. 20, 1 Fig. 97—Untersuchungen iiber die Organe der Lichtempfindung bei niederen Thieren. I]. Die Augen der Plathelminthen, insonder- heit der tricladen Turbellarien. Zeit. f. wiss. Zool., Bd. 62, pp- 527-582, Taf. 27 u. 28, 3 Fig. Hopeg, C. F., anp A1kens, H. A.,’95—The Daily Life of a Protozoan: A Study in Comparative Psycho-physiology. Amer. Jour. Psychol., vol. vi, no. 4, pp. 524-533. Hoae, C. ’97—On Bipalium kewense, Moseley. Trans. Nat. Hist. Soc., Glasgow, n. s., vol. v, pt. I, pp. 53-54. Houmes, S. J. ’o1—Phototaxis in the Amphibia. Amer. Jour. Physiol., vol. v, no. 4, pp. 211-234. ’0o3—Phototaxis in Volvox. Biol. Bull., vol. iv, pp. 319-326. The Reactions of Planarians to Light 159 Homes, S. J., ’05a—The Selection of Random Movements as a Factor in Photo- taxis. Jour. Comp. Neurol. and Psychol., vol. xv, no. 2, pp. 98-112. ’o5b—The Reactions of Ranatra to Light. Jour. Comp. Neurol. and Psychol., vol. xv, pp. 305-349- Hott, E. B., and Ler, F. S., ’01—The Theory of Phototactic Response. Amer. Jour. Physiol., vol. iv, no. 9, pp- 460-481. Irma, I., ’84—Untersuchungen tiber den Bau und die Entwickelungsgeschichte der siisswasser Dendrocoelen (Tricladen). Zeit. f. wiss. Zool., Bd. 40, pp. 359-464, Taf. 20-23. JAnincuin, E., ’96—Beitrage zur Kenntnis des Turbellarien Auges. Zeit. f. wiss. Zool., Bd. 62, pp. 250-288, Taf. 10-11, 7 Texthg. Jamieson, J., 82—The Influence of Light on the Development of Bacteria. Nature, vol. xxvi, pp. 244-245. Janowsk1, T., ’90—Zur Biologie der Typhusbacillen. Die Wirkung des Sonnen- lichts. Centralbl. f. Bakt. u. Parasitenk., Bd. 8, Nr. 6, pp. 167- 172; Nr. 7, 193-199; Nr. 8, 230-234; Nr. 9, 262-266. Jennies, H. S.,’04a—The Theory of Tropisms. Carnegie Inst. of Washington, Pub. no. 16, pp. 89-107. ’o4b—Physiological States as Determining Factors in the Behavior of Lower Organisms. Carnegie Inst. of Washington, Pub. no. 16, pp- 109-128. ’o4c—The Method of Trial and Error in the Behavior of Lower Organ- isms. Carnegie Inst. of Washington, Pub. no. 16, pp. 237-252. KennEL, J. von, ’88—Untersuchungen an neuen Turbellarien. Zool. Jahrb. Abt. f. Anat. und Ont., Bd. 3, pp. 447-486, Taf. 18-19. Lana, A., ’84—Die Polycladen (Seeplanarien) des Golfes von Neapel und der angrenzende Meeresabschnitte. Fauna u. Flora d. Golfes yon Neapel. XI. Monographie. x+688 pp. 39 Taf. Leipzig. Lennert, G. H., ’91—Beobachtungen an Landplanarien. Arch. f. Naturg., Jahrg. 57, pp- 306-350. Lemy, J.,’58—Note on Finding Rhincodemus sylvaticus at Night. Proc. Acad. Nat. Sci., Philadelphia, 1858, pp. 171-172. Linu, F. R., ’°o1r—Notes on Regeneration and Regulation in Planarians. Amer. Jour. Physiol., vol. vi, no. 2, pp. 129-141. Logs, J., ’90—Der Heliotropismus der Thiere und seine Uebereinstimmung mit dem Heliotropismus der Pflanzen. Wirzburg, 8°, 118 pp. ’93a—On the Influence of Light on the Periodical Depth-migrations of Pelagic Animals. Bull. U. S. Fish Com., vol. xiii, pp. 65-68. 160 Herbert Eugene Walter Logs, J., ’93b—Ueber kiinstliche Umwandlung positiv heliotropischer Thiere in negativ heliotropische und umgekehrt. Arch. f. ges. Physiol., Bd. 54, pp. 81-107, 6 Holzsch. *94—Beitrage zur Gehirnphysiologie der Wiirmer. Arch. f. ges. Physiol., Bd. 56, pp. 247-269, Fig. 1-4. Mast, S. O., ’03—Reactions to Temperature Changes in Spirillum, Hydra and Fresh-Water Planarians. Amer. Jour. Physiol., vol. x, no. 4, pp- 165-190. Maupas, E., ’87—Sur la puissance de multiplication des Infusories ciliés. Comp. Rend. Acad. Sci., Paris, tom. 104, pp. 1006-1008. Morean, C. L., ’0o—Animal Behavior. London, 8°, 344 pp. Morean, T. H., ’03—Evolution and Adaptation. New York, 8°, x+470 pp- Mose ey, H. N., ’74—On the Anatomy and Histology of the Land Planarians of Ceylon; with Some Account of their Habits, and a Description of Two New Species, and with Notes on the Anatomy of Some European Aquatic Species. Phil. Trans. Roy. Soc. London, vol. clxiv, pp. 105-171, pl. 10-15. NaceL, W. A., ’94—Beobachtungen iiber den Lichtsinn augenloser Muscheln. Biol. Centralbl., Bd. 14, Nr. 11, pp. 385-390. Ortmanns, F., ’92—Ueber die photometrischen Bewegungen der Pflanzen. Flora, Bd. 75, pp. 183-266, Taf. 4. PansInI, ’89—Action de la Lumiere solaire sur les micro-organismes. Ann. Inst. Pasteur, tom. 3, pp. 686. Parker, G. H., ’02—The Reactions of Copepods to various Stimuli, and the Bearing of this on Daily Depth Migrations. Bull. U.S. FishCom., IgOI, pp. 103-123. ’03—The Phototropism of the Mourning-cloak Butterfly, Vanessa antiopa Linn. Mark Anniversary Vol., pp. 453-469, pl. 33- Parker, G. H., anp Arkin, L., ’or—The Directive Influence of Light on the Earthworm Allolobophora fcetida Sav. Amer. Jour Physiol., vol. Vv, NO. 3, pp. 151-157 Parker, G. H., anp Burnett, F. L., ’00o—The Reactions of Planarians, with and without Eyes, to Light. Amer. Jour. Physiol., vol. iv, no. 8, Pp. 373-385, fig. 1-4. Peart, R., ’03—The Movements and Reactions of Fresh-water Planarians: A Study of Animal Behavior. Quart. Jour. Micr. Sci., vol. xlvi, pp. 50957 74: Peart, R., and Cote, L. J.,’02—The Effect of very Intense Light on Organisms. Michigan Acad. Sci., Third Report, pp. 77-78. The Reactions of Planartans to Light 161 Rant, E., ’01—Ueber den Phototropismus einiger Arthropoden. Biol. Centralbl., Bd. 21, Nr. 3, pp. 75-86. Raum, J., °89—Der gegenwartige Stand unserer Kenntnisse iiber den Einfluss des Lichtes auf Bakterien und auf den thierischen Organismus. Zeit. f. Hyg., Bd. 6, pp. 312-368. Roux, E., ’87—De l’action de la lumieére et de |’air sur les spores de la bactéridie du charbon. Ann. Inst. Pasteur, tom. I, pp. 445-452. ScHouTEDEN, H., ’02—Le Phototropisme de Daphnia magna Straus (Crust.). Ann. Soc. ent. Belgique., tom. 46, pp. 352-362. Scuuttze, M., ’51—Beitrage zur Naturgeschichte der Turbellarien. Erste Abt. Greifswald, 4°, 78 pp. 7 Taf. SeKeERA, E., ’03—Einige Beitrage zur Lebensweise von Vortex helluo (viridis M. Sch.) Zool. Anz., Bd. xxvi, no. 708, pp. 703-710. Situ, A. C., ’02—The Influence of Temperature, Odors, Light and Contact on the Movements of the Earthworm. Amer. Jour. Physiol., vol. vi, no. 7, pp. 459-486. StrasBurGeRr, E., ’78—Wirkung des Lichtes und der Warme auf Schwarmsporen. Jena. Zeit., Bd. 12, pp. 551-625. Tow te, E. W., ’0o—A Study in the Heliotropism of Cypridopsis. Amer. Jour. Physiol., vol. ili, no. 8, pp. 345-365. Tynva tL, J., ’778—Note on the Influence Exercised by Light on Organic Infusions. Proc. Roy. Soc., London, vol. xxviii, pp. 212-213. Verworn, M., ’o1—Allgemeine Physiologie. 2te Aufl. Jena, 8°, 631 pp. Voict, W., ’04—Ueber die Wanderungen der Strudelwiirmer in unseren Gebirgs- bachen. Verh. naturh. Vereins preuss. Rheinlande, Westfalens und Regierungsbez. Osnabriick, Jahrg. 61, pp. 102-178. Watter, H. E., ’06—The Behavior of the Pond Snail; Lymnzus elodes Say- Brooklyn Inst. Arts and Sci., Cold Spring Harbor Monographs, VI, 35 pp- Warp, H. M., ’94—Further Experiments on the Action of Light on Bacillus anthracis and on the Bacteria of the Thames. Proc. Roy. Soc., London, vol. lvi, pp. 315-374. Wuirman, C. O., ’99—Animal Behavior. Wood’s Hole Lectures for 1898. Boston, pp. 285-338. WitueEmt, J., °o4—Beitrage zur Kenntnis der Verbreitung und Biologie der Siisswassertricladen. Zool. Anz., Bd. xxvii, Nr. 11, pp. 355-365; Nr. 12, 369-375. Wi.eM, V.,’91—La vision chez les gasteropodes pulmonés. Comp. Rend. Acad. Sci., Paris, tom. 112, pp. 247-248. 162 Herbert Eugene Walter Woopwortn, W. McM., ’97—Contributions to the Morphology of the Turbel- laria: I]. On some Turbellaria from Illinois. Bull. Mus. Comp. Zool. Harvard Coll., vol. xxxi, pp. 1-16, pl. 1. Yerkes, R. M., ’99—Reaction of Entomostraca to Stimulation by Light. Amer. Jour. Physiol., vol. iii, no. 4, pp. 157-182. ’03.—A Study of the Reactions and Reaction Time of the Medusa Gonionema Murbachii to Photic Stimuli. Amer. Jour. Physiol., vol. 1x, no. 5, pp. 279-307. ZELENY, C., '05—The Rearing of Serpulid Larva with Notes on the Behavior of the Young Animals. Biol. Bull., vol. viii, no. 5, p. 308. REGENERATION IN COMPOUND EYES OF CRUSTACEA BY MARY ISABELLE STEELE, M.A. Witu Sixteen Pirates AND Two Ficures IN THE TExT SRM TEC OCILICESO MI cice svete ec iozarai ores eras teysle ate royens Tenctesel sve ie sie sc\siers, oe stay 0ia,oye0u oh, arare/are\s,\ur0 ofove\aseisherslens 164 MibmmIVeth ods andtmaterial ses sctra cen cicrea ns etoietelsievers Totelefer cece sete 2's. ohclelovesc nrcis, yeloioieteretelsies eloierele 165 iMit 'iheagaenlcuhi Ge Pes sas aoecpacpede aso conan ad soccer e EEE EC erOE Seo cuSadaneaneacEs 168 IV The preliminary regenerative processes...... 26200. 2. 0 cee sence ese ete ese c teers eee ees 172 A Healing of the wound and formation of new cuticle.............-.--.0.00-02.0-055 173 Bae Removaljotthenny ured tissue cjasiajis rete aie\a/aicisvace ete eyeieljs.o sjoim Aiaaisieie) s/s cfm ie/njelsinis[n elelaielais 243 Tue Journat or ExreriMENTAL ZOOLOGY, VOL. V, NO. 2. 164 Mary Isabelle Steele I INTRODUCTION The regeneration of the Crustacean compound eye has been made a subject of observation by a number of investigators. But for the most part end results alone have been described. Very few details of the processes involved in regeneration have been given. ‘To give an accurate description of the histogenesis of the regenerated structures is the chief aim of this paper in the belief that it will contribute something toward a more accurate under- standing of the more general problem of regeneration. The problem has separated itself into three main divisions: first, the regeneration of a functional eye; second, a search for the causes of no regeneration and abnormal regeneration, and third, observations upon the heteromorphic regeneration which may follow the removal of the entire eye. Herbst (’96, 00) and Morgan (’98) have made the principal observations upon the regeneration of the Crustacean compound eye. But so far as the particular phases dealt with in this paper are concerned Herbst’s descriptions are not sufficiently detailed to be of especial assistance in this work. Herbst’s observations upon the regenerated heteromorphic structures are somewhat more extensive and in some respects furnish an excellent basis for comparison with the results to be discussed in this paper. For the most part, however, where results have been compared with the work of others the comparison is made between the regenera- tive and embryonic development and between normal and regen- erated structures. The terminology employed is in great part that used by Parker (91) in his work upon the compound eyes of Crustacea, it being the terminology in most general use. A few minor deviations have been made but no wholly new terms have been introduced. The following plan has been followed in this paper. First, a brief description of the normal adult eye is given to furnish a basis for comparison. ‘This is followed by a description of the prepara- tory regenerative processes which in reality constitutes one of the most important phases of the subject. Then the regenerative processes proper are described under the three main divisions already suggested. Regeneration in Compound Eyes of Crustacea 165 The experimental part of the work was begun at the University of Pennsylvania. Observations upon living material were also carried on during two summers at Woods Hole, Mass., and one summer at Cold Spring Harbor, L. I. The greater part of the detail work of examining the preserved material has been done at the University of Missouri during the present year. In closing I wish to express my thanks to Dr. E. F. Phillips and Dr. D. B. Casteel for the care of experiments; to Dr. E. G. Conklin, University of Pennsylvania; Dr. C. B. Davenport, Carnegie Institute at Cold Spring Harbor, and Dr. George Lefevre and Dr. W. G. Curtis, of the University of Missouri, for their interest and valuable suggestions during the course of the work. II Marerrat anp Metuops The small hermit crab, Eupagurus longicarpus, the common shrimp, Palzmonetes vulgaris, and the sand shrimp, Crangon vulgaris, afford the greater part of the material used in the series of experiments to be described in this paper. For comparison two species of crayfish, Cambarus virilis and C. gracilis, a species of fresh water Ascellus, the common (wood-louse), Oniscus, and the fresh water Gammarus were used. Other emacs also were experimented upon but since no decisive results were obtained they need not be considered here. The work has been confined chiefly to the eyes of the forms used although experiments upon the appendages, particularly the . antennz, were conducted at thesame time. “These were, however, largely for the purpose of comparing relative rates of regeneration of the different parts, especially the rate of regeneration of the appendages as compared with that of the eyes. The experiments upon the eyes consisted in either the removal of a part of the eye or of the whole eye. ‘The part removed varied greatly in the different series of experiments and more or less in individuals of the same series. A limited number of experiments upon Palzemonetes included the removal of both eyes or the removal of one eye with a part of the brain; these operations in most cases resulted fatally. The effect of splitting the eye was 166 Mary Isabelle Steele also tried upon two series of Palamonetes, each series was com- posed of a considerable number of individuals. Results, however, were not particularly different from those obtained after removing a part of the eye. A total of 600 Palamonetes and hermit crabs had either one or both eyes operated upon. A much smaller number of Crangon and crayfish were used. No accurate account of the Ascellus, Oniscus and Gammarus was kept. More than 50 per cent of the Palamonetes and hermit crabs died immediately, or within a short time, after the operation, many of them dying within a few minutes after the eye was injured. Of the survivors about 58 per cent lived through one or more moults. Forty-two Crangon had the eye operated upon and of these one died of the operation. The crayfish used were for the most part C. gracilis, measuring from 12 to 15 mm. in length, that had moulted but once after hatching. Palamonetes and Crangon moult once about every ten days to three weeks. The hermit crabs moult much less frequently, often but once in two or three months. The hermit crabs regenerate, however, as rapidly as Palemonetes or Crangon. Considerable difficulty was experienced in keeping the animals alive and in keeping individual records. Finally the plan of keeping each animal in a separate finger bowl was adopted. This method was fairly satisfactory except in very warm weather. Then the water became warm and unless it was changed often the animals soon died. Chopped-up bits of clam were fed to them two or three times a week. Great care had to be used in warm weather for the water became foul, if the food was left in more than three or four hours, and caused the death of the animals. In spite of all precautions various accidents occurred which resulted in the death of promising material. Twice attempts were made to transfer the experiments from Philadelphia to Woods Hole or vice versa with disastrous results in each case. The failure was due in part no doubt to the extreme warm weather. For although every known precaution was taken most of the animals died within twenty-four hours. A number of the ordinary fixing fluids were used to preserve Regeneration in Compound Eyes of Crustacea 167 material. Among those most frequently used were Fleming’s osmic fixative, Perenyi’s fluid, Kleinenberg’s picro-sulphuric, picro-aceto-sulphuric, Petrunkewitsch’s fluid and alcohol acetic. Other fluids also were used and boiling water was tried. ‘The best results were obtained from picro-aceto-sulphuric and Perenyi’s fluid. In any case it is difficult to obtain a fixative that does not shrink the inner tissues from the chitin, the regenerated tissues being much more easily affected in this respect than normal tissues. The embedding was done altogether in hard parafiine 54° to 58° C. melting point. It was necessary to embed the material for a long time in order to cut it without tearing the chitin from the softer tissues. [he most satisfactory infiltration was obtained by placing the objects first in equal parts of oil and paraffine, leaving them on top of the water bath over night and then replacing the oil and paraffine with pure parafiine, leaving them on top of the water bath from eight to ten hours longer. Finally they were put in the parafiine bath from one to two hours. Even after the most thorough infiltration it was well nigh impossible to obtain complete series of good sections, because of the difficulty of cutting through the different textures of the material. In dehydrating preparatory to embedding, cedar or bergamot oil was usedin preference to xylol as these oils made the tissues less brittle. The chief stains employed were Fleming’s triple stain and Heidenhain’s iron hematoxylin. Various counter stains were used with the iron hematoxylin but the most generally satisfactory were acid fuchsin and orange G. In most instances no attempt was made either to soften the chitin or to remove the pigment before the eyes were sectioned. As a rule, however, the material was fixed shortly after a moult so that the chitin was as soft as could be obtained. Any sort of a reagent used to soften the chitin seemed to be more or less inju- rious to the softer parts, particularly the regenerating tissue. When the chitin was removed, however, 2 per cent nitric acid in 70 per cent alcohol was found to be most satisfactory. It was generally disadvantageous to remove the pigment because this destroyed many of the landmarks both in respect to the regener- 168 Mary Isabelle Steele ating tissue and the condition of the remaining old parts. Sections were sometimes depigmented on the slide. For this Mayer’s chlorine method was used. Table I gives in brief a record of the individuals operated upon, the character of the operation and the end results. III Tur Normat Aputt EYE Before taking up the discussion of the regeneration of the eye it will be perhaps well to give a brief description of the structure of Text Fig. 1 The lines, ab, a-c, a-d, a-e, a-f, a-~g and a-h represent approximately the different levels at which a part or whole of the eye has been removed. When the cut was at level as low as a-e a part of the optic ganglion was usually involved. In Palemonetes the eye never regenerated when the cut came as low as the level a-e. Hermit crabs may regenerate an eye from the level a-f and a hetero- morphic appendage from the level a-h. the normal eye. The nature and extent of the operations and the subsequent changes in the eyes will then be more easily under- stood. The eye of Palamonetes has been concisely described by Parker (’97) and the following description is adapted from his. It is to be noted that the eye structure of the three forms experi- mented upon is practically the same. Regeneration in Compound Eyes of Crustacea 169 TABLE I. PALAEMONETES | No. of | Died Lived | Showed Series | individ- Operation | Date from and _regenera- | uals | operation| moulted | tion I 16 part of cornea removed May 10 12 4 ° TS || £2. part of cornea removed july II 8 bs I Ti 38 part of cornea removed July 20 29 3 | 8 IV 40 part of cornea removed July 28 28 II 8 Vv 15 part of cornea removed July 30 3 12 ° VI 16 part of cornea removed August I 12 z: I VII 15__| partof cornea removed August a 10 2 3 Vill 12 part of cornea removed November 5 5 5 ° Ix ez 7am part of cornea removed | November 9 8 9 ° ibe ||) 4G entire cornea removed | March 10 4 7 ° |) 38 entire cornea removed July 20 8 9 I Ty, || 30 entire cornea removed July 29 28 2) ° v | 30 entire cornea removed July 30 20 9 ° VI 13 entire cornea removed August 4 8 5 ° VII 12 entire cornea removed December 30 II I ° entire eye removed | | I 12 both eyes | January I 4 6 ° II 12 entire eye | March 5 9 Hie | ° 100% 1) ae both eyes March 5 12 On eo DV; 18 both eyes March 10 10 5 ° Vv 45 entire eye April 19 12 20 ° I 20 eye split March 5 5 12 3 I 28 eye split May 24 nO) Ee | ° | HERMIT CRABS ee | ara, | cornea removed March 25 2 8 | 3 I 12 | cornearemoved May 26 2 7 | 2 TIL 25 | cornearemoved July 9 2) eer O nel em nis I 15 entire eye removed | May 26 I | 72.) | I I 25 entire eye removed July 9 | Aan era 9 Ut 8 entire eye removed October 16 oe || fo) IV 12 entire eye removed November 27 | 2 6 | ) CRANGON Ty) 20 part of cornea removed August 4 | ° 19 10 rE 22 eye removed August 4 | I 19 I Note.—Removing the entire corneal portion of any of the forms was always accompanied by the removal of part of the optic ganglion, In many cases as much as half of it. It frequently resulted that all of the corneal portion subsequently degenerated after a part of it had been removed, and it also frequently happened that the removal of the entire corneal portion resulted in the loss of the whole eye stalk, consequently the above table can be used only as a very general indication of the character of the operation. 170 Mary Isabelle Steele The compound eye may be regarded as that part of the optic apparatus contained in the eye stalk. It consists of a large num- ber of ommatidia occupying the distal end of the stalk and a series of four ganglia which extend through the axial portion of the stalk. This series of ganglia for present purposes may be regarded as a compound ganglion composed of four rather distinct sections or ganglionic masses united to each other by nerve fibers. The ommatidia are connected with the distal section of the optic ganglion by the retinular nerve fibers. ‘The optic nerve passes inward from the proximal section of the ganglion to unite the eye with the brain. “The basement membrane forms a sort of partition between the ommatidia and the optic ganglion. ‘The transparent chitinous covering over the ommatidial region is known as the cornea. Each ommatidium is composed of the following cells: two cor- neal hypodermal cells, four cone cells, two distal retinular cells, eight proximal retinular cells one of which is rudimentary, and a variable but small number of accessory pigment cells. Black pigment granules are contained in both proximal and distal retinule and are found only in these cells. “Uhe yellowish pigment is confined exclusively to the accessory pigment cells. “The differ- ent cells enumerated above give rise to the structures that consti- tute a complete ommatidium. ‘The two corneal hypodermal cells secrete the square corneal facet which covers the outer surface of the ommatidium. Immediately beneath the corneal hypodermal cells is the crystalline cone formed by the four cone cells. The nuclei of these cells are located in their distal ends. ‘The main body of the cone appears as a dense hyaline secretion. Proximally the cone is less dense in structure and tapers to a slender stalk lying between the coneandrhabdom. ‘The rhabdom, according to my observations, is a swollen spindle-shaped structure proximal to the inner ends of the cone cells. ‘The distal retinular cells lie near the inner end of the cone; the proximal retinule surround the distal end of the rhabdom. ‘The proximal processes of the retin- ulz extend over the rhabdom and pass through the basement membrane as the retinular nerve fibers to enter the optic ganglion below. The accessory pigment cells lie both above and below the basement membrane. Regeneration in Compound Eyes of Crustacea 7 Before closing the description of the normal eye mention should be made of another point. In young individuals in each of the species examined in this series of experiments there is present a growth zone in the ommatidial region. From this zone the num- ber of ommatidia is increased as the animal grows older and increases in size. In longitudinal sections cut in a horizontal plane this zone is apparent on the inner edge of the eye as a narrow band of elongated cells situated above the basement membrane and between it and the completely developed ommatidia. In some instances partially differentiated ommatidia can be recog- nized in this growth zone. This zone has been mentioned by Parker and others. It is briefly described in this connection now because it is probable that in some instances ommatidia that have apparently regenerated have in reality developed from the growth zone. The terminology used in the above description and in the dis- cussion of the regenerating eye is that used by Parker. ‘The series of optic ganglia described by him as occupying the eye stalk, however, have been referred to in this paper as different divisions of a single ganglion, it being thought that the matter could be treated with less confusion in this way. Also, the distal portion of the eye stalk, called the retina by Parker, is referred to here as the ommatidial portion. It is composed of a large number of individual ommatidia and the more general use of the term, retina, does not imply all the structures composing the ommatidia. The structure of the rhabdom as described in the present paper is not in full accord with Parker’s description. While the above description applies especially to Palazmonetes it is sufficiently accurate for the other forms described in this work to serve all purposes. The most marked differences in the struc- tures of the eye of Palamonetes and of the other forms used are as follows: In Crangon and in crayfish the eye stalk is much shorter and proportionately greater in diameter than in Palamonetes. Also, in Crangon some small glands are found located just below the basement membrane. Hermit crabs have long slender eye stalks similar to those of Palamonetes except that at the base and occupying the dorsal inner edge there is a small pointed squame 172 Mary Tsabelle Steele bearing a number of sensory hairs. (Fig. 37, 0-5q.) And finally, there are no accessory pigment cells in the hermit crab’s eye. These cells are very conspicuous in the eyes of Palamonetes. IV THe PRELIMINARY REGENERATIVE PROCESS The preparatory stages leading to the regeneration of an eye stump consist chiefly of the following processes; the healing of the wound, the removal of the injured tissues and the active prolifera- tion of new cells of a comparatively undifferentiated character. In removing any part of the eye the injury to the remaining soft, inner tissues is considerable. Especially is this true when the cut passes through the ommatidial region. Much of the tissue sur- rounding the wound is crushed and torn out of place. On this account the process of healing over the cut surface is much more difficult to follow than the healing of the wound after an antenna or leg has been removed. Before taking up the description of the preliminary regenerative process it will be perhaps of interest to give in brief the immediate effects of the operation. The death of the animal which so fre- quently follows close upon the operation seems often to be due chiefly to nervous shock. It cannot be caused by loss of blood alone for usually there is no profuse bleeding. When the eye of Palzmonetes is operated upon the animal often turns over and over ten to forty times as soon as it is released and returned to the water. Many of the animals die before they succeed in nghting themselves. Others lie upon their sides several hours after they have ceased revolving and die without showing any normal activi- ties or regaining their equilibrium. It very seldom happens that animals which whirl over and over many times after the operation ever recover from its immediate effects. “These apparently help- less motions indicate that the operation has caused the loss of equilibrium. Crayfish sometimes exhibit these uncontrolled whirling move- ments. Similar movements are also noted in fresh-water Gam- marus and in Ascellus. In Gammarus and Ascellus these move- ments are executed after the removal of the antennz or some of \ Regeneration in Compound Eyes of Crustacea 173 the mouth parts. This shows that the effects are not specific- ally connected with operations upon the eye. Whatever its ini- tial cause the effect is transmitted to the whole nervous system. In many cases the operation seemed to affect the animal more seriously when only the upper part of the eye was removed than when the entire eye was cut off. Frequently the relative number of survivors was greater in the latter than in the former case. In other instances the animal did not seem greatly affected by the operation regardless of whether the whole or a part of the eye was removed. (See Table I.) The immediate visible effects upon the eye may be briefly described as follows. As soon as any part of the corneal covering is removed or even as soon as a rent is made in it a considerable amount of the soft, viscous, inner tissue flows out through the opening. It is perhaps carried out by the escaping blood. Much of the pigmented, retinular tissue seems to escape, perhaps because it is softer and more viscous than the other tissues. After an hour or so the remaining, inner tissues are seen bulging out and above the general level of the surface. This is probably on account of the destruction of the normal tension of the tissues due to the changed pressure conditions at the wounded surface. A similar appearance is obtained when the surface injury consists of a rent torn in the cornea with a needle. A HEALING OF THE WOUND AND FORMATION OF NEW CUTICLE In a few hours after the operation most of the pigmented tissues have disappeared from the surface of the wound and the swollen surface takes on a whitish appearance. ‘This white swollen surface is apparent for several days. Not until the fourth or fifth day 1s there any sign of the characteristic red-brown crust which generally forms over wounds in Crustacea. Sections of an eye fixed six and a half hours after the operation show no definite indications of the healing of the wound. A great deal of the broken and mangled tissue lies outside the wound and hanging to the cuticle about its edges. Inside, the tissues are twisted and misshapen. At the edges of the wound there are 174 Mary Isabelle Steele slight indications that it is preparing to close over. But the quantity of material lying inside and out makes it impossible to determine what tissues are taking part in closing the wound. During the next twelve or fifteen hours the changes are still not clearly defined. ‘The interior still presents a rather badly confused mass of injured tissue. Near the edges of the wound, however, there are evidences that the hypodermis has begun to push outward to cover the cut surface. For the most part ‘the wound shows a smooth even surface which indicates that the passage outward of the injured tissues has ceased and that a sort of equilibrium had been established. The mass of tissue closing the cut seems to be made up of a few hypodermal cells and cytoplasmic strands, a considerable accumulation of blood cells and the nuclei of the breaking down tissues of the eye. Around the edges of the cut occasional strands of hypodermis with a very few nuclei can be distinguished. Sections of an eye fixed about forty hours after the operation show the beginning of crust formation. Almost the whole surface has been covered. Judging by the reaction to stains, the part which may be considered the matrix of the crust is formed by an attenuated, chitinous secretion of the hypodermis. In this matrix are embedded numerous nuclei of the injured tissues together with a great many blood cells and a few hypodermal cells. In some parts the crust is sharply marked off from the underlying tissues by a space filled with coagulated plasma. Over one part of the wound the crust is not yet fully formed. At this point hypodermal strands containing elongated flattened nuclei are seen stretched across the space still uncovered. The strands appear in two or three layers with very few nuclei in each layer. No definite centers of cell proliferation can be recognized at this time. After the crust has covered the cut surface it continues to increase in thickness for two or three days, then hardens, turns a bright reddish brown color and remains over the stump until a moult has occurred. ‘The crust takes no further active part in the healing and regenerative processes. Fig. 46 represents in a semidia- grammatic manner the crust formed over the wound in a crayfish eye about sixty hours after the operation. ‘The crust is continu- i Regeneration in Compound Eyes of Crustacea 175 ous with the inner surface of the cuticle covering the rest of the eye. There are still masses of the injured inner tissues that have been excluded by the formation of the crust clinging to its outer surface. ‘The old tissues in the interior of the eye stump have shrunk back from the crust leaving a considerable space occupied by coagulated plasma. ‘The old cuticle and the matrix of the crust both stain deeply either with orange G or acid fuchsin. No distinct cuticle can be recognized for several days, from six to eight, after the operation. In the eyes of Palamonetes that have moulted seven or eight days after the infliction of the injury a cuticle which corresponds approximately in thickness with the cuticle covering the remainder of the eye has formed over the wound. ‘This new.cuticle is much looser in texture than the old cuticle. Regeneration does not take place so rapidly in crayfish as in the marine forms examined so that a new cuticle is somewhat longer in forming. Frequently a considerable space intervenes between the overlying crust and the cuticle which has formed beneath it. This is probably due to the recovery of the tissues of the stump from their early swollen condition during which time they were gorged with blood and occupied more than their normal amount of space. It is not unusual to find considerable spaces between different layers of the new cuticle as if a shrinking of the tissues had taken place during the process of forming the new layers of cuticle. The shrinking of the interior tissues without doubt also accounts in part for the folds and wrinkles which often appear in the cuticle over the wound. The secretion of the new cuticle which grows over the wounded surface begins some little distance back from the cut edges of the old cuticle and is continuous with its inner layers. Fig. 47 1s a semidiagrammatic representation of the relation of the old and the new cuticle and the exclusion of the broken down tissue by the development of the new cuticle beneath it. Only eight nuclei appeared beneath the new cuticle in the section as shown in Fig. 47. This figure is taken from an eye that had been injured by tearing through the cuticle with a needle. Reference to the figure shows that very little of the old cuticle had been removed. Fig. 50 shows part of a section near the edge of the wound, Fig. 51 part of a 176 Mary Isabelle Steele section near the center of the wound. Both figures show the distribution of the hypodermal nuclei beneath the new cuticle that had formed over the cut surface of the eye of a Palamonetes ten days after the operation. Examination of Figs. 47, 51 will show that a new cuticle may be secreted before a complete hypodermis can be recognized. Recently hatched Cambarus gracilis, 12 to 15 mm. in length had eyes injured by tearing through the cornea with a sterilized needle. The eyes operated upon in this manner were fixed at different times, varying from eleven to thirty-five days. All of the eyes, however, were fixed before a moult had taken place. In this way it was possible to determine the precise position of the original injury. Sections of such eyes show one point conclusively, at least for Cambarus gracilis. ‘That is that the proliferation of new cells begins from the hypodermis immediately surrounding the rent. From the edges of this proliferating center new cells push out to replace the cells that were removed or that have broken down. Previous observations made upon the regeneration of the eye in crayfish (Steele ’04) indicate that crayfish probably do not regenerate a functional eye. It appears, however, that the prelim- inary regenerative processes are essentially the same in crayfish as in the other forms examined. It is frequently the case that a much greater proportion of the soft inner tissues are destroyed than of the outer cornea or the hypodermal layer beneath it. When the cut is made the retinule and the lower ends of the cone cells press out of the wound and leave the outer ends of the cones and the hypodermis practically undisturbed. Such a condition is particularly noticeable in eyes that were operated upon by tearing the cornea with a needle. In such cases the hypodermal cells 1m situ secrete the new cuticle. This cuticle is, however, without corneal facets, a fact which shows that while the operation neither removed nor caused the disintegration of the hypodermal cells it still affected their activity to such an extent that they no longer function in their usual specialized manner. ‘They now function as the ordinary hypo- dermal cells over the general surface of the body. Sections frequently show a morphological transformation of the Regeneration in Compound Eyes of Crustacea 177 Ps TEV corneal hypodermal cells in situ that are engaged in the secretion of the new cuticle. In the normal adult condition the pair of corneal hypodermal cells that belong to each ommatidium appears as much flattened cells crowded between the distal ends of the cones and the corneal facets. Their nuclei stain faintly and appear to be slender oval bodies lying flat against the cuticle. As the distal ends of the cones in an injured eye break down the nuclei of the corneal hypodermal cells enlarge, become rounded, stain deeply and in every way show signs of increased activity. Fig. 52 includes a series of figures showing the transformation of the corneal hypodermal nuclei into the larger, more deeply staining type seen in the regenerating eye. a and / inthis figure represent the corneal hypodermal cells as they appear in the normal omma- tidia. The other figures of this series, c,c’,d and e, show corneal hypodermal cells, belonging to ommatidia that have degenerated either wholly or paral oie c, c’, e the distal ends of “he cones still remain almost intact and in c the nuclei of the corneal hypo- dermal cells appear but little larger than those associated with nor- mal ommatidia. In c’ and e, however, the nuclei of the corneal hypodermal cells are much enlarged, stain deeply and the cyto- plasm surrounding them appears granular and loosely reticular. One nucleus to the left of the figure in d appears to be preparing to divide amitotically. That the nuclei shown in d are the trans- formed nuclei of the original corneal hypodermal cells is deter- mined by the fact that on either side of these nuclei are others still associated with partially disintegrated cones. Their original character is also suggested by the fact that they are grouped in pairs. A regenerated, rather than a transformed, hypodermis over the ommatidial region never shows the nuclei arranged in pairs in the early regenerative stages. “That the new cuticle has been secreted by ieee Paneionued hypodermal cells is shown by the relations of the two structures. The cytoplasmic strands of the hypodermis are continuous with the inner layers of the cuticle (Fig. 52d). Of course it is not absolutely proved that the transformed hypodermal cells take part in the later regenerative processes. This could not be done without examining a very great number of 178 Mary Isabelle Steele stages. But even the examination of a series of sections from the same eye will show that the indications are strongly in favor of the view that these hypodermal cells remain active and constitute the hypodermis during the subsequent regenerative processes. Fig. 53 1s taken from the same eye as d in the series shown in Fig. 52. In the later figure it is evident that regeneration is in progress. The hypodermal cells, however, show a tendency toward a paired arrangement indicating that they are the original hypodermal cells. After a part of the corneal covering has been removed it is evident that an entirely new hypodermis must be regenerated over the wounded surface. In crayfish it was seen that active cell proliferation began near the edges of injured hypodermis and that new cells pushed outward from these centers. ‘The early stages have been examined in a number of eyes of Palemonetes. The centers of cell proliferation in this form are not so apparent. The nuclei which in the early stages appear beneath the cuticle that covers the wound are very few and lie far apart. “Their number increases not chiefly by the repeated multiplication of nuclei at the edges of the wound but by the repeated division of the nuclei that are pushed out onto the wounded surface. During these early stages while the nuclei are actively dividing the cytoplasm is very loose and reticular and the cell boundaries are indistinguishable. In many cases the new cuticle which becomes apparent after the first moult is somewhat definitely separated into a dense outer portion and a semifibrous inner division. The inner division often appears as an interlacing network of fine fibers, many of which can be traced into the hypodermis beneath (Fig. 53). The loose incompact character of the hypodermis over the wounded surface probably indicates a high degree of activity of the hypo- dermis in this region. It is apparent from the foregoing description of the formation of the hypodermis covering the wound that the regenerated hypo- dermal cells may arise in two ways. They may arise by a trans- formation of the old corneal hypodermal cells im situ in which they assume a less specialized réle or they may arise by the migra- tion of a limited number of hypodermal cells from the edges of the Regeneration in Compound E yes of Crustacea 179 cut, which later multiply until a complete hypodermis is formed. In either case the first new nuclei must be contributed by the remaining hypodermal cells. Whenever the cut has not removed the entire ommatidial portion the remaining corneal hypodermal cells must assume a somewhat less specialized role in order to form the first new nuclei of the regenerated hypodermis. B REMOVAL OF THE INJURED TISSUE The fact that the inner tissues of the eye are so much softer than the chitinous outer covering renders it impossible to operate upon the eye in any way or to remove any part of it without serious injury to the remaining softer tissues. It is evident that, before any considerable regeneration can take place, the injured tissues must be either repaired or removed. All the observations made upon regenerating eyes tends to show that none of the injured tissues except the hypodermis ever repair themselves. Sections of an eye fixed six and one-half hours after the operation show that considerable changes have already taken place in the injured tissues. The effect is particularly noticeable in the retinula. Many of the retinular nuclei have become separated from the pigmented retinular processes and appear as rounded bodies surrounded by a dense mass of cytoplasm. ‘These are irregularly scattered among the other tissues. Some parts of the interior have become fairly clear of the broken down structures and are occupied chiefly by coagulated plasma. In other parts of the eye the injured tissues lie in confused masses. The changes in the next twelve or fifteen hours do not show much advance over the earlier stages. ‘The interior still presents a badly confused mass of broken down tissue. In some parts, however, the cone nuclei appear larger, the bodies of the cones have begun to dissolve and the number of rounded retinular cells appear somewhat more numerous than in the earlier stages, their nuclei showing irregularities in shape (Fig. 75a and b). During the earliest stages, six to sixty hours after the operation, the ommatidia that are still intact always appear bent and twisted out of their normal positions. Later stages show that these 180 Mary Isabelle Steele : uninjured ommatidia have regained their original shape and position. ‘This temporary contortion of the ommatidia seems to be due to the immediate effects of the operation, reduction of pressure, destruction of normal tension relations, etc. That in the later stages they appear normal again indicates that they have adjusted themselves to the new conditions imposed upon them by the operation. Besides those ommatidia that are actually injured by the operation a large portion of those remaining frequently degenerate. The destruction of the tissues of the eye is, consequently, much more extensive than the original injury. ‘This fact is strikingly illustrated by eyes in which the original injury consisted in thrust- ing a needle into the ommatidial portion. In several such in- stances the entire, or almost the entire, eye has degenerated. Instances of this kind have been observed in the eyes of sev- eral Palamonetes and also in the eyes of young Cambarus gracilis. Similar phenomena have been observed in eyes of fresh water Gammarus. To be sure the eyes of Gammarus are quite small which may account in a measure for the fact that in six or eight eyes examined in serial sections only one showed any of the old ommatidia intact. In Cambarus gracilis there were two instances out of six in which none of the ommatidial portion remained. In one of the cases the entire eye had degenerated; not even the vestige of the stalk remained. All of this degeneration took place in about thirty days and without a moult. In the other case all of the ommatidial portion and more than half of the optic ganglion had degenerated during the same period. In other eyes of the same series very little degeneration followed the operation. In these extreme cases it seems probable that some infection played a part. In the degeneration and removal of the injured tissues the retinula) degenerate most rapidly. In many cases they break down within the first few hours after the operation. ‘Their long pigmented processes become separated from the cell body and collapse into shapeless masses of pigment, which become scattered through the other tissues. The greater part of this pigment finally gathers in clumps near the level of the basement mem- "i Regeneration in Compound Eyes of Crustacea 181 brane. Although the retinulze are the first to collapse their remains are the last to be gotten rid of. Evidently the pigment is absorbed and removed only with difficulty. Often in regenerating eyes that have completely regenerated new ommatidia much of the old pigment remains. The cell bodies of the retinulz after losing their pigmented processes appear as large nuclei surrounded by a narrow zone of condensed cytoplasm. Within a short time these cells become scattered widely through the eye. After a few days their nuclei appear irregularly shaped and soon afterward become conspicu- ously polymorphic (Fig. 75a and b). In the usual course of events these retinular cells disintegrate and disappear. But under some conditions they apparently remain and later multiply and give rise to an abnormal development of tissue that secretes pig- ment. The rhabdoms and the inner ends of the cones also degenerate within a short time after the retinula. The cones continue to dis- solve from the proximal ends distally. The last parts to disappear are the outer ends in which are embedded the cone nuclei. Before the dissolution of the cones is complete the cone nuclei appear greatly enlarged and stain deeply. Their enlarged appearance is probably largely due to the disintegration of the cone substance from about them. As the disintegration of the tissues proceeds the interior of the eye becomes filled with a granular mass containing scattered nuclei and masses of pigment. This granular mass which is usually very much vacuolated is made up of the remains of the old tissues together with more or less of coagulated plasma and blood cells. Sometimes the remains of the old cones appear as long tapering bands of granular material extending from the periphery inward. Fig. 51 represents a small area of the disintegrated ommatidial structures as it appeared in the eye of Palamonetes ten days after the operation. Only two or three nuclei lie close beneath the cuticle and a few others lie scattered deeper in. Most if not all of these more deeply located nuclei are the remains of the old omma- tidial structures. In the lower part of the figure are some old 182 Mary Isabelle Steele pigment remains. ‘The granular mass occupying the greater part of the figure is made up chiefly of partially dissolved cones. The slender strands which can be traced through the granular mass are in part at least made up of new cytoplasm. At certain points the cytoplasmic strands are seen to be continuous with the inner layers of the cuticle. Fig. 48 is taken from a section of an eye of Palamonetes seven days after the removal of part of the ommatidia. A surface view from which this figure is taken is shown in Fig. 1. Practically all the material represented in Fig. 48 except the cuticle is made up of disintegrated ommatidia. The long band extending inward shows approximately the position of a former cone. ‘The large vacuolate spaces in the upper part of the figure are old nuclear remains. In the process of disintegration the nuclei at first enlarge and stain deeply. Later the nuclear contents disappear although the nuclear membrane persists for some time longer, often becoming shrunken and folded into a variety of shapes. It has been seen that the disintegration of the injured tissues begins immediately after the operation and that the greater part is accomplished in from one to three days. ‘The distal ends of the cones alone remain intact for a much longer time, in some cases from two to three weeks and occasionally even longer. ‘The removal of these disintegrated tissues is much slower than their dissolution. Regeneration proper may and usually does begin within a few days after the old structures have broken down and progresses simultaneously with their removal (Figs. 49,50). One part of the eye may show ommatidia differentiating while another region is still occupied by disintegrated old structures. The individual differences in the rate of regeneration of such frequent occurrence is probably largely dependent upon the variations in the length of time required for the removal of the old structures. This prob- ably also accounts for the fact that regeneration does not take place uniformly throughout the same eye. ‘The part of the eye that gets rid of the injured tissue soonest regenerates first. The above observations apply in general to all the forms used. Crangon, however, offers a significant exception in that the injured Regeneration in Compound Eyes of Crustacea 183 tissues disintegrate much more slowly than in Palzmonetes, her- mit crabs or crayfish. C METHOD OF CELL DIVISION In the earlier preliminary stages of the regenerative processes it is impossible to distinguish cell outlines. We should there- fore speak of nuclear divisions, perhaps, instead of cell divisions. In later stages the cytoplasm becomes differentiated about the individual nuclei. In all cases of the regeneration of the eye the nuclei are increased by amitotic division. Before a definite hypodermis is established the nuclei can be seen in various stages of constriction, separating off new nuclei for the development of the future underlying structures. ‘There seems to be no perfectly definite manner in which the constriction and separation of a nucleus into two parts takes place. One or two characteristic forms, however, appear so frequently as to be readily distinguished. The two most usual types are seen in the three nuclei occupying the extreme right of Fig. 53. When a nucleus divides in a plane parallel to its long axis it usually assumes the form of the outer one of the three referred to. The formation of the notch on one side gives the nucleus a peculiar heart-shaped appearance which seems characteristic and is easily recognized. Figs. 49 and 53 show nuclei of this same type further advanced in division. ‘The other type referred to is represented by the other two nuclei of the three at the extreme right of Fig. 53. Apparently these two nuclei were recently formed by the longitudinal division of one of the heart- shaped nuclei, like the one lying beside them. Each of these two is now dividing unequally by a transverse constriction. A dividing nucleolus can sometimes be seen but more frequently a definite nucleolus cannot be distinguished. When nucleoli are seen in dividing nuclei they usually appear with a darkly staining strand of material connecting them. In sections of young Cam- barus gracilis eye, nine and sixteen days after the operation, two nucleoli can be recognized in many of the nuclei. In those nuclei that are dividing one nucleolus lies in each part. In all except in the later stages of regeneration of the eye nuclei dividing amitot- 184 Mary Isabelle Steele ically can be found in great abundance. But at no time are mitotic divisions seen. In a careful examination of a large num- ber of sections only one cell has been found that suggested the possibility of its being in mitotic division. ‘The appearance of this one suggests a late anaphase or an early telephase. Conse- quently it is not certain that this is mitotic division. “This nucleus appears near the left edge of Fig. 53 at K. In any case it must be admitted that amitosis is the regular method of cell division in regenerating eyes of the forms studied, since in every specimen examined during the stages of cell divisions amitosis has been observed and mitosis has not been seen. That such is the case is somewhat unexpected since Miss Reed (o4) found mitotic division abundant in the regenerating leg of a crayfish. Miss Reed, however, observed that there were no mitotic figures during the early stages of regeneration, although new cells were being eapily formed. Perhaps we may infer from this that amitosis took place in the regenerating leg of the crayfish during the preparatory stages at ee But one would hardly expect such differences in cell division in forms so closely related for example as crayfish and Palemonetes. No observations have been made upon histogenesis in the regenerating appendages of hermit crabs, Palemonetes or Crangon. Hence it cannot be said whether or not the eye furnishes an unique exception to the regeneration of other parts in these forms. Recently, however, many have come to regard amitosis as a phenomenon of more frequent occurrence than has been generally supposed. Meves (’g1 and ’94), McGregor (’99), Child (’04 and ’07), all describe amitosis as a normal phenomenon and con- sequently no longer accept Vom Rath’s view that a cell is nearing its final dissolution when it begins to divide amitotically. On the other hand many of the more conservative investigators are unwilling to admit that amitosis occurs as a normal phenom- enon and believe that the apparent cases of amitosis can be explained on some other grounds. But in all the instances hitherto described amitosis has been found occurring along with mitosis. In the present case, however, all of the cell divisions are amitotic and they all take place in cells derived from the hypodermis; in Regeneration in Compound Eyes of Crustacea 185 these respects the regenerating eyes investigated in this series of experiments offer a unique instance in their method of cell division. V REGENERATION OcCURRING AFTER DESTRUCTION OF DistaL Part oF EYE Under this heading will be discussed the results obtained from eyes injured in varying degrees but not exceeding the destruction of more than the distal two sections of the optic ganglion. ‘The injury originally inflicted varied from tearing the surface of the cornea with a needle to cutting off the whole top of the eye so that at least the first part of the second section of the optic ganglion had been removed. Frequently, however, as has been explained in Section IV the part of the eye ultimately lost was much greater than that originally removed. For in cutting the eye with scissors or tearing it with a needle much of the tissue surrounding and under- lying the wound was so injured that it afterward degenerated. In the series of experiments to be described under this division nearly two hundred Palamonetes, twenty Crangon and fifty hermit crabs were used. Fifty per cent of the Palamonetes, 5 per cent of the Crangon and Io per cent of the hermit crabs died of the operation. Of those that survived the loss of blood and the nervous shock of the operation a considerable number died from other causes without having undergone a moult. Forty-five Palamonetes, seventeen Crangon and twenty-nine hermit crabs, however, lived through at least one moult. Among these a num- ber showed no regeneration either from surface examinations or from sections. The greater number of those that gave surface indications of regeneration and for comparison a number that showed no regeneration have been sectioned and examined. A REGENERATION OF THE FUNCTIONAL EYE Experience has shown that the number of days an experiment covered serves to indicate only in the most general way the stage of regeneration. While the time element naturally constitutes a most important factor the rate of regeneration is also dependent upon the season of the year, the age and the physiological activity 186 Mary Isabelle Steele of the individual and perhaps other factors not so apparent. Besides individual differences displayed between members of the same species there also appeared to be differences in the ability to regenerate and the rate of regeneration. among the three species chiefly used, these differences being much more marked in the regeneration of an eye than in the regeneration of an appendage. The hermit crabs seem to completely regenerate an eye in shorter time than either Palamonetes or Crangon. But the final results were similar in all three forms. The few significant differences will be pointed out and discussed later. rt Entire Preparations of Regenerating Eyes A careful examination of the regenerating eye frequently reveals a number of important features and is absolutely essential to the later interpretation of the sections. Therefore outline surface drawings have been made of all eyes later sectioned. In most cases the normal eye was drawn in connection with the regener- ating eye for comparison as to size, shape, etc. Frequently the two eyes were sectioned together. These surface views which had better be regarded as optical sections were drawn with a camera after the eyes had been brought into oil preparatory to embedding. Figs. 1 to 25 represent various stages of regenerating eyes. Fig. 1 shows dorsal and ventral views of a Palamonetes’ eye seven days after an operation which removed the ventral corneal surface and immediately after a moult. A comparison of the dorsal and ventral views shows that the injury is confined chiefly to the underside. In shape the eye is practically normal. It is not even flattened on the underside as the thinning out of the pigment near the center would seem to indicate. As is usually the case the injured eye is much smaller than the uninjured one. Not only the region operated upon but the whole eye decreases in size after the operation. ‘This indicates that however localized the operation may be the effect is much more extended. In this particular case the part of the eye proximal to the injury measures only three-fourths of the length of the same region in the unin- jured eye. Regeneration In Compound Eyes of Crustacea 187 Fig. 3 represents the ventral view of a Palamonetes eye ten days after being injured. In this case also the injury extends across the ventral side of the ommatidial portion. The pigment of the broken down ommatidia can be seen scattered in flakes and patches through the upper part of the eye. From the dorsal side the eye appeared nearly normal but sections show that almost the entire eye is in process of degeneration. The specimen represented in Fig. 4 shows a regenerating eye nineteen days after the removal of almost all of the ommatidial portion. It is readily seen that very little remains except the eye stalk. The pigment patches are remains of the original eye. Across the end of the stump the cuticle is wrinkled and folded, indicating that comparatively little new tissue has been formed and that the cuticle follows more or less closely the rough uneven out- lines of the wounded surface. When the entire ommatidial portion has been removed or has degenerated regeneration seems to be considerably slower than when a large part of it remains uninjured. ‘The two specimens shown in Figs. 9 and 1o afford a striking illustration of this fact. Both of these eyes were operated upon at the same time. Each animal moulted twice, the first time on the same day and the second, a day apart. Both were fixed in picro-acetic at the same time, thirty-two days after the operation. In Fig. g the injury involved only the posterior ventral side, less than one-half of the ommatidia. While in Fig. 10 the injury included all of the structures lying distal to the basement membrane. Examination of the sections shows new ommatidia completely differentiated in Fig. 9 while in no part of Fig. 10 are they yet defined. The next specimen, Fig. 20, presents a rather striking appearance and suggests immediately that the regeneration taking place is not altogether of the normal type. ‘This is a thirty-eight day specimen and belongs to the same series of experiments as the preceding two specimens. All of the part distal to the dark pigmented band is regenerated tissue. [he pigment consists chiefly of the remains of the old retinulz. It is evident even from a surface view that no ommatidia have developed. Sections show, however, that on one side the differentiation of cones is beginning (Fig. 69). Here it 188 Mary Isabelle Steele may be well to mention a fact that has been observed a number of times. The regeneration of the new ommatidia never presents a uniform stage of differentiation in any case whether or not all of the old ommatidia have disappeared. In fact it may be possible to select several stages from the same eye and sometimes two or three stages from the same section. An eye thirty days after the injury, Fig. 6, shows some interesting features in comparison with the one just described. From a ventral view this eye shows no signs of remaining ommatidia and from the dorsal side only a very few cones and facets are evident. The upper part of the eye is transparent. Below this transparent area are scattered patches of pigment representing the remains of the old eye. Sections show that no new ommatidia have been entirely differentiated, that the retinular cells have differentiated and are establishing connections with the optic ganglion, that new cone nuclei are being separated from the hypodermal nuclei, that on one side a few of the old ommatidia remain and that in the tissues lying nearest the old ommatidia new cones are being developed. The most striking feature presented by this specimen is the clearness with which the connections between the retinulze and the ganglion cells can be made out. ‘These connections will be discussed at length in the consideration of the detailed structures of the regenerating eye. Fig. 17 shows the ventral view of a regenerating Palamonetes eye from a thirty-five day specimen. Examination of the dorsal side shows that a large number of the ommatidia on that side appear uninjured, the greater part of the injury being confined to the ventral side as indicated by a surface examination. Sections show, however, a gradation:of regeneration from a stage in which there is no signs of cone differentiation to the complete formation of a new ommatidium. It was said above that hermit crabs regenerated an eye more rapidly than either Crangon or Palamonetes even in instances where a considerable portion of the optic ganglia had been removed. Figs. 5, 12 to 15 show regenerated eyes of hermit crabs after one or more sections of the optic ganglion have been de- stroyed. Fig. 12 had at least the distal division of the ganglion Regeneration in Compound Eyes of Crustacea 189 removed. ‘The regenerated eye shown in the figure was developed within hive hrees days and after only one moult. This moult occurred on the thirty-second day and the eye was fixed in Perenyi the thirty-third day. Sections show that ommatidial structures have been fully differentiated although incompletely developed. One point is particularly noticeable in these sections; the omma- tidia are very much shorter thaninthe normaleye. ‘This condition was probably caused by the mechanical pressure of the covering cuticle which forced the developing ommatidia into less space than they would otherwise have occupied. Fig. 13 shows the eye of a hermit crab regenerated from a stump in which not more than half of the optic ganglion remains. ‘The regeneration in this eye took place in thirty-eight days. One moult occurred twelve days after the operation. Although extremely small the eye 1s practically perfect except that the corneal facets have not yet developed. Figs. 14 and 15 show two other regenerated eyes of hermit crabs forty-one and sixty-seven days respectively after the operation. Sections of the forty-one day eye do not show ommatidia as fully developed as the thirty-eight day specimen previously described. The sixty-seven day specimen shows the eye complete in all its details even to the corneal facets. Whether a younger regenerated eye might not show the corneal facets has not been determined since no specimens were available between the forty-one day and the sixty-seven day specimens. A noticeable feature in all the regenerated eyes of the hermit crabs is their small size in comparison with the normal eyes. It is probable that the regenerated eyes would have increased in size if the experiment had covered a longer period of time. Sections of the eye shown in Fig. 12 indicate that the definitive size has not been reached. For, lying outside the fully formed ommatidia are others in the process of development. In the case of the eye shown in Fig. 13 sections do not show any indications of partially developed ommatidia and it may be that this eye would never have reached the size of the normaleye. Perhaps this is what we should expect since the amount of nervous tissue present is considerable less than is normal. It is more difficult to interpret the actual condition of a regener- Igo Mary Tsabelle Steele ating eye in Crangon than in either Palamonetes or hermit crabs. The whole of the eye stalk in Crangon is thickly covered with branching pigment cells. So that even after the eyes are brought into oil very little can be seen in detail by examining them in toto. Figs. 7 and 8 thirty-one and thirty-two days, respectively, are fairly characteristic of the regenerating eye of Crangon. From all that can be determined from the outside, regeneration seems practically complete in each of these cases. Sections show, however, that comparatively little regeneration has taken place. A fuller discussion of the regenerating eye of Crangon will be taken up elsewhere. 2 Details of Development of the Regenerating Ommatidia The complete regeneration of the ommatidial portion of the eye involves three stages. ‘These can be separated from each other rather sharply although they overlap more or less. The first stage consists in getting rid of the broken-down tissues and the healing of the wound; second, the active proliferation of new cells; and third, the differentiation of the new ommatidia. The first and second stages have been discussed in Section IV. a Regeneration of Retinulze All the observations support the conclusion that the regenerated ommatidia are derived entirely from the hypodermis. Before the hypodermis covering the end of the stump has been clearly differen- tiated, however, the proliferation of cells for the new structures lying below has begun. So that at the same time hypodermal cells are dividing in two planes, one at right angles to the periphery to increase the number of hypodermal cells, and the other parallel to the surface. The inner nuclei of the latter division migrate inward and become the first retinular cells. As they migrate they become elongated with their long axes radially arranged. Fig. 54 shows the early stages of the separation of retinular nuclei, some separating from the nuclei at the periphery and others migrating in. At a comparatively early date these retinular nuclei have migrated a considerable distance below the surface Regeneration in Compound Eyes of Crustacea IgI and may be seen in a relatively well defined row (Fig.68). Soon after the retinular nuclei have been separated from the nuclei at the surface they themselves begin to divide (Fig 68) in a plane at right angles to their plane of original division. ‘These longitudinal divisions may begin before the nuclei have reached their definitive position. ‘This division continues until a band composed of many nuclei has been formed. Figs. 60 and 62 show portions of such bands. Here and there single nuclei are found lying much nearer the basement membrane than the retinular band (Figs. 60, 62). These are occasionally seen constricted. It has been impossible to determine with certainty the fate of these scattered nuclei but there are evidences which suggest that they become the nuclei of the accessory pigment cells. The retinular nuclei even in the early stages of their migration stain much more deeply than the nuclei at the periphery. But at this stage no definite cytoplasmic outlines can be distinguished. Very faintly staining delicate strands of cytoplasm, however, can be found extending between the retinular nuclei and the periphery. These strands form an intermingled network and with the nuclei lie in a granular substance. Soon after the nuclei have reached their definitive position the cell bodies of the retinule can be recog- nized. Each nucleus appears surrounded by more or less definite strands of cytoplasm which are radially arranged and extend outward toward the periphery and inward to the basement mem- brane (Figs. 60, 65). These can now be definitely recognized as retinular cells. Usually it is easier to see the proximal than the distal strands for at an early stage these lower processes begin to secrete pigment and are consequently more conspicuous (Fig. 65). At the stage represented by Fig. 62 there is only the merest begin- ning of pigment deposition. ‘The fibers are but little differentiated from their background which still seems to be composed largely of a homogeneous granular material, probably to a great extent coagulated plasma. : At a stage such as 1s represented by Figs. 60, 62, in which there is only the merest beginning of pigment secretion delicate cyto- plasmic processes can be traced from the retinular nuclei inward 192 ; Mary Isabelle Steele through the basement membrane beneath which they can be found branching over the ganglion cells. Fig. 62 shows several isolated retinular cells and their proximal processes which are seen extend- ing through the basement membrane and reaching to the ganglion cells beneath. There is no evidence that these retinular fibers are directly connected with the ganglion cells; they seem merely to twine around them. When the fibers have reached the basement membrane they may pass directly through it as the two shown in the left side of Fig. 62, or they may extend along the upper face of the basement membrane before entering the ganglionic mass below. ‘The retinular processes frequently branch shortly before entering the basement membrane or just as they emerge below it. No special nerve methods were employed yet numbers of these fibers are seen branching among the ganglion cells and many are readily traced from the retinular nuclei to the basement membrane (Fig. 65). It is only in very favorable specimens that the fibers can be traced, however, through their whole length. b Regeneration of the Crystalline Cones Up to and including the stages shown in Fig. 65 and described above there is no evidence of any differentiation of the crystalline cones. A definite hypodermis, however, has been formed with an increased number of nuclei, from which other nuclei are separating, These are the cone nuclei (Fig. 63). The formation of the cone nuclei does not appear to take place in a uniform manner. The usual method, however, is by the division of the hypodermal nuclei in a plane parallel to the periph- ery, the inner nuclei thus formed being the cone nuclei. But since in every ommatidium there are four cone nuclei and only two hypodermal nuclei, it is evident that either the hypodermal cells must divide twice or the first cone nuclei must themselves divide in order to make the cone nuclei just twice the number of the hypo- dermal nuclei. Observations indicate that in some cases the second pair of cone nuclei arise by the division of the first pair. In other cases, however, it is uncertain whether they arise in this manner or whether they arise from the hypodermal nuclei. It is Regeneration in Compound Eyes of Crustacea 193 probable, however, that they are formed by the division of the first pair of cone nuclei as can be determined in some cases. In the early stages of cone formation the inner surfaces of the hypodermal cells lose their distinctive outlines. For at this time there is no clear line of demarcation between the hypodermal and cone cells. Fig. 63 represents one of the earliest recognizable stages in crystalline cone formation. The hypodermal cells are more or less definitely grouped into pairs and it is readily seen that cone and hypodermal nuclei are not wholly separate. Extending inward from the cone nuclei are very delicate strands of cyto- plasm. ‘These strands seem to group the hypodermal nuclei into pairs and by their branching and crossing form a much vacuolated network. Ata stage slightly more advanced (Fig. 64) the cytoplasm of the cones has begun to assume a more definite cone shape. ‘There is still, however, no distinct line of separation between the hypo- dermal and cone nuclei. Neither does the cone mass show the boundaries of its component cells. In stages a little later the cone cells begin to show individual outlines and the cytoplasm appears more condensed. Cone formation is practically complete, how- ever, before the corneal hypodermal and cone. cells show a distinct line of separation (Fig. 66). As the cone cells differentiate the cytoplasm becomes less and less vacuolate and gradually assumes a dense granular appear- ance. ‘The cytoplasm is most condensed just below the nuclei and decreases in density proximally. In longitudinal sections of the cones the cell boundaries appear distinct from their outer ends inward to the outer retinule. At this point the cell boundaries become indistinct and the cone tapers rather suddenly into a slender stalk which extends to the distal end of the rhabdom, where it ends abruptly (Fig. 66). At a somewhat more advanced stage the boundaries between the cone and corneal hypodermal cells become distinct, and the cone secretion takes on the dense homogeneous deeply staining appear- ance characteristic of that in mature cones (Fig. 67). 194 M ary Isabelle Steele c Regeneration of the Rhabdoms There is no indication that the retinular cells have begun to secrete the rhabdoms until after the cones have been distinctly outlined, although the retinula themselves become clearly differen- tiated before there is any indication of the cones. Not until the differentiation of the ommatidia has reached a stage intermediate between these stages shown in Figs. 64 and 66 can the anlagen of the rhabdoms be recognized. The rhabdoms first appear as slender homogeneous rods. Each rod is of uniform diameter throughout its length, and is distinguishable from the inner ends of the cone cells only by the fact that it stains slightly deeper and shows no divisions which indicate that it is composed of more than one cell (Fig. 66.) The rhabdoms show no signs of the character- istic spindle-like form and the complicated system of transverse plates so noticeable in the normal adult eye until after the last stage in the differentiation of the cones (Fig. 67). Even at the stage shown in the preceding figure the rhabdom does not show a normal appearance of its spindle form and the pigmented exten- sions of the retinular do not cover it so completely as in normal adult ommatidia. It is evident from the preceding description and accompanying figures that all the structures necessary to a completely regenerated eye have been laid down. It is also seen that with the exception of the corneal facets the regenerated ommatidia are practically identical with those of the normal adult eye. A specimen of later stage, however, shows both the corneal facets and the definitive form of the rhabdom, so that the regenerated ommatidia present a perfectly normal appearance even to the minutest detail (Fig. 76). All observations show that the differentiation of corneal facets could not become evident until after at least two moults. The first cuticle which is developed is formed before a continuous hypodermis has grown over the wounded surface and before any regeneration of ommatidia has begun. Corneal hypodermal cells are not differentiated as such until all of the other ommatidial structures have been laid down. ‘Therefore the secretion of corneal facets constitutes the final process in the regeneration of Regeneration In Compound Eyes of Crustacea 195 ommatidia. Since this is true, several moults may occur before the corneal facets differentiate, and at least two must take place. 3 Comparison of Ommatidia in Regeneration and Ontogeny In comparing the regenerating with the embryonic eye it 1s necessary to consider them only from the beginning of ommatidial differentiation, since there can be no exact parallel between the preliminary stages of regeneration and the mode of the origin of the embryonic eye. These two processes are similar, however, in the respect that the ommatidia in both cases develop from the hypodermal cells. All these observations upon the regenerating eye give evidence that the cells which take part in the formation of the new ommatidia are derived primarily from the hypodermal cells that cover the wounded surface. That the ommatidia of the embryonic eye develop entirely from the hypodermis is the con- clusion of most observers. ‘There is no further agreement, how- ever, in details except in the case of embryonic eyes that are described as arising without invagination, e. g., the compound eye of the honey bee as described by Phillips (05) and of the lobster (Parker ’g0), in which it was found that the ommatidia are developed from a single epithelial layer and consequently from morphologically similar cells. It has been seen that the first cells differentiated from the hypo- dermis in the regenerating eye are the retinular cells. “This can be regarded as being in agreement with the conclusion of Phillips that the retinulz constitute the morphological center of the ommatidium. At any rate the retinula are in each case differen- tiated before the cone cells can be recognized, none of the cells originally separated from the hypodermis to form the retinulz ever take any part in the formation of cones, and finally the cone cells arise peripheral to the retinule. The differentiation of the regenerating ommatidia, described in a preceding section, and of the embryonic eye, as described by Kingsley (’87), may perhaps be regarded as presenting a parallel. Kingsley finds the nuclei which go to make up the cones and re- tinulz arranged in radial rows and that the outer and hence the 196 Mary Isa belle Steele later formed nucleus of each row contributes to the formation of a cone while the remaining nuclei form the retinule. Unlike the rows of nuclei described by Kingsley and the spindle shaped groups of cells described by Phillips the retinulz of the regenerating eye do not appear to become separated into definite groups before the development of the cones. In sections from the same eye there may be groups of retinula somewhat distinctly separated from each other and other retinula which constitute a continuous band for a considerable distance. Figs. 60 and 61 illustrate these opposite cases. But even when groups can be recognized there is no certain indication that a group belongs to a single ommatidium. The group may contain a fewer or a greater number of cells than belong to a single ommatidium. Besides the retinule continue to divide occasionally up to the time the cones are differentiated. From the evidence furnished by a num- ber of different specimens it appears that the definite separation of the retinul into groups does not take place until after the cones are well advanced in their development. As the cones differen- tiate from the periphery inward the retinulae become grouped about them. As this grouping continues the retinular processes become more and more slender, perhaps largely as a result of mechanical pressure. The development of the cone as shown by my observations is the result of intra-cellular secretion. In this respect it agrees with the embiyonic development of the cones in the eye of the honey bee as described by Phillips. It is directly opposed to the method described by Patten (87), Kingsley (’87) and Watase (’89), who regard the cones as the result of the extra-cellular secretion of the cone cells. The evidence furnished by the regenerating eyes of Palamonetes, Crangon and hermit crabs agrees with the observations of those who do not find the cone and the rhabdom to be developed i in the embryonic eye as continuous structures. Some investigators regard the rhabdom as merely an inward prolongation of the cone cells. Kingsley finds such a relationship in the embryonic develop- ment of the eye of Crangon. Patten regards the cone as extending from the hypodermis to the basement membrane and as differen- \ Regeneration in Compound Eyes of Crustacea 197 tiating at the lower end into the rhabdom in most cases. But in Vespa he describes the inward prolongation of the cone cells as enclosing the rhabdom. Parker (’00) describes the prolongation of the cone cells in Homarus as extending to the basement mem- brane and inclosing the rhabdom in the same manner. ‘The description of the relation of the rhabdom to the cone in Crangon, given by Kingsley, and applied to Crustacea in general by Patten, does not agree with the facts presented by the regenerating eyes of Crangon, Palzmonetes and hermit crabs. Obviously, however, this interpretation is in accord with that of Phillips in the case of the honey bee and that of Grenacher (’74), both of whom find that the rhabdom is developed as a secretion of the retinulz, and do not find the cone cells extending as slender processes beyond the distal end of the rhabdom. Concerning the source and manner of the innervation of the ommatidia the results obtained in this study of the regenerated eye agree only with those observers who, like Parker (’g1) and Phillips (05), regard the retinular cells as hypodermal sense cells which send nerve fibers into the ganglion below. It is true that no special nerve methods were used in this work upon the regenerating eye. But in some specimens at least the prolongation of the retinu- lar processes into fibers which penetrate the optic ganglion is clearly evident (Fig. 62). In many other cases the processes can be traced from the retinular nuclei to the basement membrane and similar processes are found branching among the ganglion cells below it. But in no case is there the slightest evidence that the ganglion cells are sending fibers upward to the regenerating ommatidia. Consequently there seems to be no room for reason- able doubt that the retinulz form the nerve endings of the omma- tidia. In this particular these results differ from those of Patten, Kingsley and other workers on the embryological development of the eye of certain Decapods. ‘These investigators regard the nerve connections as being formed by the extension of processes upward from the ganglion cells, through the rhabdom and into the cone. The observations made in this work seem neither to uphold nor to oppose the views of those who find that the ganglion cells send 198 Mary Isabelle Steele processes upward into the retinula during the embryonic devel- opment. For it is not inconceivable that the innervation of the ommatidia of a normal eye should be accomplished by the upward growth of processes from the ganglion cells to the retinulz and that in the regenerating eye it should be accomplished by proc- esses growing inward from the retinulz to the ganglion cells. That this is not impossible is suggested by the fact that in regener- ation tissues are sometimes developed from the same germ layer while they arise from different germ layers in embryonic develop- ment. Several instances are known where muscles in regenerated appendages arise from the hypodermis although normally they are of mesodermal origin. Miss Reed (’04) finds this to be true in the regenerating leg of the crayfish. Ost (’07) notes the same phenomenon in the regenerating antennz of Oniscus. It is recognized, then, that certain tissues originating normally from different germ layers may arise in a regenerating organ from the same germ layer. It would be at least possible that, although the nerve connections between the optic ganglion and the retinulz arise as processes from the ganglion in the development of the normal eye, they might arise as processes from the retinulz in the regenerating eye. As even in this case they would develop from the same germ layer although from different parts of it. The possibility that the nerve connections may have arisen differently in the embryonic eye and in the regenerating eye is conceivable. Yet it seems that the evidence obtained from a comparative study of the normal adult eye and the regenerating eye suggests that the nerve processes develop from the retinula in the normal eye just as in the regenerating eye. The preceding pages show hae the dev iogaede of the regener- ating compound eye corresponds in a general way with the embry- onic development of the compound eye. ‘They also show that the observations made upon the regenerating eye do not agree entirely with those of any one worker upon the embryonic development of the compound eye of Arthropods. In many respects, however, there is a close similarity between the development of the regener- ating eyes of Palamonetes and hermit crabs and the process of Regeneration in Compound Eyes of Crustacea 199 differentiation in the embryonic eye of the lobster as described by Parker (Joc. cit.) Further, these observations upon regenerating eyes agree with those of Phillips upon the developing compound eye of the honey bee in regard to the order of appearance of the retinulz and cones, in the method of innervation of the ommatidia and in regard to the relation of the cones and rhabdoms. In the developing eye of the honey bee, however, Phillips finds the rhab- doms partially differentiated before there is any indication of the cones. On the other hand in the regenerating eyes of hermit crabs and Palamonetes the cones are definitely formed before any thabdoms can be recognized. The variations, however, which have been noted between the developing compound eye of the honey bee and the regenerating eyes of Palamonetes and hermit crabs, cannot be regarded as fundamental. Such differences are perhaps not more marked than those that would be noted if the embryonic development of the same eyes were compared. It is scarcely necessary to add that these observations on regen- erating eyes are in several respects quite at variance with the observations of Kingsley and Patten, who find the rhabdoms developed from an inward prolongation of the cone cells, the cones formed as extracellular secretions and the ommatidia innervated by nerve processes coming from the optic ganglion and penetrating the rhabdoms and cones. 4 Differences in the Regeneration of the Eye Among Pala- monetes, Crangon and Hermit Crabs Reference has already been made to the fact that certain differ- ences in the regenerating eye appear among Palaemonetes, Cr angon and hermit crabs. The rate of regeneration and the ability to regenerate varies greatly in these different genera although in the most essential particulars the regeneration of the ommatidia is similar. It has been seen that hermit crabs may regenerate an eye after the removal of half the optic ganglion. But neither Palamonetes nor Crangon regenerate a perfect eye if the injury includes any part of the optic ganglion. It has also been seen that the differentiation of the ommatidia takes place more rapidly in 200 Mary Isabelle Steele the hermit crabs than in either of the other forms. The only structure in the eye of the hermit crab which apparently does not regenerate perfectly is the basement membrane. ‘This membrane, however, is but slightly developed in the normal eye. It is not strange, therefore, that it appears imperfect in the regenerated eye. The regenerated eyes of the hermit crabs in these experiments have developed from a level below the basement membrane. In every case they present a clearer and more normal appearance than the regenerated eye of either Palemonetes or Crangon. This is largely due to the fact that in the eye of the hermit crab there are no shapeless masses of old pigment scattered among the regener- ated tissues as is usually the case in Crangon and Palemonetes. The absence of the yellow accessory pigment cells in the eye of the hermit crab also tends to give to the ommatidia a distinct and orderly arrangement. ‘The absence of this accessory pigment is, however, not due to incomplete regeneration. “The normal eye of the hermit crab, unlike that of many Decapods, contains no accessory pigment cells. These pigment cells are very abundant in the eyes of Crangon and Palamonetes and tend to make the ommatidia less clear. The most significant difference between Palamonetes and Crangon seems to be in the rate of regeneration after similar injury. External appearances would indicate that Crangon regenerates more rapidly than Palamonetes. But a comparative study of the section shows that the reverse is true. In almost every individual in the series of twenty Crangon operated upon the injury was slight. “The wound healed rapidly, the animals moulted frequently and externally there was every indication that regeneration was rapidly taking place. An exam- ination of the sections, however, shows that in none of them has there been any considerable regeneration. On the other hand, in each case much of the old injured tissue remained in a semi- broken down condition. Sections of the eye represented in Fig. 8 show that one part of the eye had not been injured below the level of the outer retinular cells. ‘The proximal ends of the retinule still remain intact although thirty-one days have elapsed since the operation. A Regeneration in Compound Eyes of Crustacea 201 continuous hypodermis has not yet been formed. A considerable area between the cuticle and the outer retinule is occupied by a granular structureless mass. Just to one side of this area is a region in which none of the old ommatidia appear but in which there are new ommatidia almost completely formed. No more than five of these appear in any one section. ‘These lie near the basement membrane toward the inner edge of the eye. Con- sequently these new ommatidia lie next the growing zone, always present in the eyes of young individuals as described in Section III. It is apparent then in this particular case that it is impossible to determine conclusively whether the new ommatidia are regenerated or normally developed ommatidia. In other cases, however, new ommatidia are found developing in positions where it is evident they are regenerating ones. From the evidence obtained by an examination of a number of regenerating eyes of Crangon it seems that the rate of regeneration depends largely upon the rate of removal of the injured tissue. The failure of the old tissue to degenerate prevents the regeneration of new structures. Since the injured ommatidia, although they fail to break down for a considerable time at least, are incapable of regeneration in themselves. We should perhaps expect the cones to be incapable of any sort of regeneration for the cone nuclei have been destroyed and the constructive metabolic activity of a cell apparently depends largely upon the nucleus. ‘The nuclei of the retinula, however, have not been injured and still retain much of their normal appearance. But there is no evidence that the retinulz ever take part in normal regeneration. A comparative study of the regenerating eyes of Palaemonetes and Crangon shows that the difference in the rate of regeneration is in reality largely a difference in the rate of degeneration of the injured tissues. In Palamonetes the injured tissues usually break down rapidly and are quickly removed. In Crangon they persist indefinitely. Hence regeneration in Palamonetes begins soon after the injury, and new ommatidia may be almost fully developed in shorter time than is required by Crangon for the removal of the injured ommatidia. The specific case of Crangon perhaps suggests that if all of the ommatidia had been completely 202 Mary Isabelle Steele removed by the operation, that regeneration would have followed more rapidly. An inference which is supported by the observa- tions of Zeleny ('05), who finds that in regenerating appendages of crayfish, an increase of the injury, increases the rate of regener- ation. On the other hand, the same inference cannot be applied to the regenerating eyes of Palamonetes; for as has already been pointed out for this form new ommatidia differentiate more rapidly when a part of the old ommatidia remain uninjured. 