ae ibtstel an tH ae i} ¢ 2%. SA ar redaete tetas ceserise + Ses ee So ge eget Fe = - eee Hi tFifot mnie oy, y, : piensa tet aie Mite NTA SH bd Sint HAD of - Hai he ea aeaNt pees ens bass 4 ¢ E Rha AGT a ni Ores Wahi tte ne a ths Sy wll stil k HaHa maith NF PP Veag ie ie toast ti eet 4, Br ae i Pye). } flat . | eRe stati 4} tlt Site Ai ‘ ori sa poe se K: Ws ES ee seh Githiints He if Mf ee ay a} y ire a + nati ae CT Se (bet athe a 4 ae Ad ‘i aie fia ig ii it Niheonat thet ae t amen i het 1 eta ie ane fie i ee use nN aN t REY RE PY by: Hid RAL iG it he ve fe ay ih att i i : ys fH t 4 nas ui ait Ds Bp i is M4 pine Mth a i 44) ee feed he a nf es Ait eacea A if dtp) Paes! Tae a Digitized by the Internet Archive in 2009 with funding from University of Toronto http://www.archive.org/details/journalofexperim04broo PH ESIOURNAEL OF EXPERIMENTAL ZOOLOGY EDITED BY WILLIAM K. BROOKS FRANK R. LILLIE Johns Hopkins University University of Chicago WILLIAM E. CASTLE JACQUES LOEB Harvard University University of California EDWIN G. CONKLIN THOMAS H. MORGAN University of Pennsylvania Columbia University CHARLES B. DAVENPORT GEORGE H. PARKER Carnegie Institution Harvard University HERBERT S. JENNINGS CHARLES O. WHITMAN Johns Hopkins University University of Chicago EDMUND B. WILSON, Columbia University AND ROSS G. HARRISON Yale University Manaacine EpitTor VOLUME IV THE JOURNAL OF EXPERIMENTAL ZOOLOGY BALTIMORE 1907 CONTENTS No. 1—February, 1907 CHARLES RussELL BARDEEN Abnormal Development of Toad Ova Fertilized by Spermatozoa Exposed foamesocninen mays. With Five Plates........220...s228 sa.e% 6 ose: I Wiiiam B. Hers An Ecological and Experimental Study of Sarcophagide with Relation to ite meeatis. NVI SEVEN PIOUS. io ieee rosa dee pee ah tee cee. 45 Sara WHITE CULL eyuvenescence as the Result of Conjugation...........0.2....020.05..%- 85 GerorGE LEFEVRE Artificial Pathenogenesis in Thalassema Mellita. With Six Plates....... gl Jacques Lors Semen tic | heory Ob LTopisms. 2300550. sas acts sees es jee eolent A 151 Frank W. BaNncrort The Mechanism of the Galvanotropic Orientation in Volvox............. 157 No. 2—June, 1907 Cuares R. StocKarD The Influence of External Factors, Chemical and Physical, on the Development of Fundulus Heteroclitus. With Seventeen Figures...... 165 O. C. GLASER Movement and Problem Solving in Ophiura Brevispina. With Five UPR ES SRS a Pe eT gn 203 IsapEL McCracken Occurrence of a Sport in Melasoma (Lina) Scripta and its Behavior in Seem ft CMMENENATE: 222-03 bled a aie eele Se se oe Paces oa dx 221 Ross GranviLLE Harrison Experiments in Transplanting Limbs and their Bearing upon the Prob- lems of the Development of Nerves. With Fourteen Figures.......... 239 E. G. SPAULDING The Energy of Segmentation. An Appplication of Physical Laws to © SEE Veo SES ree PR Re a 283 No. 3—September, 1907 A. J. GOLDFARB Factors in the Regeneration of a Compound Hydroid, Eudendrium Ramosan: With two Pipureés:< >. stn. ~'2 2s ). The pineal gland projects above the third ventricle and its tissue is partly degenerated. ‘The infundibulum is small and thin-walled. The lateral walls of the mid-brain are thick and project inward so as to nearly obliterate the aqueduct. ‘The hind-brain is rela- tively normal, although flattened from front to back (sections c and d). ‘The spinal cord in places is fairly normal, in places the cells from the walls of the neural tube fill or nearly fill the central canal. Organs of Special Sense—The nasal organs are fairly well dif- ferentiated. ‘They lie, relative to the brain, posterior to the nor- mal position. ‘The pigment layer of the retina is separated by a space from the sensory layer. ‘The latter is not well differentiated. The optic stalk is not patent but contains no nerve fibers. A lens is present. [he auditory vesicles are simple in form. Peripheral Nerves—Cranial and spinal ganglia are moderately well developed. The nerve fibers cannot be readily traced. 38 | Charles Russell Bardeen Alimentary Canal—The mouth 1s open. Lips and jaws are rudimentary. The pharynx is dilated and contains a sac-like protrusion through its floor from the pericardial cavity (section c). The gills are rudimentary and are contained within a cavity on each side which is formed by an opercular fold open behind (sec- tiond). “The cesophagus hasa lumen into which irregular vesicular spaces open. ‘The tracheo-pulmonary process is short. ‘The gut passes to the left of the liver and pancreas, then curves across the body in front and finally extends straight back. The lumen is not distinct. Neither liver nor pancreas is well differentiated The body cavity is greatly dilated. Heart and Blood Vessels—The heart is an S-shaped tube. The ductus arteriosus and ventricle are thick-walled; auricle and sinus venosus are thin-walled. There is a slight amount of blood in the lumen. The pericardial cavity projects into the mouth, pushing the floor of the pharynx ahead so that the heart comes to lie literally in the mouth. ‘The only definite blood vessels are two vessels which appear to be posterior cardinal veins. ‘These are dilated and anastomose with one another in several places. Genito-Urinary Organs—The tubules of the pronephros are greatly dilated. The Wolfian ducts are irregular in form. That on the right side appears to be missing in places. Skeleton and Connective Tissues—The chorda dorsalis is appar- ently normal. ‘There is an excessive amount of connective tissue. Some of the cartilages of the head are fairly well differentiated. Musculature—Some of the muscles of the head are fairly dis- tinct. The myotomes are moderately normal, although in places the muscle cells are somewhat scattered. Skin—In the region of the head there are especially large masses of projecting epithelium (section a). Larva No. 8, Experiment II External Form—The long axis is fairly straight. The general appearance is that of a normal embryo soon after the tail has grown out. The tail, however, is somewhat shrunken. The sucker is fairly normal. Abnormal Development of Toad Ova 39 Central Nervous System—The olfactory lobes extend forward between the nasal fosse. ‘The tissue of the olfactory lobes is somewhat degenerated. ‘The pineal gland consists of a rounded hollow vesicle with a short much dilated stalk. The ventral wall of the third ventricle is abnormally thick. The infundibulum is relatively normal. ‘The walls of the midbrain are also fairly nor- mal but contain some degenerated pigmented cells. The optic lobes are beginning to be differentiated. “The hindbrain and spinal cord are relatively normal. Organs of Special Sense—Vhe medial walls of the nasal fossz are thick but contain many pigmented and degenerated cells. The eyes are very abnormal. ‘The optic stalks are greatly dis- tended. ‘The pigment layer of the retina is irregular. The sen- sory layer consists of a mass of degenerated cells. The lens is differentiated. ‘The auditory vesicles are simple in form. Peripheral Nerves—The sensory ganglia are moderately well developed but the cells are many of them abnormal. Nerve fibers cannot be readily followed. Alimentary Canal—The mouth is open. Lips and jaws are clearly marked, although not highly developed. Gill slits are patent, but the gills are not well developed and seem to contain no blood vessels. Opercular folds extend over the anterior. portion of the gill region on each side. [he cesophagus has a lumen. The tracheo-pulmonary process is short and branched. The stomach lies at the left of the liver anlage. The gut extends straight back. Neither liver nor pancreas is well developed, although masses of cells indicate their anlages. No blood spaces are found in the liver. Heart and Blood Vessels—The heart consists of an S-shaped tube, but the walls are not normally differentiated. ‘There is some blood in the ductus arteriosus. ‘There are apparently a few blood vessels present but there is no well developed vascular system Genito-Urinary Organs—The tubules of the pronephros are slightly dilated. Skeleton and Connective Tissues—The chorda dorsalis is rela- tively normal. The cartilages of the head are fairly well developed. 40 Charles Russell Bardeen There is a slight increase over the normal amount of connective tissue in the body. The pigment cells are abnormally scattered about. Musculature—Vhe muscles are fairly well differentiated in the head. ‘he more anterior of the myotomes are fairly normal but in the posterior half of the embryo they are not well developed. Skin—In many places the skin shows abnormal outgrowths especially about an irregular opening into the body cavity. SUMMARY AND CONCLUSIONS Toad spermatozoa removed from the body begin to lose both motility and fertility within half an hour. Both motility and fertility last much longer on cool than on warm days and some- what longer in unexposed than on spermatozoa exposed to the Roentgen rays. On cool days the power of fertilizing lasts in some of the spermatozoa over two hours. When only ten or fifteen per cent of the control eggs are fertilized as a rule few or none of the eggs placed with the exposed sperm are fertilized. When only a few eggs are fertilized by the exposed sperm as a tule these eggs do not develop beyond the gastrula stage, but occasionally one may develop into an abnormal tadpole. When the spermatozoa have been well exposed to the rays and yet are still capable of fertilizing a considerable number of eggs, the eggs thus fertilized develop at first apparently normally or even better than the control, but beyond the gastrula stage the development begins