5 Comparison of Normal and Regenerated Eyes It seldom happens that a regenerated eye appears altogether normal either from external or internal examination. Externally they frequently appear abnormal in shape and are always smaller than the opposite eye. These external abnormalities are, however, of no especial importance except in so far as they may indicate internal conditions. A common external feature which is sugges- tive of internal conditions is the irregular arrangement of the pig- ment masses. Internal examination shows that these masses are frequently remains of broken down retinulz. Besides the pigment remains of the old retinula, a number of other abnormalities may appear which make it difficult to interpret sections correctly. The arrangement of the retinule makes it difficult to group them into distal and proximal rows of nuclei as can be done readily in the normal eye. It is quite possible in cross sections to select ommatiia in which eight retinulz, the typical number, can be counted but in other cases this number cannot be recognized owing possibly to the suppression of the eighth retinular cell, nehicha is eidimeneauy 3 in the normal eye. The difference in the length of normal and regenerated ommatidia 1s quite noticeable in many cases. The regenerated ommatidia are often much shorter. The new ommatidia might have grown to normal size, however, had the experiment covered a longer period of time. The remaining significant difference between the normal and regenerated eyes is that in some regenerated eyes the optic ganglion is not complete. This difference appears only in the hermit crabs as these forms may develop an eye after half the optic ganglion has {i Regeneration in Compound Eyes of Crustacea 2.03 been removed. In no case is there any evidence that the optic ganglion regenerates. Consequently this difference would remain unchanged. B CASES OF ABERRANT REGENERATION OF OMMATIDIA In study of the regenerating eye several cases of aberrant regen- eration of ommatidia have come under observation. One case deserves especial mention. Fig. 36, a and b represent ventral sur- face views of an injured right eye and the normal left eye of a Palemonetes. Judging both from the surface indications and an examination of the sections the right eye must have been cut off at a level corresponding approximately with the line a—b shown in the figure of the normal eye. A cut at this level would remove the upper two sections of the optic ganglion and injure the third. _It would also cut across the heavy muscle band lying in the posterior part of the eye stalk. The experiment covered thirty days, one moult taking place ten days after the operation. Casual surface examination was sufficient to show that a rather large amount of new tissue had formed and that a spot of pigment had developed on the ventral side of the stump. The growth of such a large amount of new tissue is quite unusual when so much of the ganglion has been removed. A study of the sections gives additional information regarding this new tissue. In Fig. 56 the tissue lying between the periphery and the broken line extending from x to y represents approxi- mately the amount of new tissue. Careful examination shows a difference in the character of the differentiacon of the regener- ated tissues in the different regions. ‘his figure is from a section so near the dorsal surface that but little of the nerve tissue appears. Near the right side of the figure a conspicuous section of the old muscle band is seen. Just distal to the muscle band the new tissue is more dense and compact than in the remaining part of the regenerating tissue. “This band of new tissue (mt) is composed of fibers extending inward from the periphery and joining end to end with the fibers of the old muscle band, thereby forming a contin- 204 Mary Isabelle Steele uous band and reéstablishing the broken connection between the muscle and the chitinous covering of the stalk. Fig. 57 of the same series represents a section deeper in from the dorsal surface so that parts of the optic ganglion are apparent. A few scattered spots of pigment are also present. Here again the regenerated tissue shows differentiation into strands of fibers in the part lying beyond the remains of the old muscle band. In other regions there is a loose network of fibers with scattered nuclei. A difference, however, in the appearance of the nuclei in different regions of the regenerated tissue can be observed. From a to b the nuclei are small and inconspicuous, constituting uniformly granular masses and staining with but little more intensity than the fibers which extend inward from them. The nuclei lying between the points } and c, on the contrary, are conspicuous, stain” deeply and are more than twice the size of those lying between aandb. Here also a number of nuclei are seen lying below the periphery which show a tendency to extend straight inward from the periphery. Fig. 58 shows the upper part of a section that lies so near the ventral surface that it is entirely outside most of the optic ganglion. In this figure no part of the muscle band appears. Conspicuous masses of pigment are present in this section as well as a great number of relatively large deeply staining nuclei. Since the eye was cut longitudinally from the dorsal to the ventral side the sections near the ventral side are approximately tangential, so that many of the nuclei that appear to be deep in from the periphery are in reality near to the surface. ‘This needs to be kept in mind in interpreting the figures. It is a conspicuous fact that many of the nuclei shown in Fig. 58 resemble in shape and appearance the retinular nuclei found in sections of regenerating eyes. Further resemblances between these and retinular nuclei are their tendency to stain deeply and the general direction of their long axes which is at right angles to the surface. These facts considered in relation to each other leave but little doubt that these elongated nuclei represent the retinular elements in a regenerating eye. At one point in Fig. 37 (c.c.) the rudiments of two crystalline cones have appeared. This is additional evidence that an eye is regenerating, imperfect and Regeneration in Compound Eyes of Crustacea 205 é s abnormal as it may be. The pigment shown in this figure repre- sents the maximum amount seen in any section. For the most part it presents no definite arrangement but lies in irregular masses clustered within a fairly well defined area. In a few sections, however, a part of the pigment shows a tendency toward a normal arrangement as if the pigment granules were contained within the processes of the retinular cells, and rudimentary ommatidia can be recognized (Fig. 59). Any attempt to explain the phenomena presented by the eye under discussion may appear somewhat premature, since it pre- sents a practically unique case so that but little data for comparison is available. In the first place this is the only well established case of any attempt of Palamonetes to regenerate ommatidial structures after the removal of a large part of the optic ganglion. It seldom happens that any regeneration takes place from the eye stump of Palamonetes when no more than half of the optic gan- glion remains. Before attempting to explain the phenomena, therefore, it is well perhaps to examine the results of other observa- tions that may suggest an explanation. From the evidence obtained from the study of normally regener- ating eyes the indications are that the first new regenerated tissue is largely of an indifferent character, 7. ¢., capable of giving rise to different structures, as determined by conditions more or less external to itself. It has been seen that in the regenerating eye the primary hypodermis gives rise to the cells which develop into the different structures of the ommatidia. In the earliest begin- ning of differentiation, if a cell divides so that the plane of division is at right angles to the surface the two resulting cells are hypo- dermal cells. If on the other hand the plane of division is parallel to the surface the inner one of the pair thus formed becomes a retinular nucleus and the outer one remains hypodermal in char- acter. At this stage the only apparent difference between the two nuclei is in their respective positions. In later cell generations when the division plane is parallel to the surface the inner nuclei of the pairs formed become crystalline cone nuclei. “Thus we have cone nuclei, retinular nuclei and hypodermal nuclei indistinguish- able except for their relative positions. Apparently the subsequent 206 Mary Isabelle Steele differentiation of each is conditioned by their relative positions in respect to each other, to the surface and to the old parts present. That the relation of the old tissues to the regenerating tissues is a determining factor in the regenerating of the new structures has been maintained by several workers. Child (?04) who especially emphasizes the idea, says: “The fate of the new material must be regarded as depending essentially upon its relation to the old parts.” That the regeneration of one structure may be dependent upon the presence of erother has been shown by Lewis (oz). He found that a lens could be developed from any part of the ectoderm of a frog embryo by transplanting the optic vesicle and allowing it to come in contact with the ectoderm. In this case it appears that the actual contact of the two tissues constituted a determining factor and that the new conditions have arisen on account of the new relations of the two tissues. From the instances referred to the following inferences may be drawn—first, newly regenerated tissue is largely indifferent in character; second, the differentiation of the new tissue is largely conditioned by its relation to the old tissue. “The above inferences may be used in suggesting an interpretation of the special phe- nomena under eon nderuian: To begin, it is evident that the particular individual now being considered exhibited a more than usual degree of physiological and regenerative activity. In no other way could we account for so aici more new tissue than is ordinarily regenerated by a stump of this length. The sections show that a large part of the new tissue still presents an undifferentiated appearance, although in certain regions differentiation has begun. Just above the cut end of the old muscle band (Fig. 35) the new tissue appears thickened, arranged in definite fibers and is apparently continuous with the muscle band suggesting that connections between the muscle and the chitinous covering of the eye had been reéstab- lished. Sections passing through the stump near its ventral side show rudimentary ommatidia in process of development. Just why ommatidia should appear on the side of the old stump is not at first sight apparent. One possible explanation of this phenom- Regeneration in Compound Eyes of Crustacea 207 enon has, however, suggested itself. The differentiation of new ommatidia appears to depend largely upon the reéstablishment of connections between the optic ganglion and the new tissue. ‘That such is the case is suggested by the fact that in regenerating eyes the retinular processes reach the optic ganglion before the cones begin to differentiate and before any pigment is deposited i in the retinulz. Further, according to Parker (/oc. cit.) inthe embryonic development of the lobster’s eye from the earliest beginning of differentiation there is a connection between the ommacaial region and the optic ganglion and this connection is never lost at any stage in the development of the eye. Since the pigment appears in the retinulz after they have formed connection with the ganglion its presence in this stump (Fig. 36a) may be an indication that the connections between the optic ganglion and the new tissues have been formed and consequently that further development of ommatidial structures has been initiated. This suggested explanation does not of course give any reason why ommatidial development should begin on the ventral side rather than elsewhere. The following Seueiaa.s is suggested. Cross and longitudinal sections of the roswell eye stump oe that the optic ganglion extends somewhat nearer the surface toward the anterior ventral side. On this account perhaps rudimentary ommatidia have developed on the ventral side first because the distance between the new tissue and the optic ganglion was shorter so that nerve connections were more quickly established in that region than elsewhere. Other cases of aberrant regeneration have also come under observation. These, however, have been produced apparently by pathological conditions. One or two of the more interesting cases will be described. Fig. 19 represents dorsal surface views of a Palamonetes eye thirty-one days after the removal of the upper part of the ommatidial portion. ight days after the operation a moult occurred. A second moult pecceed fifteen days later and it was seen that the greater part of the ommatidial structures had disappeared. Fig. 19 shows a mass of pigment Just distal to the optic ganglion. Above this pigment there is a considerable area of transparent tissue which shows no external evidence, however, of 208 Mary Tsabelle Steele being differentiated into ommatidia. Sections of this eye show a considerable development of abnormal pigment. At one side are seen a limited number of regenerating ommatidia that have failed to differentiate normally. The arrangement of the abnormal pig- ment is of a character frequently seen in short stumps but not generally found where the injury, as here, has not involved the optic ganglion. The probable cause of this abnormal pigment deposition will be considered in another section. ‘The point of chief interest here is that we find two sorts of development going on side by side, one region developing normal structures and a contiguous region developing abnormal structures. Another similar, though rather more exaggerated, case is fur- nished by a small hermit crab (Fig. 25). The eye had been in- jured before the animal was brought into the laboratory. No moult, however, had taken place since the injury. About two weeks after being brought into the laboratory the crab moulted and three days later was killed. Fig. 25 represents the ventral view both of the injured and uninjured eye three days after the moult. It is seen that the injury involved the whole ommatidial portion and appar- ently a part of the optic ganglion. ‘The distal surface of the regenerating eye presents a very irregular outline. On the inner border a peculiar protuberance has developed and two separate pigment areas are apparent. The upper of these two areas suggests that small ommatidia have been regenerated. Externally the lower pigment mass suggests no probable explanation of its character. Sections of this eye show a number of interesting points. In the first place it is demonstrated that the upper and smaller pigment area belongs to the retinule of small but almost completely developed ommatidia. All the structures of a typical ommatidium have been differentiated with the exception of the corneal facets and the spindle shaped enlargement at the base of the rhabdom. The lower pigment area is an abnormal pigment deposition similar to the preceding case. The protuberance developed on the inner border of the eye seems also to be formed of abnormal tissue. _ Its interior is entirely made up of a loose irregular network of tissue containing a number Regeneration in Compound Eyes of Crustacea 209 of faintly staining nuclei. These tissues resemble very closely the depigmented tissues of the abnormal pigment masses. ‘This resemblance suggests that possibly the two abnormal ape sEauee® have had a common origin. A case partially resembling the one just described was observed in a green shrimp, Palemonetes viridis. The original injury consisted in the removal of a small part of the top of the eye. “The eye was operated upon August 1. On the ninth of the same month the animal moulted and was preserved. Figs. 17 and 18 represent dorsal and ventral views of the injured eye. The dorsal surface of the whole eye is shown in Fig. 17. Fig. 18 represents the distal end of the eye from the ene side Ate under greater magnifica- tion. A pigmented mass similar in general outline to the retinular area in the normal eye is visible through the transparent outer tissues. Distal to the pigmented portion is a considerable area of transparent tissue with flecks of pigment scattered through it. The distal contour of the eye is irregular because of a swelling or protuberance similar to the one on the hermit crab’s eye previously described. This eye also shows an unusual development of new tissue considering the time in which it was produced. Sections of this eye show that the optic ganglion had not been injured, that not all of the ommatidia had been removed and that a considerable part of the old pigment remained. The ommatidia that were left have almost completely degenerated, however, and the whole distal portion of the stump is filled with a complicated network of faintly staining cells. There is also absolutely no regularity in cellular arrangement, as is seen in normally regener- ating eyes. For the most part the nuclei are scarcely distinguish- able from the cell-body. Although here and there are scattered nuclei which stain more deeply. There are evidences in some cases that these are nuclei of disintegrating ommatidial structures. Some sections show remains of old cones associated with the darkly staining nuclei. Comparison of others of these deeply staining nuclei with the nuclei of partially depigmented cells shows a similarity between the two which suggests that the former belong to cells in which pigment secretion has lately begun. It is evident that most of the pigment masses present are the 210 Mary Isabelle Steele remains of the old ommatidia although they are greatly scattered through the new tissue. A few dense cysts of new pigment, however, have been formed and other pigment secreting centers have begun to appear. From these observations it seems apparent that had the animal lived the entire mass of tissue sooner or later would have been densely packed with pigment cysts and that very probably new eye structures would not have regenerated. For we have seen in the preceding two cases that an abnormal secretion of pigment stopped, apparently, ommatidial regeneration after it had begun. It does not seem too much to assume then that in this case normal regeneration. of tissues would have been precluded by such an abundant development of abnormal tissue. C EYE STUMPS THAT SHOW AN ABNORMAL DEVELOPMENT OR NO REGENERATION The instances described in the preceding section apply partic- ularly to those unusual types in which the ommatidia have begun to regenerate and this process has been more than balanced by opposing factors. “This leads naturally to a consideration of cases in which there is either no regeneration or only an abnormal development of pigment. I. Abnormal Development of Pigment Most of the examples of abnormal pigment secretion were afforded by Palzmonetes in which the optic ganglion was more or less injured. Usually in any of the forms studied eye stumps that contain no more than half of the optic ganglion show no normal regeneration aside from the cuticle and hypodermis. Any attempt to regenerate other tissues produces either scattered strands of connective tissue or abnormal masses of pigment. “These pigment masses most frequently appear collected in nodules or cysts and are usually enclosed in a sort of connective tissue sheaths. Fig. 43 represents an eye stump of Palamonetes, showing one of these pigment depositions. Fig. 72 shows an outline section through this stump from which the relation of the pigment to the normal issues can be readily made out. Fig. 43 shows in detail the Regeneration in Compound Eyes of Crustacea 211 appearance of the pigment outlined in Fig. 72. An examination of Fig. 73 shows that the deposition of pigment appears to begin at several centers. These centers gradually increase in size. There also seems to be a tendency for the several centers to fuse with each other. It is further seen that the pigment cells or masses vary from very large to very small areas. A study of depigmented sections suggests that these smaller pigment bodies arise in one of two ways: first, by an out-pocketing of the cytoplasm, which after becoming distended with pigment separates from the parent mass and second, by an unequal division of the cell. It is possible that the latter is the true method for all cases. But it was not possible to determine this point with cer- tainty. When a pigment cell has become gorged with pigment the nucleus is much changed and distorted. And even after the most thorough depigmentation it cannot always be identified. Consequently it may be that the smaller masses, in which no nuclei are visible, are not mere masses of cytoplasm that have been constricted off but are the result of unequal cell division. Fig. 75f shows a small group of depigmented pigment bodies. In the larger masses nuclei are visible. In the smaller bodies nuclei cannot be determined with certainty. The amount of pigment within a cell varies. Some cells contain only a few scattered granules while others are so completely filled that they appear to be black homogeneous masses. In these more densely filled cells the pigment appears to have fused into solid brittle masses that can be crushed like starch grains. g in Fig. 75 represents one of these masses after it has been crushed. The pigment is dissolved from the sections with the greatest difficulty. Mayer’s chlorine method was generally used for this purpose. But in removing the cyst-like depositions of pigment it was found that alternate treatment with the chlorine method and with one-twentieth per cent KOH in 70 per cent alcohol gave equally as good and more rapid results. Even with this treatment twelve to twenty-four hours were required to remove the pigment from sections 6 thick. Frequently this failed to dissolve the dense pigment masses. In Fig. 75 the dense, crushed pigment mass g lies in the same section with the group of depigmented 212 Mary Isabelle Steele cells shown in f. It frequently happens that not all of the tissues included in what may be regarded as a single pigment region are pigmented (Fig. 73). The unpigmented tissues, shown in the area represented in this figure, contain but few recognizable nuclei. Here and there are cells that show a few pigment granules and occasionally small groups of such cells. These facts together with the general appearance of the tissue suggest that eventually the entire area might have become packed with pigment. Figs. 29, 32, 44 represent eye stumps of Palamonetes that show somewhat different types of these abnormal pigment formations. Fig. 32 presents a rather unusual type. Externally the pigment appears as thickly scattered granules instead of a dense black mass as in most cases. Sections of this stump show a small quan- tity of new tissue lying at the distal end and alongside the nerve stump. ‘The cells composing the new tissue are closely packed, large, granular, and their nuclei do not take up iron haematoxylin at all. Along the side of the eye stump a number of small pigment cysts appear but for the most part the cells of the new tissue are not yet densely pigmented. Many of them, however, show numerous pigment granules. This particular specimen shows less of the connective tissue-like, fibrous network than is usually found in the pigment areas. Apparently this stump shows an early stage in abnormal pigment secretion. The other two cases figured show dense masses of pigment. Fig. 44 presents a single compact mass. In each case sections show the pigment arranged in the characteristic cysts, such as are seen in Fig. 73. One additional fact of interest is shown in Fig. 44. The pig- ment cysts in this case do not lie wholly above and distal to the remains of the optic ganglion but are embedded in the end of the optic stump. Apparently the upper part of the ganglion stump has degenerated and given place to the pigment. ‘This is not an unique instance as several other stumps have presented a similar phenomenon. ‘There was one case in particular in which there were several small pigment cysts embedded in different portions of the remains of the optic ganglion. ‘The ganglion, in this case had almost entirely degenerated, apparently. ‘Vhis animal had been preserved in alcohol, however, and it was consequently impossible Regeneration in Compound Eyes of Crustacea 213 to determine just how much of the abnormal appearance of the tissue was due to degeneration before the death of the animal and how much was due to disintegration after its death. It is evident that this sort of pigment development, whatever may be its cause, does not belong to the normal regeneration of an eye. Further, it appears probable that the causes leading to its formation are of such a nature that they inhibit the true regenera- tive process. The last two cases described in the preceding section furnish evidence of this. In the eyes shown in Figs. 19 and 20 regeneration of normal ommatidia had begun but was limited by some opposing factor. These causes not only inhibit the true regenerative processes after they have begun but it is also probably true that they even prevent true regeneration from beginning. All the comparative evidence that we have indicates that in the case of the Palamonetes viridis previously described (Fig. 17) a new eye would never have developed. The whole distal end of the stump was filled with a mass of abnormal pigment depositing cells. Although this case is striking it is not exceptional. Similar conditions have been found in varying degrees in other eye stumps. There is sufficient similarity in all the cases of abnormal pigment deposition to indicate that they have in certain respects a common cause. It is important to point out some of these similarities in greater detail. A striking resemblance exists between the broken down retinula of an injured eye and the pigment secreting cells. In the early stages of the disintegration of the ommatidial structures the nuclei of the retinule frequently become separated from the retinular processes. Each nucleus becomes surrounded by a rounded mass of cytoplasm which apparently has no connection with other structures. ‘The nuclei become polymorphic and not infrequently appear divided. As the disintegration proceeds these rounded nuclear cells usually disappear, but, as mentioned in a preceding section, the broken down masses of pigment remain. These rounded remains of the retinulz can be identified from a few hours up to sixteen days after the injury. They are always seen a few hours after the injury although they may not always be present in eyes examined in a week to two weeks after the opera- 214 Mary Isabelle Steele tion. ‘This shows that in some cases they disintegrate much more rapidly than in others. In some eyes examined twenty-five to thirty-five days after the operation similar rounded cells with polymorphic nuclei are found in numbers, increasing by amitotic division. In still other cases, cells of this character containing ‘pigment granules are found. Fig. 75a, b, c, d, e, f, g represents a series of groups of these rounded cells with polymorphic nuclei. ‘These groups were taken from crayfish, hermit crabs, Crangon and Palamonetes, represen- ing in all seven species. ‘The first three groups, a, b and c, show the appearance of breaking down retinula, seventeen hours, thirty-nine hours and sixteen days, respectively, after the injury; d, e and 7 show the secretion of abnormal pigment as found in eyes ten, twenty-three and sixty-seven days respectively after the injury; 7 represents a group of depigmented cells that were so filled with pigment that without depigmentation no structures were visible. An examination of this series cannot fail to show the similarity between the breaking down retinulz and the pigment secreting cells. Particularly is this so if it is remembered that, except a and b, no two groups are taken from the same species. These facts taken together have suggested that the immediate cause of the pathological pigment secretion is the abnormal activity of old retinulz which have not completely broken down. It has already been mentioned that after an eye has been operated upon the pigment from the injured retinule frequently becomes greatly scattered among the other tissues. Not only does the retinular pigment become scattered but in some cases the rounded retinular cells, also, are found considerable distances down the stalk and on the side opposite the injury. ‘These instances were observed in eyes examined from fifteen to twenty days after the injury. It seems probable that some of these metamorphosed retinular cells become embedded in other tissues, then later divide amitotically and begin to secrete pigment. ‘The nodules of pig- ment previously described are the result. In some cases the multiplication of these pathological cells takes place rapidly so that large areas are occupied by them. Fig. 70, which is from a section of the eye shown in Fig. 20, represents suchacase. A Regeneration in Compound Eyes of Crustacea 2155 relatively large amount of new tissue was regenerated by this eye, and sections show that normal regeneration had begun (Figs. 68, 69). Anew hypodermis was completely differentiated and on one side the differentiation of ommatidia was taking place (Fig. 69). The greater part of the new tissue was made up, however, of cells of a character known to be abnormal. ‘The hypodermal cells are practically the only cells that appear normal. With the exception of a small area on one side almost the whole of the interior is filled with rounded cells containing polymorphic nuclei. Fig. 71 represents a partof Fig. 70 more highly magnified and showing the structure in greater detail. A comparison of these two figures with the series shown in Fig. 75 cannot fail to show a striking similarity. Altogether there is strong evidence that the failure of the old retinulz to disintegrate completely is the immediate cause of abnormal pigment deposition, in many cases at least. Further, there is some evidence that regenerating retinula may sometimes become involved in the abnormal secretion of pigment. Group d, Fig. 75, represents a probable case of this sort. The group was taken from the regenerating part of the eye in a region where there is positive evidence that some normal regeneration is taking place. A few of the cells in the group still show but few pigment granules and show elongated nuclei, characteristic of regenerating retinulz (Fig. 75d, ret.n.) The appearance of the network of tissue with which many of these pigment nodules are associated still remains to be accounted for. Evidences which point to the origin of this are not so numer- ous as are the evidences that the old retinulz form the centers for the pigment secretion. In some cases these pigment nodules are found embedded in the hypodermis of the eye stalk, in other cases in the membrane surrounding the optic ganglion. In such instances it seems probable that the fibrous network supporting these nodules is due to the hypertrophy of the normal tissue immediately surrounding the pigment deposits. In those cases, however, where a great mass of this fibrous network developed it seems to have had a different origin. ‘There are three particularly striking instances of the unusual development of this abnormal 216 Mary Isabelle Steele tissue, each furnished by a different form. Sections taken from a crayfish eye fixed sixty-two and a half hours after the operation show a condition that is apparently an early stage in such develop- ment. At this stage the network is not yet compact and ts found in chains of elongated cells, showing nuclei dividing amitotically. These chains of cells run in all directions but do not appear to develop from the hypodermis. Some sections show these chains extending from the injured retinule which still surround the remains of the old cones (Fig. 74). This suggests that the old retinular cells are undergoing rapid multiplication. The chains of cells found in the crayfish eye differ in the follow- ing respects from the network of abnormal tissue found in Pala- monetes viridis and hermit crab, shown in Figs. 17 and 25. In the cases of the hermit crab, Fig. 25, and of Palamonetes viridis, Fig. 17, the cells constituting the network are no longer recog- nizable as chains and the nuclei no longer stain deeply nor appear to be dividing. These differences may be accounted for by the following facts. First, the whole available space in the stump of the eye was completely filled with the network in the eyes of hermit crab and Palemonetes viridis and the chains of cells had become so completely interwoven that their original character could no longer be recognized. Second, it is probable that the nuclei no longer stain deeply because the cells have ceased active division. It has been shown that the cells cease to divide actively soon after the secretion of pigment begins. In these cells the secretion of pigment had begun. Assuming that these apparent differences have been accounted for we may now turn to their likenesses which suggest a similarity of origin. The most suggestive likeness is that in each of the three forms the abnormal tissue appears to have developed outward from the base of the wounded area rather than inward from the periphery. The most striking evidence of this is the fact that masses of old pigment appear near the periphery as if they had been carried outward by the growth of the new tissue. These abnormal tissues, in the case of hermit crabs and Palamonetes, lie close against the cuticle and several layers of the cells are flattened as if they were the oldest cells and had been pressed against the \ Regeneration in Compound Eyes of Crustacea 217 cuticle by the multiplication of the cells beneath. In these cases a true hypodermis is not distinguishable. These facts suggest that the migration and pathological development of the old retinulz are responsible for most if not all of the cases of abnormal pigment deposition. This of course does not explain what induces this pathological development. The initial cause of this development, in the cases where an abundant network of tissue has developed, was perhaps due to some infection at the time of the operation. ‘This is suggested by the fact that, in the crayfish eye described above, a great deal more tissue had developed abnormally in sixty-two hours than is usually developed normally in ten days or two weeks, and by the fact, also, that in the eye of Palamonetes viridis a very unusual amount of new tissue had developed during the first nine days after the injury. The more frequent cases in which the pigment secreting cells appear as rounded cells, containing polymorphic nuclei similar to the disintegrating retinular cells, seem to be produced by causes somewhat different. In some specimens examined some time after the injury these cells show no signs of rapid multiplication. It seems probable that these cells are old retinule that have retained one of their characteristic functions, the secretion of pigment. Since it is an observed fact that the old retinulz become metamorphosed and wander to different parts of the stump where they have been found dividing amitotically. While the above facts are strongly in favor of the conclusion that the abnormal pigment-secreting tissue is due to the development of old retinular cells yet the proof is not absolute. A series of stages of this development, not more than two days apart, would have to be examined in order to be certain of the absolute truth of this tentative conclusion. 2 Eye Stumps that Show No Regeneration It now remains to consider the other phase of the subject out- lined in this section; namely, those cases in which there is no regeneration further than the healing of the stump. A number of these cases present anomalies in that there is no apparent reason 218 Mary Isabelle Steele for their failure to regenerate. There were among the hermit crabs several parallels between those that regenerated an eye and those that did not, so far as conditions were concerned. In a series of fourteen hermit crabs that had the ommatidial portion of the eye removed five regenerated an eye and nine did not. All were kept as nearly as possible under the same conditions. The part of the eye removed in the original operation was about the same for each individual. All were operated upon at the same time in the same way. Some of those that regenerated an eye and some that did not moulted upon the same day after the operation. Consequently the physical condition of these specimens were apparently similar. Compare Fig. 13 and Fig. 21. Each of the hermit crabs from which these figures were taken moulted twelve days after the operation. The hermit crab from which Fig. 13 was taken was killed at the end of thirty-eight days and the other at the end of sixty-seven days. The latter lived nearly twice as long yet it shows no signs of regeneration. More of the optic ganglion remains in the stump shown in Fig. 21 than in Fig. 13. Again, compare Fig. 15 with Fig. 21. The two crabs from which these figures were taken were operated upon at the same time, moulted approximately upon the same dates and were killed sixty-seven days after the operation. ‘The stump shown in Fig. 21 shows no regeneration while the one shown in Fig. 15 has regenerated an eye perfect in all of its details. The number of cases might be multiplied but these given are sufficient to show the parallels presented by individual cases. Instances of this sort are confined in great part to hermit crabs. A number of shrimp, however, failed to regenerate even when the optic ganglion was not injured. The same is true for Crangon which in several instances failed to regenerate normally even after the removal of only a small part of the eye. In most cases sections of such eyes that did not regenerate show no recognizable pathological conditions. In the case shown in Fig. 21, however, there was found what seemed to be the beginning of pathological pigment development. Externally there were no signs of pigment formation. ‘The regenerated tissue consisted of a Regeneration in Compound Eyes of Crustacea 219 heavy cuticle, a hypodermis and some loose strands of tissue extending from the hypodermis to the distal end of the stump of the optic ganglion. Grouped at the end of optic ganglion stump and scattered in the loose tissue above it were a few cells of the characteristic pigment-secreting type. But none of these cells had yet become densely filled with pigment (Fig. 75¢). It seems rather improbable that so little abnormal tissue in which scarcely any secretion of pigment had taken place could have been the sole cause in the prevention of normal regeneration. Particularly is this true when it is remembered that instances have been observed in which practically complete ommatidia were regenerated in eyes containing great masses of abnormal pigment (Figs. 19, 25). VI REGENERATION AFTER REMOVAL OF THE GREATER PART oR ALL OF THE Optic GANGLION There now remains for discussion those cases in which the whole or most of the eye stalk was removed and consequently either all or the greater part of the optic ganglion. Palamonetes, Crangon and Heemit crabs will each be considered independently since Mihie differences presented by them are such as to require separate treatment. A HERMIT CRABS Of a total of sixty hermit crabs operated upon twelve died as a result of the operation, a loss of 20 per cent as against 55 per cent of the Palamonetes after a similar operation. ‘Thirty-six of these remaining crabs moulted from one to three times and lived from twenty-three to one hundred and ninety-four days. These thirty- six crabs fall into two groups: those that regenerated an antenna- like appendage in place of an eye and those that showed no particu- lar regeneration. I Regeneration of Heteromorphic A ppendages Ten crabs in all regenerated an appendage from the old eye stump. In but one case was more than thirty-two days required for the appendage to become apparent. All of these appendages are very small none exceeding in length the normal eye stalk. It 220 Ma ry Isabelle Steele is probable, however, that they would have increased both in diameter and length had the experiment covered a longer period of time. None were distinguishable before the occurrence of a moult. In each case recorded the appendage appeared after the first moult. Most of these appendages were definitely segmented after the first moult, in some instances several segments being developed within twenty-one or two days. But none show any indication of being divided into parts corresponding to the exo- and endopodites. Figs. 23 and 30 show two appendages that were present twenty- one and twenty-two days respectively after the injury. Neither appendage exceeds in length the squame at the base of the normal eye stalk which measures but little more than one-fourth of the whole length of the normal stalk. One appendage bears a con- siderable number of large tubular hairs. ‘The other shows none whatever. Each is seen to consist of several segments. Five segments are distinctly visible in Fig. 23 while in Fig. 30 there are six or seven though they are not distinctly differentiated. In Fig. 30 the appendage projects outward at a broad angle. Fig. 23 1s unique in that it curves in toward the median line nl suggests in its general shape and position the squame at the base of the opposite eye. The bifid tip of this appendage is probably due to some injury that occurred at the time of the moult. This explan- ation is suggested by an examination of the specimen. Figs. 34 and 45 represent two other appendages that appeared twenty-nine and thirty-two days, respectively, after the injury. These types differ somewhat from the preceding two. It is to be noted in both cases that the original operation did not include the squame at the base of the eye. ‘This is a good indication that at least a part of the proximal segment of the optic ganglion was left. The specimen shown in Fig. 34 was sufficiently transparent so that the optic nerve could be observed extending into the base of the segmented appendage. ‘The specimen from which Fig. 45 was taken was fixed in Flemming and the consequent darkening of the tissues prevented an accurate determination of the length of the stump of the optic nerve. But the nerve stump could be seen extending well into the base of the new appendage. Regeneration in Compound Eyes of Crustacea 221 The appendage shown in Fig. 28 developed in twenty-four days with the intervention of one moult. It is of interest because of the indications that the optic nerve has extended through almost the entire length of the regenerated appendage. It is also of interest because of the ganglionic swelling that appears to be associated with the nerve in its distal half. Fig. 41 shows an unique type in that the appendage is curved closely back until the free end almost touches the head. Although this appendage is made up of several segments it was rigid from its first appearance. The remaining examples of these appendages are of approxi- mately the same character as those figured. They belong chiefly to the type shown in Fig. 34, except that two of them show a larger number of tubular hairs. One of these belongs to a speci- men that moulted twice and was not killed for sixty-seven days after the operation. ‘The regenerated appendage shows but little advance over those that were fixed at the end of half that time. It is still no longer than the normal eye stalk and shows no greater number of sensory hairs than are seen in Fig. 30. ‘The additional facts obtained from an examination of the sections will be referred to at the close of this section in the general discussion of their significance. 2 Cases that Show No Espectal Regeneration As was stated above out of the thirty-six crabs that moulted one or more times only ten developed heteromorphic appendages while twenty-six showed no particular regeneration. ‘The propor- tion 1s a little more than 30 per cent to a little less than 70 per cent in favor of those that showed merely a healed over stump. The stumps that show no actual regeneration present a variety of shapes and characters. None of them, however, show any signs of pathological pigment development. From all appear- ances the failure to regenerate in most instances was due to a lack of sufficient regenerative activity to produce the new tissue neces- sary. In some cases where the eye was taken off even with the head the wound healed over leaving a smooth surface, not so much as a slight elevation marking the former position of the eye. In 222 Mary Isabelle Steele most instances, however, a longer or a shorter stump remained. It is impossible to determine by surface examinations how large the stump was originally for it decreases in size after the operation. Sometimes the stump of the optic nerve and ganglion shrinks to two-thirds of its original volume. Fig. 31 shows a short rounded stump which evidently contains a part of the proximal segment of the optic ganglion. ‘The stump, originally as broad as the base of the opposite eye, has, after one moult twenty-three days after the removal of the eye, shrunk to one-half of the original mass. The remains of the optic nerve seem to come flush against the end of the stump, showing that no new tissue has been developed distal to it. Fig. 39, thirty-two days after the injury, shows a stump more than one-third the length of the normal eye. Yet sections show no indication that any definite structure is being regenerated. It is useless to multiply figures on this phase of the question. They only serve to show how completely is lacking any indication of regeneration. The following table will serve to show that time cannot be considered the chief factor in regeneration. Oh of Experiment begun Closed Days | Moults Regeneration specimen | I October 16 June 30 | 106 one none 2 November 27 — June 48 | 158 | one none 3 October 16 April 2/| 194 | one none 4 May 26 = June 16 | 21 one segmented appendage 5 May 26 = (| July 4 39 one none 6 July 9 September 3) 56 | one none 7 July 9 August 3 | 2 one segmented appendage 8 July 9 (August 16 39 one segmented appendage 9 July 9 September 14 67 one none 10 July 9 |August 6 28 one segmented appendage The last five examples given in the preceding table are taken from the same series. Evidently the conditions here were more favorable than usual. The original number of the series was twenty-five. Six of these died either from the effects of the oper- ation or soon afterwards. Of those remaining five others were lost through an accident. Out of the fourteen for which there is a Regeneration in Compound E yes of Crustacea 223 complete record nine developed heteromorphic appendages and all of them within thirty-three days. Two of the remaining five, which moulted at the end of twenty-one days and then died, might perhaps have developed an appendage had they lived through a second moult. Each of them showed a very small bud where the eye had been removed. B CRANGON I Regeneration of Heteromorphic A ppendages In some respects Crangon appears to be a favorable form for experimental work. ‘They are less disastrously affected by the operation than the others worked upon. ‘The entire eye was removed from twenty-two Crangon and not one of the number died from the effects of the operation. This entire number was of the same series. ‘The experiment covered a period of thirty-three days, August 3 to September 4, inclusive. During that time with one exception each individual moulted at least once and fourteen moulted a second time. ‘Three of those that moulted but once were eaten by their comrades soon after the moult. The evident hardiness of the Crangon and the frequency of the moults would seem to be favorable conditions for regeneration. Results, however, show only one individual that regenerated a heteromorphic appendage, the others showing no regeneration. Fig. 38a and b shows surface views of this one regenerated ap- pendage. The animal which developed it moulted on the fourth day after the operation. At that time there was no evidence of regeneration. Seventeen days later another moult occurred and an appendage of six segments, with sensory hairs near the tip, appeared. The appendage measures four-fifths of the length of the eye on the opposite side and projects forward at the same angle. The outline of the optic nerve can be seen extending through the proximal half of the appendage. 2 Cases that Show No Especial Regeneration Twenty-one out of twenty-two Crangon showed no regeneration. Four of these died within nine days after the operation and so 224 Mary Isabelle Steele perhaps should not be counted either way. ‘There were then seventeen negative cases against a single positive case. The eye stalks of Crangon are very short and the sections of the optic ganglion are crowded very close together, and extend well into the base of the stalk. Hence it not infrequently happened that a part of the ganglion remained in the stump. A number of these stumps have been sectioned and none of them show any regenerated tissue except the hypodermis and cuticle. Spots of pigment are often seen at the end of the stump but since the whole stalk of the normal eye is heavily pigmented this does not seem to be significant. Figs. 24, 37, 40 and 42 show a variety of appear- ances which the stumps presented. The accompanying table shows the number of moults which occurred. No. of 4 | spec- Ep oot First moult | Second moult | Third moult | Date of death sees = begun ation imen 8 August 3 August 9 August 21 none September 2) none 10 August 3 August 10 August 22 none September 4 | none 13 | August 3 August 12 August 28 none August 28 | none 15 August 3 August 4 August 13 August 19 September 2. none 17 August 3 August 14 August 29 none | September 4] none In some cases the eye stump is extremely short while in others it is longer so that a part of the ganglion remains. All of the specimens included in the table except No. 17 have been sectioned but none of them show any signs of regeneration. Sections of No. 10 show that nearly half of the optic ganglion was left but no regeneration is taking place. A very much folded and wrinkled cuticle with short hairs projecting from it covers the stump. Even No. 8 (Fig. 37), short as it appears, is found to contain the proxi- mal end of the optic ganglion. In this case the stump has merely healed over but no new tissue has developed. In several other instances not shown in figures the eye had been totally removed so that not even a short stump is visible. In most such cases the cuticle is wrinkled over the spot where the eye had been. The wrinkles and folds on some of the stumps figured shows the com- mon tendency. These folds are chiefly due probably to the shrinking of the inner tissues of the stump. Regeneration in Compound Eyes of Crustacea 225 C PALAEMONETES Out of nearly three hundred Palzmonetes not a single individual regenerated any sort of an appendage when all or nearly all of the optic stalk was removed. It is true that more than 50 per cent of them died from the operation or soon after. Often half or two- thirds of a series died within twenty-five or thirty minutes after the operation, and in some instances the proportion was still greater. (See Table 1.) Palamonetes were by far the least resistant of any of the forms operated upon. There were, however, over sixty individuals that lived from twenty to one hundred and twenty- four days and moulted from one to three times. Considering the results of these experiments it may be said that Palemonetes vulgaris does not regenerate an antenna-like append- age in place of an eye. Herbst would, perhaps, insist that these results were due to a lack of time or to a failure to remove all of the optic ganglion. ‘This latter objection in many cases could not be urged. The eye stalk in Palamonetes is long and the optic nerve extends well into its base. And in these experiments the eye was so completely removed that not even the vestige of a stump remained. Consequently there was no possibility of leaving any part of the optic ganglion. Part of the brain even was removed with the eye in two series. In regard to the other objection naturally there is no positive proof that results might not have been different in a longer period of time. There are strong reasons, however, for believing that time would have made no essential difference. Chief among these reasons is the fact that in the regeneration of any other organ Palemonetes needs but little more time than the hermit crabs and less time than Crangon. In three parallel series of experiments upon the regeneration of the first antenna after its total extirpation it was found that Pale- monetes regenerates a first antenna as quickly and as perfectly as either Crangon or hermit crabs. In another parallel series of experiments upon the regeneration of the second antenna it was found that Palamonetes regenerates this appendage rather more rapidly than either hermit crabs or Crangon. Palamonetes may regenerate a first or second antenna in about thirty days. Neither 226 Mary Isabelle Steele hermit crabs nor Crangon regenerate these appendages in less time. In the regeneration of a functional eye it was seen in a previous section that hermit crabs regenerate rather more rapidly than Palzmonetes but that Palamonetes regenerate more rapidly than Crangon. Palamonetes may regenerate a functional eye in thirty to thirty-five days. It is seen, therefore, that appendages are regenerated by Palzmonetes in approximately the same time as they are regenerated in hermit crabs and Crangon. Consequently it does not seem to be assuming too much to express the conviction that a greater amount of time would have made no essential difference in the results of these experiments in which the entire eye of Palaemonetes was removed. Below 1s a brief table showing results obtained by removing the entire eye of Palamonetes. This table does not include all the individuals of any one series but it is entirely representative. No. of spec- | Pexpeximent First moult | Second moult Eaperoen Days esos nen | begun | closed ation | I November 5 | November 27 | December30 | February 3 Cle} none 2 | November g | November January | March 13 124 | none 3 November 9g | November January February 12 95 | mone 4 January 1] January 15 | February 7 February 2 55 | none 5 January 1 | February 20} none | February 28 59 | mone 6 | March 5 | April 24 | none | April 24 50 | none 7 | April 19 | May 5 | June 8&25 | July 3 71 | none 8 May 10 | June none | July 4 55 | none 9 July ~ 10 | July 26 | August 6 | September 2 54 none 10 July 10 | July 27 {August 9\ September 14 66 | none \ September 6/ | II July 20 | July 30 | none August 18 29 none 12 July 10 | July 18 | July 31 | August 2) 24 none 13 July 20 | July 30 | August 14 | August 15 27 none 14 July 20 | July 30 | August 10 | August II 22 none 15 July 20 | July 24 | none | August 20 30 none 16 July 30 | August 8 | August 27 | September 4 36 none It will be seen that Palamonetes have been under observation practically every month in the year. ‘The results in each instance are negative. Fig. 35a, b, c represent some of the stumps that show new tissue distal to the nerve stump. Most of the cases, however, Regeneration mn Compound Eyes of Crustacea 2,2,7 regardless of the time of the experiment and the size of the stump, are similar to the one shown in Fig. 33, No. 13inthe table. Fig. 33 shows the nerve stump flush against the healed end. ‘The indica- tions from sections and surface views are not such as to lead one to expect that further regeneration would have ever taken place. Fig. 26 represents the only stump that even suggests the develop- ment of a heteromorphic structure. As the table shows this specimen lived only twenty-four days, during which time it moulted twice, and regenerated the tiny mass of new tissue repre- sented by the darkly stippled portion of the figure. ‘The eye was completely removed, the cut coming at the level of the attachment of the eye to the head, represented in Fig. 26 by the line a—b._ Fig. 27 represents a more highly magnified view of the stump. The remaining figures in the series, Fig. 35a, b,c, show the maximum regeneration, yet in none of these cases did the experi- ment cover more than thirty days. Apparently regeneration in most cases proceeds to the forming of the hypodermis and cuticle, which may be extended slightly beyond the nerve trunk by loose strands of connective tissue, and then stops. Fig. 35c shows more than the usual amount of new tissue. The line a—b represents the level of the union of the eye with the head. The unshaded central part of the eye stump shows the remains of the optic nerve; the shaded peripheral portion shows the new tissue. Neither sections nor surface examinations give the slightest evidence of the regeneration of nerve fibers, or of any special differentiation of the regenerated tissue. D THE HISTOLOGY OF THE HETEROMORPHIC APPENDAGES The microscopic structure of the antenna-like appendages has not been considered in great detail because suitable material has been wanting. In the whole series of experiments only ten hermit crabs and one Crangon ever regenerated a heteromorphic append- age in place of the excised eye. Several of these died and were preserved in alcohol. From such material no detailed results were obtainable. Again, the only sections of any particular interest and value are longitudinal ones. “These heteromorphic append- 228 Mary Isabelle Steele ages were so small and so curved that it was almost impossible to obtain satisfactory longitudinal sections. A few points of interest, however, have been observed. ‘These for the most part serve to corroborate the observations of Herbst rather than to add to them. An examination of the appendages in toto show that the old optic nerve either extended as a nerve trunk through the greater part of the length of the regenerated appendage or that other structures were developed in the new appendage which appeared to be continuous with the old optic stump (Figs. 30, 34, 38). Sections confirm the observations made from surface examinations. A large number of intermediate stages would be necessary, however, to determine whether the regeneration of the nerve trunk had been from the optic nerve stump outward or whether peripheral regeneration had developed nerve fibers inward which unite with the optic nerve stump. ‘The fact, however, that the nerve trunk appears more distinctly differ- entiated in the proximal part of the appendage than in the distal may probably be regarded as an indication that the regeneration proceeds from the proximal end outward. Sections of these appendages show that the interior is chiefly occupied by nerve cells and fibers. “The nerve fibers appear to be continuous with the nerve fibers of the old optic nerve stump. The nerve cells are grouped into ganglion-like masses which are scattered pretty generally through the length of the appendage. The brain sheath 1s continuous a the loose fibrous sheath which envelops the mass of nerve cells and fibers. Fig. 77 shows a somewhat diagrammatic section through the brain and the proximal end of a heteromorphic appendage that developed within sixty-seven days after the operation. It was necessary to combine two sections in order to show the continuity of the optic nerve with the nerve trunk of the appendage. ‘There can be no doubt, however, that they form a continuous structure. One feature is noticeable both in sections and in whole prepara- tions. That is, that the optic trunk leading to the regenerated appendage is much smaller in diameter than the one opposite. This fact suggests that probably only a part of the fibers of the optic nerve tere persisted (Figs. 34, 38) Regeneration in Compound Eyes of Crustacea 229 Large blood sinuses are present in the heteromorphic append- ages. Aside from these no tissues are apparent except the hypo- dermis and the fibrous sheath which encloses the nerve bundles and the nerves themselves. In some instances muscles are found in the base of the appendage but these are probably remains of the base of the eye stump. Material has been insufficient to make detailed observations upon the character of the masses of sensory cells found in the ganglion-like groups throughout the appendage. Sufhcient obser- vations have been made, however, to warrant the conclusion that they are concerned with the innervation of the hollow sensory hairs. In a few instances processes have been traced into the bases of the hairs which open by a wide mouth into the interior of the appendage (Figs. 78 and 79). Herbst has considered the microscopic structure of these hetero- morphic appendages in considerable detail and has examined a number of different stages. He describes the nerve cells as grouped into spindle-shaped ganglia with groups of nerve fibers extending from each end of the spindle-shaped masses, the distal bundle of strands being connected with the sensory hairs while the proximal bundle passes inward toward the brain. None of the stages examined by Herbst were younger than about six months, however, and consequently any structures that had developed would likely be much more definitely organized than in the appendages examined in this series of experiments. Herbst considers that these ganglion-like groups of cells have developed from the hypodermis and that in the earlier stages they have no direct connection with the brain. In later stages, how- ever, he describes the proximal bundles of the several ganglia as uniting to pass inward to the brain. But in most cases at least he considers that there is no union with the old optic nerve and con- sequently that the connection of the appendage with the brain is secondary. He mentions the similarity between these epithelial sense cells and those found in the first antenna, homologizing the sensory hairs which are found on the appendage with the olfactory sete found upon the first antenna. Finally he comes to the con- clusion that both in form and structure the heteromorphic append- 230 Mary Isabelle Steele age shows that it should be regarded as a rudimentary first antenna. ‘The structure of the heteromorphic appendage regener- ated by the hermit crab agrees in certain respects with the observa- tions of Herbst upon the structure of the heteromorphic append- ages regenerated by other forms. In other respects the obser- vations made upon hermit crabs are not sufficiently extensive to have any particular weight either way. The most significant difference, however, between these observations and Herbst’s is in regard to the relation of the old optic nerve stump to the nerve bundles extending through the appendage. In the heteromorphic appendages regenerated by the hermit crabs there are several cases in which there can be no doubt as to the continuity of the optic nerve with the nerves in the appendage (Figs. 28, 34, 38). Further, these are found in stages younger than any spoken of by Herbst. The continuity of the optic nerve stump and the nerve trunk of the heteromorphic appendage will be considered in all of its aspects in the general consideration of the problem of such hetero- morphic regeneration. E GENERAL CONSIDERATION OF REGENERATION FOLLOWING REMOVAL OF ENTIRE EYE It has been seen in some cases that hermit crabs and Crangon regenerate an antenna-like appendage in place of an eye. On the other hand, the species of Palamonetes used in this series of experiments has never shown any indication of such regeneration. In view of this, the question w hich naturally arises is w hy do we not find antenna-like appendages growing from the eye stumps of Palazmonetes vulgaris, when hermit crabs and Crangon kept under the same condition do regenerate these structures, and when the phenomenon is of pretty g Sonera occurrence among the Decapods. Herbst has observed the development of an antenna-like append- age from the eye stumps of a number of stalked- -eyed Crustacea belonging to different families. He has even secured a few cases of this heteromorphosis in another species of Palzmonetes (P. varians). Morgan (’g9) was the first to make the observation for Regeneration in Compound Eyes of Crustacea 231 hermit crabs and a like phenomenon has been noted for three species of crayfish, Cambarus virilis and C. gracilis (Steele ’o4) and the blind crayfish, C. Pellucidus testi (Zeleny ’06). Widespread as the phenomenon appears to be, however, no satisfactory explanation of the cause of such heteromorphic regeneration has yet been suggested. Also an explanation of the negative cases, that is, where no particular regeneration takes place, is equally wanting. In the explanation of any phenomenon it is essential that negative cases be taken into account before any general conclusions are drawn. As has been pointed out above, even among the hermit crabs where the heteromorphic appendages appeared most frequently, in by far the majority of cases no regeneration took place. ‘There was in these experiments a single series of hermit crabs in which nine out of fourteen individuals regenerated a heteromorphic appendage. In the light of this, we should perhaps be safe in concluding that for hermit crabs failure to regenerate may often be due to external conditions. But this would still explain nothing for Crangon and Palamonetes. All of the Crangon experimented upon belonged to the same series and were kept as nearly as possible under precisely the same conditions. Yet but one out of the original twenty-two developed an antenna-like appendage notwithstanding there were fourteen others that lived as long or longer and moulted as frequently. In so far as it was possible to determine the question, the physiological activity of the fourteen that showed no regeneration was equal to that of the one individual that did regenerate the appendage. Extensive series of Palamonetes were operated upon at the same time with the hermit crabs, and were kept under similar conditions. Yet, as has been seen not one regenerated the antenna-like append- age. From this it appears evident that, whatever variations in results may be accounted for by differences in external conditions, the primary answer to the question must be sought elsewhere. It may be objected that in operating upon the eye the entire optic ganglion was not always removed. ‘This, however, could not be offered as an objection 1 in every case. In all three of the forms there were many instances in which not a vestige of the optic ganglion remained, and yet no regeneration reamed: Besides 232 Mary Isabelle Steele there is evidence that in some cases hermit crabs regenerate an antenna-like appendage when part of the ganglion has been left in the eye stalk. We have then the following conditions for hermit crabs at least. First, when the cut comes at a level which leaves as much as two sections of the optic ganglion intact an eye may regenerate (a—b, text Fig. 2). Second, when the cut is made at the base or slightly above the base of the eye stalk (c—d, text Fig. 2) so that little or none of the optic ganglion remains the regeneration of an antenna-like appendage is possible. Lastly, if the eye is removed at a level intermediate between a—b and cd (text Fig. 2) no regeneration follows. Text Fig.2 The line a-b represents approximately the level from which a hermit crab may regen- erate an eye. From the level of the line c-d or below it a heteromorphic appendage may regenerate. No regeneration takes place from intermediate level, e.g., from the level e-f. It is possible, perhaps even probable, that the character of the hypodermis differs more or less at these different levels. It is even conceivable that the hypodermis should be capable of one sort of regeneration at the level a—b or above it, and of another sort at the level c-d or below it; but there is certainly no apparent reason why no regeneration whatever should take place if the eye is removed at a plane intermediate between these two levels. So far as careful microscopic examination can determine there is no difference in the hypodermal cells underlying the cuticle proximal Regeneration in Compound Eyes of Crustacea 233 to the basement membrane. Whatever differences in character may exist between the hypodermal cells over different regions of the eye, the results of this whole series of experiments suggest the inference that presence or absence of a maximum amount of the optic ganglion is a controlling factor in determining the character of the regeneration. The fact that no regeneration takes place from levels intermediate between a—b and c-d is in itself evidence that internal conditions are different at these intermediate levels than from a higher or a lower level. So far as the optic ganglion may be a controlling factor the difference in conditions may be due either to a difference in the character of the ganglion cells or to the reduced ganglionic mass. From the structure of the optic gan- glion (Parker ’90 and Kenyon ’g7) it is probable that not until the lower level c—d has been reached have the peripheral terminations of the optic nerve fibers been seriously interfered with. Both Parker and Kenyon mention the fact that a part of the optic nerve fibers have their cellular origins located in the brain. The fact that this heteromorphic appendage never regenerates except from this lower level suggests that there may be a causal connection between the regeneration of the heteromorphic appendage and the destruction of the distal terminations of the optic nerve fibers. With their peripheral terminations destroyed there might probably be a tendency on the part of the optic nerve fibers to grow outward and form new terminations. Since their natural terminations, the cells of the optic ganglion have been destroyed it seems probable that the fibers of the optic nerve stump would behave like those in a nerve stump of an ordinary appendage, e.g., a leg or antenna. This in itself might have a tendency to induce any new tissue that regenerated to differentiate into the form of some sort of append- age. That this heteromorphic appendage should be antenna-like in form seems probable for two reasons. First, it is the natural tendency of all Arthropod structures to divide into segments. Second, the simplest form of joint found in any appendage is in the antenna. Further, this appendage, although antenna-like, shows a much greater variety in form than any ordinary regener- ated appendage and the joints formed are often irregular and 234 Mary Isabelle Steele incomplete. This fact suggests that the regeneration was not influenced by a fixed set of internal conditions. In the usual cases of regeneration and embryonic development, whatever the determining factor or factors may be, it is recognized that we may expect certain structures to appear in connection with a given set of external and internal conditions. In the development of this heteromorphic appendage, however, conditions seem more variable. As a consequence it shows con- siderable variety of form. In some cases the appendage is but little more than a slender horn-like projection, in other cases the appendage may be curved inward toward the median line, project forward at the angle of the eye or curve backward until the free end touches the margin of the head. (Compare Figs. 23, 38 and 41.) Again from the very first moult the appendage may appear as a single flagellum-like structure or as a pair. None of the hermit babe however, have regenerated a heteromorphic appendage composed of two flagellum- like parts. But in my_ previous observations upon cray rash (Joc. cit.) two or three instances were noted in which the appendage appeared double at the time of the first moult. Herbst has also noted what he regards as an endo- podite and exopodite in several instances. ‘The appearance of the single structure in some cases and the double one in some others can perhaps be explained by the supposition that the nerve fibers become separated into two masses in some instances and remain as a single trunk in others. Miss Reed (/oc. cit.) found that when the stump of the leg of a crayfish or hermit crab was split longitu- dinally in some instances two legs were regenerated from a single stump and in other cases only one. Sections of such legs showed that the end of the nerve stump had been split in the cases in which two legs regenerated and that the nerve stump had not been split when only one leg was regenerated. A similar result might follow in the development of the heteromorphic appendage if the nerve trunk became separated into two bundles by the interpo- sition of another sort of tissue. An explanation of the antenna-like form of the heteromorphic appendage having been suggested, attention should now be directed toward an explanation of its inner structure, which is also Regeneration in Compound Eyes of Crustacea 235 found to be antenna-like. ‘That its inner structure should be antenna-like might be expected since its innervation is associated with a region of the central nervous system that is particularly concerned with the innervation of the special sense organs, and since its outward form is antenna-like it is rather to be expected that the inner structure would also conform more or less to the antenna type. It seems evident that the ganglionic groups of sense cells which are found in the heteromorphic appendage, belong to the general peripheral nervous system found so widely distributed among the different Arthropods. The groups of cells and the associated sensory hairs are equivalent to the “ Hautsinnesorgane”’ of vom Rath (’94). Ost (loc. cit.), however, does not regard these sense cells as true ganglion cells, as Herbst does. In the regenerating antenna of Oniscus, Ost finds the nerve fibers regenerating from the central stump and the groups of sense cells differentiating from the hypodermis. ‘The regenerating nerve fibers come from the end of the nerve stump, extend to the periphery and intermingle with the sense cells. Bethe (’96) considers that the peripheral nervous system of Arthropods differs both in function and origin from the central nervous system. Holmgren (’95) regards it as a sort of sympathetic system. That cutting the peripheral terminations of the optic nerve may induce the regeneration of a heteromorphic appendage seems to receive some support from the results obtained by Zeleny upon the blind crayfish. Although reduced in size the optic ganglion is still present in the rudimentary eyes of blind crayfish. On the other hand the ommatidial structures are entirely wanting. So long as the vestigial eye remains undisturbed there seems to be no tendency toward the development of an antenna-like organ. But when the optic ganglion is removed a heteromorphic appendage appears. Such appendages are apparently functional as sense organs and Zeleny concludes that in the blind crayfish a non- functional organ has been replaced by a functional one. The suggested explanation for the outgrowth of the hetero- morphic appendage also carries with it an implied explanation of the non-appearance of a heteromorphic structure in place of a 236 Mary Isabelle Steele somatic appendage. ‘The nerve trunk of an appendage is asso- ciated with ganglion cells only at its central end, not with ganglion cells at its peripheral end, as distinguished from the optic nerve in its relation to the optic ganglion, consequently in removing an appendage no parts have been removed that would not be likely to again regenerate in a similar manner. While in animals as highly specialized as the hermit crabs we do not find the ganglion parts of the nervous system regenerating. For the negative cases that appear after the entire optic ganglion has been removed, it is evident that no real explanation can be offered until a more adequate understanding of the process of growth and development has been reached. Although we may fully recognize the fact that great differences exist in the physiolog- ical activity of the various individuals and that the external condi- tions are subject to numerous variations, these facts alone will not account for the great number of negative cases which result. In addition to these it seems necessary to recognize an individual variation in the quality of the tissues. Nothing short of some specific inherent individual difference seems sufficient account for the fact that only an occasional hermit crab regenerates a hetero- morphic appendage. The ability to regenerate a heteromorphic appendage in place of an eye which appears as an individual variation in hermit crabs and Crangon and other genera seems to be entirely wanting in at least one species of Palamonetes. Or if not entirely wanting it appears so rarely that even after a great number of experiments and observations it is apparently dbeede In summing up the foregoing discussion it is apparent that a weight of responsibility has been placed upon the nervous system. Numerous observations, however, have left no doubt that the nervous system does exercise an important physiological influence upon the other tissues of the body, both in ordinary growth phenomena and in regeneration. Child (04) observed in oper- ating upon Leptoplana that if more than half of the cerebral ganglion was removed a new head did not regenerate. “This was true regardless of the plane in which the cut was made, a fact which seems to indicate that the mass of nervous material is an important factor in the case of Leptoplana at least. Wilson (03) Regeneration in Compound Eyes of Crustacea 227, discovered that after the larger chela of Alpheus had been removed, cutting the nerve in the smaller one prevented it from growing into the form of the larger one when, however, the large chela had been removed and the nerve in the small one left intact, the small chela developed into the form of the large one. It has been noted above that Miss Reed found she could obtain the regeneration of the double chelea. The experiments of Schaper (’98), Harrison (03), Barfurth (or), Goldstein (’04) and others have shown that the early stages of embryonic development and of regeneration are apparently independent of the nervous system. But the same experiments have also shown that the later stages of growth and differentiation are very largely influenced by the part efihe nervous system which normally innervates the regenerating or developing parts; other instances might be mentioned but a sufhcient number have been given to convince one that in very many instances there is an important connection between the part of the nervous system immediately concerned and the regeneration of the other tissues and structures. VII REGENERATION AFTER SPLITTING THE Eye Loncirvu- DINALLY Several series of Palamonetes were operated upon by having the eye split longitudinally (Table 1). Although in the regener- ation of any part of the eye the new tissue is derived from the hypodermis the results obtained from the experiment of splitting the eye seem to indicate that injury to the optic ganglion is of great importance. In many cases at least splitting the eye could not have resulted in serious injury to the hypodermis yet in-no case did regeneration follow if the optic ganglion had been injured. Whether or not regeneration follened the operation apparently depended upon the depth of the split. If the split extended through the ommatidial portion only and the optic ganglion remained uninjured, the ommatidial portion degenerated and new ommatidia were in some cases regenerated. On the other hand, if the split extended into the optic ganglion the whole ommatidial portion and the whole or part of the optic ganglion degenerated. 238 Mary Isabelle Steele In some cases not even a vestige of the eye remained; in others, stumps of considerable length persisted. But in no cases where the split extended into the optic ganglion was there any sign of regeneration. Figs. 16 and 22 represent instances in one of which an eye regenerated and in the other there were no signs of a regenerating eye. The specimen from which Fig. 16 was taken lived sixty-five days after the eye was split. The regenerated eye is about six- sevenths of the length of the normal eye. Sections show that new ommatidia have regenerated. ‘The eye is not altogether normal in structure, however. The eye stump shown in Fig. 22 was taken from an individual that lived seventeen days after the operation. Apparently the entire optic ganglion has degenerated. There are no definite indications of regeneration. ‘The stump of the optic nerve tapers to a point, perhaps indicating that degener- ation is still incomplete. The stump is little more than one-third the length of normal eye. No additional facts of importance were gained from the experi- ment of splitting the eye. “These results obtained serve chiefly as additional proof that an injury to any part of the eye 1s followed by widespread degeneration of the tissues and that in the case of Palemonetes, after an injury to the optic ganglion usually no regeneration takes place. SUMMARY In summing up the results of the experiments discussed in this paper the following points are to be noted: 1 The death of the animal which so frequently follows imme- diately upon the operation is perhaps due rather to its effect upon the nervous system than to loss of blood. 2 The healing of the wound takes place by the formation of a provisional crust over the cut surface and later by the development of a new cuticle beneath this crust. a The crust is formed of hypodermal cells and a chitinous secretion. Intermingled with this are blood cells and the cells of the injured tissues. From two to three days are required for the formation of the crust. Regeneration in Compound Eyes of Crustacea 239 b The new cuticle is secreted before a continuous hypodermis has formed over the wound. It is continuous with the inner layers of the cuticle over the eye stump. 3 New hypodermal cells over the ommatidial region may arise in two ways; either by the transformation 7m situ of corneal hypo- dermal cells into less specialized, actively multiplying hypodermal cells, or by the proliferation of new hypodermal cells inward from the edges of the cut. 4 Any injury to the eye is always accompanied by extensive degeneration of the remaining tissues. Sometimes the entire eye suffers destruction. 5 The rate of regeneration is considerably affected by the rate of disintegration and the removal of injured parts. 6 Active regeneration may be in progress at the periphery while deeper below the surface the injured structures are not yet removed. 7 In the regeneration of an eye all of the new structures arise from the hypodermis. 8 Multiplication of cells takes places by amitotic divisions. 9 The cells for the retinule are the first to differentiate from the hypodermis. ‘Their differentiation may begin before a con- tinuous hypodermis has developed. 10 The retinular nuclei move inward from the periphery, elongate and divide along their radial axes, and extend proximal processes through the basement membranes to the optic ganglion. Thereby nervous connections are established in the regenerating region. 1r Not until after the retinular processes have extended into the optic ganglion is the differentiation of cones established. The cones differentiate from the periphery inward. 12 The rhabdom is developed from the inner ends of the retinu- lar cells and is at first present as a slender homogeneous rod of uniform diameter, which extends from the inner end of the cones to the basement membrane. The spindle shaped enlargement of the rhabdom does not appear until after all the other parts of the ommatidium have been differentiated. 13 The hypodermis does not become a true corneal hypodermis 240 M ary Isabelle Steele and secrete corneal facets until after all of the other ommatidial structures have been differentiated. Corneal facets are never apparent until after more than one moult has taken place. 14 Ommatidia do not differentiate at a uniform rate in all parts of the regenerating eye. 15 In Palamonetes regeneration of perfect ommatidia does not take place if the optic ganglion has been injured. -Hermit crabs may regenerate a perfect eye after removal of as much as half the optic ganglion. Crangon regenerates an eye very slowly, even when the optic ganglion is uninjured, but there are evidences that ommatidia may differentiate after a part of the optic ganglion has been removed. 16 The rate of regeneration is quite variable in all the species experimented upon, but both hermit crabs and Palamonetes however may regenerate ommatidia within thirty-five to forty-five days. 17 Splitting the eye of Palazmonetes is not followed by regen- eration if the split extends into the optic ganglion. 18 In the breaking down of the injured ommatidia the pigment secreting cells become widely scattered, and the old pigment persists for a long time. Frequent cases of abnormal development of pigment also occur. “There are evidences which indicate that this abnormality is due to the pathological development of the broken down retinulz. 1g After removal of all or nearly all of the optic ganglion, hermit crabs may regenerate a heteromorphic appendage in place of the excised eye. There is, however, apparently a level from which neither an eye nor an antenna-like appendage will regen- erate. 20 The nerve-trunk of the heteromorphic appendage forms a continuous structure with the stump of the optic nerve. 21 Removal of the entire eye of Crangon may also be followed by the regeneration of an antenna-like appendage. 22 In no case was there evidence that Palzmonetes vulgaris possessed the ability to regenerate a heteromorphic appendage after the removal of the entire eye. 23 The results of this entire series of experiments points to the Regeneration in Compound Eyes of Crustacea 241 following conclusion. ‘The regeneration which takes place from any level is largely influenced by the presence or absence of the whole or a part of the optic ganglion. BIBLIOGRAPHY BarFurtH, D., ’o1—Ist die Regeneration vom Nervensystem abhangig? Verh. d. Anat. Ges., 1gor (p. 197-201). BreTtHE, ALBRECHT, ’96—FEin Beitrag zur Kenntnis des peripheren Nervensystems von Astacus fluviatalis. Anat. Anz., 1896, Bd. xi (p. 31-34). Cup, C. M., ’04—Amitosis in Moniezia. Anat. Anz., Bd. xxv (p. 545-558). ’o7—Amitosis as a Factor in Normal and Regulatory Growth. Anat. Anz., Bd. xxx, nos. 11 and 12. Go.psteEI, K., ’o4—Kritische und experimentelle Beitrage zur Frage nach dem Einfluss des Centralnervensystem auf die embryonale Entwickelung und die Regeneration. Arch. f. Entw.-Mech., Bd. xvi, H. 1, 1904. ‘og—Die Abhangigkeit der Muskulatur vom Centralnervensystem wah- rend derEmbryonalzeit. Eine Erwiderung an Herrn Prof. Neumann. Arch. f. Entw.-Mech., Bd. xviii, H. 4. : Grenacuer, H. ’74—Zur Morphologie und Physiologie des facettirten Arthro- " podenauges. Gdtten Nachrichten (p. 645-656). Harrison, Ross G., ’03—On the Differentiation of Muscular Tissue when Re- moved from the Influence of the Nervous System. Am. Jour. of Anat., vol. 2, no. 2. Hersst, C., ’96—Ueber die Regeneration von antennenahnlichen Organen an Stelle von Augen. I. Arch. f. Entw.-Mech., Bd. ii, H. 4. *g6—Ueber die Regeneration von antennenahnlichen Organen an Stelle von Augen. II. Versuche mit Sicyonia Sculpta. Vierteljahrsschr. Naturf. Gesellsch. Ziirich. Jahrg., 41 (p. 435). *oo—Ueber die Regeneration von antennenahnlichen Organen an Stelle von Augen. III. Weitere Versuche mit total Exstirpirten Augen. IV. Versuche mit theilweise abgeschnittenen Augen. Arch, f. Entw.-Mech., Bd. ix, H. 2 (p. 215-293). Herrick, F. H., ’89—The Development of the Compound Eye of Alpheus. Zool. Anz., xii, no. 303 (p. 164-169). Hormcren, E., ’95—Zur Kenntnis des Hautnervensystems der Arthropoden. Anat. Anz., Bd. xii, no. 19 (p. 449-457). Kenyon, F. C.,’g6—The Brain of the Bee. Jr. of Comp. Neur., vol. vi. ‘97—The Optic Lobe of the Bee’s Brain in the Light of Recent Neuro- logical Methods. Am. Nat., xxxi (p. 369-376). 242 Mary Isabelle Steele Kincstey, J. S., ’87—The Development of the Compound Eye of Crangon. Jour. Morph., i, no. 1 (p. 49-66). McGrecor, J. H., ’99—The Spermatogenesis of Amphiuma. Jour. of Morph. Sup. to vol. xv. Meves, g1—Amitotic Division of Spermatagonia of Salamandra. Anat. Anz., Bd. Vi. Ost, J., °06—Zur Kenntnis der Regeneration der Extremitaten bei den Arthropo- den. Arch. f. Entw.-Mech., Bd. xxi, H. 3 (p. 289). Parker, G. H., ’90—Histology and Development of the Eye in the Lobster. Bull. Mus. Comp. Zo6l., Harvard, vol. xx, no. 1 (p. 1-60). *g1—The Compound Eyes in Crustaceans. Bull. Mus. Comp. Zodl., Harvard, xxi, no. 2 (p. 45-140). *95—The Retina and Optic Ganglion in Decapods, Especially in Astacus. Mitth. Zool. Stat. Neapel., vol. xii, H. 1 (p. 1-73). *97—Photomechanical Changes in the Retinal Pigment Cells of Palamon- etes and their Relation to the Central Nervous System. Bull. Mus. Comp. Zoél., Harvard, xxx (p. 275-300). Patren, WiiiiaMm, ’86—Eyes of Molluscs and Arthropods, Mitth. Zool. Stat. Neapel, vol. vi (p. 542-756). *$87—Studies on the Eyes of Arthropods. 1. Development of the Eyes of Vespa. Jr. Morph., vol. i (p. 193-226). Purtuips, E. F., ’05—Structure and Development of the Compound Eye of the Honey Bee. Proc. Acad. Nat. Sci., Philadelphia (p. 123-157). Ratu, O. Vom, ’87—Zur Kenntnis der Hautsinnesorgane und des sensiblem Ner- vensystems bei den Arthropoden. Zeit. f. Wiss. Zool., Bd. 41, H. 2 (p- 499-589). Reep, Marcaret A., '04—The Regeneration of the First Leg of the Crayfish. Arch. f. Entw.-Mech., vol. xviii, H. 3 (p. 307-316). Rosenstapt, B., ‘96—Beitrage zur Kenntnis des Baues der zusammengesetzten Augen bei den Dekapoden. Arch. Mikr. Anat., Bd. xlvii (p- 748-770). ScHaPER, A., g8—Experimentelle Studien an Amphibienlarven, Erste Mitteilung. Haben kinstlich angelegete Defekte des Centralnervensystems oder die vollstandige Elimination desselben einen nachweisbaren Einfluss auf die Entwickelung des Gesammtorganismus junger Froschlarven / Arch, f. Entw.