to become retarded and at the time of hatch- ing, as the tail begins to grow out, marked deformities appear in the larvae. These deformities are visible externally and are still more striking when the internal structure is examined. The illustrations given on Plates II to V illustrate these deformities more readily than they can be described in words. While they vary considerably there are certain features characteristic of most of the tadpoles. General Development Growth of the tadpole is inhibited beyond the stage which in- tervenes between hatching and the time when it should begin to Abnormal Development of Toad Ova 4I swim. Thus when the control tadpoles of the same age as the experiment tadpoles are equivalent in general form to he 12 mm. tadpole of the frog described at some vetoes in Marshall’s well- known text-book, the development of he control tadpole is, as a rule, more nearly similar to the newly-hatched tadpole of the frog described by Marshall. External Form The head is usually abnormal in shape, the anterior end appear- ing shrunken. ‘The ccelom isin many of the tadpoles abnormally distended. ‘The tzil is usually short, more or less deformed and is often bent in a dorsal direction. Internal Structure The vascular system is little developed in any of the experiment embryos. The heart usually is S-shaped but is rudimentary in form and may have no continuous lumen. In some embryos the wall of the ventricle is thickened by muscle cells but in none are there strong trabeculz in the ventricle. ‘The chief arteries seem in none of the embryos to be completely developed, although in some there are here and there traces of them. ‘The chief veins are likewise in none of the embryos completely developed although in one embryo the cardinal veins are large. In the liver the capil- laries are sometimes well, sometimes but slightly developed. There are relatively a very few blood corpuscles in any of the em- bryos. These lie in some of the scattered vascular anlages. It is uncertain whether the blood had circulated in any of the embryos, but in some of them it is fairly certain that no circulation was established. In all of the embryos the spaces in the tissues indi- cate a considerable amount of lymph either free in the tissues or confined in lymph vessels. Of the central nervous system the brain is the part most con- stantly and deeply affected, but the spinal cord in many of the em- bryos is markedly deformed. The abnormalities consist partly of failure of development or tissue differentiation, partly of irregular growth of tissue, pigmentary degeneration of nerve cells and the 42 Charles Russell Bardeen filling of the central canal with partially degenerated cells. In one embryo the hind brain and anterior part of the spinal cord are exceedingly rudimentary on one side. Of the organs of special sense the eye exhibits the greatest deformities. “The nose and ear are as a rule rather rudimentary than markedly deformed. ‘The eye, however, usually shows a patent optic stalk connecting with a space between the pigment and sensory layers of the retina, a lack of differentiation in the sensory layer, and a more or less highly differentiated lens resting against the sensory layer. The abnormalities in the alimentary canal are exceedingly variable and may affect any or all parts. ‘The mouth is in all instances patent, the lips and jaws rudimentary. The pharynx and gills vary much in structure in the different embryos. As a rule there are traces of internal gills and of the opercular folds but the gills, owing to lack of development of the vascular system, are rudi- mentary. The cesophagus is patent in some, closed in other of the embryos; the stomach as a rule lies at the left of the anlages of the liver and pancreas. The latter structures are seldom highly developed. The rudiments of the lungs are slightly devel- oped. The intestines may be more or less coiled, but are in none of the embryos highly developed and in some are very rudimen- tary. In many embryos the abdominal cavity is greatly distended while the gut 1s rudimentary. The pronephric tubules are usually greatly swollen in places and this dilatation is also frequently found in the Wolfian ducts. There are seldom distinct traces of the metanephric tubules. The myotomes, when not well developed, usually consist of muscle cells somewhat scattered about in the surrounding mesen- chyme. ‘The muscles of the head are usually more or less dif- ferentiated. The mesenchyme of the embryos is considerably greater in amount than in normal tadpoles. The cells seem to be spread apart by fluids in the tissues. ‘The cartilages of the head and the chorda dorsalis are relatively normal. The ectoderm in most of the tadpoles shows in places outgrowths of an irregular nature. ‘These may be extensive villus-like pro- Abnormal Development of Toad Ova 43 cesses. In one instance marked ingrowth of processes from the ectoderm occurred. The cells of the tissues appear for the most part clear in outline. Many of the cells of the central nervous system seem to have under- gone a pigmentary degeneration. Numerous cells in most of the tissues show mitotic figures. I have been unable satisfactorily to determine whether or not there are abnormalities in these figures. ‘There is an abnormal number of cells with two or more nuclei. A striking feature of the experiment embryos is the irregular dis- tribution of the pigment cells. “They are much more irregularly distributed through the tissues than in the normal embryos. There is a striking resemblance between tadpoles which de- velop from ova fertilized by sperm exposed to the Roentgen rays and the tadpoles exposed directly to radium irradiation by Schaper*! This shows clearly that injuries produced in nuclei may be carried through many generations of cells in an individual and finally give rise to deformities corresponding with those due to direct irradiation. Bohn* found that the rays of radium rapidly enfeeble or kill the sperm of strongylocentrotus lividus, but that the eggs appear more fertile after exposure. He does not describe the effect of exposure of the germ cells on subsequent development. Herbst®* found by treating the sperm of sea-urchins with fresh water, alkalies and potassium-free salt water and then fertilizing ova of a different species with the sperm thus injured that the ova sometimes developed as if injured but that there was no evidence that the specific hereditary factors transmitted by the spermatozoa were altered. Further studies are necessary to determine if the hereditary factors carried by the sperm may be specifically influ- enced by irradiation. D. T. Macdougal* has shown that in some plants mutations may be produced by injecting radium prepara- tions, sugar solutions and solutions of calcium nitrate and of zinc sulphate into the ovaries. This most important work sug- ®1Schaper: Anat. Anzeiger, xxv, p. 298, 1904. Levy: Archiv f. Entwicklungsmechanik, xxi, p. 130, 1906. 8G. Bohn: Comptes Rendus de l’Acad. des Sciences, Paris, cxxxv, p. 1012, 1085, 1903. “SHerbst: Archiv fiir Entwicklunsmechanik der Organismen, xxi, p. 293, 1906. “MacDougal; The Popular Science Monthly, September, 1906, p. 16. 44 ~ Charles Russell Bardeen gests that the hereditary factors contained in spermatozoa might be so altered as to produce specific variation in the individuals springing from ova which they fertilize. The effects of altering the normal course of development of vertebrates by electrical, magnetic, chemical or mechanical agents applied to the whole organism, have been shown by Dareste,® Féré,® Roux,’’ and many others, to be seldom confined in a specific way to an organ or group of organs, although some organs, like those composing the nervous system, are especially sensitive to all such factors. [he experiments with irradiation show that although some tissues are much more susceptible to the rays than others, there are wide differences in the effects of the rays on different individuals. In conclusion, I desire to express my thanks to my colleague, Professor B. W. Snow, for the use of the Roentgen ray apparatus belonging to the Department of Physics, and for his aid in con- ducting the exposures. EXPLANATION OF PLATES I to V. On these plates there are represented outlines of the external forms and transverse sections through the body of one control and five experiment larve. Descriptions of the larve represented are to be found in the text as follows: Plate I, p.21 Plate IV—A, p. 31 Plate IT, p. 24 Plate 1V—B;- p237 Plate III, p. 27 Plate eeics Sc Sekai Seer 2 Cae is Oo Vee ee SR i ais Seen ee 3 Garpls. sabes ose Sti bos seeks eee Fee I Totals. ave 5 feces Oo SR eee ae 45 Total weight, 4.65 kilograms. Ecological and Experimental Study of Sarcophagide 47 This gives an average of 103.3 grams. ‘The result on the whole was disappointing, fewer fish having been cast up than during similar surfs. A second weighing trip over the same ground was made about two weeks later, conditions being similar. ‘The results of this trip are as follows: Sree RG SIA INOUCM BASS, oS. < sis. d clo wos 28S ms wooed Oa ae Bos 441 Le BEDE ak NS a ee 18 TES SEE ety 1 Sage ar ra re 50 | TSE tly S.slehee StS EME iS SO rea a ae 12 EE ered frst hes isso a.aith a neato bs asi oaecee dear tase IO US EEL ERIS tEAM i Sti ey ea rh ee 7 et ee ie aA Sh ci 6 coop Ro dn ese 6 |) a2 hrs Sr —.414 | 21 | 7.30a.m. $ |) 132 5 |.26.4| —-.2 | 1zhrs |—.01r6 | oom 21 | 7.30p.m. 8) | 397 5 | 27-4 | +1.0 | 12zhrs |-+.083 + .315 | 22 | 7.30a.m. 9 | 133 5 20200) cus ee bese Out — 2a 22 | 7.30p.m. Salis 5.26.6 -o | 12hrs 000 | -00 23 7.30a.m. IO | 133 5 | 26.6 .O | 12hbrs 000 -00 23 | 7.30pm. | 10 | 133 5 26:6 0) |) 12.hrs. |) 7000.) -00 24 | 7.30am. | i | 133 5 | 26.6 .o | 12hrs | .000 | .00 24 | 7.30p.m. II 133 5 | 26.6 .o | 12hrs o00 | .00 ZAI Relea | Gee | Pre 5 | 26.4 —.2 | 12hrs —.016 | —.06 25 | 7.30pm. | 12 | 128 Beh 52 08) econ ei2ihTs 91) ——2OoG | = 5 26 | 7.30a.m. | 13 | 128 5 | 25.6 | .o | 12zhrs -000 | 00 | 27 | 7.g0am. | 14 | 124 5. | 24.8 | —.3 | 2qhrs |—.066 9) ) aga 28 a.m. 15 | 5 iy Nf eee eae ee ngs: 2) | = imagines * Prepared in part after Minot (91). Ecological and Experimental Study of Sarcophagide 59 One set of Sarcophaga was weighed at two-hour intervals and three sets at four, six and eight-hour intervals, respectively. After the individuals were well in the pupal stage all weighing was car- ried on every twelve hours until the imagines emerged. TABLE III Showing the average changes in weight of Sarcophaga sarracenia Riley, from the larva at extrusion to the adult flesh fly (Set No. 8) I 2 3 4 5 6 7 | 8 9 10 II x rs o @ o a Pa | 2 te = | Feels) Bll aa 2 als as] © 33) Be 2 P SS) £ lemlPS| 2/SS/5 8s See Sa 8 BSE) 2 a =: el a Oj< [4 ic amatinae| Fe oa ~ 13, |10.30a.m. || 7 20 2 Qj = 13 | 1.30p.m. 3 15 26 06 | 3hrs 020 10 13 | 7-30p.m II 10 ad 84 | 6hrs 140 53-84 14 | 1.30a.m 15 22 10 ya Tat) |) 6.hrs 183 16. 