-Mech., Bd. vi (p. 157-197). *98—Expermental Studies on the Influence of the Central Nervous Sys- tem upon the Development of the Embryo. Jr. of the Boston Soc. Med. Sci., January, 18098. STEELE, Mary I., ’04—Regeneration of Crayfish Appendages. Univ. of Mo. Studies, vol. 11, no. 4. Regeneration in Compound Eyes of Crustacea 243 Warase, S., ’89—On the Structure and Development of the Eyes of the Limulus. Johns Hopkins Univ. Circ., viii, no. 79 (p. 34-37). *g0—On the Morphology of the Compound Eyes of Arthropods. Studies Biol. Lab., Johns Hopkins Univ., iv. no. 6 (p. 287-334). Wison, E. B., ’°03—Notes on the Reversal of Asymmetry in the Regeneration of the Chele in Alpheus Heterochelis. Biol. Bull., vol. 4 (p. 187- 214). ZELENY, C., ’05—Compensatory Regulation. Jr. Exp. Zodl., vol. it, no. 1 (pr 102). ’o6—Regeneration of an Antenna-like Organ in Place of the Vestigial Eye of the Blind Crayfish, C. Pellucidus Testi. Sci., vol. Xxiil (p. 527)- EXPLANATION OF PLATES Outlines of all figures were drawn with the aid of a camera. In all of the detailed figures the nuclei were also drawn in with the camera. The magnification is given with the explanation of each figure. As far as possible the figures illustrating the different phases of the subject are numbered according to the number of days the experiment covered. Reference letters-used a. anterior. n.ct. new cuticle. a.cp anterior border of carapace. n.e. normal eye. a.p.c. abnormal pigment cells. n. nucleus. bm. basement membrane. nr. nerve trunk. br. brain. a.ct. old cuticle. br.sh. brain sheath. cm. old muscle. c.c. crystalline cones. op.n. optic nerve. c-hy. corneal hypodermis. 9.sp. optic squame. c.p. coagulated plasma. p. posterior. Ch crust. pt. pigment. ct. cuticle. pt.cs. pigment cysts. e.s. eye stump. ret, retinula. cf. corneal facet. ret.n. retinular nuclei. gl. ganglion. rh. rhabdom. het. heteromorphic appendage. rt. regenerated tissue. hy. hypodermis. seg. segments. hy.tr. transformed hypodermis. s.ct. sub cuticle. m. muscle. sm. sensory hairs. Pirate I Fig. 1 Palemonetes, seven days. One moult seven days after operation. a, Dorsal view, and b, ventral view. Most of ommatidia removed from ventral side. Pigmented portion appears disorganized. Injured eye measures about four-fifths length of normal eye. X 45. Fig. 2 Young Palemonetes, seven days. One moult seven days after operation. Dorsal view. Eye operated upon by thrusting needle into top of ommatidial portion. Nearly half of the ommatidia destroyed. Injured eye measures about three-fourths length of normal eye. X 45. Fig. 3. Palemonetes, ten days. One moult seven days after operation. Ventral view. Part of ventral ommatidial portion removed. Pigment irregularly scattered throughout ommatidial region. Very few uninjured ommatidia remains. X 45. Fig. 4 Palemonetes eye, nineteen days. One moult eighteen days after the operation. Nearly whole ommatidial portion was removed. Pigment patches remains of old ommatidia. New tissues can be seen arranged in strands on interior edge. X 45. Fig. 5 Hermit crab, twenty-five days. Regenerated eye, one moult twenty-four days after opera- tion. At least one section of optic ganglion removed. Regenerated eye five-eighths length of normal eye. X 45. Fig. 6 Palemonetes, thirty days. First moult seven days after operation; second moult twenty- one days later. Ventral view. Whole ommatidial region destroyed. Upper part of regenerated tissue perfectly transparent. Irregular patches of old pigment remains seen in lower part of ommatidial region. Regenerating eye three-fourths length of normaleye. X 45. Fig.7 Crangon, thirty-two days. First moult eighteen days after operation; second moult fourteen days later. Dorsal view. Operation removed upper ommatidial surface. Remains of old pigment apparent. Interior of eye shrunk away from cuticle. Injured eye four-fifths length of normal eye. X 60. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE I Mary Isapette STEELE Tue Journat or ExperiMentar ZobLoGy, VoL. v, No. 2 Pirate II Fig. 8 Crangon, thirty-one days. One moult sixteen days after_injury. Dorsal view. Operation removed small part of inner anterior ommatidial surface. Injured eye four-fifths length of normal. xX 45. Fig. 9 Palemonetes, thirty-three days. First moult fourteen days after operation; second moult nine days later. Dorsal view. Injury chiefly on posterior ventral edge. Injured eye four-fifths length of normal eye. X 35. Fig. 10 Palemonetes, thirty-three days. First moult fourteen days after operation; second moult ten days later. Whole of ommatidial region destroyed. Regenerating eye four-fifths length of normal eye. X 35- Fig. 11 Palemonetes, thirty-five days. First moult sixteen days after operation. Ventral view. Part of ommatidia removed from ventral side. Upper end of eye more pointed than usual. Pigment appears unevenly distributed. Regenerated eye seven-eighths length of normal eye. 45. Fig. 12 Hermit crab, thirty-three days. First moult thirty-two days after operation. Dorsal view. Operation removed all of ommatidia and part of optic ganglion. Small, complete, new eye regenerated. New ommatidia shorter than normal. New eye two-thirds length of normal eye. X 45. Fig. 13 Hermit crab, thirty-eight days. First moult twelve days after the operation. Dorsal view. Ommatidial region and nearly half of the ganglion removed. Very small but perfect eye regenerated. Regenerated eye four-sevenths length of normal eye. X 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE II Mary Isapette STEELE Tue Journar or ExperiMENTAL ZoLoGy, VOL. v, No. 2 Prate Ll Fig. 14 Hermit crab, forty-one days. First moult forty-one days after operation. Dorsal view. Operation destroyed whole ommatidial portion and upper part of ganglion. Complete eye regenerated. Eye is about two-thirds length of normal eye. X 45. Fig. 15 Hermit crab, sixty-seven days. One moult forty-five days after operation. Whole omma- tidial portion and upper part of ganglion destroyed. Regenerated eye fully differentiated. Regener- ated eye two-thirds length of normaleye. X 45. Fig. 16 Palemonetes, sixty-five days. Moulted six days after operation. Eye split. New omma- tidia regenerated. Regenerated eye six-sevenths length of normal eye. 45. Fig. 17 Palemonetes Viridis, nine days. One moult nine days after operation. Shows irregular development of upper end of eye. Pigment scattered irregularly. X 45. Fig. 18 Ventral view of top of eye shown in Fig. 40. Regenerated material appears loose and reticular. X go. Fig. 19 Palemonetes, thirty days. Dorsal view. First moult eight days after operation; second moult fifteen days later. Ommatidial portion wholly destroyed. Upper part of eye transparent. Regen- erated tissue forms loose reticulum. No external signs of differentiation of ommatidia. Injured eye seven-eighths length of normal eye. X 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE II Mary Isapette STEELE Tue Journar or ExperimeNntat ZobLoGy, vou. v, No. 2 Pirate IV Fig. 20 Palamonetes, thirty-eight days. First moult ten days after operation; second moult four- teen days later. Ventral view. Entire ommatidial region destroyed. Dark band represents remains of old pigment. Regenerating eye five-eighths length of normal eye. % 45. Fig. 21 Hermit crab, sixty-seven days. Moulted twelve days after operation. Stump shows large part of ganglion remaining but no signs of regenerating eye. Loose shreds of new tissue developed distal tothe stump. 45. Fig. 22 Palamonetes, seventeen days. Moulted seventeen days after operation. Eye split. Almost entire eye degenerated. Stump about one-third length of normal eye. go. Fig. 23 Hermit crab, twenty-one days. Regenerated heteromorphic appendage and base of nor- mal eye. One moult twenty-one days after operation. Appendage segmented; curves inward toward the median line. X go. Fig. 24 Eye stump, thirty-two days. First moult ten days after operation; second moult sixteen days later. Pigment patches near distal end of stump. Short stiff hairs on end of stump. Stump is two-fifths length of normaleye. 45. Fig.25 Hermitcrab. Injured when found. One moult. Ventral view. Distal end of eye shows regenerating ommatidia at ». Abnormal pigment developed at pt. Abnormal protuberance on the inner edge. Injured eye three-fifths length of normal eye. 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE IV Mary IsapeL_e STEELE Tue JourNat or Exrerimentat ZoOLoGy, vor. v, No. 2 Pirate V Fig. 26 Palemonetes, twenty-four days. First moult ten days after operation; second moult four- teen days later. Right eye removed at level of line a-b. Small bud of newtissue. t, Regenerated from stump. Cuticle removed from stump. % 45. Fig.27 Stump shown in Fig. 26 more highly magnified. New tissue darkly shaded. X 125. Fig. 28 Hermit crab, twenty-four days. One moult twenty-two days after removal of eye. Hetero- morphic appendage. Ventral view. Cuticle heavy, appendage irregularly segmented. Sensory hairs developing near tip. Nerve trunk visible beyond proximal half of segment. Nerve trunk small in diameter as compared with Fig. 30. Appendage three-fourths length of normal eye. go. Fig. 29 Palemonetes, twenty-five days. One moult. Outline of normal eye and eye stump show- ing abnormal pigment. Pigment in a number of small masses on upper distal end of stump. X 75. Fig. 30 Hermit crab, twenty-two days. Moulted twenty-two days after operation. Heteromor- phic appendage. Dorsal view. Show segments, sensory hairs and nerve trunk extending beyond prox- imal half of appendage. Appendage one-third length of normal eye. 125. Fig. 31 Hermit crab, twenty-six days. One moult twenty-six days after operation. Dorsal view of normal eye and healed over stump. Stump three-eighths length of normaleye. X 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE V Mary IsapeLLe STEELE Tue JourNaL or ExperiMENTAL ZOOLOGY, VOL. V, NO. 2 Prate VI Fig. 32 Palamonetes, twenty-seven days. First moult sixteen days after operation; second moult ten days later. Outline of normal eye and eye stump showing abnormal pigment. Pigment appears as granular area on inner dorsal surface. Eye stump four-fifths length of normaleye. 45. Fig. 33 Palemonetes, twenty-seven days. First moult ten days after operation; second moult four- teen days later. Ventral view of stump and normal eye. End of optic nerve stump flush against the cuticle. Optic nerve reduced in size, two-sevenths lengths of normaleye. X 45. Fig. 34 Hermit crab, thirty-nine days. Moulted twenty-nine days after operation. Dorsal view of normal eye and heteromorphic appendage. Shows optic squame in connection with appendage. Optic nerve stump extends through proximal half of appendage. 35. Fig. 35 Series of Palemonetes eye stumps after removal of greater part of eye. a, Eye stump with small quantity of new tissue developed beyond end of optic nerve stump. Stump measures one-third length of normaleye. X45. b, Eye stump that shows no regeneration. Twenty-ninedays. Moulted ten days after operation. One-fourth length of normal eye. 45. c, Eye stump showing an un- usual development of new tissue. Moulted ten days after operation. Stump two-sevenths length of normal eye. X 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE VI Mary IsaneLte STEELE Tue Journat or ExrertMeNTAL ZoOLoGY, VOL. v, NO. 2 Prate VII Fig. 36 a and b, Palemonetes, thirty days. One moult. Optical section of normal and regener- ating eye. Ventral view. a, Shows regenerating eye; b, normal eye for comparison. Operation apparently removed eye near level of line a-b on normal eye. Regenerating eye shows considerable new tissue and pigment spot on ventral side. Heavy cuticle over end of stump. X 45. Fig. 37 Crangon eye stump, thirty-two days. First moult two days after operation; second moult nine days later; third moult sixteen days later. Stump one-third length of normal eye. No regenera- tion. X 45. Fig. 38 Crangon, thirty-two days. First moult four days after operation; second moult seventeen days later. a, Optical section of heteromorphic appendage and outline of normal eye. Dorsal view. 45. b, Ventral view of heteromorphic appendage more highly magnified. Shows six segments and sensory hairs developed on the inner distal edge. Nerve trunk apparent through greater part of length. Appendage measures four-fifths length of normal eye. go. Fig. 39 Hermit crab, thirty-two days. One moult thirty-two days after operation. Ventral view of stump showing no regeneration. Stump two-thirds length of normal eye. X 45. Fig. 40 Crangon eye stump, thirty-one days. First moult seven days after operation; second moult twelve days later. Stump measures one-third length of normaleye. X 45. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE VII Mary IsaneLtte STEELE Tue JourNnar or ExperiMENTAL ZOOLOGY, VOL. V, NO. 2 Pirate VII Fig. 41 Hermit crab, thirty-two days. One moult thirty-two days after operation. Dorsal view of normal eye and heteromorphic appendage. Appendage small and sharply curved backward. Shows several segments and a few sensory hairs. go. Fig. 42 Crangon eye stump, thirty-two days. First moult eight days after operation; second moult twelve days later. Cuticle folded and wrinkled. Short hairs on end of stump. Stump about one-half length of normal eye. No regeneration. X 45. Fig. 43 Palemonetes, thirty-eight days. Two moults. Eye stump and outline of normal eye. Ventral view. Shows abnormal pigment spot. Eye stump one-half length of normal eye. X 45. Fig. 44 Palamonetes, thirty-eight days. First moult sixteen days after operation; second moult eighteen days later. Eye stump showing abnormal pigment which appears as a single solid mass on upper anterior border of eye stump. Stump about two-thirds length of normaleye. X 45. Fig. 45 Hermit crab, thirty-nine days. One moult. Heteromorphic appendage. Dorsal view. Nerve trunk distinct in proximal part of appendage. Appendage three-fifths length of normal eye. X go. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE VIII Mary Isanette STEELE Tue Journar or ExreriMENTAL ZOOLOGY, VOL. V, NO. 2 Prate IX Fig. 46 Semidiagrammatic section through the top of eye of Cambarus virilis sixty-two and one- half hours after removal of part of ommatidia. Shows relation of cuticle and protective crust (cr.) Shows broken down tissue excluded by crust also. Below crust is space from which inner tissues have shrunk. Space occupied by coagulated plasma. go. Fig. 47 Semidiagrammatic section through upper part of eye of Cambarus gracilis showing con- tiguity between regenerated and old cuticle. Also shows broken down tissues excluded by development of cuticle. Eye operated upon by tearing small hole in cornea with needle. 450. Fig. 48 Section from eye shown in Fig. 9. Broad band of new cuticle developed. Few regenerated nuclei present. All the tissue shown below cuticle degenerating remains of old ommatidia. 1350. Fig. 49 Section from eye shown in Fig. 3. Shows new cuticle with no hypodermal cells beneath it. Shows amitotically dividing nuclei. 1350. Fig. 50 Part of section from eye shown in Fig. 3. Section from near edge injured area. Hypoder- mal nuclei much more numerous than in Fig. 51, which is taken from a section near center of injured area. X 600. Fig. 51 Section from eye shown in Fig. 3. New cuticle but no well defined hypodermis yet formed beneath cuticle. Granular masses and pigment patches are remains of degeneratingommatidia. X 430. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE IX Mary IsapeL__e STEELE as % 1 48 5 ® 5 Tue Journat or ExperIMENTAL ZOOLOGY, VOL. V, NO. 2 Prate X Fig. 52 a,b,c, d and e shows series of figures from an eye of Palemonetes representing transforma- tion of corneal hypodermal cells into active regenerating hypodermis. Shows normal corneal hypoder- mal cells together with corneal facets and tops of cones ina andb. Transformation of cells from resting corneal hypodermal cells to active regenerating cells in c and d. These forma continuous series. X 1350. Fig. 53 Section of eye of Palemonetes from which part of the ommatidial region was removed. Shows new cuticle and reticularsubcuticle. Transformed hypodermal cells in process of amitotic divi- sion. To the left a single cell which may be undergoing mitotic division. Experiment covered twenty- three days. Moulted twice. Section taken from same eye as series in Fig. 52. X 1350. Fig. 54 Sections from same eye as Figs. 52 and 53. This section shows cells separating from the hypodermis and also early stages of differentiation of retinule. Outline of cuticle and subcuticle shown. X 660. Fig. 55 Section from eye shown in Fig. 10. Hypodermal cells differentiated. Amitosis taking place in these and the deeper lying cells. The deeper cells are regenerating retinular cells. 1350. REGENERATION IN COMPOUND EYES OF CRUSTACEA PUATRE OX Mary Isapecte STEELE Tue JourNar or ExreriMENTAL ZOOLOGY, VOL. V, NO. 2 Pirate XI Figs. 56-59 Taken from sections of regenerating eye shown in Fig. 36. Fig. 56 from section near dorsal surface. Regenerated tissue lies for the most part peripheral to the broken line xy. Cuticle torn and inner tissues shrunken from it. Old tissues show parts of muscle bands and small groups of ganglion cells. Distal to muscle band new tissue seems differentiated into fibers. go. Fig. 57 Shows section deeper below surface than Fig. 35. Same features as in preceding figure. In addition a few small pigment masses. An increase in size of the nuclei in region from} toc. Cuticle not shown. X go. Fig. 58 Represents upper part of tangental section near the ventral surface. Nuclei increased in size and number over those in preceding figure. Rudimentary ommatidial elements apparent in new tissue. Figure composed entirely of regenerated tissue except small group of ganglion cells. X go. Fig. 59 Rudimentary ommatidia from the eye shown in Fig. 36. Sections oblique so that entire ommatidium cannot be recognized. Shows distal ends of cones, retinular nuclei and pigmented proc- esses which appear to be retinule. > goo. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE XI Mary Isapette STEELE Tue Journar or ExreriMENTAL ZOOLOGY, VOL. VY, NO. 2 Pirate XII Fig. 60 Group of retinule from eye shown in Fig. 6. Proximal retinular processes are seen extend- ing to basement membrane. Two of the processes can be traced through below the basement mem- brane. X goo. Fig. 61 Taken from section of eye shown in Fig. 16. Shows group of retinule. Nuclei in out- line and proximal processes shown. X 1350. Fig. 62 Group of retinule from eye shown in Fig. 6. Shows proximal retinular processes pene- trating basement membrane and twining among ganglion cells below. 1350. Fig. 63 Taken from section of eye shown in Fig. 11. Shows early stage in differentiation of crystal- line cones. Cone nuclei are being separated from hypodermal nuclei. Hypodermal nuclei are grouped in pairs. Delicate strands of cytoplasm extending inward from pairs of nuclei. X 1350. Fig. 64 Taken from section of eye shown in Fig. 11. Shows more advanced stage of cone differen- tiation than Fig. 29. Cell outlines becoming defined but hypodermal and cone cells not distinctly sepa- rated. Section somewhat oblique so that the four cone nuclei are visible. Distal retinular processes extending between the cones. Lower ends of cones not yet differentiated. X goo. PLATE XII REGENERATION IN COMPOUND EYES OF CRUSTACEA Mary Isapeitte STEELE por, feet OR agit, Tue JourNaL or ExreriMENTAL Zo6LoGy, VOL, V, NO. 2 Pirate XII Fig. 65 Taken from section of eye shown in Fig. 11. Shows retinule with their distal processes extending to hypodermis. Shows early pigment deposition in proximal processes. Hypodermal cells shown in outline. XX 1350. Fig. 66 Taken from section of eye shown in Fig. 11. Ommatidia completely differentiated except spindle shaped enlargement of the rhabdom. Distal ends of cones not yet differentiated completely. Cone at left of figure cut obliquely. Retinule not altogether normal in their distribution. X 1350. Fig. 67 Regenerated ommatidium from eye. Shown in Fig. 11. Rhabdom still not quite normal in appearance. X 1350. Fig. 68 Sections from an eye shown in Fig. 15. Shows differentiated hypodermis and retinular nuclei beginning to assume their definitive position, Hypodermis and retinule both show dividing nuclei. X 1350. PLATE XIII REGENERATION IN COMPOUND EYES OF CRUSTACEA Mary IsapeLLe STEELE NO. 2 vy Tue JourNar or Exrerimentar Zodocy, vor. Pirate X1V Fig. 69 Regenerating cones from eye shown in Fig. 20. Most of regenerated part of the eye occu- pied by abnormal tissue. Abnormal cells mingled with the normally regenerating structures. Com- pare Figs. 68-71. X 600. Fig. 70 Part of section of eye shown in Fig. 20. Most of the cells abnormal polymorphic nucleate cells except those comprising the hypodermis. 450. Fig. 71 Right hand edge of Fig. 70 more highly magnified. Shows dividing hypodermal cells at upper edge and cells with polymorphic nuclei in interior. Three retinular nuclei at right edge of figure. X 1350. Fig. 72 Outline of section through stump shown in Fig. 43. Shows location of pigment spot with reference to other structures in stump. X 125. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE XIV Mary IsaspeL__e STEELE Tue Journar or Experimenta ZoGLoGy, vot. v, NO. 2 Pirate XV Fig. 73. Detailed representation of pigment area shown in preceding figure. Shows collection of pigment cells within cysts. goo. Fig. 74 Part of section from Cambarus virilis sixty-two and one-half hours after operation. In left part of figure lower part of cone and upper end of rhabdom. Remainder of figure occupied by chains of abnormal cells apparently developing from disintegrating retinule. A few cells show pigment gran- ules. X gIo. Fig. 75 Group of cells which show polymorphic nuclei. a, Group of disintegrating retinule from Cambarus virilis seventeen and one-half hours after operation; b, group of disintegrating retinule from Cambarus virilis thirty-nine hours after operation; c, disintegrating retinule from Cambarus gracilis sixteen days after operation; d, group of abnormal pigment cells from Crangon twenty-three days after operation; e, group of abnormal pigment cells from hermit crab sixty-seven days after opera- tion; f, group of depigmented pigment cells from pigment cyst in eye of stump of Palemonetes thirty days after operation; g, outline of crushed pigment body lying in same group with depigmented cells shown inf. X goo. i PLATE XV REGENERATION IN COMPOUND EYES OF CRUSTACEA Mary JIsapectte STEELE Tue JourNAL or ExperIMENTAL ZOOLOGY, VOL. v, NO. 2 Pirate XVI Fig.76 a, Corneal facet and upper end of cone from fully regenerated ommatidium; }, fully regener- ated rhabdom a and b both taken from regenerated hermit crab eye shown in Fig. 15. Ommatidia of hermit crabs much more slender than Palemonetes ommatidia. X goo. Fig. 77 Hermit crab, sixty-seven days. Section through brain and proximal end of heteromorphic appendage. Shows continuity of optic nerve and nerve trunk of appendage. Slightly diagrammatic. xX 1265. Fig. 78 Section through distal end of heteromorphic appendage, showing strands of fibers sm-f. extending to the sensory hairs, and groups of sensory cells, s.c. XX 125. Fig. 79 Detail drawing of small part of section shown in Fig. 78. Shows sensory cells and fibers sn.f. in connection with bases of two sensory hairs. 750. REGENERATION IN COMPOUND EYES OF CRUSTACEA PLATE XVI Mary IsapeL__e STEELE Se : S tS! fo] Lv is Y © 0 000001 C 000-0009 Tue Journar or ExperIMENTAL ZOOLOGY, VOL. V, NO, 2 ON SOME PHENOMENA OF COALESCENCE AND REGENERATION IN SPONGES’ BY H. V. WILSON With Four Ficures I In a recent communication I described some degenerative and regenerative phenomena in sponges and pointed out that a knowl- edge of these powers made it possible for us to grow sponges in a new way. ‘The gist of the matter is that silicious sponges when kept in confinement under proper conditions degenerate in such a manner that while the bulk of the sponge dies, the cells in certain regions become aggregated to form lumps of undifferentiated tissue. Such Taras or plasmodial masses, which may be exceed- ingly abundant, are often of a rounded shape resembling gem- mules, more especially the simpler gemmules of marine sponges (Chalina, e. g.), and were shown to possess in at least one form (Stylotella) full regenerative power. When isolated they grow and differentiate producing perfect sponges. I described more- over a simple method by which plasmodial masses of the same appearance could be directly produced (in Microciona). The sponge was kept in aquarium until the degenerative process had begun. It was then teased with needles so as to liberate cells and cell agglomerates. ‘These were brought together with the result that her fused and formed masses sale | in appearance to those produced in this species when the sponge remains quietly in aquarium. At the time I was forced to leave it an open question whether the masses of teased tissue were able to regenerate the sponge body. During the past summer’s work at the Beaufort Laboratory? 1 Published with the permission of Hon. Geo. M. Bowers, U. S. Commissioner of Fisheries. 2T am indebted to the director of the station, Mr. H. D. Aller, for his kindly aid in supplying all facilities needed in the course of my investigation. Tue Journat or ExreriMENTAL ZOOLOGY, VOL. V, NO. 2. 246 H. V. Wilson I again took up this question and am now in a position to state that the dissociated cells of silicious sponges after removal from the body will combine to form syncytial masses that have power to differentiate into new sponges. In Microciona, the form especially worked on, nothing is easier than to obtain by this method hundreds of young sponges with well developed canal system and flagellated chambers. How hardy sponges produced in this artificial way are and how perfectly they will differentiate the characteristic skeleton, are questions that must be left for more prolonged experimentation. Taking up the matter where it had been left at the end of the preceding summer, | soon found that it was not necessary to allow the sponge to pass into a degenerative state, but that the fresh and normal sponge could be used from which to obtain the teased out cells. Again in order to get the cells in quantity and yet as free as possible from bits of the parent skeleton, [ devised a substi- tute for the teasing method. The method adopted is rough but effective. Let me briefly describe the facts for Microciona. ‘This species (M. prolifera Verr.) in the younger state is incrusting. As it grows older it throws up lobes and this may go so far that the habitus becomes bushy. The skeletal framework consists of strong horny fibers with embedded spicules. Lobes of the sponge are cut into small pieces with scissors and then strained through fine bolting cloth such as is used for tow nets. A square piece of cloth is folded like a bag around the bits of sponge and is immersed in a saucer of filtered sea-water. While the bag is kept closed with the fingers of one hand it is squeezed between the arms of a small pair of forceps. The pressure and the elastic recoil of the skeleton break up the living tissue of the sponge into its constituent cells, and these pass out through the pores of the bolting cloth into the surrounding water. The cells, which pass out in such quantity as to present the appearance of red clouds, quickly settle down over the bottom of the saucer like a fine sediment. Enough tissue is squeezed out to cover the bottom well. The cells display amceboid activities and attach to the substratum. Moreover they begin at once to fuse with one Coalescence and Regeneration in S ponges 247 another. After allowing time for the cells to settle and attach, the water is poured off and fresh sea-water added. ‘The tissue is freed by currents of the pipette from the bottom and is col- lected in the center of the saucer. Fusion between the individual cells has by this time gone on to such an extent that the tissue now exists in the shape of minute balls or cell conglomerates of a more or less rounded shape looking to the eye much like small inverte- brate eggs. Microscopic examination shows that between these little masses free cells also exist, but the masses are constantly incorporating such cells. The tissue in this shape is easily handled. It may be sucked up to fill a pipette and then strewn over cover glasses, slides, bolting cloth, watch glasses, etc. The cell conglomerates which are true syncytial masses throw out pseudopodia all over the surface and neighboring conglomerates fuse together to form larger masses, some rounded, some irregu- lar. The details of later behavior vary, being largely dependent on the amount of tissue which is deposited in a spot, and on the strength of attachment between the mass of tissue and the sub- stratum. Decidedly the best results are obtained when the tissue has been strewn rather sparsely on slides and covers. The syncytial masses at first compact and more or less rounded, flatten out becoming incrusting. They continue to fuse with one another and thus the whole cover glass may come to be occupied by a single incrustation, or there may be in the end several such. If the cover glass is examined at intervals, it will be found that aii ouertion is gradually taking place. “The dense homogeneous syncytial mass first dev elops at the surface a thin membrane with underlying connective tissue (collenchyma). Flagellated cham- bers make their appearance in great abundance. Canals appear as isolated spaces which come to connect with one another. Short oscular tubes with terminal oscula develop as vertical pro- jections from the flat incrustation. If the incrustation be of any size it produces several such tubes. The currents from the oscula are easily observed, and if the cover glass be mounted 1n an inverted position on a slide the movements of the flagella of the collar cells may be watched with a high power (Zeiss 2 mm.). 248 H. V. Wilson This degree of differentiation is attained in the course of six or seven days when the preparations are kept in laboratory aquaria (dishes in which the water is changed answer about as well as running aquaria). Differentiation goes on more rapidly when the preparation is hung in the open harbor in a live-box (a slide preparation inclosed in a coarse wire cage is convenient). Sponges reared in this way have been kept for a couple of weeks. The currents of water passing through them are certainly active and the sponges appear to be healthy. In such a sponge spicules are present, but some of these have unquestionably been carried over from the parent body along with the squeezed out cells. The old question of individuality may receive a word here. Microciona is one of that large class of monaxonid sponges which lack definite shape and in which the number of oscula is correlated simply with the size of the mass. While we may look on such a mass from the phylogenetic standpoint as a corm, we speak of it as an individual. Yet it is an individual of which with the stroke of a knife we can make two. Or conversely it is an individual which may be made to fuse with another, the two forming one. To such a mass the ordinary idea of the individual is not applic- able. It is only a mass large or small having the characteristic organs and tissues of the species but in which the shape of the whole and the number of the organs are indefinite. As with the adult so with the lumps of regenerative tissue. They have no definiteness of shape or size, and their structure is only definite in so far as the histological character of the syncytial mass is fixed for the species. A tiny lump may metamorphose into a sponge, or may first fuse with many such lumps, the aggregate also pro- ducing but a single sponge although a larger one. In a word we are not dealing with embryonic bodies of complicated organization but with a reproductive or regenerative tissue which we may start on its upward path of differentiation in almost any desired quan- tity. A striking illustration of this nature of the material is afforded by the following experiment. ‘The tissue in the shape of tiny lumps was poured out in such wise that it formed con- tinuous sheets about one millimeter thick. Such sheets were then cut into pieces, each about one cubic millimeter. These Coalescence and Regeneration 1n S ponges 249 were hung in bolting cloth bags in an outside live-box. Some of the pieces in spite of such rough handling metamorphosed into functional sponges. Even where the embryonic bodies of sponges have a fixed structure and size, as in the case of the ciliated larva, the potential nature as displayed in later development, is not fixed in the matter of individuality. Such a body (see p. 10) may form a single individual or may fuse with some of its fellows to form a larger individual differing from the one-larva sponge only in size. It is then in spite of its definiteness of shape and size, essentially like a lump of regenerative tissue in that whether it develops into a whole sponge or a part of a sponge depends not on its own structure but on whether it is given a good opportunity of fusing with a similar mass. A parallel case to the coalescence of larvae is afforded by the gemmules of fresh water sponges. Mr. M. E. Henriksen in a manuscript account submitted to me a year ago, describes the fusion of gemmules to form a single sponge. In the preceding description I have passed over the question as to the precise nature of the cells which combine to form the masses of regenerative tissue. On this point as on the histological details in general I hope to have more to say later. Nevertheless the phenomena are so simple that observation of the living tissue reveals much, probably indeed all that is of fundamental impor- tance. If a fairly dense drop of the squeezed out tissue be mounted at once and examined with a high power (Zeiss 2 mm., comp. oc. 6), the preparation is seen to consist of fluid (sea-water) with a few spicules and myriads of separate cells. The cells fall into three classes. 1 ‘The most conspicuous and abundant are spheroidal, red- dish, densely granular, and about (8#/in diameter. These cells which can be nothing but the unspecialized, amceboid cells of the mesenchyme (ameebocytes or archzocytes), put out hyaline pseudo- podia that are sometimes elongated, more often rounded and blunt. 2 here is also a great abundance of partially transformed collar cells, each consisting of an elongated body with slender flagellum. he cell is without a collar, the latter doubtless hay- 250 H. V. Wilson ing been retracted. In the freshly prepared tissue the flagella are vibratile, the cells moving about. Soon however the flagellum ceases to vibrate. 3. The third class is not homogeneous. In it I include more or less spheroidal cells ranging from the size of the granular cells down to much smaller ones. Many of these are completely hyaline, while others consist of hyaline protoplasm containing one or a few granules. Fusion of the granular cells begins immediately and in a few minutes time most of them have united to form small conglomer- ate masses which at the surface display both blunt and elongated pseudopodia. ‘These masses soon begin to incorporate the neigh- boring collar and hyaline cells. One sees collar cells sticking fast by the end of the long flagellum to the conglomerate mass. Other collar cells are attached to the mass by short flagella. Still again only the body of the collar cell projects from the mass while there is no sign of the flagellum. Similarly spheroidal hyaline cells of many sizes are found in various stages of fusion with the granularconglomerate. In sucha preparation the space under the cover glass is soon occupied by innumerable masses or balls of the kind just described, between which continue to lie abundant free cells, some collar cells, others hyaline. Practically all the granu- lar cells go to make up the balls. The play of pseudopodia at the periphery of such balls, which results in the incorporation of free cells and in the fusion of balls to form larger masses, is easily watched. Along with such a cover glass preparation it is con- venient to have some of the squeezed-out tissue in a watch glass of sea-water. In the watch glass preparation it is instructive to watch with a two-thirds or one-half objective the fusion of the cell conglomerates to form masses like those strewn on covers, slides, ete: ((p: 3): These observations on the early steps in the formation of the masses of regenerative tissue make it plain that such masses are composed chiefly of the spheroidal, granular cells (amcebocytes or archeocytes), but that nevertheless other cells, collar cells and more or less hyaline cells also enter into their composition. I may recall the fact that in the formation of regenerative masses in a Coalescence and Regeneration im Sponges 251 degenerating sponge,? the evidence from sections, which is the only evidence available in the case, points to the conclusion that the collar cells help to form the syncytial tissue of the masses. The question of interest lying at the heart of this matter may be so formulated: can particles of the Microciona protoplasm dif- ferentiate into functional collar cells and, when the occasion arises, change back into unspecialized masses capable of com- bining with other masses of unspecialized protoplasm to form a regenerative body? ‘The facts to which I have just alluded sup- port this idea, and indicate that the immediate problem is one worth pursuing farther as a good case of temporary differentiation of protoplasm in the metazoa analogous to the temporary speciali- zation of the cell individual which occurs in such colonial protozoa as Protospongia.‘ As far as the amcebocytes are concerned it is certain that they have great regenerative power. Weltner in a recent paper? has emphasized the importance of these unspecialized cells in the processes of growth and regeneration. _ His conclusions which refer directly to fresh water sponges, are that in a growing sponge, in a sponge regenerating new organs after its winter period of simplification, and in the regeneration of a sponge from a cutting, the amcebocytes are the all-powerful elements in that they give rise to all the new tissues formed, He further alludes to the fact that such reproductive bodies as the gemmules of fresh water sponges and the buds of Tethya (according to Maas) are only groups of amoebocytes; further that the gemmules of Tedania and Esperella described by Wilson as developing into ciliated larve, and the similar bodies found by Ijima in hexactinellids, are such groups. I may add that the presence of such groups of unspecialized cells in the hexactinellids has recently been con- firmed by the master in sponge-morphology, F. E. Schulze, who recognizes the probability of their reproductive nature and gives 3 A new method by which sponges may be artificially reared, Science, n. s., vol. xxv, no. 649, 1907 * Metschnikoff, Embryologische Studien an Medusen, p. 147, 1886. 5 Spongilliden-studien V. Zur Biologie von Ephydatia fluviatilis und die Bedeutung der Ameebocyten fiir die Spongilliden. Archiy fiir Naturgeschichte, 73 Jahrg., 1 Bd., 2 Heft, 1907. 252 H. V. Wilson them a new name, that of sorites.° It is clear then that in many sponges reproductive bodies are formed by the association of unspecialized amoeboid cells. But there is nothing in this fact which precludes the possibility that the groups of amcebocytes are in part recruited from transformed collar cells and other tissue cells, such as pinacocytes (flat cells of canal walls), that have undergone regressive differentiation into an unspecialized amoeboid condition. Cells analogous to the amcebocytes of sponges are found else- where in the metazoa, e. g., in the ascidians.’_ It would be inter- esting to know what capacity, if any, for development they have, when freed from the parent (bud) and collected together in sea- water. If I shall here briefly record some experiments which gave only negative results but which under circumstances admitting of a wider choice of species, ought to yield returns of value. These experiments were based on the assumption that if the dissociated cells of a species will recombine to form a regenerative mass and eventually a new sponge, the dissociated cells of two different species may be made to combine and thus form a composite mass bearing potentially the two sets of species-characteristics. It is clear that such an organism would be analogous to one produced by an association of the blastomeres of the two species. Pending the successful carrying out of this experiment, it would be idle to discuss further the nature of the hypothetical dual organism. In my own experiments three sponges were used: Microciona, Lissodendoryx and Stylotella. The three are all monactinellids, but Microciona is the only one in which the skeleton includes any considerable amount of horny substance. Dissociated cells of Microciona and Lissodendoryx were mixed, and again dissociated cells of Microciona were mixed with those of Stylotella. In each case the experiment was performed at two different times, and a considerable number of admixtures, in watch glasses and on 6 Wissensch. Ergebn. d. Deutsch. Tiefsee-Exp. 1898-99. | Hexactinellida, pp. 213-15. Jena, 1904. 7 Comp. Hjort’s and Lefevre’s papers on budding in ascidians. Coalescence and Regeneration in Sponges 253 cover glasses, was made. The preparations were examined at short intervals with the microscope. The cells of these three species are colored very differently, and are therefore easily dis- tinguished, at least as soon as fusion sets in and little masses of cells begin to be formed. In all the experiments the cells and cell-masses of a species combined, and not the cells of different * species. Thus in the admixture of Microciona and Lissoden- doryx, Microciona regenerative masses and Lissodendoryx regen- erative masses were produced. Similarly when Microciona and Stylotella cells were mixed, the resultant masses were pure, some Microciona, some Stylotella. The Microciona masses in these experiments were hardy. They continued to develop and in some preparations metamorphosed. ‘The cell masses of the other two species while they reached a considerable size were not hardy, most dying soon although some began the process of metamor- phosis. These three species are so unlike that there was little ground in the beginning for the expectation that coalescence would take place. Possibly as in the cases where fusion of egg and sperm of different species is induced through some alteration in the physio- logical state of the protoplasm, so the regenerative cells and cell masses of different species may be made to combine under abnor- mal conditions. The more promising task is however to find allied species and subspecies, the regenerative tissue of which will combine under natural conditions. Such forms, I take it, should be sought among the horny sponges and the monactinellids with abundant horny matter. Il The tendency to fuse so vigorously displayed by the cells and cell masses of regenerative tissue led me to examine into the power that larva have to fuse with one another and the capacity for development in the resultant mass. Delage and others have remarked on the .ot infrequent occurrence of fusion between sponge larve. Delage® says that he has often observed two or 8 Embryogénie des Eponges. Arch. de Zool. Exp. et Gén., p. 400, 1892. 254 H. V. Wilson several larva unite to form a single sponge “which has from the start several cloacas.”’ I find that this power to fuse displayed by the larvz is one that is easy to control. Fusion between larve will readily take place if they are brought in contact at the critical time when the ciliated Figs. 1, 2,3, 4 Composite masses produced by the fusion of larve. The stippled ends and areas are in nature blue, and represent the ends of the component larve. The body of the mass is white. Fig. 1 shows a mass composed of four larva which has just united with a mass composed of five or six larve. In Fig. 2 more than ten, probably about twenty, larve have combined. In Fig. 3 about six larve have combined. In Fig. 4 the original quadruple mass composed of four radiately arranged larva, has been extended in one direction by the addition of a pair of larve, and in the opposite direction by the addition of two pairs of larve. Figs. 1 and 3 X 44; Figs. 2 and 4 X 22. epithelium is being replaced by the permanent flat epithelium. At this time they will fuse in twos or threes or in larger number up to and over one hundred (Figs. 1-4). The smaller composite masses composed of as many as five or six larvae metamorphose into perfect sponges. The larger masses composed of many Coalescence and Regeneration in Sponges 255 larvae did not metamorphose in my experiments but experience with the regenerative tissue suggests that such masses would metamorphose if certain mechanical difficulties due to the great size of the mass were removed. Possibly this might be accom- plished by cutting a flattened sheet composed of some hundred larve (such as I have produced) into pieces and inducing the pieces to metamorphose separately. I may now describe some of the details in this process of larva- fusion. In a species of Lissodendoryx used the larva is of the following character. It has the usual ovoidal shape with a posterior protuberant non-ciliated pole. The anterior pole is somewhat truncated and is sparsely ciliated. The rest of the body bears the usual thick covering of cilia. As seen with reflected light the bulk of the body is dead white, the posterior pole deep blue, and the anterior pole bluish. This coloration is not abso- lutely fixed for the species, but the larve used in my coalescence experiments were all of this character. Within twenty-four hours after liberation the ciliated larve are creeping (remaining in con- tact with the bottom as they swim) over the bottom of the dish. Some are now put in deep round watch glasses and with pipette and needle coaxed together into a clump. Fusion soon begins and on the next day plenty of composite larva are present. The larvae fuse endwise, for the most part in pairs. ‘The compound larva so produced owing to its weight has a very feeble locomotory power. Using pairs chat are nearly motionless, larvae may be brought together (coaxed with needle) and arranged in a desired position on a cover glass for instance. In gna cases fusion results before the separate masses move apart. In this way, selecting an instance, | have added to one arm of a quadruple mass a pair of larva, and to the opposite arm two pairs (Fig. 4). For the purpose of bringing about the fusion of many larve the following simple method is convenient. Suppose that we have the larva in a parafine-coated dish, and they are in a late “creep- ing” stage. Small excavations, 2-3 mm. deep and 4~5 mm. wide, are now made in the parafhne, and with the pipette the larvae are driven into the holes. They lie here in numbers up toand over one hundred, crowded together and heaped upon one another. 250 H. V. Wilson Fusion begins soon and the larve are gradually converted into a flattened cake. The larger cakes thus made measured four by three millimeters. The body of such a cake is a continuous flattened mass in which there is no indication of the component larvae, but the rounded ends of the larvze that have last fused with the general mass remain for a time distinguishable. Owing to their blue coloration the ends of the larvae may be recognized in these and the other compound masses even after the outline of the larva has been completely lost. As already stated the smaller compound masses metamorphose without difficulty. The coalesced larve may be made to attach to cover glasses, slides, etc. Larger masses composed of about twenty larvae underwent a partial metamorphosis. Such masses were laid upon bolting cloth to which they readily attached. The largest masses were hung in small bolting cloth bags in a live box. Whether owing to bad handling or more probably to some inherent difficulty, they did not metamorphose but soon died. The ease with which larvae of the same species may be made to fuse together suggests that larvae of different species might like- wise be induced to coalesce. Some experiments along this line could not fail to be of interest. IV In the tendency to fuse with the production of a plasmodium, the dissociated cells of sponges resemble the amcebocytes (amce- bula) of the mycetozoa and Protomyxa. ‘The regenerative power of the plasmodium has an interest both theoretical and economic in itself. But it is the tendency to fuse displayed by the cells that have been forcibly broken apart, which constitutes the fact of most general physiological importance. Discarding for the mo- ment the word “cell” and speaking of the protoplasm of a species as a specific substance, the phenomena may be restated to advan- tage in the following way. A mass of sponge protoplasm in the unspecialized state typically exhibits pseudopodial activities at the surface. In lieu of more precise knowledge it is useful to regard the pseudopodia as struc- tures which explore and learn about the environment. Oncoming EQUILIBRIUM OF ANIMAL FORM? BY HANS PRZIBRAM Biologische Versuchsanstalt, Vienna Witn Ten Ficures If the equilibrium of a mass be disturbed, the body will alter its position in regard to the surrounding neighborhood till it again gets into a position of equilibrium. If the form of an animal be altered through amputation of certain parts, the equilibrium of mass may or may not be altered therewith, the animal either being able to maintain its position or having to alter its posture for readjusting its equilibrium. But this is not all that may happen: it may restore its form, too, after some time, thus tending toward a new equilibrium of form, till it has reached a new stable condition. This may involve three regulatory processes: Regeneration of lost parts from the cut surface; reduction of existing parts in contact with the cut surface; compensation at parts of the body not touched by the amputation. The study of regeneration has long received much attention; of late reduction too has been studied more fully, especially in the lower animals, whilst compensation as a means of restoring animal form seems first to have been pointed out by me. Its study has been taken up especially in America, where Zeleny found additional cases (observed independently of my work), and Wilson, Morgan and Emmel have studied different aspects of compensatory regulation in crustaceans. Having found the principle of compensatory regulation illus- trated in the chel of Alpheus, I have been looking for other analo- gous cases and have found the same process in other crustaceans, especially Callianassa and the common Crabs, Portunus and 1 Read before the International Congress of Zodlogy at Boston on August 22, 1907. Tue JourNar or ExperIMENTAL ZOOLOGY, VOL. V, No. 2. 260 Hans Prabram Carcinus, as I demonstrated at the last session of the German Naturalists’ Association. It will be remembered that this com- pensatory regulation consists in the hypertypical growth of the smaller claw of the first pair of thoracic limbs, after autotomic removal of the big chela, whilst a hypotypical small chela regener- ates. That such a “transposition” or “reversal” need not be restricted to this pair of appendages I am now able to show in Typton spongicola (Fig. 1), where the second pair of thoracic limbs is developed into asymmetrical chele. After removal of the bigger chela, which, by the way, may be situated normally at the wen or at the left side of the body, reversal of the chet of this second pair is brought about (Fig. 2), the process being in all respects analogous to that inthe first pair of chelze in Alpheus: But not only may the means of regeneration and compensation be clearly shown to occur in this case, but also reduction is involved to an appreciable degree, especially if the crayfish 1s starved during the experiment. Then each moult shows the shedding of a smaller skin and the animal is at the end of the transposition in all dimensions smaller than at the time of the operation. Thus, as in the lowest animals, a proportionate diminution of the whole form may be produced as regulation proceeds, the only difference with Morgan’s “morphallaxis” lying in the bigger fragment necessary for reconstruction. A curious instance of “compensatory reduction” was met in some experiments on cutting the nerves in crabs. As Morgan has reported, the chela generally degenerates after this operation. In a few cases, however, I was fortunate enough to get a further growth of the limb. In these the terminal joint of the big or crushing claw was removed (Fig. 3),and in one instance regenerated in a rather reduced state; but also the corresponding dactylopodite of the smaller or nipping claw lost its differentiation (Fig. 4). Compensatory reduction may also occur in animals other than crustaceans and in other regions than in the chela. Megusar, working on regeneration in beetles (Coleoptera), amputated one of the two slightly differing jaws of the Hydrophilus larva. Whilst there are normally two feeth on the inner side of each jaw (Fig. 5) the larva appeared after the moult succeeding the amputation 261 Equilibrium of Animal Form fre. 2. FLGH: f16. 4. HIGHS: FIG. 6. Fie. 5. 262 Hans Przbram of but one jaw with but one tooth on the non-operated jaw, the regenerating one showing no teeth (Fig. 6). A tendency ‘toward a quick restoration of a symmetrical con- dition was also found several times in experiments of Miss Zuel- zer on the regeneration and moulting periods of the isopod, Asellus aquaticus. When both long antenne were removed, simultaneously, but at different levels (Fig. 7) they would be apt to appear regenerated to an equal length (Fig. 8) even after the next moult, though not having yet attained their normal length. In some way these conditions may be related to the different rate with which appendages regenerate from different levels. It is not necessary that the two correlated parts be symmet- rically situated. The male of the water-newt, Triton vulgaris, produces in its state of courtship a crest along its back and around its tail, which has a ragged appearance (Fig. 9). Kammerer cut the tail off to test the regenerative power of the secondary sexual characters. He found that at first the tail appears with smooth not ragged edges. But this is not all: the crest on the back had also lost its ragged appearance, thus conforming with the outline of the new dorsal rim of the tail (Fig. to). It is not shed or resorbed, but keeps the height of the courtship crest. The object of my paper is to emphasize the similarity of these processes of regeneration, reduction and compensation, in lower and higher animal forms, and their relation to the reéstablishment of the equilibrium of animal form. Equilibrium of Animal Form 263 FIG.Y. Fie. /0. 264 Hans Przibram LITERATURE Emme L, V. E.—Regeneration and the question of symmetry in the big claws of the lobster, Science, n. s., xxvi, 19, vil, 1907 (with literature). Kammerer, P.—Regeneration sekundarer Sexualcharaktere bei den Amphibien. Archiv f. Entwicklungsmechanik, xxv, 1907. Mecusar, F.—Die Regeneration der Koleopteren. A. f. Entwm., xxv, 1907. Morcan, T. H.—Notes on Regeneration. (Transpositional or compensator- regeneration of the large chela in some Crustacea, p. 169). Bio- logical Bulletin, vi, 1904. Przizram, H.—Die “Scherenumkehr” bei dekapoden Crustaceen, zugleich: Experi- mentelle Studien tiber Regeneration. Archiv. f. Entwicklungs- mechanik, xxv, 1907 (with literature). Witson, E. B.—Reversal of asymmetry in Alpheus heterochelis. Biological Bul- letin, iv, 1903. ZeELENY, CH.—Compensatory regulation. Journal of Experimental Zodlogy, ii, 1905 (with literature). Zueizer, M.—Ueber den Einflus der Regeneration auf die Wachstumsgeschwindig- keit von Asellus aquaticus, a. +4 > rat 7 cect ot lel | ONES et sca We (et cot OA: eet? eae eee ices len | Group A 10.5) -048| 8.4/1.5 | r79len.2 2. -059/21.0,2.5 |.119/15.5/3. -080/30.5|2.5 |.082/24.0/2. -034.29.01 .75| .060 35-014. © |.185|10.8|1.6 |.148 |.194 16.5/2.5 -I51|17.5|2-5 |.143 16.5|2.0 |.121 104 28.5 3.5 +123]25.0|2.5 |-100/25.0]1.5 |.060 /-134 26.02.85 .110|28.0/2.0 | .072|28 .0|2.15].077 -071/42.0/5.0 |.119/36.014. esas -122/34.5]3-0 | .087 32.5/5-0 154 | | | | .66 .058 \2.65|.212 3.241.144 13+171-157| 2.40|.117 2.45|.212 Group B 17.0|1. GroupC 22. 5|1. GroupD 29.01. 00 won oxrnoo Group EE |42.0)3. Average * Not strictly comparable with the others because the regenerating surface is at a deeper level. The average lengths of the regenerated arms are 1.66 mm. for the individuals with one removed arm, 2.65 for those with two removed arms, 3.24 for those with four, 3.17 for those with six, 2.40 for those with eight and 2.45 for those with the whole mouth apparatus removed. Since the individuals of a group are not exactly alike, and since the amount of regeneration is dependent on the size, the specific amount, 7. ¢., the regenerated length divided by the disk diameter was obtained in each case. The average specific amount of regeneration for each degree of injury is given in Table Lin italic type. It is seen that from a minimum for the case with the lowest injury it increases to an optimum when six arms are removed, beyond which it again decreases. ‘Thus with one removed arm the specific amount is .058, with two arms removed it is .112, with four .144, with six .157, and with eight ie A series of individuals with different extents of removed mar- 268 Charles Zeleny gins was studied with respect to the question of relation of degree of injury to the rate of regeneration, but because of distortions involving the whole umbrellar region, no adequate data were obtained. 2 The Effect of Successive Injury on the Rate of Regeneration The study of successive injury was confined to the margins. As in the last mentioned case there was considerable trouble with distortion of the disk. Eight individuals however were without distortion and could be used for the present purposes. The experiments on successive injury come under two heads. In one series a number of individuals of a size were chosen. In half of these the whole margin was removed and allowed to regenerate for twenty-nine days, at the end of which time it had nearly completed its regeneration. ‘This margin was then removed for the second time at the same hour that it was removed for the first time in the others. After twelve days the animals were killed and a measure- ment of the new margins gave a direct comparison of first and second regeneration as shown in Table II. TABLE II Width of regenerated margin in millimeters First regeneration Second regeneration a 8 2 1.6 Average... .:... 3 1.2 The second regeneration shows a decided advantage over the first. In a second series the first and second regenerations were com- pared within single individuals. A part of the margin of each individual was removed, and after it had nearly completed its regeneration It was removed again at the same time that a similar segment from another part of the circumference was removed for the first time. A direct elimination of indviduality was thus obtained. On account of individual differences in the method each case is described separately. In each, however, the first operation Regeneration in Cassiopea 269 came on June 17, and the second on July 16, twenty-nine days later. Twelve days were allowed after the second operation. In individual A about one-sixth of the margin was removed. The regenerated margin was 2.0 mm. in width twenty-nine days later. It was then removed a second time and a similar segment was removed from a part of the circumference which had not been injured. Twelve days later the animal was killed and showed a second regeneration width equal to 1.8 mm. anda first regenera- tion width of 1.6 mm. Fic.1 Fic. 2 Fig. 1 Outline of disk of Cassiopea xamachana in which segment B is a second regeneration and segment A a first regeneration of the margin. Fig. 2 Outline of the disk of Cassiopea xamachana in which the unbroken line between B’ and lower A is a second regeneration of the margin to be compared with a first regeneration shown between B and lower A. In individual B the character of the operation was the same as in A but the removed portion of the margin was one-fourth of the whole for each regeneration, and the fepenerced margin was 2.7 mm. wide as compared with the feared 4.2 mm. ie this case the second regeneration width is 2.0 mm., and the first 1.1 mm. The relative widths are shown in Fig. 1, where A represents the first regeneration and B the second. In individual C as shown in Fig. 2 in semidiagrammatic form the original outer circumference is shown by the dotted line and 270 Charles Zeleny outer unbroken line between B and A. At the first operation the half of the margin shown in the figure by the part to the right of a line between 4 and Ad’ was removed. This was allowed to regen- erate until it had a width of 3.5 mm. as compared with 4.5 mm. in the uninjured margin. The second operation was then made and consisted of a cut along the heavy line between B and B’ including five-sixths of the whole margin. This second operation included a strip of regenerated margin, and anequal strip of margin which had previously been uninjured. After twelve days the animal Fic. 3 Fic. 4 Fig. 3 Regeneration of inner and outer margins in a non-pulsating individual of Cassiopea xama- chana. Compare with Fig. 4. Fig. 4 Regeneration of inner and outer margins in a pulsating individual of Cassiopea xamachana. Compare with Fig. 3. was killed and the average width of the second regeneration (B’ to lower 4) was 1.4 mm. as compared with a first regenera- tion width (B to lower 4) of .8 mm. In all three individuals there is a well marked difference between the second and the first regeneration in favorof theformer. The character of the operation to a very large extent eliminates chance of individual error, and makes the result reliable. The five individuals of the other series strengthen the conclusion that in Cassiopea a second regeneration of the margin occurs more rap- idly than the first. Regeneration in Cassiopea PL 7/ A The Effect of Rhythmical Pulsation of the Disk on the Rate of Regeneration f I am indebted to Dr. A. G. Mayer for the method of obtaining a rhythmically pulsating disk. The margin with Its sense organs and the mouth apparatus is removed and the ventral surface is cut as shown in Figs. 3 and 4. Such a disk remains quiet if undisturbed, but a rhythmical pulsation can be produced by an electrical stimulation at X. This pulsation in some cases con- tinues for several days after a single stimulation. In an animal operated on in this way there are two regenerating surfaces, an outer one to replace the sensory margin and an inner one to replace the mouth apparatus. In my experiments a comparison was made between those with pulsating and those without pulsating disks, to see whether pulsation has any effect on the rate of regen- eration. TABLE III WIDTH OF REGENERATED WIDTH OF REGENERATED INNER MARGIN OUTER MARGIN Duration of Duration of pul- — = — regeneration sation in stimu- Stimulated Unstimulated | Stimulated Unstimulated in days lated individuals individual individual individual individual 1.0 2.5 aly 1.0 5 3 days 1.6 2.3 obs 1.0 5 4% days 1.3 6.0 6 A] 5 no pulsations 6.0 1.8 1.0 6 5 2 days 1.6 2.8 5 1.0 4 8-16 hours 2.8 2.8 1.0 1.0 4 8-16 hours 2.4 | 1.0 4 few minutes 1.2) | “5 4 no pulsations Two factors must be considered as entering into the result of the present experiment, first the stimulation of the animal as a result of the electrical shock, and second the pulsation resulting from this. It was very difficult to get pulsation that would con- tinue for a considerable period of time. In all cases it stopped before the completion of the experiment. The data are given in Table III, which shows that the effect of the stimulation and pulsation is inhibitory on the whole. In the table, as in the Def Charles Zeleny experiments, the individuals, except the last two are grouped in pairs, one member of each pair being stimulated, and the other one unstimulated. In four of the six pairs of cases the unstimulated individuals show a greater amount of regeneration than the stimulated ones. In one the two are equal and in the sixth the stimulated is greater than the unstimulated. DISCUSSION I The Relation of the Degree of Injury to the Rate of Regeneration The results obtained in Cassiopea agree with the general rule [ have found to hold true in the arms of Ophioglypha lacertosa and the chela of Cambarus propinquus and which Ellis finds in the legs of Mancasellus macrourus. The rate of regeneration of a removed appendage is determined not only by the character and position of the cut surface, but also by the character and extent of other injuries received at the same time. The rate increases with added injuries to other parts of the body up to an optimum which represents the amount of injury most favorable for regeneration. Beyond this point added injury causes a decline in the rate of regeneration. In the case of the arms of Cassiopea the optimum comes when six of the eight arms are removed. 2 The Relation of Successive In ury to the Rate of Regeneration ee § Careful investigations of the rate of regeneration after successive injury are rare in the literature. [he general statement is, how- ever, frequently made that the rate and character of regeneration are unaffected by successive injury within very wide limits. I have been able to find descriptions of three cases which do not agree with this statement. Vanlair finds that the sciatic nerve of the dog regenerates more rapidly after the second than after the first removal. Driesch finds in Tubularia that the development of the aboral hydranth is more rapid after a second than after a first removal. He also makes out an interesting case of effect of successive re- moval on the character of regeneration in Antennularia. The free basal end of the stem of this hydroid, after a first removal, develops 2 0 OY i , Sar 2 RD Rege neration in Cassiopea 273 stolons alone. After a second removal it develops one or more slender stems as well as the stolons, after a third removal two or three strong stems and only one or a few slender stolons, and finally after a fourth removal no stolons at all, the whole growth consisting of one or two very stout stems. To further test the question of the influence of successive injury upon regeneration, I made a study of the problem in the chelz of the Gulf-weed crab, Portunus Sayi, in the Scyphomedusan Cassiopea xamachana and in several other forms. ‘The report on the first of these is now in press, and the data show that while the second regeneration is greater than the first, nevertheless, when the age factor is eliminated the twoare exactly alike. In Cassiopea, however, in which the age factor is also eliminated, the margin shows in every case a greater rate of regeneration after the second than after the first removal. The material on several other forms is now being worked up. 3 The Relation of Functional Activity to the Rate of Regeneration In view of recent discussions concerning the relation of form regulation to functional activity a comparison of pulsating with non-pulsating disks of Cassiopea is of special interest. Con- trary to the general view that functional activity is an aid in effecting form regulation, it was found that pulsating individuals, with two exceptions, showed a slower rate of regeneration than non-pulsating ones. The result indicates that there is need of further investigation along this line before general conclusions are made. 274 Charles Zeleny LITERATURE LIST Driescu, H., ’97—Studien tiber das Regulationsvermogen der Organismen. I. Von den regulativen Wachstums—und _ Differenzirungsfahig- keiten der Tubularia. p. 393, Archiv. f. Entw. Mech. d. Organis- men, Bd. v. Exris, M. M., ’07—The influence of the amount of injury upon the rate and amount of regeneration in Mancasellus macrourus (Garman). Biol. Bull., vol. xiii, no. 3. Mayer, A. G., ’06—Rhythmical pulsationin Scyphomeduse. Carnegie Institution of Washington, Publication no. 47. Vantalr, '94—Archives de Physiologie normale et pathologique. Zeveny, C., ’°03—A study of the rate of regeneration of the arms in the brittle- star, Ophioglypha lacertosa. Biol. Bull., vol. vi, no. 1. o5—The relation of the degree of injury to the rate of regeneration. Jour. of Exp. Zodl., vol. ii, no. 3. In Press. Some internal factors connected with the regeneration of the chela in the Gulf-weed Crab, Portunus Sayi. Carnegie Insti- tution of Washington. STUDIES IN ADAPTATION I THE SENSE OF SIGHT IN SPIDERS BY ALEXANDER PETRUNKEVITCH, Ph.D With Six PrLates Whether the organization of an animal is the result of contin- uous adaptation to the conditions of life, or whether the animal adapts itself as best it may to the use of organs in the possession of which it has been put by nature, an impartial observer sees everywhere the remarkable balance between structure and function. The few apparent incongruities which we meet with in all groups of the animal and vegetal kingdoms, only serve to confirm the rule, since they are due lareely © to our lack of knowledge of the conditions under which such organs exist to their best eee or of the role which they played in the phylogenetic development of the species. But however this may be, we should expect to find, and there surely must be a marked difference between the origin of organs by means of which the animal communicates actively with the surrounding world and that of such organs or structures as lie entirely outside of the personal activity or con- trol of the animal. It 1s very probable that many, perhaps the majority of arthropods are entirely unaware of the presence on their bodies of so-called decorative structures, designs or colors and whether a male possessing a slight variation in such a structure, design or color, will or will not be preferred by a female, is to my mind entirely a matter of chance. In this con- nection I can only confirm the observation of earlier naturalists that females often choose defective males in the presence of other in every respect perfect ones. I have caught in Europe a pair of large beetles (Prionus coriarius), in coitu in the presence of two other perfect males, when the male in question had Tue JourNar oF ExpertMENTAL ZOOLOGY, VOL. V, NO. 2. 276 Alexander Petrunkevitch one eliter and one leg malformed and showed other minor defects. Upon investigation the male proved to have normal spermatozoa and the female laid an extraordinarily large number of eggs. Similar cases are common among all insects as well as other animals. I have made the same observation in the beetle, Lucanus cervus, the males of which have the beautiful horns on their heads. Of the beautiful moth, Samia cecropia, over a hun- dred of which I allowed to develop in a cage for other experi- ments, many couples with entirely defective and malformed wings’ were found in coitu. Among orb-weaving spiders are several species in which during the mating period the female allows several males to remain in her net, of course at a good distance from her, and it is a common occurrence for defective males to be accepted by the female while perfect ones will be chased away for their dear life or will even fall victims to the voracity of the much stronger female. Protective structure or coloration belongs to the same group of phenomena and here a great number of cases might be recorded where an animal remains entirely unaware of the protection afforded by its form and color if in the proper sur- soundings and of the danger of disregarding this. I have seen butterflies that would be protected by the leaf-like coloration of their wings if they should sit motionless, instantly caught on the wing by large dragonflies, and in the West Indies I could observe regularly a hemipteron of the group Emesidw, which would resemble a dead twig if in an appropriate position, swinging to and fro on the four hind legs in the center of a large leaf. All this I bring here by no means with the object of denying the existence of adaptation or of the principle of selection itself, which only one who has never observed nature outside of the laboratory can do. But I want to show that in all those instances where the animal itself cannot make use of a new variation or mutation, in the possession of which it has been placed by nature, its advanta- geous character will be a mere matter of chance, and selection, if it take place, must be greatly retarded in its progress. Quite different must it be with those structures which the animal is able to put to active use, however unconscious it may be of the advantage it enjoys over other individuals of the same species. The Sense of Sight in Spiders 277, A caterpillar that is able to construct a better protected cocoon than its fellows, an animal that can run faster from danger, a bird of prey that distinguishes its victim at a greater distance, all these will evince their superiority over other individuals of the same species upon the first occasion. Here, then, opportunity is given for a more rapid accumulation of characters and an acceler- ated selection. Several years ago I studied, together with Dr. von Guaita, although with a different object in view, the stridu- lating organs of One These organs are undoubtedly not of a protective character as the insects in question are safer when silent. This is sufficiently shown by the fact that they stop their cry instantly upon the approach of danger. The origin of the stridulating organs has therefore to be sought in some advantage which they may afford the sexes during the mating period. That they are a rather late acquisition is evidenced by their late appear- ance, only with the last moulting, while the rapidity of selection which in this case must still have been retarded by the danger arising from their possession, is apparent from the complete absence of any traces of similar organs previous to the last moult- ing. The great variability in the number of teeth on the chord or bow of the stridulating apparatus, which permits of difference of pitch in the tones produced by friction, the fact, with other words, that the stridulating organs are in respect to their details, not absolutely fixed structures, seems to be due to this antagonism between the advantage afforded by the organs in the relavon of the sexes toward aan other and their dieade antage In respect to diminished safety from the aggression of enemies. But how is it with organs in which the advantage arising from their perfecting will be immediately exploited to its fullest extent by the animal possessing them? With this question before me I commenced my research on the sense of sight in spiders. The results of this research which now extends over more than a year, | bring in this paper. Two factors in the life of spiders have left a deep impres- sion on their organization, the first, that they subsist exclusively by means of prey and the second, that the external sexual organs, 278 Alexander Petrunkevitch the organs of copulation in the male, are entirely separate from the internal sexual organs. If we may judge of the amount of food required by the quantity of insects captured in the orb-net of a single spider, a quantity that is sometimes appalling, it was necessary for the spiders, in order to satisfy this want, to develop as they indeed have done, either the instinct and the engineering capacity for constructing nets or else the instinct for hunting their prey freely on the ground or on plants. ‘Two different directions are thus given to the development of the whole spider group and we should naturally expect to find differences in the structure of the corresponding organs of sense and of the spinning apparatus. But while in higher animals, as for example, birds of prey, all organs of sense are developed to a remarkable degree of perfection, we cannot say the same for the hunting spiders. It is a matter of common observation that vultures discern their prey from a distance amounting at times to several thousand feet and they are undoubtedly able to scent a carcass hidden among bushes at a considerable if not so great a distance as this. Their ears are sensitive to high and low sounds of a very small amplitude and only the sense of touch which could scarcely be of use to them, is little developed. It is otherwise with the hunting spiders. They possess doubtless, a very fine sense of touch, the whole body being covered with hairs and bristles sensitive to the slightest stimulus. In regard, however, to the organs of hearing and smell we are as yet without definite results of any kind. At any rate these two senses are very little developed. Even the ques- tion as to the existence of such organs is to my mind far from being settled. Whether the organs discovered by Dahl and the lyriform organs have anything to do with these senses must be determined by new investigation. I myself have made some experiments on the sense of hearing, thus far without any definite results. I alsorepeated the experiments of Pritchett on the sense of smell in large lycosids but even to such irritants as formalde- hyde, osmic acid and acetic acid, they did not respond soreadily or so quickly as to the slightest touch with the end of a silk thread. As the sense of touch 1s solely protective, there remains only the sense of sight to guide the spiders on their hunting trips. The The Sense of Sight in Spiders 279 splendid experiments of Mr. and Mrs. Peckham on the jumping spiders (Attide), during the mating period, have clearly demon- strated that the sexes recognize each other by the use of their eyes alone, the male remaining unaware of the presence of a female, nor will he perform the peculiar, characteristic love dance, if his eyes are covered with paint. But is the sense of sight in spiders as sharp as we should be led to expect by comparison with that of birds of prey f This question has not yet been answered with sufficient certainty owing to the difficulty of experimenting, Plateau came to an enacts negative conclusion, asserting that spiders possess an indistinct and very poor vision, being unable to discern objects beyond a distance of from 8 to 10 cm. Forel was of the same opinion, while the more ingenious experiments of the Peckhams leave no doubt that the males of the Attids recog- nize their females at a distance of about 30 cm. and moreover that they distinguish colors. But whether spiders can see beyond this distance and how sharp their vision is, the experiments of the Peckhams also, leave unanswered. In fact it would be impos- sible to answer this question merely by observing the behavior of spiders during an experiment or in nature. A close anatomico- physiological study is required and only by combining the experi- ment on the living specimen with its after examination may one reach a satisfactory answer. [This answer I hope to have given in the present research. Keeping in mind all that has been said in the preceding pages, we may conclude that the study of the eyes of spiders and of their sense of sight, as examples of adaptation of the first kind, pos- sesses several advantages as well as certain disadvantages and these we have next to consider. “The advantages are: 1 That the sense of sight is beyond any Spake the only sense that guides hunting sniders on their hunting excursions and in finding the females during the mating period. 2 That the eyes of spiders are organs which are for each species definite in number and position on the cephalothorax. That the sense of sight is a sense common to the great majority of lower and higher animals and that some analogizing is therefore not only admissible but may be of great value. 280 Alexander Petrunkevitch The disadvantages are: 1 That beyond some experiments of Plateau and the Peck- hams, nothing definite is known in regard to the acuity of vision in spiders and no method has been brought forward for its study. 2 That the eyes are complicated organs, consisting of a refrac- tion and a perception apparatus, each of which and the separate parts of which may possibly be capable of adaptation and which have therefore to be considered separately. Let us now begin with a closer study of the ocular group on the cephalothorax of various adult spiders. THE POSITION OF THE EYES ON THE CEPHALOTHORAX The number of eyes in the great majority of spiders is eight and the group which they form on the cephalothorax is so characteristic for the different families that it was for a long time used as a systematic character of great value. The group con- sists usually of two or three rows, more rarely of two or three smaller groups containing three or two eyes each. Of importance for systematics are the length of each row, the distances between the eyes taken in relation to their diameters, the form of the row, whether straight or bent in the middle forward or backward, as well as the shape and to some extent the color of the eyes. Even a superficial observer will notice in addition, that both in the jumping and the ground spiders (Attida and Lycoside), two of the eyes on the forehead are larger than the others, especially in the former. But the reason for such a configuration in the eye- group was never sought for and the apparent similarity in the eye-groups of spiders belonging undoubtedly to different families, led to the conclusion that the value of the eye-group as a sys- tematic character had been overestimated. The position of the eyes on the cephalothorax becomes more comprehensible when we begin to study the directions of their respective axes and the angles that these axes form with the three chief planes of the body. Unluckily we at once meet with great difficulties even though we choose species having perfectly round eyes. ‘The first of these is to ascertain with exactitude the posi- The Sense of Sight in Spiders 281 tion of the two planes intersecting the plane of symmetry of the body. After experimenting for a long time I decided upon the following method. The spider is killed in alcohol and then kept in it for several days. When the muscles have become entirely rigid, so as to allow of no change in the shape of the cephalo- thorax while drying, the abdomen is severed from the cephalo- thorax through the petiolus with a sharp pair of scissors. Next the pars labialis, together with the laminz, the palpi and the chela, is carefully removed with forceps. At the same time care must be taken not to tear the chitinous bridge connecting the opposite sides of the cephalothorax immediately behind the chelz since this would lead to a flattening of the cephalic part and consequent distortion of the true angles. ‘The legs are now carefully removed leaving the cox alone in their normal position attached to the sternum. Next a thin line is drawn with chinese ink on a very thin layer of Canada balsam that has been spread with the finger over a small cover-glass. It is best to use rather a thick cover- glass and to draw the line quite across it parallel to two edges. A drop of fish glue is now put in the center of the glass and on this the cephalothorax is placed with the sternum toward the glue and gently pressed until all the cox and the sternum are in contact with the glass. The glass now represents one of the two planes intersecting the plane of symmetry at right angles. I shall call it the horizontal or foundation plane. It is only approximately parallel to the surface of the earth or to that of any object upon which the spider may be, since the spider is able to raise itself up on either front or hind legs, in this way changing the angle between the foundation plane and the horizon. At the same time it is essential that the plane of symmetry should coincide with the black line on the cover-glass. This is accomplished by using a needle under the microscope at a magnifying power of about twenty diameters. It is now easy to determine the plane of sym- metry by taking the point midway between the front middle eyes, the central, longitudinal groove of the cephalothorax when it is a species possessing this groove, and the point at an even distance between the two chitinous plates of that part of the petiolus which remains with the cephalothorax after the abdomen has been 282 Alexander Petrunkevitch severed. The third plane which I call the vertical or transverse plane and which intersects the two other planes at right angles, is geometrically determined by these. After the cephalothorax has been prepared in this manner and thoroughly dried in a warm place, the cover-glass upon which it is fixed is laid upon the center of a square glass plate, which is at the same time the center of a circle drawn upon the surface of the plate with radii forming angles each of which measures ten degrees. If we now use an eye-piece having two lines intersecting in the center at go°, we can readily measure the angles that the eye-axes form with the plane of sym- metry. To know how this is done we must remember that the eyes of spiders have each a lens, the outer surface of which forms part of the surface of a sphere. When we look at a lens in the direction of the eye-axis, it appears to us as a circle while if we look as it under an angle of less than go°, its outer surface appears to us as two curves intersecting each other at two points. If we place the eye in the center of the microscopic field and move the eye-piece until one of its lines falls upon these two points of inter- section of the curves of the lens and the other passes through a point midway between them, then the latter line represents the projection of the eye-axis on the horizontal or foundation plane and the angle that it forms with the plane of symmetry can be read directly from the scale. More exact results would of course be obtained by using a goniometer ocular such as is made by Zeiss, but even the arrangement I have described here, the ocular with intersecting lines, gives an error of not more than a few (3 to 4) degrees. When the angles have in this way been measured a drawing is made with the aid of an Abbe drawing apparatus of the entire eye-group and the eye-axes are then drawn in, in accordance with the data given by the measurements. We obtain in this way a correct figure of the projection of the eye-group on the horizontal or foundation plane. In order to gain a clear pic- ture of the position of the eyes it is necessary to make two other projections, one on the plane of symmetry and the other on the vertical plane. ‘Lo accomplish this I put a small drop of beeswax in the center of the circle on the glass plate into which the edge of the cover-glass holding the cephalothorax is pressed. This The Sense of Sight in Spiders 283 cover-glass is held in a vertical position with the aid of two straight angles which are removed as soon as the wax is hard enough to keep the glass in position. Another method, much simpler and perhaps just as good, consists in adjusting the cover-glass until its upright edge forms a straight line with its own reflection in the slide. The angles are now measured in the same way as before. The drawing is of course made under the same magnif- cation and the upper edge of the glass holding the cephalothorax is likewise drawn. It represents a cross-section through the horizontal or foundation plane. To make a drawing of the projection on the vertical plane one proceeds in the same manner with that difference that the cover- glass carrying the cephalothorax is placed on its other edge. Two facts become immediately apparent upon comparing the drawings made in this way, of eyes of spiders belonging to different families. First, that there is not a single pair of eyes which are focused upon a single point like, for instance, the eyes of man. On.the contrary, the axes of all eight eyes are so directed as to form divergent angles with each other. Second, that not only do the positions of the eyes on the cephalothorax of spiders belong- ing to different families differ from each other, but the axes of the same eyes in different spiders do not lie in the same direction but form with the three planes of the body, the planes of projection, angles differing considerably from each other but fixed for each species. Yo make this clear let us look at the drawings (Figs. 1-9) representing the eyes of three hunting spiders, Lycosa nidicola—a common large ground spider of morhenn America, Phidippus tripunctatus—a large ] jumping spider belonging to the same region and Heteropoda venatoria—a samiGnsien fropiel and Sricecel spider of very large size, belonging to the family Heteropodidz and resembling i in habitus the crab spiders (Thomis- idz) among which it was Saat placed by earlier systematists. Translating the results of these measurements into common language, we may thus describe the positions of the eyes in the three spiders. The anterior middle eyes in the jumping spider, Phidippus tripunctatus, are directed forward and a little outward and downward. In Lycosa nidicola more outward and consider- 284 Alexander Petrunkevitch ably upward. In this respect Heteropoda venatoria resembles Lycosa more than it does Phidippus, since its anterior middle eyes are also directed frontward but still more outward and con- siderably upward although not so much so as in Lycosa. TABLE I Projection on the horizontal or foundation plane. S e Figs. 1,2 and 3 The angles which the axes of the eyes form with the plane of symmetry. The right side with a plus, the left with a minus sign AME | ASE PME PSE Phidippus tripunctatus.......-......... | Se 8 ZI 67 95 MY cosaynidicol acim jer (e cpese csc erelater-lelesniste | a= 12 24 24 94 Heteropoda venatoria...............+0- + 32 27 ° | 109 TABLE II Projection on the transverse or vertical plane. See Figs. 6,8 and Q The angles which the axes form with the plane of symmetry. Signs as before AME ASE | PME | PSE Phidippus tripunctatus..............+-. ae P 57 | 63 EY.COSaLDIAICO] ay terete ators ie) =1=:¢Fo sie l=|= cisisiets s ae ; 155 ? 56 Heteropoda venatoria.............+.++. | Be ? 8 | 71 TABLE Ul Projection on the plane of symmetry. See Figs. 4, 5 and 7 The angles which the axes of the eyes form with the horizontal or foundation plane. Zero in front of the head, 180° at the back. Positive quantities for eyes looking upward, negative for those looking down- ward. | AME | ASE PME PSE Phidippus. tripunctatus. 2.52. .4..s-0+--- —2 | ° +<90 + >go Tey cosa midicolavier.ciseieveys:e1eve'=isi elas stete/=tnalo +14 | —18 + 10 + >90 Heteropoda venatoria...........