63 14 30a.m. | 21 53 10 al 3-1 | 6hrs 516 23-45 14 | 1.30pm, | 27 | 112 TO} )||ptir-2 5-9 | 6hrs .983 18.54 Pee 7sopm. | 33 | 185 Io | 18.5 Wea | 6hrs | 1.216 10.86 Agel t.goa.m. | 39 | 222 TOW e222 a7 |) “Gils 616 3534 15 | 7.30a.m 45 | 298 Io | 29.8 7.6 6hrs | 1.266 5-703 15 | 1.30p.m 51 | 300 IO | 30.0 2 6 hrs .033 III 15 | 7-30p.m 57 | 246 Gea Oe) |e TOLD 6hrs | 3.200 10.67 16 | 9.30a.m. | 71+] 451 5 | 90.2] 41-0 | 14hrs | 2.928 5-95 migrated 19 | 7-30a.m. | 141+] 459 TO} |'4'5-9|—-44-3| |) 7Obrs | .632 a7 OL pupated in days | 19 | 7.30p.m. |. 6 | 439 TO) /|/43/-9) |! —2.0 | i2hrs |—.166 gti 20 | 7.30a.m 7. || 432 ite) || avgy Sohal, ae 12hrs |—.o58 Shige 20 30 p-m Ney TOM 42/7) ean ea hrs |—loan —.095 21 | 7.30a.m 8 | 423 TORN PAZ Oe |e— Au | elehTs i038 O77, 21 | 7.30p.m 8 | 420 POM NA22ON| Neg 1zhrs |-—.025 —.059 22 | 7.30a.m 9 | 420 Io | 42.0 .o | 12hrs .000 000 22 | 7.30p.m. 9 | 414 TOM MAD 40.16 12hrs |—.os50 et) 23 | 7-30a.m. | 10 | 411 ney || “Wiaie |} Sag) || edness — .06 23 | 7-39p.m. | 10 | 409 TONNE 4059) 20 a2 ars | Ss Or6 —.039 24 30a.m. Ir | 411 TOs | 40-0 it} 2 12hrs |-++ .o16 + .039 24 | 7.30p.m. II | 407 LOM AOS 7/0 ad: 1zhrs |—.033 —.08 25 | 7-30a.m. 12 | 404 TORN |b 4Or4 a eens 12 bts O25 — .061 25 | 7-30p-m. | 12 | 403 1) |P4@agh |) S=ch 1zhrs |—.oo8 ==. 02 2 7-30 a.m 13 | 401 ney) || elie |] =e, |) Wales | -S.euls —.039 2) 7.30 a.m 14 | 399 10; 1939.9] ——.2 |" 2a hrs’ ||— Joo8 —-O2 7-30 a.m Hoees99) ||| LO) "| 39.9 o | 24hrs 000 00 30 a.m 17 393 |} Skog |) sale 48 hrs |—-.o12 ——.08 a.m 18 Als a Le) ASols || sets 24hrs = — imagine | | | | 60 William B. Herms Sarcophaga sarracenie endured the experimental conditions far better than Lucilia as was evident from the number of imagines resulting in each case—go per cent in the former and only 162 per cent in the latter, including the set which perished entirely. The greatest mortality occurred during the period between migration and pupation, though a good portion never completed the period of pupation after having entered it—about 40 per cent in Lucilia and about 6 per cent in Sarcophaga. In order to secure a fair weight it was necessary to brush each larva clean with a camel’s hair brush. This great amount of handling had practically no effect on Sarcophaga, which are quite hardy, but on the other hand Lucilia were seriously affected.® However, in the latter case it is well worth noting that the sur- vivors are very near the mode of normal frequency for larvz at migration, for pupz and for imagines (cj. normal frequency curves, Fig. 7, with growth curve, Fig. 2). This growth as indicated by the curves was checked by the growth of larvae, both in time and weight, feeding on fish out of doors under normal conditions, and it was found that the time is the same, and that the average weight of migrating larve varied but 0.5 mg. ‘This difference would, undoubtedly, have been entirely wiped out had a larger number of individuals been used. Discussion of Curves Lucilia caesar (Fig. 2): The figure shows a curve derived from a set weighed every two hours for the first twenty-four hours after hatching, every four hours during the following twenty-four, every six hours for the next twelve, etc., as indicated in Table II, column 2, according to which this curve is constructed. ‘The curve begins with the weight of ten eggs (I mg.) This egg stageis represented — by the dotted line and covers a period of about twelve hours. At this point the larve hatched and were weighed three hours after the emergence of the first individual with a weight of 4 mg., for twenty or 2 mg. for ten, as shown by the curve. The average increase in weight during the first three hours is 50 per cent which °This may not have been entirely due to the handling; the same mortality for this species may possibly exist in natural conditions. Ecological and Experimental Study of Sarcophagide 61 is followed during the next two hours by an average decrease of 17.5 per cent. [his decrease was shared by three sets, and two further sets showed zero growth, which was probably due to the fact that the weighing was done a half to three-quarters hour later, during which time the larve gained enough to bring the pro- cess from negative to zero. Neither does the sixth set show the decrease because it was weighed still later. Five of the six sets of Lucilia were thus affected, as is shown by Table I.’ 450 Cees 4a5 6) 7 8 9 10 1k 12 13°14 15) 16 Fig. 2 Growth curve of Lucilia cesar Linné, derived from Table II, column 4, on the basis of 10 individuals; x= days; y = milligrams. The dotted line at the beginning of the curve represents the egg stage; the solid line represents the larval period with the apex as the point of migration; the broken line represents the period of pupation and the drop at the end, the loss of the pupa cases. Migration is represented by the apex of the curve when the larvz reach an average weight of 40.4 mg. and this for a feeding period of about sixty hours (57+hours), an average hourly in- crease of .7070 mg. or an increase of 40,400 per cent of the orginal weight. Immediately after migration there is a marked loss in weight. TLoss of weight probably due to loss of moisture in crawling on dry surface of receptacle. | | William B. Herms 62 Since there is no appreciable faecal discharge and because the loss is quite gradual, it is assumed to be due to loss of moisture. This loss is less pronounced after the fourth day until the seventh day when pupation takes place. broken line. The pupal period is marked by the Here there is an abrupt loss in weight. early period of transformation from larva to pupa the case changes During this in color from a pale yellow to the characteristic chestnut of the This change in color, due to the hardening of the chitin, is quite rapid. advanced pupa. In the eighth a there is another drop oe = Soest Pr i : EGER Ht siiocts : =o BEALESESEIECE siitos Scenesenseeesesé_ aaa lI ae eae ea ° IO II 12 oe 14 =e 16 17 18 19 8 a Uf 7 : F 6 Growth curve of Sarcophaga sarracenie Riley, derived from Table III, column 4, based on Fig. 3 The solid line represents the larval period with apex as = milligrams. i) the point of migration; the broken line represents the pupal period and the drop at the end, the loss of 3; «= days 10 individuals; the pupa cases. which is followed in the ninth day by a considerable increase. This is also evident in the remaining sets weighed at about this This increase of weight shared by practically all pupz weighed, is probably due to an addition of moisture extracted from the air, or there may be a correlation to internal metabolism. At this point in the life history the metamorphosis 1s wonderfully During the first day or two of pupation the individual | same age. rapid. Ecological and Experimental Study of Sarcophagide 63 seems to disintegrate, forming a mass of fluid matter, but in the third day there is a rapid organization, which results in an indi- vidual that could be easily recognized as a fly, even at this early period. All pupz were kept in glass vials about two-thirds full of sand in which they were buried. ‘These vials were covered with netting (bobbinet), and were kept near a window which was kept open most of the time together with other windows in the room. ‘That the sand, which was comparatively dry, might extract moisture from the air and in turn transmit it to the pupz is very probable. The increase in weight above mentioned is followed by a nearly corresponding decrease during the next twelve hours, which re- sulting weight is held for two days and a half. At the end of this time the weight again gradually decreases until during the last day of pupation when it is uniform. ‘The abrupt drop at the end of the curve represents the casting off of the pupa cases. Sarcophaga sarracenie (Fig. 3): This curve is based on the growth of a series weighed every six hours during the feeding period (see Table III, column 2). The curve shows several interesting features. In the first place the increase in weight is really prodigious, beginning with an average weight of 0.2 mg. (10 larva equals 2 mg.) and increasing to an average of 90.2 mg. in seventy-one hours, an average hourly increase of 1.270 mg. or an increase of 45,100 per cent of the original weight. This aver- age for larve of this species is not high since many individuals which were weighed for another purpose ranged from 150 to 200 mg. (75,000 to 100,000 per cent) and over and were probably de- veloped from larve weighing no more than the above at extrusion. This species did not show a decrease in weight at the beginning of the feeding period as Lucilia. Another remarkable feature is the large and rapid drop in Weight after migration. ‘This loss is just about one-half of the average original weight of the migrating larve. The pupal period is characterized by a small but comparatively regular loss in weight. + Percentage increment curves (Figs. 4 and 5): Since applying the percentage increment method used by Minot (’g1) the writer William B. Herms 64 OO a ee ee ee ee rere . Sr HEH HHH HE EEE EEE EEE EEE Fh ahehete-o—}—}—}—}—t 4 -~ ee eee eee | ne PCH | (| oO ° ° SSG gS) SEY NS Oo 7O 7 RS 3s tS Bs BS S sS SSS RES E 3) BS | [e} ° 8 ° fe) oO + oA a vt fon os) ~ \o ~ + on a oo ~ is convinced that this is the true method of growth measurement Fig.4 Growth curve of Lucilia cesar Linné, based on percentage increments. ; #4 S5 25 Se 235 ae} ° bo H as = © — Yn ieee oo ge. ote (9°) wog Bl co} yet s 2 B A Ssdddossdsosst tastdocat eccat eosiiecsifossifit eisttaedt 2 g : S 7 = E & & - FRRRLHIMGEEE EEE aaseeed EttitT EEEEEEE EH : PELE BE eet HH BHAI PEE mw A wor ss gf Ss oo HEHEHE Hee 3 ee 8s ee. fence 2 #2 "Ep & Es 06 (hes as é A eee E cae is oe ee ee Bot SS - fe Bas ° Sr ° o w OP ar Lo 3 g ue Ono n x © “a Ss) 7-4 ma. & & NS BE pe 5 ae Oo 8 & Ge Sng eee ae : eee Egy ai af SE og & co = fac} > ° san * ae) ne anc} e0 f= S ae S « & % sce 5538S Sod ~ 38 & REGS oes c ox ea ne om le) ad v - 7d o SS fet uv 7 Ba Ln v aera — ss a Se e aoe e bes Pelee = s aa 8 EESE clusions reached by Minot (’91) also apply here, viz: II, column 10; x = days; y = percentage increments. which gradually diminish with age. ably clear manner. of pupation. sions: 65 rity. also, “‘the 6 c Ecological and Experimental Study of Sarcophagide tends to be followed by an opposite compensating irregul Second, The variability diminishes with the age,” As a proof for this ge weight of migrating Viz: 38.6, 40.4, 39.4, 36.0 The avera irregularity of growth of an individual is very great” and “each individual strives to reach a particular size.”’ latter statement, see Table I. Lucilia larve for the five sets shown is, and 39.0 mgs. 13 Bek thy as) Gp ats} a0) 12 LOy rr 9 Fig. 5 Growth curve of Sarcophaga sarracenie Riley, based on percentage increments, from Table III 8 derived ra- Crosses represent point of mig: column 10; x = days; y = percentage increments. > tion and point of pupation. IV CORRELATION OF FEEDING PERIOD WITH FOOD SUPPLY g period of That there exists a correlation between the feedin the larve and the food supply, My atten- To illustrate, one a fish was exposed for fly there can be no doubt. tion was called to this fact time and again. instance follows: On July 28, 1905, no ee m5 TH _ SB Umiae + - 2 cB) 5 i?) Ee vo ay Be} ve a =>} eggs and about one thousand were sec o’clock, p. m., not a large number. ging early July 29. As early as noon of July 31 (54 hours feeding), the larvz on the over several pieces of the flesh, the larvae emer 66 William B. Herms smallest piece had eaten it clean and had migrated, while almost all larvee on the larger pieces continued feeding until about noon of August I (67 to 72 hours feeding) when practically all migrated. A few, however, continued feeding on the unconsumed flesh and the last larva did not migrate until after noon of August 3. That the larve on the last two pieces of fish did not continue eating until the flesh was consumed, indicates that there is an optimum when enough nourishment has been taken to pass through the metamorphosis to best advantage, and at this point migration takes place. ‘That this optimum 1s not always reached is shown by the early migration because of food shortage, and this does occur at times out of doors, since eggs and larva may be deposited in such large numbers on a single fish that an early migration is necessary. [hus we have sharp competition which is further augmented by the pressure of four species, each more voracious than the other if that were possible. “Therehas been, undoubtedly, a gradual adaptation in the past to the existing conditions. In order that the flies might exist in the locality, there must have been a conformation to the food supply, 7. ¢., a race of flies which can adapt itself to an inconstant, somewhat periodic supply of food would survive. Further there must have been an adaptation of the four species already mentioned to each other, a species which required a comparatively long feeding period could not well exist with a species whose feeding period was of short duration. ‘The food would invariably be consumed by the quick feeder, while the former would suffer starvation. As it is, all larva migrate at about the same time, 7. e., when the fish is consumed which usually requires a fairly uniform time, as above noted. A glance at the larve at this time will reveal the fact that there are two general sizes of larvz present, very large ones and small ones. All have fed during approximately the same time, yet. Sarcophaga sarra- ceniz nee attained its enormous size while Lucilia caesar 18 uni- formly smaller. The greatest variation in size is found in the former. ‘The larve of the screw-worm fly are usually smaller in average than Lucilia cesar. There is probably also a correlation with the surf producing storms, which is discussed in full later in this paper. Another Ecological and Experimental Study of Sarcophagide 67 factor which requires a hasty consumption of the flesh is the dry- ing out caused by the sun, and also the decay from putrefaction. The above general observations led to the following described systematic experiments bearing on the subject of correlation with food supply. V EFFECTS OF OVER AND UNDERFEEDING The observations described in the last chapter were interpreted to mean three things, viz: First, That there exists an optimum at which a certain larval weight is reached, which is the weight best adapted to pupation and emergence as imagines. Second, That if feeding is carried on beyond this optimum point, pupation is hin- dered or even death may ensue. ‘Third, There must be a point below the optimum at which the larve can barely pupate, and have not strength enough to carry through the pupal period or will even die before pupation. ‘To determine the facts and to find the critical point between death or survival, the series of experiments about to be described was arranged. On August 12, 1905, a large German carp, quite fresh, was washed up. ‘This carcass was immediately taken from the beach, placed near the laboratory and exposed for eggs of Lucilia caesar. In an hour and a half (between 10 and 11.30 o’clock, a. m.), about eleven thousand eggs were deposited, the gross weight of which was 1083 mg. ‘The fish was taken into the laboratory, the viscera removed, the body cut into six pieces, and the eggs roughly divided into six masses. “These masses were then placed upon the pieces, which in turn were placed in separate boxes or compartments. Two of these sets were used for histological material, and conse- quently only four series (series I, 2, 3, 4) were used in the experi- ment. The eggs hatched early August 13. The plan was to take the flesh away from a portion of each set at consecutive inter- vals of six hours each, allowing one series to migrate normally, weighing the larve at end of feeding period. ‘Then the weighed larvee were placed in separate dishes so that the same lot could be weighed again about the middle of the period of pupation, and the adults when the wings and bodies were dry. The remainder of each series was retained as a check or in case of accident. 68 William B. Herms This first experiment was supplemented by a second one a little later (August 28) and carried out on exactly the same plan (series 5, 6). To show the results of this series of experiments more clearly the following table is presented: TABLE IV Showing results of Feeding Experiments on Lucilia cesar, the common green flesh fly Feeding ' Series No; | Period | Mean Weight of Lae) a7. Weisht of Pupe | Mean Wemhtat ieee ee at End of this Period | in Hours (1) optimum) 60-72 38 .183(256) 30.283(289) | 22 .283(208) 2 | -a- "| 35-68 (50) 24.76 (50) | 18.44 (47) 3 bo Se | 31-06 (50) 22.38 (48) | 17.54 (48) 4 t <,he48 22.14 (50) 11.81 (48) 8.08 (44) 5 Lie brakes Co 17.06 (50) (39) 12.38 (31) 6 36 8.82 (57) 9-34 (49) | 7-15 (20) 7 retarded 33-59 (64) (39) | 21.00 (25) 8 60-78 46.73 (65) dead larve — 9 55 36.78 (50) larve having migrated = early Note 1—Weight was carefully taken in milligrams. Note 2—Figures in () denote number of individuals weighed. Note 3—The same individuals were weighed in each series except in cases mentioned below. Note 4—See Fig. 6 for relative sizes of imagines. EXPLANATION OF TABLE IV Series 1—Ihe term optimum as here used represents larve in ‘greatest frequency and also corresponds to the normal frequency of the group which was shown by the weight taken of individuals feeding out of doors, as stated above. ‘This series accordingly forms a convenient basis for comparison in reference to the re- maining series. “The normal frequency curves (larvae, pupz and imagines) of Lucilia caesar are based on this series. ‘The feeding period here indicates that the larve migrated after eating for from 60 to 72 hours. This optimum weight for larve (mean 38.183 mg. and mode 37.00 mg.) represents the point at which the chances are best for pupation and emergence as adults. From this point either way the chances diminish, most rapidly, of course, at the extremes. ‘The pupa cases of optimum forms and Hours BEG GEG pa- Shanota- hase @GR HA. @epeee sete. Fig.6 Cut showing relative sizes of imagines as produced by feeding experiments. Number of the series on the left, and number of hours that the larve were permitted to feed on the right. See explana- tion of series on pp. 68 to 72. 