++-eeeeeee + 8 | +8 + 92 The anterior side eyes in Phidippus are so directed that their axes are parallel to the horizontal or foundation plane and turned a little sidewise. In Lycosa they are directed a little more out- ward and at the same time downward. In Heteropoda they are directed still more outward but at the same time upward. The Sense of Sight in Spiders 285 The posterior middle eyes in Phidippus are directed consider- ably toward the side and upward. In Lycosa they are directed much more frontward and also upward, while in Heteropoda they look straight upward. ‘Their axes are in this case almost perpendicular to the horizontal plane. The posterior side eyes in Phidippus are directed sidewise, considerably upward and a little backward. In Lycosa they are directed in the same way sidewise and backward but considerably more upward, while in Heteropoda they are directed considerably less upward and much more backward. I studied the positions of the eyes as I have described them here, originally in perfectly ripe females. Six individuals of each species were measured and showed only slight differences in the angles. Since the method employed is not an absolutely exact one, it is difficult to say whether these differences depend upon variation in the position of the eyes or upon defects in the method itself. However, the following facts speak rather for defective measurements than for natural variation. I expected to find that during the post-embryonic development the position of the eyes on the cephalothorax would change with each moulting, approaching more and more nearly to that of the adult female, which should be considered an adaptation to the particular_life of the spider. An observation that made this seem still more probable has been made by various scientists at different times, that the eye-group in young spiderlings occupies a relatively larger part of the qoute oes: than it does in the adult spiders. Nevertheless there seemed to me to be occasion for a more thorough study of the eye-group and the eye-axes in spiders of different ages. I had in my possession perfectly ripe females of the three species I have mentioned, several specimens of Heteropoda just before the final moulting, others that had to moult twice and three specimens of very small spiders that had still to moult at least three times before attaining maturity and also a cocoon filled with very young spiderlings. I kept sev eral females of Lycosa nidicola for a time in large glass jars and preserved both mother and the spiderlings which were in part taken from the cocoon, in part killed while on the back of the mother just as they were about to 286 Alexander Petrunkevitch leave her. I had, besides, several unripe specimens of unknown age. I had also Phidippus in the same stages as Lycosa. It 1s not difficult to obtain these since the female of this species make a tent of web in which she lays the eggs, afterward guarding the young ones for a considerable time. | caught in addition several females of Pardosa nigropalpis with young ones on the back. The method employed is as follows: Instead of measuring each eye and the distances between the eight eyes in each spider, I make with the aid of the Abbe apparatus a drawing of the entire eye-group of an adult female. Leaving the drawing in the same position on the drawing table, | remove the adult female and substitute a younger one. I then try different objectives and oculars until the image on the paper of the eye-group of the younger spider is of the same size as the drawing of the adult. A sheet of clean paper is now put in place of the one with the drawing and the new drawing is made. In this way drawings are obtained of all stages. The angles of the eye-axes are next measured and the axes drawn in on the corresponding figures. In all cases, begin- ning with the young spider at the time when it is ready to leave the mother in order to commence its own, independent existence and ending with the mother herself, the configuration of the eye- group and the angles of the axes proved to be the same and the drawings made of them on paper, when superposed and examined against the light, coincide absolutely. But this does not apply to the youngest spiderlings, those taken directly out of the cocoon. Although in such spiderlings the eye-group is in general very nearly the same as in the adults, careful measurements show dif- ferences which, while not appreciable to the unaided human eye, are nevertheless of great importance. I give these measurements here (Table [V) but shall discuss them farther on when exam- ination of the fields of vision will reveal more clearly their significance. It is unfortunately still more difficult to measure the angles in such spiders than it is in older ones so I give only figures of which | am certain. If we compare these tables with those for the adult females we shall at once notice the following differences. In the youngest spiderlings of Phiddippus and Lycosa the anterior middle eyes The Sense of Sight in Spiders 287 are directed a little more outward and in Phidippus a little more downward also, than the same eyes in the adult. In Heteropoda, on the contrary, the anterior middle eyes of the spiderling are directed much more frontward than in the adult. The anterior side eyes in Phidippus and Lycosa are directed much more toward the side than in the adult, in Heteropoda much more toward the front. The angles of projection on the two TABLE IV SHOWING THE ANGLES THAT THE AXES OF THE EYES FORM WITH THE THREE PLANES OF THE BODY IN VERY YOUNG SPIDERLINGS a Projection on the horizontal or foundation plane. Compare Figs. 4, 5 and 7 AME ASE PME PSE Phidippus tripunctatus................. Be 10 30 ; 88 Tey COSAapNIGICOl aiatetsrasstetesayelo\sjevs/=\ele-=iche"= lai* | + p 30 25 85 Heteropoda venatoria.................. Be II 22 18 60 b Projection on the transverse or vertical plane. Compare Fig. 9 AME | ASE PME PSE Phidippus tripunctatus................. ee ? ? ? ? Wyycosamidicola'ejereroforcinvessi-\-)-ite)atsjej-isiere% se ; ? ; 68 Herteropoda-venatoniaz: «20.2.0. ose So ? ? ? ? c Projection on the plane of symmetry. Compare Figs. 4, 5 and 7 = ——— AME ASE PME PSE i} Pinot pine tee ete eek tewe is = ° | ; ? ycosamidicolacnaesscteaitictsts Seveciecte= P ? | + 30 (Hetero podaivenatortasjetatcicleisie + cstie'siciieieie ce ; ? =e 7 ? other planes could not be ascertained for these eyes. It is impos- sible to study the posterior middle eyes in the young spiderlings of Phidippus at all, on account of their extreme minuteness. In Lycosa the projections of the axes of these eyes on the horizontal plane is approximately the same as in the adult but their pro- jection on the plane of symmetry shows that they are directed upward at an angle about three times as great as in the adult. 288 Alexander Petrunkevitch In Heteropoda spiderlings they are directed somewhat sidewise and nearly upward but more toward the front, while in the adult they are directed straight upward and a little backward. The posterior side eyes in the spiderlings of all three species, especially in Heteropoda, are directed a little frontward instead of backward and in Lycosa less upward, also, than in the adult. The relative sizes of the eyes and of the distances between them are also different for spiderling and adult. Tables V, VI, VII and VIII may serve to illustrate this. TABLE V Lycosa nidicola. Diameter of eyes in millimeters y f | AME | ASE PME PSE Moo thertage ict crctaaiete tate atte ere astars aptors | 0.361 | 0.279 | 0.689 0.541 Spiderling, ready to leave mother............ 0.074 | 0.057 | 0.148 0.115 Spiderling, taken out of acocoon........... ‘| 0.049 0.038 0.115 0.115 Or in proportion to the anterior middle eyes which we take as unit of comparison | AME ASE | PME PSE Mother Sakai. ceiccin tceats cin ieiere rote wie eres I | 0.77 1.90 | 1.50 Spiderling, ready to leave mother........... I 0.77 2.00 1.55 Spiderling, taken out of acocoon............ I 0.77 | 2.30 2.30 TABLE VI Heteropoda venatoria. Diameter of eyes in millimeters | Length of | | | | cephalo- | AME | ASE PME ESE | thorax Adiultsterma lessees cyele ne lesei-eeynatera' te | 10.6 0.45 0.75 0.55 | 0.70 Immature female before last moulting....| 6.8 | 0.28 0.55 0.40 | 0.50 Two moultings before maturity.......... | 5-4 0.22 0:43 - 1OT gl aee Or in proportion to the anterior middle eyes | AME ASE | PME | PSE | AGN Manel Ds gone anese aeonenuaanpoenaadd I 1.66 1.22 Te55 Immature female before last moulting....... I | 1.93 | 1.43 1.78 Two moultings before maturity............. I | 1.95 1.54 | 2.04 The Sense of Sight in Spiders 289 TABLE VII Pardosa nigropalpis. Diameter of eyes in millimeters | Length of | cephalo- | AME ASE PME | PSE thorax | | Mother pe tee jem tetstarsciatere starter cenit tet rys- 2.8 O.11S O.1IS 0.328 | 0.246 Spiderling, ready to leave mother......... 0.85 0.049 0.049 0.115 | 0.820 Or in proportion to the anterior middle eyes AME ASE PME | PSE IMothensemmcrrerprisc cir ate aerate aie. se | I I 2.85 | 2.13 Spiderling, ready to leave mother.......... | I | I 2.34 | 1.67 TABLE VIII SHOWING THE DISTANCES BETWEEN THE EYES AND BETWEEN THE EYES AND ThE EDGES OF THE CEPHALOTHORAX Lycosa nidicola. Measurements in millimeters mes & Gis 4 | g28 s a 22 5 | eee | cra o Bb o ae son oO 8 ac) o 2107 a ‘34 we! % af J ok a5 $=2n 2 $yaes ain 3S | a2 St ist Oo. tS Se a bo $)2° oid =e) 28 23a as Boos oO eo H aa 8 I a Adultfemale rc ryatlerisie oleate 8.1 6.0 aie) Tear 2.2 Spiderling taken out of a COCOOD. Hane jate-nioe che tietseyavel- 0.984 0.771 0.459 0.197 0.180 Or in proportion to the length of the cephalothorax as a unit | Adultstemalesmrrremtertctasrs I 0.74 0.27 0.13 0.27 Spiderling taken out of a (elke coagapubeondvecnuK I ‘ 0.78 0.46 0.20 0.18 Heterapoda venatoria. Measurements in millimeters INE Gn Alsace seouane on | 8.0 8.0 3-2 1.4 1.6 Spiderling taken out of a | | (Selsey GnoohogdacenroododG 1.017 ©.go2 0.525 0.180 0.131 Or in proportion to the length of the cephalothorax ; a“ 7 2 Adil teitern aleryeterecciareterstsie ave I Xie | 0.4 0.17 0.20 Spiderling taken out of a | | COCGOMs seratata pistes terete hrcte I. 0.88 | 0.51 0.17 0.12 290 Alexander Petrunkevitch By a comparison of all these tables we may now gain consider- able light on the question as to what happens to cephalothorax and eye-group during the post-embryonic development. When the spiderling sheds its first skin in the cocoon, the eye-group occupies almost the whole breadth of the cephalothorax which is comparatively very low. The first change, of which I shall speak farther on, is in the directions of the eye-axes. When the spiderling leaves the cocoon, that is, before the next moulting takes place, the eye-axes have become fixed in the positions which they will occupy during the whole life of the growing and mature spider. Whatever change takes place from the time the spiderling leaves the cocoon, is only in the relative size of the space on the cephalo- thorax occupied by the eye-group and to a certain extent, in the relative sizes of the eyes themselves. The cephalothorax grows more rapidly than the eye-group,so that the latter occupies with each moulting a relatively smaller part of the cephalothorax. We shall presently see that this also is of advantage to the spider. THE MAXIMUM *ANGLE AND THE FIELDS OF VISION While handling under the microscope a dried out cephalo- thorax from which all organs and muscles had been removed to permit of the study of the endoskeleton, I chanced to notice that the faint images, visible in the eyes, of the trees which grow before my laboratory window, were not of the same size. I very soon found that boiling or even keeping in a cold solution of potassium hydrate so changes the optical property of the eye-lens that it becomes entirely intransparent so that it is necessary to use some other method. I have finally adopted the following one. The spider is killed in strong alcohol from which it is at once removed. The abdomen, all appendices and the sternum are then removed and the organs and muscles filling the cephalothorax are care- fully taken out with a forceps. Next the inside of the cephalo- thorax must be cleaned under water with a soft brush. I use a small camel’s hair brush of the kind used for water colors but cut the hairs quite short. With this brush it is possible to remove all the remaining muscles as well as the vitreous bodies of the The Sense of Sight in Spiders 291 eyes. Care must be taken, however, not to remove the black pigment ring surrounding each lens on the inside surface of the cephalothorax and forming a sort of iris, as this would make a difference in the measurements. Even in spiders that have been kept for a long time in alcohol the lenses are often still so transparent that one may see the images formed by them, but such lenses are yellow and the images rather poor. If on the contrary the eyes are prepared in the manner just described, it is scarcely possible to give an idea of the beauty of the little images. The lenses are then entirely colorless and transparent and the images render correctly color and line. For the study of the fields of vision, each eye together with a little of the surrounding chitin must now be cut out of the cephalo- thorax. ‘The lens is then placed with its inner surface on a small drop of liquid on a slide and in a hanging position examined under the microscope through the slide. By this means three advan- tages are gained. First, the object examined sends its rays through the lens in the normal direction so that the eye of the examiner is substituted for the retina of the spider; second, the observer looks in the direction of the eye-axis or at least very nearly in that direction; and third, the outside surface of the lens remains dry, limited by the air alone, as is the case with the living spider. More difficulty is presented by the fact that the refraction- coefhcient of the vitreous body is not known. However it is sometimes possible to prepare an eye fresh with the vitreous body in its natural position and the retina cut off with the aid of a razor. We are then able to measure the image of a scale at a given distance. But since the vitreous body coagulates too rapidly to be used in the study of the maximal and minimal angles of vision, we have to use in its stead a drop of water and also of some liquid possessing a higher refraction-coefhcient than the vitreous body. With the latter we obtain a somewhat larger image than in reality and may overestimate the acuity of vision. Such a liquid I found in a mixture of equal parts of pure glycerine and egg albumen. ‘The results obtained by the use of water we may then employ for control, to guard us against the opposite extreme of an underestimation of the acuity of vision. 292 Alexander Petrunkevitch The proportion between the two media is according to my measure- ments as 4 : 5, 7. ¢., that an image will occupy four divisions of a scale if water is used as the medium of suspension as against five divisions when glycerin-albumen is used. The microscope also must be arranged in a special manner. The diaphragms with the mirror and the Abbe lens must be removed. ‘The instrument is then placed on a high box open to the window, with a long slit occupying the whole space between the legs of the stand. Next a scale in the form of a cross with right angles is drawn on bristol board. Each arm of the cross is two centimeters wide and consists of alternating black squares like those on a checker-board. For the sake of convenience as well as to avoid error, the numbers are written in each white square in roman numerals in one direction and arabic in the other. This cardboard is now placed under the microscope so that the distance between it and the spider’s eye is exactly 10 cm. Excessive light around the eye is excluded by means of a small diaphragm or a black paper with a small round hole arranged so that the spider’s eye hangs directly in the middle of the hole. Of course any objective with small magnifying power may be used. As for myself I either use the a* or the A achromatic system of Zeiss and the compensation ocular 6 with the ocular micrometer. This micrometer is adapted to apochromats but may just as well be used with achromats if one ascertains the size of each division. In my instrument each division of the micrometer with the A objective corresponds to 0.0164 mm., while the correction of the a* lens makes possible a magnification where each division will correspond to 0.1 mm. .The light reflected from the white bristol board on which the scale is made, is sufhcient to give a perfect image of the scale in the microscope. The scale is so placed that the center of the cross falls exactly on the axis of the spider’s eye. | found that the eyes of the hunt- ing spiders are quite round and that the maximum angle of vision is therefore the same in each direction, 7. ¢., the limit of the field of vision in such eyes 1s a circle representing the circumference of the base of a cone. In order to find the maximum angle in each case, there remains only to read on the image of the scale The Sense of Sight in Spiders 293 in the eye, how many centimeters are visible. Since the dis- tance between the spider’s eye and the scale equals 10 cm. a simple calculation will give the value of the angle in question, or it is still simpler and entirely sufficient for our purpose, to draw on paper an isosceles triangle, the base of which must be as many centimeters long as are visible on the image of the scale in the eye and its height ro cm. ‘The angle can now be directly meas- ured. When the angles have been measured for all the eyes, they are represented on the drawings in each _ projection, showing the field covered by each eye. Optically the angle depends upon the curvature of the lens and the refractive coef- ficient of the substance of which it consists. But what is of interest for us here is the general fact that the larger the spider’s eye, the smaller, as a rule, is its field of vision. If we compare the drawings of corresponding projections in different spiders after the maximum angles of vision have been introduced, we cannot fail to recognize the remarkable relation between the particular life of the spider and the position of its eyes. In order to make this clear I must state here that which I shall prove farther on, that the larger the spider’s eye, the sharper Its vision or power of distinction. Let us begin with an examina- tion of the projection on the horizontal or foundation plane (Fig. 1). We see that the largest eyes in Phidippus are the anterior middle ones covering a field of 40° each or both together about 55°, owing to the fact that their axes are a little divergent. Each of the anterior side eyes also covers a field of 40° or both together 3°, 1. e., more than the entire field covered by the AME! But the ASE are considerably smaller than the AME. The minute eyes of the second row, the posterior middle eyes, cover a field of 62° and the PSE one of 48° or the whole eye-group covers about 240° of the horizon. The projections on the other two planes (Figs. 4 and 6) show in addition that the PME and the PSE guard chiefly the sides of the spiders, leaving about 52° in the vertical plane and more than 80° at the back on the dorsum, entirely unguarded. ‘This is the only direction from which the ‘These are abbreviations commonly used by arachnologists. AME stands for anterior middle eyes; ASE for anterior side eyes; PME for posterior middle eyes, and PSE for posterior side eyes. 2.94 Alexander Petrunkevitch jumping spider can be taken unawares by an attacking enemy. We know besides from the behavior of the spider that when- ever an insect or anything else approaches it from the side, it immediately turns toward the intruder as though with the desire to see it better by using its front eyes. Lycosa nidicola is a spider that lives on the ground under stones, making excursions in the grass. Its manner of walking like that of all ground spiders, is distinctly straight forward and we find that the largest eyes, the posterior middle eyes, are so situated as to guard the front of the animal. In the projection on the horizontal (Fig. 2) plane they together cover a field of 48°, 1. e., considerably less than the four eyes of the front row, which cover all together a field of 77°. Between the eyes of the second and those of the third row there is an unprotected area of about 7°, or remembering that the drawing is considerably enlarged, we may say that an object I cm. sq. will be invisible within the space of these 7° as soon as it is farther than 8 cm. from the spider. The presence within this area of a spider of the same species could be already noticed at a distance of about 25 cm., quite sufficient to protect against sudden onslaught. The posterior side eyes which are second in size, guard the spider at the sides and back. ‘Thus the entire eye-group covers about 253° of the horizon and leaves unprotected a space on top and at the back. In Heteropoda (Fig. 3) the largest eyes are the posterior side eyes The four front eyes cover a field of 145°. Between them and the posterior side eyes there is an unprotected area similar to that in Lycosa, of about 10°. Or since an adult of Heteropoda covers with extended legs about 8 to 10 cm. in each direction, a spider of the same species approaching it within this unprotected area, would become visible at a distance of about 40 cm. The sides are therefore very well guarded especially when we consider that the largest eyes are used in their protection. ‘The eyes of the front row together with the posterior side eyes cover with the inter- ruption mentioned, fully 267°. The dorsal surface of this spider is extraordinarily well protected as compared with the two pre- ceding spiders. There remains an unprotected field of about The Sense of Sight in Spiders 295 10° in front of the posterior middle eyes and about 55° of unpro- tected field at the back behind these same eyes. In the projection on the plane of symmetry (Fig. 7) the eyes cover .152° or in the normal position of the spider, on a wall, 125°, as may be readily understood from the drawing. In the vertical plane (Fig. 8) the eyes of Heteropoda cover a field of 193°. This spider lives in build- ings where it runs along the walls and ceilings hunting insects and other spiders and it is distinctly crablike in motion. The com- paratively large fields of vision in this species are possibly to be accounted for in connection with the habit of the spider to remain quiet during the day and to begin its activity at dusk. But this does not obscure the fact that the sides of this spider are better protected than the front. It was next necessary to ascertain whether or not the fields of vision vary in spiders of the same species at different ages. With this object in view many spiders were examined, always with the same result, 7. ¢., from the time when the eyes assume their permanent position on the head of the spiderling, the maximal angles of vision and the fields covered by these eyes are the same as in the mature female. As to spiderlings taken directly from the cocoon, I am sorry to say that I was unable to make any observations upon them. ‘They are so small and their chitin so soft that it is impossible to prepare them in the manner described and I have not as yet devised another method. But assuming that their maximal angles of vision are the same, which is indeed very probable, we may readily see the advantage in the changes of direction in the eye-axes as I have described them. A glance at the accompanying drawings will make this clear. ‘Vo attain their permanent position the axes of the AME in Phidippus move upward and inward. This slight upward change makes it possible for an image to be formed of an object on the central part of the retina of an adult on the same plane, a good deal farther away than is possible in the eye of the spiderling. We shall see farther on that the central part of the retina is much more sensitive than the periphery. The change inward tends to the same end as the change upward and the final position of the anterior middle eyes in Phidippus allows therefore of a more perfect distinction of 296 Alexander Petrunkevitch objects in front of the spider. In the same eyes in Heteropoda the direction of the axes changes in the opposite sense, 1. ¢., out- ward and this change serves to bring about a better discernment of objects considerably at the side of the median line, while the same eyes still guard the front sufficiently. In the posterior middle eyes of Lycosa, the most sensitive ones in this spider, the direction of the axes changes to one more downward and inward thus serving to protect better the front. The direction of the axes in the posterior side eyes of the same spider changes in such a way that the adult eyes look farther backward. In both PME and PSE this change takes place at the expense of the field protected in the young spiderling, which now becomes relatively exposed. And here the advantage of a slower growth of the eye- area as compared with the growth of the rest of the cephalothorax, becomes evident. Indeed if the eye-group should occupy in the adult spider relatively the same portion of the cephalothorax as it does in the youngest spiderling, the unprotected field would become in consequence of the change in the direction of the axes, so large that the presence within it of an object even larger than a spider of the same species, would remain entirely unnoticed. In Heteropoda the change in the direction of the axes of the PME is in exactly the opposite sense to that in the same eyes in Lycosa and affords more protection to the dorsal surface. At the same time in the axes of the PSE in Heteropoda, the change of direction is in the same sense as in Lycosa and Phidippus. But this change is considerably more marked in Heteropoda with the result that in the adult spider the eyes cover fully 267 degrees instead of the (probable) 166° in the spiderling. This idycncape cannot be gained however, without the formation of an unprotected field. Again, as in Lycosa, this field would have been much larger but for the difference in growth between the eye-group and _ the cephalothorax. THE LIMIT OF VISION If we examine under small magnifying power at once all eight eyes of a cephalothorax, freshly prepared and suspended on a drop of glycerin-albumen as I have described, we shall remark The Sense of Sight In Spiders 297 that the four pairs of eyes form four pairs of images differing from each other in size. As a rule we shall find that the largest eyes form the largest images. The question at once occurs, are all eyes equally sensitive notwithstanding that they form images differing in size, or are the larger eyes more sensitive than the smaller ones? The surest way to find the answer to this question is to determine the minimum angle of vision for each eye. But how are we to do this when we do not even know with sufhicient exactitude the distance at which under normal conditions a spider recognizes another of the same species. ‘The experiments of the Peckhams, in spite of their ingenuity, still admit of too great range for error, to be utilized in a study of the normal auele of vision. Of what advantage, then, would be a similar experiment but with some of the eyes ailectencd with paint? It could serve merely to control another method, a method of comparative morpholog gy. We have to start from the proposition that the physiology of the nervous system is analogous in the other animals and man, a proposition which few are disposed to admit, but here the experience gathered in many fields and from observations made on different animals, comes to our aid, an experience that leaves scarcely any room for doubt that the stimulation of a single nerve-ending transmits to the central nervous system a single sensation only, whether or not the stimulus itself is a simple one or in reality composed of many contemporaneous stimuli. Thus, as 1s well known, in order to perceive two pin-pricks as two dis- tinct sensations, it is necessary that they should be applied to two separate nerve-endings as otherwise the sensation 1s that of a single prick and again, when the i image of two stars falls on only one cone of the retina of the unaided human eye, the eye perceives but a single star. These are well known facts w hich justify us in saying that in the spider’s eye two rods must be stimulated by light rays in order that the image of two points should be produced. But here the analogy ends. How strong the effect produced and whether the corresponding image in the brain is of the same kind as in man, we cannot know. We cannot know whether a spider sees colors as we do, whether green appears to it in the same way as it does to us, although we do know from the experiments of 298 Alexander Petrunkevitch the Peckhams that spiders are able to discriminate between colors. Neither can we know whether gradations of light and shade are the same for the spider as for us nor how great the amplitude of the light wave, which would be required to produce the same effects as in us. Nevertheless we do know that an image is formed in each spider eye; we do know that the four pairs of images differ from each other in size; we do know that the more rods covered by the image the more detail can be perceived by the eye. We may thus work on a fairly safe basis. Let us first examine the images as they appear under the microscope. When a black square 1s placed under the microscope so that the axis of the eye is perpendicular to the center of the square, it is not possible to detect any spherical aberration by common means. But if we place the eye so that the black square lies considerably to one side of the eye-axis, the aberration at once becomes appreciable. For the accompanying drawing (Fig. 11) an eye of Lycosa was put so that it was a little outside the center of the microscopic field and the axis of the eye formed a more or less sharp angle with the slide, while the black square, each side of which was 5 cm. long but having one side prolonged into a straight black line, the whole made with chinese ink on a plate of milk- white glass, was absolutely parallel to the slide. The Zeiss draw- ing table was carefully arranged beforehand so that a small square placed under the microscope and drawn with the aid of the Abbe apparatus, gave on paper a perfect square. Then drawings of the image in the spider’s eye of the black square were made from different positions of the eye, obtained by revolving the table of the microscope on its axis. It 1s clear from the drawing that in this case the base and one side of the square are especially distorted. Is the spider’s vision then distorted of all things that lie out of the axis of the eye? It is impossible to know but I believe that the spider forms a true idea of objects, first because the distortion in each eye of a pair is in the opposite sense to that in the other eye of the same pair, thus offsetting it, and second, because the retina is not a plane but is of very complicated form differing in different eyes and for different species. Generally speaking we may compare The Sense of Sight in Spiders 299 the retina to a boat or canoe in some eyes and to a deep bag in others. This may be ascertained not only from sections. It is sometimes possible to remove the entire retina intact from the vitre- ous body whose proximal end fills it out and to examine it in toto under the microscope. The vitreous body also varies in shape in different eyes and is usually considerably elongated in the direction of the eye-axis in those eyes which form the largest images. The vitreous body is especially long in the anterior middle eyes of jumping spiders, Phidippus tripunctatus, for example, and is shaped something like a long cone with its base which is concave, toward the lens, its axis being at the same time the axis of a conical hole which extends through its entire length. This hole also is largest at the lens and much smaller jat the retina and may be seen in sagittal and cross-sections. It is also sometimes visible in young spiderlings of the jumpers, where it presents a likeness to a pupilla. In life it is probably filled out with a liquid. Roughly speaking the size of the image is in direct proportion to the size of the eye but measurements on discrepancies which must be due to differences in the curvatures of the lenses. The following table illustrates this. TABLE Ix Ratios of diameters of eyes and of images to the diameter of the AME and its image. Compare with Fig. 10 | AME ASE PME PSE { ey Phidippus tripunctatus........... ¢ BG : | 9-595 orree Sone \ image I | 0.4285 0.0800 0.3143 Heteropoda venatoria........ steal he : | i are 1-55 \ image I | 1.75 1.50 2.00 ( a e Tey cosainidicOlaisse cc-.2)= eis .5 ale srs}s sis Q 5 ; S27 See oe \ image I °.8 2.266 1.8666 } This table shows that while there is a dependence of the size of the image upon the size of the eye, this dependence is not of such a kind as to allow of definite conclusions in regard to the smallest angle of vision from measurements of the eyes and 300 Alexander Petrunkevitch images in one individual and of the elements of the retina in another. I have therefore applied two other methods. The one consists of choosing two individuals of the same species, having eyes of the same diameter and preparing one for the study of the image, the other for sections through the retina. This is possible and yields good results when one has a large quantity of living spiders. By far the better method, the one which I now use exclusively and upon which the conclusions I[ have reached in this research are based, consists in preparing the spider in such a way as to obtain from the same individual at the same time the lenses intact for the study of the size of the image and the retina for sectioning. This is perfectly feasible even in young spiderlings although it requires not a little patience and experience. I pro- ceed in the following manner: a spider is killed by a cut across the middle of the cephalothorax so as to allow the fixing liquid to penetrate as rapidly as possible. ‘Thus far I have obtained the best results for the purpose with the picro-formalin mixture of Bouin and my sublimate modification of the Gilson liquid. The cephalothorax is allowed to remain for six hours in the fixing liquid and is then transferred in the usual manner to 70 per cent alcohol. ‘The sternum and the mouth parts are now carefully removed with a fine pair of scissors. ‘Then the cephalothorax is placed in a low dish containing alcohol and held by the side with forceps while a thin and flexible spatula is carefully introduced between hypoderm and chitin. Pushing the spatula slowly for- ward it is possible to separate the entire chitinous part of the cephalothorax from the underlying hypoderm with all its muscles and organs. The vitreous bodies of the eyes remain with the retinas attached by the optic nerves. ‘These are now separated from the remaining organs, carefully noted to exclude error and placed in separate dishes in parafhn in the usual manner. They are then sectioned with a microtom, either parallel to the eye-axis or perpendicular to it. ‘The sections are depigmented in chlorine gas dissolved in 70 per cent alcohol and after washing stained in Heidenhein’s haematoxylin. The eye-lenses of the same individ- ual, which have been removed together with the tergum, in the manner described, are now placed in water and carefully cleaned The Sense of Sight in Spiders 301 with the brush on both sides. ‘They are then cut separately out of the cephalothorax and suspended, each first on a drop of water and then on one of glycerin-albumen and the image which each forms of a 10 cm. square at a distance of 10 cm. is measured with the aid of the ocular micrometer. “These measurements may then be directly compared with those obtained from the sections through the retina. In measuring the distance between the rods I use the highest power only and count how many microns are occupied by ten rods. This is essential since the distances between the rods are apt to vary alittle. Besides, the TABLE X An adult Lycosa nidicola, small individual. Average distance between the centers of the two rods in micromillimeters Toward the At the In the center : P a periphery periphery INN Up sSoooonddoodo oonbudedeauane secnudc 8 12 | 15 TAS Eyer eet tarata neice Neo cr easetartane atane SieyHine 6 10 | 12 1D No So conpescoade bdenbobognaacarooeder 8 12 | 15 1D oncaucacdoccodo gdp copDuOboOmoOUnON 9 16 21 rods are larger and farther apart at the periphery, gradually becom- ing smaller and lying closer together toward the center. I do not give a drawing of this but Table X affords sufficient demonstration. It seems to me, in view of the strange shape of the retina, that we may form an idea of the acuity of vision in the spider’s eye only from images that cover the central part of the retina alone. In order to diminish the possible error, | measured the image of 10 cm. as mentioned, but divided the result by ten so as to find the size of the image of 1 cm. from 10 cm. distance. In this way we obtain the following table: TABLE XI Phidippus tripunctatus AME ASE PME PSE Size of the image of 1 cm. at 10 cm. distance. Eye suspended on glycerin-albumen...............--. Le 115 49 10 36 Distances between the centers of rods in the center | w Ofithe retinal. cee rtatarctssject placed at the same distance from each of the three eyes. In the human eye this image will occupy 57 cones, in the anterior middle eye of Phidippus nearly seven rods, while in the posterior middle eye of Lycosa the image is only a little longer than the diameter of a rod. THE SENSE OF SIGHT IN SPIDERS PLATE VI ALEXANDER PETRUNKEVITCH PHIDIPPUS LYCOSA AME PME : Secrccccocooccsoosoccooeoos i a ea i Tue JourNat or ExperimentAL ZOOLOGY, VOL. V, NO. 2 ay THE PHYSIOLOGY OF THE NERVOUS SYSTEM OF THE RAZOR-SHELL CLAM (ENSIS DIRECTUS, CON.) BY GILMAN A. DREW With One Pirate The razor-shell clam is a particularly favorable lamellibranch for the study of the functions of the ganglia, because: (1) It is very active and responds rapidly to stimuli. (2) Each ganglion supplies nerves to organs that are so active that one can hardly fail to see movements, even when the stimulation is slight. (3) The animal is so narrow that the shell valves can be wedged apart enough to allow all operations and experiments to be performed without removing the animal from its shell. (4) The ganglia with their commissures, connectives and chief nerves, all lie so superficially they can be seen without cutting the animal more than to separate the fused margins of the mantle lobes and the inner lamella of the inner gills, and to expose or cut almost any one of them requires only the cutting of a thin outer covering that cannot cause a mutilation that needs to be taken into account in the results that are obtained. Before discussing the functions of the different ganglia it 1s desirable to study the activities of the animal as a whole and to become acquainted with the responses of the various portions of the body when the organs that are subject to external stimuli are stimulated. The habits of the animal have already been discussed in another paper, but in studying the effect of stimuli it is necessary to know something of the normal life of the animal, and accordingly a brief statement of its habits are desirable here. ‘The animals are best known on mud-flats that are exposed at low tide, but they are 1The habits and movements of the Razor-shell Clam, Ensis directus,Con. Biol. Bul., vol. xii, no. 3, 1907. Tue Journat or ExPeRIMENTAL ZOOLOGY, VOL. V, NO. 3 312 Gilman A. Drew known to occur at moderate depths. Dr. K. Kishinouye writes me that the Japanese “fishermen catch razor-shell clams from the bottom of the sea, ten or more fathoms in depth, by means of slender spears that are weighted at their upper ends and held at the end of a rope.’ Inasmuch as the catch is made by simply pulling up and dropping the spear and is dependent upon acci- dentally striking the clams, they must be fairly abundant to make such a method profitable. “The animal lives embedded in the mud almost perpendicularly, with the siphon end usually barely protruding above the surface. Occasionally specimens are found, when the mud-flat is bare, with half or more of their shells exposed, but, judging from observations of specimens in aquaria and of other specimens in their native mud-flats that had not been dis- turbed and were covered with water, I am inclined to believe that this is not a usual position, and is probably assumed as the result of the stimulation of the heat of the sun. When an animal is disturbed, as by jarring the mud or by stimulating the exposed siphons, it almost instantly disappears into the areal This is evidently its means of escape from enemies. The burrowing is done by means of a remarkably long, active, cylindrical foot, Figs. 1 and 2, 7, that can be protruded from the anterior end of the shell to a distance equal to more than one-half of the length of the shell. When fully extended the end of the foot is swelled to form a knob that serves as an anchor for the animal to draw itself into the mud. ‘The fact that the animal disappears so promptly after it is disturbed indicates that the foot is probably kept somewhat extended when the animal is at rest in its usual position. The margins of the mantle lobes are fused together so that four openings into the mantle chamber are left. “Two of these are the openings of the siphons, Fig. 1, bs and cs, the third is the opening through which the foot is protruded, and the fourth is a small opening about midway on the ventral margin, Fig. 1, vo. What function is performed by the last mentioned opening is not clear. With an expanded animal in a dish of sea-water it is easy to demon- strate that a current of water enters this opening. ‘This is to be expected as the opening leads into the branchial chamber, into Nervous System of the Razor-shell Clam 313 which water is constantly passing, and there is no reason why water should not enter this opening as well as the branchial siphon. When the animal is embedded in the mud however, free admis- sion of water through this opening is not to be expected. ‘The opening is surrounded by well developed tentacles that are similar in appearance to those around the siphons and, like them, very sensitive to tactile stimulations. Stimulation of these tentacles always cause the animal to close its shell and usually, this may be the mechanical effect of suddenly closing the shell, the slight protrusion of the foot. The foot is almost immediately retracted into the shell again and remains retracted unless stimulation 1s continued. When the stimulation is continued the foot is alter- nately protruded slightly and retracted, and occasionally, when the animal is held anterior end downward, burrowing movements are started. On each side of the line of fusion of the mantle lobes are very small papilla that are probably also very sensitive to touch. The whole region is very sensitive but whether sensation is more acute on the papilla than on the general surface was not determined. Posteriorly, from the ventral opening to the branchial siphon, the fused mantle margins are very thick and muscular. Anteriorly, to the opening through which the foot is protruded, the margins are loosely attached by their epithelial cells. The extensive fusion of the margins of the mantle keeps mud out of the mantle cham- ber during burrowing, and forms a device for expelling strong jets of water. Around the opening through which the foot is protruded the margins of the mantle are much enlarged to form muscular, thin-edged scrapers or valves, Figs. 1 and 2, c, that keep mud from being drawn or forced into the shell when the foot is withdrawn and the shell is forced down into the mud. It will be convenient to refer to this portion as the collar. ‘The collar is very sensitive to touch and when stimulated is drawn tightly against the sides of the foot. When the foot is withdrawn it turns in over the end and so closes the shell. Strong stimulation of the collar when in this position, causes the margins to be drawn still further in and thus reflected into the shell. 314 Gilman A. Drew The siphons are the most exposed, and apparently the most sensitive to stimuli of any portion of the mantle. They are sur- rounded by sense tentacles and, in the expanded animal, protrude a short distance beyond the posterior end of the shell. “Tentacles occur all over the branchial siphon and fringe its margin. The cloacal siphon has tentacles around it and on its sides but its edge is very thin and does not bear tentacles. When stimulated the siphons contract and are withdrawn between the posterior bor- ders of the shell valves. As has already been mentioned, the stimu- lation of the siphons of a specimen that is embedded in the mud is the signal for its disappearance. A very slight touch, such as might be given by a drifting weed or a piece of dirt, will cause an instant withdrawal of the siphons but may not cause the animal to burrow. Ifthe stimulation is repeated, burrowing is quite sure to follow promptly. When the animal is removed from the mud, stimulation of the siphons when not long continued simply cause their complete withdrawal and the closing of the shell with the foot retracted. Continued stimulation, especially when accompanied with or preceded by the stimulation of the tenacles around the ventral mantle opening, and with the animal held with the anterior end pointing down, cause the foot to be protruded, swelled at the end and withdrawn in a manner similar to the movements of burrowing. If the stimulations are continued, these movements are usually repeated until a dozen or more complete thrusts and withdrawals have been made. The foot, which is also periodically exposed to external stimuli, is likewise very sensitive. Stimulating its surface causes its with- drawal but it is never thrown into burrowing activity as the result. When the foot is withdrawn, the collar closes in over it, and if stimulation has been more than slight the siphons are retracted and the shell is closed.~ From the foregoing it will be seen that a reasonably strong stimu- lation of any portion of the exposed animal affects it as a whole and may cause either complete retraction into the shell and the contraction of the muscles that close the shell, or may institute movements that are intended for escape into the mud. The latter movements seem never to be caused by the stimulation of either Nervous System of the Razor-shell Clam 315 the foot or the collar, but only by stimulation, usually when in the proper position, of the posterior or ventral mantle region. The habits of the animal are such that these regions are most likely to give warning of the presence of enemies. If instead of applying reasonably strong and repeated stimuli, such as would be caused by stroking or pricking, very light and short stimuli are given, such as may be given by barely touching a tentacle with the side of a needle, a different result may be obtained. With a specimen lying in a dish of sea-water it is possible, by repeated slight stimuli, to cause the siphons or the foot and collar to be withdrawn without visibly affecting other portions. The foot and collar are so intimately associated, touching each other as they do, that it is very hard to cause the retraction of one with- out the other, but it is possible to cause a marked change in one without appreciably changing the other. Before considering reactions further it is desirable to give atten- tion to the nervous system. The three pairs of ganglia that are usually present in lamelli- branchs are all well developed, but there are no other definite ganglia. ‘There seem to be a few scattered ganglion cells about the branchial nerves and a few others in sensitive portions of the mantle, but on the whole the nerves, commissures and connectives are remarkably free from ganglion cells. Although small ganglia are reported to be present on the cerebro-visceral connectives of Solon,” a very closely related form, I find no trace of such ganglia in Ensis, either in the dissections of mature individuals or in the serial sections of individuals about two centimeters long. The cerebral ganglia, Figs. 1 and 2, cg, lie directly ventral to the anterior foot muscles and anterior to the mouth. ‘They are far apart and are connected by a narrow commissure, Fig. 2, cc, in which ganglion cells do not seem to be present. Each cerebral ganglion is joined to the corresponding visceral and pedal ganglion by connectives, Figs. 1 and 2, cve and cpc, in neither of which are ganglion cells abundant. Posteriorly each cerebral ganglion sends a nerve to supply the labial palps of the same side, Fig. 2, 2 Lankester’s A Treatise on Zoology, part 5, Mollusca (Pelseneer); Cambridge Natural History, Mollusca. 370) is Gilman A. Drew lpn. Dorsally and anteriorly a nerve is continued to the corre- sponding anterior foot muscle, Fig. 2, ajn. Anteriorly a large nerve that soon branches starts forward. A portion of the first branch of this nerve bends ventrally to the margin of the mantle lobe and is continued posteriorly as the circum-pallial nerve, Fig. 2, cpn. The remainder of this first branch is continued forward toward the collar. The second branch from this large anterior nerve, aan, supplies the anterior adductor muscle. It is not always given off at exactly the same point in different specimens, and it sometimes happens, as in the case of the specimen shown by Fig. 2, that the origins of the nerves on the two sides are not symmetrical. The examination of serial sections and physiological experiments both indicate that these are the only nerves that supply this large muscle. The remainder of the large anterior nerve is eonamied anteriorly and sends numerous Beemehes to the collar reigon of the mantle. The nerves of the two sides are continuous in front of the anterior adductor muscle so a complete connection between the two cerebral ganglia is formed, just as the circum-pallial nerves connect the cerebral and visceral ganglia of their respective sides. It may be well to state here that, while such anatomical connec- tions undoubtedly exist between these ganglia, repeated experi- ments have failed to show the possibility of sending a nervous impulse from one ganglion to another by either of these connec- tions. Possibly neurones from the two ganglia overlap in their dis- tribution so there may be more pple coordination between portions that work together. Upon cutting the inner lamellz of the inner gills where they are joined together, and pushing them to their respective sides, the visceral ganglia, Fig. 2, vg, are immediately seen. They lie just anterior to the posterior adductor muscle, sometimes, as in the specimen shown in Fig. 2, with their posterior ends overlapping the anterior border of the muscle. ‘The visceral ganglia are closely fused, so there is only a slight constriction between them. The commissural fibers are distinctly visible in sections but ganglion cells cover them entirely. As already indicated each visceral ganglion is joined to the corresponding cerebral ganglion by a connective, Figs. 1 and 2, cvc, that runs along the side of the body, Nervous System of the Razor-shell Clam 27 and by a circum-pallial nerve, cp, that follows the margin of the mantle lobe. There is no indication that sensory impulses ever travel from one ganglion to the other through the circum-pallial nerve. All of the cerebro-visceral association fibers seen to be contained in the cerebro-visceral connectives. Soon after leaving the visceral ganglion each pallial nerve gives rise to a branch that supplies the posterior adductor muscle, Fig. 1, pan. It is then continued posteriorly and ventrally, sends many branches to the siphonal region and then turns anteriorly along the border of the mantle as the circum-pallial nerve, cpm, which joins the cerebral ganglion. What service is performed by this connection is not clear unless it is to afford overlap for the distribution of the motor fibers from the two ganglia to the margin of the mantle. Sensory fibers from the siphons all seem to go to the visceral ganglia, and from the tentacles around the ventral opening in the mantle, to the cerebral ganglia. For the siphons this is easily determined by cutting the pallial nerves between the nerves that supply the siphons and the visceral ganglia, when stimulation of the siphons causes no response, and for the tentacles around the ventral open- ing by cutting the circum-pallial nerves between the tentacles and the cerebral ganglia, after which stimulation of the tentacles causes no response. If the cut is made between the siphons and the ventral tentacles, the effect of stimulating either portion seems entirely normal. “The motor fibers of these nerves are hard to experiment with but it is evident that most of the mantle muscles posterior to the ventral opening are supplied by fibers from the visceral ganglia. Anterior to this opening, along the path of the circum-pallial nerves, the muscles are not very well developed. A branchial nerve, Fig. 2, bn, leaves each visceral ganglion to pass anteriorly and laterally to the united lamellae of the corresponding pair of gills. “The physiology of these nerves has not been studied. Upon stimulation of the isolated visceral ganglia, slight contrac- tions of the posterior foot muscles have been observed, but this may have been caused by escaped current from the electrodes. Nerves from the visceral ganglia to these muscles have not been found. ‘The supply of nerves to the heart and the cardioinhibi- tory action have not been studied. 318 Gilman A. Drew The pedal ganglia lie in the foot, very near the point where its dorsal border joins the visceral mass, Figs. 1 and 2, pg. They are more deeply embedded than are the other ganglia, but parts of them may sometimes be seen when the foot is pressed ventrally and posteriorly. “lo expose them it is necessary to remove the over- lying tissue, which includes a thin layer of muscle. Like the vis- ceral, the pedal ganglia are closely fused and their connecting commissure is covered by ganglion cells. The cerebro-pedal connectives, Fig. 2, cpc, may be seen throughout their extent with- out cutting. Like the other connectives they seem to be free from ganglion cells. ‘he pedal ganglia supply the nerves to the foot. From each ganglion three large nerves pass toward the end of the foot and one or two extend ventrally. From the postero-ventral surface of each ganglion a nerve passes posteriorly and laterally to the side of the foot. The pedal nerves are not easily reached without rather extensive dissection and their individual actions have not been studied. As they are the only nerves to the foot they must contain both motor and sensory fibers. In stimulating the different ganglia directly, it was found that an electric stimulation that could just be distinctly felt by the tongue was most satisfactory as it did not cause mutilation and the results did not give evidence of escaped current. With all commissures and connectives intact, the stimulation of any ganglion visibly affected the whole animal, but the relative time of the contraction’ of different parts differed according to the ganglion stimulated. “Thus when the visceral ganglia were stimu- lated, the siphons responded immediately, the collar and anterior adductor muscle later, and the foot slightly later still. This could be noticed without the use of recording instruments and indicates that an appreciable time is taken in transferring from one set of fibers to another; much longer than in the higher animals. Organs connected directly with the ganglia stimulated always responded first and those that were stimulated through association centers later. The majority of the experiments performed were to determine: (1) The organs that received nerves from each pair of ganglia. (2) Whether each pair of ganglia individually govern the move- Nervous System of the Razor-shell Clam 319 ments of the organs it supplies with nerves, or whether some of the ganglion are accessory and dependent. (3) Whether all connec- tives carry impulses in both directions. (4) How far it was possi- ble to have impulses transferred through association centers that would not normally be concerned with the impulses in uninjured animals. Stimulation of the ganglia directly and of the nerves that leave the ganglia, and stimulation of sensory areas and the nerves that supply the sensory areas were all used. In discussing the various nerves that leave the ganglia, mention has been made of the organs they supply, so it is only necessary to summarize here. When the cerebral ganglia are separated from the other ganglia by cutting the cerebro-pedal and cerebro-vis- ceral connectives and the circum-pallial nerves,* stimulation of the ganglia causes contraction of the anterior adductor muscle, the anterior foot muscles, the collar and at least a portion of the margin of the mantle. The functions of the nerves to the labial palps were not determined. When the visceral ganglia are separated from the others by cutting the cerebro-visceral connectives and the circum-pallial nerves, stimulation of the ganglia causes contraction of the siphons and of at least a portion of the mantle margins, and feeble contrac- tions of the posterior adductor muscle. Slight contractions of the posterior foot muscles have also been observed, but as the ganglia lie very near them and no nerves have been found enter- ing them from these ganglia, it seems probable that the slight observed contractions were due to escaped current. ‘The pos- terior adductor muscle when stimlated directly did not contract more than when the ganglia were stimulated. It seems that it does not function much in closing the shell. Its office is satisfac- torily filled by the thickened, united, posterior margins of the man- tle lobes. When the pedal ganglia are separated from the others by cut- 5 Although the circum-pallial nerves did not seem to be able to carry impulses from one ganglion to the other, they were cut in these experiments to make sure that the ganglia were isolated. They were cut far from the ganglia that were to be experimented upon, so the effect of their motor fibers could be determined. 320 Gilman A. Drew ting the cerebro-pedal connectives, stimulation of the ganglia causes vigorous contractions of the whole foot, including the ante- rior and posterior foot muscles. It was not determined whether the foot muscles were affected throughout or only in part. They become so intimately connected with the general musculature of the foot that a complete contraction of the foot would necessarily involve them. The above experiments show the organs that are supplied with motor nerves from each pair of ganglia, but they do not indicate whether the contraction was in each case caused by stimulating cells in the ganglia themselves or by stimulating fibers that might be passing through them from other ganglia. By isolating the different ganglia and stimulating sensory areas connected with them, motor cells can be proved to be present in the cerebral and visceral ganglia. Stimulating the sensory surfaces was generally accomplished by stroking with the point of a seeker or pencil. If muscular organs connected with the ganglia con- tracted upon such stimulation, motor cells must be present in the ganglia. After cutting the cerebro-visceral and cerebro-pedal connectives and the circum-pallial nerves (the latter near the siphons), strok- ing the tentacles around the ventral mantle opening caused con- traction of the anterior adductor muscle, and both sides of the collar. After separating the visceral ganglia from the others by cutting the cerebro-visceral connectives and the circum-pallial nerves (the latter near the cerebral ganglia), stroking the siphons apparently caused slight contractions of the posterior adductor muscle and strong contractions of both sides of the posterior mar- gins of the mantle. Stimulation of one of the pallial nerves elec- trically, caused the siphons, posterior adductor and both of the posterior mantle margins to contract. “These experiments indicate that sensory cells end in both the cerebral and the visceral ganglia and that stimulating these fibers causes disturbances in the motor cells in the same ganglia that cause the contractions mentioned. With the pedal ganglia results are not so easily obtained. When these ganglia are separated from the others by cutting the cerebro- pedal connectives, the foot immediately loses its rigidity, and any Nervous System of the Raz cor-shell Clam 221 amount of stimulation of the surface of the foot, electrically, chem- ically or mechanically results only in the contraction of muscle fibers in the immediate vicinity of the point of stimulation. The foot never makes a movement as a whole and will remain motion- less for hours, probably until it dies. ‘This seems to mean either that there are no motor cells in the ganglia or that the sensory fibers have no endings or collaterals in the pedal ganglia but are continued directly through these ganglia to the cerebral ganglia. I am inclined to believe that motor cells are present in the pedal ganglia and that the sensory fibers pass directly through them without endings or collaterals, for the following reasons: (1) Microscopically the ganglia show an abundance of ganglion cells and it seems more reasonable to believe that, in such a muscular organ, they are not all sensory, especially as the action of sensory cells so placed, if motor are not present, would have to be referred to the cerebral ganglia before movement could be effected. (2) When the cere- bro-pedal connectives are cut the foot responds with contractions. These have the character of tetanic contractions that would more probably come from the action of disturbed nerve cells than from the single stimulus caused by cutting motor fibers. If such move- ments could be caused by the stimulation due to cutting fibers only, then the cutting of the pedal nerves (below the pedal ganglia) should cause them, but beside the single twitch caused at the instant of cutting no movements follow this operation. (3) If one of the cut cerebro-pedal connectives is stimulated, the foot as a whole, both sides, responds, apparently with a complete, normal contraction. ‘The course of the fibers in the ganglia have not been traced, but the effect is not what we would expect if the action is the result of the stimulation of only half of the motor fibers that go to the foot. It is much more easily explained by supposing that impulses have been sent to association cells which cause the motor cells of the foot, contained in the pedal ganglia, to act. Stimula- tion of the nerves that leave one of the pedal ganglia, after the pedal ganglia have been removed, does not cause complete contrac- tion of the whole foot, as it should if the ganglia themselves have had no effect. Whatever the arrangement, there can be no question that the 322 Gilman A. Drew pedal ganglia are deficient in originating power, and that when the pedal are separated from the cerebral ganglia the foot will not by itself, or as the result of surface stimulation, execute movements more than are to be accounted for by the local direct stimulation of muscle fibers. “To make sure that the movements did not come from stimulating the ganglia, they were entirely removed and still sumulation of the surface of the foot gave exactly similar contrac- tions. From the foregoing experiments it would seem that both cerebral and visceral ganglia are able to receive impulses and to direct the movements of certain organs with which they are connected, when they are entirely separated from the other ganglia, and that the pedal cannot act by themselves. ‘This is somewhat surprising but possibly the habits of the animal may account for it. Appar- ently the cerebral ganglia are central for the nervous system. This is indicated by their connections with the other ganglia as well as by experiments. “They would then have charge of the special activities of the whole animal, as well as of the special organs in their immediate vicinity which they supply with nerves. The visceral ganglia govern over organs that are in constant activity, . organs that must give warning of the approach of enemies. They must give warning to the cerebral ganglia and then be ready to cover the retreat by closing and withdrawing the siphons and con- tracting the posterior margins of the mantle and posterior adduc- tor muscle. The cerebral ganglia may now take charge of the advance with the aid of the efficient accessory pedal ganglia. They have little more to do during periods of burrowing. Dur- ing the life of the animal the foot is not in a position that it will be called upon to give such an alarm very often if ever, and dur- ing burrowing the cerebral ganglia can devote nearly their whole attention to the process. It is desirable in directing a retreat of this kind to have a general in charge that is in constant communica- tion with outposts that may give information regarding the enemy. The cerebral ganglia have such communications; the pedal gan- glia only indirectly. The cerebral ganglia are the only pair that are far enough apart to allow the cutting of the connecting commissure without injuring Nervous System of the Razor-shell Clam 323 the ganglia themselves. With these ganglia the operation Is very simple as they are connected by a a long narrow commissure that is distinctly visible throughout its lenge, With the commissure cut it was found that certain activities were delayed or otherwise interfered with. Stimulating either ganglion apparently caused complete contraction of the foot. The same result was also obtained when the cerebral ganglion that was not to be stimulated was removed previous to the experiment, or, as already noted, if both cerebro-pedal connectives were cut and the cut end of one of them was stimulated. These experiments again show that to cause the action of the foot it is only necessary to stimulate one pedal ganglion, which sets up the necessary association impulses. If both cerebral ganglia had been left connected with the pedal or visceral ganglia, although separated from each other, it would have been possible that stimulation of one resulted in the stimula- tion of the other through the pedal or visceral ganglia. Similar experiments were tried to determine the effect upon the visceral ganglia when one cerebral was stimulated after the cerebral com- missure had been cut. The siphons and posterior mantle margins of both sides always contracted completely, even though the un- stimulated cerebral ganglion was separated by cutting connectives, or was removed. ‘The results were thus similar to those obtained for the pedal ganglia. With the cerebral commissure cut and the two sides of the collar separated to guard against a possible transfer of impulses through the anterjor pallial nerves, although no evidence could be found that such transfer of impulses could be made when the nerves were intact, moderate stimulation of one side of the collar caused only the contraction of the same side of the collar with imperfect con- tractions of the anterior adductor muscle and possibly slight con- tractions of the anterior foot muscle of the same side, with the usual retraction of the foot siphons, etc. Strong and continued stimulation however caused contraction of the other side of the collar as well. Evidently the impulses that affected the cere- bral ganglion that has control of this side of the collar, must have passed by way of either the pedal or the visceral ganglia. Experi- ment indicated that the impulses can be transmitted either way but 324 Gilman A. Drew that the response is quicker and more marked by way of the pedal than by way of the visceral. Much the same results were obtained by stimulating the tentacles on one side of the ventral mantle opening after the cerebral commissure was cut, as were obtained by stimulating one side of the collar. Other experiments were tried to determine to what extent impulses can be made to travel over association fibers in other than what would seem to be the usual ways. It was found that stimu- lating one ganglion of a pair readily affected to its fellow gan- glion and that the disturbance could be readily passed on from this ganglion to others provided the transfer was of a nature that was probably usual. For example, if the right cerebro-vis- ceral connective was cut and the right posterior pallial nerve was stimulated, all of the organs connected directly with the visceral ganglia on both sides responded, and, a little later, the organs con- nected with both sides of the cerebral ganglia and the foot re- sponded. If the left cerebro-pedal connective of the same speci- men is also cut and the same stimulation is given, the response of the foot is delayed slightly but not long. In the last case it has been necessary to send impulses from the right to the left visceral ganglion, from the left visceral ganglion to the left cere- bral ganglion, from the left cerebral ganglion to the right cerebral ganglion, and from this to the right pedal, which in turn must stir the left pedal to action with it. It will be noticed that all transfers in this experiment are in directions that may be sup- posed to be usual, either from one ganglion to another of a pair, or by way of warning from the visceral to the cerebral, or from the cerebral tothe foot. These impulses are sent so readily that in one case it was found that by stroking the siphons of a specimen that had been operated on in the manner described, regular burrowing movements were instituted. Stimulation of the ten- tacles on the left side of the ventral mantle opening gave results almost as quickly as on uninjured specimens. Here again the impulses from one ganglion to another are in usual directions. On other specimens, when the left cerebro-pedal connective was cut and the left side of the collar was stimulated, the foot responded without delay. Impulses were moving in usual direc- Nervous System of the Razor-shell Clam 325 tions. If now the cerebral commissure is also cut and the left side of the collar is then stimulated, or for that matter if the left cerebral ganglion is stimulated directly, the foot responds with convulsive contractions only after considerable delay. In some cases no response could be obtained. In this experiment im- pulses must be passed from the left cerebral around through the visceral to the other cerebral and from this to the pedal before the foot was stimulated. “The path cannot be considered usual and the action is both delayed and modified. Itis interesting to find that the centers are able to respond at all in this roundabout and unusual way. SUMMARY 1 This form is very satisfactory for experimental study of the physiology of the nervous system because of its shape and activity, and the ease with which its nervous system may be seen and oper- ated upon. 2 Continued stimulation of any portion of the body will in time have its effect on all of the ganglia. 3 Certain organs like the siphons, collar and foot, may be so gently stimulated as to cause them to be withdrawn without dis- turbing organs that receive their nerves from other ganglia. 4 The relation of ganglia of a pair is quite intimate. Stimu- lating nerves connected with one causes organs connected with both to respond promptly. 5 Association fibers by which ganglia communicate with each other are found only in commissures and connectives. Although the anterior pallial nerves are united so that a connection 1s formed between the cerebral ganglia, and the circum-pallial nerves connect the cerebral and visceral ganglia of corresponding sides, there 1s no evidence that the ganglia are able to communicate through them. 6 Both cerebral and visceral ganglia are provided with sensory and motor cells. The pedal ganglia are apparently dependent upon the cerebral for initiative. When the pedal ganglia are isolated from the others, stimulation of the surface of the foot causes only local responses due to the direct stimulation of muscle fibers. It would seem that the sensory neurones have neither end- 326 Gilman A. Drew ings nor collaterals in the pedal ganglia but are continued to the cerebral ganglia. 7 Impulses may pass in both directions through any of the commissures and connectives. - 8 Stimulation may cause impulses to be sent by roundabout connections when the usual connections are destroyed, but the stimulation must be of considerable duration and the result is often considerably delayed. University of Maine January 7, 1908 EXPLANATION OF PLATE The anatomy of Ensis directus, Con. Fig. 1 A specimen as seen from the right side with both valves of the shell, the right lobe of the mantle, the right labial palps and the right gills removed. Represented with the siphons and collar extended and the foot slightly protruded. Avery common position. Drawn from the study of dissec- tions and serial sections and enlarged to about one and one-third natural size. Fig. 2 A specimen as seen from the ventral surface with the mantle margins cut and the shell valves wide apart, and the foot forced posteriorly and to the right side of the animal. The inner lamellz of the inner gills have been separated and pushed to the sides. The ganglia and nerves in this figure have unintentionally been made alittle too large. Drawn from dissections with a few details added from the study of serial sections. Enlarged to about one and one-half natural size. aa anterior adductor muscle h heart aan anterior adductor muscle nerve /p labial palp af anterior foot muscle Ipn labial palp nerve afn anterior foot muscle nerve m mouth apn anterior pallial nerve bn branchial nerve posterior adductor muscle posterior adductor muscle nerve bs branchial siphon pfm posterior foot muscle e collar pg pedal ganglion ce cerebral commissure pn pedal nerves cg cerebral ganglion ppn posterior pallial nerve cpe cerebro-pedal connective r rectum cpn circum-pallial nerve s crystalline style cs cloacal siphon st stomach cve cerebro-visceral connective vg visceral ganglion f foot vm visceral mass g gill vo ventral mantle opening NERVOUS SYSTEM OF THE RAZOR-SHELL CLAM PLATE I Giiman A. Drew Tue Journar or ExperiMeNTAL ZoULoGyY, VOL. Vv, NO. 3 —s7~ THE INFLUENCE OF GRAFTING ON THE POLARITY OF TUBULARIA BY FLORENCE PEEBLES With Twenty-Six Ficures The experiments made by Loeb, Driesch, Morgan, and others, have demonstrated that by closing the oral end of a piece of the stem of Tubularia the development of the aboral hydranth is hastened. The same result was obtained by Morgan (’03) when he bent long pieces in the middle, or ligatured them so that the ccenosare of the two ends was completely separated. Morgan and Stevens (’04) have shown further, that the formation of a hydranth at the aboral end of a piece produces a change in that region, so that when this hydranth is removed, the piece is more likely than before to develop another aboral hydranth. The object of the experiments described in this paper was primarily to determine what influence grafting exerts upon the polarity of Tubularia mesembryanthemum, but one experiment led to another until the investigations extended to a study of some of the factors of regulation. Last spring it was my privilege, through the generosity of the “Association for Maintaining the American Woman’s Table at Naples,” to spend two months at the Zodlogical Station, during which time I carried on the experiments described in the following pages. It gives me great pleasure to express here my gratitude to the Association, and also to Prof. Anton Dohrn, and the other members of the staff at Naples, for the courtesies extended to me during my stay. In earlier experiments in grafting (00 and ’o2) two compo- nents having the same diameter were selected, the two cut sur- faces were applied, and held together until the ccenosarc united. This method proved so tedious that a new one was adopted. In Tue JouRNAL or ExPERIMENTAL ZOOLOGY, VOL. V, NO. 3 328 Florence Peebles order to use this method one component must be slightly smaller than the other so that one end of the smaller one may be inserted in an end of the larger one. The two components were usually telescoped so that they lapped about one millimeter. I TWO LONG PIECES GRAFTED TOGETHER IN THE SAME DIRECTION Experiment rt. The first series of experiments consisted 1n graft- ing together two pieces from the same region of two different individuals, so that the aboral end of one piece was inserted in the oral end of another piece from which the hydranth had just been NOS -s x “ A a Fic. 1 Fic. 2 Fic.4 Fic. 6 Fic. 8 removed. Each piece measured about 3 cm. (Fig. 1), not includ- ing the hydranth of the distal piece which was not removed until the day after the graft was made. ‘The first cut (1) removed the old hydranth and about 4 mm. of the stem from the distal com- ponent, the second cut (2) was made through the proximal com- ponent a short distance back of the line of union, and the last cut (3) removed the basal end from the proximal component. ‘These three pieces (Figs. 1 and 2) I shall call respectively 4, B and C. Their individual behavior after this second operation will first be considered, and then they will be compared in order to see the relative rate of development of the hydranths. Grafting of Tubularia 329 (4) The piece designated 4 (Fig. 2) consisted of the major part of the distal component, with a short distal piece of the prox- imal component grafted in the same direction, on its aboral end. If the long and short piece (Fig. 2, 4) act as one, we should expect a new hydranth to form first at the oral end (Fig. 2, X) and later at the aboral end (Fig. 2, Y). If, however, the two components retain their individuality this result would not follow, for the oral end of the small component and the aboral end of the large com- ponent have had a start of twenty-four hours. Forty-seven grafts were made, the results from these are given below in the table. TABLE: A Hy dranthsratexehrets later atid eon erypateyeteleraatelerst-etcheievereicte lauetstaiescyatesstele\sfolere sfoieleverststsiavets/avers 18 Hydranths ates noneylater atl idleiamclalelnccsies cies sVersiclerniets yore 4: Hydranths simultaneously at ¥ and Y 6 My dranthsirstia tele lateratexe sete tsyefovcteteleteterareiteralatelerete ialeleveieist ersierereletevelinictelatereveisistelelst= 6 Eby dramthe\atite One ya te atarale ata alela nisletesoreierstete) > w 3 3 2 2 UETLEULELITLAT| reduced, the double pieces resembling a single one in their behav- ior. Pieces of the same length as 4 and C were cut from corre- sponding regions of stems which had not been grafted. A com- 2 |2 |2 2 Q 1 i 1 PRS ele Me TABLE 8 2a | es Lm {pt x is 2 2 2 2 1 1 1 1 r rx B | 1 1 rn Ss 1 1 3 3 1 1 1 2 c aA 1 4 2 il 2 i parison of their behavior with that of these pieces cut from the grafted stems shows a marked difference in the number of hydranths 336 Florence Peebles developed. A much larger per cent of aboral hydranths formed on 4 inthe control, and also more oral hydranths on C in the con- trol. This seems to indicate that the process of hydranth forma- tion at the grafted ends of these pieces affects their later behavior. Table 7 and Table 8 show very definitely, that the results after the second operation are not modified by those following the first unless the period between the two operations is sufficiently long for them to get a start. In the first place (Table 7) the number of oral hydranths on the distal half of the original distal component, is much larger than that on the proximal half, while there are more aboral hydranths on the proximal half than there are on the distal half. ‘This is not the case (Table 8) when the second opera- tion follows the first after a very brief period. Secondly, the number of aboral hydranths on the proximal end of the original proximal component is much larger than the number on the distal half of the same component, and the number of oral hydranths on its distal half is greater than on the proximal half. ‘This is not the case when the time between the first and second operations is reduced. Experiment 4. In the fourth series of experiments the two components were grafted together in the way described above, but this time the level of the second cut was changed, making the piece 4 (Fig. 10) consist of the greater part of the distal compo- nent while B (Fig. 10) was made up of a short basal piece of the distal component grafted on the oral end of the proximal compo- nent. The third cut (Fig. 10, 3) was made nearer the aboral end of the proximal component thus making C shorter than in the preceding experiment. A series of sixteen grafts of this descrip- tion are represented in Table 9. If we compare the rate of development of hydranths on 4,B, and C we find oral hydranths on 4 and C appearing at about the same time, also those on the oral and aboral ends of B. The per cent of hydranths on the oral and aboral ends of 4 and C was about the same as that in Experiments 1 and 3. The proportion of double hydranths at the region of the graft was larger, and also the num- ber of aboral hydranths on the distal end of the proximal compo- nent. When this experiment is modified by decreasing the time Grafting of Tubularia 337 between the first and second operations (Table 10) no hydranths formed on the oral end of the proximal component in piece B, but a large number of the distal short pieces formed hydranths. A TABLE 9 Soa eet tame cee enaemn ea Peale a ate. teeta 1 4 b 2 |2 4 la le ae 2 2 2 P 2 2 2 2. fo) Lar dite a rh bE |e op a pe IRS ree Hh very small number of hydranths developed on the aboral ends of A, in this series not one. The oral hydranths on A and C appeared about the same time. TABLE 10 i 1 1 1 1 1 1 i 1 ly A {1 Ss 1 A il [ * Me + + + + | + |B 2 2 2 ‘1 2 2 1 3 eat eit in 1 o 2 1 1 338 Florence Peebles 2 TWO LONG PIECES GRAFTED TOGETHER BY THEIR ORAL ENDS Experiment 1. Two pieces of stem, each measuring about 3 cm., were grafted together by their oral ends (Fig. 11). The grafts were left undisturbed for twenty-four hours, then the double piece was cut at two levels (Fig. 11, 7 and 2) so that each compo- nent was halved. The distal halves of each were united by their oral ends forming the piece B while the proximal halves of each (4 and C) had their oral ends exposed by a fresh cut, their aboral ends having a start of twenty-four hours. New oral hydranths formed on 4 and C at practically the same time, the aboral usually preceding the oral by a few hours, or forming simultaneously with it. The piece B developed a very small proportion of hydranths at its free aboral ends, but in nearly every graft double heads formed, one on the oral end of each component; these emerged and finally pulled apart. These oral hydranths were much slower in developing than the oral hydranths on the free ends of 4 and C. When the second operation followed a few hours after the first, the percentage of hydranths formed at the line of union was greatly reduced. ‘The free aboral ends rarely developed hydranths at the same time, one usually preceded the other by six or eight hours. The oral hydranths on 4 and C in this experiment formed before the aboral hydranths with very few exceptions. Experiment 2. In a second series of experiments in which the two components were grafted together by their oral ends (Fig. 12), the level of the first cut was changed so that 4 consisted of the major part of one component, while in B instead of the components being equal in length, one was much longer than the other (Fig. 12, A and B). A period of eighteen to twenty-four hours elapsed between the first and second operations. Table rr gives the results from eighteen grafts. These practically represent the en- tire series of experiments so that it is unnecessary to give other tables. If we consider the rate of appearance of the hydranths we see that the percentage of aboral hydranths is very large, and that they appear before the oral hydranths of the same piece with few exceptions. [he number of hydranths on the oral end of the Grafting of Tubularia 339 large component of the graft is greatly reduced, while a relatively large number form on the aboral ends of the smaller component. The pieces 4 and C, upon which the aboral hydranths appeared before the oral, were kept until the first hydranths were lost, and new ones developed. Without exception, the second set of oral hydranths appeared first, even in cases where the aboral hydranths had developed one to two days earlier than the oral ones. This shows without doubt that the polarity is not permanently reversed. Experiment 3. In this experiment the second operation fol- lowed the first after six hours The results show that the aboral ends had not had a sufhcient start to produce hydranths before TABLE 11 hy ey ee} 22S) A Les eet ale ol a 2) 2) (2 2 (2) c Dy te ay ie ee re TE py be Ee Tp the oral ends. None of the grafts formed hydranths on their oral ends, instead, the majority developed a hydranth first at the aboral end of the short piece, and several hours later at the aboral end of the longer piece. 3 TWO LONG PIECES GRAFTED TOGETHER BY THEIR ABORAL ENDS It has been demonstrated by many observers that a piece cut from the stem of Tubularia, if sufficiently long, shows marked polarity, 1. e., a hydranth develops first on the oral end of the stem and then later on the aboral end. I have shown that when two 340 Florence Peebles long pieces are grafted together in the same direction, they may act as one piece, a hydranth forming first at X and later at Y (Fig. 2) but a large per cent form hydranths simultaneously at the oral ends of each component (Fig. 4). The question naturally arises as to the result if two pieces are grafted together by their aboral ends so that the two oral ends are exposed. Experiment r. ‘Two pieces each measuring about 3 cm. with the hydranths still attached, were grafted together by their aboral ends (Fig. 13). At the end of twenty-four hours the hy- dranths were removed, together with 1 cm. of the stem. The graft (4) consisted of two components of equal lengths whose aboral ends had been united for twenty-four hours and whose oral ends were exposed at the same moment by a fresh cut. This experiment was repeated, with modifications, several hundred times. ‘The results can not be combined in tables without much repetition, therefore I shall merely give one representative series. TABLE 12 ria bh 1 ff ft Pi. ee f(t 14 j2 j2 jd fl 2 bt 2 |2 1 2 |2 1 1 n it 1 1 aL In this series no hydranths formed at the graft line. In most cases the two components acted as one piece forming a hy- dranth at one end only, or first on one end, and then on the other. Whether stolons would have developed later, I can not say, for the pieces were kept only four or five days. At the end of that time although two sets of hydranths developed on some pieces there was no sign of stolon-formation or pushing apart at the aboral ends. That one piece is influenced by the other seems evident from the results. Only five grafts formed hydranths on each end at the same time, while ten produced one first at one oral end, and then after one to two days, at the other oral end. Ten more formed hydranths at one of the oral ends and nothing at the other. Experiment 2. Ina second experiment one component was cut close to the line of graft (Fig. 14,2). In this way, it was thought, Grafting of Tubularia 341 that the long piece might exert some influence on the short one. The results from one series of twenty-seven grafts is shown in Table 13. Nine of the long pieces formed hydranths on the oral ends first, while this took place only twice in the shogt pieces. Nine of the long pieces formed hydranths while no new hydranths appeared on the short ones. Experiment 3. In order to determine if the distance from the original hydranth had any influence on the rate of development, a set of experiments was made where the pieces were grafted so that when cut the oral end of one component should be much nearer the original hydranth than the oral end of the other (Fig. 15). The results showed no difference in the time of develop- ment of the oral hydranths on the two components. TABLE 13 fPppeP 2 |2 \2 1 2 2 2/8 2 2 Yh ah DUTT teil Tee tt oitte ne oe deta Koike BE hit a oh Experiment 4. In the fourth experiment the pieces forming the graft were cut off at equal lengths, very close to the line of union (Fig. 16), each piece measuring 2 to 4 mm. With scarcely any exceptions (about three out of forty), the hydranths formed first on the oral end of the inner piece ( Fig. 16, X) and if one formed at all on the outer piece (1°) it appeared at least one day later. The results of these four series of experiments seem to me to be of peculiar interest, and they are not without weight in consider- ing the problem of polarity. Why should a compound (grafted) individual with two oral ends exposed at the same time, in a large majority of cases develop a hydranth first on one oral end, and and then on the other? Shall we say that there is such a thing as polarity in this new double individual, or shall we say that all the “hydranth-forming material” has been carried to one end so that development at the other was delayed? It is evident from experiments that I shall describe later, that the direction of the 342 Florence Peebles current has nothing to do with the order of appearance of the hydranths on the ends of the grafts. We must seek an explana- tion elsewhere. I believe that it requires a large amount of energy to construct a new hydranth. In order to produce sufficient energy certain metabolic processes are set up. ‘These processes must begin as soon as the wound closes. If the condition of the stem is such that sufficient energy can be produced, hydranths are formed at once, if not, the development is delayed until there is enough energy. When two pieces are grafted together, some of this energy is expended in healing the wound, and uniting the pieces. If there is a large enough quantity left over, or already in the pieces, hydranths develop at once at the two oral ends, but if there is not a large enough amount present, one end is delayed until the hydranth has been completed. This hypothesis may serve to explain the hastening of the aboral hydranth after one has been formed or is about to form in that region. If the hydranth has formed there may be energy left over, if it is about to develop there is a large amount of energy present. Under normal condi- tions some stems contain more energy, or vitality. The preceding tables show that from the two original components a large number of hydranths may develop as many as eight, while from others only one or sometimes none, appear. ‘The conditions of the experiment are apparently the same; the results can be explained in no other way than that one individual possesses more energy than another. I do not believe that there is any one material whose presence modifies the result, it is the state of all the materials at the time of the operation. 3. GRAFTING A SHORT DISTAL PIECE ON THE BASAL END OF THE SAME STEM In an earlier paper (00) I described a series of experiments in which a short distal piece of the stem was grafted in a reverse direction, on the proximal or aboral end of a long piece. The results which I obtained from a small number of experiments, seemed to indicate that the long piece influenced the rate of develop- ment of a new hydranth on the short piece, for the tentacle ridges on the short piece did not appear until after the hydranths had Grafting of Tubularia 343 emerged from the oral end of the longer piece, a region which was nearer the basal end than the short piece. “The number of experi- ments was almost too small to warrant any definite conclusion. I have repeated this experiment a number of times and have finally come to the conclusion that the major component does not in- fluence the minor one unless it shares in the formation of the hydranth which develops at that end. LE ht Fic 17 Fie. 18 Experiment 1. A short piece measuring about 1.5 mm. was cut from the distal end of a long (3 cm.) piece from a region near the hydranth. This small piece was then grafted on the aboral end of the long piece (Fig. 17) so that the oral end of each piece was exposed, the aboral ends united. ‘Twenty-seven grafts are represented in Table 14. The dotted line shows the original position of the small piece. TABLE 14 a ee a AT De PIO retells Meg's SU Olea aa Tate) SO Tg eo Ne a(t a re lee he ee Ooty Bie Te ea re fT ott} seri Seer eae) aL eer) | Tie L TAU Tete EME UTE) pies al ETE | PT TRG PE PPC Mb taTFE Mb U rk. iby 1 1 Lie le 2 le 2 2 HT 2 le be lt 2 es Eleven of the twenty-seven grafts formed a hydranth first at the oral end of the major component (P), and later at the oral end 344 Florence Peebles of the minor component (D). Eleven developed oral hydranths on the long piece and nothing on the shorter one. “T'wo developed hydranths on the short piece only, and two on the aboral end of the long piece only. In one case new hydranths appeared simul- taneously on the two exposed ends. In all grafts where a long and a short piece are united, the formation of a new hydranth 1s always slower in the short piece. In six of the eleven grafts that developed the hydranth first at the oral end of the major compo- nent, the hydranth that formed later on the other end came partly from the long piece and partly from the short one, i. e., the distal tentacles developed in the minor component, and the proximal in the major component. In these the development was always slower. 4 GRAFTING A SHORT BASAL PIECE ON THE DISTAL END OF THE SAME STEM In order to test the influence of a long distal piece on a short basal one, a second series of experiments was made. ‘This time the hydranth was cut from a piece of stem measuring 3 to 4 cm. From the basal end of this piece a short piece (2 to 3 mm.) was cut off, and grafted in the opposite direction, on the oral end of the same stem (Fig. 18). The results from these experiments were not surprising, Here no influence seemed to be exerted by the major component, as the following table shows. TABLE 15 Waite le Peale ;; . 1 fa 1 D 1 2 i Lie 1 No very definite conclusions can be drawn from this table, or from any of the other series in this experiment. ‘The major com- ponent apparently took no part in the formation of the hydranth in the smaller piece. From constant observation of the behavior of grafts composed of a short and a long piece, I am inclined to believe that the size of the short piece has more to do with the rate Grafting of Tubularia 345 of regeneration than contact with the major component. The only cases where it seems to me we are justified in looking for the influence of one upon the other is where the hydranth is developed partly in the long piece and partly in the short one. This did not take place in any of the experiments. § THE INFLUENCE OF THE CURRENTS ON REGENERATION AFTER GRAFTING Experiment 1. Two short pieces grafted in the same direction. It was suggested to me by Professor Morgan, that the current in the two pieces of which a graft is composed, may have something to do with the order of regeneration in the outer ends of grafts. Soon after a piece is cut from the stem of Tubularia, the wound closes, and rapid circulation begins. The current is easily seen coursing up one side of the piece and down the other. When two pieces are united the currents do not always flow from one piece into the other. Instead, the current may be seen flowing up one piece and turning at the line of union as if stopped by a membrane, and continuing down the other side. Frequently, however, the current continues to flow up one side and on into the other piece. In Fig. 19 (E and F) these two cases are shown. In £ the current is continuous; in F it is separate in the two pieces. Small pieces, measuring 1.5 to 2 mm. in length, were cut from different stems from a region at least 1 cm. back of the hydranth. One piece was inserted in the other so that the oral end of one overlapped the aboral end of the other (Fig. 19, 4). At the end of twenty-four hours each graft was carefully examined under sufh- cient magnification to detect the direction of the currents. All of the grafts in which the circulation was continuous from one com- ponent to the other (Fig. 19, £) were put in lot 4, while those in 346 Florence Peebles which the circulation of one piece was distinctly separate from that of the other (Fig. 19, Ff) were isolated in lot B. Another lot (C) consisted of those in which the circulation was irregular, and the last (D) in which no circulation whatever could be detected. The rate of development of the hydranths is shown in Table 16. TABLE 16 ae B {0} D he 1 1 ji 1 Loni tel rib a |e) | 2 2 2 2 The behavior of the grafts is practically the same whether the circulation is continuous or not. In each case a hydranth devel- oped first on one piece and then on the other or on one piece only. Experiment 2. Two short pieces grafted together by their oral ends. Pieces of the same length as those in the preceding experi- ment were grafted together by their oral ends, so that the aboral ends were exposed. They were separated as before into lots 4, B,C and D. The results from one series are shown in Table 17- TABLE 17 A B c D \ fled 1 Oe iti het bie tt Use lls 1 2 1 2 |2 Of these thirteen grafts four developed simultaneously on the aboral ends. Three formed one first on one end, then on the | other, five formed one on one end and none on the other, and one developed a hydranth on one end and a stolon on the other. A control experiment was made in which single pieces, the same length as the entire graft, were cut. Out of eleven pieces, eight formed a hydranth on one end and nothing on the other; two developed a hydranth first on one end, and then on the other, and one devel- oped a stolon on one end, and a hydranth on the other. Experiment 3. Two short pieces grafted together by their aboral ends. ‘These pieces were the same length, and taken from the same region of the stem, but were grafted by their aboral ends Grafting of Tubularia 347 (Fig. 19,C) so that the oral surfaces were exposed. A series of twenty-four grafts is represented in Table 18. TABLE 18 A B if} 1 TY rT ee {le suleche. te 2 Out of twenty-four grafts, twelve developed a hydranth first on one end, then on the other in spite of the fact that in some the current was continuous in the two pieces and not in others. Only one developed hydranths simultaneously on the free ends. Eleven formed a hydranth on one end only. These experiments, as a whole, show that double pieces usually form one hydranth only (Fig. 19, D), or first one and then another later on the other end, regardless of the direction of the graft, or the flow of the currents. c D Lit [Le Me (ttt te 2 2 2 6 THE EFFECT OF INTERRUPTING THE NORMAL PROCESS OF HYDRANTH FORMATION Driesch (’97) first showed, in his researches on Tubularia, that when the formation of the hydranth is interrupted by separating the two tentacle ridges shortly after they appear, the method of completing the hydranth is not always the same. He has described four methods of regulation: (1) The ‘‘ Regenerationsmodus”’ where the hydranth emerges with the original proximal tentacles, and later develops a new distal row; (2) the “ Ersatzanlagemodus”’ where the coenosarce in front of the proximal row elongates and a new distal row appears before the hydranth emerges; (3) the “ Auftheilungsmodus”’ where the prox mal tentacles divide, form- ing the distal row from their distal ends; (4) the “ Auflésungs- modus” where the proximal row disappears entirely and a new anlage forms. I repeated these experiments (’00) suggesting that the difference in the method of completion of the hydranth on the proximal piece was due to the degree of differentiation of the pri- mordia. If the distal row of tentacles was removed soon after the red material had begun to collect in the two rows, the fourth method was invariably followed, 1. e., the first proximal row dis- 348 Florence Peebles appeared and the complete new anlage developed. If on the other hand the two rows were separated later after they were well defined, the first method of completion was followed. The prox- imal piece is, therefore, as Driesch has shown, capable of com- pleting itself in a distal direction. The small piece (4) bearing the distal row of tentacles, is also able to complete itself distally, but as far as | am aware no one has found that such a piece is capable of forming new proximal tentacles, thus completing itself in a proximal direction. It seemed to me worth while to repeat this experiment 1n order to find out at what time the distal piece (Fig 20,4) becomes so highly differentiated that it is no longer able to complete itself proximally, and also to observe the other methods described by Driesch, especially the “ Auftheilungsmodus” which | had never seen, although I had repeated the experiment more than a hundred times. In order to find out the exact time when the small piece (Fig. 20, 4) becomes too highly differentiated to complete itself, it was necessary to remove the tip of the stem before the tentacle ridges were visible. Driesch tried this on forty-five pieces, allowing about twenty-four hours to intervene between the first and second operations. Out of these forty-five pieces, thirty-seven developed one row of tentacles, two formed a double row, and six a complete hydranth without a stalk. Five of the six pieces that formed a complete hydranth were “zu gross,”’ therefore he considers that there were only three out of forty which developed more than the distal row. He concludes that there is therefore a definite time before the anlage appears, when the character of the further development of the smaller tip is determined. In order to deter- mine the exact position of the tentacle ridges before they are visible on the outside, it was necessary to measure a number of anlagen, then to take the average length. Fifty pieces were cut from dif- ferent stems, and left undisturbed until the anlage was visible on each. Measurements were then made, first from the tip of the stem to the base of the proximal tentacles (Fig. 21, P), and second from the tip of the stem to the base of the distal tentacles (Fig. 21, D) the average length of P was 1.7 mm. of D .6 mm. Grafting of Tubularia 349 Experiment 1. Removal of the region of the distal tentacles before the ridges appear. The hydranth and about 5 mm. of the stem below it were removed from twenty-seven pieces of stems whose length averaged 3.5 cm. These pieces were left undis- turbed for eighteen hours. At the end of this time the circulation in the oral end was rapid but no ridges had been laid down. A piece .7 mm. long was then removed from the oral end of each piece (Fig. 22,4), thus separating the region in which the distal tentacles would appear later, (4) from the longer proximal piece (B) on which the proximal row would have developed. The pieces were isolated and their later behavior observed. ‘The following table gives the results from twenty-five pieces, two of the twenty-seven having been lost. SERIES 1 A Completes hy dram the ater eyes tetas =a) slew lene svete fw = ere 2s incl SEES Gdy erste eee cle tei te 5 Distal'tentacle som yicer. jatar sYoy-seyepetois ee) «: shat atsvats ove sis lojaretove)\etelolel nt evstehete Var eieie eretereetantsys 10 IND VEEN Cirsb) No yn cgi occocHonuedoddne Seanoonppusuehosbansespocpaneaeore cece 10 MBq tal bmataen beret 2 yot 2 -\a)o era toyet=, oYei Se charatehohekitoheyopois) eteieis a ctarag tate vatenshekane heen 13 Motalemium berks) ac /elsrsniéiaie/s cel e/s,2 5 (ots 2 Boca pagde 22 Again it will be seen by comparing this series with Series 2 A that some of the’ proximal and distal pieces retained the original anlage which had been started before the second operation, the larger number, however, developed a new anlage. Series 3 and 5 B will not be given as the pieces developed new hydranths in the usual way. : In the fourth series, the second operation was postponed until after the proximal ridges were discernible, but the distal had not yet appeared. 352 Florence Peebles SERIES 4 B Old proximal row retained, new distal in front. .......-..0 +e eee ee cece reece eee eees 22 IN Git Ea oucanob po ddodoobpUgnton nnd pauabUSobUeucOnGope ROBO AE DUDE On ioonds. 8 Wotalsmumber’ 2; 1a csc/sic/saysle : Se etevere ayeyete etek ete crescent ais asat ale! eiaty afeqstetensveteete 30 The method of completion of the hydranth on the proximal piece (B) when the original row persisted, in Series 1 and 2 was that described by Driesch as the “Ersatzanlagemodus,” a new distal row developed in front of the proximal row before the hydranth emerged. That in Series 4 was Driesch’s “ Regenerationsmodus”’ where the hydranth emerged first, and later a new row of distal tentacles developed. I agree with Child (’o7c) that there 1s no essential difference in the two methods. CONCLUSIONS These experiments prove that before the ridges have become visible, changes have taken place in the tip of the stem that render it when isolated incapable of producing a complete hydranth. These changes however do not take place until several hours after the wound has closed. If the tip is removed shortly after closure of the wound, even as long as twenty-four hours after the first operation, it is still possible for the isolated tip to form a com- plete hydranth. After the proximal tentacle anlage has appeared the distal piece when isolated, even if the tentacle ridges are not visible, does not form a complete hydranth. — Instead the distal tentacles and mouth develop. 7 REVERSING THE DISTAL ROW OF TENTACLES In earlier experiments (’00) I have shown that when the tip of the oral end of a long piece is cut off, reversed, and grafted back again at once, that the new hydranth develops in exactly the same region as it would have if the piece had not been removed. The distal row of tentacles appears in the small piece and the prox- imal row in the longer one. In the following experiment the tip was not removed until after the distal row of tentacles had appeared (Fig. 20) The piece 4 was then removed and grafted in / Grafting of Tubularia 258 the reverse direction on the proximal piece (Fig. 24): This brought the distal tentacles into a position which was slightly dif- ferent from their normal one. ‘The space in front of them, 1. e., between them and the end was greater than before. In a large number of experiments the pieces united, the tentacles completed themselves, and the normal hydranth emerged. In about 10 per cent of the grafts, a most interesting result was observed. ‘The original distal tentacle ridges remained as distinct bands, while in front of them at the cut surface, a new row of distal tentacles and hypostome developed, the hydranth finally emerged with the stripes running from the base of the new distal tentacles to the proximal ones (Fig. 25). No reproductive organs formed on these pieces. The bands persisted until the hydranths dropped off. This serves to demonstrate that the small distal piece A when connected with the proximal piece is capable of developing new structures after the tentacles have been laid down. It also shows that the “red stuff” which is seen in the ridges, is not used again when a new row of tentacles is laid down. Stevens (’02) and Child (07c) have observed in the proximal piece, after separation of the two ridges, that the “red stuff’? which was originally in the proximal row forms a mass at the end of the stem, and when the hydranth is completed the mass is ejected. 8 REMOVAL OF THE ENTIRE PRIMORDIUM Driesch (’02) found when he removed the early anlage of the hydranth by a cut just below the proximal row, that in a large number of cases the ridges disappeared, and a new anlage devel- oped, the latter being much reduced in length. Thus the length of the “reparation area”? bore a definite relation to the length of the piece. Child (’07a) has also shown that there is a reduc- tion in length of the primordia in short pieces, but the reduction in length is not proportional to the reduction in the length of the piece. I have made a series of experiments in which long pieces were cut from the stems of different individuals, and after the tentacle aniage appeared, the distal end of the piece was removed by a cut below the proximal tentacles (Fig. 26, X). I hoped to find that at al 354 Florence Peebles a fixed distance below the primordium the stimulus of the cut would result in the formation of a short aboral hydranth while the oral ridges were fading out and the new short ones were appearing. The results were as follows: When the space below exactly equaled the length of the primordium (Fig. 26) the hydranth con- tinued to develop without any sign of delay caused by the cut. No hydranth developed on the aboral end for this region became the stalk. When the space below the primordium was equal to one-half the length of the primordium the result was the same. The cut was then made close to the base of the proximal row of tentacles. In about one-third of the pieces the original anlage disappeared and a new one much shorter than the first appeared. No aboral hydranths formed on these pieces. un) Miz J hula i ni ew) PP oe i es ee Fic.20 Fie.21 Fic.22 Fic.23 Fic. 24 Fic. 25 Fic. 26 Q THE INFLUENCE OF THE CONCENTRATION OF THE SEA-WATER ~ Loeb’s earlier work (’92).0n Tubularia brought to light the fact that the concentration of the sea-water is a definite factor in the regeneration of new hydranths. He found after testing various strengths, that sea-water diluted to 66 per cent was the optimum strength for growth. If the solution was weaker or stronger the growth was retarded. Snyder (’05) has also tested various strengths of sea-water, and has found that in Tubularia crocea when the sea-water was diluted not only was greater growth observed but a larger number of aboral hydranths developed. My own experiments confirm those of Loeb and Snyder, but ~~ Grafting of Tubularia 355 some of the results which I obtained indicate that the great increase in size of the hydranths and the rapidity of their formation in dilute sea-water is due to something more than the difference in osmotic pressure. The concentration of the sea-water in the Bay of Naples is estimated at 3.8 per cent. If this water is diluted to 2.5 per cent, growth is increased. Herbst (’04) found that artificial sea-water of the same strength as that of the Bay of Naples was more favor- able for growth of sea-urchin larvae than normal sea-water. I followed Herbst’s formula and made a solution of artificial sea- water. I found that growth in this solution was more rapid and also that the hydranths formed in this solution were larger than usual, and lived longer. When the artificial solution was diluted with distilled water, the result was very different from that obtained from diluted water from the Bay. ‘There was no increase in growth or in rate of regeneration. ‘The results showed that the solution was not as favorable for growth as normal sea-water. I concluded from this that osmosis could not be the only factor in determining the increase in growth. _ It is not altogether improb- able that organic substances in the Bay of Naples exert a retard- ing effect on growth. When these are excluded by preparing the pure artificial sea-water their retarding influence is abolished, and the growth which we consider so unusual is really no more than the normal rate under optimum conditions. ‘This would explain why diluting artificial sea-water does not produce the same result as diluting that which comes directly from the Bay of Naples. Child (’07c) has also made a study of the effect of diluting the sea-water. He concludes that the diluted medium increases the energy of the processes which involve hydranth formation. Since I made my experiments with dilute sea-water Child’s work has been published. As my results are practically the same as his I shall omit a description of the experiments. SUMMARY 1 When two individuals are grafted together changes at once take place in the region of the graft. These changes may not be 356 Florence Peebles visible externally, but they exert some influence on the rest of the pieces, whether they result in the formation of new structures or not. 2 When the aboral end of a piece of the stem of Tubularia is stimulated through grafting to produce a hydranth before the oral end, the change in the polarity is not lasting for when a second set of hydranths develop, the piece assumes its original polarity. 3 When a short and a long piece are grafted together the influence of the longer piece is shown only when the new hydranth comes from the two pieces, a row of tentacles forming in each. 4 Short pieces grafted together in any direction usually develop a hydranth on one of the free ends only, or first on one end and later on the other. The result is the same whether the currents in the two pieces are continuous or not. 5 Ifthe tip is removed from the end of a stem on which a new hydranth has just begun to develop, before the ridges are visible, the small piece is capable of forming a complete hydranth. If the tip is removed after the ridges are laid down, the piece devel- ops one row of tentacles. The proximal piece completes itself distally by forming new tentacles on its tip, either before or after emerging from the perisarc. If the tip is removed before the primordia appear the proximal piece usually forms new primordia without delay. 6 If, after the appearance of the primordia, the piece in which the distal row of tentacles develops is reversed and grafted back on the proximal row, the hydranth completes itself in the normal manner. Sometimes the distal piece develops a new rowof distal tentacles in front of the original red ridges which persist after the hydranth emerges. 7 When the entire primordium is removed by a cut just below the base of the proximal tentacles, the ridges frequently fade out and a new primordium develops which is Fas shorter than the original one. A cut 2 to 3 mm. below the primordium does not affect its later development. Such pieces do not form aboral hydranths. 8 Dhiluting normal sea-water in which pieces of Tubularia are placed, increases the rate of growth and the percentage of new Grafting of Tubularia 357 hydranths formed. Artificial sea-water has the same effect, but when the artificial sea-water is diluted these favorable results do not follow. Bryn Mawr, Pa. December 27, 1907 BIBLIOGRAPHY Cuttp, C. M., ’07a—An Analysis of Form-Regulation in Tubularia. II. Differ- ences in Proportion in the Primordia. Archiv f. Entw.-Mech., 23, 1907. o7b—An Analysis of Form-Regulation in Tubularia. V. Regulation in Short Pieces. Archiv. f. Entw.-Mech., 24, 1907. ’o7e—An Analysis of Form-Regulation in Tubularia. VI. The Signif- cance of Certain Modifications of Regulation: Polarity and Form-Regulation in General. Archiv f. Entw-Mech., 24, 1907. Driescu, H. ’97—Studien tber das Regulationsverméogen der Organismen. I. Von den regulativen Wachthums und Differenzirungsfahig- keiten der Tubularia. Archiv fiir Entwickelungsmechanik der Organismen. 5. 1897. ’o2—Studien tiber das Regulationsvermogen der Organismen. 7 Zwei neue Regulationen bei Tubularia. Archiv f. Entw.-Mech., 14, 1902. Hersst, C., ’04—Ueber die zur Entwickelung der Seeigellarven nothwendigen anorganischen Stoffe, ihre Rolle und ihre Vertretbarkeit. 3 Theil. Die Rolle der nothwendigen anorganischen Stoffe. Archiv f. Entw.-Mech., 17, 1904. Logs, J., ’92—Untersuchungen zur physiologischen Morphologie der Thiere. 11 Organbildung und Wachstum. Wurzburg, 1892. ’o4—Concerning Dynamic Conditions which Contribute toward the Determination of the Morphological Polarity of Organisms. University of California Publications, vol. 1, nos. 16 and 17, 1.904. Morean, T. H., ’03—Some Factors in the Regeneration of Tubularia. Archiv f. Entw.-Mech., 16, 1903. Morean, T. H., and Stevens, N. M., ’04—Experiments on Polarityin Tubularia. Journal of Experimental Zodlogy, vol. 1, 1904. PEEBLES, ‘0O—Experiments in Regeneration and in Grafting of Hydrozoa. Archiv f. Entw.-Mech., 23, 1900. °o2—Further Experiments in Regeneration and Grafting of Hydroids. Archiv f. Entw.-Mech., 14, 1902. 358 Florence Peebles Snyper, C. D., ’05—The Effects of Distilled Water on Heteromorphosis in a Tubu- larian Hydroid, Tubularia crocea. Archiv f. Entw.-Mech.,, 19, 1902. Stevens, N. M., ’o2—Regeneration in Tubularia mesenbryanthemum. 11 Archiv f. Entw.-Mech 15, 1902. A’ “sSLUDY OF THE GERM CELLS OF CERTAIN DIPTERA, WITH REFERENCE TO THE HETERO- CHROMOSOMES AND THE PHENOMENA OF SYNAPSIS BY N. M. STEVENS With Four Pirates INTRODUCTION In connection with previous work on the spermatogenesis of the Coleoptera (’05, ’06), the germ cells of the common fruit-fly, Drosophila ampelophila, were examined in the autumn of 1906. The difficulties encountered in handling this material led to the study of the spermatogenesis of several other flies. The results will be presented in accordance with the following scheme: Calyptrate Muscidz. Muscine. t Musca domestica. 2 Calliphora vomitoria. 3 Lucilia cesar. Sarcophagine. , 4 Sarcophaga sarraciniz. Anthomyiine. 5 Phorbia brassica. Acalyptrate muscid. » 6 Scatophaga pallida. 7 ‘Tetanocera sparsa. 8 Drosophila ampelophila. Syrphide. 9g Eristalis tenax. METHODS With this material it was found that the best results could be obtained from fresh tissue mounted in Schneider’s aceto-carmine. Tue Journat or ExreriMeNTAL ZoOLoGy, VoL. v, No. 3 360 N. M. Stevens The testes (or ovaries) of adult flies were dissected out in physio- logical salt solution and immediately transferred to a drop of aceto- carmine on a slide. The cover-glass was pressed down with a needle to break the capsule of the testis and spread the cells. All excess of stain was removed with filter paper, and after ten or fifteen minutes, the preparation was sealed with vaseline. Such preparations may be studied to the best advantage after twenty- four hours, as the chromatin gradually acquires a deeper tint. They remain in good condition for several days, but are, of course, not permanent. The method has several advantages besides that of enabling one to examine a large amount of material in a limited time. The aceto-carmine fixes and stains instantly without the shrinkage incident to the usual treatment with fixing fluids, alco- hols, xylol and parafhn, necessary in order to obtain sections. Then, one is able to study the whole cell with all of the chromo- somes present and uncut, which is an obvious advantage for work of this kind. The chromatin stains much more deeply than any other cell element, but the achromatic structures are not always well brought out, and they have been omitted from most of the figures, as this investigation is concerned primarily with the hetero- chromosomes and the method of synapsis. In favorable prepara- tions of this kind, with good light, it is possible to get as accurate camera drawings as from sections stained with iron hematoxylin. RESULTS OF INVESTIGATION I Musca domestica In many respects the spermatogenesis of this fly resembles that of Tenebrio molitor (Stevens ’05), Odontata dorsalis (Stevens ’06) and the other Coleoptera which have an unequal pair of heterochromosomes. ‘There are, however, no synapsis, synize- sis Or spireme stages in the spermatocytes, nor are tetrads ever formed. In the prophase of spermatogonial mitoses one finds five pairs of long slender chromosomes, the members of each pair either lying parallel to each other or twisted together (Fig. 1). The members of the additional unequal pair are usually separate The Germ Cells of Diptera 361 (Fig. 1, h, and h,). Apparently a side-to-side pairing or conju- gation of homologous chromosomes, with the possible exception of the unequal pair, occurs preliminary to each spermatogonial mitosis. [he twelve chromosomes separate, and each divides longitudinally in metakinesis. Whether they pair again in the telophase or not until the prophase of another cell-division is not evident. The heterochromosomes remain condensed and are found side by side during the whole growth stage, while the other chromo- somes pass into a more or less diffuse condition (Fig. 2). In the prophase of the first spermatocyte mitosis there are five thick V-shaped chromosomes and a pear-shaped mass of chromatin which in metakinesis proves to be the unequal pair of hetero- chromosomes (Fig. 3). The V-shaped chromosomes all divide longitudinally and the larger and smaller heterochromosomes separate as seen in Fig. 4. In the interval between the first and second divisions a nuclear membrane forms, but the chromosomes do not change greatly. Figs. 5 and 6 show the two kinds of daughter nuclei, one containing the larger, the other the smaller heterochromosome. In the second spermatocyte mitosis the V- shaped chromosomes again divide longitudinally and the hetero- chromosomes divide as shown in Figs. 7 and 8, so that in all stages they are clearly distinguishable from the ordinary chromosomes. The resulting spermatozoa fall into two equal classes, dimorphic as to the heterochromosomes, as in similar cases among the Hem- iptera and Coleoptera. In most of the flies studied there was no difficulty in finding odgonia in which the number and relative size of the chromosomes could be determined. Only one such was found in Musca, that shown in Fig. 9. Here a part of the chromosomes are still paired; others have separated, but the members of each pair of ordinary chromosomes are not far apart; while the two equal heterochromosomes are on opposite sides of the group (h). Here again we have what may be regarded as a partial synapsis of homologous chromosomes. The relation of the heterochromosomes in the two sexes is the same as in many of the Coleoptera (Stevens ’05 and ’06) and the Hemiptera heterop- tera (Wilson ’o5 and ’o6), an unequal pair (large and small) in 362 N. M. Stevens the male and an equal pair of large heterochromosomes in the female. An egg which is fertilized by a spermatozo6n containing the smaller heterochromosome produces a male, while one which unites with a spermatozoon containing the larger heterochromo- some produces a female. Although there is no distinct synapsis stage visible in the devel- opment of the spermatocytes of Musca domestica, the method of synapsis is without doubt indicated by the side-to-side pairing of chromosomes of equal length in the prophases of both spermato- gonial and odgonial mitoses. The final synapsis is a closer union of the homologous chromosomes, and the first spermatocyte division separates the members of each pair instead of dividing each chromosome as in the spermatogonia. 2 Calli phora vomuitoria The chromosomes in this species are similar to those in Musca domestica. Both members of the unequal pair of heterochromo- somes are smaller, as may be seen in a spermatogonial metaphase (Fig. 10). Pairing of homologous chromosomes is also evident here. In the growth stage (Fig. 11) the heterochromosomes are associated with a plasmosome as in many species of Coleoptera. Two views of the metaphase of the first spermatocyte mitosis are shown in Figs. 12 and 13, and an anaphase in Fig. 14. Two metaphases and an anaphase of the second division appear in Figs. 15, 16 and 17. The equal pair of heterochromosomes in the female is clearly shown in two odgonial metaphases (Figs. 18 and 19). In this case we have further evidence of the side-to- side pairing of homologous chromosomes in the spermatogonia and oogonia. 3 Lucilia Cesar Only a few specimens of this species were captured and the series of stages is incomplete. No spermatogonial or odgonial metaphases were found. In the growth stage a pair of m-chro- mosomes is present with an enormous heterochromosome bivalent The Germ Cells of Diptera 363 (Fig. 20). The metaphase of the first spermatocyte division is shown in Figs. 21 and 22, and prophases of the two kinds of second spermatocytes in Figs. 23 and 24. The spermatozoa would evi- dently be dimorphic as in the other species. 4 Sarcophaga sarracinie The three species of Diptera whose spermatogenesis has already been described belong to the sub-family Muscinz, while Sarcoph- aga is a member of the sub-family Sarcophagine. The number of chromosomes in Sarcophaga is the same as in the other species, I2 somatic and 6 reduced, and the heterochromosomes closely resemble those in Calliphora. The spermatogonial plate (Fig. 25) shows the 12 chromosomes paired, but separated ready for metakinesis, and one chromosome shows the division line. In the growth stage (Fig. 26) the pair of heterochromosomes comes out clearly in the midst of diffuse and irregular masses of faintly stained chromatin. In these flies the ordinary chromosomes become much branched or diffusely granular in the growth stage but do not unite to form a spireme of even width as in so many forms. Whether or not they unite end-to-end at any stage before or after synapsis I cannot say. A prophase and an equatorial plate of the first spermatocyte mitosis may be seen in Figs. 27 and 28, and the metaphase and anaphase in Figs. 29, 30 and 31. The polar views of the metaphase of the second mitosis (Figs. 32 and 33) of course show dimorphism as to the heterochromosomes (Ay, h,). Equal division of all of the chromosomes follows as in the three preceding species. Figs. 34 and 35 were drawn from adjacent odgonia in metaphase to show the close longitudinal pairing of the chromosomes and their later separation before metakinesis. The equal heterochromosomes are usually found together in the middle of the plate and each one is evidently equiv- alent in size to the larger heterochromosome of the spermato- gonia and spermatocytes (Figs. 25 to 33). Fig. 36 is from an ovarian follicle cell. The four figures, 25, 34, 35 and 36 show the pairing of homologous chromosomes in spermatogonia, O6gonia and somatic cells. 304 N. M. Stevens 5 Phorbia brassica Only one male of Phorbia was obtained and only four stages drawn; but these indicate precisely the same conditions as in the other species examined. Phorbia belongs to the sub-family Anthomyine. Fig. 37, a growth stage; 38, a prophase; 39, a metaphase; and 40, an anaphase, show clearly the presence of an unequal pair of heterochromosomes resembling those of Musca domestica. 6 Scatophaga pallida 7 Tetanocera sparsa The chromosomes of Scatophaga and Tetanocera resemble each other so closely in number, form and behavior that they will be considered together. Fig. 41 is a spermatogonial prophase of Scatophaga; and Figs. 42 and 43, spermatogonial prophase and metaphase of Tetanocera. All show equally paired ordinary or V-shaped chromosomes and unequally paired heterochromosomes. Figs. 44 and 45 are prophase and metaphase of the first spermato- cyte of Scatophaga, Figs. 46 and 47 the corresponding stages for Tetanocera. In both species it will be seen that there is a close resemblance between the paired condition of the chromosomes in the prophases of a spermatogonial division and of a first spermato- cyte mitosis. In general the chromosomes were larger in the spermatogonia (Figs. 41, 42, 43) than in the spermatocytes (Figs. 44, 45, 46, 47), but frequently prophases of spermatocyte mitoses could be certainly identifed as such only by the metaphases in the same cyst and the growth stages in the neighboring cysts. The only actual observable difference between the synaptic con- dition in the spermatocytes and the spermatogonia is the behavior of the pairs in the following mitosis: in the spermatogonia the members of the pairs separate in metaphase (Fig. 43), and each divides in metakinesis; while in the spermatocytes the members of each pair remain closely associated in metaphase (Figs. 45 and 47) and separate in metakineses (Fig. 48), but do not divide until the The Germ Cells of Diptera 365 second spermatocyte mitosis, though they frequently show the preparatory split in the anaphase (Fig. 49). We have here an unusually clear demonstration of the essential facts of synapsis and reduction, together with the rather unusual phenomenon of conjugation of homologous chromosomes in cells outside the sphere of maturation. Prophases of the second spermatocyte mitosis in Scatophaga appear in Figs. 50 and 51, and metaphases in Tetanocera in Figs. 52 and 53. An odgonial prophase and anaphase are given in Figs. 54 and 55, and a late prophase for Tetanocera in Fig. 56. These two species as well as the one following belong to the Acalyptrate Muscide. 8 Drosophila ampelophila Drosophila has been placed at the end of the list of Muscidz because of the peculiarities which occur in the behavior of its chromosomes and the difficulties which have been encountered in their interpretation. While in Sarcophaga all the stages neces- sary for a description of the behavior of the heterochromosomes of both sexes were found in the course of a few hours’ work on perhaps ten or twelve preparations, satisfactory results in the case of Drosophila have been obtained only after a prolonged study extending over more than a year and involving the dissection and examination of some two thousand individuals. Sectioning the material has never given satisfactory results. Hermann’s platino- osmic solution and Worcester’s formal-sublimate gave the best fixation, but the division stages are so scattering that permanent preparations, even if good fixation were secured, seemed less prac- tical than the aceto-carmine method, which is much quicker and gives clearer pictures of the mitotic phenomena when they are present. Spermatogonial mitoses are not abundant, and cells in which perfectly clear equatorial plates can be studied are exceedingly rare. [he chromosomes in prophase are paired and twisted together in such a manner that it has been impossible to make an intelligible drawing of them in an early prophase. In Fig. 57, a 366 N.M. Stevens late prophase, two small spherical chromosomes and four larger elongated ones are distinctly paired while the members of the unequal pair (/,, h,) are separated. For a long time it was impossible to be sure that an unequal pair was present, as fore- shortening in the case of one chromosome (/,) was possible, but recently a comparatively large number of good spermatogonial plates has been secured in which the inequality in length of one pair is clearly demonstrated. No case has been found in which the members of this pair appeared to be equal. Figs. 58, 59 and 60 show exceptionally clear cases, and Fig. 61 shows a peculiar folding of the chromosome /,, whose significance may be apparent as we proceed to consider the unequal heterochromosome bivalent of the first spermatocyte. In Drosophila the heterochromosomes cannot be demonstrated in the growth stages of the first spermatocyte. In some sections from Hermann material stained with thionin the plasmosome (p) and some of the chromosomes appeared as in Fig. 62 in cysts adjacent to the spermatogonial cysts. In later growth stages nothing definite, except the immense plasmosome, can be made out in regard to the contents of the nucleus. The earliest pro- phase of division is the appearance of the chromatin massed together, usually on one side of the nucleus, while the plasmo- some may be in the middle or on one side of the nucleus (Fig. 63). In aceto-carmine preparations the chromosomes first appear in early prophase, scattered through the nucleus, faintly stained and irregular in outline (Fig. 64). The plasmosome may be broken up at this time or it may appear intact in the spindle. Figs. 65 and 66 are later prophases in which the chromosomes are completely condensed. The unequal heterochromosomes are h, and hy. Fig. 67 shows the three equal bivalents, and the unequal hetero- chromosome pair in its simplest form, in the metaphase of the first spermatocyte mitosis. Fig. 68 shows slight modifications of this form from other cells of the same cyst. The most common form of this pair is seen in Figs. 69 and 70, where there are two equal V-shaped elements and a third portion (x) which in many cases looks like a separate element, and for a time the group was thought to be trivalent; 1. e., made up of two equal V-shaped The Germ Cells of Diptera mg07, chromosomes and a smaller odd chromosome. ‘This belief was strengthened by the appearance of many metaphases and ana- phases (Figs. 70, 71, 72) where the third portion of the figure (x) seemed to be on the point of separating from the V-shaped element next to it. This opinion was not confirmed however by the composition of the spermatogonial or oégonial equatorial plates, nor was it possible to demonstrate with certainty a separate ele- ment corresponding to x in the polar plates of the first spermato- cyte mitosis or in the second spermatocyte. Fig. 72 1s one of several cases where the portion x seemed to be separated from the two other elements of the group, but the separation'must have been only apparent, for one much oftener finds an anaphase like Fig. 73 where the separation of the heterochromosome group into two unequal parts is certain (iu, h,). Sometimes the anaphase is like Fig. 74, where more or less spherical masses replace the usual V’s of the heterochromosome group. Often all of the chromo- somes except the smallest pair show in the metaphase that they are elongated and V-shaped (Fig. 75), and in late anaphases (Fig. 76) the elements of the two largest bivalents are usually divided and the daughter chromosomes separated, often crossed. Both here and in the second spermatocytes it is often difficult or impossible to distinguish the heterochromosomes from the others. In the telophase the chromatin forms a dense mass which loses none of its staining quality and is soon resolved into the already divided chromosomes of the second spermatocytes (Figs. 77, 78, 79). A greater or less degree of elongation together with twisting and fore-shortening makes it impossible to measure or even esti- mate with any accuracy the relative length of the chromosomes, so as to distinguish the two classes of second spermatocytes as to size of heterochromosomes. Figs. 78 and 79 are two equa- torial plates from the same cyst where the corresponding chromo- somes are probably a—a, b—b, and h,—h,. All of the chromo- somes divide in this mitosis. The odgonial metaphases are perfectly clear, and four equal pairs of chromosomes are always present (Figs. 80, 81, 82). In the metaphase they are usually grouped in pairs, and in the pro- phase they are closely approximated and twisted. In fact this 368 N. M. Stevens prereductional pairing of homologous chromosomes was first noticed in the odgonia and ovarian follicle cells of Drosophila. An attempt was made to ascertain whether such a pairing occurs in embryonic cells. Very little evidence was obtained. In the prophase of one mitosis paired chromosomes were found (Fig. 83). Fig. 84 is the equatorial plate of a segmentation stage. In both cases the pairs appeared to be equal. Q Eristalis tenax A considerable number of these flies were captured on some late blooming mustard plants in October. The material was in exceptionally favorable condition, and a complete series of draw- ings was obtained. ‘The outer wall, or capsule, of the testis was thinner and more permeable to fixing fluids than in most of the other species studied and it was therefore possible to work with both sections and aceto-carmine preparations. ‘This fly belongs to the family Syrphide, but the chromosomes in most respects resemble those of the Muscide. The heterochromosome bivalent is different in form from that of any of the Muscidz described above; it however consists of a larger and a smaller component united in a somewhat different way from the corresponding ele- ments in Drosophila. Among the spermatocytes, several follicle cells were found in mitosis; the chromosomes of one such is shown in Fig. 85. The spermatogonial chromosomes are paired in prophase but sepa- rate and form a flat plate in the metaphase as seen in Fig. 86, where the two heterochromosomes (/,, h,) are conspicuously unequal in size. In this form there is a distinct synizesis stage, as shown in Fig. 87, from a section of material fixed with Gilson’s mercuro-nitric fluid and stained with thionin. ‘The cysts in which this stage occurs border upon the spermatogonial region of the testis. The outline of the chromosomes is visible and in the next stage the chromosomes are distinctly bivalents. Later they become more diffuse, but do not appear to form an even spireme at any stage. Fig. 89 is a growth stage, showing the heterochro- mosome group (/), a pair of m-chromosomes and the other chro- The Germ Cells of Diptera 369 mosomes in a loosely branched condition. Fig. go ‘s an early prophase in which the heterochromosome pair is very compact and deeply stained, while the other chromosomes are granular and denser in some parts than in others. A later prophase (Fig. 91), from a section, shows the heterochromosome pair assuming the cross-shape which we find in the later metaphase. Fig. 92 is a polar view of the equatorial plate of the first spermatocyte; and Figs. 93 and 94, side views of the spindle to show the cross-shaped heterochromosome bivalent in two positions. Here the crosg (Fig. 94), instead of having opposite arms equal, as in cross- shaped tetrads composed of equal elements, has one of the ver- tical arms longer. It is evident from Figs. 93 and g5 that the longer arm is the smaller heterochromosome, while the remainder of the cross is the larger member of the pair. The larger ele- ment is folded in the same manner as in Drosophila (Figs. 66 and 67) but the smaller element is attached by one end instead of by the middle as in Drosophila. The second spermatocyte mitosis pro- ceeds as in the other forms and presents nothing of especial interest. Dimorphism of the spermatozoa is foreshadowed by the first sper- matocyte anaphases (Figs. g6 and g7). In the female the clearest figures were obtained from ovarian follicle cells (Figs. g8 and 99). The pairs are equal and comparison with the spermatogonial chromosome group (Fig. 86) indicates that the equal heterochro- mosome pair 1s one of the two longest. The general results for the nine species of flies are the same; 1. e., an unequal pair of heterochromosomes in the male leading to dimorphism of the spermatozoa, and a corresponding equal pair in the female, each equivalent to the larger heterochromosome of the male: also a prereductional pairing of homologous chromo- somes in the prophase of mitosis in spermatogonia, o6gonia, and ovarian follicle cells. DISCUSSION rt Sex Determination So far as I know there is no published work on the heterochro- mosomes of the Diptera. The literature on the heterochromo- 37° N.M. Stevens somes in other orders of insects has recently been so fully dis- cussed in a paper by A. M. Boring (’07) that it seems hardly nec- essary to go into the subject exhaustively here. “The dimorphism of the spermatozoa resulting from the maturation of the male germ cells of the nine species of Diptera considered in this paper is of the same character as that described by the author for 36 species of Coleoptera (see note, p. 49, Stevens ’06), and by Wilson (05 and ’o6) for several species of Hemiptera heteroptera. The dimorphism is brought about by the presence in the spermatogonia and spermatocytes of an unequal pair of heterochromosomes, while in large numbers of other insects such dimorphism is due to the presence of an odd chromosome in the male germ cells. These flies have proved to be exceptionally favorable material for demon- strating the occurrence in the female germ cells and somatic cells of a pair of chromosomes, each equivalent to the larger hetero- chromosome of the male. Here, as in similar cases previously described, it is perfectly clear that an egg fertilized by a spermatozo6n containing the smaller heterochromosome produces a male, while one fertilized by a spermatozoon containing the larger heterochromosome develops into a female. The material does not, however, throw any fur- ther light on the question whether the dimorphic spermatozoa are themselves in some way instrumental in determining sex in these insects; or whether sex is a character borne by the hetero- chromosomes and segregated in the maturation of the germ cells of each sex. If the latter supposition is true, sex is probably determined by the dominant heterochromosome of the egg, and fertilization is selective as has been shown in previous papers (Wilson ’o5, ’06; Stevens ’06, p. 54; Nowlin ’06; Boring ’o7). The only hope of determining whether sex is a Mendelian char- acter seems at present to lie in breeding experiments with forms that may be shown by cytological study to be favorable. It is probable that in some cases at least, other characters may be so correlated with sex that their behavior in heredity may throw light on the sex question. As to the proportion of sexes in these flies, a few figures may be given for Drosophila ampelophila. In the autumn and winter The Germ Cells of Diptera 371 of 1906-07, Drosophila was bred in the laboratory on two kinds of food, grapes and bananas. As the flies were dissected for the cytological work, a record was kept of the numbers of each sex; 1551 were so recorded. Of these 759, or 48.94 per cent were males; 792, or 51.06 per cent females. ‘The records of the grape- fed and the banana-fed flies were kept separately. The total number of grape-fed flies dissected between November 1 and March 19 was 787, 404 or 51.33 per cent being males, and 383 or 48.67 per cent females. The banana-fed flies between October 30 and December 3 numbered 764, 355 or 46.47 per cent males, and 409 or 53.53 per cent females. In the total number there were 2.02 per cent more females than males, in the grape-fed 2.66 per cent more males than females, and in the banana-fed 7.06 per cent more females than males. These differences are probably not significant, but if sex is a Mendelian character, the numbers for the two sexes should of course be equal unless food produces some discriminating effect on the development of either individuals or eggs of the different sexes. It has always been a noticeable fact that the banana-fed flies were larger and more robust than those fed on grapes; this however applies to both sexes. In mass cultures it is not possible to tell whether failure of many of one sex or the other to reach the adult stage in differ- ent cultures might account for the discrepancies in numbers observed with the two kinds of food. Castle and his co-workers (’06 p. 772) found the sexes about equal in three families of the sixth inbred generation of a grape- fed series, and the remarks which follow the table indicate that they regard the normal proportion as near equality. Monkhaus’ results on sex in Drosophila seem not yet to be in print, except for a brief report in the Year Book of the Carnegie Institution. An attempt was made to ascertain the normal proportion of the sexes for the adults of Musca domestica. When caught by hand 58.33 per cent were females, but when a wire trap baited with sugar or molasses was used, only 46.53 per cent were females. The results need no comment. Cuénot states that the normal proportion of males to females 372 N.M. Stevens in Lucilia caesar, Calliphora vomitoria and Sarcophaga carnaria is approximately equal, and his experiments show that neither amount nor kind of food given to the larvae has any marked effect on the proportion of the sexes in the first or second generation, but here as elsewhere in such experiments the number of eggs that did not hatch is not noted, and this may be the critical point. It is evident that more experiments are needed in which the fate of all of the eggs of isolated pairs of flies is determined. 2 Synapsis In the spermatogenesis of most insects synapsis involves an end-to-end union of homologous chromosomes, and tetrads of various forms are commonly found in the prophase of the first spermatocytes. In these flies no tetrads have been observed and as a rule nothing comparable to the synizesis, bouquet or spireme stages of other forms is apparent. In these respects the germ cells of the Diptera resemble the oogonia of sagitta (Stevens ’03 and ’o5) and the male and female germ cells of the aphids (Stevens ’o5 and ’o6). In the oogonia of Sagitta the chromosomes pair side-to-side in an early stage, while in the spermatogonia of the aphids the pairing occurs as a prophase of the first spermatocyte mitosis. ‘The indications are that in the flies the chromosomes are already paired side-to-side at the beginning of the growth stage (Figs. 87 and 88), but the pairs do not appear to unite end- to-end to form a spireme. In some cases the members of the pairs are perfectly fused in the prophase of the first spermatocyte (Figs. 3 and 27); 1n others the bivalents are clearly such in both prophase and metaphase (Figs. 44 to 46). The first spermatocyte division is without doubt reductional for both ordinary chromosomes and heterochromosomes. Perhaps the most interesting point in the whole study is the pairing of the chromosomes in cells somewhat removed from the sphere of the reduction process. This was first noticed in the oogonia of Drosophila, and was also found to occur in the ovarian follicle cells, the spermatogonia and some embryonic cells. This is not an occasional phenomenon, but one which belongs to every The Germ Cells of Diptera 292 odgonial and spermatogonial mitosis. In many cases the pro- phases of spermatogonia and first spermatocytes resemble each other very closely, the members of each pair being twisted together in both. In the spermatocyte we get complete synapsis and reduction; in the spermatogonium only a foreshadowing of reduction, and abundant proof that synapsis is here a side-to-side pairing of homologous chromosomes, and the first spermatocyte division a separation of univalent chromosomes, and not a longitudinal or quantitative division of two chromosomes united end-to-end. ‘The relation of the heterochromosomes to each other in synapsis varies greatly with differences in form and size. One is tempted to suggest thatif homologous maternal and pater- nal chromosomes in the same cell ever exert any influence on each other, such that it is manifest in the heredity of the offspring, there is more opportunity for such influence in these flies than in cases where pairing of homologous chromosomes occurs but once in a generation. Possibly experiments in cross-breeding of flies may bring out some interesting facts in heredity. Nore. A preliminary statement in regard to the chromosomes of Drosophila was made at the International Congress of Zodlogists in Boston, August 21, 1907. The question as to whether an odd chromosome or an unequal pair of heterochro- mosomes was present in the male cells was then unsettled. LITERATURE CITED Borine, A. M., ’07—A Study of the Spermatogenesis of Twenty-two Species of the Membracidw, Jassida, Cercopida and Fulgoride. Jour. Exp. Zodl., vol. iv. Castie, W. E., and others, ’06—The Effects of Inbreeding, Crossbreeding, and Selection upon the Fertility and Variability of Drosophila. Mus. Comp. Zoél., Harvard, no. 177. CuEnor, L., ’99—Sur la détermination du sexe chez les animaux. Bull. Sci. de la France et Belg., vol. xxxii. Nowuin, W. N., ’06—A Study of the Spermatogenesis of Coptocycla aurichalcea and Coptocycla guttata, with especial reference to the Problem of Sex Determination. Jour. Exp. Zodl., vol. iii. 374. N. M. Stevens Stevens, N. M., ’03—On the Ovogenesis and Spermatogenesis of Sagitta bipunc- tata. Zool. Jahrb., vol. xviii. ’o5—Further Studies on the Ovogenesis of Sagitta. Zool. Jahrb., vol. xxi. ‘o5—A Study of the Germ Cells of Aphis rose and Aphis cenothere. Jour. Exp. Zol. vol. 11. ’o5—Studies in Spermatogenesis with especial reference to the “Accessory Chromosome.” Carnegie Inst., Wash., Pub. 36. — ; ’o6—Studies on the Germ Cells of Aphids. Carnegie Inst., Wash., Pub. 5. ‘o6—Studies in Spermatogenesis. I]. A Comparative Study of the Heterochromosomes in Certain Species of Coleoptera, Hemiptera and Lepidoptera, with especial reference to Sex Determination. Carnegie Inst., Wash., Pub. 36, no. 2. Witson, E. B., ’0o5—Chromosomes in Relation to the Determination of sex in Insects. Science, n. s., vol. xx. ’o5s—Studieson Chromosomes. I. ‘The Behavior ofthe Idiochromosomes in the Hemiptera. Jour. Exp. Zodl., vol. 1. ‘o5—Studies on Chromosomes. II. The paired Microchrosomes, Idiochromosomes, and Heterotropic Chromosomes in the Hem- iptera. Jour. Exp. Zodl., vol. il. *o6—Studies on Chromosomes. II. Sexual Differences of the Chromo- some Groups in Hemiptera, with some Considerations on Determination and Inheritance of Sex. Jour. Exp. Zodl., vol. i. Pirate IL Sarcophaga sarracinia Fig. 25 Spermatogonium, metaphase. Fig. 26 First spermatocyte, growth stage. Fig. 27 First spermatocyte, prophase. Fig. 28 First spermatocyte metaphase. Fig. 29 First spermatocyte, metakinesis. Figs. 30 and 31 First spermatocyte, anaphase. Figs. 32 and 33 Second spermatocyte, metaphase. Figs. 34 and 35 Odgonia, metaphase. Fig. 36 Ovarian follicle cell, metaphase. Phorbia brassica Fig. 37 First spermatocyte, growth stage. Fig. 38 First spermatocyte, prophase. Fig. 39 First spermatocyte, metaphase. Fig. 40 First spermatocyte, anaphase. Scatophaga pallida and Tetanocera sparsa Fig. 41 Scatophaga, spermatogonium, prophase. Fig. 42 Tetanocera, spermatogonium, prophase. Fig. 43 Tetanocera, spermatogonium, metaphase. Fig. 44 Scatophaga, first spermatocyte, prophase. Fig. 45 Scatophaga, first spermatocyte, metaphase. Pirate II Scatophaga and Tetanocera (continued) Fig. 46 Tetanocera, first spermatocyte, prophase. Fig. 47 Tetanocera, first spermatocyte, metaphase. Fig. 48 Scatophaga, first spermatocyte, anaphase. Fig. 49 Scatophaga, first spermatocyte, anaphase. Figs. 50 and 51 Scatophaga, second spermatocyte, prophase. Figs. 52 and 53 Tetanocera, second spermatocyte, metaphase. Fig. 54 Scatophaga, odgonium, prophase. Fig. 55 Scatophaga, odgonium, anaphase. Fig. 56 Tetanocera, odgoniaum, prophase. Drosophila ampelophila Fig. 57 Spermatogonium, late prophase. Figs. 58-61 Spermatogonia, metaphase. Fig. 62 First spermatocyte, early growth stage. Fig. 63 First spermatocyte, very early prophase. Fig. 64 First spermatocyte, prophase. Figs. 65 and 66 First spermatocyte, late prophase. Fig. 67 First spermatocyte, metaphase. Fig. 68 Heterochromosome pairs. Figs. 69-71 First spermatocyte, metaphase. Figs. 72 and 73 First spermatocyte, anaphase. PLATE III a®,. ie ei a = A we. Parte IV Drosophila (continued) Fig. 74 First spermatocyte, anaphase. Fig.75 First spermatocyte, metaphase. Fig. 76 First spermatocyte, anaphase. Figs. 77-79 Second spermatocyte, metaphase. Figs. 80-82 Odgonia, metaphase. Fig. 83, Chromosomes from embryonic cell. Fig. 84. Chromosomés from segmentation stage. Eristalis tenax Fig. 85 Chromosomes of follicle cell of the testis. Fig. 86 Spermatogonium, metaphase. Fig.87 First spermatocyte, synizesis stage. Fig. 88 First spermatocyte, growth stage immediately following synizesis stage. Fig. 89 First spermatocyte, later growth stage. Fig. 90 First spermatocyte, early prophase. Fig. 91 First spermatocyte, late prophase. Figs. 92-95 First spermatocyte, metaphase. Figs. 96 and 97 First spermatocyte, anaphase. Figs. 98 and 99 Chromosomes of ovarian follicle cells, prophase and metaphase THE GERM CELLS OF DIPTERA PLATE IV M. Srevens | y V 74 e vt "a R. a Y 42 >) abe or rN h. AA 76 80 81 82 a a mm ae te Od y gm &e gy A oh A pe eo L ea 5 et From the Marine Biological Laboratory, Wood's Hole, and the Laboratory of Physiological Zodlogy, University of Pennsylyania MOMENTARY ELEVATION OF TEMPERATURE AS A MEANS OF PRODUCING ARTIFICIAL PARTHE- NOGENESIS IN STARFISH EGGS AND THE CON- DITIONS OF ITS ACTION BY RALPH S. LILLIE I INTRODUCTION Exposure of mature eggs of Asterias forbesii to the influence of cold sea-water (about 4° to 7°) for somewhat prolonged periods (1 to 7 hours) was first shown by Greeley,! at Wood’s Hole in 1901, to be followed by cleavage and production of larvae on return to normal temperatures. Greeley also experimented with tem- peratures higher than the normal, exposing eggs (taken from the same dishes as those used for experiments with cold) to tempera- tures of 31° to 37° for similar periods of time (1 to 7 hours); but the results of this treatment were purely negative, the eggs merely absorbing water and undergoing a change which he described as liquefaction. He concluded, somewhat sweepingly, that “seg- mentation of the starfish egg cannot be produced by raising the temperature of the sea-water.” He found later (summer of 1902) that temperature was an important factor in the production of parthenogenesis by hypertonic solutions,’ the time of exposure decreasing (within a certain range of temperatures) as the tempera- ture rose, a result confirmed by Lyon* at Naples in the fall of 1902 for species of Strongylocentrotus and Arbacia. But eleva- tion of temperature alone, unaccompanied by other treatment, remained ineffective; moreover, “at 30° it was found impossible to produce artificial parthenogenesis in Asterias or Arbacia with any of the solutions used.”’ In the earlier paper Greeley had 1 Greeley: American Journal of Physiology, vi, p. 296, 1902. 2 Greeley: Biological Bulletin, iv, p. 129, 1903. 3 Lyon: American Journal of Physiology, ix, p. 308, 1903. Tue JourNnaL or ExPERIMENTAL ZOOLOGY, VOL. V, NO. 3 76 Ralph S. Lillie Oo treated with incredulity Delage’s account of successful experiments with higher temperatures, ascribing the results to the effects of agitation and not of simple elevation of temperature. In Greeley’s own experiments “when great care was exercised in handling the eggs not a single segmentation was produced.” ‘The criticism, however, was Te funded for it was clear from Delage’s papers’ that his eggs were exposed to the high temperatures at a time— namely, e BG, maturation—when agitation is quite ineffective in producing parthenogenesis. It is not until the eggs have been mature for some time that this result appears;? while warming, as Delage expressly afirms, is most effective during early matura- tion stages. In Greeley’s second paper he again cites Delage’s experiments, but without comment. Evidently his intention was to return to the subject. Since the appearance of Delage’s papers in 1go1 there seems to have been little further investigation of the influence of rise of temperature in exciting celepme of unfertilized eggs. “The theoretical possibility that development could thus be iaceed was incidentally adverted to by Loeb® some years later: if the sperma- tozoon acts by introducing positive catalysers into the egg, thus accelerating the chemical processes on which the initiation of devel- opment depends, a similar acceleration with similar consequences ought to follow simple elevation of temperature. Loeb has also more recently emphasized the importance of the temperature factor in the production of parthenogenesis by the use of hyper- tonic solutions.* But no further experimental contributions have appeared toward the solution of the question whether—and under just what conditions—elevation of temperature can in itself initiate the development of unfertilized eggs. The a prior: probability that such would be found to be the case must have seemed strong when the high temperature-coefiicient of the acceleration of chem- ical processes was considered: a five or sixfold acceleration of at least certain of the reactions occurring in the egg-substance would ‘Delage: Comptes rendus, cxxxiii, p. 348, 1901; Archives de zoologie expérimentale et générale, 3me Sér., ix, p. 285, 1gor. 5 Mathews: American Journal of Physiology, vi, p. 142, 1902. 6 J. Loeb: University of California Publications, Physiology, vol. ii, p. 158, 1905. 7 J. Loeb: Biochemische Zeitschrift, vol. 1, p. 183, 1906. Artificral Parthenogenests in Starfish Eggs Biel follow warming to 35° or 40°; and a fundamental change in the properties of the system and possibly a removal of the conditions impeding spontaneous development might reasonably be expected to result from such treatment. The failure of investigators in this country to obtain partheno- genesis by elevation of temperature appears the less accountable since Delage’s descriptions are at least sufficiently definite to have suggested a procedure quite different from the one which was actually employed and proved ineffectual. “Thus Delage writes® “La température peut, a elle seul, surtout appliquée brusquement a un stade critique, dont il va étre question, déterminer la parthéno- génése chez Asterias.”’ This critical stage is described as the time (approximately) at which the nuclear membrane of the ger- minal vesicle disintegrates allowing the nuclear contents to enter the cytoplasm; this event determines the time at which “merogonic”’ fertilization becomes possible, and also artificial parthenogenesis by heat: “at this moment the eggs of Asterias can be made to develop parthenogenetically by simple immersion in water warmed to 30° to 33°."" The lack of exactitude in this description con- sists chiefly in the failure to assign any definite limit of time to the action of the warm sea-water. As will be seen below, this is a matter of importance, since too long and too brief exposures alike fail to produce the desired effect and lead simply to abnormal changes resulting in the early disintegration of the egg. It is clear, however, that the eggs in Delage’s experiments were warmed for only a short period; in fact, he recommends placing the eggs in warm sea-water contained in small vessels (cuvettes) which may be rapidly cooled in running water.!? In Greeley’s experi- ments the eggs (1) were allowed to mature—a necessary condition for the production of parthenogenesis by cold, action of acids, agitation, or hypertonic solutions, but one which precludes the possibility of development by simple warming (as will be seen below); and (2) were exposed to the high temperatures for periods of an hour or more; whereas exposure to 35° for 60 or 70 seconds * Delage: Comptes rendus, cxxxiii, p. 348, 1901. ® Delage: Comptes rendus, /oc. cit., p- 348. 10 Delage: Archives de zoologie expérimentale et générale, Joc. cit., p. 309. 378 Ralph 8. Lillie is sufficient, at the proper time during the maturation period, to produce development. It is not surprising that the eggs failed to develop under these conditions. The general outcome of Greeley’s own work on the influence of temperature changes on protoplasm appears to have led him to doubt the possibility of producing parthenogenesis by elevation of temperature. He had found that cold, by inducing loss of water, exercises on protoplasm an action similar to that of a hypertonic solution, which was already known to produce parthenogenesis; and it must have seemed to him scarcely possible that warmth, which affects the protoplasm in a precisely opposite manner from cold, could have the same influ- ence on the developmental process. It is also evident that his experiments on the action of high temperatures in inducing par- thenogenesis were less complete than those made with cold; evidently his studies of the influence of temperature-conditions on development were cut short while they were yet unfinished. II EXPERIMENTAL My own experiments were begun in the summer of 1906, at a time when | was unaware that Delage had succeeded in producing development by this means. ‘The idea with which the study was begun was that possibly a slight change in the aggregation-state of certain of the protoplasmic colloids might be a determining condi- tion of development, and that such a change might be induced by a momentary heating of the eggs. Heat coagulation produced by momentary heating followed by rapid cooling was, according to Corin and Ansiaux, a reversible process." Such a slight and re- versible coagulation might conceivably without injuring the egg so change the state of the egg substance as to cause development to be resumed. It soon became evident, however, that even transitory exposure to temperatures of 45° to 50°, the lowest at which heat coagulation could be expected, was rapidly injurious, inducing breakdown of the eggs without any developmental changes. On the other hand, brief exposure to temperatures of 35° to 38°—in 4 Corin and Ansiaux: Bulletin de l’académie royale de Belgique, xxi, p. 345, 1891. The results of Corin and Ansiaux have since been rendered doubtful by Pauli: Beitrage zur chemischen Physiologie und Pathologie, x, p. 53, 1907. Artificial Parthenogenesis in Starfish Eggs 379 general the optimum for enzyme action—gave extremely promising indications. ‘The remainder of the investigation was then devoted to determining the influence of such temperatures acting for vari- ous brief periods. In the following experiments the eggs were exposed for brief periods (varying from a fewsecondsto several minutes) to the action of sea-water previously warmed to a definite temperature. ‘The procedure employed is as follows: the eggs are transferred at a known period after removal from the animal to a small beaker in which a thermometer is placed; sea-water at a temperature slightly above that selected for the particular experiment (e. g., 35°) is then added rapidly to the small beaker in quantity sufficient to bring the temperature to the desired point; this temperature is maintained constant during the definite time-period of the experi- ment by partly immersing the small beaker, whenever necessary, in a larger vessel of water at somewhat higher than the experi- mental temperature. After the lapse of the selected time-period (e. g., 70 seconds) the contents of the small beaker are suddenly transferred to a large volume of sea-water at normal temperature contained in a finger-bowl. The temperature of the eggs is thus suddenly reduced again to the normal. It may safely be assumed, when one considers the small volume of each egg and the correspondingly large surface for thermal interchange with the medium, that during at least the greater part of the period of immersion in the warm sea-water the eggs have themselves been at the same temperature as the medium. ‘The agitation involved in the two transfers is unavoidable with this procedure; but at the stages with which I have worked—mostly early maturation— mechanical shock is in itself ineffective in causing development. Mere transfer from one dish to another produces no visible result. The effects observed are therefore to be ascribed wholly, or at least inchief part, to the change in the thermal conditions prevailing in the egg-system. Experiments with Arbacia Eggs The results with sea-urchin eggs have been almost entirely negative so far as concerns production of development by momen- 380 Ralph S. Lillte tary elevation of temperature. In the earliest experiments, eggs were exposed for a few seconds to temperatures supposedly high enough to cause partial coagulation of a portion of the colloidal constituents of the protoplasm. “Temperatures of 45°, 50°, 55°, and 60° were allowed to act for periods ranging from 5 to 60 sec- onds. No noteworthy changes followed such treatment; swelling and disintegration resulted from exposure for even brief periods to the higher temperatures. A few eggs showed membranes similar to fertilization membranes after exposure to 45° for a few seconds, and occasionally some cleavages were found. The great majority of eggs so treated died without showing any developmental change. Treatment that resulted favorably with Asterias eggs also gave imperfect or negative results with Arbacia. Eggs were'exposed to 35°. 37-5°, and 40°, for periods ranging from five seconds to two minutes. In the most favorable experiments a few eggs showed membranes and irregular cleavages; but development never proceeded beyond a stage of a few cells, and the great major- ity of eggs always remained apparently unaffected. I have also attempted to induce cleavage in unfertilized Arbacia eggs after the artificial production of a fertilization membrane by the method introduced by Loeb, viz: treatment for one to two minutes with a mixture of 3 cc. 3, acetic acid and 50 cc. sea-water. Eggs so treated become, as in the case of Strongylocentrotus investigated so thoroughly by Loeb, far more susceptible to the develop- ment-inducing action of hypertonic sea-water; but the results after warming to 35° for periods of 20, 30, 40, 60 and go seconds, within 10 to 15 minutes after membrane-formation, were in no observable respect different from those obtained with the same eggs after simple treatment with acidulated sea-water without warming. A certain proportion of such eggs always undergo cleavage, usually irregular, but development rarely proceeds farther than an early stage ofafewcells. — A striking phenomenon, which I have frequently observed in sea-urchin eggs treated in the above manner with acidulated sea- water, seems entitled to special mention here, namely, the appear- ance of active amoeboid movements of the egg-protoplasm, at times surprisingly energetic in character. “he movement appears most Artificial Parthenogenests in Starfish Eggs 381 active about three or four hours after treatment with the acid- ulated sea-water. The following record will illustrate: July 15, 1907, 12.37 p.m. Unfertilized sea-urchin eggs were placed in a mixture of 50 cc. sea-water + 3c. N. acetic acid; one portion (A) was transferred to normal sea-water after one minute, a second (B) after 1m. 30s. At 4:30 p. m., lot A showed numerous irregularly shaped eggs in which active amoeboid movement was in progress. In many eggs the movement was so energetic that the actual contractions of the cell-surface and the protrusion of pseudopodia were plainly visible; many even exhibited an active crawling or squirming movement,suggestive of sluggish muscular contractions. In many eggs small portions of the surface protoplasm were constricted off—small beadlike protuber- ances like polar bodies being especially numerous. Transitions between irregular ameeboid masses and distinct though irregular cleavage stages were not uncommon; the latter also showed continual and active changes of form. Lot B showed essentially similar conditions. The temperature of the water in the dishes was 25°. This observation seems interesting on account of the unusually active nature of the amceboid movements. “The assumption of irregular amoeboid forms by various eggs is familiar to most experimentalists,” and is especially frequent in starfish eggs. But active crawling movements of the above kind have, so far as | am aware, not hitherto been described in these eggs. ‘The theoretical interest of the phenomenon consists chiefly in the very clear indica- tion which it affords that the form-changes in cleavage are of essen- tially the same nature and due to the same conditions as are the ordinary amceboid movements of cells; these last, as may be inferred from the closeness with whichthey may be artificially simulated, are almost certainly due—at least as regards their main features—to local (possibly electrically conditioned) changes of surface tension. The above transitional condition between amceboid movement and cleavage supports strongly the view that the change of form in normal cell-division is also due to surface-tension changes, which differ from those causing amceboid movements only in the very regular and symmetrical distribution of the areas of lowered sur- face tension. Experiments with Starfish Eggs A Conditions of Formation of Fertilization-membrane Exposure to temperatures of 45° and higher caused mature star- fish eggs to become coarse and opaque within 20 minutes or less. ” Especially energetic amoeboid movements are seen in abnormally developing parthenogenetic eggs of Chetopterus; cf. F. R. Lillie, Archiv f. Entwicklungsmechanik, xiv, p. 487, 1902. 82 Ralph S. Lille Oo No membrane was formed. In one series of experiments, eggs in early maturation stages (at which time membranes are most readily formed) formed in some instances membranes on exposure to 45° for 15 seconds; exposure to the same temperature for 30 seconds was followed by breakdown without membrane-formation. Temperatures of 45° and higher are thus rapidly destructive, as in the case of sea-urchin eggs; but very brief exposures may produce some of the effects (as membrane-formation) of more favorable conditions. Temperatures of 40° and lower were then tried. The earliest visible effect of brief warming at such temperatures is the forma- tion of the fertilization membrane. The production of this mem- brane appears to be associated with the removal of certain hin- drances to further development (p. 385), and accordingly it may be regarded as the first visible sign of developmental changes in the egg. The structure is produced with remarkable ease by momentary exposure of eggs to the action of warm sea-water; yet it is significant that temperatures above a certain maximum (ca, 45°), acting for more than a few seconds, fail to cause its produc- tion. Apparently some ferment-action, rather than the direct effect of the heat, is concerned. It also fails to be produced at 30° unless possibly the exposure is very prolonged. I have made few observations with temperatures lower than 35°. The temperature relations of this phenomenon ought perhaps to be more thoroughly investigated. The following table summarizes the results of three series of experiments covering a considerable range of temperatures. “They illustrate very typically some of the conditions of membrane-pro- duction in starfish eggs. Artificial Parthenogenesis in Starfish Eggs TABLE I Series I. Fuly 30, 1906 383 | RESULT Time of | Temperature exposure | degrees A Eggs warmed before separa~| B Same eggs warmed four hours eae tion of first polar bodies after removal from animal | 30 15 | No membranes formed 30 | No membranes formed | 35 15 | No membranes | No membranes formed 30 | Membranes in more than half A few membranes 40 15 All form membranes Most form membranes 3° | All form membranes Almost all form membranes 45 15 All form membranes Almost all form membranes 30 None form membranes; eggs soon No membranes; eggs soon dis- disintegrate integrate 50 15 No membranes; early disintegra- | tion 30 | No membranes; early disintegra- tion Series II. August 1, 1906 RESULT Temperature = = = degrees Expose A Eggs warmed during mat- B Warmed 24 hours after uration process completion of maturation 33 I5s. No membranes | Practically no membranes 305. No membranes Practically no membranes 6os. Almost all form membranes A few membranes 2m. Almost all form membranes A few membranes 35 15s. A few imperfect membranes Considerable number membranes 305. Most form membranes Practically all form membranes 1and2m. | All form membranes All form membranes 37-5 5s. A few imperfect membranes A few membranes 15s. Practically all form membranes| Practically all form membranes 30S. All form membranes Practically all form membranes Im. All form membranes Practically all form membranes 40 5s. Practically all form membranes 15 and 30s. All form membranes and 1m. 384 Ralph S. Lillie TABLE I—Continued Series III. August 6,1906. Eggs warmed during maturation process Temperature | Exposure RESULT degrees | 33 30 and 60 s. | Nomembranes 2m. | Fair number of membranes 34 | 305s. | No membranes 1 and2m. All form membranes 35 30s. Most form membranes I andzm. All form membranes 36 30s. Almost all form membranes 1and2m._ | All form membranes 37 | 15, 30, 6os. | All form membranes 38 | 15, 30, 60s, | All form membranes In general the above observations may be regarded as typical, though I have found some variability in the readiness with which eggs from different animals form membranes. But with star- fish eggs in the early maturation period membrane-formation rarely or never fails if eggs are exposed to temperatures between 33° and 40° for the periods indicated as optimal in the above table. The result is remarkably constant, even if the subsequent cleavage and development should prove abnormal or should altogether fail. The facility with which the membrane is produced varies also in eggs from the same animal at different intervals after removal; in general the early maturation stages, before the first polar body is separated, are most favorable; after the completion of matura- tion, membrane-production is less regular and constant, and more prolonged exposures to the high temperature are necessary. This change is possibly to be correlated with the change in suscepti- bility to parthenogenetic development under this form of treatment, which also diminishes after maturation is completed, as [ shall describe later. The minimum time of exposure necessary for membrane-pro- duction is shown by the above experiments to decrease rapidly with rise of temperature until a certain limit is reached. At 33° expo- sure must be prolonged to two minutes: at 34° the minimum lies somewhere between 30 and 60 seconds; at 35° between 15 and 30 Artificial Parthenogenesis in Starfish Eggs 385 seconds; at 37.5° between 5 and 15 seconds, and at 40° momen- tary exposure (5 seconds) produces membranes in practically all eggs. [hese temperature-relations point to an underlying process hae undergoes unusually rapid acceleration with rise of tempera- ture, until a certain optimum is reached (apparently in the neigh- borhood of 40°), after which heat acts unfavorably. Exposure to 45° for 30 seconds fails, as seen above, to produce membranes and acts destructively on the eggs, although briefer exposure (15 seconds) may be effective. The actual separation of the membrane may be readily studied. Within to to 15 minutes after return to normal sea-water it appears as a wavy or crenated layer adhering closely to the egg-surface; this layer gradually detaches itself as the sea-water enters he space next the cell-surface, and with the resulting distension the inequal- ities disappear; after 20 to 25 minutes (at 20° to 22°) the membrane is uniform and normal in appearance, though still very near the egg-surface. The process may be characterized as secretory in nature, and it appears to be dependent on a partial solution of the superficial lipoid layer of the egg; this is indicated by its ready production through the action of ‘the various fatty acids and fat- solvents. The above temperature-relations appear to indicate, in the case of production by warming, a dependence on some enzyme action. Ifa simple solution of certain substances at higher temperatures were the determining condition, the high tempera- ture-coefhicient of acceleration, as well as the failure of tempera- tures above 45° so to act, would be unintelligible. On the other hand, the assumption of dependence on some process accelerated by an enzyme with an optimum temperature of 38° to 40°, and rapidly destroyed at 45°, would account for the above relations. Certain hydrolytic cleavages may be concerned, possibly a saponi- fication resulting in a partial solution of the surface layer; the production of the same effect by the action of fat-solvents or alkalis becomes readily intelligible on such an assumption. An important significance has been ascribed by Loeb to the process of membrane-production in sea-urchin eggs. After mem- brane-formation, however induced, the condition of the egg is altered in such a manner that relatively brief exposure to hyper- 386 Ralph 8. Lillie tonic or hyperalkaline solutions is sufficient to produce normal development.** Even without such after-treatment, eggs in which membranes have been produced frequently cleave and under cer- tain conditions may form blastulz; usually, however, such eggs undergo breakdown or cytolysis within a few hours. Since this aterare as well as the cleav age, 1S dependent on the presence of free oxygen, the inference is drawn that in some manner, possibly by removal of anticatalytic substances, membrane-formation leads to an acceleration of oxidation processes in the egg; these if properly directed i fertilization—lead to cell division and development; otherwise they result in the destruction of the egg. Membrane-formation has thus an important significance in development. My own observations on the starfish egg in some respects sup- port this conclusion, though they can scarcely be said to do so uniformly. ‘That the process of membrane-formation is not essen- tial to cleavage has been known for some time; Loeb’s early studies in artificial parthenogenesis supply instances of cleavage without formation of fertilization membranes, and he cites other instances in a later paper.* It is also possible artificially to suppress mem- brane-formation without destroying the possibility of cleavage in the following manner: Eggs were placed I5 minutes after renova in KCN solution in sea-water, and remained here 20 OUI. they were then washed in normal sea-water and warmed to 35° for 70 seconds; these eggs formed no membranes although a con- siderable proportion underwent irregular cleavage. On another occasion the same suppression of membrane-formation without pre- vention of cleavage was observed in eggs exposed to 34, KCN for only two hours. Alchough cleavage is thus to a certain degree independent of membrane-formation, nevertheless normal cleavage and development certainly do appear to be facilitated by the separa- tion of the membrane. In the above cited experiments develop- ment stopped short at an early stage, and | have never found eggs to develop to an advanced stage under this form of treatment without the formation of amembrane. On the other hand, when- 18 Loeb: Joc. cit., also Archiv fiir die gesammte Physiologie, cxviii, pp. 181 and 572, 1907. 4 Loeb: University of California Publications, Physiology, vol. ii, p. 153, 1905. Artificial Parthenogenests in Starfish Eges 387 ever mature eggs are treated in such a way as to form fertilization membranes—whatever method is used—a certain proportion are always found to undergo cleavage. Another observation that [ have frequently made appears to favor the idea that there is a cor- relation between membrane-formation and the acceleration of oxi- dation processes in the egg. I have found uniformly that the coagu- lation and disintegration which follow when mature eggs are allowed to remain for some hours in normal oxygen-containing sea-water, occur much more rapidly in eggs that have formed membranes than in those that remain without this structure. “Thus, warming eggs during early maturation to 35° for 25 or 30 seconds induces membrane-formation in a large proportion—usually about one- half—but not in all of the eggs; practically all of the eggs so treated die at an early stage; if theyare examined after 18 hours, those with membranes are invariably found to be in an advanced state of disintegration, the entire space enclosed by the membrane _ being filled with a mass of loose granular detritus; those without mem- branes, on the other hand, although coagulated and opaque, are sull compact and undisintegrated. ‘The same contrast between eggs with and without membranes in the rate and character of the disintegration is seen when the membranes are formed by the action of ether or a fatty acid. ‘This result, which I have found with perfect uniformity throughout the present investigation, shows that eggs which have formed membranes, yet without under- going normal development, exhibit less resistance to the disinte- grative action of the post-maturation oxidative changes than do those lacking these formations. It is possible that the greater cytoplasmic activity of the eggs with membranes (as shown by the production of pseudopodia and the irregular cleavages and other form-changes) may facilitate the disintegrative process; the effect may also conceivably be dependent, at least in part, on simple mechanical conditions: the change in the closely adhering surface- layer of the unaltered egg, due to the removal of the membrane- forming substance, would probably facilitate the action of any disintegrative agency. One might suggest that the mechanical resistance to surface-changes, including cleavage, is lessened by the formation of a membrane, and that the significance of the 388 Ralph 8. Lillie process in facilitating developmental changes may possibly lie here. The membrane is readily formed by brief exposure, during or after the maturation stage, to the action of sea-water containing xylol or ether; and such eggs show the typical irregular form- changes and cleavages; | have however not yet obtained free- swimming blastule from eggs thus treated. ‘Treatment for one or two minutes with a solution of 3 cc. 34; acetic acid in 50 cc. sea- water produces perfect membranes, and | have frequently obtained a small proportion of swimming larvz from eggs so treated. The effect must be regarded as due to the lipolytic action of the fatty acid and not as a general effect of acidity (or increased concentra- tion of hydrogen ions) since mineral acids—H,SO, and HNO,— used similarly fail to produce the least sign of a membrane." The ability of mature eggs to form membranes as a result of momentary warming shows a certain periodical variation, as will be shown in more detail later (cf. pp. 400, 403). In general the disso- lution of the germinal vesicle is an important condition, although immature eggs may form perfectly typical membranes under certain conditions (p.407). Again,as already shown, membrane-formation by heating becomes more difhcult after maturation is complete. On the other hand, treatment with a fatty acid appears to produce membranes with equal readiness at any time after maturation has begun. Thus I have subjected successive portions of a single lot of eggs to the action of the above acetic acid solution at 10 minute intervals throughout the entire course of maturation (until the separation of the second polar body) and again an hour later, without finding any decided difference in effect at the different periods; a small proportion of blastula was obtained in every one of the ten experiments of the series except the first (treated 10 minutes after removal from animal). The largest proportion of larvze was obtained from eggs treated previously to the separation 18 Compare Loeb: Joc. cit., and Dynamics of Living Matter, 1906, p. 172 16 Loeb (Joc. cit., cf. also Dynamics of Living Matter, p. 170) found the same difference between fatty and mineral acids. Lefevre, however, finds that in Thalassema mineral acids produce membranes with the same readiness as do fatty acids. Here apparently some other action than the directly lipolytic is involved. Cf. Lefevre: Journal of Experimental Zodlogy, vol. iv, p. 106, 1907. Artificial Parthenogenesis in Starfish Eggs 389 of the first polar body; still, no such well-defined periodicity was found as with the experiments on the effects of warming (pp. 396, et seq.) The appearances indicate a difference in the conditions of the membrane-forming process—the acid acting by a simple lipolytic action on the surface layer, while the effect of heating depends on acceleration of an enzyme action, as already suggested. Variations in the quantity or in the condition of the enzyme would affect the results of warming without altering those due to the simple action of a fat solvent. Later I shall give some further observations on membrane- formation in starfish eggs. The process certainly seems to be correlated with a change in the developmental capabilities of the egg. It does not however necessarily lead to an acceleration of the oxidations in the egg, as is shown by the fact that immature eggs in which membranes have been formed show no increased disposition to undergo the typical oxidative coagulation or cyto- lysis (p. 408); yet under certain conditions (after maturation has begun) such an accelerated oxidation does seem to result and to constitute an important condition of development, as already indicated. ‘The experiments to be described later show, however, that only a small part of the effects of momentary warming can thus be accounted for. In the starfish egg, in fact, repression rather than acceleration of oxidations seems to be an important condition of the initiation of the developmental process, although this latter once begun naturally requires free oxygen for its con- tinuance (p. 413, et seq.) B- Development of Eggs after Momentary Warming Membrane-formation is followed after a more or less prolonged interval by a series of form-changes in the egg; these under favor- able conditions take the form of regular cleavage. It must be regarded as significant that the most manifold and irregular changes of form may occur, with all gradations between the pro- trusion of pseudopodia and the assumption of various irregular uncleaved forms or the production of irregular and unequal cleay- ages and fragmentations on the one hand, and the normal process 390 Ralph S. Lillie of typically regular and equal cleavage on the other. The irregu- larities are extremely various, and it is difficult to assign any def- nite conditions to the production of any particular kind. They seem largely due to changes occurring in the cytoplasm independ- ently of the nucleus; in other words, there is frequently a lack of correlation between nuclear and cytoplasmic activities in the warmed eggs; certain processes are initiated in both, sometimes leading to nuclear division independently of cytoplasmic division, at other times to the apparently independent assumption of irregular forms on the part of the cytoplasm, with the production of irregular pseudopodia, usually followed by subdivision of the cytoplasm into unequal cleavage cells or still smaller enucleate fragments. Such fragmentation is very typical of eggs that have been warmed for too prolonged periods; the formation of small bead-like pro- tuberances which then separate from the rest of the cell-body is an especially frequent phenomenon. ‘These conditions as a rule reach their height about three or four hours after warming, at a time when the first cleavages usually begin to appear in regularly dividing eggs. The production of protoplasmic processes at times shows remark- able peculiarities, particularly in eggs derived from animals late in the season or otherwise unfavorable. ‘The proportion of irregu- lar form-changes is also greater in eggs warmed after maturation is complete (p. 402). A slightly prolonged warming often leads to the production of numerous long slender close-set pseudo- podia of clear protoplasm, of a uniform length sometimes equal to that of the egg-radius, imparting a prickly or radiating appear- ance to the entire structure; irregular fusions may take place between these processes as in the pseudopodia of Foraminifera.