70 William B. Herms beyond are very chitinous, making a comparatively rigid shell, which affords the optimum of protection. On the other hand, the farther below the optimum, the less rigid the case, until at the lowest extreme it is a mere flimsy covering. [his shows that here the least possible energy 1s expended, while the greatest amount possible is stored up for the trying transformation from larva to imago. Series 7—Among the optimum series there were certain larve which were very slow in pupating. While others of the same series pupated August 18 to 23, these particular ones had not by August 27, the date on which they were weighed to ascertain the reason for such tardiness. This weighing, in which 64 larva were taken, showed a mean of 33.59 mg., which in itself does not show a high average (cf. series 1, larva). If we consider the lapse of time since migration (twelve days) and the loss of moisture under- gone during this time, we must see that this average is after all high, and that these larvze were beyond the optimum weight, and for this reason pupation was deferred. (For loss of moisture cj. curve of growth for Lucilia cesar.) Of these 64 larvae 39 pupated by September 9, giving us twenty days in the larval period against a normal of about six days; the rest were dead. Furthermore, at the present writing only 25 adults have emerged out of the 39 pupz. The mean of these adults is 21 mg., which shows pointedly that the heaviest larve had been eliminated or had dried out sufficiently, resulting in adults very near the optimum weight (cj. adults in series 1). Series 8—The sixty-five larve in this series were picked from the same fish upon which the optimum series was feeding, and con- sequently could not have been poisoned. All the dead larva with- out exception were taken and weighed so that no selection was possible. The larvae were carefully examined for injuries, but only one was found to show an injury, the smallest of the lot (25 mg.) In this case it was decided that death was due to this cause and the weight was not included in the result shown. Thus we have what was expected, 7. e., that continued feeding without regard to the optimum would eventually result in death. The mean for this series 46.73 mg. is extremely high and a com- Ecological and Experimental Study of Sarcophagide 73 parison with the normal frequency curve for larva shows that the _aberrent forms beyond the optimum shown in the curve are either just within the above mean or just beyond. It is also interesting to know that 63 per cent of the total number of dead larve are beyond 45 mg. in weight. Series g—It will be seen that these larve migrated five hours sooner than other migrating individuals, but that the weight on the other hand is nearly equal to the optimum. ‘This corroborates the statement made above that some individuals eat more rapidly than others, and again this series augments the evidence in favor of an optimum weight. . Series 2—In this series the optimum time limit for feeding is practically reached, but, since these larvae were still feeding at sixty hours, the flesh was taken away, and we see that the optimum weight had not yet been reached by several milligrams. This indicates again, when compared with series g, that certain larve feed faster and reach the optimum more quickly than others. Series 3—¥ood taken away after feeding fifty-four hours. Pupation takes place readily and promptly and adults emerge on time, but are short weight and small in size. Series 4—Food taken away after feeding forty-eight hours. Pupation takes place as above, also same for imagines. For those series below the optimum down to and inclusive of forty-two hours, the average time between migration, or in these cases time of taking food away, and pupation, is very much more regular and nearer the normal time, 7. ¢., about four days. ‘This sein: be expected, since in the normal series we have many larve pach have gone beyond the optimum weight and pupation is consequently retarded. On the other hand, in the series just mentioned the larve are practically all within the optimum, consequently the time for the entire life history of each individual conforms more closely to the normal and there is no dragging out of the prepupal period. Series 5—In this series from which the food was taken away at forty-two hours, all the original weighed larve died because of an accident. As a result the pupal weight is not given, but the weight of the adults was taken from the accessory series. The number of individual larve in the accessory series was not taken, but only thirty-nine pupz resulted, and from these only thirty- 72 William B. Herms one adults emerged, whence the number used in this series. The mean weight shows a remarkable selection, since the average weight of the adults is greater in this series than in the preceding " series, and yet the mean weight of larvz is less, that is to say, the lighter weight larvee did not produce adults. Series 6—In this series with a feeding period of thirty-six hours we reach the lower limit at which adults could be secured. The smallest fully developed flies of Lucilia cesar weighed about four milligrams, though several emerged that weighed but half that much, but died before the wings were spread. Several also died in struggling to free themselves from the pupa cases. As men- tioned before the pupa cases of these light forms are very flimsy and are also gummy, thus making it more difficult for the young imagines to emerge from such a case. Note 1—The weights given above for imagines are of flies that had no opportunity of feeding, the weight being taken within an hour or two after emerging from the cases. For comparison the individual weight was taken of sixty-four specimens of Lucilia cesar, regardless of sex, feeding out of doors on fish. The range of weight was quite wide, varying from 14 mg. to 72 mg., however, only three were above 60 mg. The average weight of this group is 38.17 mg. As compared with the adults of the optimum series (22.28) the above average seems very great. This can be accounted for by the presence of numer- ous Ova in various degrees of development within the females and also by the presence of liquid food that is contained within the alimentary canal of each individual. Note 2—It was the plan of the writer to secure the normal frequencies of Sarcophaga sarracenia, but the great variation in weight required more individuals than there was time for weighing. After weigh- ing several hundred larve at migration, the result showed a range of weight from 75 mg. to 227 mg., and in no class were there more than eight variates. No less than five thousand larve would be necessary to establish the normal frequency of this species. This wide variation in larve at migration means also a wide variation in pupe and imagines, which was also partly worked out. Tables showing the distribution of frequencies in Lucilia cesar: TABLE V Larve at migration Classes.. 25|26|27|28|29|30]31|32|33/34135136137138/39140141/42143| 44145146147|48|49|50|51/52153154155156157 Variates | 1| 1| 1| | 3] 6| 4{ 7| 7/13|18|25|30|24]23]22|16|15|15|11| 5] 2] 2] 2] 1] 0 o| 1] o| of of of 1 TABLE VI Pupe weighed about the middle of the pupal period Classes........ 18|19|20|21|22|23|24|25|26|27|28|29|30|31|32|33/3435|3/37138|39140141142143/44145/46 Vanates 26 3.01. 1| 1| 2] 2] 6] 9/13/16/18]19]31|20|20\19|20|17|16| 9|16{10| 6| 7] 5] o| 4| 1| o| of x TABLE VII Imagines weighed two or three hours after the wings were spread; no food taken Classes? ce e-ecdan eee 13|14|15|16|17|18|19|20|21|22|23 |24|25|26|27|28|29|30|31|32/33134135136 Watiates! 1c 5 ons een eee 2| o| 3] 4{11]10|16]22|24]23]21 |21|16]11| 8] 7] 3] x] x] 2| 1] 1] of x Ecological and Experimental Study of Sarcophagide 73 TABLE VIII Constants based on Tables V,VI, VII, and derived by the usual formule: [4-=E 2, £,- 20.6745 © eee E, = + 0.6745 /n Fag C= GX 100% Eo = + 0.6745 [142 Cayr | Larve Pupe Imagines 0) MEE) eae 256 289 208 ECON 38.183 +.180 30.283 +.197 22.283 +.176 MAR(MOdE)) oe. eee ce dsce ees 37/- 28. ; 21 10022. (ee mdindexonVariab.))....2.-..... 4-293 4.127 4.974 4.139 3-769 4.124 fe(@cet-of Variab.).=.-:.26-. + 11.24 +.33 16.42 +.47 16.91 +.56 *See Table VII. pees 43 37° 41 45 49, 17 25°25 29 33°37 41 45 «13 17 25 25 29 33 Fig.7 Normal Frequency curves of Lucilia cesar Linné, based on Tables V, VI and VU, respec- tively. The classes based on weight in milligrams are shown along the abscissa and the number of f variates along the ordinate. The first curve, based on 256 individuals, is of larve at migration; the __ second curve, based on 289 individuals, is of pupz at about mid-pupal period; the third curve, based _ on 208 individuals, is of imagines shortly after emergence. b a Note—It is practically impossible to weigh all pupz at the same stage, which accounts for the irregu- _ larity of the derived curve. The weight of the imagines was taken several hours (2 to 3) after _ merging from the pupa cases, It should be borne in mind that these imagines had not taken food before weighing, This seems to be the best plan for the derivation of a curve with a reasonably _ uniform basis. 74 William B. Herms VI CORRELATION OF LIFE HISTORIES TO THE SURF-PRODUCING STORMS As already intimated the surf-producing storms occur at com- paratively regular intervals, a low surf taking place about every three days, a heavier surf every six or seven Oe. and a still heavier surf every fourteen or fifteen days. Now the very fact that the life histories of Compsomyia macellaria, Sarcophaga assidua, Lucilia cesar, and Sarcophaga sarraceniz cover, respectively, a period of eight or nine days, twelve or thirteen days, fourteen or fifteen days, and eighteen or nineteen days, seems to indicate a peculiar coinci- dence, if nothing more. The factor 3 plays an important réle, viz: 3 X 3, 3X 4,3. x 5 and 3x6, which corresponds in general to the occurrence of the surf. Does this signify anything, or is it merely coincidence? In an endeavor to interpret this and to have a working basis, a table was secured of the northeast and east winds (1. e., the surf- producing winds on Cedar Point) prevailing at Sandusky during May, June, July and August for the years Ig0I to 1905, inclusive. This table was furnished through the kindness of Mr. E. H. Nimmo, Director of the U. S. Weather Bureau Station at San- dusky, Ohio. This table confirms in general the observations made above. In it are given the dates on which the prevailing winds were from the direction favorable for a surf. For the four months named above in the years 1904 and 1905 are found the following intervals between storms in days: 1904—3, 14, 7> 6, 3> 6, 35 Q; 12, 6, 6, Q; 6, 6, 3 6. 1g05—Q; 12, 4, 3> 3> 3> 2, 8, 4. 6, 3> 6, 6, 6, Q> 6, 6, 3> Q. It can readily be seen that 3 is again the prominent factor. Fish are only cast up in quantities by a surf, and a surf is alone caused by a prevailing wind from the northeast or east. ‘These fish are practically the sole food for the flies along the beach and this is especially true of those individuals living on the narrow strip of sand called Cedar Point, upon which the laboratory is situated. Were adults to emerge from their pupa cases at a time when no fish or very few fish were present on the beach the proba- Ecological and Experimental Study of Sarcophagide 75 bilities are that such individuals would suffer starvation. If this were often repeated the tendency would be to impair the vigor of the species, especially by interfering with the normal egg-laying habit. ‘This latter would certainly be the case if the usual number of adults were to emerge with an undersupply of food present upon which the eggs could be deposited. ‘The large number of larve for the short supply of food would result in producing smaller indi- viduals, which has been proven by experiments. ‘That the sarco- phagids given in the list are normal, as compared with individuals of the same species breeding elsewhere, 1s evident to the most super- ficial observer, and they are certainly not less numerous. Considering the above facts and also bearing in mind that the food supply is influenced by the comparative regularity of the surfs, there seems then to have been somewhere in the past an adaptation to the surf-producing storms. When the adult fly emerges from the pupa case it 1s likely to find available food on the beach, or has but a very short time to wait for it. Then since egg deposition and food supply are so intimately connected, eggs are deposited and the cycle begins anew. As soon as the liquids have been sucked from the accessible parts of the fish, egg-laying ceases and the remainder of the work is left to the larve. ‘The presence of juices would then seem to be a gauge for regulating the number of eggs and young larve deposited on one fish by the females. ‘This will recall the state- ment made above that eggs are seldom if ever deposited on fish that have become dry, which fact should also be borne in mind in connection with the adaptation to the surfs. If fish were to lie around for any length of time before the flies emerge, the juices would be dried up by the sun, and the fish would become unfit for food. However, it is very probable that the adults would after all deposit eggs on the dry fish. Lack of food for the adults would necessarily be a serious menace to the species. Under conditions as they now exist a drying out of a fish by the sun would not likely occur, since the flies would not permit a single fish to dry out thus. The assertions relating to this are based on laboratory experi- ments, 7. ¢., drying out a fish in the laboratory and then placing it outside in reach of flies. 76 William B. Herms That which is of principal interest in regard to this correlation of life histories to the surf-producing storms 1s the brief interval _ between the storms, represented by the factor 3 or 6, and this with its relation to the days required for development with each species, viz: Compsomyia about g days, Sarcophaga assidua about 12 days, Lucilia about 15 days, and Sarcophaga sarraceniz about 18 days. Further, it must be remembered that the life history of each species for this locality covers a comparatively definite period, which is a necessary consideration in this matter of correlation. When eggs or larvze were collected, very little chance was involved in predicting the date on which the imagines would appear. The writer made use of this factor in his experiments with the three most abundant species. It would also be useless to speak of a correlation to the surf producing storms if the life histories of the species studied here corresponded to the life histories of the same species in localities remote from a beach. From the literature consulted the following data was secured relative to the latter. Compsomyia macellaria: Morgan ('90) gives (August 18 to August 29-30) zz to 12 days; Francis (90) larval stage about a week and pupal stage from g to r4 days, a total of from 16 to 21 days. Sarcophaga assidua: Howard (00) gives (July 3 to July 25) 22 days, also (July 9 to July 18-26) 9 to 17 days. Lucilia cesar: Howard (00) gives (May 12 to 29) 17 days. Sarcophaga sarracenie: Howard (’00) gives (May 12-30) 18 days; (July 2-29) 27 days; (June 6-17) rr days; (June 13-26) 13 days; (July 7-21) 14 days; (July 9-22) 13 days; (July 24 to August 9-11) 16 to 18 days; Kellogg (’05) gives ro to 12 days, Howard (02) ro days. One can readily see from the above citations that there is a marked variation in each species, and that these periods do not coincide very closely (excepting the first period in the last-named species) with the results secured in these studies. It must, how- ever, be admitted that more extensive and systematic work should be done relating to the question under discussion. Conditions as stated above may lead to the impression that egg deposition is a direct result of the presence of food within the Ecological and Experimental Study of Sarcophagide 77 alimentarycanal. But this is evidently not a necessity, as is shown by the following observations. In carrying on experiments it was always a matter of concern from the first to guard against outside larvae. On several occasions fish were covered by a screen of netting to keep out flies, but the females of Sarcophaga sarraceniz invariably deposited their young on the netting and these then found the fish without much difficulty. While carrying on experiments indoors the flesh was kept in Petri dishes and covered with like dishes. Several times it happened that a female of Sarcophaga sarracenie gained en- trance to the room through the door and deposited larve on the outside of the dishes. Not being able to get at the flesh the larvee perished. Furthermore, in such cases where the head of the fish was hooded with cloth, the females of Lucilia caesar deposited eggs very freely on the cloth and also on the loose ends of the string used to tie the hood. ‘These observations led the writer to be- lieve that it is not necessarily the presence of food within the ali- mentary canal that stimulates egg deposition. In this connection, however, it might be interesting to note, that no eggs were secured from individuals of Lucilia casar kept under confinement with plenty of accessible food. ‘The flies crawled about on the fish apparently sucking the juices, but all died in a short time. Con- finement very probably was the cause. ‘This evidently agrees with experiments on the house fly cited by Howard (’00), viz: “T am inclined to believe that what may be termed the psycho- logical influence of confinement, even in so large an enclosure as the one used in the 1898 experiments, alarmed the flies, caused their early death, and prevented them from obeying their natural instincts and performing their natural functions.” Vit EXPERIMENTS UPON THE TROPISMS OF FLY LARVAE The. following experiments and observations are not intended to cover the topic of tropisms and their relation to fly larva with any degree of thoroughness. The object of this final chapter to the general paper is to present a statement of experiments made on movements in reaching the food and in migration, including a preliminary discussion. 