‘7 These cytoplasmic activities seem to have little directly to do with nuclear influence; separated enucleate portions of protoplasm may also undergo irregular form-change or subdivide still further. Other instances of specific change of form in enucleate portions of eggs have been described by several observers. It seems clear that the 17 Such conditions seem frequent in abnormally developing parthenogenetic eggs; compare especially the accounts of F. R. Lillie for Chetopterus, /oc. cit., p. 487; also of Lefevre for Thalassema, /oc. cit., p- 109. Artificial Parthenogenesis in Starfish Eggs 391 cytoplasm possesses a certain formative capacity of its own;'8 this under the above abnormal conditions may g give rise to structures having very definite peculiarities, but quite foreign to the normal development. Under favorable conditions a large proportion of eggs undergo regular cleavage and development. ‘The following series rable IL) will illustrate; he eggs (all from a single lot) were exposed to temperatures of 35°, 36°, 37° and 38°, for varying brief periods during the early maturation period (between 20 and 45 minutes after removal from the animal, before separation of the first polar body). The susceptibility varies somewhat within this period; but, as will be shown later, warming may produce develop- ment at any time between the dissolution of the germinal vesicle and the formation of the first polar body (after which time it becomes increasingly difficult to incite development by this means). Within at least the greater part of the period of exposure covered by this series the susceptibility to development by warming varies relatively slightly, and the condition of the eggs may be regarded as essentially uniform throughout. Later experiments will be described in which the variation in susceptibility at different periods during maturation is itself made the subject of study (cf. pp. 396, et seq.) Eggs from the same lot were treated in a precisely similar manner on the afternoon of the same day, from 2.36 to 3.06 p.m. All had matured in the typical manner. The result was quite different. Membrane-formation was less uniform and required a more prolonged exposure to the respective temperatures, and although in favorable experiments a considerable proportion of eggs underwent cleavage, mostly irregular, not a single swimming larva was obtained. ‘This kind of experience has been aoe Ihave never succeeded, afterthe completion of maturation, in bring- ing unfertilized eggs to the free-swimming stage. “The eggs invari- ably either fail to cleave, or cleave more or less irregularly, usually after undergoing rregular form-changes, and die at an early stage. 18 Compare Wilson’s account of the phenomena in the isolated enucleated polar lobe of Dentalium; cf. also the references in his paper to analogous phenomena in echinoderm eggs. Wilson: Journal of Experimental ZoGlogy, vol. 1, p. §3, 1904. 39 Bs Ralph S. Lillte TABLE II August 8, 1906. Eggs were removed from the animal at 10:15 a.m., and treated as follows: Tempera- ture of sea-water in the dishes, 23° Interval ntact Temperature and removal from er time of exposure RESULT minutes (ca.) | deg. sec. | 20 | 35 30 | No membranes formed. No cleavage 21 35. 40 | Ca. 50 per cent form membranes; many cleavages, mostly | irregular; no blastule obtained 22 35 50 All form membranes; numerous regular cleavages; a few blastule obtained 24 35 60 | All form membranes. Cleavage largely regular; blast- ule more numerous than in Experiment 3 All form membranes. Cleavage more regular and rapid 25) 35,5 792 | than in Experiment 4; good proportion gastrule 27 35 80 | Similar to 5; good proportion blastule and gastrule 29 | 36 15 | Practically no membranes (one seen). No cleavage 29.5 35 20 | Allform membranes. Mostly irregular cleavage. No larve 30 | 36-30 Similar to 8 31 36 40 | More favorable; large proportion regular cleavages and a | fair number of blastule and gastrule 32 36 50 Somewhat more favorable than 10; a considerable number of blastule and gastrule 19 20 | 36 a7; 10 A few membranes formed; no cleavage found 37 37 15 | Most form membranes: cleavage mostly irregular; no blastule obtained 38 37 20 All form membranes; mostly irregular cleavages, a few regular; no larve found 39 | 37 30 More favorable than 14; good proportion regular cleavages; large number blastule and gastrule obtained, and a few good Bipinnarie 40 37 40 Fewer regular cleavages than in 15; a fair number of larve obtained 41 38 5 Hardly any membranes (2 or 3 seen). No cleavages found 41 38 10 Almost all form membranes. Cleavage irregular or incom- plete. No larve 42 38 15 All form membranes and cleavage is less irregular than in 18. Some regular cleavages, and a few blastule and gastrule obtained 43 38 20 Large proportion of irregular cleavages and a fair propor- tion give swimming blastule and gastrule; a few reach Bipinnaria stage Artificial Parthenogenesis in Starfish Eggs 393 The time of early maturation (before the separation of the first polar body),is apparently a critical period for the production of this type of parthenogenetic development. The same has been found true by Delage." A similar series of Spa vata on July 24, 1907, with the three temperatures 35°, 30° and 37° and a somewhat different range of exposures gave in general the same result, summarized in Table TET. TABLE IT Fuly 24,1907. Eggs were removed at 2:15 p.m., and treated as follows Interval after | Temperature and | removal exposure eS | minutes (ca.) | deg. sec. To] 30 | 35 60 | Good proportion of regular cleavages, and fair number blastula and gastrule zy 31 35 7° | Cleavage more rapid and regular than in 1; large number | active larva obtained yi 33 ie te) In general like Experiment 2: somewhat less favorable; | numerous larva 4 34 | 35 go | Smaller proportion of regular cleavages and fewer larve than in Experiments 2 and 3 5 37 | 36 50 All form membranes, but cleavage is mostly irregular; no larve 6 38 36 40 Larger proportion cleavages than in 5, largely regular. No | larvee 7 40 36 50 Fair proportion of regular cleavages, fair number of blas- tule and gastrule obtained 8 | 42 | 36 = 60 Cleavage less regular and slower than in 7; good many | | blastule and gastrule y | 45 37 20 All maturing eggs form membranes, relatively few cleav- ages; no larve obtained 10 46 37 30 Large proportion of regular cleavages; fair number of blastule and gastrule obtained II 47 37 40 Fewer cleavages, slower and less regular than in Experi- | ment 10; eggs mostly adopt irregular shapes without cleaving: a few larve 12 | 48 37 50 Eggs form membranes and adopt irregular shapes with | slender pseudopodia; no cleavages found. No larve 19 Delage: Joc. cit. 394 Ralph S. Lillte From the above experiments it appears that the optimum time of exposure to the high temperatures is shorter the higher the tem- perature employed. In the above two series the best results were obtained at 35° with 70 seconds exposure, at 36° with 40 or 50 seconds, at 37° with 30 seconds, and at 38° with 20 seconds. In general the same relative favorability of different periods of expo- sure at different temperatures was found in several other similar series of experiments. The decrease in the optimum time of exposure with a given increase in the temperature is somewhat surprisingly rapid, the difference of three degrees between 35° and 38° reducing the optimum exposure from 70 to about 20 seconds. If the process in which the initiation of development depends is a purely chemical one, this would indicate an extraordinarily high temperature-coefhicient of acceleration. The conditions in a heterogeneous system like protoplasm must, however, be recog- nized as peculiar; rise of temperature, in addition to accelerating the specific chemical processes (usually about threefold for a rise of 10°), has a certain effect (which under some conditions may be very considerable) in altering the surface of interaction between the colloidal bodies concerned and the other chemical substances reacting with them; increased subdivision of the colloidal particles following a rise of temperature would in itself accelerate the reac- tion; and the total acceleration would be measured by the product of this increase in the surface of interaction into the specific acceler- ation of the process itself through the rise of temperature; this com- pound value might exceed many times (as apparently in the case under consideration) the simple temperature-coefhicient of acceler- ation of the purely chemical process. “The results with membrane- production also indicate a very high temperature-coefhcient. So far as regards development I have as yet made no special endeavor to determine the optimum periods of exposure at temperatures above and below those cited. ‘The favorable periods for tempera- tures of 39° and 40° would undoubtedly be found very short, while at 34° and 33° exposures would be more prolonged. “Temperatures so low as 30° would in all likelihood be found effective with sufh- cient time of exposure, as Delage’s*® results indicate. 20 Delage: Joc. cit., p. 307+ Artificial Partheno genesis in Starfish Eggs 395 The different temperatures do not however seem equally favor- able, and more larve are obtained at 35° and 36° than at 37° and 38°; in other words, with the higher temperatures warming seems more likely to produce abnormal results. My experience has been that temperatures of 35°, acting for somewhat longer than one minute, have usually given the best results. ‘This is illustrated by the foregoing tables and again by the following four experiments. In this series eggs from a single animal were exposed to 35° for 60, 70, 80 and go seconds, respectively, at a period of 30 to 35 minutes after removal from the animal. The results were as follows: 1 | 35° 60s. | Agood proportion reach gastrula stage and a few form early Bipinnarie 2 | 35° 70s. | A large proportion reach advanced gastrula stage and a considerable number form early Bipinnarie 3 35° 80s. | Somewhat less favorable than Experiment 2; a considerable number form | advanced gastrula, and a few early Bipinnarie 4 35° gos. | Fewer larve obtained than in the above three experiments and none reach early | Bipinnaria stage An exposure of 70 seconds to 35° again proves most favorable. In all of the following experiments on the determination of the period of greatest susceptibility to this treatment I have accordingly employed uniformly an exposure to 35° for 70 seconds; such treat- ment if applied at a favorable period (best at some little time before the separation of the first polar body) almost invariably results in the production of a good and sometimes a high proportion of larva. There appears however to be some variation in the opti- mum period of exposure to a given temperature in eggs from differ- ent animals and at different seasons of the year. “Thus on Septem- ber 6, 1906, eggs were treated as follows during early maturation (17 to 37 minutes after removal): 35° for 70, 75, 80 and 85 seconds; 36° for 40, 45, 50, 55 and 60 seconds; 37° for 20, 25, 30 and 35 seconds. Cleavage was most nearly normal and a certain rather small proportion of larva was obtained with 35° for 85 seconds, 36° for 50, 55 and 60 seconds (the last best), and 37° for 30 and 35 seconds (both about equally good). With the other exposures the cleavage was slower and less regular and no swimming larve 396 Ralph 8. Lillie resulted. Another portion of the same eggs was similarly treated in the afternoon about three hours after completion of maturation; a large proportion failed to form membranes, cleavage was either irregular or failed altogether, and not a single larva resulted. This series was less favorable than those tabulated above, and the optimum exposures were considerably more prolonged. The dif- ference in the time of year may be a factor of importance; this, however, can only be determined by further experiment. On the whole, when normal eggs are used a given temperature has a quite well defined optimum period of exposure which can be de- termined with considerable accuracy. Since the temperature- coefficient of acceleration of a given process may afford valuable indications as to its essential nature, a more exact redetermi- nation of the optimal periods of exposure through a greater range of temperatures may yield valuable results. I hope at some future time to make more extended and exact determina- tions of the above and similar relations. Susceptibility to Warming at Different Periods During Maturation The foregoing experiments had shown that momentary warming has a favorable action in inciting parthenogenetic development during, but not after, the period of maturation. It remained to determine more precisely the limits of the period of susceptibility to this form of treatment, and the variation in favorability within this period. For this purpose in each of the series of experiments tabulated in Table IV the eggs of a single female were employed; successive portions of these were warmed to 35° for 70 seconds, beginning about five minutes after removal (at a time when the germinal vesicle had undergone no visible alteration), and thereafter at regular five minute intervals until the separation of the first polar body had taken place. The condition of the eggs at the time of warming was observed in each case in a “control” portion kept under the microscope throughout the entire period of the series. With good eggs from a single female the maturation proc- ess progresses with almost uniform velocity in all eggs; the numer- ous eggs of each portion may thus be considered practically uni- Artificial Parthenogenesis in Starfish Eggs 397 form so far as regards the stage at which the treatment was applied; all portions were treated exactly alike. The results indicate the existence of a fairly well defined period of maximum susceptibility, lasting for a certain period preceding the completion of the first maturation division. favorable. Thereafter conditions become rapidly less The results of three satisfactory series of experiments are summarized in the following table: TABLE IV The temperature of exposure was 35° and the time 70 seconds in all cases. The condition of the eggs at the time of exposure in each experiment is indicated by the italicized portion enclosed in parentheses. The time itself (interval after removal of eggs from animal) is given in the second column Time after DENS No. | removal 7 = | minutes | Series I | Series IT Series III | (ca.) | July 31, 1907 August 7, 1907 August 12, 1907 I | 5 | (Germinal vesicle intact and| (Germinal vesicle intact) (Germinal vesicle intact) | unaltered) Practically all) Practically all remain) Almost all remain im- | eggs remain immature; immature; no develop-| mature with intact ger- | | no development ment | minal vesicle | | | 2 | 10 | (Germinal vesicle still un (Germinal vesicle still in-| (Vesicle still intact) A changed in most) Most) tact) Most remain im-| fair proportion mature | eggs remain immature; a) mature; a few mature) and develop. Some form | few form membranes and) but none develop larve, " mostly feeble | develop; a few feeble) blastule | blastule obtained | | | 3 | 15 | (Outline of vesicle ecrmine (Outline of vesicle becoming) (Germinal vesicle becoming | | indistinct in most eges)) indistinct in a fair pro-| indistinct in a fair pro- Most form membranes) portion) A few eggs| portion) A large pro- and develop. Many ac-| mature and cleave. No) portion mature and de- tive blastule formed | larvae obtained velop; a fair proportion | | form blastule and gas- | trule 4 20 | (Germinal vesicle is indis-| (Vesicle disappearing in| (About one-half of the eggs A | tinct in nearly all) | larger proportion form | larve than in Experiment 3,and reachmore advanced stage (gastrule formed) about one-third of the eggs; rest remain imma- ture) Larger propor- tion cleavages than in Experiment 3; no larve are maturing, vesicle in- distinct) A fair propor- tion develop as in Ex- periment 3 Ralph S. Lillie TABLE [V—Continued No. Time after| RESULTS removal minutes Series I Series IT Series III (ca.) July 31, 1907 August 7, 1907 August 12, 1907 25 (Vesicle almost disappeared (As in Experiment 4; only (Germinal vesicle disap- in nearlyall) Morefav- one-third or so maturing) _ pearing in ca. one-half of orablethanExperiment4; Largerproportioncleave the eggs) Large num- larger proportion larve than in 4; fair number ber good blastule and blastule gastrule formed 30 (Vesicle barely distinguish (About one-thirdofthe eggs (About one-half of the eggs able in most) Similar fo} maturing) Cleavage as) maturing) Larger pro- Experiment 5 | inExperiment 5;nolar- portion of Jarve than in ve found 5 35 (Vesicle indistinct in most) (Like 6: one-third matur-| (Like6) More favorable | Numerous larve formed; ing) Cleavage rather than 6; large proportion more favorable than Ex-- more regular than in of good gastrule formed periment 6 | 65 fair proportion form blastule 40 (Region of germinal vesicle) (One-third mature; a (As before; first polar spin- | almost indistinguishable) polar spindle visible as dle visible in about one- Larger proportion cleave clear area at surface of half of the eggs) Large and more regularly than egg) A few blastule| proportion of eggs form in 7; next morning most) formed; less favorable gastrule as in 7; rather eggs in blastula or early) than7 | more favorable gastrula stage 45 (Like 8) Most eggs form (As in 8) Less favorable) (Like 8) Cleavage slower vigorous gastrule by next’ than 8; no larve and fewer larve formed morning; some reach Bi- than in 8; still very fa- pinnaria stage later. Very vorable | | | favorable culture ° Vesicle quite invisible in (Like8 andg) Unfavor- | (Like 8; polar bodies not 5 q Shs practically all eggs) Less, able; eggs take irregular) yet separated) Less fa- favorable than 9; till as) shapes; no larve vorable; cleavage more large proportion form vig- irregular and few form orous larve a few of which larve reach early Bipinnaria) stage Artificial Parthenogenesis in Starfish Eggs 399 TABLE IV—Continued Time after RESULTS No. | removal |——— : = minutes Series I Series II Series HT (ca.) July 31, 1907 August 7, 1907 August 12, 1907 II 55 (Like1o) Numerousactive) (Polar bodies not yet sepa- (No polar bodies as yet) | larve, some reaching) rated) Cleavage irregu-) Cleavage slow and more early Bipinnaria | lar; no larve formed irregular than in 10; not a single larva | 12 60 (Polar bodies not yet sepa-| (Polar bodies beginning to, (Polar bodies beginning to rated) Decidedly fewer| separate) Cleavage ir- separate) Cleavage ir- larve and less activethan| regular; no larve | regular; no larve in 11 | | 13 65 First polar bodies have sep-| (First polar bodies sepa-| (Polar bodies separated in arated in many eggs)| rated in the mature eggs.. maturing eggs) No de- Cleavage much retarded) 1. e., one-third of whole) velopment and only afew small ir-/ Similar to 12 regular blastule obtained | 14 17 (Polar bodies separated), (Like 13) Like 13; no, (Like 13) No develop- Cleavage delayed and ir-) larve ment regular; very few larvae; feeble and abnormal In the above three series a large number of eggs reached the larval stage, Series I being especially favorable. In two other similarly conducted series the eggs proved inferior, only about 10 per cent undergoing maturation. In the first of these series August 10, 1907, larva were obtained only from eggs warmed at periods corresponding to Nos. 5 and 8 of the above series; the second, August 13, proved somewhat more favorable, blastula resulting from eggs exposed at periods corresponding to Nos. Asis Os 7> 8 and g above with the optimum at Nos. 6 and 7. The suppression of maturation in eggs warmed within 5 to 10 minutes after removal also resulted in Shock series. Three other similar series—July 29, July 30 and August 3 suppression of maturation in eggs heated directly after removal. The eggs in these series were inferior and no larve were obtained; 400 Ralph S. Lillie but in the best experiment, that of July 29, the largest proportion of regular cleavages—and in general the most favorable condi- tions—was found in eggs warmed at stages corresponding to Nos. 8, 9 and 10 of the above table. In the series tabulated above the following chief uniformities are apparent: (1) Warming shortly after removal (within 5 to 10 minutes, before the germinal vesicle has undergone any apparent change) has the effect of completely preventing maturation; the germinal vesicle remains intact and the egg remains permanently in this condition until disintegration sets in. Such eggs behave in the same manner as do eggs that fail to mature from any other cause—they remain clear and apparently unaltered at a time when mature eggs have undergone complete coagulation and dis- integration. (2) Warming at any time after the germinal vesicle membrane has begun to dissolve and before the separation of the first polar body may lead to development and production of larve; the proportion of eggs that develop is at first small, but increases rapidly; in general the conditions for development steadily improve until a certain stage is reached—about 40 to 45 minutes after removal at normal summer temperature; at this time the suscepti- bility of the eggs is greatest and momentary warming is followed by development and the production of active larve in a large pro- portion. ‘Thereafter the susceptibility rapidly declines; and at the time of separation of the first polar body warming results chiefly in abnormal development and larve are rarely obtained. Warming in later periods is still less favorable. In the following two series eggs were treated as above at 10 minute intervals until after the formation of both polar bodies, and after this less fre- quently until about five hours after removal. Both series proved favorable and showed good agreement; larva were most numerous from eggs warmed 10 to 15 minutes before the separation of the first polar body; at the time of separation few or none were obtained, and thereafter conditions became progressively more unfavorable with lapse of time. After both polar bodies had separated the eggs not infrequently failed altogether to cleave or even to produce membranes—a result which agrees with those of the earlier experiments already cited. Artificial Parthenogenesis in Starfish Eggs 401 TABLE V Series I, Aug. 18, 1907. . Eggs were removed from Series II, Aug. 20, 1907. Eggs were removed at animal at 11:12 a.m., and warmed to 35° for 70 9:30 a.m., and warmed to 35° for 70 seconds after seconds at following intervals after removal. the intervals indicated Interval RESULT | Interval RESULT I Ism. (Germinal vesicle beginning 7m. (Germinal vseicle unchanged) to show signs of dissolu-| Practically all eggs remain tion) A few eggs reach 1 in immature state; no de- blastula stage |} velopment | 3 25m (Germinal vesicle indistinct | 17m (A few eggs show beginning | in large proportion of eggs) I maturation) A few feeble Large number form blas- | blastule formed tule and gastrula | 3 35m (Germinal vesicle invisible 27 m. (A small proportion of eggs in most eggs) Good pro- maturing) Considerable portion blastula and gas- | number of blastula ob- trule tained 4 45m | (First maturation spindle at 37 m. (Only 5 to 10 per cent eggs surface; no polar bodies | maturing) Fair number separated) Good propor- | of blastule and gastrule tion blastule and gastru- | le | 5 55m. | (First polar bodies separat- | 47 m. (First maturation spindle at ing) A good many eggs surface of maturing eggs) seem to have cleaved reg- More favorable than Ex- ularly but no larve were periment 4; good propor- obtained portion form blastule and | gastrule | 6 | th.gm. | (Al maturing eggs have first 57 m. (Polar bodies not yet sepa- | polar bodies) Mostly ir- rated) A fair proportion | | regular cleavages and frag- | of larve; rather fewer than | | mentation; no larve | in Experiment 5 | | } 7| th.1gm. | (Second polar bodies not yet i 1h.7m. | (First polar bodies beginning | separated) After 8 hours | | 10 separate in maturing most eggs still uncleaved; | | eggs) A few larva; less a few irregular cleavages; | favorable than 6 no further development 402 Ralph 8. Lillie TABLE V—Continued | Interval RESULT Interval RESULT 8 | th.25m. | (Second polar bodies separated 8 rh.17m. | (First polar bodies separated | in large proportion of eggs) in maturing eggs) Cleav- | Like Experiment 7; after 8 age irregular; hardly any | hours mostly uncleaved, or reach swimming stage; a | irregular in shape; some single blastula found are coagulating; no devel- opment | | 9) rh.35m. | (Al maturing eggs with two 9 | th.27m. | (Second polar bodies not yet polar bodies) After 8 hours | | formed) After 7 hours egg eggs uncleaved ; largely | irregularly cleaved or ir- without membranes and | regularly shaped and un- | in process of coagulation; | cleaved; no larve formed | no development | 10 rh.4sm. | (Like g) After 8 hours un- 10 1h.37m. | (Second polar bodies in al- | cleaved and_ irregularly most all maturing eggs) | shaped; largely without After 7 hours markedly | membranes and coagulat- | different from Experiment ing; no development | 9; most mature eggs have | | | membranes but are un- | cleaved; a few irregular | | cleavages or fragmenta- | | tions; good many coagu- | | | lating; no development II | th.58m. | After 8 hours eggs areirregu-| 11 | th.47m. | (Al maturing eggs with two | lar and uncleaved; many | | | polar bodies) Similar to 10; are without membranes | | mostly with membranes and coagulating but uncleaved and irregu- | lar; some are coagulating | | 12 | 3h.35m. | Marked difference from 11; 12) 1th.57m. | (Likerz) After 7 hours eggs | after 6 hours practically are irregular, uncleaved or all eggs are without mem- | fragmented; no develop- branes and coagulated ment 13) 4h.5m, | Similarto 12: after 5 hours 13 | 2h. 7m. | Similar to 12 eggs are without mem- branes and coagulated 14 | 4h.35m. | Similar to Experiments 11 | 14 | 2h.22m. After 7 hours eggs un- and 12. cleaved, irregular or frag- | | mented Artificial Parthenogenesis in Starfish Eggs 403 TABLE V—Continued Interval RESULT Interval RESULT rsa subs The same as above; eggs do|| 15 | 4h. 44m _ | After 4 h. 15 m. almost all not form membranes and mature eggs found coarsely coagulate within 5 hours coagulated and in process of disintegration, mostly (but not all) without mem- branes || 16} 5h. 35 m. | After 34 hours mature eggs | are in process of coagula- tion; only a few have mem- branes I find no record in my notes of the condition of the membranes in Experiments 12 to 14 of Series II. In this series membrane-formation is not so completely absent as in Series I in the stages succeeding the completion of maturation; but presumably the proportion of eggs without membranes increased steadily from Experiment 11 to Experiment 16 where only a few were formed. In both of the above series the eggs show a progressively increas- ing inability to respond to momentary warming after the maturation process is complete. The proportion of eggs that fail to form mem- branes also increases with the lapse of time in the post-maturation stages. Cleavage becomes irregular or fails altogether; the curious result also appears in the experiments made some time after the complete separation of the second polar body that the tendency to coagulation, typical of mature unfertilized eggs, is markedly accelerated. ‘This effect is conspicuous in eggs warmed at a stage of four or five hours after removal, as seen in both of the above series; the coagulative process is well advanced in such eggs within three or four hours after warming; while in mature eggs not exposed to this treatment, or warmed at anearlier stage, coagulation does not become evident until some hours later. The process, as Loeb has shown, is oxidative in nature; warming in post-matura- tion stages has thus the effect of accelerating oxidations leading to a coagulative disintegration of the egg-substance. An earlier experiment performed with another object in view shows the same result: Eggs were removed at 11:30 a.m. August 5, 1907; about 404 Ralph S. Lillie 3 h. 45 m. later they were warmed to 35° for the periods indi- cated. I 35° 60 s. (3:15 p.m.) | 3h. later (6:15) only a few eggs have membranes; most are unaltered; no cleavage 2 35° 65 s. (3:16 p.m.) | At 6:20 condition similar to Experiment 1; only a few eggs have mem- branes; these are dead and in process of coagulation s. (3:18 p.m.) | At 6:21 the mature eggs are all opaque and coagulated; only a few ww we nm ° x ° have membranes; these uniformly show the greatest disintegration 75 Ss. (3:20 p.m.) | At 6:21 condition of eggs similar to Experiment 3 5 | 35° 80s. (3:22 p.m.) | Similar to Experiments 3 and 4 In the unfertilized unwarmed control at 6:23 p.m. the eggs remained quite uncoagulated; warming has thus hastened the coagulative change in the mature eggs, and especially in those with membranes, which uniformly showed the most advanced disinte- gration, What are the conditions of this varying susceptibility to the above form of treatment at these different periods in the life of the unfertilized egg? One event, occurring shortly after the removal of the egg from the animal to normal oxygen-containing sea-water, seems of fundamental significance, viz: the dissolution of the membrane of the immature egg-nucleus or germinal vesicle. This event naturally must precede the maturation divisions that follow; but quite apart from this it seems to form the condition of a profound change in the properties of the egg-cytoplasm. Delage”! has found that enucleate egg-fragments of Asterias are insuscepti- ble to fertilization before the germinal vesicle has undergone visible change; but that very soon after its membrane has begun to show indication of dissolution, merogonic fertilization first becomes possible; a little later, when the membrane has become invisible— although the area of the former germinal vesicle may still be seen, often with nucleolus Waar the: fragments of cy toplasm are com- pletely and readily fertilizable. Tiese observations demonstrate that the essential feature of maturation, so far as the cytoplasm is concerned, is not the separation of the polar bodies, but simply the removal of the barrier between the nuclear and the cytoplasmic 21 Delage: Archives de zoologie expérimentale et générale, Sér. IIT, T. 9. p. 285, 1901. Artificial Parthenogenesis in Starfish Eggs 405 areas; they also show that the nuclear membrane acts as a semi- permeable membrane with reference to certain substances con- tained within it. The critical event, therefore, which conditions this remarkable change in the properties of the cytoplasm is, according to Delage, the passage of certain nuclear constituents (suc nucléaire) into the cytoplasmic area (Joc. cit., p. 289). These substances he suggests may either change the osmotic pressure of the cytoplasm, or may influence the rate of oxidations, or may be of the nature of particular electrolytes or enzymes. He was unable to produce by artificial means any developmental changes in such egg-fragments. The precise nature of the change induced in the egg-cytoplasm in consequence of the dissolution of the germinal vesicle is as yet unknown. The fact that the egg, if not fertilized within a few hours, readily undergoes an oxidative change involving a coagula- tion of the cytoplasmic colloids seems to point to an acceleration of oxidations in that region—due possibly, as lately suggested in an interesting paper by Mathews,” to the liberation of oxidases formerly confined to the nuclear area; if oxidases are of nuclear origin, as certain facts seem to indicate,” such a consequence would naturally follow; the periodic dissolution of the nuclear membrane in mitotic cell-division would thus have the significance of providing for the distribution of the oxidases (synthesized in the nucleus) throughout the cytoplasmic area; this would naturally result in a periodic acceleration of oxidation processes in the cell. Lyon” has in fact shown that the production of carbon dioxide by the dividing egg follows a rhythm parallel with that of the nuclear division; and Loeb* has connected these oxidations with the syn- thesis of nucleins from the compounds of the cell-protoplasm— a process which is likewise characterized by a rhythm parallel with that of the mitotic process. This general interpretation, though suggested by quite different = Mathews: American Journal of Physiology, vol. xviii, p. 94, 1907. * Cf. the references in my paper On the Oxidative Properties of the Cell-nucleus, in American Journal of Physiology, vol. vii, p. 412, 1902. * Lyon: American Journal of Physiology, vol. xi, p. 52, 1904. % Loeb: Biochemische Zeitschrift, vol. i, p. 183, and vol. ii, p. 34, 1906. 406 Ralph S. Lillie considerations, is in striking agreement with the view propounded by Conklin” some years ago in his studies of karyokinesis in the Crepidula egg. Some of his conclusions on the physiology of this process should be quoted. ‘‘The nuclear membrane appears to permit the passage of materials inward but not outward during the resting period; whereas the escape of nuclear material into the cell is brought about by the disappearance of the nuclear membrane during karyokinesis.”” In Crepidula there can be demonstrated cytologically “a very extensive interchange of material between the nucleus and the cytoplasm;” “a large part of that most characteristic nuclear substance, the chromatin, passes into the cytoplasm in the form of oxychromatin during every cell-cycle, while a relatively small part is reserved for the purpose of repro- ducing the daughter-nuclei.” This passage of nuclear material Gewreele aucleo- -proteid in nature) into the cytoplasm is re- garded as a fundamentally important condition of the subsequent game undergone by the latter. These phenomena appear to be characteristic of mitosis in general and essentially similar conditions have been described for a penance of cells. In the starfish egg by far the greater part of the chromatin is set free in the cytoplasm during the first maturation division.?’? In Chetopterus also the greater part of the germinal vesicle consists of a “residual substance” which is set free in the cytoplasm at the first maturation-division and plays an important part in the future development.’ It is natural, in view of the probable nucleo-proteid nature of at least certain enzymes, to regard the above “‘oxychromatin” or “residual material’? as consisting—at least in part—of the ferments con- cerned in the chemical processes—largely oxidative in their nature as shown clearly by the conditions in the starfish-egg—that deter- mine the later characteristic changes in the cytoplasm. The ascertained cytological facts are thus in essential harmony with the above hypothesis. Whether the change in the cytoplasm depends primarily on increased oxidations or on other conditions is scarcely decided as % Journal of the Academy of Natural Sciences of Philadelphia, second series, vol. xii, pt. 1, 1902. °7 Wilson and Mathews: Journal of Morphology, vol. x, p. 334, 1895. 28 Cf. F. R. Lillie: Journal of Experimental Zodlogy, vol. iii, p. 153, 1906. Artificial Parthenogenesis in Starfish Eggs 407 yet. My own observations agree with Delage’s and those of later observers in indicating that the dissolution of the nuclear membrane is in some way associated with a well-defined alter- ation in the capacity of the egg for further development. Mo- mentary warming previously to this event not only fails to result later in cleavage, but it has the effect of completely preventing the change in question and with it the entire maturation-proc- ess. On the other hand, as already seen, the same treatment applied at any time after the beginning of the maturation-changes (until the separation of the first polar body) may lead to develop- ment and the production of larvae. The properties of the cyto- plasm thus must undergo a profound change the nature of which remains to be determined. One normal sequence of the dissolution of the germinal vesicle is a change in the reaction of the egg-cytoplasm toward membrane forming agencies. Membrane-formation now promptly follows warming, or the entrance of a spermatozoon, or the momentary action of a fatty acid or fat solvent; while in the immature egg this structure is usually not formed under these conditions.** Yet, al- though as a rule eggs that remain permanently immature as above do not form fertilization membranes on warming, this is by no means invariably the case. I have recorded numerous instances in which momentary warming has produced perfectly normal membranes in immature unfertilized eggs. In general such eggs belonged to lots that were unfavorable as regards capacity for development, so that the membrane-production may be considered as evidence of a certain abnormality. The following instance will illustrate: in the series of August 2, 1907, cited above, in which eggs were warmed at 5-minute intervals (until the separation of the first polar body in those maturing), the majority failed to mature, and the developing mature eggs in no case reached the free swimming stage. In this series most of the permanently im- mature eggs, after subjection to the momentary warming process, formed quite typical uniform membranes indistinguishable from those found in fertilized mature eggs; this was especially true of 29 Cf. Loeb: University of California Publications, Physiology, vol. ii, p. 150, 1905. 408 Ralph S. Lillie those warmed at periods of 10 to 15 minutes after removal; after an hour (at which time most mature eggs had formed polar bodies) the proportion of immature eggs that formed membranes had declined considerably, and at later stages only a few were formed. This is not in the least an isolated observation, but 1s fairly typ- ical of what I have frequently observed; the ability of immature eggs to form membranes seems in general ‘best marked shortly after removal, and diminishes after an hour or more in sea-water. An observation made on the same lot of eggs showed that sperma- tozoa may also induce membrane-formation in immature eggs; sperm was added at 1 h. 25 m. after removal; the next morning a fair proportion of the mature eggs had formed larva; and nearly all of the immature eggs showed perfectly typical sharply defined fertilization-membranes; otherwise these eggs remained unchanged. Spermatozoa are known to enter immature starfish-eggs,*° but typically to produce no membranes. Under certain conditions however, not definitely understood (eggs “over-ripe”’ or otherwise not quite normal), membranes may be formed as just seen, either by spermatozoa or through an artificial agency. The explanation may be as follows: normally the possibility of membrane-forma- tion depends on the passage of certain substances from the nucleus to the cytoplasm, since the beginning of maturation is a prerequi- site; in the above eggs however the permeability of the germinal vesicle membrane is abnormal, so that the substances necessary to the membrane-formation, which ordinarily are unable to tra- verse the nuclear membrane, are now able to effect this passage and to enter the cytoplasm. ‘The latter then reacts to heat or the entrance of the spermatozo6on by forming a membrane in the man- ner characteristic of mature eggs. It is interesting also to note that such immature eggs show no other change in their properties; they remain clear and unaltered for prolonged periods and show no greater tendency to disinte- grate than do normal immature eggs—a fact apparently contra- dictory of Loeb’s view that the separation of the membrane in- volves an acceleration of oxidative processes in the egg. In mature 39 Wilson and Mathews: Journal of Morphology, x, p. 319, 1895. Artificial Parthenogenesis in Starfish Eggs 409 eggs, however, there is an obvious difference in the conditions; the entire contents of the germinal vesicle—not only those sub- stances that can pass the nuclear membrane—have become mingled with the cytoplasm; and in fact mature eggs differ from immature eggs in undergoing the typical disintegration much more rapidly after forming membranes, as shown above. It is quite possible that for the oxidations concerned in the post-maturative disinte- gration of the cytoplasm there is needed the presence of specific papers: derived from the nucleus—e. g., oxidases, or enzymes or proferments of some other kind, or certain activ ators—and that these substances merely find better conditions for their activity after the separation of the fertilization-membrane than before. In their absence membrane-formation would in itself effect no esential change in the condition of the cytoplasm. Membrane- formation alone is thus quite ineffective—unless accompanied by certain other and independent changes—in accelerating oxidations in the egg-cytoplasm. The effect of momentary warming in preventing the dissolu- tion of the germinal vesicle is curious and difficult to explain. The process itself, as shown by Loeb" some years ago, depends on oxidations, since it 1s prevented by acidulation of the sea-water or by depriving the eggs of free oxygen. One of his observations seem analogous to Tet one under discussion: exposure of unripe egg even temporarily (as for 15 minutes), to acidulated sea-water (100 cc. sea-water + 5 cc. *; HNO,) prevented the eggs from matur- ing after retransfer to normal sea-water. An oxidative process therefore which normally leads to the dissolution of the nuclear membrane within a few minutes after the eggs are laid, 1f checked before that time is ordinarily not resumed and the eggs remain immature. But why should temporary warming at this stage produce a similar result? The expectation would be that by such treatment the oxidations, as well as the other chemical proc- esses in the egg, would be accelerated, and that a process like maturation, dependent on oxidations, would be furthered rather than prevented. Evidently warming, during the brief period that * Loeb: Archiv fiir die gesammte Physiologie, xciil, p. 59, 1902. 410 Ralph §. Lillie normally precedes the solution of the nuclear membrane of the immature egg, in some manner inhibits the oxidations on which this change depends. Just why this effect should result remains for the present obscure; possibly several distinct chemical proc- esses are concerned, having different coefficients of acceleration by rise of temperature; at the higher temperature the available oxygen may enter into a quite different reaction from that on anal the maturation-change depends; the latter would then be prevented through a deficiency of available oxygen. Ovxidations in one set of processes may easily involve reductions in another if the supply of free oxygen is limited. What is remarkable is that maturation is prevented permanently by warming at this stage. Warming after the germinal vesicle has broken down has no effect on the course of maturation, the polar bodies forming in the usual manner; and after this process is complete the eggs, as already seen, may proceed to cleave and develop without fertiliza- tion. Apparently conditions unfavorable to maturation produce a permanent prevention of the process only if they act during the brief period immediately following the deposition ‘of the eggs; this is for some season a critical stage, and if the maturation process is not then begun it fails altogether. In harmony with this inter- pretation is the well known fact that starfish eggs which show no signs of maturing by twenty minutes or so alter removal from the neta to Roane sea-water remain immature permanently. The effects of momentary warming at stages succeeding the disso- lution of the germinal vesicle vary, as just shown, according to the exact period at which the treatment is applied. As already seen, membrane-formation and development may result from warming very soon after the vesicle begins to lose its distinct outline. The conditions are at first unfav orable, only a small proportion of eggs forming membranes, and still fewer dev eloping to a free-swimming stage. In general, as indicated by Table IV, the proportion of favorably developing eggs shows a progressive increase until an optimum stage 1s reached—usually about 15 or 20 minutes before the separation of the first polar body; warming at the time of separation of this polar body rarely results in larvae, and in later stages the conditions become steadily less favorable with lapse of time. Artificial Parthenogenesis in Starjish Eges 411 The conditions of this change of susceptibility are at present unknown. I have endeavored to determine if a similar variability exists in respect to fertilization by spermatozoa; and the result has appeared that although normal fertilization is possible through- out a far greater period in the history of the egg (namely, at any time after maturation has begun until several hours after its completion) a very similar variation in the degree of susceptibility to the fertilizing influence does in fact exist. Conditions for fertili- zation by spermatozoa are at their best during the maturation period, at or about the time of separation of the first polar body; and later they become less favorable. There is thus a certain parallelism between the conditions of artificial fertilization by momentary warming and of normal fertilization by spermatozoa. ‘The fol- lowing table gives the results of two series of experiments. Sper- matozoa were added to successive portions of eggs, taken in each series from a single female, at the indicated intervals after removal from the animal. ‘The condition of the eggs at the time of fertili- zation is indicated by the italicized portion in parentheses. Four other similar series of experiments were performed with, in general, very similar results. In all of these the most favorable time for fertilization was either before or about at the time of the separation of the first polar body; eggs fertilized at periods of one to three hours after the completion of maturation gave few or no larve, and these were mostly abnormal. ‘These experiments agree in indicating that the egg gives the best response to the fertilizing influence of the spermatozo6n at or near the time of separation of the first polar body. After the separation of the second polar body the proportion of developing eggs undergoes rapid decline. It is however possible for eggs at such stages to give normal larvz on fertilization, although the optimal conditions are found at earlier stages. On comparison with the results of momentary warming a cer- tain agreement is seen. ‘The egg responds best to both fertilizing influences at or near the time of separation of the first polar body although rather before than after this event in the case of warming This agreement is of some further interest as indicating that the essential determining conditions of the initiation of the develop- 412 Ralph 8S. Lillie mental process are similar in normal and in artificial fertilization. Further and more precise analysis of these conditions is needed; in particular, examination should be made of the susceptibility of eggs, at different periods during and after maturation, to the TABLE VI Series I. August 24, 1907 I 35 m. | (First polar body not yet separated) Practically all mature eggs form | blastule and gastrule | 57 m. | (First polar body about to separate) A large proportion of larve; | seems less favorable than Experiment 1 3 1h.17m. | (First polar body separated) Favorable; practically all eggs form active larve 4 | 1h. 50m. (Mature eggs have both polar bodies) Somewhat less favorable than Experiment 3; a large proportion of good larve 5 3h. 10m. (1 h. 15 m. after separation of 2d polar body) Marked contrast to Experiment 4; most eggs dead and coagulated next morning; only a few larve Series II. September 2, 1907 Tr 50m. (First polar body not yet separated) Practically all mature eggs form | larvae (First polar body in all maturing eggs) Very uniform and normal looking lot of larve; next morning are mostly active ; early gas- trule we rh. 55m. (Both polar bodies in all maturing eggs) Less favorable than Experi- | ment 2; larve less numerous and less well developed; a consider- able number small or otherwise abnormal 4 2h. 55m. (More than one hour after separation of 2d polar body) Unfavorable; | relatively few larve and these mostly abnormal; most eggs dead | and coagulated next morning | 4h. 30m. (Nearly 3 h. after 2d polar bodies) Still less favorable. Most mature eggs are dead and coagulated next morning; a few larva, mostly small, thick-walled, or irregular in shape. None normal 6 5h.25m. A few larve; most eggs dead and coagulated next morning fertilizing influence of momentary warming in dilute potassium cyanide solutions. This method, as will shortly be shown, pro- * duces results far superior to those obtained by simple warming in normal sea-water; and it is possible that after the completion of Artificial Parthenogenests in Starfish Eggs 413 §S maturation eggs may be found to respond to some such form of treatment. As yet I have made no investigation of these relations. Probably the most appropriate form of treatment will be found to vary at different stages, according to the physiological condi- tion of the egg. The experiments about to be described indicate that the state of oxidation of the egg-protoplasm is a most impor- tant factor; and it seems not unlikely that the above differences 1n response at different periods may be found to depend largely on varying conditions of oxidation at different stages. Effects of Combining Momentary Elevation of Temperature with the Action of Cyanide Solutions The supposition that momentary elevation of temperature pro- duces its effects on the eggs through an acceleration of oxidation processes suggested itself early in anes investigation. ‘The beauti- ful experiments of Loeb® had shown the importance of the presence of oxygen in the action of hypertonic solutions on the Strongylo- centrotus egg. I therefore tested the effects of warming starfish eggs under conditions that exclude the influence of accelerated oxidations. For this purpose sea-water containing potassium cya- nide to +49 concentration was employed. In this medium intra- cellular oxidations are greatly retarded if not almost altogether suppressed, as shown by the fact that mature eggs remain for days without undergoing the typical coagulative disintegration, which, as Loeb has shown, is dependent on oxidations. In the following experiments the eggs were warmed to 35° for 70 seconds while in KCN sea-water, to which they were transferred in some cases directly from normal sea-water, in others from 5%, KCN in which they had been allowed to lie for varying periods of time. After warming, the eggs were transferred in some experiments directly to normal sea-water, in others to 3%), KCN at normal tem- peratures, whence, after varying intervals, they were transferred to sea-water. The influence of previous treatment with cyanide solutions on the development of eggs warmed momentarily in normal sea-water ® Loeb: Biochemische Zeitschrift, vol. i, pp. 189, 1906, ef seq., and preceding papers in University of California Publications. 414 Ralph S. Lillie was first tested. Sea-water containing KCN in 355 to pe Con- centrations acts in the same manner as sea-water deprived of its dissolved oxygen by a current of hydrogen or otherwise; the mat- uration process is checked, and may be resumed on retransfer to sea-water if too long aninterval has not elapsed. As shown above, after the maturation-process has progressed beyond a certain stage, starfish eggs become less and less susceptible to the influence of momentary warming. It can be shown that the process (what- ever its nature) which deprives the egg of this susceptibility is retarded or prevented along with the maturation by the addition of cyanide to the sea-water. ‘This is illustrated by the following experiments : Eggs were placed August 21, 1907, 20 to 25 minutes after removal from the animal, in sea-water containing 3755 KCN. In this solution they were left for 2h. 30 m. ‘They were then trans- ferred to normal sea-water (which was changed to free the eggs of adhering cyanide) and portions were warmed to 35° for 70 seconds at successive intervals of ten minutes until the appearance of the first polar body. At the close of the period of exposure to the cyanide solution the eggs were almost all in an early maturation stage with invisible germinal vesicle. Maturation was resumed in normal sea-water; the polar bodies began to separate after an interval of 1 h. 30 m.; a certain delay in the resumption of the process is thus indicated. Eggs were warmed at the following intervals after return from cyanide solution to normal sea-water and the results were as tabulated in the following table: TABLE VII I 5m. | Eggs form membranes and some reach well-advanced cleavage | stages. No larve formed 2 15 m. More favorable; a considerable number of larvae formed 3 | 25 m. Considerable number of active larve 4 35 m. Seems rather less favorable than Experiment 3; still a good propor- tion form larve 5 45 m. Less favorable; only a few larve 6 thse isams | Unfavorable; no larve formed 1h.35 m. Unfavorable; no larve Artificial Parthenogenests in Starfish Eggs 415 An experiment of August 17 showed a similar result: Eggs placed ten minutes after removal in 37-5 KCN, left in this solution two hours, then washed for 10 minutes in normal sea-water and warmed, gave a considerable number of larve. In these experiments the eggs were not warmed directly in the cyanide solution; but were first transferred to normal sea-water and then after an interval, subjected to the warming process in the latter medium. The largest proportion of larvae developed from eggs warmed within 15 minutes to half an hour after this trans- fer (Experiments 2 to 4); later the conditions became less favorable. The failure to reach advanced stages in Experiments 5 and 6 may seem to contradict the rule found above that optimal conditions for parthenogenesis are found at a time approaching that of the separation of the first polar body. ‘The influence of the cyanide must, however, be taken into account; as will be seen later the presence of cyanide during the warming process improves the conditions greatly, and the greater favorability in the earlier experiments in all probability depends on the relative briefness of the period succeeding removal fromthe cyanide solution. These eggs were thus exposed at a relatively favorable stage of matura- tion while still toa certain degree under the influence of the cyanide. Sucha combination of circumstances would be favorable to devel- opment. In the experiments now to be described the eggs were exposed to the high temperature while in the cyanide-containing sea-water. In the first series they were placed in 559 KCN solution at an early maturation stage, and after varying intervals were warmed to 35° for 70 seconds in the same solution and then transferred directly to sea-water. The result has appeared uniformly that under such conditions a far larger proportion of eggs develop, and develop- ment is more rapid and more nearly normal, than in eggs warmed in normal sea-water without the cyanide treatment. 416 Ralph 8S. Lillve The following series will illustrate: TABLE VIII August 24,1907. Eggs were removed at 2:55 p.m. and after 30 minutes in normal sea-water were transferred to eat KCN in sea-water. In this solution, after the intervals indicated, successive portions were warmed to 35° for 70 seconds, and immediately transferred to normal sea-water. In each experiment the eggs were allowed to settle and the sea-water was changed and this washing process was repeated a second time Period in KCN : RESULT before warming | I 30m. | Almost all eggs form larvae, largely more or less irregular blastule; | | some reach early Bipinnaria stage 2 | 50m. Larger proportion of active and normal larve than in Experiment 1; practically all mature eggs form larve of which many reach the early Bipinnaria stage 3 th.1om. Rather less favorable than Experiment 2; many larve reach early Bipinnaria 4 th.jom. Majority of mature eggs form larve a good many of which are small and thick-walled; very active swimmers. A fair proportion reach early Bipinnaria 5 2h.35m. Sharp contrast to Experiment 4; all eggs die in an early stage. No larve formed . . . 2 . Control warmed in normal sea-water: Three portions were warmed in normal sea-water (without previous cyanide treatment) at respectively 30, 40 and 50 minutes after removal. All three formed numerous active larve; the conditions, however, were decidedly less favorable than with the cyanide- treated eggs; most eggs died in early stages, development was slower, and the resulting larve were less active and normal than in the favorable cyanide cultures. Sperm-fertilized controls: Sperm was added to five successive portions at 35 m., 57 m., 1h. 17 m., 1h. 50m. and 3 h. 10 m. after removal; numerous active ]arve were obtained in all but the last; on the whole, the best sperm-culture was inferior to the best cyanide-culture and reached less advanced stages of development. A second series on August 27 gave similar results though the eggs were not so favorable. The result was, however, all the more striking since the best cyanide cultures were found to give a larger proportion of active normal larve than were obtained with sperm fertilization, even at the most favorable time. Artificial Parthenogenests in Starfish Eggs 417 TABLE IX August 27. Eggs were removed at 9:45 a.m.; left 30 minutes in normal sea-water; then transferred to song KCN, and after the designated intervals warmed to 35° for 70 seconds in this solution, from which they were transferred directly to sea-water ; this was changed twice to remove all traces of cyanide Time in KCN | solution beats | I 45 m. All maturing eggs form membranes and cleave to an advanced stage. Only a few form blastule; these are relatively feeble and abnormal 2 60 m. | The majority of mature eggs form blastula; larve are largely abnor- mal, with walls of unequal thickness; the number of active and normal larva is greater than in the sperm-fertilized control 3 | rh.25 m. | Less favorable than Experiment 2. Eggs mostly stop short in early cleavage stages; only a few larve obtained 4 1h. 45 m. | Still less favorable. Eggs cleave irregularly and very few form blas- | tule 5 2h.15m. | Like Experiment 4; a few feeble abnormal blastulae 6 2h.45m. | Eggs stop short in early cleavage; no larve 7 3h. 50m. | Like Experiment 6. Cleavage irregular; no larve Control warmed in normal sea-water: Three portions warmed respectively 30, 40 and 50 minutes after removal gave only a few small abnormal blastule. Sperm-fertilized control: Portions were fertilized 30 m., 49 m., I h. 11 m., 1h. 27 m., rh. 55 m., 2 h. 55 min., 4 h. 20 m. after removal; in the best cultures (30 m. and 49 m.) only one-third to one- half of the eggs formed blastule of which a large proportion were abnormal. In experiments 3 to 7 many eggs cleave irregularly and stop short in early cleavage stages. A remark- able peculiarity of such eggs is that after 24 hours the blastomeres still remain clear and uncoagulated and apparently living, though undergoing no further cleavage. This condition is in striking con- trast to the fate of eggs fertilized either normally or artificially without cyanide treatment and whose development also ceases in early stages; in such eggs the blastomeres rapidly undergo the typical coagu- lative disintegration characteristic also of mature unfertilized eggs. The cyanide has apparently per- manently modified the cell-protoplasm in such a manner as to check or prevent the oxidations on which this breakdown depends. A third similar series (August 23) should also be mentioned briefly. In this series the control eggs, warmed in normal sea- water without previous cyanide treatment, gave no swimming larve; and the sperm-fertilized eggs gave only a few, from a por- tion fertilized about 40 minutes after removal; these fertilized later (Ih. 10 m., 1 h. 25 m., rh. 50m, 3 h. 15 m.and 5 h. 15 m.) gave no larve. The eggs were thus typically “unfavorable.” A portion of the unfertilized eggs was placed in 3;%;5 KCN 30 min- 418 Ralph §. Lillie utes after removal, warmed to 35° for 70 seconds after the indicated intervals in the cyanide sea-water, and then transferred as above to fresh sea-water. ‘The results were as follows: 25 m. in KCN | Large proportion of vigorous larve formed z 55 m More favorable than Experiment 1; after 24 hours numerous blastule and gastrule were present 1 h. 30 m. | Less favorable; relatively few blastule were formed and these were 3 3 | y | mostly abnormal | | 4 2 h. 50 m. Unfavorable; very few eggs reach blastula stage A fourth series (August 29) gave an even more striking result. Eggs were placed, 40 minutes after removal, in >; KCN, warmed to 35° for 70 seconds after the following intervals in this solution, then transferred to normal sea-water which was changed as usual. The control eggs warmed in sea-water at 40, 50 and 60 minutes after removal gave only a few blastula, the great majority dying and disintegrating at an early stage. In the best of the several sperm-fertilized portions only one-third to one-half of the mature eggs formed blastulae which were largely feeble or otherwise abnormal. ‘The results were as follows: I 35 m.in KCN | Next morning the dish was full of vigorous normal-looking blastule and early gastrule; condition much better than in the best sperm- fertilized control 2 60 m. Decidedly less favorable than Experiment 1; a good pro portion of eggs form larvee, but these are less active and normal than above 3 1h. 35 m. Still less favorable; nevertheless a large proportion have formed larve; these are largely irregular in form and somewhat feeble in movement In each of the above four series of experiments a far larger proportion of eggs produced larvae after treatment with cyanide for an appropriate length of time than after simple warming un- accompanied by such treatment; and the development was more nearly normal and resulted in the production of larger and more vigorous larve. ‘[heresults were indeed comparable in the best in- stances to those obtained with normal sperm-fertilization; in fact, in the last twoseries better conditions were obtained with the artifici- ally fertilized eggs than with those fertilized in the natural manner. Artificial Parthenogenesis in Starfish Eggs 419 It is noteworthy that a certain time of exposure to the cyanide solution—apparently about one hour or somewhat less—produces optimal conditions for development; after more prolonged expo- sure warming tends to result in abnormal development; in Tables VIII and IX the proportion of eggs that reach a larval stage is seen steadily to diminish with increase in the time of exposure to the cyanide beyond an hour or so, and the larve tend to become thick walled, irregular in shape, or otherwise abnormal. After exposure for more than two hours to the cyanide few eggs develop to a free-swimming stage. ‘This change in the condition of the eggs points to the existence of certain processes other than oxida- tions which continue unchecked in the presence of cyanide; there are no doubt hydrolyses of various kinds, and it may reason- ably be inferred that both kinds of processes are concerned in the changes that render the egg capable of parthenogenetic development. Suppression of oxidations for a time, during which the hydrolyses proceed unchecked, appears then to be favorable to bringing the eggs into a condition in which they respond readily to momentary warming; but if the hydrolyses unaccompanied by oxidations are allowed to proceed too far, lack of codrdination in the succeeding developmental processes seems to result, as shown by the increased proportion of unfavorably developing eggs. Normally a certain balance between the oxidative and the hydrolytic processes must exist; possibly a disturbance of this balance may be an important condition in the initiation of the developmental process. Such an interpretaton is at least suggested by the foregoing results. It should be pointed out that simple exposure to cyanide solu- tions without warming has no influence in initiating development in these eggs—at least under the above conditions. In the second of the two series tabulated above a portion of eggs was transferred from the cyanide solution to sea-water, without warming, at the time of each experiment of the series. None of these eggs formed membranes or showed any other sign of development and all were dead and coagulated next morning. The momentary elevation of temperature is thus essential. Since hydrolytic processes are relatively unaffected by cyanide, we may infer that hydrolyses are accelerated to at least four or five times the original velocity during 420 Ralph §. Lillie the period of warming—probably to an even greater degree, since the above results on membrane-formation (pp. 381, et seq.) indi- cate a much higher temperature-coefhcient of acceleration for such processes under the conditions prevailing inthe cell. Indications, then, seem to point to an acceleration of hydrolytic processes, combined with a repression of oxidations, as an important con- dition in the initiation of development in these eggs. That hy- drolyses are in fact accelerated seem to be here ii the condi- tions of the membrane-formation; this event occurs quite normally in the cyanide solution; it appears to be dependent on a hydrolysis which is greatly accelerated by a rise of temperature; and presum- ably other hydrolyses in the egg would be similarly affected by the same change of conditions. Membrane-formation seems to afford a clear proof that certain processes, not dependent on oxida- tions, are markedly accelerated by momentary warming, and that certain critical changes in the developmental capabilities of the egg- protoplasm may result from such momentary acceleration. Naturally it is impossible for eggs treated as above to develop while remaining in the cyanide solution; the transfer to oxygenated sea-water is indispensable. ‘This transfer however need not be im- mediate. It is possible to keep eggs, after warming under the above conditions, in cyanide sea-water or a certain not too prolonged period before transfer to sea-water. No visible change occurs during the stay in the cyanide solution, but on transfer to normal sea-water development proceeds normally. Indeed, under cer- tain conditions such after-treatment with cyanide has proved highly favorable to development as the following experiments illustrate : In these experiments the eggs, after remaining for a certain time in cyanide-containing sea-water, were warmed momentarily as above and brought to normal temperature in that medium; then after an interval they were returned to normal sea-water. A cer- tain stay in the cyanide solution after warming proved in every case decidedly favorable. Artificial Parthenogenesis in Starfish Eggs 421 The following series will illustrate : TABLE X Series I. September 7. Eggs were transferred 30 minutes after removal from the animal, to sea-water containing ,M.. KCN; after an interval of ca. 40 minutes they were warmed in this solution to 35° for 70 seconds; thence transferred to cyanide solution at normal temperature; from this, after the designated intervals, portions were transferred to normal sea-water. Exposure to KCN after Spent RESULT 1 | (control) 0 (to sea-water | Not favorable; comparatively few larve formed directly ) 2 5m. A striking contrast to the control; nearly all mature eggs form active larve; the majority of these gastrulate and many reach the early Bipinnaria stage 3 1om. Similar to Experiment 2; very favorable; numerous early Bipinnarie result (with mesenchyme and with the three intestinal divisions plainly marked) 4 20 m. A very good vigorous lot of gastrula were obtained, but rather less favorable than in Experiments 2 and 3; relatively few reach ad- vanced stages Controls warmed in normal sea-water 35, 45 and 55 minutes after removal gave considerable numbers of good larve. Sperm-fertilized control, fertilized one hour after removal, gave also a large proportion of larve, though fewer than in the best experimental cultures; development was also less rapid. A number of eggs were left in the KCN solution after warming until next morning (23 hours); they were then clear and uncleaved and all had typical membranes. On transfer to normal sea-water none underwent development, and next day all were dead and disintegrated. In the above series of experiments a marked increase in favor- ability resulted from the brief after-treatment with cyanide. In those next to be described a greater range of exposure to the cya- nide solution was employed; otherwise the procedure was the same. 422 Ralph S. Lillie TABLE XI September 2,1907. The eggs were left in sea-water for 45 minutes after removal; then transferred to .M > KCN for one hour, warmed to 35° for 70 seconds in this solution, retransferred to ,M., KCN at normal temperature, and thence, after the designated intervals, transferred to normal sea-water which was changed twice to remove all cyanide. | i] [Interval between warm- ing and return to Condition after 24 hours pormal sea-water 1 | o(control; directly to sea-| Large: number of normal well-advanced gastrule water after warming) 2 4m. A decidedly larger proportion of swimming larve than in Experiment 1; numerous normal gastrule are formed 3 14m. | A large proportion of larva; on the whole less uniform and less advanced than in Experiment 2 4 24m. | Like Experiment 3 but with somewhat larger proportion of abnor- malities; still, many good active gastrule 5 44m. Distinctly less favorable than Experiment 4; fewer larve than in Experiments 3 and 4, mostly blastule or imperfect gastrule 6 64m. Similar to Experiment 5; a good proportion of larva, largely abnor- mal; fair number of gastrule 7 1hr.24m. Relatively unfavorable; a smaller proportion of larve and these mostly small thick-walled blastule; relatively few gastrula, which are less advanced than in above experiments 8 2h. 34m. Considerable number of thick-walled blastule, but fewer than in Ex- periment 7. No regular blastula and no gastrule. Many eggs have stopped short in early cleavage stages 9 23 h. Development stops in early stages and eggs disintegrate; none reach larval stages Controls warmed in normal sea-water, 50 and 65 minutes respectively after removal, gave a fair num- ber of blastule after 24 hours, of which a few were beginning to gastrulate. As compared with Experi- ments 1 to 6 above, the larve are fewer and in a less advanced stage of development. Of the sperm-fertilized controls, those warmed within 1 h. 30m. after removal gave a large number of normal active larve. On examination, after 24 hours, of eggs left in the cyanide solution, all were found with membranes, round, clear, uncoagulated and uncleaved; many, however, showed little pseudopodia-like projections, and frequently small portions of the surface-protoplasm had become detached from the egg. While cleavage is impossible in the KCN solution, there appears nevertheless to have been some slight cyto- plasmic activity in these eggs. In a third series similar conditions were found; in this series the eggs were unfavorable and very few larve resulted even in the best sperm-fertilized control. A relatively small proportion of eggs formed larve in the best experiments; still, exposure to 53,5 KCN Artificial Parthenogenests in Starfish Eggs 423 solution for some minutes after warming gave decidedly better re- sults than were obtained from eggs transferred directly to sea-water without after-treatment with cyanide. The eggs were removed from the animal at 10 a.m. September 4, 1907; at 10:35 they were placed in 5; KCN; and after 55 minutes were warmed to 35° for 70 seconds and then replaced in cyanide solution at normal temperature, whence, after the intervals used, they were transferred to normal sea-water. Here the eggs brought into sea-water directly after warming in cyanide solution gave no larvae; while eggs after exposed to cyanide for only 5 minutes yielded consider- able numbers of good gastrulz, proving in fact more favorable than the best sperm-fertilized control; 10 minutes after-treatment on the other hand gave few larvae; and eggs left respectively 20, 35 and 50 minutes in cyanide after warming gave successively fewer and fewer; while none resulted with after-exposures of th. 10m., rh. 30 m., 2h. 50 m. and 4h. 20 m. These experiments indicate clearly that checking of oxidation processes during a certain interval after warming acts favorably under certain conditions; if this interval is prolonged for more han a few minutes conditions become rapidly less favorable, possibly, as suggested above, in consequence of the progress of certain hydrolytic processes unaccompanied by oxidations. ‘The striking increase in the proportion of developing eggs under the treatment used above, and also in the rate and normality of the development, suggests strongly that anaérobic conditions—at least at certain stages—form an important factor in the initiation of development in starfish eggs. Oxygen is necessary to the developmental process itself; but the internal changes that impart to the egg the distinctive power of automatic development seem best induced under conditions that must very effectually prevent most intracellular oxidations—at least those conditioned by the presence of enzymes. ‘The above results indicate therefore that momentary elevation of temperature—assuming that its essential action 1s the acceleration of chemical processes in the egg-sub- stance—must affect primarily other processes than the oxidative; in brief, acceleration of these processes, presumably hydrolytic in nature, simultaneously with a suppression of oxidations, appears 424 Ralph S. Lillie in some manner to result in changes leading to the initiation of development. After-treatment with cyanide also acts favorably in the case of eggs that have been warmed in normal sea-water without previous exposure to cyanide solutions. The following experiments will illustrate : TABLE XII September 9, 1907. Eggs were removed at 10:15 a.m. and the majority began to mature. After 43 minutes they were warmed in normal sea-water to 35° for 70 seconds. One portion (A) was then transferred to normal sea-water; a second portion (B) to 5 — KCN solution, and from this portion were transferred at the following intervals to sea-water eee ye Interval in KCN solution | before transfer to RESULT sea-water I © (control A; untreated) Almost all eggs die in early cleavage stages; only one or two blastule with KCN) | found 2 5m. | Most eggs die, but a distinctly larger proportion reach the blastula 10 m. | Similar to Experiment 2; larve are decidedly more active, numerous, Ww | | stage than in the control and these are better developed and well-developed than in the control; some have entered the early gastrula stage after 24 hours 4 | 25m. | Conditions are still more favorable; larva are more numerous and | more typical than in Experiments 2 and 3 5 60 m. Similar to 4; good many early gastrule (mostly more or less abnor- mal) after 24 hours A portion of eggs fertilized with spermatozoa about one hour after removal gave a good proportion of larve; largely small and thick-walled or otherwise abnormal. The eggs were thus not especially favor- able. The proportion of eggs developing to blastule and farther, while not large in the above series, was decidedly increased by the after-treatment with cyanide, and development proved both more rapid and more nearly normal in eggs thus treated. The best conditions were found in Experiments 4 and 5. Too prolonged after-exposure to cyanide affects the egg injuriously, the propor- tion of abnormal larve being greater in Experiment 5 than in Experiment 4. A repetition of this experiment, with a larger range of exposure to cyanide, gave a similar result (Table XIII). Artificial Parthenogenesis in Starfish Eggs 425 TABLE XIII September 10, 1907. Eggs were removed at 10:30 a.m.; the majority underwent maturation. After 45 minutes the eggs were warmed to 35° for 70 seconds as usual; a portion (for control) was placed immediately in normal sea-water; the remainder in KCN solution, whence, after the intervals M Zo000 indicated, portions were tran ae 10 sea-water Time in KCN solution | RESULT ™ 0 (control) Nearly all eggs are dead after 24 h., but a few blastule and gastrule | have developed. (A second portion of eggs warmed about 65 m. after removal also gave a few larve, mostly irregular blastule) 2 6 m. Most eggs die but larve are distinctly more numerous and active than in the control; a fair proportion are gastrulating after 24 hours. 3 | 11m. | Similar to Experiment 2; a large proportion of larve are gastrulating after 24 hours. 4 21m. Rather less favorable than Experiments 2 and 3; a fair number of larve formed 5 36 m. Similar to Experiments 2 and 3; a fair proportion of larve are gas- trulating after 24 hours 6 | 59m. A good proportion of larve after 24 hours, largely well formed | early gastrule ap || 1h. 30m. Unfavorable; no larve found 8 4h. Unfavorable; no larve The Sperm-fertilized control (sperm added 40 m. after removal) gave a large proportion of gastrule largely abnormal—irregularly shaped, thick-walled, or sluggish. Very few have gastrulated by 24 hours. Here also a decided increase in favorability followed after- exposure to the cyanide solution for a not too prolonged period. The results however were less favorable than in the experiments where eggs were exposed to cyanide for some time previously to warming eid were warmed in the solution. We may infer that while suppression of oxidations for a certain period after warming is favorable to development in eggs which have previously been well exposed to oxygen, this treatment differs from the preceding in certain very essential particulars, the nature of which requires further analysis. ‘Treatment with cyanide previously as well as subsequently to the momentary warming is essential if the most favorable conditions are to be attained. On reviewing the general outcome of the experiments described in this section we are led, first, to the conclusion that the entire series of events leading to the initiation of development in these eggs includes the changes preceding and following the warming 426 Ralph S. Lillie process, as well as those immediately induced by the latter. Secondly, all of these changes appear to proceed best under con- ditions of lack of oxygen—in other words, to be essentially anaé- robic in their nature. A predominance of anaérobic processes in the changes initiating development implies that an important part is played here by reductions (in the chemical sense), since anaérobic metabolism is always accompanied by the production of strongly reducing substances. ‘The possible part played by such reductions in the processes of cell-division and growth has been discussed by Mathews in the paper already cited; and the above general result is therefore consistent with his view that the production of asters (regarding this phenomenon as an essential feature of mitosis) is the expression of localized reducing processes. I can however hardly see my way clear to the conclusion that the momentary elevation of temperature under anaérobic conditions acts essentially by accelerating reductions and thus producing astral areas. Whilethis isa possible interpretation, it can, asalmost purely speculative, serve no particular purpose at present until confirmed or disproved by experiment. Moreover, in the sea- urchin egg the conditions seem of quite an opposite nature. Still, so far as regards the main chemical conditions of the partheno- genetic initiation of development in the starfish egg, the above results appear to indicate very definitely a subordination of oxida- tive processes to those of some other nature. This conclusion, while opposed to that reached by Loeb in the case of the sea-urchin egg, is in harmony with the recent experi- mental results of Delage® with the starfish. In this form partheno- genetic development through the action of carbon dioxide was found to be best obtained in thie absence of oxygen; a high concen- tration of oxygen in the carbon-dioxide-containing sea-water proved definitely unfavorable; and, in general, the lower the pro- portion of oxygen present, the better were the results obtained. Thus the initiation of development through this means, as well as through momentary warming, appears dependent on processes of an essentially anaérobic nature. Precisely contrary relations 33 Delage: Comptes rendus, vol. 145, p. 218, 1907. Artificial Parthenogenesis in Starfish Eggs 427 were found by Delage in the case of Strongylocentrotus, as had already been determined by Loeb; here the presence of oxygen in the hypertonic solutions is favorable to development. We have thus a striking contrast between the two forms in respect to the part played by oxygen in the initiatory process. ‘This contrast cannot be explained at present; it can only be referred to deep- seated constitutional differences between the two eggs. One fur- ther consideration is suggested and should be emphasized here: it must be recognized clearly that the physiological conditions underlying the initiation of development—i. e., the bringing of the egg into a condition in which it becomes capable of automatically passing through its characteristic ontogenetic cycle—may be of quite different nature from those on which the developmental proc- ess itself depends. ‘This is seen in the fact that notwithstanding the contrast in the conditions of the initiatory process, both the above eggs require the presence of free oxygen for their develop- ment. Unexplained constitutional differences between species play a part here, and we are not yet in a position for broad general- ization. Nothing but further exact investigation of the conditions of artificial parthenogenesis in eggs of different groups can be expected to bring to light the fundamental conditions common to the different types. For the solution of this problem a system- atically inductive ._procedure seems safest at present. SUMMARY 1 Momentary exposure of the eggs of Asterias forbesi, dur- ing the early maturation period, to temperatures of 35° to 38° results in the formation of typical fertilization membranes, fol- lowed by the development of many eggs to a free swimming larval stage. 2 The favorable duration of exposure to the above tempera- tures is very brief, with a well-defined optimum for each tempera- ture; this optimum is approximately 70 seconds for 35°, 40 to 50 seconds for 36°, 30 seconds for 37°, and 20 seconds for 38°. A very rapid rate of decrease in time of exposure with rise in tempera- ture is thus indicated, a rise of three degrees above 35° apparently 428 Ralph 8S. Lillie tripling the velocity of the process or combination of processes on which the initiation of development depends. The process of membrane-formation shows a similarly high temperature-coefh- cient of acceleration. 3 The responsiveness of eggs to this treatment varies greatly at different periods in the life of the egg. Warming within five minutes after the removal of the eggs from the animal is ineffective, and has the effect of preventing permanently the dissolution of the germinal vesicle. Warming at any time between the beginning of nuclear dissolution and the separation of the first polar body may result in development and the production of larva; the most favorable period is some little time (10 to 20 minutes) before the separation of the first polar body. Warming subsequently to this event tends to produce abnormal form changes or irregular cleav- age; after maturation is complete the effect is mainly to acceler- ate the coagulative change characteristic of mature unfertilized eggs in presence of oxygen. 4. Maturing eggs placed in 5, KCN solution retain for several hours their susceptibility to development by the above means. A stay of a certain duration in cyanide solution followed by momen- tary warming in this solution and transfer tosea-wateris followed by a striking increase in the proportion of favorably developing eggs. Further exposure of eggs to cyanide solution for a certain period after warming effects a still further improvement in the con- ditions of parthenogenetic development. Eggs thus treated with cyanide approximate closely, in the rate, character, and favorability oftheir development, to normally fertilized eggs. 5 Since the essential action of the above dilute cyanide solutions is to prevent intracellular oxidations, the inference is drawn that anaérobic processes play an important part in the series of changes leading to the initiation of development in starfish eggs. Sup- pression of oxidative combined with acceleration of hydrolytic and reducing processes is indicated as a condition of the initiatory process in these eggs. THE SEX RATIO AND COCOONING HABITS OF AN ARANEAD AND THE GENESIS OF SEX RATIOS?! BY THOS. H. MONTGOMERY, Jr. Witn Two Ficures This communication presents a study of the numerical propor- tions of the sexes in Latrodectus determined for 41,749 newly hatched young, with briefer observations on such proportions in other spiders; then an account of the general cocooning habits; next an attempt to show that different species of organisms prob- ably have different sex ratios, with an explanation of the origin of such differences. I LATRODECTUS MACTANS FABR. This is the largest North American Theridiid and it was selected partly because of the ease with which it may be kept, but more particularly on account of the great degree of sexual dimorphism: with the adult males and females so different in form and size it was anticipated that the sexes might be distinguished at the time of hatching, and this hope was realized. At Austin, Texas, where I have been observing these spiders, the web of this ‘species is found usually on the ground beneath a stone or log, sometimes several feet up within a crevice of a rock wall. The female remains at the upper portion of the web, and uses a niche or cranny as a retreat. Her web is composed of unusually powerful threads, capable of holding the strongest beetles and even of sustaining small stones; indeed I allowed my captives to fasten down with it the glass plates serving as covers for their cages, and this they did so firmly that the glass would not fall when the cages were inverted. Adult males I have found ‘ Contributions from the Zodlogical Laboratory of the University of Texas. No. 89. Tue JourNat or ExperiMENTAL ZOOLOGY, VOL. V, NO. 3 430 T hos. /5b Montgomery, tie only on the webs of females and only from December to February. Here, accordingly, the beginning of the reproductive season is in the early portion of the year. But on the Colorado river about sixty miles northwest of Austin I collected adult males in August. It would seem then that different groups of individuals show dif- ferent mating periods. Methods It was my object to determine not only the general sex ratio of this species, but also the ratio for each successive cocoon of a given spider. Therefore it was necessary to keep females through a whole reproductive season. That the captivity of the mothers did not produce abnormal results will be shown later. Only by controlling individuals in this way can one obtain accurate notes of the times of making and hatching of each cocoon, and also pre- vent the cocoons from being parasitized. Early in March and April of this year (1907) I collected a number of females and these I have kept until the autumn, up to the close of the time of ovi- position. ‘To each of them was devoted a separate cage of paste- board, most of these cages about three inches high, and most of them triangular with each side about three inches long. A photo- graphic glass plate was used as a cover, and another asa base. All these cages were kept together in a portion of my study where no direct sunlight reached them, and upon the cover of each was laid a paper card that excluded most of the light entering from above; by lifting this card one could look into the web without injuring it or disturbing the inmate. The bodies of their victims, when they have sucked them dry, the spiders drop out of the web; fort- nightly, accordingly, I pulled out the bottom glass plates so as to remove these accumulations. Cocoons were also taken out from below, by removing the same plates. “Thus the upper portion of the snare where the spider awaits her prey and where she devours It, was never disturbed, and to spare this web as much as possible, food was admitted through a small hole in one side of the cage, this hole being otherwise closed by a cork. Only living food is accepted, and for this I used large house flies caught in the usual wire traps; sometimes the diet was varied Sex Ratio of an Aranead 431 by larger insects. ‘The spiders were given equal amounts of food, and from the beginning of the experiment until August 8 all were richly fed, and daily except in the colder portion of the spring when food was hard to obtain, so that each of them averaged probably five or six blue bottle flies a day, quite the equivalent of the amount in a state of nature. Between August 8 and 29 they received only two meals, for I was absent; and in September they received only three good meals. Up through the first week in August, which marked the close of reproduction with most of them, these spiders were kept under natural conditions of light, temperature and amount of food. The healthy and active con- dition of the captives until the middle of August, and the large number of cocoons they produced, evidenced the favorable circum- stances under which they were maintained. Each spider received a separate number, and each cocoon the number of the mother together with the cocoon letter; thus the first cocoon of spider 2000 was 2000A, the second, 2000B, and so on. No cocoon was removed from a cage until several hours after its construction, for when just made the danger of injury to the eggs is greatest; each was lifted out as gently as possible with a pair of forceps, placed in a bottle covered with perforated paper and kept there until the young emerged; these were preserved in 80 per cent alcohol within twenty-four hours of hatching. It is necessary to kill the spiderlings before their first postnatal moult, else they commence to attack each other. ‘This isolation of the cocoons is the only method for preventing the young from dis- persing and so becoming lost at the time of hatching. Three series of females were kept: (1) twelve individuals whose young were allowed to hatch for the computation of the sex ratio; (2) five individuals whose eggs were preserved twenty-four hours after oviposition to test possible voluminal differences; and (3) two individuals kept to test parthenogenesis. More than this number I could not keep well fed. No deaths occurred until August 29, and the nine deaths from then on were probably due to insufhicient feeding commencing with the second week of August. 