78 William B. Herms Chemotaxis Two experiments were tried with reference to chemotaxis. First Ex periment— Thirty-six larvae between four and five hours old, were placed in a small vial 4.3 cm. in Re and a piece of hsh weighing about one gram was placed 1.5 cm. from the bottom of ee al 12.7 cm. in length. While ce was being done, care was taken that the flesh did not come in contact with the sides of the long vial. After the larva were shaken to the bottom of the smaller vial, the two were put mouth to mouth horizontally onatable. ‘This took place at 9.05 o’clock, a.m., June 30, 1905. In three or four minutes. there was a decided movement toward the mouth, but because of the unevenness of this region, the move- ments became scattered. At this juncture a short piece of paper was placed like a bridge inside the vials connecting them. The following table shows the results of the experiment. Un- fortunately the data taken were not sufficient to make a complete table. Second Experiment—A small piece of fish weighing about two grams was placed in the center of a large sheet of heavy white paper, then young larve were put at different distances from the flesh, after dipping them partly (posteriors) in glycerine s so that a trail would be left in crawling. The first larva was placed with head toward the meat at g cm. distance, and reached the food in four minutes after taking a some- what winding course. The wind was favorable in this case. A second larva was placed at a distance of 11 cm. with its head away from the flesh and the wind at right angles. ‘This larva started at 10.32 o’clock, a. m. and after a very circuitous route, circling frequently though always drawing nearer and never going beyond, reached the food at 10.52 0’clock. ‘Time, twenty minutes. The course of this larva took it considerably to one side of the flesh almost to the starting place of the first larva from which point the two paths to the food were almost parallel. A third larva was placed at a distance of 11 cm. on the wind- ward side of the food. After a great deal of traveling, making many circles and stopping frequently like the other two, it reached Ecological and Experimental Study of Sarcophagide 79 TABLE VI Showing result of the experiment above indicated Time Started ® ; pot June 30, a.m., Time Arrived ume React REMARKS Larve in Minutes ghey = 9-05 vials placed = I 9.20 9.30 oI 2 9.27 9-31 4 3 9.28 9.31 3 . 9-31 9-37 6 5 52 9-34 Z 6 9-32 II .00-12 .00 go+ 7 93 9-39 a 8 no record 9-39 betw. 6-7 9 | no record 9-40 betw. 6-8 10 no record 9.42 betw. 6-10: II no record 10.00 betw. 6-28 12 || norecord 9.52 betw. 6-20 13 9-45 9-51 6 14 9-45 g-51-10.15 betw. 6-30 15 9.46 9-51-10.15 betw. 7-30 16 9-47 g.51-10.15 betw. 8-30 17 9-47 9-51-10.15 betw. 8-30 18 no record g-5I-10.15 19 | no record g.5I-10.15 20 no record 9-51-10.15 | z 21 no record g-51-10.15 | 22 larve arrived by 10.15 22 no record 10.15-10.30 23 no record 10.15-10.30 24 no record 10.15-10.30 25 no record 10. 15-10.30 26 ~~ —snorecord 10. 15-10.30 27 larve arrived by 10.30 27 no record 10.30-I1.00 28 | no record 10.30-I1.00 29 no record 10.30-II.00 30 10.40 10.30-II.00 5 30 larve arrived by 11.00 ar. no record II .00-12.00 Observations interrupted 32 no record II .00-12.00 by lecture at 11.00 33 no record 10.00-12.00 34 no record II .00-12.00 35 no record II .00-12.00 36 / no record II .00-I2.00 36 larve (all) arrived by 12M. Note—When the larve started toward the food, they hastened on, stopping once in a while to sway the head about in the air for the purpose of orientation. If the flesh was not reached by means of a direct route, as for instance along the upper side of the vial, the larve crawled to the end, then down and 8o William B. Herms a point farther away from the flesh but on a line with it, and the starting point of the firstlarva. ‘This process required an hour and four minutes and the larva died ,at this place apparently from exhaustion. A fourth larva placed 3} cm. to one side of the starting point of the first also failed to find the food. Its course led it farther away and finally off the paper. Discussion—F¥ rom the above two experiments it will be seen that there are two factors involvedin findingthe food. First, the primary stimulation of the larve by meansofthe food. Whateverthenature of this stimulation may be, and whatsoever the internal mechanism involved, the process which underlies the turning of the larva in an effort to draw nearer to the food, may be termed chemotaxis. The second factor is the swaying of the head from side to side or in an arc of acircle. ‘This the larva does for the purpose of orienta- tion, and the process may be termed, according to Holmes (’05), “Selectionof Random Movements,” or,accordingto Jennings (’04), “Trial and Error Movements.’’ Both processes cover the case equally well. The writer has been unable to detect any dissimi- larity between the two theories, as applied to the behavior of fly larve. The larvz stop frequently in their course, sway the head as above indicated, also circle frequently while crawling. ‘The same course may be pursued again or there may be a change in direction which is generally the case after a pause. It is clear that an over- production of random movements is involved; that a selection is made from these, depending on the force of the stimulation, and that the larve are thus guided on their way. On the other hand, it may be said that the larve reach the food successfully because they pause frequently in the course and sway the head about in order to try the conditions, then when they change the course, it back to the food. Comparatively few found it necessary to do this, since the more direct route was naturally along the lower side. The smallest larye seemed to have the most trouble in reaching the food. One very small larva (No. 6) remained within a distance of 2 cm. from the flesh for over an hour and a half. No larva left the flesh to return to the smal!er vial, though once in a while one started away, but always to return in afewseconds. The larve were under observation all day and all evening. Ecological and Experimental Study of Sarcophagide 81 is evident that an error has taken place. ‘Therefore, it does equally well to apply the “Trial and Error Theory.” The following quotation from Jennings (’05), p. 475, apparently ; makes little distinction, if any, between the two theories just men- tioned: “We perform movements which subject us to various con- ditions, till one is found that relieves the difficulty. We call the process searching, testing, trial, and the like. In the lowest and highest organisms the injurious condition acts as a stimulus to produce many movements, subjecting the organisms to various conditions, one of which is selected.” The use of the terms “selected” and “ selection” which frequently recur in the paper above quoted should not be overlooked with- out a thought as to their significance. ‘The first impression 1s that these terms imply intelligent choice on the part of the organism, but the author (Jennings) undoubtedly expects the broader inter- pretation, such as expressed by the term selection when applied to a magnet. This also holds equally well for the theory of the “Selection of random movements.” ‘There is in reality no intel- ligent choice involved; the organism responds reflexly to the stim- ulus, either positively or negatively. Each of the three theories of animal behavior evidently ex- plains much, but the writer believes, at least in reference to his - own experiments cited above and others cited below, that the tropism theory is not sufficient without either the second or third theory, and vice versa. It is still largely a matter of theory whether animal behavior can be so readily explained. Even in fly larve we have to deal with what seems to be a death feint, and that in itself leaves much to be explained. Phototaxts Observations were made on the same larve used in the first experiment. In the evening of the same day (June 30, 1905,) on lighting the lamp, the larve were noticed to leave the flesh at once and hasten toward the side of the vial nearest the light. This took them 4.5 cm. away from the food. Changing the angle between the vial and the light or rolling the tube over always resulted in a readjustment on the part of the larve. Moving the 82 William B. Herms lamp from one side of the tube to the other resulted likewise. Gradually increasing the distance between the vial and the light resulted in a return to the food when a maximum of thirteen feet was reached. On decreasing the distance again, the larve once more left the food when a distance of nine feet was reached. This experiment was repeated several times with like results. The lamp used was an ordinary oil lamp with small (No. 1) wick turned up fairly well. “Ihe adjustment to what was appar- ently the exact point of greatest photic stimulation was very remarkable, as was the almost frantic effort to gain this point when the angle was changed. Here we have an example of positive phototaxis overcoming the action of positive chemotaxis which is surely not a useful reaction. Stereotax1s The larve when placed in a receptacle which was ridged, preferred to crawl in the grooves. In one instance larve were kept in a bottle which had a convex bottom; on examination later, all were found in a circle wedged in close together around the margin of the bottom, with heads down and posteriors extended. On several occasions larve were found crawling in the crevices of the floor, and some of these were wedged in so tightly that it was a task to extricate them without injury to the larve. Positive stereotaxis 1s a prevailing phenomenon in the lower orders and fly larvae are no exception. Geotaxis Fly larve are positively geotactic, the