432 Thos. H. Montgomery, fr. Cocooning Habits The mating I have not observed, but it probably takes place about the beginning of the year when the adult males are found upon the webs of the females. Wild cocoons are to be discovered as early as February. The cocooning season extends, at Austin, from that month continuously into August. I conclude that it usually terminates in August, for only eight cocoons were made by my spiders after the eighth of that month; of these two were made in September and one in October. Further the last cocoons of a series, especially such dating from the middle of August, are frequently infertile; compare on Table I the last cocoons of 2013, 2016, 2021 and 2031. Females live on after the cocooning season provided they are well nourished. The completed cocoon is not quite globular but somewhat pyri- form, the upper portion having a short stalk to attach it to the object that overarches the web; when fresh it is snow white, when older, yellowish, and its outer coat is markedly resistant and firm. The process of cocooning and oviposition has much resemblance to that of Theridium? but Latrodectus is less specialized in that she applies the thread mainly by direct application of the spinner- ets and rarely by manipulation of the fourth leg pair. The case as seen in the making of cocoon 2020E was as follows: At 5:37 p-m. the mother was seen working at the base, a disc of flossy silk then only 2 mm. in diameter; she hung below it, holding its edges with her three posterior pairs of legs while with her first pair she suspended herself from the web; she was then making 52 appli- cations of her spinnerets per minute. The base was completed at 6:04, an inverted cup with a diameter equal to that of the fin- ished cocoon. Oviposition, with rapidly repeated uplifts of the abdomen against the concave surface of the base, lasted from 6:04 to 6:16. The construction of the cover of the cocoon occu- pied from 6:16 to 8:19; for the first ten minutes the fourth pair of legs were used to comb out the thread before each application of the spinnerets to the cocoon, after that time these legs were no ? Compare Montgomery: The oviposition, cocooning and hatching of an Aranead, Theridium tepi- dariorum C. Koch. Biol. Bull., xii, 1906. Sex Ratio of an Aranead 433 more employed to handle the issuing thread. The rapidity of the applications of the spinnerets was found to be as follows: from 6:16 to 6:25, 78 applications per minute, from 6:25 to 6:45, 120 applications per minute, from 6:45 to 7:30, 125 applications per minute, from 7:30 to 8:19, 140 applications per minute. At 8:19 she ceased suddenly, perhaps from exhaustion, then spun again at the rate of 108 applications per minute from 8:28 to 8:32. The cocoon was then completed, and the final touches were to anchor it firmly in the web after cementing it to the roof. Now each time the spider applies her spinnerets to the cocoon she draws out a thread having a length of 5 mm. (the length of the fourth tibia); multiplying the distance of such a thread by the number of applications of the spinnerets, the astounding fact is reached that in spinning the cover alone of the cocoon the spider employs a thread having a total length of about eighty meters. “The mus- cular energy employed is very great, being a rapidly repeated uplifting of the heavy abdomen. Another spider worked on the cover of a cocoon for one hour and fifty minutes, and two others for five hours each. Oviposition usually takes place in the morning before 6:30 o'clock, and a little later than that one usually finds the process of cover making. In 143 cases oviposition was between midnight and 7 a.m., in eleven cases between 7 a.m. and noon, in eight cases between noon and 6 p.m., and in only one case between 6 p-m. and midnight. The young make their own way out of the cocoon, usually through a single circular aperture that they make probably by biting; they emerge in rapid succession, and unlike the adults are positively phototropic. In eight cases the hour of emergence was between midnight and 6 a.m., in twenty-eight cases between 6 a.m. and noon, in sixty cases between noon and 6 p.m., and in thirty cases between 6 p.m. and midnight. ‘The afternoon at its hottest hours, between 3 and 5 of the clock, is the most frequent time of hatching. The young do not commence cannabalism until after their first postnatal moult, and the time of this varies 434 Thos. H. Montgomery, ‘fr. with the temperature as well as with the individual spiderling. Most of those cocoons laid from the middle of June on hatched in nineteen or twenty days, and two in as short a time as seventeen days; longer intervals are characteristic of eggs laid earlier in the year, and the earliest cocoon always takes the longest time to hatch; this is readily seen on comparing successive cocoons in Table I, and shows that the rate of development depends directly upon the temperature. The total number of cocoons raised by those seventeen spiders that furnished series of them was 187, an average of eleven to each spider. One individual formed eight cocoons, one formed nine, four formed ten each, four formed eleven each, five formed twelve each, while two formed thirteen each, the range thus extending from eight to thirteen. The time interval between successive cocoons varies with the month, so probably with the amount of nourishment, it being shortest in July and August; such intervals may be easily compared from the data given in Table I. Sexual Dimorphism and the Sex Ratio On the North American continent there are two good species of the genus Latrodectus Walck., L. mactans Fabr. and L. geo- metricus Keys., as I have convinced myself by a study of the material in the United States National Museum; for the oppor- tunity of examining this collection my thanks are due to the cour- tesy of Mr. Nathan Banks. ‘This collection contains specimens of mactans from California (San Bernardino, Tulare county, Clemente Island), Texas, New Mexico, Nebraska, District of Columbia, Colorado, Georgia, North Carolina and Oregon; while Marx’ states that it occurs also in Pennsylvania, Ohio and Utah. Though mactans shows this wide distribution and is every- where of rather confined local occurrence it does not appear to have split into geographical races. The sexes of mactans show a marked dimorphism both in size and color, as seen in the following comparison. 3 Catalogue of the described Aranee of temperate North America. Proc. U.S. Nat. Mus., xii, 1890 Sex Ratio of an Aranead 435 Adult female. Maximum dimensions: abdomen from anterior convexity to spinnerets, 12 mm.; first leg, 21 mm., second leg, 15 mm.; third leg, 13 mm.; fourth leg, 20 mm. Rufous black is the color of the cephalothorax, sternum and legs, and only the meta- tarsi and tarsi are lighter. “The abdomen is deep black, enormous, nearly globular and arched on all surfaces except the ventral; it is marked only by a broad red mark on the venter behind the epi- gynum, a short red median band just dorsal to the spinnerets, and (rarely) traces of other red spots along the dorsum; at the dorso-anterior border are one or two narrow transverse red marks. In a few specimens from more northern localities the dorsal red spots were prominent, and in one only a pair of oblique red bands on each side. The female is thus shining black with a few red markings. Fia. 1 Adult male. As shown by the accompanying figure the male is much smaller than the female, with elongate abdomen; in this figure the abdomina are shown from ventral and lateral views and the stippled areas denote the light markings; he has also propor- tionately longer and more slender legs. “The cephalothorax and sternum are both pale brown with darker borders, the cephalo- thorax also with a darker median stripe. The legs are yellow; the distal ends of the femora and patella are darker, and there are two rings of the same dark color at the distal ends of the tibiz. The abdomen in seven specimens has a broad white dorso-median 436 Thos. H. Montgomery, fr. band extending from near the anterior end to the spinnerets; at the antero-dorsal boundary a transverse white band that extends down on both sides; on each side behind the latter are two oblique white bands; all these bands are narrowly edged with black. On the venter there is a broad white mark which ts on each side bor- dered by black, and on each side of the spinnerets are three oblong black spots. The remainder of the abdomen is pale brownish flecked with white. In one male there was a red line inclosed within the dorso-median white band of the abdomen, and deep black filling all the spaces between the abdominal white bands.‘ The males in the instar just preceding the adult stage have the abdomen larger and more rounded, the legs proportionately thicker, and the abdomen colored like the young female, namely, dull or deep black, with a medio-dorsal white band including a red one, an arched transverse white band anteriorly, two oblique narrow white bands on the sides, and a broad white band on the venter. The newly hatched of both sexes are yellow with black stripes on the abdomen as follows: two narrow parallel stripes along the dorsum, two broader ones on the venter, and three (often broken) oblique stripes on each side. Thus the color of the adult male retains the color pattern of the young much more than does the adult female, for the latter becomes to great extent deep black. In color, size and activities the male is decidedly more embryonic. So far all observers, with the exception of Doumerc, have held the newly hatched of spiders to be sexually indistinguishable. At that period the genital plates are quite simple, and the pedipal- pal tarsus of the male is not different from that of the female. On sectioning the spiderlings of L. mactans I could not distinguish 4J have described mactans rather more fully than might seem necessary for our present purposes, but this is called for on account of the present confusion with regard to the American species. For this reason I will give briefly the characters in which geometricus differs from mactans. The male of geo- metricus differs only in having two pairs of black spots just behind the middle of the dorsum. The female of geometricus differs from the female of mactans in having the dimensions of all parts of the body slightly smaller but the abdomen much smaller, the legs pale colored with dark rings; in geometricus also the general color of the abdomen is pale brown, more rarely black, and always marked by lighter markings on the dorsum and sides. Further, the cocoon of geometricus has the surface beset with numerous slender, cylindrical villi, each from 1 mm. to 1.5 mm. in height, while the cocoon of mactans is quite smooth. L. geometricus occurs in California and Jamaica. Sex Ratio of an Aranead 437 ovaries from testes, for each is simply a small paired chord of germ cells of an early generation. But careful comparisons dem- onstrate that there are two constant forms of the newly hatched spiderlings, with the following differences. 1 Individuals which have the abdomen wider and deeper, with the dorsum much more strongly arched and the pedicel placed further back. In such individuals the abdomen is almost always distinctly larger. These are females. 2 Individuals which have the abdomen narrower and less deep, with the dorsum only moderately arched or not infrequently flattened or even indented. Such individuals have in almost all cases the abdomen smaller. These are males. 2 2 yj i IIe ie ee) ea OC ier lesmen Cee le ee Oe Ge 1a er Fic. 2 Intermediates do not occur between these two groups. In Fig. 2 I have illustrated such differences by showing in outline the abdomina of spiderlings from several cocoons. ‘To the left is the number of the particular cocoon, and on a line with it outlines of the abdomina of the largest and smallest females, and of the largest male of that cocoon, each abdomen shown on lateral and ventral view. It is hardly necessary to add that plane drawings cannot 438 Thos. H. Montgomery, te represent these differences as clearly as the rounded originals do. Now these are like the form and size differences of the abdomina of the adults; the adult female has the more arched abdomen with the pedicel placed further back, while the male has the narrower and flatter abdomen with the pedicel situated further forward. Therefore I conclude, and no reasonable objection can be enter- tained to this opinion, that those spiderlings most resembling in these particulars the adult females are females, and those most similar to the adult males are males. And it will be recalled that inter- mediate individuals are not found, and one can separate rapidly and unhesitatingly the spiderlings from a given cocoon into two lots. I had hoped to be able to distinguish the sexes at a still earlier stage of growth, by constant differences in egg sizes, and for this purpose preserved the eggs (all at the age of about twenty-four hours) of the cocoon series of five different spiders. But this expectation was not realized. The eggs in any cocoon differ somewhat in volume, not greatly, but these size differences form a graduated series and not two unbridgable groups. The sexual differences of the hatched spiderlings being constant only with regard to form but not always with regard to size of the abdomen explains why we do not find female eggs always larger and so dis- tinguishable from male eggs. The following table represents the proportions of the sexes in those cocoons of which the young were preserved shortly after hatching. The first column gives the number of the mother and the letters designating her successive cocoons. ‘The second gives the day of oviposition, and the third the time interval between oviposition and hatching. The succeeding columns state the number of unhatched eggs, of males, of females, and the male ratio: under the male ratio | understand the quotient obtained by dividing the number of males by the number of females. Sex Ratio of an Arnead 439 TABLE I Cocoon Oviposition gine i Utes fon g ratio hatching eggs 2000 days A April 11 41 18 333 39 8.5 B May 5 32 19 228 57 4. Cc 25 25 II 270 75 3-6 D June 5 20 ° 147 121 1.2 E 12 20 71 247 47 5.2 F 22 19 10 140 95 1.4 G July 1 19 5 | 132 165 8 H 8 19 II 146 102 1.4 I 16 19 42 145 94 1.5 if 25 19 136 92 69 1.3 K Aug. 4 18 63 60 76 i) 2002 A April 14 48 193 5) 71 09 B May 3 31 ° 280 27 10.3 Cc 21 25 ° 243 33 73 D 31 22 16 208 53 3-9 E June 9 19 9 186 68 2.7 F 15 20 I | 261 25 10.4 G 23 19 9 292 21 13.9 H July 1 20 4 325 20 16.2 I 8 20 RB | 242 18 13.4 J 20 19 I | 176 15 Tl. 7, 2003 B 404 2 202. Cc April 15 40 2 300 46 6.5 D May 8 33 ° 301 9 33-4 E 23 25 ° 355 20 19 /89/ 13 June 3 21 ° 313 37 8.4 G 10 20 ° 319 18 17.7 H 17 21 ° 331 12 27715 I 25 20 ° 316 17 18.5 J July 2z 20 ° 493 18 27.4 2008 A 239 6 39-8 B April 15 40 222 13 47 2 Cc May 6 33 ° 256 38 6.7 D 20 25 ° 246 25 9.8 E 29 22 ° 309 22 14. 440 Thos. H. Montgomery, fr. TABLE I—Continued Cocoon Oviposition pcre Unhatched | fot 2 ratio hatching eggs | | 2008 days | | F June 7 19 3 367 31 11.8 G 12 19 ° 355 27 13.1 H 18 | 20 ° 250 50 ine I 25 | 19 | oC 307 28 10.9 il July 1 19 | 2 287 40 Wied K 7 19 60 | 237 15 15.8 2012 | A 8 526 93 5.6 2013 A 646 6 107.6 B April 11 41 | 33 577 48 12. € May 7 32 3 316 15 21. D 24 26 I 430 9 47-7 E June 4 21 I 404 20 20.2 EF 12 20 ° 398 10 39.8 G 23 20 I 432 12 36. H 30 20 3 319 16 19.9 I July 16 20 3 432 12 36 J 26 20 2 282 9 31.3 K Aug. 5 2014 A 525 I 525. B April 15 38 30 284 34 8.3 iC May 13 29 I 366 19 19.2 D 28 23 ° 474 24 19.7 13; June 8 20 o 448 40 TIe2 F 17 20 I 393 ef) 39-3 G 26 20 5 395 24 16.4 H July 2 20 6 432 22 19.6 I 12 19 2 367 9 40.7 J 24 19 3 328 16 20.5 2015 A ° 519 27 19.2 B May 6 32 ° 419 15 27.9 (e 23 25 ° 507 21 24.2 D June 3 20 ° 441 22 20. E 12 20 ° 440 4 110. F 22 20 ° 405 18 22.5 Cocoon 2015 [rate Mae? 0 |h:(e9) 2029 PAS KY POR wAeaAw azamoaOw? Oo4ZzZ4RAGHH mots na wn Oviposition July 1 Aug. 6 Aug. 7 April 11 May 19 June 13 Sex Ratio of an Aranead 441 TABLE I—Continued Time to Unhatched a 9 Siratio hatching eggs days 20 I 486 17 28.5 20 I 399 8 49-8 19 ° 348 14 24.8 18 ° 316 5 63.2 18 I 276 14 19.7 40 I 286 33 8.6 37 3 278 29 9-5 pe) 3 ZO], 49 4-2 25 | 5 280 45 6.2 23 18 172 59 2.9 20 10 193 21 9.1 20 6 191 21 9. 21 121 59 42 1.4 20 76 105 35 3. 28 143 10 ° 10. | 32 2 208 32 6.5 399 16 24.9 40 23 533 7 7-5 25 4 311 37 8.4 20 4 287 73 3-9 21 ° 311 40 Fe 20 5 204 50 73 20 21 214 94 2.2 19 3 282 89 3-1 | 19 I 210 72 2.9 | 17 3 42 72 .58 41 I 643 25 25.6 36 3 387 18 215 26 ° 434 45 9-6 20 4 412 58 7.1 19 3 429 18 23.8 Thos. H. Montgomery, fr. TABLE I—Continued 442 Cocoon Oviposition 2021 G June 22 H 30 I July 9 J 18 K 30 L Aug. 6 M Sept. 25 2022 A B May 6 c 24 D June 3 E 10 F 17 G 25 H July 2 I 10 J 19 K 27 L Aug. 8 2023 A April 27 2030 A B May 13 2031 A B April 16 & May 11 D 25 3} June 4 F II G 18 H 26 I July 2 yi II K 18 L 28 M Aug. 24 Time to hatching days 19 Unhatched eggs » On W >A a ~ OrIAnNwW FAN Ff imal i 332 392 194 265 118 9 G ratio 24 16.2 33 Io. 16 23-3 Io 39-2 56 3+4 45 537 6 98. 10 24.7 31 TIT 28 Te4 51 4.8 32 8.4 32 9.2 29 IZ0E 24 10.6 44 532 26 9-5 52 12 96 1.2 7 58.5 42 4-7 61 12 47 6.2 60 2.8 28 39 50 -98 9 1.6 33 13 Sex Ratio of an Aranead 443 Notes to Table I Where a blank is left in the column headed ‘‘oviposition” it indicates that the cocoon was found when the mother was captured, the day of oviposition therefore unknown; where a blank is left in the column ‘‘Time to hatching” it signifies either that this was not determined, or else that the eggs did not hatch; where a blank occurs in the column ‘“‘unhatched eggs” it signifies that these were not counted, while a blank in the remaining columns indicates that the eggs did not hatch. Certain cocoons need further explanation, as follows: 2002K. Made September 18, too late to te entered into this table; it did not hatch. 2003A. Proportion of the sexes not given because many escaped at hatching. 2012A. A wild cocoon the mother of which was not secured. 2013K. Eggs killed by mold, the only such accident. o16A,B. Cocoons hatched when found, perhaps made the preceding season. 2016C. Found April 6 and hatched April 17, but the bottle of young dried up so that they could not be counted. 2017. Other cocoons of this spider were used for the egg series. 2020A. A wild cocoon that proved to be parasitized. 2021A. A wild cocoon found empty. 2023, 2030. Other cocoons of these spiders were used for the egg series. In the succeeding table these data are so summarized as to bring out the reproductive differences of the several spiders entered in the previous table, the “totals” of the third and seventh columns being averages. TABLE II 1 | Average no. | | Spider | No. of cocoons |of obec) Unhatched 5; 9 Average that hatched eggestoa | eggs o ratio cocoon 2000 II 35 386 1.940 940 2 2002 | 10 23.6 36 2,220 351 6.3 2003 9 22 2 39132 179 17. 2008 II 26. 287 2,866 329 8.7 201z I 8 8 526 93 5.6 2013 10 4-7 47 4236 157 26.9 2014 I0 5. 50 4,012 19 20.1 2015 II 27 3 45556 165 27-5 2016 10 44.4 444 1,681 334 Ss 2017 I 2 2 208 32 6.5 2020 * 10 6.4 64 2,793 614 5 2021 II G52 59 4,246 348 12.2 2022 12 9.9 119 3,468 365 9-5 2023 I 42. 42 118 96 1.2 2030 2 ° ° | 609 49 12.4 2031 7 160.3 1122 | 599 288 ae Totals | 127 22.6 | 2871 | 375210 4539 8.19 444 Thos. H. Montgomery, fr. From these data we infer the following conclusions: 1 The average male ratio (number of the males divided by the number of the females) is 8.19, determined from a count of 41,749 newly hatched spiderlings. Among the progeny of a par- ticular female this ratio was never lower than 1.2 nor higher than 27.5. [oa cocoon the average number of hatched males is 292.9, of hatched females, 35.7, and of unhatched eggs, 22.6; the average number of eggs to a cocoon is 351.2. Of the total of 127 cocoons entered in this computation, only 8 showed a male ratio of less than one, and from only one (2016M) did only males emerge and no females, this being the only “unisexual” cocoon. The highest male ratio in any cocoon, excluding the case of 2016M just cited was 202 (in 2003B); in eighteen cocoons the male ratio was 30 or higher. 2 The objection might be raised that the above average male ratio of 8.19 might not be the normal one for the species, but might be induced by the life of captivity of the mothers. Therefore I have considered separately the ratio in cocoons made in the natural state and brought into my study to hatch out. Such cocoons are the following of Table I: 2008A, 2012A, 2013A, 2014A, 2015A, 2020B, 2022A, 2030A, 2031A. These present a total of 3866 males and 223 females, giving the average male ratio of 17.3, con- siderably higher than the ratio 8.19 obtained from the total of cocoons [ raised. Whether this difference is due to difference in the mode of life of the mothers, or rather so the fact (to be brought out later) that the male ratio tends to be highest in the first cocoon of a series, I cannot say. ‘These figures would show at least that the high male ratio of captive cocoons cannot be ascribed to artificial conditions, and indeed there is no reason for thinking that the imprisonment of the mothers could affect this ratio. 3 It will be noticed that the male ratio was determined for each cocoon, accordingly also for the average of all cocoons, from the spiderlings that hatched out because I could not distinguish the sexes before the time of hatching. ‘That is, the male ratio of those eggs that did not hatch could not be ascertained, and this is the single disturbing error in the above calculations. ‘Table I furnishes the number of unhatched eggs for each cocoon, and Sex Ratio of an Aranead 445 Table II the average number for each series, for the 127 cocoons from which young emerged; 2871 eggs were infertile in the 127 cocoons from which hatched 41,749 spiderlings. Were those undeveloped eggs all males, the male ratio would be increased to 8.8; were they all females, decreased to 5.1; yet there is no prob- ability of either of these extreme cases. For when the male ratio is unusually high (30 or higher) the number of unhatched eggs to a cocoon is small, and where the male ratio is unusually small, a good case of which is the series 2031 of Table I, the number of unhatched eggs is generally but not always very high. ‘Therefore it is probable that a large proportion of such undeveloped eggs are males, and consequently the error introduced by such eggs is probably a small one. Next, as to the cause of lack of development of certain eggs. Mechanical jarring of the freshly laid eggs of spiders has been proved to be fatal to them ever since the observations of Herold, so that the handling of the cocoons in the removal from the cage to the hatching bottle may have prevented the development of some eggs. Yet I believe the arrest of development was rarely so induced, for I took great pains to handle all cocoons with extreme gentleness and, as we see from the third column of Table II, the average number of unhatched eggs to a cocoon varies with the dif- ferent mothers which would not be the case were it due to acci- dent. Probably, therefore, infertility of eggs is due to lack of fertilization; and indeed the frequent happening of the last cocoons of a series proving most infertile is to be ascribed to the supply of spermatozoa becoming exhausted. ‘To test whether normal parthenogenesis occurs | raised two immature females to maturity without benefit of males; one moulted in March and the other in April, which brought them to the mature condition with fully formed epigyna, and both were allowed to live, with good feeding, until August 30. One of them laid no eggs at all; the other made a cocoon containing a few infertile eggs on June 18, and on June 26 dropped on the floor of the cage, without constructing a cocoon, a mass of eggs that also proved infertile. “These two individuals had not been impregnated, and all the eggs laid by them were 446 : Thos. Ef. Montgomery, vig infertile, which renders unlikely the occurrence of parthenogen- esis in this species.° 4. When we compare the male ratios of individual cocoons given in the most right hand column of Table I no constant relation is found between this ratio and the position of a particular cocoon within a series. ‘The ratio may or may not be markedly different in immediately successive cocoons as well as in cocoons at the extremes of the particular series. The male ratio does not reg- ularly increase from the first to the last cocoon, nor does it reg- ularly decrease. Yet it will be noticed that of the twelve larger series of cocoons of Table I, in seven of them the first made cocoon of the series shows the highest male ratio (series 2000, 2003, 2008, 2013, 2014, 2020, 2022). 2 OBSERVATIONS ON OTHER ARANEADS The only observer who has furnished detailed data upon the sexual relations in spiders is Doumerc.® In the autumn of 1839 he captured a mature female of Theridion triangulifer and liber- ated it in a room where it formed a web ina window frame. On April 23 following it made a first cocoon from which only males emerged and on May 10 a second cocoon from which hatched only males. On June 16 she was impregnated by a male, on June 26 made a third cocoon from which only females emerged and on June 28 a fourth cocoon from which came out only males. Doumere does not state how he distinguished the sexes of these spiderlings; and for such “unisexuelliparous”’ species he con- cludes that there are two uteri, one for males and the other for females, and that oviposition does not take place from both at the same time. In collections adult females are usually more numerous than adult males, due simply to the fact that males in their mature condition are usually short-lived. But during the mating season males are more abundant in most species, certainly among the 5 Compare Montgomery: On parthenogenesis in spiders. Biol. Bull., 1907. ® Notice sur les Cocons a pontes unisexuellipares de l’Aranéide Theridon triangulifer, Walck. Ann. Soc. Entom., France. 9, 1840. Sex Ratio of an Aranead 447 Argiopids, Theridiids and Lycosids. In genera like Filistata, where the males are rare and where normal parthenogenesis seems to occur, it may be that the male ratio is smaller than 1. For Theridium tepidariorum C. Koch I found? a size differ- ence or dimegaly of the eggs, and concluded it was probable that females emerge from the larger eggs and males from the smaller ones; I showed also that some cocoons contained only large eggs and others only small ones, that therefore some may furnish only females and others only males, and that usually one kind of egg greatly predominates in number. Unfortunately this species is not found at Austin so that I could not keep females to test this point, and in my collection I have only a few bottles of newly hatched young. ‘These specimens exhibit, however, two kinds of individuals that do not intergrade; ones with larger and more arched abdomina, others with smaller and flatter abdomina; since these correspond respectively with the differences of the adult females and males, I take them to be females and males, respectively. Using this criterion I found the proportion of the sexes of the young from four cocoons to be as follows: TABLE II Cocoon Unhatched eggs fou g co Ratio ; | 1058 numerous 3 112 | -03 1059 I 431 | ° | 431 1060 | ° 186 p 93- | 1061 | 9 | 237 5 | 47-4 These examples are too few to allow any conclusions beyond the one that in a particular cocoon one sex greatly predominates. These relations are quite similar to those found by Doumerc for T. triangulifer, a particular cocoon having a predominance of a particular sex, but quite different from the relations in Latrodectus. Therefore there is probably not a common male ratio for all species of spiders. 7 Probable ‘dimorphism of the eggs of an Aranead. Biol. Bull. xii, 1907. 448 Twos, Jal Montgomery, Yt. 3 THE ORIGIN AND FIXATION OF SEX RATIOS There are obviously two sides to the question of sex determi- nation: the one, the process that regulates kind and succession of the sexes of one offspring-unit (totality of offspring of one parent), the other, the process of differentiation of the sexes and the origin of sex ratios. The second question is essentially phylogenetic, but if it should be proved that there is a distinct inheritance of sex then this phylogenetic aspect will come to have an important bearing on the other. We are now concerned with the question of the origin of sex and of sex ratios. The opinion is fairly general that in most animal species, those exhibiting parthenogenesis excluded, there is no great disparity in number between the sexes. ‘This has followed from a consider- ation of the numbers of the sexes in man, for so far in no other animal except the horse, is there available any computation of the sexes based upon a count of large numbers of individuals at birth. Only such computations have value, it is hardly necessary to add, that are founded upon the count of the sexes at the time of birth or earlier because the mortality after birth frequently varies with the sex, the males in certain lower animals being more short-lived and less resistant. We have found that in Latrodectus the male ratio is 8.19, there being born more than eight males to every female. For man, where the statistics are the only ones more numerous than those of Latrodectus, the proportion of males to females in 10,864,950 births is 1036 to 1000, the male ratio there- fore 1.03, as taken from the compilation presented by Pike in his Table III.* According to Morgan® for “‘still-born infants, fully formed, but not alive,’’ Quetelet found 133.5 males to 100 females, a male ratio of 1.33, and Bodio (from whose statistics I compute an average) a male ratio of 1.31. Darwin?’ tabulates the births of English race horses, 25,560 in all, giving 99.7 males to 100 females. For Bufo lentiginosus King" found in individuals that had com- 8 Pike: A critical and statistical study of the determination of sex, particularly in human offspring. Amer. Nat. 41, 1907. 9 Experimental Zodlogy. New York, 1907. 10 The descent of man, new ed., New York, 1886. 1 Food as a factor in the determination of sexin Amphibians. Biol. Bull., 1907. Sex Ratio of an Aranead 449 pleted their metamorphosis 241 males to 259 females, a male ratio of slightly less than 1; and other students have constated a low male ratio in Amphibia. These examples, based on cases where relatively large numbers of individuals were counted before the age of maturity, and all with the exception of Bufo at or before the time of birth, are sufficient to indicate that different species have different sex ratios, and that the sex ratio may be a quality of the species. Now when there is a male ratio of 8.19 as in Latrodectus, such a proportion of the sexes can be explained neither upon Newcomb’s theory of chance” nor yet upon Castle’s idea of the Mendelian inheritance of sex. Some other explanation is called for and, as I shall proceed to argue, it is probably to be sought in the factor of selection coupled with segregation. In the first place the distinction of the sexes is a difference of reproductive power. The female is the reproductive individual, while the male is not reproductive but impregnatory, for females can reproduce without males, as in cases of parthenogenesis, but males are unable to reproduce of themselves. The male is dis- tinctly the less important organism for the perpetuation of the race. ‘There is no reasonable proof for the formula of Geddes and Thompson" that the male is more katabolic and the female more anabolic, for that is merely an unfounded statement. The sexual difference is one of degree of reproductive ability. Probably in the earliest racial species all individuals reproduced in equal measure; this is probable simply because sexual dimor- phism implies a rather advanced differentiation, therefore one that should have developed more or less gradually. ‘The origin of sex 1s most easily concerved in the following manner. Indi- vidual variation within a species affects so far as we know every quality, therefore reproductive ability would be a quality subject to variation. Among the fluctuants of a species in which sex had not yet become pronounced there would be a series of individuals ” A statistical inquiry into the probability of causes of sex in human offspring. Carnegie Inst. Publ. 11, 1904. 8 Castle: The heredity of sex. Bull. Mus. Comp. Zool. Harvard, 1903. 4 The evolution of sex. London, 1897. 450 T hos. Ee Montgomery, hfs extending from such with the greatest to such with the least repro- ductive ability; the former would be incipient females, the latter, incipient males. With this difference in reproductive ability would certainly be associated metabolic differences, indeed the latter would probably occasion the former. Even racially older than observable sex difference is the process of conjugation, which at the start had no immediate connection with reproduction. Conjugation has been fixed by selection in that it aids the race by strengthening the reproductive individuals and making them more efficient in generation, conjugation being in certain Protozoa a form of nourishment. Conjugation being then of use to the race by strengthening or stimulating the reproductive individuals, selection would preserve those segregations of individuals in which variation with regard to reproductive ability is most marked, for in such species con- jugation would be most effective by occasioning the most diverse substance intermixture. Selection coupled with segregation would in time tend to eliminate the means, as the least useful in conjugation, and to preserve those individuals most dissimilar in reproductive capacity. This would terminate in a group of repro- ductive individuals, females, and of fertilizing individuals, males, not connected by intermediates. With regard to the sex ratio of a particular species, Pike (/.c.) concludes that it may “be looked upon as one of the physiological adaptations of the species, determined by the conditions of its existence. * * * If sex is hereditary, we might reasonably expect that the relative numbers of male and female births in any species would be those which, after deducting the early deaths, would confer upon the species at the period of sexual maturity of its individuals the greatest advantage in the struggle for existence so far as the production of young is concerned.”” This explanation is to my mind entirely just, and the factors would be, to carry the idea out further than Pike did, those of selection and segregation. Those species continue that survive in the struggle for life, and this struggle is the endeavor to insure offspring. Selection oper- ates by removing those species whose reproductive ability can- not successfully meet this struggle. Therefore selection would Sex Ratio of an Aranead 451 preserve, ceteris paribus, those races in which the females are either most reproductive or else most caretaking of their young, and in which there is at the same time a sufficiency of males to insure the needed fertilization of the eggs. Within a given species the male ratio would be subject to individual variation: some females would produce a preponderance of males, others of females. In the breeding area of such a species different groups of individ- uals would come to show different male ratios, just according to the productive peculiarities of their females, and in agreement with what we understand of the action of segregation or physiological selection in general. There would be groups with an unneces- sarily large male ratio, others with the male ratio injuriously small, others with the male ratio just rightly proportioned to the number of females to be impregnated. An excessively high male ratio would be a waste of males, and too low a male ratio a waste of eggs because then all the eggs could not become fertilized; in both these cases there would be an overplus of individuals that would not be of service in procreation. Accordingly, selection would preserve such groups of individuals in which the male ratio is most nicely proportioned, most closely proportioned, to the number of females needing to be impregnated. It would pre- serve them because they would leave the most offspring. The other segregations of individuals would become eliminated because they include a waste of energies and individuals. Selection and segregation would certainly be efficient factors, while it is more doubtful whether heredity would also play a part. What the male ratio would be in a particular species would vary with different conditions, and particularly with differences in the mode of life of the sexes. Where the sexes are most alike in general habits of life, where internal impregnation of the female is necessary and where the male cannot impregnate more than one female, the proportion of the sexes would be most equal. Where the males are physically quite as strong or even stronger than the females, and where the male has the habit of impregnating several females, it might be that the male ratio would sink below 1; whether polygamous gregarious species should be reckoned here we cannot say offhand, for the number of males born should be higher 452 T hos. H. Montgomery, Fr. than the number that mate, seeing that many may be killed by direct competition. More generally the male lives a simpler life than the female, is less active both physically and psychically, less fit for the struggle for existence, such as is the male in the spi- ders we have been considering; in such cases many males die before reaching maturity, and for such species the male ratio would be high. Then where eggs do not require fertilization, as in par- thenogenetic generations, selection would remove the males. Thus the average male ratio of a particular species would be fixed primarily by selection and segregation: these factors would confine in rather narrow bounds the ratio of that particular species. They would keep the number of males nghtly proportioned to the number of ova that are to be fertilized, without unnecessary waste of either. And since the factor of chance and the factor of Mendelian inheritance cannot explain certain specific sex ratios, it is at least suggested that these factors may also fail to determine sex within the offspring unit.® 3Tt may have some statistical value to append a count of 8796 adult individuals of the Rose chafer, Macrodactylus subspinosus, that I made during June, 1901, on individuals collected from one small garden near West Chester, Pa., and which gives a male ratio of 1.31: Date of 2 June 20 824 638 21 642 683 24 1568 1135 25 678 474 27 623 432 28 65 5 4 445 Totals 4989 3807 THE CHROMOSOMES IN DIABROTICA VITTATA,: DIABROTICA SOROR AND DIABROTICA 12-PUNC- TATA A CONTRIBUTION TO THE LITERATURE ON HETEROCHROMO- SOMES AND SEX DETERMINATION BY N. M. STEVENS With Turee Prates In Publication No. 36 of the Carnegie Institution of Washington, the spermatogenesis of a number of Coleoptera was described, and discussed with reference to the determination of sex. The study of the Diabroticas was begun at Cold Spring Harbor in the summer of 1906, and I wish to express my gratitude to Dr. C. B. Davenport for the privileges granted me both at the Carnegie Institution for Experimental Evolution and in the research labora- tory of the Brooklyn Institute. I am also much indebted to Miss Isabel McCracken of Stanford University for material of Dia- brotica soror, prepared with the greatest care and sent to me in December, 1906, and March, 1907. The same methods were used as in previous work on the germ cells of the Coleoptera. The germ glands were fixed in Gilson’s mercuro-nitric fluid, in Flemming’s strong chromo-aceto-osmic solution, and in Hermann’s platino-aceto-osmic fluid. Sections 5 thick were stained with iron hematoxylin or with thionin. The aceto-carmine method was used for long series of Diabrotica soror and Diabrotica 12-punctata. DIABROTICA VITTATA In the majority of the Coleoptera previously studied (85.7 per cent), an unequal pair of heterochromosomes was found. ‘The Diabroticas have an odd or unpaired heterochromosome, resem- Tue JourNAL or EXPERIMENTAL ZOOLOGY, VOL. V, NO. 4 454 N. M. Stevens bling in this respect the Lampyride and Elaterid as well as many of the Orthoptera and Hemiptera. In the spermatogonial equatorial plate of Diabrotica vittata, we find 21 chromosomes (PI. I, Fig. 1) of various sizes and shapes. If x be considered the heterochromosome, the others can be mated, forming ten equal pairs. In sections stained with iron hamatoxy- lin the division of the testis into several definite regions is very striking. The resting spermatogonia hold little of the stain, while the chromatin of the spermatocytes in synizesis and synapsis stages is very black, and again the spireme stage is pale. The synizesis stage here, as in several other Coleoptera (Stevens ’06), appears to be a prolonged telophase of the last spermatogonial mitosis. Fig. 2 shows the appearance of the short, crowded chromatin loops in synizesis. Following this stage comes a period in which the chromosomes are uniting in synapsis, and one finds many nuclei similar to Fig. 3, some of the loops still short as in Fig. 2, others longer and showing a sharp angle or a knob at the point of union of two chromosomes. ‘There is no such definite bouquet stage as in many forms, but one next finds a stage in which there are irregularly disposed loops with many free ends and some sharp angles like those in Fig. 3 (Fig. 4). In this stage the hetero- chromosome (x) is for the first time evident, condensed against the nuclear membrane. ‘This stage rapidly goes over into the spireme stage (Fig. 5), where all of the chromosomes except the heterochromosome (x) seem to be united into a single spireme thread, and the points of union are no longer visible. The spireme is very pale and the heterochromosome therefore very conspicuous. There is nothing unusual in the prophases of the first division. The spireme segments and splits longitudinally, the daughter elements separate as in Fig. 6, then unite again and form rods, dumb-bells, V’s and rings (Fig. 7). The chromosomes in the spindle (Fig. 8) are so attached to the spindle fibers that in meta- kinesis they separate into their univalent components, and go to the poles as short thick V’s which mass together but soon separate for the second division without any definite rest stage. The unpaired heterochromosome (x) is of course connected with only one pole of the spindle and does not divide in this division. Fig. The Chromosomes in Diabrotica 455 g is the equatorial plate with the heterochromosome (x) at a differ- ent level from the other chromosomes. Equatorial plates of the second division are shown in Figs. 10 and 11, the heterochromo- some (x) appearing in Fig. 10, and not in Fig. 11. All of the chromosomes divide in this division giving equal numbers of spermatids and spermatozoa containing ten and eleven chromo- somes, respectively. The spermatids (Figs. 12 and 13) contain a chromatin nucleolus (7), which is certainly not the heterochro- mosome, since it is found in all of the spermatids. As the head of the spermatozodn becomes more and more condensed, the nucleolus gradually decreases in size and finally disappears (Figs. 14 and 15). The ripe spermatozoén has a very long slender head (Fig. 16) which stains intensely black in contrast with the earlier gray stages (Figs. 14 and 15). DIABROTICA SOROR AND DIABROTICA 12-PUNCTATA Diabrotica 12-punctata of the eastern United States and Dia- brotica soror of the Pacific coast states resemble each other so closely that one might easily be mistaken for the other. Both are greenish yellow or yellowish green with twelve black spots on the elytra. Kellogg describes Diabrotica soror as yellowish green and Diabrotica 12-punctata as greenish yellow. The color varies considerably with the age of the beetle. Diabrotica 12- punctata averages larger,shades more on the yellow, and the under side of the abdomen is green or yellow while in Diabrotica soror itis black. The color of the abdomen seems to be the one exter- nal character by which the two species can always be distinguished; for the size, ground color, and size and fusion of spots are extremely variable in both species. A small amount of material of Diabrotica 12-punctata was col- lected at Bryn Mawr, Pa., in October, 1906. On examining the sections, it appeared either that the species was polymorphic as to its germ cells, or that there must be two or more sub-species or varieties, and possibly hybrids. It was too late to obtain more material of this kind, so, through the kindness of Miss McCracken, a supply of Diabrotica soror was secured for comparison with 456 N.M. Stevens the eastern species; and in the summer and autumn ‘of 1907, 100 males of each species were studied by means of aceto-carmine preparations. The character of the chromosomes in the male germ-cells of the two species is precisely the same. About 50 per cent of the individuals examined have nine equal pairs of chro- mosomes and an unpaired heterochromosome, while the remain- ing 50 per cent have one, two, three or four additional small hetero- chromosomes. DIABROTICA SOROR T ype I The stages in the spermatogenesis of the first type are in most respects similar to those of Diabrotica vittata. The sperma- togonial metaphase has nineteen chromosomes (Fig. 17), the unpaired chromosome (x) being the largest The synizesis and synapsis stages are similar to those of Diabrotica vittata, but less conspicuous in sections and the stages are less clear. The changes that occur between the telophase of the last sperma- togonial mitosis and the pale spireme stage (Fig. 18) prob- ably take place much more rapidly in this species. A polar view of the metaphase of the first spermatocyte division is shown in Fig. 19, a lateral view in Fig. 20, and a late anaphase in Fig. 21. The odd chromosome is usually found at or near one pole of the spindle in the metaphase (Fig. 20). The bivalents are similar to those of Diabrotica vittata, and the first division separates their univalent components. In preparations from Hermann material the chromosomes of the daughter plates (Figs. 22 and 23) often begin to show a vesicular condition and in telophase the heterochromosome (x) forms a vesicle by itself, while the other nine chromosomes are blended together (Fig. 24). Fig. 25 is a later stage taken from a cyst in which some second spermatocyte spindles were present, while Fig. 24 was from a cyst containing a few first spermatocyte spindles. Half of the nuclei in these cysts of course contain no heterochromosome. ‘The rest stage between the two divisions is more pronounced than in Diabrotica vittata where the chromosomes are simply massed together in telophase, The Chromosomes in Diabrotica 457 and separate for the second division without the formation of a nuclear membrane. The second spermatocyte equatorial plates are shown in Figs. 26 and 27, the heterochromosome (x) appearing in Fig. 26. All of the chromosomes divide in this division, giving, as usual, two equal classes of dimorphic spermatozoa. ‘The spermatids and spermatozoa are similar to those of Diabrotica vittata. The chromatin nucleolus is found in the earlier stages but is not visible in stages corresponding to Figs. 14 and 15, and the head of the mature spermatozoén is only about one-half as long. T ype Ila About two-thirds (33 out of 100 males collected at Mountain View, Cal.) of the individuals belonging to the second type have one additional small chromosome, making twenty in the spermatogonia (PI. II, Fig. 28). Vhe additional chromosome appears as a second heterochromosome in the growth stages (Fig. 29, s). In the first spermatocyte spindle the larger heterochro- mosome (x) is found, as usual, near one pole of the spindle, while the smaller one (s) may be in the equatorial plate (Fig. 30) or on either side of it (Figs. 31, 32, 33), closely associated with x or as widely separated from it as possible (Figs. 33 and 31). Fig. 34 is a polar view with the two heterochromosomes near one pole of the spindle. “The small chromosome may or may not divide in the first division. In some individuals it almost always (possibly always) divides as in Fig. 35 later than the other chromosomes. In other cases it may be found undivided between the daughter plates (Fig. 36), outside of one of them (Fig. 37), or it may be concealed in the general polar mass of chromatin. In the telo- phase and brief rest stage (Figs. 38 and 39) it is often quite dis- tinct from the remainder of the chromatin. Whether it divides in this mitosis or goes undivided to one pole or the other seems to be a matter of chance, depending perhaps on the part of the spindle which it happens to enter in the prophase. It seems to be much less automatic in its behavior than the other chromosomes. ‘This peculiarly erratic behavior of the small heterochromosome in the 458 N. M. Stevens first division gives, or may give, in the same individual six differ- ent kinds of second spermatocytes with reference to this chro- mosome (s), while there are, as usual, two kinds with reference to the large heterochromosome (x). If the small chromosome goes undivided to the same pole with the odd chromosome (x) (Fig. 33), we have second spermatocytes containing nine and eleven chro- mosomes (Figs. 40 and 41); if it goes undivided to the other pole (Figs. 31 and 37), the resulting second spermatocytes each contain ten chromosomes, one showing the large the other the small hete- rochromosome (Figs. 42 and 43); while if it divides, the second spermatocytes contain ten and eleven chromosomes (Figs. 44 and 45). As might be expected one finds two conditions in the second spindle. Either a small daughter chromosome is found outside of the equatorial plate (Fig. 46), or the small chromosome which has not divided in the first division, divides in the second (Fig. 47). Both conditions may be found in the same cyst. It is, of course, in only a few favorable spindles that it is possible to see the small chromosome actually dividing, but the metaphases are readily separated into two classes, one where all of the chromosomes are in the equatorial plate (Fig. 48) and another in which one small chromosome, which from its form and size is evidently a daughter chromosome from the first division, appears outside of the plate and often quite near one pole (Fig. 46). It is therefore quite cer- tain that the small heterochromosome divides in either the first or second division but not in both. Clear daughter plates of the second division have never been found. The conditions described above lead to the production of two equal classes of spermatozoa with reference to the large hetero- chromosome (x) and four classes, which may be quite unequal, with reference to the two heterochromosomes. CG A : ne variable numbers. Equal numbers (ASS SS) II : i ‘ ithe ) variable numbers. The Chromosomes in Diabrotica 459 If s always went to the same pole with x in the first division the classes of spermatozoa would be as follows: hae) UII OR ect as Equal numbers If s always went to the opposite pole from x, we should get the following results: ss @) $B S GL top mae Equal numbers If s always divided in the first spermatocyte division, there would be four equal classes of spermatozoa: I C ike } equal numbers. Equal numbers + } equal numbers. A study of seventy or more individuals of this kind gives the impression that the small heterochromosome most often divides very late in the first division, but it 1s certain that there is consider- able individual difference. In some cases nearly every anaphase of the first division shows s dividing; in others, it is rarely or never seen dividing in the first spindle, and as stated above, all of the various possibilities have been found in one individual. T ype IIb Fifteen out of the same 100 males of Diabrotica soror had two small heterochromosomes in addition to the eighteen ordinary chromosomes and the large heterochromosome x. ‘These are shown in a spermatogonial plate (Fig. 49). The three hetero- chromosomes may also be seen in a growth stage (Fig. 50), a prophase of the first division (Fig. 51), lateral and polar views of the metaphase (Figs. 52 and 53) and an anaphase (Fig. 54). Fig. 55 is an equatorial plate of the second division. When two small heterochromosomes are present both may go to either 460 N. M. Stevens pole of the first division spindle, one to each pole, or one or both may divide as in Fig. 54. The resulting combinations in the spermatozoa are as follows: ) | : \ variable numbers. Equal numbers ct | eed ag variable numbers. +x+a2s | Type Ilc Three out of the same 100 specimens had three small hetero- chromosomes, as shown in Fig. 56, a growth stage, 57 and 58, metaphases of the first division. T ype IId One individual had four such small heterochromosomes which may be seen in Figs. 59-65, growth stages and first spermatocytes. Here one may find all of the possibilities with respect to division and distribution of the small chromosomes. The possible com- binations in the spermatozoa are therefore as follows: + variable numbers. o> N — OOO 0 O + 4 i) + aN & Equal numbers variable numbers. +++4+4+ es 4 & + N G The Chromosomes in Diabrotica 401 There were no spermatogonial plates of type IIc and Id which could be counted, and in no case, though many ovaries have been fixed and sectioned and others examined with aceto-carmine, has it been possible to determine the number and character of the chromosomes 1n the female. DIABROTICA I2-PUNCTATA Exactly the same conditions as to the small heterochromosomes prevail in Diabrotica 12-punctata collected at Bryn Mawr, Pa., as in Diabrotica soror at Mountain View, Cal. Out of the first 100 males examined in October, 1907, 51 had no small hete- rochromosome, 35 had one, 11 had two, 2 had three and 1 had four, while in Diabrotica soror the numbers for the five corre- sponding classes were 48, 33, 15, 3, I. A few figures only will be given for Diabrotica 12-punctata. As in many other Coleoptera, spermatogonial equatorial plates in which the chromosomes are well enough separated for accurate counting are rarely found. ‘The one shown for Diabrotica soror, type IIb, in Fig. 49, was drawn from an aceto-carmine prepara- tion in which the chromosomes had been separated by pressure on the cover-glass. Figs. 66 and 67 are spermatogonial plates of Diabrotica 12-punctata, type | and type Ila, drawn from sections. There is some overlapping here, but no doubt as to the number in either plate. Growth stages for the five classes are shown in Figs. 68, 69, 70, 71 and 72. The larger size of both nucleus and chromosomes in Fig. 72 is due to its having been drawn with the same power from an aceto-carmine preparation. ‘These figures also serve to show something of the diversity of form of the odd chromosome (x). When no small heterochromosome is present it usually is nearly spherical (Figs. 18 and 68). Where one or more of the small chromosomes are found, it is as a rule somewhat elongated (Figs. 69 and 70), often irregular in form (Fig. 72), or much elongated and bent in U-form. Whether this difference indicates some influence exerted by the presence of the smaller heterochromosomes, or marks the individuals containing the small chromosome as a separate species is not at present clear. 462 N. M. Stevens In both Diabrotica soror and Diabrotica 12-punctata the small heterochromosomes are usually quite closely associated with the larger one (x) in the growth stages, but this is by no means invari- ably true. It is not at all unusual to find them separated in some cells and in one individual it was noted that the two were more often widely separated. (Figures to illustrate this have been thrown out for lack of space.) Fig. 73 shows a metaphase of the first spermatocyte from the one individual of this species which had four small chromosomes. Figs. 74 and 75 are anaphases from the same section, from an individual with two small chromosomes, showing in one case (Fig. 74, 5, and s,) both dividing, in the other (Fig. 75) one dividing (s,) and the other (s,) passing undivided to the same pole with the odd chromosome (x). In general, these small chromosomes are remark- ably uniform in size. One case, however, was found among the aceto-carmine preparations where an unusually small one was constant for the individual (Figs 76-78). This very small chro- mosome was not found dividing in the first spermatocyte and it could not be followed in the second division. In one cyst the spireme was segmenting, later than usual, into dumb-bell shaped bivalents (Fig. 77), as in Tenebrio molitor (Stevens ’05, Pl. 6, Figs. 177-179). As inthe other species of Diabrotica it has not been possible to find favorable stages for counting the chromosomes in the female. One may be able to do this by breeding the insects and working with the tissues of the larva or pupa. Judging from similar cases where the female number is known (for the Coleoptera, Elater I, Fig. 229, Pl. 13, Stevens ’06, and Photinus pennsylvanicus, figures not yet published; Anasa tristis and other Hemiptera, Wilson ’o5 and ’06; Peeciloptera, Fig. 283, Pl. 8 and Fig. 294, Pl. 9, Boring 07), we must suppose that the female number for Diabrotica vittata is twenty-two and for Diabrotica soror and Diabrotica 12- punctata, type I, twenty. Since the small heterochromosomes seem to be as likely to go to the spermatozoa which receive the odd chromosome (x) as to those which lack it, it would appear probable that the conditions with reference to the small hetero- chromosomes 1n the female are the same as in the male, and more- The Chromosomes 1n Diabrotica 463 over it 1s perfectly possible for more than four to occur in either male or female, as will be seen from the tables on p. 8. DISCUSSION. Sex Determination For the present it is necessary to assume that the number of chromosomes in the female bears the same relation to the number in the male as in other cases among the Coleoptera and Hemiptera where an odd or unpaired Hi eae enaorie is present in the male. The division products of the unpaired chromosome pass to one-half of the spermatozoa and these spermatozoa fertilize the eggs which develop into females; while the spermatozoa which lack the odd chromosome fertilize the eggs which produce males. ‘This still seems to be as far as we can safely go in dis- cussing the relation of the odd chromosome to sex determination. This chromosome is uniform in its behavior in the three species of Diabrotica, and it seems clear that it alone of the heterochro- mosomes described can have any connection with the determina- tion of sex. T he “ Supernumerary” Chromosomes The small heterochromosomes in Diabrotica 12-punctata were first seen in some first spermatcoyte spindles by Miss Anne M. Lutz of the Carnegie Institution of Experimental Evolution, Cold Spring Harbor, more than two years ago, but the matter was not followed up. Prof. E. B. Wilson, in a recent communication (Science, n. s., vol. 26, no. 677, p. 870), has given the name “supernumerary” chromosomes to certain additional heterochromosomes in Meta- podius (Hemiptera), and perhaps that name is as good as any other for the additional small heterochromosomes which appear in variable numbers in about 50 per cent of random collections of Diabrotica soror and Diabrotica 12-punctata. As in Meta- podius the number of supernumeraries is constant for the individ- ual. In Metapodius the supernumeraries are described as accom- panying a pair of idiochromosomes with which they frequently unite to form a compound element in the second spermatocyte. 464 N. M. Stevens In the Diabroticas they are present with a larger unpaired hetero- chromosome, and there is no evidence that they are ever united with it. The most puzzling characteristic of the supernumeraries in the Diabroticas is the fact that they may in the same individual divide in either maturation division, and when two, three or four are present, each one may divide in either spermatocyte division, thus giving great diversity in the chromatin content of the sper- matozoa. In Metapodius the supernumeraries are described as dividing in the first division. Occasionally, as in Figs. 64 and 73 two of the supernumeraries seem to be paired in the metaphase of the first division, but this is probably accidental, as it is not constant in any individual. The only other known case among the Coleoptera at all resem- bling this is that of the steel-blue flea-beetle, Haltica chalybea, which has a large and a small heterochromosome which are often widely separated in the metaphase of the first spermatocyte (fig- ures not yet published). In the anaphase, however, the two heterochromosomes are found between the two daughter plates, and one goes to each second spermatocyte. This is merely a case of late pairing and the distribution of the division products of the two heterochromosomes to the spermatozoa is the same as in other cases of an unequal pair of heterochromosomes. The first lot of Diabrotica 12-punctata were dissected out and all fixed together, so there was no opportunity to connect differ- ences in the germ cells with differences in external characters of the insects, if such existed. In the California material obtained in December, 1906, and March, 1907, from Miss McCracken, each beetle, after dissecting out the testis or ovary, was preserved in alcohol, and later placed in the vial with its germ gland. In the December lot there was a difference of 1 mm. in the length of the Elytra, some measuring 4.5 mm., others 5.5 mm. All of the smaller beetles had the odd heterochromosome only, the others one supernumerary additional; and it was quite naturally sup- posed that there might be two distinct species or varieties, one of which had only the large unpaired heterochromosome, the other an unequal pair of heterochromosomes. In the March lot, one exception occurred—a small beetle had the additional small chro- The Chromosomes in Drabrotica 465 mosome, and there were several individuals of intermediate size which had one or more supernumeraries. In June, July and August, 100 males of Diabrotica soror were studied in California. The length of the elytron from origin to tip, measured to the nearest fourth of a millimeter in a straight line—not over the curve —was recorded for each beetle, the insects numbered and kept for future reference. After the hundred had been collected, measured and studied, they were arranged in series according to nuclear type. It was at once evident that the insects of type I and type II were about equally variable in size, and that all of the varia- tions in fusion of spots occurred in each type. In fact no constant difference in external characters could be detected which might indicate two species. Thinking that possibly the variability in size might be different in the early and late broods, two lots of 100 each were collected November 1 and December 1, 1907, and measured without regard to the character of the germ cells. The results are given in a table below. Meanwhile too males of Diabrotica 12-punctata had been collected, measured, and the testes studied in aceto-carmine in October, 1907. ‘This species is somewhat less variable than Diabrotica soror, but the two types with reference to the supernumerary chromosomes show the same kind of variations. The variability in length of elytra of the different species and types is shown in Tables I and I1. TABLE I Diabrotica soror. Length of elytroninmm...........| 3 |3-25| 3-5/3-75| 4 |4-25|4-5 14-75] 5 5.25 5-55-75 Lyfe WOGouceeacane oobbanuse 1 | 4 | 14 4/16] 2] 6 | 1 | | 48 Day ovismemeferrstsistctetetyetsterises | I | CFU) Gul ean) peal Mel re I eye! | 33 Tb 2h struc een meeceniarae s lees 2 | 6 2 |} | | 15 MEO Gecognoapdoonseonbee | I 2 | | less INGLY io sessoonecuonobaedon | I | | leer | | | | June; 23\toyAupusti7cieicl--ileyasre-1s leek ° 5 | 0° | 27 | 11 | 34 3 | 16 I | 2 | | 100 Novemberstastosaeaeecenisostren ret Ase he Gh || pS eye roller at) Sl Be || | 100 Decémberhin cyrtaernterteeertetes I 9 9 | 25 | 17 | 23 | 11 4 I | 100 Watals . steisotster-fotelsteisisterare einen I} oO} 6| 2} 43 | 35) 81 | 40 | 66 | | 7 I | 300 466 N. M. Stevens TABLE II Diabrotica 12-punctata Length of elytroninmm.......... | 3 |3-25) 3-5/3-75| 4 [4-25] 4-5]4-75] 5 [5-25] 5-5]5-75 Ay PEWS OF Ssrereinre fats evatel acre evexa/otse | ale} 7 | 23 I 2) 51 UIE 00h an aah omOHASdAGcagen I 7) 4] 3 a7 I 35 Ube azysmemisctiecietsh tasters ates I I 9 II MUCH ENS oecoshOH OSs OAbOnOp I | I 2 EAS ab serereineete niente mcreracere| | | ee I r | | | pL Otale eres sietetors ehesets e1< (eit asvieisves= I Di] 2 rg) orgie ral Siltaulliaro It will be seen from the tables that Diabrotica soror is somewhat more variable and averages smaller in early summer than in late autumn; also that there is a possibility of two or three intergrading groups. The latter fact would not, however, seem to have any significance with reference to the supernumerary chromosomes, since in Diabrotica 12-punctata (Table II) the curve of variability is very steep with one mode at 5 mm. ‘The t00 specimens of Diabrotica 12-punctata included in Table II were collected on some late goldenrod in one corner of a field on October 3, 4 and g; the first 100in Table I, in one rose garden in small collections extend- ing over about six weeks. In both lots, most of the insects had recently emerged, and the conditions of temperature and nutrition under which they had developed could not have varied very greatly. The one significant result so far as the supernumerary chromo- somes are concerned is the parallel series of numbers for the five types of the two species—Diabrotica soror, 48, 33, 15, 3, I and Diabrotica 12-punctata, 51, 35, 11, 2, 1. Were it not for this par- allelism of results in the two similar but geographically widely separated species,tone might suppose the presence of the super- numeraries to be accidental, due perhaps to an irregularity in the breaking up of the spireme or to imperfect metakinesis some- where in the history of the male or female germ cells. The behav- ior of the supernumeraries in the growth stages of the spermato- 1 Diabrotica 12-punctata occasionally ranges into California, but belongs more especially to the eastern half of the United States, being perhaps most abundant in the Mississippi Valley. The Chromosomes in Dtabrotica 467 cytes would suggest that they might have originated in a detached portion of the odd chromosome (x), but such a supposition is not borne out by their later behavior in the maturation divisions, nor is there any evidence of an unequal pair among the other chromo- somes indicating accidental separation of a part of one chromo- some. The only evidence I have that the supernumeraries might be chromosomes in the process of development or degeneration is the one individual (Diabrotica 12-punctata, No. 83 of the lot of 100 collected in October, 1907) in which one very small supernumer- ary was observed (Figs. 76-78). In other cases there seemed to be remarkable uniformity in size without regard to the number present. If at some period in the past history of the race before the eastern and western species separated one supernumerary arose in any way, its peculiar habit of division, sometimes in one, sometimes in the other maturation division, may have given rise to the propor- tional numbers of the different types in the two species. Or it may still be possible, as was surmised earlier in the study, that (1) there will prove to be two distinct types (varieties or species) in each of the present species, one having the large unpaired hetero- chromosome only, the other having an unequal pair of hetero- chromosomes like that in Haltica, and that (2) the irregularities in time of division and the consequent peculiarities in number and distribution of the supernumeraries in Diabrotica are to be attrib- uted to hybridism. If this should prove to be true it would indi- cate little or no hereditary value for these supernumeraries or for the smaller members of the unequal pair in other Coleoptera. A careful biometrical study of several external characters may bring to light some differences which can be associated with the pres- ence or absence of the supernumeraries. The only other differ- ence in the chromosomes of the two types seems to be a varia- tion in the form of the odd chromosome (x). In type I it is usually nearly spherical in growth stages, while in type II it is more or less elongated. Until the material is investigated further, it hardly seems worth while to discuss at any greater length the hereditary significance of 468 N. M. Stevens 3 the supernumerary chromosomes or the possible results of their irregular distribution. It however seemed advisable to publish the results which have been obtained, as considerable time must elapse before more material can be worked over; and it is to be hoped that another summer’s work in California with breeding experiments and collections from different localities may furnish the data which are now lacking, and clear up the whole matter. SUMMARY 1 Dhiabrotica vittata has twenty-one chromosomes, ten pairs and an unpaired heterochromosome which behaves like the odd chromosome in other Coleoptera and in the Orthoptera and Hemiptera homoptera, dividing in the second spermatocyte divi- sion, but not in the first. Synapsis occurs at the close of the syni- zesis stage. A chromatin nucleolus is present in all of the sper- matids. 2 Dhiabrotica soror and Diabrotica 12-punctata both have in all cases nineteen chromosomes, nine pairs and an unpaired hetero- chromosome, which divides like that in Diabrotica vittata. About 50 per cent of the individuals examined have only nineteen chromosomes, the remaining 50 per cent have from one to four additional or “supernumerary” chromosomes which divide in either spermatocyte division, not in both, and may therefore give rise to from four to ten different kinds of spermatozoa with refer- ence to their chromatin content, in the same individual. The percentage of individuals containing no supernumerary chromo- some, one, two, three, or four supernumeraries, is nearly the same for the two species—48, 33, 15, 3, 1 for Diabrotica soror at Moun- tain View, California, and 51, 35,11,2,1 for Diabrotica 12-punctata at Bryn Mawr, Pa. It has not as yet been possible to associate the different nuclear types with variations in any external char- acter. Biological Laboratory of Bryn Mawr College Bryn Mawr, Pa. Nore—A part of the facts concerning the chromosomes in Diabrotica soror were given at the Inter- national Congress of ZoGlogists in Boston, August 21, 1907. The Chromosomes in Diabrotica 469 LITERATURE CITED Borine, A. M. ’07—A Study of the Spermatogenesis of ‘Twenty-two Species of the Membracida-, Jassida, Cercopide and Fulgoride. Journ. Exp. Zo6l., vol. 4, p. 469. Srevens, N. M. ’o5—Studies in Spermatogenesis with Especial Reference to the ‘Accessory’? Chromosome. Carnegie Inst., Wash., Pub. 36. ‘o6—Studies in Spermatogenesis. II. A Comparative Study of the Heterochromosomes in Certain Species of Coleoptera, Hemiptera and Lepidoptera, with Especial Reference to Sex Determination. Carnegie Inst., Wash., Pub. 36, II. Wiuson, E. B. ’o5—Studies on Chromosomes. II. The Paired Microchromo- some, Idiochromosomes and Heterotropic Chromosomes in the Hemiptera. Journ. Exp. Zodl., vol. 2, p. 507. ‘o6—Studies on Chromosomes. III. Sexual Differences of the Chromo- some Groups in Hemiptera, with some Considerations on Deter- mination and Inheritance of Sex. Journ. Exp. Zodl., vol. 3, p.t. o7—The Supernumerary Chromosomes of Hemiptera. Science, n. s., vol. 26, p. 870. DESCRIPTION OF PLATES Figs. 1 to 48, 50, 54 and 66 to 71 were drawn from sections with 2 mm. obj. and 12 oc.; Figs. 49, 51 to 53 and 55 to 65 from aceto-carmine preparations with 2 mm. obj. and 6 oc.; Figs. 72 to 77 from aceto- carmine preparations with 2 mm. obj. and 12 oc. The magnification of all of the figures was then doubled with a drawing camera, and the plates reduced one-half. Lettering used on Plates x = the unpaired, “odd” or “‘accessory”” chromosome. n = the chromatin nucleolus of the spermatids. s =a “supernumerary” chromosome. Sty 525 53, S4 = I, 2, 3 OF 4 supernumerary chromosomes in the same individual. Prate I Diabrotica vittata Fig. 1 Spermatogonial metaphase, twenty-one chromosomes. Fig. 2 First spermatocyte, synizesis stage. Fig. 3. First spermatocyte, synapsis stage. Fig. 4 First spermatocyte, postsynapsis stage. Fig. 5 First spermatocyte, spireme stage. Figs.6and7 First spermatocytes, prophase. Figs. 8 and 9 First spermatocytes, metaphase. Figs. 10 and 11 Second spermatocytes, metaphase. Figs. 12 to 15 Spermatids. Fig. 16 Ripe spermatozoén. Diabrotica soror. Type I Fig. 17 Spermatogonial metaphase, nineteen chromosomes. Fig. 18 First spermatocyte, spireme stage. Figs. 19 and 20 First spermatocytes, metaphase. Fig. 21 First spermatocyte, anaphase. Fig. 22 and 23 First spermatocyte, daughter plates. Figs. 24 and 25 Second spermatocyte, rest stage. Figs. 26 and 27 Second spermatocytes, metaphase. THE CHROMOSOMES IN DIABROTICA PLATE I N. M. STFVENS xuy y -wtee Xe as se On : 1 @.° L) = C73) werd i 5 ¢ : 8 dns af = syle" aime em oo es ae 10 a fe : : 15 WE pe pee CSE Soe : is 19 = ee Ke a) 00S ‘ of 00 16 ed se 22 23 Prate II Diabrotica soror. Type Ila Fig. 28 Spermatogonial metaphase, twenty chromosomes. Fig. 29 First spermatocyte, spireme stage. Figs. 30 to 34 First spermatocytes, metaphase. Figs. 35 to 37 First spermatocytes, anaphase. Fig 38 First spermatocyte, telophase. Fig. 39 Second spermatocyte, rest stage. Figs. 40 to 45 Second spermatocytes, metaphase, polar view. Figs. 46 to 48 Second spermatocytes, metaphase, lateral view. Type IIb Fig. 49 Spermatogonial metaphase, twenty-one chromosomes. Fig. 50 First spermatocyte, spireme stage. Fig. 51 First spermatocyte, prophase. Figs. 52 and 53 First spermatocytes, metaphase. Fig. 54 First spermatocyte, anaphase. Fig. 55 Second spermatocyte, metaphase. THE CHROMOSOMES IN DIABROTICA PLATE II N. M. STEVENS Prate III « Diabrotica soror. Type IIc Fig. 56 First spermatocyte, spireme stage. Figs. 57 and 58 First spermatocytes, metaphase. Type IId Figs. 59 and 60 First spermatocytes, spireme stage. Fig. 61 First spermatocyte, prophase. Figs. 62 to 64 First spermatocytes, metaphase. Fig. 65 First spermatocyte, anaphase. Diabrotica 12-punctata Fig. 66 Spermatogonial metaphase, nineteen chromosomes. Fig. 67 Spermatogonial metaphase, twenty chromosomes. Fig. 68 First spermatocyte, spireme stage, no supernumerary. Fig. 69 First spermatocyte, spireme stage, one supernumerary. Fig. 70 First spermatocyte, spireme stage, two supernumeraries. Fig. 71 First spermatocyte, spireme stage, three supernumeraries. Fig. 72 First spermatocyte, spireme stage, four supernumeraries. Fig. 73 First spermatocyte, metaphase, four supernumeraries. Figs. 74 and 75 First spermatocytes, anaphase, two supernumeraries. Figs. 76 to 78 First spermatocytes, unusually small supernumerary. THE CHROMOSOMES IN DIABROTICA PLATE III N. M. Stevens s > 2 x ae eer we <4 ae @ Or. 3S a 3 © e° oS e = =x d Af: C@eq « ee _ is 3 56 uet 58 a = e---s ce et @ = $ x x @ 1° i ae ae > 19° : \ M | | ENG pee — ° 66 ex * 68 From the Department of Comparative Anatomy, Harvard Medical School, Boston, Mass. THE EXPERIMENTAL CONTROL OF ASYMMETRY AT DIFFERENT SEAGES IN GHE DEVELOPMENT OF THE LOBSTER: BY VICTOR E. EMMEL INTRODUCTION The asymmetry of decapod crustacea has recently been studied by Przibram (or, ’02, ’05, ’07), Morgan (’04), Zeleny (’05) and Wilson (05). This asymmetry is manifest in the first pair of claws or chela—one of which is larger and frequently structurally different from the other. It has been found in some cases that if the large chela is removed, a small one may regenerate in its place. At the same time, the original small chela on the opposite side of the body may grow into a large one. ‘This transposition of chela is known as “reversal of asymmetry.” A complete reversal of asymmetry follows the amputation of the large chela in the adult of Alpheus as shown by Przibram and Wilson. On the other hand, sucha reversal is not obtained in similar experi- ments with the adults of the hermit crab and lobster as found by Morgan and Przibram. This I have confirmed in regard to the lobster by experiments with over 200 adults, in none of which was there obtained a transposition of the chelz. The previous studies have dealt only with adult animals. In view of this fact it seemed desirable to investigate the establish- ment of asymmetry at various stages in the growth of the lobster— one of the forms in which reversal does not occur in the adult. This has been done with the results about to be described. The work was carried on at the Experiment Station of the Rhode Island Commission of Inland Fisheries, and I desire to express my indebtedness to Dr. A. D. Mead for generously permitting me to Tue JourNar or ExperiMentaL ZoOLoGy, VOL. Vv, No. 4 °472 Victor E. Emmel use the apparatus and the excellent material available at the lobster hatchery. NORMAL DEVELOPMENT OF THE CHELAE A brief description of the normal development of the chelz of the lobster may aid in understanding the nature of the present experiments. In the adult lobster, one of the two chele, either the right or the left, is a rather long slender nipper claw, and the other is a larger and more massive crusher. Each claw consists of a movable jaw, the dactyl, and an immovable jaw or index. In very young lobsters the right and left chelz appear alike. During the first three larval stages they are embryonic in character. The claws are relatively short and broad; the index is smaller than the dactyl and both index and dactyl are beset with long hairs or bristles. During the fourth and fifth stages, the claws have become long and slender but are still alike. Characteristic tactile hairs and pointed teeth appear and the claws now begin to resemble the adult nipper type. At about the sixth stage however a divergence in the differen- tiation of the two chele becomes apparent. In one of the chelz the nipper characters continue to develop. This claw retains the long slender form characteristic of the adult. ‘Tactile hairs are distributed in a dense fringe on each side of the dentate mar- gin. The pointed cutting teeth are arranged in a linear series for each jaw with the exception of a stout displaced tooth about midway in the dentate margin of the index. In marked contrast to this development of the nipper, the other claw becomes wider, broad tubercle-like teeth develop, and the tactile hair of the nipper type gradually disappears in successive moults. Thus the adult crusher claw comes to be characterized by the almost entire absence of tactile hairs, and the presence of broad crushing teeth; and bya form larger and more massive than that of the nipper. The final result is the establishment of the adult asymmetry. In the development of the lobster therefore there is a series of larval and adolescent stages, in which there is a transition from Control of Asymmetry 473 symmetrical to asymmetrical chele. In this asymmetry the crusher occurs as frequently on one side of the body as on the other. PLAN AND METHOD OF THE EXPERIMENTS The present experiments were made in the following stages of the lobster’s development—the stages being designated as first, second, etc., according to the number of moults since the time of hatching: The second larval stage; Fourth stage; Fifth stage; Twelfth stage or lobsters a year old; Adult lobsters. All of these experiments attempt to determine to what extent asymmetrical differentiation of the chele can be controlled by amputation. In lobsters, as is well known, an injured limb is thrown off spontaneously or autotomously, separating along a cer- tain “breaking plane” near the basal joint. It was found necessary to exercise great care in the mutilation and rearing of the delicate larval lobsters. The chela was removed, under a small hand lens when necessary, by grasping the tip of the limb with a pair of forceps. In the older lobsters the chela promptly separates at the breaking plane. In the younger lobsters however the separation is not so readily obtained and a gentle pull may be required. The most difficult period in which to keep the lobsters alive is during the second and third stages. After several failures with ordinary aquaria the best results were obtained by keeping the animals in a current of fresh sea-water. This was accomplished by means of a rather elaborate apparatus built in the pool of a wooden float. The bottoms of pulp pails were removed and replaced by “bobbinet” cloth with meshes small enough to prevent the escape of the lobsters. A second cover, or false bottom, of mosquito bar was also found necessary—not to confine the lobsters but to prevent the ever present shrimp from pulling them out through the meshes. These pails were then suspended in the water of the pool. In 474 Victor E. Emmel each pail was placed a small paddle not unlike a boat propeller, consisting of a vertical shaft with two horizontal blades at its lower end. Each paddle was kept in motion by proper gearing with a “live shaft.”” The blades were beveled so as to give an upward movement to the current of water. In this way it was possible to rear a small per cent of the mutilated lobsters through the critical larval stages. After the fourth stage, the lobsters were placed in a floating car divided into small compartments. Each lobster was kept in a separate compartment and a careful record made of its mutilations, moults and regenerations. EXPERIMENTS WITH LOBSTERS IN THE SECOND STAGE On July 24, 1906, two groups of lobsters were. mutilated. These lobsters had all hatched from the egg within about four days. In Group A, the right chela, and in Group B, the left chela was removed from each specimen. In spite of an exceedingly great mortality, thirteen specimens were reared beyond the sixth stage. After each moult the regenerated chela was inva- riably removed. Thus the limb on the opposite side was given a great advantage for growth in order to learn whether this chela could be made to differentiate into a crusher. The results are shown in the accompanying table. This table includes also the data from a supplementary experiment made during the following summer. In this experiment great difficulty was likewise experi- enced in rearing the mutilated animals, for out of 200 larval lobsters from which the right chela was removed, only three speci- mens lived beyond the sixth stage. From the data for these sixteen lobsters it will be observed that when the chel had differentiated far enough to display asymme- trical characters, the claws which regenerated after amputation were all nippers; at the same time, the claws which were not mutilated, being thus given the greater advantage in growth, were all crushers. Control of Asymmetry 475 TABLE I. Larval lobsters in the second stage. Original chele symmetrical. Group A Right chela removed r FINAL ASYMMETRY OF CHELAE Number of subsequent No. Date of first moult Moults Date Right Left I July 24, 1906 siX Sept. 29, ’06 | nipper crusher 2 24, 1906 six Oct. 6, ’06 | nipper | crusher 3 24, 1906 six Sept. 29, ’06 | nipper crusher 4 24, 1906 six | Nov.8,’06 | nipper | crusher T 5 June 12, 1907 four | Aug.2,’07 | nipper | crusher T 6 12, 1907 four + (*) | Sept.21,’07 | nipper | crusher 7 12, 1907 four + Sept. 21, ’07 | nipper | crusher Group B Left chela removed | | 8 July 24, 1906 | six Oct. 27,’06 | crusher | nipper 9 24, 1906 six Sept.29,’06 | crusher | nipper 10 24, 1906 six May 31,707 (?) nipper f 11 24, 1906 six Oct. 19,’06 | crusher | nipper 12 24, 1906 six Oct. 19,06 | crusher nipper 13 24, 1906 six Oct. 19,706 | crusher | nipper 14 24, 1906 six Oct. 19,06 | crusher) nipper 15 24, 1906 . six July 12,’07 | crusher, nippert 16 24, 1906 six Sept. 22,’06 | crusher| nipper | *In specimens Nos. 6 and 7, the regenerated right chela was not removed after the moult to the sixth stage on July 14, and on account of unavoidable absence, a record of further moults was not kept. } These specimens, unfortunately, died before there was a clearly developed asymmetry of the chelz. In Nos. 4 and 5, the general appearance of the left claw and the characteristic double tubercle dentition at the base of the jaws indicated that these claws were differentiating into crushers. No. 10 however died on May 31 without having differentiated asymmetrically. } It may be of interest to note that No. 15 was much slower in its differentiation than the other speci- mens. At the close of the experiment in 1906, this animal showed no evidence of asymmetry. It was kept through the winter and after three more moults during the following summer, this lobster, in harmony with all the others in Group B, developed a crusher claw on the right chela. LOBSTERS IN THE FOURTH STAGE At the fourth stage the lobster has made a marked advance toward the adult form. The chela however are still alike and symmetrical. 476 Victor E. Emmel On July 25, 1907, seventeen specimens were mutilated on the day following the moult to the fourth stage. Each lobster was mutilated by the autotomous removal of the right chela. The results are shown in Table II. TABLE II Lobsters in the fourth stage. Original chele symmetrical. Right chele removed H | | | | | | | | FINAL ASYMMETRY OF No. | Stage Mutilation | Stage | Mutilation | Moult to CHEE | | | Stage VI |- $< | | Right Right | I | July 16 nipper | crusher 2 16 nipper | crusher 3 j 16 nipper | crusher Ava| + |) eee ic 16 nipper crusher 5 2 va | z S 16 nipper crusher 6 4 ak 2 aa a, 16 nipper crusher 7 | & 5 = ze 16 nipper crusher 8 § 3 bee 4 16 nipper | crusher 9 = ta} E 3 (?)* nipper crusher 10 3 2 Mw 5 (?) nipper crusher Il 2 2 & g, (?) nipper crusher 12 a a) 2 & (?) | nipper crusher 3 5 3 as = (?) | nipper crusher 14 3 re 3 a (?) nipper crusher 15 s s 4 (?) nipper crusher 16 | (?) | nipper crusher 17 @) nipper crusher | | 1 * Specimens 9 to 17 moulted to the sixth stage a few days after July 16, but the date of moult was not recorded. It is readily seen that these results show a marked uniformity. Without exception the mutilated claws became nippers; the claws which were not mutilated became crushers. LOBSTERS IN THE FIFTH STAGE The fifth stage is especially important because, normally, at the next moult asymmetrical characters are displayed. Consequently, during this period there must be a rapid progress in the differen- tiation of the chelz. Control of Asymmetry 477 The lobsters were mutilated by the autotomous removal of the right chele July, 1907. The mutilations were made about four days after the moult to the fifth stage. Through the kindness of Dr. A. D. Mead and his assistant, Mr. L. N. Wight, some of the lobsters were kept alive until the final data could be obtained in September. The results are shown in Table III. TABLE II Lobsters in the fifth stage. Original chele symmetrical. Removed right chela four days after moult to fifth stage Fuly, 1907 FINAL ASYMMETRY, SEPTEMEER 21 No. : Stage * Right chela Left chela | I seventh (?) nipper crusher 2 seventh (?) nipper crusher * 3 seventh ( ?) nipper crusher 4 seventh (?) crusher nipper 5 seventh ( ?) crusher nipper 6 seventh (?) crusher nipper 7 seventh ( ?) crusher nipper 8 died 9 aicd 10 died * By September 21, these lobsters were apparently all in the seventh stage, although the stage cannot be positively stated because a record of all the moults was not obtained. These results are in marked contrast with those obtained for the preceding stages. Instead of all the uninjured claws pro- ducing crushers, they produced three crushers and four nippers. At the same time, the regenerated claws, instead of being all nippers include three nippers and four crushers. Since in the adult lob- ster the crusher appears about as frequently on one side of the body as the other (and this is equally true of the nipper), it appears that the normal development was not modified by the removal of one chela. Evidently therefore during the fifth stage, in which the-chelz are apparently still symmetrical, the controlling influence of such amputations upon symmetry disappears. 478 Victor E. Emmel A point which should receive further investigation for this stage is the time of mutilation with reference to the moult. It will be observed that in the above experiments the mutilations were made several days after moulting. In another experiment the left chela was removed from a number of lobsters on the day in which they had moulted to the fifth stage. Only four of the specimens lived until the chele displayed asymmetrical characters, but it is interesting to note that for each lobster, a crusher devel- oped on the right or uninjured chela, and that the regenerated claw was a nipper. ‘This result indicates that possibly the asym- metry may be controlled at this stage, provided that the mutila- tions are made sufhciently early. IMMATURE LOBSTERS, A YEAR OLD With the assistance of Mr. E. W. Barnes, superintendent of the Experiment Station of the Rhode Island Fish Commission, we succeeded in keeping about thirty-five lobsters, hatched in July, 1905, until the following summer. At this time these year- ling lobsters were all in about the twelfth stage and averaged two inches in length. ‘The asymmetry of the chelz is clearly developed at this age. But when it is recalled that the lobster does not attain sexual maturity until about the fifth year, it will be readily appre- ciated that these yearling lobsters were still quite immature and at a period of rapid growth. It seemed desirable therefore to ascertain the degree of stability which the asymmetry. may have attained at this age as compared with the adult. The experimental results obtained are shown in Table IV. In this experiment on yearling lobsters, 15 were mutilated by the autotomous removal of the crusher chele (Group A), and 14 were mutilated by the removal of both chele (Group B). In both groups the regenerated claws were again amputated. In no case in either group did these mutilations and consequent regen- erations reverse the original asymmetry. Each yearling lobster retained its original right or left handed arrangement of the chelz. It should be added however that in the case of the crusher Control of Asymmetry 479 chela the regenerating claw is not always a characteristic crusher from the outset, but frequently displays, at first, characters inter- mediate between those of a crusher and a nipper.! TABLE IV Lobsters a year old. Original chele asymmetrical. Group A Crusher chela removed* | RELATION OF CHELAE AFTER TWO ORIGINAL RELATION OF CHELAE | REGENERATIONS AND TWO MOULTS LOBSTERS Right Left Right Left | 8 specimens crusher nipper | crusher nipper : F | F 7 specimens nipper crusher nipper crusher Group B Both chele removed* 8 specimens crusher nipper crusher nipper 6 specimens nipper crusher nipper | crusher * After the first moult the regenerated chele were again removed from each lobster. ADULT LOBSTERS Przibram (or, ’02) and Morgan (’o4) have already observed that in adult lobsters a typical reversal of asymmetry as the result of amputation and regeneration of the chela has not been found. In the course of my experiments over 200 adults were mutilated by removing one or both chela. In no case did a crusher develop on the side which had originally carried a nipper, and the same was true, vice versa, for the nipper. As in the yearlings, but not to the same extent, the regenerating crusher chela is not always at first distinguishable as such, but may present characteristics intermediate between the nipper and crusher (Emmel, ’06?). Also in certain very rare cases, symmetrical chele of either the nipper or crusher type may regenerate in place of the amputated asymmetrical limbs (Emmel, ’06%, ’07?.) The fact to be empha- sized however is that in these adult lobsters, the amputation of 1 Compare Emmel ’o6* and Przibram ’o7, p. 291. 480 Victor E. Emmel neither one nor both chele produced a reversal of the original asymmetry. DISCUSSION Until recently the phenomenon of reversal of asymmetry or compensatory regulation, which Przibram and Zeleny found Alpheus, was not supposed to occur in such forms as the lobster or hermit crab. It appeared that these species were characterized by a “direct regeneration” of the original asymmetry. But the discovery that under certain conditions the adult lobster might regenerate a crusher from the stump of the amputated nipper Rhee (‘Emmel ’06?) demonstrated that in the lobster, at least, both sides of the body might still retain the potentiality for the more highly differentiated type of crusher claw. The present results which show that asymmetrical differentiation can be controlled at early stages of development in the lobster, suggest that similar relations in the development and stability of asymmetry may be found in other crustacea.’ Various theories have been advanced concerning the factors which determine right or left asymmetry, and which may be dis- cussed on the basis of the preceding experiments. Herrick (’o05) studied the shrimp Alpheus, and concluded that asymmetry of the chelz in Alpheus and also in the lobster “is probably one of direct inheritance, all members of a brood being either right or left handed. ‘That is to say, the normal position of the toothed or crushing claw is not haphazard, but is predeter- mined in the egg”’ (p. 225). Conklin (’03, ’05), without discussing inheritance, shows how inverse symmetry may be determined in the egg. He found rea- son for believing that the cause of inverse symmetry, which occurs regularly among some species and occasionally among all, man included, is to be found in the inverse organization ofthe egg, and ?It is interesting to find this suggestion already anticipated by Przibram. In his important mono- graph published in the Archiv. f. Entw.-Mech., Bd. 25, 1907, p. 310 (received while the present paper was being written), he discusses the question,‘‘Is the possibility of the reversal of chela present in those forms which have hitherto shown no reversal ?”” He concludes that in some of these (Callianassa and Carcinus) the asymmetrical relations may be altered. ~*n = Control of Asymmetry 4 that this inverse organization may be due to the maturation of the egg at opposite poles in dextral and sinistral forms (’05, p. 10). On this basis, the alternate appearance of right or left asymmetry in the lobster might be regarded as cases of inverse symmetry “resulting from slight alterations in the localization of germinal substances in the unsegmented egg.” Morgan (’07) does not venture to decide between the possibili- ties of “inheritance” and the structure of the egg, as determining right or left handedness in various species. He says “both possi- bilities seem to exist in the egg; but whether this can be referred to alternate dominance and recession, or to purely local conditions that arise during segmentation, is unknown” (p. 165). It is evident that the present experiments at least demonstrate that the asymmetry of the adult lobster is not necessarily inherited nor even predetermined in the egg. However, the question still remains as to what factors in normal development determine right or left asymmetry. No evidence was found that the occur- rence of right or left asymmetry in the lobster can be referred to germinal units having “alternating dominance and recession.” The fact that in early development a crusher can be produced on either side of the body by the amputation of the opposite chela, indicates that the factors which control asymmetry become opera- tive after hatching. What these factors are, or how they may be released by the amputation of a limb, is not known. We can merely refer to the fact that in early stages of development of the lobster the asymmetry of the chela can be experimentally controlled. When asymmetry. has once been normally established, similar experiments no longer reverse it. That the accidental loss of a limb in the young lobster may play an important role in determining the asymmetry of the adult isnot improbable. For the autotomy of a chela during the exigen- cies of moulting or as the result of injury, is a common occurrence, especially among young lobsters. In an examination of several thousands of fourth stage lobsters it was found that a large per cent of the animals had lost either the right or left chela. In these the right or left asymmetry would not be inherited or due to the structure of the egg. 482 Victor E. Emmel Przibram (’07),in his recent extensive work on “Die Scherenum- kehr bei decapoden Crustaceen,”’ found experimentally that a reversal of chelz could be obtained in six genera and eleven species of crustacea, including forms in both the Macrura and Brachyura. He then inquires whether this capacity for reversal of asymmetry is a constant characteristic for a given species. Here he finds that the statement that a reversal of asymmetry always follows the amputation of the crusher (““K — schere”) in a crustacean requires modification. For it appears that the readiness with which reversal occurs varies inversely with the size of the animal. Specimens of Athanas, Alpheus, Typton, Callianassa, Carcinus and Portunus, which are under 10 mm. in carapace length, showed a quick and complete reversal of asymmetry. But, on the other hand, the larger specimens showed a decrease in the tendency toward transposition of the chela, so that when the large chela is removed the chela on the opposite side retains its original form. It appears therefore that the relations found in the control of asymmetry at different stages of development in the lobster, are also true for other crustacea with asymmetrical chela. For both in the lobster and in the forms described by Przibram, the various stages in their development form a series beginning with a com- plete control of asymmetry and ending with the disappearance of such control. In other words, the possibility for experimental control and reversal of asymmetry seems to be correlated in some way with the degree of differentiation or development, so that the greater the degree of development the more stable is the asymme- try of the organism. SUMMARY 1 Inthe first four stages of the lobster’s development, a crusher may be produced on either the right or the left side of the body by the autotomous amputation of the chela on the opposite side—the regenerated chela becoming a nipper. 2 During the fifth stage, although the chele are apparently still symmetrical, the possibility for such experimental control disappears. Control of Asymmetry 483 3 In later stages of development, when the asymmetry of the chele has become established, the amputation of one or both chelze does not produce a reversal of the original asymmetry. 4 The results of these experiments indicate therefore that the factors which control the asymmetry of the lobster become operative after hatching and are correlated with conditions of growth after the organism leaves the egg. No indication was found that the occurrence of right or left relations of asymmetry in this species can be referred to germinal units having “alter- nating dominance.” It appears also that these relations are not due to an “inverse organization of the egg,”’ for it is evident that up to the fifth stage right or left asymmetry can be produced at the will of the experimenter. January, 1908 LIST OF LITERATURE Conxuin, E. G. ’03—The Cause of Inverse Symmetry. Anat. Anzeiger, Bd. xxiii, 1903, p. 577-588. ’05—The Mutation Theory from the Standpoint of Cytology. Science, n. s., vol. xxi, 1905, p. 525-529. Emmet, V. E. ’05—The Regeneration of Lost Parts in the Lobster. 35th Report of Inland Fisheries of Rhode Island, 1905, p. 81-117, with two plates. ’06'—The Relation of Regeneration to the Moulting Process of the Lob- ster. 36th Report of Inland Fisheries of Rhode Island, 1906, p. 257-313, with four plates. ’06*—Torsion and Other Transitional Phenomena in the Regeneration of the Cheliped of the Lobster. Journ. Exp. Zodl., vol. ili, 1906, p. 603- 618, with two plates. ’06'—The Regeneration of Two Crusher Claws following the Amputation of the Normal Asymmetrical Chele of the Lobster. Archiv f. Entwick-Mech., Bd. xxii, 1906, p. 542-552, with two plates. ’o7'—Regeneration and the Question of ‘Symmetry in the Big Claws of the Lobster.’ Science, vol. xxvi, 1907, p. 83-87. ‘o7*—Regenerated and Abnormal Appendages in the Lobster. 37th Report of Inland Fisheries of Rhode Island, 1907, p. 99-152, with ten plates. 484 Victor E. Emmel Hap ey, P. B. ’05—Changes in Form and Color in Successive Stages of the Amer- ican Lobster. 35th Report of Inland Fisheries of Rhode Island, 1905, p- 44-80, with eleven plates. Herrick, F. H. ’95—The American Lobster. Bull. U. S. Fish Commission., 1895, 252 pp., with 54 plates. ’o7—Symmetry in the Big Claws of the Lobster. Science, vol. xxv, 1907, P- 275-277- Morean, T. H. ’04—Notes on Regeneration. Biol. Bull., 1904, vol. vi, p. 159-172, with four figures. ’°o7—Experimental Zodlogy, New York, 1907, 454 pp. PrzipraM, H. ’o1—Experimentelle Studien tiber Regeneration. I. Arch. f. Entw.- Mech. Bd. xi, p. 321-345, with four plates. *o2—Experimentelle Studien uber Regeneration. II. Arch. f. Entw.- Mech., Bd. xiii, p. 507-527, with two plates. ’07—Die Heterochelie bei decapoden Crustaceen. III. Arch. f. Entw. Mech., Bd. xix, p. 181-247, with six plates. ’o7—Die Scherenumkehr bei decapoden Crustaceen. Arch. f. Entw. Mech., Bd. xxv, p. 266-345, with five plates. Witson, E. B. ’03—Notes on the Reversal of Asymmetry in the Regeneration of the Chelz in Alpheus heterochelis. Biol. Bull., vol. iv, p. 197- 214, with three figures. ZeELENY, C. ’05—Compensatory Regulation. Journ. Exp. Zodl., vol. ii, p. 1-102, with twenty-nine figures. THE PHYSIOLOGICAL BASIS OF RESTITUTION OF EOsSt,PARDS BY C. M. CHILD Hull Zotlogical Laboratory, University of Chicago, Chicago, Ill. With One Ficure In a series of “Studies on Regulation” which have appeared in Roux’s Archiv and the Journal of Experimental Zodlogy during the last five years, and in certain other papers (Child ’o6a, ’06b), I have attempted to point out the essentially functional character of form-regulation and have defined regulation in general as a return or approach to physiological equilibrium after such equi- librium has been disturbed or altered (Child ’o6a). According to this idea form-regulation and functional regulation are both essen- tially the same thing. Itis perhaps unnecessary to state again here what I have repeatedly stated, viz: that the term “functional” is used in this connection in its widest sense as equivalent with “dynamic” or “physiological” and so includes all dynamic fac- tors in organic life. In other words, the problem of form-regulation is a physiological problem and not a problem swt generis as Driesch and various other authors have maintained. Let us consider for a moment what these assertions imply as regards the factors concerned in the determination of any par- ticular structure. If we assert that a given structure is altered or determined by functional conditions does not this assertion neces- sarily involve the idea of relation to its environment, intra-organic or extra-organic or both? As a matter of fact the very essence of the term “functional” as employed in these papers is to be found in the interrelation or correlation between the different parts of the organism and between the.organism and its extra-organic environment. Tue Journat or EXPERIMENTAL ZOOLOGY, VOL. V, NO. 4 486 C. M. Child That this could fail to be evident to any reader of these papers had not occurred to me until a recent paper by Prof. S. J. Holmes (Holmes ’07) came to my notice. ‘This paper is a restatement of the author’s symbiotic theory of form-regulation and a reply to certain criticisms of my own (Child ’o6a) of an earlier statement of this theory (Holmes ’04). Holmes maintains that my sugges- tions concerning the nature of form-regulation do “not contain any general principle of explanation for that functional substitu- tion and equilibration upon which it is assumed that form-regu- lation depends. But I suspect that when his theory comes to be developed so as to supply this missing element it will involve the assumption of some such symbiotic relation between the parts of the organism as I have assumed ” (p. 424). If I understand this assertion, it involves a serious misapprehension of my position. I have insisted again and again in my work on form-regulation in the interrelations or correlations between parts—in fact, certain of my papers have been concerned chiefly with showing that such rela- tions existed. Moreover, it is in consequence of the existence of such relations that I regard form-regulation as essentially a func- tional process. Even in my earliest papers positive statements on this point were made. ‘Thus, for example, on p. 219 of No 1 of my Studies on Regulation (Child ’o2) in a consideration of the general body-form of Stenostoma I wrote: “Every organism is what it is because of the relation of all its parts to each other and to the rest of the world. If any of these relations are changed the organism is changed.” And again in No. IV of the Studies (Child ’o4a) in a discussion of “formative factors:”’ ‘ All the com- plex activities of which organisms are capable are ‘formative factors:’ when we can view all of these in their complex inter- relations, then and then only shall we ‘understand’ organic form.” Also in No. V (Child o4b): “The factors of organic form include all the activities of organic substance as well as the environmental factors in varying degree. Indeed, in most cases, if not in all, we may regard organic form as the visible effect upon the protoplasm of functional factors in the widest sense”’ (pp. 468-469). In the later papers these interrelations are still more strongly emphasized. I have preferred not to designate them as symbiotic relations since Physiological Basts of Restitution of Lost Parts 487 I cannot see that anything is gained by the use of this term. More- over, it seems to me that many of the correlations are not really symbiotic at all, except in so far as they may be mutual with respect to complex parts of the organism. For example, the mere mechanical union with other parts is undoubtedly in many cases one factor in preventing parts of the organism from undergoing regulation into wholes. But I cannot see that it serves any useful purpose to call such factors symbiotic relations. It seems pref- erable therefore to maintain that these relations, or as I believe we may more properly call them, correlations, are physically and chemically of all sorts possible in the material and environment in which they exist. Moreover, while many of them are undoubt- edly mutual,i. e., reciprocal, at least as regards complex parts, others are just as certainly largely or wholly one-sided so far as form is concerned. It seems scarcely necessary to enlarge further upon this point. As regards the existence of relations between parts as an essential feature of regulation Holmes and I agree perfectly. As regards form-regulation we differ, in that it seems to me difficult or impossible to account for the facts on the basis of symbiotic relations, even in the widest sense. Holmes’ illustration of the process of regeneration is as follows: “Let us imagine an organism made up of a number of differen- tiated cells, each of which derives some advantage from some substances produced by the contiguous cells, and giving out some substance upon which the contiguous cells are more or less depend- ent. We will suppose that in addition to these differentiated cells, there are scattered through the body numerous indifferent or embryonic cells whose multiplication is held in check by the others, but which upon the removal of any part respond to the functional disturbance by growth and multiplication near the place of muti- lation. We may represent our organism by the following diagram in which the differentiated cells are represented by the larger circles 4, B, C, etc., and the indifferent cells by the smaller circles between them. Each cell such as 4 contributes something util- ized by B, G, and F, and derives something in return from each of these sources. Now suppose 4 is removed: the indifferent cell lying nearby, no longer held in check by the same stimuli, 488 C. M. Child begins to grow and develop. What line of differentiation will it most naturally take? Owing to the symbiotic relation existing between the cells differentiation in the direction of 4 will be most favored as this secures it the advantages which 4 received. In other words, this will be the direction of development along which social pressure will tend to guide it. And the result will be a regen- eration of the missing part” (Holmes ’o4, p. 282; ’07, pp. 420, 411). In 1906 (Child ’o6a) I criticised this illustration on the ground that if the cells or parts were mutually dependent, 1. e., if sym- biotic relations existed between them, removal of any one of them, e. g., 4, would bring about changes in the others in consequence of which their influence upon the undifferentiated cell substi- Fic. 1 tuted for 4 would be different from what it was originally, and hence the undifferentiated cell or part would develop—not into another 4 but into something else. There is logically no escape from this conclusion. ‘The removal of 4 results in the formation of a new system different from the original and must necessarily do so, except under certain limiting conditions to be discussed below. Holmes’ reply to my criticism is as follows: “According to Child, since the removal of 4 would alter B, G, F, etc., not only something different would be developed in place of 4, but the whole complex, according to my theory, would be profoundly altered. How far this tendency will result in a modification of these cells depends on the plasticity of the organism and the degree Physiological Basts of Restitution of Lost Parts 489 of mutual dependence of the parts—factors of course which vary in different organisms. But Child overlooks the fact that accord- ing to the symbiotic relation assumed, the other cells C, D, £, etc., tend to keep B, F, Gin their original condition. In so far as these remain in their original state, their influence on the indifferent tissue in the region of 4 will tend to mold it in the direction of the missing parts. In so far as B, G and F are modified through the loss of the missing part, their influence on the tissue in the region of A will come to be modified, and they will in turn modify the cells lying next to them. But, as there is a tendency for the modi- fications produced by the loss of 4, to spread successively to other parts, there is also a tendency, according to my theory, toward the checking and reversal of this process. Ifthe loss of 4 tends to modify B, F and G, the presence of £, C and D tends to hold them in place, and in so far as these are maintained through this influence they tend to mold the tissue in the position of 4 into the form of the missing part; and in so far as this is so molded, its modifying influence on B, F and G is diminished”’ (Holmes ’07, pp. 425, 426). I am unable to see that this argument shows that something like A may be generally replaced. Undoubtedly the modifying in- fluence of 4 upon the contiguous cells or parts B, F, G, is lessened by the presence of other cells or parts, E, C, D, but it is not reduced to zero in any case where the relations between parts are mutual. The balance between the “tendency for the modification produced by the loss of A to spread” and the opposite tendency simply deter- mines how great or how small the modification shall be. Some- thing more or less like 4 may undoubtedly be produced in many cases, but according to this hypothesis we should expect that the regenerated part would differ more or less widely from the original part in most cases. In fact, if the restoration of a part removed is purely a matter of interrelation between the various parts of the system, we must modify this hypothesis in either one of two ways to account forit. First: we may assume that the removal of the part, 4 in Holmes’ diagram, does not alter the other parts, B, G, F, etc., in any way which affects essentially their interrelations with the parts of the system. 4.90 C. M. Child In this case the undifferentiated material which in the absence of A is stimulated to develop, will develop into another 4. But in this case the relation between the original 4 and the other parts of the system is essentially one- iced and not mutual or sym- biotic. Or as a second possibility, we may assume that the rela- tion between 4 and the other parts of the system is such that removal of A produces modifications in the other parts only very slowly, while in the absence of 4 these other parts affect the undif- ferentiated cells in such manner as to bring about rapid develop- ment so that restoration is complete before the parts B, G, F, etc., have been appreciably altered by the absence of 4. Here the relations, though in the final analysis mutual, are so far as 4 and its restoration are concerned, one-sided. In short, if we accept the symbiotic theory as a basis, we can account for the restoration of a part like that removed only by additional assumptions, accord- ing to which the relations inv olved in the restoration become prac- tically one-sided rather than mutual. In my earlier criticism of Holmes’ theory (Child ’o6a, pp. 420, 421) the following statement appears: “To return to Holmes’ diagram, replacement of 4 can occur only when the relation is largely one-sided, i. e., when 4 is dependent on B--F, but these latter are not to any marked degree dependent on A. In this case, and in this case only, will the “‘social pressure” force the undif- ferentiated cell to differentiate into something like 4.”’ Holmes replies to this: ‘Where redifferentiation from new tissue is concerned, as in the present case, it is not the relation of A to B-—F, that should be more or less one-sided, but the relation of the tissue in place of 4 to this complex. This is an important distinction which Child does not seem to have considered. B—F are relatively fixed, the tissue in place of 4 is young and plastic, and more dependent so far as the direction of its differentiation is concerned, upon b-f, than these are upon it. We may grant that when regeneration occurs, the relation of dependence between the old parts and the new tissue is more or less one-sided, although the relations of the part removed may not have been. This would naturally result if the parts were relatively stable. They may be in a symbiotic relation, nevertheless, each part contributing in Physiological Basts of Restitution of Lost Parts 491 some way to the normal functioning of the others, and dependent to the extent that the removal of one part may alter only to a cer- tain degree the quality and quantity of the activity of the surround- ing parts, without producing extensive modification of structure or function” (Holmes ’07, pp. 426, 427). The first part of this argument seems to me to obscure the real point at issue. If the relation between 4 on the one hand and B-F on the other is not at least largely one-sided, removal of 4 must alter B—F, and if, as Holmes assumes, the new tissue which replaces 4 is more dependent on B—F than they on it, it becomes still more difficult to understand how the new tissue can replace A, for, so far as B-F are concerned, it does not at first take the place of 4 functionally. In the last sentence quoted, Holmes attempts to save his symbiotic theory after admitting that in regen- eration the relation may be more or less one-sided, by suggesting the existence of symbiotic relations which do not produce “‘exten- sive modifications of structure or function’? when one part is removed. It seems to me that such relations are negligible quan- tities so far as form-regulation is concerned, for if removal of a part of the complex does not produce extensive modifications of structure or function in the parts remaining, we must certainly conclude that the presence of this part is not essential for the main- tenance of the characteristic structure and function in the other parts. Evidently then this assumption does not relieve us from the necessity of assuming that the remaining parts are, so far as form and structure are concerned, practically independent of the part removed, 1. e., that the relations involved in form-regulation are largely one-sided in cases where restoration of the missing part occurs. It makes no difference whether we regard the persist- ence of B—F in essentially unchanged condition after the removal of A as due to “relative stability” or to real independence of 4. The fact remains that 4 can be restored only in case the other parts do persist essentially unchanged during the period between its removal and its restoration to a certain stage of development. And it is just as certain that Holmes’ symbiotic hypothesis cannot account for such persistence except by assuming the existence of special conditions which modify the relations between parts so 492 C. M. Child that they become essentially one-sided rather than mutual. If the restoration of a part like that removed were the exception rather than the rule, or even if it were less frequent, we might still accept the hypothesis. But a hypothesis which can account for the typical phenomena within its field, only with the aid of addi- tional special assumptions, which in this case amount practically to throwing over the hypothesis, can scarcely be regarded as satis- factory. The numerous cases already known where an animal is capable of replacing a part repeatedly after successive removals seem to me to furnish additional evidence against Holmes’ theory. Even so important a part as the head may be replaced repeatedly in many forms without appreciable change in character. It is scarcely probable, to say the least, that the relations between the head-region and other parts are one-sided in the sense that it is dependent on them, while they are independent of it. And if the relation is not one-sided in this sense, we should expect that repeated removals of the head would bring about essential changes in the other parts, even if the first removal did not. If such changes in the other parts do occur to any marked extent, it is difficult to understand how a new head like the old can be replaced time after time as the result of “social pressure,”’ for such changes in the old parts must alter the character of the “social pressure.” Here then Holmes’ theory leads us into something closely approach- ing a dilemma. Holmes continues: “If the parts B-F were more plastic, absence of 4 would naturally tend to cause greater changes in them, especially if new tissues were not produced in place of 4, which would come to assume some of the missing functions before the modification extended very far. There would then be a pro- gressive modification extending from the region of 4, which would tend to become less the farther it extended, but eventually perhaps affecting more or less the entire organism. Functional equilib- rium would then be maintained by working over the organism so that all the parts were adjusted to functioning on a smaller scale. The different methods of regulation, through morphal- laxis, regeneration and the various combinations of these proc- Physiological Basts of Restitution of Lost Parts 493 esses are, I believe, interpretable according to the symbiotic theory, and the relations of regeneration and morphallaxis to the degree of specialization of the parts which Child has elaborated, are, in fact, exactly what the theory would lead us to expect” (Holmes ’07, p. 427). Here Holmes fails absolutely, so far as I can see, to explain how and why equilibrium will be maintained. Certainly the “ pro- gressive modifications” resulting from the removal of 4 cannot bring the system back to its original condition: they must lead either to destruction of the system, or rather of the parts of it which remain, or else to a new condition of equilibrium different from the old. How does Holmes know that “functional equilib- rium would then be maintained by working over the organism so that all the parts were adjusted to functioning on a smaller scale?” What factor in the parts remaining compensates for the “progressive modifications” resulting from the loss of 4? Why should there be any compensation? ‘To none of these questions does Holmes’ hypothesis give any answer. According to the symbiotic theory as Holmes has presented it, the removal of a part is, at least in many cases, analogous to re- moval of a quantity from one side of an equation without change in the other. It is obvious that such procedure alters the value of one side of the equation in all cases except where the quantity removed is equal to zero. In short I believe that Holmes’ theory of regulation overlooks the most essentral feature in the process of replacement of a part removed. ‘This feature is the qualitative functional totipotence of the remaining parts after removal of the part in question. In other words, a part which has been removed cannot be replaced unless something remains after its removal which plays its part functionally in some degree. According to Holmes’ theory its place is taken by undifferentiated tissue, which is forced to develop into something like the part removed by the influence exerted upon it by other parts. But this undifferentiated tissue cannot exert the same influence on other parts as was exerted by the part removed. Moreover it is difficult to understand how undiffer- entiated tissue whose differentiation is held in check by the other 494. C. M. Child differentiated parts, could persist in a system such as Holmes postulated. If it has no function in the system and is not in symb.otic relation with other parts, why should it not disappear ? If, on the other hand, symbiotic relations between it and other parts exist, it should, according to Holmes, differentiate into some- thing. It is evident therefore that something besides undiffer- entiated tissue must take the place functionally of the part removed if replacement is to occur. We can most readily gain an idea of what this something is by means of a concrete example. In Planaria and various other triclads, where even small pieces are capable of replacing all parts, we find that the reactions of such pieces, while differing in degree from those of the original animal, do not differ essentially in kind After removal of the head, for example, the piece reacts in much the same manner as when the head was present, though more slowly and with less energy. In Leptoplana, on the other hand, where regeneration of a head does not occur after removal of the ganglia, the piece without ganglia is at once and clearly distinguish- able from the animal with ganglia by the character of its reactions. In Planaria then, and in the other forms where replacement of the head and ganglia are possible, the piece still retains in some degree the functional characteristics of a head-region. In remov- ing the head we have not removed the only region possessing such characteristics, but only the region which possesses them in the highest degree of any part of the animal. In Leptoplana removal of the head and ganglia leaves no part which can supply function- ally, even in slight degree, their place, and formation of a new head is impossible. We must conclude that the localization of visible structural differentiation in an organism is not necessarily coextensive with the localization of functional processes or conditions characteristic of th's region, but may be limited to the region of greatest energy of these processes or conditions. It is a well-recognized fact that the so-called functional structure of bone, tendon, etc., represents only the most frequent or most energetic functional conditions, and there is every reason to believe that similar relations exist between structure and function in many other cases. The case of Planaria Physiological Basis of Restitution of Lost Parts 495 cited above is in fact a demonstration that functional processes may be less sharply localized than the structures which represent them. The anatomical structure known as the head in Planaria is not the only region where “‘head-reactions”’ are possible, but it does represent the region where they occur with greatest energy and frequency in the normal animal. Admitting this, the question arises as to why heads do not form all along the body in Planaria, i. €., as to why structure should be thus more narrowly localized than function. The answer is not far to seek. If two parts, one of which is capable of reacting in a certain manner more rapidly and with greater energy than the other are correlated, the reaction to a given stimulus ail occur in the first part earlier and with greater energy than in the second. ‘The fact that a reaction has occurred in the first part must bring about changes in the system in consequence of which the character of reaction in the second part is altered. If structure is, as I believe, the visible expression of functional or dynamic conditions, we cannot expect that the second part, even though it possesses in some degree the same functional capacities as the first should exhibit the same structure, for the very fact of its correlation with the first part which possesses these capacities in greater degree determines that the functional con- ditions in it shall be different from those in the first part. In gen- eral terms we may say that the region where a particular functional complex occurs with greatest energy, frequency or rapidity domi- nates so far as this particular complex is concerned all other parts of the organism which possess the same capacity in less degree, and modifies their activities to a greater or less extent. Conse- quently the structure with which a particular functional complex is associated in the normal animal may be much more narrowly localized than the functional complex. In Planaria, for example, the head-structure is limited to the anterior end of the animal, while the functional capacities commonly regarded as charac- teristic of the head exist at all levels of the body. These other regions are capable of producing a head-structure, but only when isolated from the original head. It is evident then from this consideration that localization of visi- ble structure is not necessarily an exact criterion of localization of 496 C. M. Child functional capacity. In all cases where a difference in localiza- tion exists, structure is more narrowly localized than functional capacity. On this fact, which I believe to be of fundamental importance for the problem of form, depends the ability of a part to become a whole when isolated. In order to bring out clearly the difference between Holmes’ hypothesis and my own, we may make use of Holmes’ diagram (Fig. 1). According to my hypothesis, the various parts, 4, B, C, D, etc., though perhaps visibly different as regards structure, each possess the physiological properties of the others or of some of the others in some degree. B, Gand F, for example, the parts contiguous to 4, are capable in some degree of activities similar to those characteristic of 4, but as long as 4, a region of greater energy or frequency or rapidity as regards these particular activi- ties is present the correlations arising from it obscure, inhibit or modify the activities of B, G, F, so that they appear structurally and functionally to be different from 4. But when J is removed, the parts B, G, F become at once the dominating parts as regards the A-activities and the correlations between them and other parts become similar in kind, to those which previously existed between 4 and the other parts, though probably different in degree. In short B, G, F, or certain portions of them are substituted func- tionally for 4 simply because in the absence of 4 their activities must, by virtue of their constitution, be somewhat similar to those of A. No entelechy or other peculiar principle is needed to guide or determine this substitution. It occurs with the same certainty as any other physical phenomenon in all cases where these parts possess the functional capacities to which attention has been called above. ‘ According to this hypothesis, the undifferentiated cells postulated by Holmes are not only unnecessary, but could not substitute for 4 if present, because the parts B, G, F are more like A than are the undifferentiated cells and would therefore dominate in the process of substitution. The d-processes are undoubtedly in most cases, if not in all, at first less energetic or less rapid or both, than they originally were in A, and in consequence of this difference the system may regain its original condition of equilibrium in either one of two ways. If Physiological Basis of Restitution of Lost Parts 497 the other parts C, D, E are plastic, 1. e., if their activities are readily and rapidly altered by altered conditions, they will be affected by the decrease in 4-correlations following the removal of A and will undergo more or less change in response to the changed correlations, 1. e., regulation by what we ordinarily call redifferentiation will occur. If, on the other hand the parts C, D, E are relatively stable, i. e., not rapidly changed by altered conditions, they and the correlations arising from them will remain much the same as before the removal of 4. But the 4-processes in B, G, F are out of proportion to these correlations and must be quantitatively increased by them. In this case then equilibrium is regained by functional hypertrophy of the portions of B,G, F, which are the functional substitutes for 4. This is what we know as regeneration in the stricter sense, 1. e., formation of new tissue from the regions adjoining the cut surface and its visible differen- tiation with increase in size into a part like that removed. In most plants and in some animals regeneration occurs from regions more or less distant from the cut surface, simply because these regions are physiologically more like the missing part than is the region at the cut surface. As a matter of fact, since correlations in the system are at least in large measure mutual, most if not all cases of restitution are mixtures of redifferentiation and regeneration. Some change, 1. e., some redifferentiation occurs in some or in all parts of the system and some regeneration, 1. e., functional hypertrophy of the part which forms the physiological substitute for the part removed takes place. Holmes’ hypothesis fails to recognize the fundamental fact, viz: that something must remain after the removal of a part, 4, which can take its place functionally in the system in some degree. Without this the only factors which can prevent progressive depar- ture from the original condition are lack of plasticity in the parts remaining or one-sided relations between parts. As a matter of fact however plasticity is a conspicuous feature in many forms in which the regulation of parts into wholes occurs most readily, and on the other hand all the evidence indicates that correlations are in large measure mutual rather than one-sided. In those cases 498 C. M. Child where a part after isolation is incapable of becoming a whole, while the remaining parts are capable of replacing it, there is reason for believing that the correlations are more or less one- sided, i. e., the part in question has been so greatly modified by the past or present correlations arising from other parts that it has lost its totipotence and can never become a whole, but the correla- tions arising from this part have not been sufficient to modify the other parts of the system to an equal extent. Examples under this head are the appendages of arthropods, amphibia, etc. One other point discussed by Holmes requires brief considera- tion: in his first paper he selected the regulatory development of a head in Planaria as an illustration of the working of social pres- sure. In his discussion of this case differentiation is regarded as proceeding from the cut surface distally, in consequence of the social pressure exerted on the new parts by the old (Holmes ’o4, pp. 282, et seq.). In my criticism of this point (Child ’o6a, p. 421, et seq.), I called attention to the fact that in Planaria, and in other forms as well, differentiation of the regenerating tissue actu- ally proceeds in the opposite direction, 1. e., from the tip toward the base. In reply to my criticism Holmes (’07, pp. 427, 428) points out that the first visible differentiation is not necessarily the first actual differentiation, that “before any external features are produced in the development of a limb the main outlines of its differentiation may have been established through influence proceeding from its basal part, after which the tip might differen- tiate more rapidly than the intervening portion and the other visi- ble features of structure appear successively toward the base.” He also points out that in many cases the visible differentiation is centrifugal rather than centripetal and cites the case recently described by Zeleny (’07) of the antennule of Mancasellus, in which visible differentiation at first proceeds from the base toward the tip, but later in the opposite direction. He continues: “But granting that, in many cases, differentiation actually begins at the extremity and works toward the base of the regenerating organ, the process is not inconsistent with the point of view here set forth. We may suppose that the influence of the environ- ment causes the extremity of an organ to begin to differentiate Physiological Basis of Restitution of Lost Parts 499 like that of the missing part. That is only one step. We have then to account for the numerous coordinated differentiations that take place as the part develops toward the base. * * * The fact that, with few exceptions, such as the failure to regenerate the intermediate segments of the appendages, etc., the whole organ, nothing more nor less, is regenerated, and forms a congruent union with the basal part, is indicative of close interaction of the various parts of developing organs with the body of the organism at all stages of the process. “T am inclined to think that neither centrifugal nor centripetal differentiation, expresses the entire truth of the matter, but that the new part differentiates as a whole, much as organs doin embry- onic development, and at all times in intimate fictional relations with the old part, differentiation becoming accelerated in one part or another, according to special conditions’’ (Holmes ’07, pp. 428, 429). As regards most of these points my position does not differ very widely from that of Holmes. My criticism of his analysis of the case of Planaria was directed primarily, not at his hypothesis in general but merely at his failure to consider the actual facts in that case. [see no reason why the occurrence of differentiation in either direction or in both should constitute a fatal objection to his hypothesis or to my own, for such differences are merely incidental and depend on the conditions in individual cases. When my criticism was written the experimental data seemed to indicate that visible differentiation in the centripetal direction was the general rule, though by no means without exceptions, and since Holmes did not in his first paper attempt to account for this fact in any way, his hypothesis was open to criticism. I cer- tainly had no intention of maintaining that differentiation must in all cases proceed centripetally, since at that time various cases were known to me in which visible differentiation proceeded centri- fugally.t. Ido not believe however that Holmes’ suggestion that 1In his discussion of the direction of differentiation in the antennule of Mancasellus, Zeleny (’07, p 335) says: ‘Child has recently expressed the opinion that differentiation must in every case proceed from the tip toward the base and in no other way.”’ My actual statement was that differentiation from the tip toward the base is ‘‘a general ruel in cases of regeneration.” ‘This statement as it stands is 500 C. M. Child the new part differentiates as a whole, much as organs do in embryonic development is universally applicable. There are certainly many cases in which the terminal portions attain or approach their final condition of functional activity before the basal parts are formed, and in a considerable number of cases also the basal parts are replaced incompletely or not at all. In fact it seems to me that such cases might be expected to occur, for in a correlated system the conditions for the regulatory forma- tion of non-terminal regions must, at least sometimes, be largely dependent on the existence of typical functional conditions in terminal parts. If conditions in the terminal parts are more im- portant than those in the old parts as determining factors in the differentiation of intermediate parts we should expect to find the intermediate parts differentiating later than the terminal parts, but if, on the other hand, conditions in the old parts are the chief determining factors, differentiation might occur wholly in the centrifugal direction. Moreover, although I agree with Holmes that the absence of visible differentiation does not necessarily imply absence of physi- ological differentiation, I am incined to believe that the direction of progression of visible differentiation is not without significance as an indication of the direction of progression of physiological differentiation. In other words, while the absence of visible differentiation proves little or nothing with regard to physiological differentiation, its presence may prove something. I think it probable therefore that in some cases the regenerating part is not differentiated as a whole, but that its various regions are deter- mined successively in one direction or theother: in other cases it may perhaps be differentiated as a whole. It would appear that none of these possibilities conflict with either Holmes’ hypothesis or my own. undoubtedly open to misinterpretation and should have been qualified, for I was well aware at the time it was made that centrifugal differentiation occurred in various cases. In fact, I had shown in earlier papers (e. g., Child, ’o4b) that the differentiation of the intestine in regenerating parts of Leptoplana is apparently centrifugal. However I take the present opportunity to make acknowledgments to Holmes and Zeleny for calling my attention to this misleading statement, and also to make clear my teal position in the matter, which is that differentiation may occur in either direction or in both accord- ing to conditions in the particular case. Physiological Basis of Restitution of Lost Parts 501 To sum up: Holmes and I agree in that we both postulate a condition of physiological equilibrium, or rather, as I should put it a condition of oscillation or cyclical change about equilibrium, asthe basisof ourhypotheses. The chief pointof difference between us is that Holmes’ hypothesis does not, as | understand it, provide for the maintenance of or return to the typical condition, except by the assumption of relations largely one-sided, or that of lack of plasticity. While these assumptions may serve for certain indivi- dual cases, they seem to me to be totally inadequate for the analysis of form-regulation in general. According to my own hypothesis a part can be replaced only when some other part is physiologi- cally sufficiently similar to it to perform its functions at least qualitatively, if not quantitatively, after its removal. The independent formulation of two hypotheses of form-regu- lation so similar in general point of view as are Holmes’ and my own, is I believe not without significance, since agreement between different observers as regards the general nature of problems may be an indication that real progress inthe analysis of data is being made. It is desirable in such cases, and particularly in fields where the data are so varied and complex, that differences of opinion should be fully and critically discussed. For this reason I have ventured to consider at some length in the present paper the points which seem to me debatable, and to state my own posi- tion in a manner which I hope will lessen the chances of future misunderstanding. Hull Zodlogical Laboratory University of Chicago February, 1908 502 C. M. Child BIBLIOGRAPHY Cuitp, C. M. ’02—Studies on Regulation. I. Fission and Regulation in Stenos~ toma. Arch. f. Entwickelungsmech., Bd. xv, H. 2 and 3, 1902. ‘oga—Studies on Regulation. IV. Some Experimental Modifications of Form-Regulation in Leptoplana. Journ. Exp. Zo6l., vol. 1, no. 1, 1904. ‘ogb—Studies on Regulation. V. The Relation between the Central Nervous System and Regeneration in Leptoplana: Posterior Regen- eration. Journ. Exp. Zodl., vol. 1, no. 3, 1904. ‘o6a—Contributions toward a Theory of Regulation. I. The Signifi- cance of the Different Methods of Regulation in Turbellaria. Arch. f. Entwickelungsmech, Bd. xx, H. 3, 1906. ‘o6b—The Relation between Functional Regulation and Form-Regula- tion. Journ. Exp. Zodl., vol. in, no. 4, 1906. Houmes, S. J. ’04—The Problem of Form-Regulation. Arch. f. Entwickelungs- mech. Bd. xvi, H. 2 and 3, 1904. *o7—Regeneration as Functional Adjustment. Journ. Exp. Zodl., vol. iv, no. 3, 1907. ZeLeny, C. ’07—The Direction of Differentiation in Development. I. The Antennule of Mancasellus macrourus. Arch. f. Entwickelungs- mech, Bd. xxii, H.2, 1907. THE PROCESS OF HEREDITY AS EXHIBITED BY THE DEVELOPMENT OF FUNDULUS HYBRIDS! BY H. H. NEWMAN With Five Pirates AND SixTEEN Ficures 1N THE Text I Introduction........ eiaislatevel-e= Se eya eleva el epee Lelefcenetata abas/arata lala calainiayaiatapotelsveiajetetols)ateratheisiofelaletar 504 shia Materialvandtmethodse rrr yeratcry-tcrrnrarsielsteetotatatelats fel store ccieleterate aie ol (ajelsieleteketsteheleiets\-l=talat=ettcler= 506 Ai” Materials—A description of the) speciesiused << .-/ecie.e «ieiemicie le rie e)einie ciel ele niririeleinisieie 506 1 Morphological differences between the adults............-0-0 eee eee eee eee e eee 506 2 Physiological differences between the adults.......-.. 20.2000 -es eee ee ee eeeeeee 507 3 Morphological differences between the eggs and embryos.........-.----+++-005+ 507 4 Physiological differences between the eggs and embryos........-.--0++-+seeeee 508 i} NIGER EE gocbopa ns abanncias doc donde cddpoonone dadansongsradammadaavodnosondodos 509 1 Spawning behavior and sexual dimorphism.........-.-.0+0--eeeeeeee eee eeees 509 2 The importance of equalizing the physiological condition of the parents........... 510 3 The importance of equalizing the external conditions of the developing embryros .._ 512 4 The attitude that must be taken toward variability...............0-eeeeeeeeeee 512 Fe Nats Neos aon dcou ods bune=BbonooDUunoD coud beUEdepEDsDasHaddecgpsoac 514 III Description of Experiments and Presentation of Data..............0s eee eee eset eee en eee 517 A Data derived from the study of living material. .............00seeseeeceeseecenceece 517 Lp “Wy pe'series’ (Sertesil)) sists, <\at=yeta1=j stores aa Fig. 12 History of a race derived from a crooked specimen. The crookedness had disappeared in the third generation. The individual ab had the posterior end crooked (Fig. 13). When this animal was placed in the culture fluid, it became plumper, and the abnormality of form was less marked (”). When it divided the anterior product was of the normal form, while the posterior product had the posterior point slightly dis- placed toward the aboral side, but was otherwise normal (Fig. 1332) When it again divided, its progeny were both normal in form. The case of aj Belones partly in the second category, partly in the third. The body of the parent aj was small and irregular, with a broad anterior end, which bore on one angle a projecting point (Fig. 14, +). When this was placed in the culture fluid it did not divide for three days. [he body increased in size and especially in thick- Heredity in Protozoa 611 ness, and the projecting angle became more marked (Fig. 14, "). On the third day it divided; the posterior product was normal in shape, though smaller than usual, while the anterior product was extremely irregular, having the form shown in Fig. 14, *. In the next twenty-four hours this irregular structure underwent a partial division, increasing its size and irregularity of form (Fig. 14,°). The structure thus produced was double, since it had two mouths (77), both of which took food; and there were two independ- ent protoplasmic circuits for the digestion of food. During the next twenty-four hours this structure divided into two very unequal parts. One product was a short, somewhat Fig. 13 History of a race derived from an individual with a crooked posterior tip. The irregularity had nearly disappeared in the second generation; in the third (not shown) it was quite gone. irregular individual. ‘The larger product was still very irregular; it represents three united individuals (Fig. 14, ‘). The smaller product divided again, producing progeny that were normal in form, though small in size. The larger product, composed of three incompletely separated individuals, did not divide again; after two days it disintegrated. 3. The individual aq represents mainly the third category, in which the irregularity of form is increased in reproduction, till death occurs. This specimen was curved as shown in Fig. 15, a. At its first division the products did not completely separate, but formed the structure shown in Fig. 15,5. At the next division the right half divided in such a way as to produce one nearly nor- 612 H. S. ‘fennings mal free individual, while the other product remained attached to the left half. The latter underwent a partial, irregular division. Thus the result is to produce an irregular structure consisting of three fused individuals (Fig. 15, c). Fig. 14 Diagram of the history of the race derived from the irregular individual aj. In the third and fourth generations double and triple monsters appeared, with several mouths (m) and multiple proto- plasmic circuits. Two such circuits are shown by arrows at 3. Heredity in Protozoa 613 This structure underwent other partial fissions, giving the irreg- ular monster shown in Fig. 15, d. This lived for about four days, then disintegrated. 4 Instances of the fourth category, in which no divisions occurred, are given by aq' (Fig. 16,4) and am??? (Fig. 16, db). . S Fig. 15 History of the race derived from the irregular individual aq. See text. These both lived for five days without dividing or taking food; both then disintegrated. The mass ar was the result of partial fission, so that it included several partial individuals. As successive partial fissions occurred 614 TTS; ‘fennings it took various forms, of which the three given in Fig. 16, c, d, e are types. This structure took food by five or six mouths, and had a number of partly independent systems of circulation. It reached a length of 450, with a breadth of 150”. The normal Paramecia in the same culture in which it occurred showed dimensions of about 1504 x 60. This structure had therefore the bulk of about twenty normal individuals. This was kept for ten days, but finally it disintegrated. Fig. 16 Irregular individuals which do not divide farther. a and b are separate individuals that finally died. c,d and ¢ are stages taken at intervals of several days in the complex mass ar. _m, mouths. f Behavior of Mutilations in Reproduction Paramecium differs from many of the infusoria in the fact that it does not stand mutilation well. The internal contents seem very fluid, so that they flow out as soon as the ectosare is cut; the animal at once disintegrates. It is therefore difficult to study the regulation of injuries in this animal, either during the active life, or at reproduction. However, from a large number of experiments, certain results were reached that show how mutilations behave, both in ordi- nary regulation and in reproduction. Heredity in Protozoa 615 1 Mutilations in adults. Whenever the ectosarc is punctured, the internal contents flow out and the animal dies. But in a few cases mutilations were produced without puncturing the ectosare. Thus, a fine glass rod was drawn across an individual near its middle; leaving a deep constriction, while the two halves of the body were swollen (Fig. 17, a). “Fhis constriction persisted for some hours, becoming gradually less marked. The next day the animal was perfectly normal. In another similar experiment, blister-like swellings were pro- duced, and the anterior portion of the body became totally irregu- lar (Fig. 17,6). But within 24 hours the normal form was com- pletely restored. Thus it is clear that the adult Paramecium has the same power of regulating form that is so well known in Stentor and other infu- b Fig. 17 Mutilations produced by drawing the tip of a glass rod across the adult animal. See text. soria. But this can come into play only when the injury has not been of a nature to puncture the ectosarc and so to cause disinte- gration. Many attempts were made to remove only a part of the internal fluid (endosarc), without causing death. The ectosare was pierced with the tip of an excessively fine capillary glass rod.* But in all cases where any of the endosarc flowed out, the remain- der followed, and the animal died. 2 Mutilations in dividing specimens. It was thought pos- sible that specimens undergoing fission might show a different physical state of the protoplasm, such as to permit mutilations without immediate disintegration. To a limited extent this was *These can easily be made so fine that the tip is apparently not larger than a cilium of Parameciume 616 lake ss fennings found by experiment to be true. When a specimen undergoing fission is pierced with the tip of the glass rod or otherwise muti- lated, it does not go to pieces so rapidly as the adult, though in most cases it finally disintegrates. But in a few instances speci- mens thus treated survived. Thus, while the Paramecium ma was undergoing fission, its anterior half ma’ was pierced with the rod, allowing a part of the internal contents to escape. This half became distorted (Fig. 18, a) while the other half became swollen. The latter resumed later its normal form, and fission continued. The injured half a retained its distorted form (Fig. 18, 4). During growth the form became somewhat nearer normal (Fig. 18, c), but complete regu- lation did not take place in this generation. Fig. 18 History of the specimen ma, mutilated during fission. See text. During the night the irregular individual divided. ‘The anterior product was quite normal in shape; the posterior one still showed a slight irregularity of form at the posterior end. At the next fission this disappeared and both products were normal. Thus the effects of the mutilation persisted in some of the indi- viduals for three generations, then disappeared. In a number of other cases young or dividing specimens were marked with deep furrows by pressing them with the rod. These marks lasted some hours, but disappeared before the next fission occurred. In the dividing specimen mb the posterior part mb? was pierced with the glass rod, so that a part of its contents escaped, while by contraction most of the remainder of its contents were forced into the anterior half mb‘ (see Fig. 19, 6). Thus the Heredity in Protozoa 617 pierced part became very small; later it increased in size and became irregular (Fig. 19, c). The fission was never completed, this irregular part remaining attached to the posterior end of the normal individual mb. The normal part mb' divided twice, budding off, as it were, two normal individuals at its anterior end; its posterior part remained with the irregular mass attached, as in Fig. 19, d. At the next division the two components remained connected, with the irregular mass attached to the posterior end (Fig. 19, e). ee ere e) O ” Fig. 19 Effect of mutilation during fission in the specimen mb. See text. The irregular mass had itself made some attempts at fission, with the result that it became still more irregular. There was no further change for three days; then another partial fission produced the results shown in Fig. 19, f. During the next day the entire structure disintegrated. In this case the effects of the mutilation lasted for several generations, finally causing death. All together, it is clear that while mutilations may be passed on bodily to certain of the products of division for a number of gener- ations, there is no tendency for them to be inherited by all the 618 Tel eek ‘Fennings progeny; no tendency for the mutilation to be duplicated in new individuals. There is no tendency to produce a race of mutilated individuals, any more than there is in Metazoa. Regulation takes place at the time of fission, so that after several fissions the normal condition is restored. 4 Acquired Characters That Tend to be Inherited g Acquirement and Inheritance of a Tendency for the Adults to Remain United in Chains The acquired characteristics thus far described have shown no tendency to be inherited in such a way as to produce a race bearing the new character in question. We now come to a case in which such a tendency actually showed itself. The difference between this case and the others is instructive, suggesting what must be the essential nature of an acquired character that may be inherited. The characteristic in question is a tendency for the adult indi- viduals to remain united in chains. This tendency appeared in the line a, which we have already described in connection with. the transmission of a long spine (pp. 589-604); the beginnings of the characteristic now under consideration have been set forth in that description. In the process of growth the broad base of the long spine (Fig. 4,7) became drawn out, till in the individual a'?124242 it formed a ridge running along the aboral surface nearly the entire length of the body (Fig. 4, °). At the next fission it was found that the fission plane did not pass so readily through this ridge as through the remainder of the body, so that the two resulting individuals did not separate, but remained connected by a bridge passing from the aboral surface of one to that of the other (Fig. 4, 27°). The continued union of the two individuals after fission reap- peared in succeeding generations, both in the individuals formed from the region anterior to the spine (as in Fig. 4, 1°), and in those formed from the region posterior to the spine (Fig. 4,17 7°). In the eighteenth and twenty-first generations three individuals formed a chain (Fig. 20, a). In succeeding generations many such connected individuals and chains were formed. In the fif- Heredity in Protozoa 619 teenth generation | began to save all the progeny of a; up to this time only the specimen bearing the spine had been kept alive. In the large number of progeny thus obtained many variants were to be observed in the matter of interconnection. Many individ- uals were free and separate. Pairs of united individuals were very common. Chains of three to eight or more (Fig. 20) were not uncommon. ‘These longer chains were likely to break apart in the course of time, as a result of their bending and twisting move- ments. Fig. 20 Chains of individuals formed in the history of the race a, as a result of incomplete fission. There was much variation in the extent and strength of the union. Sometimes there resulted from the division of united indi- viduals specimens that were quite free. The division of free specimens often produced united pairs. In some cases the con- necting band was very thick and strong, so as to hold the two specimens inflexibly in various positions (Fig. 20, 6). In other cases the fission was so incomplete that mere partly double specimens resulted (parts of ¢, Fig. 20). Finally, the irregularities 620 H. S. fennings of fission at times went so far as to produce mere monstrosities (parts of d, Fig. 20). Such monstrosities were rare, while indi- viduals neatly united in pairs or in chains were very common. The first occurrence of such unions (Fig. 4, 1°) was on May to. Cultures were kept in watch glasses from that time till July 1 (probably about fifty generations); at that date the unions were still abundant. In fifty generations the original individual which underwent the modification causing the union would have pro- duced progeny to a number running far up into the billions. Effects of Artificial Selection On June 22 I began experiments to determine the effect of selec- tion on this peculiarity. Would it be possible by selection to pro- duce on the one hand a series showing little or no tendency to remain united, on the other hand a series in which most or all the individuals remain in united pairs ? Two selected cultures were started in watch glasses. ‘The first contained twenty individuals united two by two in ten pairs. The second contained twenty free individuals (descended from the same ancestors as the united pairs). Forty-eight hours later (June 24), both sets had multiplied to about 100 specimens. In the first set (ancestors united) there were ten united pairs. In the second set (free ancestors) there were two united pairs. From the first set | removed all the free individuals, leaving only the ten united pairs. From the second set the two united pairs were removed, leaving all free. The further history was as follows: Culture from free ancestors. On June 25 this had multiplied to 200-400; among these were three or four united pairs. I re- moved the latter and retained only 100 of the free individuals. On June 26 these had multiplied two to four times but contained no united specimens. This culture was kept for a week or so longer, but developed no more united pairs. Thus, selection had quite removed from this set the tendency to remain united. Culture from united ancestors. After the second isolation of ten united pairs (June 24), the number multiplied to about 50 in 24 Heredity in Protozoa 621 hours; among these there were eight groups of united individ- uals—some of two, some of several, specimens united in chains. The eight groups were again isolated ( June 25). Effects of Natural Selection These eight groups showed many imperfect individuals, and the groups were at a great disadvantage as compared with the free individuals. ‘This was because they are not able to swim about actively, but must lie at the bottom and move about only irregu- larly. As a result they get comparatively little food, and are not able to avoid regions where the conditions are harmful. The bac- teria multiplied much more rapidly than in the free culture, con- taining many individuals—the latter keeping down the number of bacteria by feeding on them. In consequence of these bad conditions, the united groups began to die. Some multiplied farther, all the individuals remaining united. But forty-eight hours after the isolation of the second lot of eight groups, all were dead. Thus it is easy to produce by selection a culture containing only free individuals and multiplying in the usual way. Artificial selection will likewise produce a culture of united specimens, multiplying mainly by incomplete fission. But at the same time natural selection acts; these groups die, owing to their inefliciency in getting food, keeping down the bacteria, avoiding harm, and in the performance of the general bodily functions. This extinction by natural selection of the series multiplying by incomplete fission was shown in another way. A considerable number of the progeny of a, with both separated individuals and united groups, was allowed to accumulate in a shallow watch glass. Here the united groups flourished fairly well, because the vessel was so shallow that they received plenty of oxygen and of food while lying on the bottom, while the undue multiplication of the bacteria was prevented by the numerous free individuals. Now the culture was transferred to a large vessel, three inches deep. Here the culture multiplied enormously, but all the groups of united specimens quickly disappeared. “They sank to the bottom 622 lal S. ‘fennings of the vessel, where the conditions were not such as to keep them alive, while the free individuals remained at the top and multi- plied. Thus by continued natural selection all specimens mul- tiplying by incomplete fission were removed, and in a few days the deep culture contained only normal, free individuals. In shallow cultures, on the other hand, the united groups persisted for about two months, as we have seen. In this case then we have a new characteristic, of known origin, that is inherited by many individuals for many generations, and is finally extinguished only by the action of natural selection. The many other new characteristics that we have described were not inherited (save as they were handed on directly to a single speci- men). In the one case the new feature becomes a race charac- teristic; in all the others it fails to do so. WHAT MUST BE THE NATURE OF -A NEW CHARACTER, THAT IT MAY BE INHERITED? What is the peculiarity of the characteristic that was thus mul- tiplied and inherited, and what light does it throw on the question as to what must be the nature of an acquired characteristic in order that it may be inherited ? The characteristic thus inherited was a modification of the pro- toplasm of the cell, of such a character as to cause 1t to behave differ- ently in reproduction. The other characteristics, not inherited, were not such modifications of the protoplasm as to cause it to behave differently in reproduction. Consideration of the facts of normal reproduction in the Pro- tozoa, and of heredity in general, indicates that this difference is an essential one. In order that it may be inherited (by more than one of the progeny), a characteristic must be the result of such a modi- fication of the mother cell as will cause it to behave in a certain way at reproduction. It makes no difference whether the mother cell in question is a germ cell, in a Metazoan, or a differentiated Pro- tozoan. Thus we know that in the inheritance of the seta of the Hypo- tricha, for example (Fig. 21), these are not simply handed over in Heredity in Protozoa 623 finished form, like the spine of a (Fig. 4), but are produced anew on each product of fission. The old seta and cilia degenerate and disappear as fission sets in. Inthe daughter individuals the new seta appear in a small group with a totally different arrangement from that seen in the adult parent (Fig. 21, ») and the final arrange- ment 1s reached by complicated processes of differentiation and distribution. Thus the presence of setae in the posterity could have been brought about in the beginning only by such modifica- tions of the protoplasm of the mother cell as would cause it at fis- sion to produce sete. Any change in the structure, number, or Fig. 21 Dividing Stylonychia, from Biitschli, showing at « the appearance of the new sete in a close group arrangement of the setz could result only from such a modification of the mother cell as would alter in a definite way the processes occurring at reproduction. The thing transmitted from the parent cell to the young progeny is, not the seta themselves, but the change in the protoplasm causing the production of sete in a definite way. To return to a specific problem—How then could such a local- ized appendage as the spine of a (Fig. 4) become an inherited char- acteristic? Only through such a modification of the protoplasm 624 H. 8. “‘fennings of the parent cell as would cause at fission the production of such an appendage on each of the progeny. At first thought it appears difficult to conceive how this could occur. This will be made easier, perhaps, by a consideration of the origin of certain characteristics in the race a (Fig. 4, etc.). Examples of Modifications from which New Inherited Charac- ters Might Result Let us take first the origin of the spine whose history is traced in Fig. 4. The original ancestor of the race a was without spines. But it was so deformed and modified that at the time of fission two short teeth were produced during the processes of division (Fig. 3,7). At the next fission one of these short teeth formed as it were a region of weakness, where a long spine was pushed out, as an accompaniment of the processes of fission (Fig. 3, * °). Such a region of weakness might well exist without a visible tooth to show its position; then at fission a spine would be produced in this spot. It is evident that active physical and chemical processes are in progress at the time of fission; these may easily result, under the influence of a local modification of the parent cell, in the push- ing out of a spine or other’structure of characteristic form. How such a new structure might appear in each of the progeny of each generation is illustrated in a simple way by certain other phenomena seen in the race a. As we have already set forth, the progeny of a showed after a certain period a tendency to remain united in chains. At the same time there appeared among the free progeny of a a considerable number of individuals which bore at one or both ends a spike-like point (Fig. 22). This character did not become general, but so many cases appeared that one might say that there was an inherited tendency toward this. Ob- servation of the process of fission showed that these points arose by the pulling out of the protoplasm while in the plastic condition at the time of fission; the two young animals were connected, at a certain stage, by a bridge of this plastic protoplasm. By their movements they drew this out to a long strand, which finally broke at the middle, leaving a point at the ends of the two animals. Heredity in Protozoa 625 When this happened at successive fissions, the animal bore such a point at each end. It is evident that these points are due to the same cause that pro- duced the inherited tendency to remain united in chains (as in Fig. 20). They result from the ridge of new material along the aboral side of the animal, shown in Fig. 4, °. Now, it is easily conceiv- able that this new material might be of such texture and thickness that it would always be drawn out at fission in such a way as to produce points of a definite form and size. These would then appear regularly after fission; a race of Paramecia with this as Fig. 22 Examples from the race a, of individuals having a point at the posterior end, due to the draw- ing out of the connecting band at the time of fission. a new characteristic would have been produced. ‘The spine would be hereditary, because produced anew in each generation, just as are the setae of the Hypotricha, or the organs of the Metazoa* SUMMARY AND GENERAL DISCUSSION The following general statements of the laws and principles bearing on heredity} that result from our investigation are made with direct reference to the Protozoa, and will best be grasped by keeping in mind concrete cases, such for example as those shown in Fig. 4, Fig. 20 and Fig. 22. *It is of course possible that the origin of new permanently inherited characters is not normally through mere modifications of the external parts of the cell, such as we see in our illustrative cases. Possibly there must be originally some modification of more recondite parts—nucleus, chromosomes, or the like— and that these then secondarily act upon and change the outer parts. This would add farther compli- cation, but would not change the essential point, which is, that in order that a characteristic may be inherited, it must be due to some modification that causes a change in the processes of reproduction. +For a summary of results on other matters than heredity (on the changes during fission and growth, etc.), see pp. 599-604. 626 Jake S. fennings 1 The “inheritance of acquired characters’ meets the same difficulty in the Protozoa as in Metazoa. In both Protozoa and Metazoa most characteristics acquired during the lifetime of the individual are not inherited, and such inheritance does not occur more readily in the one group than in the other. 2 The difficulty with the “inheritance of acquired characters” lies, not in the separation of soma and germ, but in the process of cell division. Ifa cell bears a structure at one end, there is no simple and direct reason why, when it divides, both the cells pro- duced should bear the structure, and observation shows that they do not, in the case of new structures. ‘There is no evident way in which a structure of this sort can overleap the barrier of cell divi- sion and appear on the other side.* If we insist on making a comparison between the condition in the Protozoa and the separation of soma and germ in the Metazoa, the following is the state of the case. If any Protozoan cell (as in Fig. 7) is to be divided at the next fission into two parts a and p, then, so far as inheritance of new structures is concerned, a stands to p as soma to germ, and reciprocally, p stands to a as soma to germ. In other words, there is no evident transmission, and no evident mechanism for transmission, of a new structure from a to p or the reverse, just as there is no evident mechanism for trans- mitting a structure from soma to germ. 3. In order that a character may be inherited (by more than one of the progeny, so as to produce a race), it must be produced anew in each generation. ‘This is what happens in the normal reproduction of both Protozoa and Metazoa. 4 In order that a new (or “acquired”’) character may be inherited, it must be the result of such a modification of the parent cell as will cause a change in the processes of reproduction; and specifically, precisely such a change in these processes as will pro- duce the character in question. ‘This is equally true of Protozoa and Metazoa. 5 Most characteristics acquired during the life-time of the * This will be most readily grasped by looking at the figure of a typical case, such as Fig. 4,%. Why, when this animal divides transversely, should there be a spine upon the posterior (left) half, as well as upon the anterior (right)? As a matter of fact, there is mor. Heredity in Protozoa 627 individual are not the result of such modifications of the parent cell as will cause a change in the process of reproduction such as to produce anew these characteristics; hence they ar: not inherited. This is true in both Protozoa and Metazoa. 6 Thus the problem of how new inherited characters arise is the same in Protozoa as in Metazoa. We may therefore work on the general problem as readily in the one group as in the other, and there is no reason why the principles reached in one group should not apply equally to the other. Thus a new line of attack on the problem is opened; in view of the rapid multiplication of the Protozoa and the ready accessybility of their reproductive cells both to environmental influences and to observation, this gives some marked advantages. 7 The search for the origin of new inherited characters (in both Protozoa and Metazoa) resolves itself experimentally into a search for agencies and processes which will permanently modify the cell in such a way as to cause it to act differently in reproduction. 8 When a given structural characteristic arises during the reproductive processes so as to appear in a given generation, that is not because the same structlire was present in a preceding g gen- eration. Often indeed it was not present before; its origin is due to some change in the constitution (chemical or structural?) of the preceding reproductive cell. Thus, the production of a spine such as we see in Fig. 4 is evidently due to a spot of weakness at a certain point in the cell body, causing a protrusion during fission. Such a structure might result from the localized presence some- where in the cell body of a certain chemical compound, which would react at a certain stage with some other substance, thus producing a spot of weakness, where a spine would be protruded. So, the appearance of the new anterior sete in the posterior prod- uct of division in the Hypotricha (Fig. 21) is evidently due in some way to the constitution of the cell. 9g Thus, then, the cause of the appearance of a certain struc- ture in a certain generation is some other peculiarity of the cell producing it; some chemical peculiarity, for example. We may generalize this by saying that the appearance in the progeny of a certain structure } is due to the existence in the mother cell of a quite different condition a. 628 Jala We ‘fennings 10 It follows from what has been set forth in the paragraphs preceding, that in the production of a new inherited character the original modification will be something quite different from the visible structural characteristic which later appears in consequence of it. The original modification will be some chemical or struc- tural change in the reproductive cell or cells that are later to pro- duce the structure in question. (By producing in Paramecium a localized change in the character of the protoplasm, a spine is later produced at that spot, etc.) The first appearance of the visible structure is one generation after the production of the modifi- cation to which it is due. 11 Not all modifications of the germ cells that result in the production of a new character in the next generation, will result in the repeated production of this character in succeeding genera- tions. In most cases, the new structure appears but once, and is not inherited. In order that the new structure shall be inherited, the original modification to which it 1s due must be transmitted to the succeeding generation of germ cells. This is by no means a matter of course; in fact, it is something not to be expected, as a rule. The cell usually, by regulative processes, throws off after a time any modification which the environment has impressed upon it. Many examples of this are seen in the foregoing pages. Certain unusual conditions of the cell result in the production, at the next fission, of a spine. But during fission regulation occurs; the unusual condition disappears, and the spine is not again pro- duced. This is doubtless the fate of most modifications of the cell. We saw, however, one modification which persisted, producing its effect in succeeding generations (pp. 618-622). Of such a nature must be all modifications which produce new inherited character- istics. It 1s easy to so modify the cell that new characteristics shall appear in one succeeding generation; to so modify it that the new characteristic shall appear regularly in succeeding generations is a totally different matter. We often hear it pointed out that heredity is not transmission, but new production; and this has been emphasized in the pre- ceding pages. But it needs to be realized that while it is true that Heredity in Protozoa 629 the inherited structure visibly appearing is not transmitted, sore- thing is transmitted, namely, the condition of the protoplasm which causes the production of the visible inherited structure. If this determining condition were not transmitted, the visible structure could not be produced in each generation. It is this “something” transmitted that lies at the basis of the figurative expression “bearer of heredity,” or the like. 12 What sort of modifications will remain permanently and be transmitted to the progeny? Evidently, only such modifications as are not removed by the regulatory processes of the cell. The modifications that are removed by regulation are precisely those which interfere in one way or another with the physiological proc- esses of the organism, while modifications which arise in harmony with, or as a result of, the normal functioning of the cell are not removed by regulation. Thus only characteristics of the latter class—namely, adaptive characteristics—will be retained and transmitted. Furthermore, it appears clear that the successive modifications in the reproductive processes induced by these adaptive characteristics must likewise be in harmony with the normal functioning of the cell, else they would be removed by the known regulatory activities of the cell. Thus all stages in the modification, including the final one, must be in harmonious adjustment to the normal activities of the organism. It would appear therefore that only the new characteristics that are adaptive will be inherited. Anything not in harmony with the normal func- tioning of the cell will be removed by regulation. 13 Let us now examine the problem of the “inheritance of acquired characters.” What processes would be required for the inheritance by the progeny of the same characteristic that has already been produced directly in the parent, by environmental action ? Keeping the Protozoa in mind, we have evidently two cases here: a If the “acquired character” is some general chemical or structural change in the parent cell—something that affects the cell as a whole—then there appears to be no special difficulty in the way of a direct transmission of this to the progeny, provided 630 al. S. fennings it is not thrown off by regulation. _ If new inherited characters of any sort are ever produced by environmental action, such direct transmission of an acquired internal modification must occur, as we have already seen (paragraph 11). In the Metazoa, it would evidently be only general changes in the germ cells that would be thus directly transmitted. b The case of a new /ocalized modification or of a definite new structure, such as a spine, which is directly produced by environ- mental action, is wholly different. As we have already seen (para- graphs g, 10, 11), in order that a new localized structure 4 shall appear in each generation, a certain other condition a must be produced in the mother cells; this condition a must be transmitted from generation to generation, and must so modify the reproductive processes as to cause, at each fission, the production of the new structure b. Now, if the new structure b was first produced directly in the parent by environmental action, and is then to be inherited, the processes required are the following. The existence of the struc- ture ) (a spine, for example), in the parent cell, must cause the production in that parent cell of precisely the “other” condition a, that is of such a nature as to so change the processes of repro- duction that they will again produce identically the character } (the spine) which had first been produced by the environment. Or, what amounts to practically the same thing, the environment must coincidentally produce two heterogeneous effects: (1) it must directly produce the structure b; (2) it must produce some permanent change a in the constitution of the cell, such as_ will so modify the processes of reproduction that they in their turn will produce the same structure 6: Such coincidental production of a complex structure b in two quite heterogeneous ways would be most extraordinary, and we have as yet no glimmering of a mechanism by which the coinci- dence could be produced. Moreover, as we have seen, in most cases (in all precisely observed cases) it is not produced; we have little if any direct evidence that it ever occurs. Yet if it ever occurred it would be of such importance that we must of course continue to be on the watch, in all experimental Heredity in Protozoa 631 work, for any evidence of it. The question, put as simply as possible, is as follows: Is there ever any mechanism or property in virtue of which, when a structural modification occurs in one part of the body, this will modify another part of the body (not in the same way, but) in such a way that this other part will, at reproduction, start up processes tending to produce a similar structural modification ? 14 The propositions thus far set forth have had direct refer- ence to the Protozoa; but in the main they apply a fortror: to the Metazoa also. The difference between the two groups as to heredity is not one of principle, but of complexity. The extreme difference in complexity may be put as follows: In the Protozoa, when a new inherited character is to appear in the adult, this requires a modification of the adult of the pre- vious generation, of such a character as to change in a definite way only the next fission and processes immediately connected with it. This requirement is sufficiently complex when we come to ask how the numerous locomotor organs of the Hypotricha, in their varied typical patterns, have arisen and become hereditary. But it is not to be compared in complexity with what we have to set forth next. In the Metazoa the requirement for the sppeotanes of a heredi- tary new structure in the adult is that the preceding germ cell shall be so modified that at the next fission the reproductive proc- esses shall be changed, but the change shall not yet be of a char- acter to produce the ultimate structures. In the next and the next, and in hundreds of succeeding fissions the processes must all be modified so as to keep in each cell the conditions for the final production of the ultimate new structure. These conditions will necessarily be different in the different cell generations, as differ- entiation occurs, and of course each of the intermediate condi- tions is something quite diverse from the final structure. At the end the new structure is produced, not by a modification in the re productive processes of one cell, as in the Protozoa, nor by the same modifications in many cells, but by the diverse modifications of thousands and thousands of cells, all so modified as to cooperate in the production of the final structure. The mind refuses the useless attempt to conceive of such complexity of change. 622 H. 8. fennings 3) As Conklin (’08) has so well set forth in a recent address, “the mechanism of heredity is merely the mechanism of differen- tiation.” The questions with which we have to deal are those as to the nature of the determining conditions and of the processes, by which the constitution of the cell changes. Perhaps the most direct study of heredity possible in the Metazoa is such a study as Conklin is making of the internal determining conditions in the differentiating cells of the developing organism. When one comes to the study of heredity in the Protozoa, this simply coincides with a study of the determining causes of differentiation. Johns Hopkins University Baltimore, Md. March 10, 1908 LITERATURE CITED Carkins, G. N. ’02—Studies on the life history of Protozoa. I. The life history of Paramcecium caudatum. Arch. f. Entw.-mech., 15, 139-186. ’06—The Protozoan life cycle. Biol. Bul., x1, 229-244. Conkun, E. G. ’08—The mechanism of heredity. Science, xxvil, 89-99. Jenninas, H. S. ’06—Behavior of the lower organisms. 366 pp. New York. Jensen, P. ’07—Organische Zweckmissigkeit, Entwicklung und Vererbung vom Standpunkt der Physiologie. 251 pp. Jena. Kettoce, V. L.’07—Darwinism today. 403 pp. New York. Roux, W. ’81—Der Kampf der Theile im Organismus. Leipzig. Tower, W. L.’06—An investigation of evolution in chrysomelid beetles of the genus Leptinotarsa. Carnegie Inst. of Washington, Pub. 26, 320 pp. Wurman, C. O. ’99—Animal behavior. Woods Hole Biol. Lectures for 1898, pp. 285-338. berrielt ie omy “ik aa Waa Deere? Wi QL The Journal of experimental 1 zoology PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET ee ee ee UNIVERSITY OF TORONTO LIBRARY (ea ee ee ti G : ae : Sess St