burrowing habit (?) being very marked. On the other hand, imagines when first emerging from the pupa cases crawl out of the sand and up nearby grasses, remaining there until the wings are spread and dry. The cut showing this also illustrates how the flies cling to the grasses with head downward. Note—The experiments and observations relating to this paper were conducted at the Ohio State University Lake Laboratory at Sandusky, Ohio, chiefly during nine weeks of the summer, 1905, and a portion of the summers, 1903 and 1904. The writer is indebted to Prof. Herbert Osborn, Director of the Lake Laboratory and Asso- ciates, Profs. F. L. Landacre and J. S. Hine; also to Dr. W. E. Kellicott, Barnard College, for the kind assistance rendered and suggestions offered during the course of these studies. Ecological and Experimental Study of Sarcophagide 83 LITERATURE CITED Davenport, C. B., ’04—Statistical Methods with Special Reference to Biological Variation. New York. Francis, M.,’90—The Screw-worm. Bull. No. 12, Texas Agric. Exp. Sta., Sept. Hine, J. S., ’04—A Note on Insects as Scavengers, etc. Jour. Cols. Hort. Soc., vol. xix, Dec., pp. 123-128. Homes, S. J.,’05—The Selection of Random Movements as a Factor in Phototaxis Jour. Comp. Neurology and Psychology, vol. xv, Mch., pp. 98-112. Howarp, LELAND O., ’00o—A Contribution to the Study of the Insect Fauna of Human Excrement. Proc. Wash. Acad. of Sciences, vol. ii, Dec. 28. ’o2—The Insect Book. pp. 429. New York. Jennincs, HERBERT S., ’04.—Contributions to the Behavior of Lower Organisms. 7th paper. “he Method of Tnal and Error in the Behavior of Lower Organisms. pp. 235-252. Carnegie Institution, Washington. ’05— The Method of Regulation in Behavior and in other Fields. Jour. of Experimental ZoGlogy, vol. ii, No. 4. KELLOGG, VERNON L., ’05—American Insects. pp. 674. New York. Minot, CHARLES SEDGWICK, ’91—Senescence and Rejuvenation (Plates II, III, IV). First Paper: On the Weight of Guinea Pigs. Jour. of Physiol., vol. xii, pp. 97-153. Morean, H. A., ’90—Texas Screw-worm. Bull. No. 2, Second Series La. Agric. Exp. Sta. NeEeEpuaM, J. G., ’oo.—Insect Drift on the Shore of Lake Michigan. . Occasional Memoirs of the Chicago Entomological Society, vol. i, No. 1. Pia VENESCENCE-AS THE RESULT OF CON JUGATION BY SARA WHITE CULL Thirty years ago Bitschli proposed the view, since confirmed by Maupas and others, that the life histories of infusoria run in cycles, and that a period characterized by binary fission is fol- lowed by another in which conjugation takes place; this latter process resulting in a thorough reorganization of the excon- jugants and a Verjiingung or rejuvenescence, which shows itself in a higher rate of cell division and, generally speaking, in renewed life activities. If conjugation does not take place nor an equivalent stimulus be given the organisms they will eventually die of what has been termed “protoplasmic old age.” Hitherto it has been supposed that both cells in conjugation were benefited by the process, a mutual fertilization taking place; but in a series of experiments made by Calkins on Paramecium caudatum, the fact was noted that when both exconjugants live, in some cases one is far more vigorous than the other, as demon- strated by the greater number of offspring in one case than in the other Dr. Calkins suggested that I should examine this point and carry out some other observations that he had already made. The work was done in the zodlogical laboratory of Columbia University in the fall and winter of 1905-06. Butschli has pointed out the striking analogy which exists be- tween conjugation and fertilization as it is seen among higher organisms and among those protozoa which show sexual dimorph- ism. In many of these forms such as the peritrichous ciliates or the coccidiida, there is a marked sex-differentiation in the size and activity of the gametes. Here in fertilization, a more or less passive individual of normal or more than normal size, a macro- 1 Studies on the Life History of Protozoa. 1 Arch. f. Entwk., Bd. xv, 1 02. Tue JourNAL or ExperiMENTAL ZOOLOGY, VOL. IV, NO. 1. 86 Sara White Cull gamete, completely fuses with a smaller cell of greater activity, a microgamete. [he complete union of two cells alongwith differences in size and activity are characters which distinguish the process of fertilization as usually understood, from the process of conju- gation, as seen in forms like Paramecium. Both processes agree in having the same essential feature, the union of the nuclei of the two cells. In the different classes of protozoa all steps may be found from conjugation in a general sense to a process exactly similar to fertilization used in a strict sense. Even the maturation phenomena which play so important a réle in the history of metazoan germ cells are represented in some sort by processes which have been observed in a few protozoa. In isogamous union, such as that which takes place in Para- mecium caudatum, two individuals of the same size and approxi- mately equal activities unite for a short time, and the ectoplasm around the mouths of the two organisms fuses to form a sort of bridge over which the nuclei pass. During the maturation phases, previous to this nuclear exchange, the micronucleus of each organism gives rise by division to four or more pronuclei. ‘Two of these are destined to be functional and the others, cor- puscles de rebut, as Maupas calls them, disintegrate. One of the two functional pronuclei passes into the other organism where it fuses with the stationary pronucleus of that cell, forming one single reorganization nucleus. From this, by repeated division arise the micronucleus and macronucleus of the rejuvenated protozoan. These organisms then proceed to reproduce by ordinary fission. The species used for the experiments described here was Para- mecium caudatum and the material was what is known as “wild.” Each conjugating pair was taken up in a fine pipette and put into a hollow slide containing some drops of the culture liquid—hay infusion—free from all other protozoa. ‘These slides were then put into moist chambers. In all cases an examination was made after the isolation of the conjugating pairs to see that they had not been separated in the process of handling, forif this precaution were not taken, one could not be sure of dealing with the results of conjuga- tion. On the day following isolation, when, in most cases, the exconjugants were swimming freely through the water, each one Rejuvenescence as a Result of Conjugation 87 Was put into a small glass vial containing liquid similar to that from which they had been taken, and these vials were marked in such a way as to indicate the connection between the various individuals. “These vials were examined and the animals counted every few days fora month, anda fresh but not a new food medium was given them each time, the same being used for all the organisms. Ninety-three pairs of these wild conjugants were isolated at different times and of that number at the end of one month repre- sentatives of sixty-five pairs, or seventy per cent, were alive. At least one of the original conjugants remained of each pair, in the majority of cases both had given rise to offspring. Forty pairs of conjugants from long-continued cultures living in the laboratory on hay infusion were isolated by Calkins (loc. cit.) and examined from time to time. Only six pairs, or twelve per cent of these paramecia were represented by living forms at the end of a month. A comparison of these observations with those now made on the “wild” material would seem to indicate that the fertility of conjugation is dependent upon the condition of vitality in the individuals pairing, for, in both cases, the medium was the same. On the other hand, the explanation may lie in the fact that both conjugants had lived in the medium for many months, so that their chemical composition was too similar to pro- duce a new compound by fusion of their nuclei, this new com- pound being, perhaps, the source of energy for reorganization. A study of the mortality of these paramecia showed that at the end of one week the strains of both conjugants had died out en- tirely in six percent of the originalninety-three pairs. After three weeks had passed thirteen pairs, or approximately thirteen per cent, had died. These facts confirm Calkins’ observation that conjugation is by no means always successful in producing reju- venescence. The point which interested me chiefly in these experiments was that of double or reciprocal fertility—do both conjugants possess new power and ability to carry on the activities of life, or is but one of them fertilized as is the case among higher organisms? ‘The statistics which were gathered with this in mind show that, at the 88 Sara White Cull end of the month, of the sixty-five pairs then represented by living cells, in twenty-seven pairs, or forty-one per cent, one of the excon- jugants only or the offspring from it were alive; in fifteen pairs, or twenty-three per cent, the progeny of one exconjugant was three times as large as that of the other; in six pairs, the descendants of the one were twice as numerous as those of the other organism; and in only five cases had both conjugants given rise to the same number of offspring. ‘The twelve remaining pairs showed a wide disparity in the number of paramecia produced by any two con- jugants. The following table shows these results in summarized form: PROGENY oF ONE ProGeny or Boru PROGENY oF ONE ProGENY oF ONE ConyuGANT TWICE CoNnJUGANTS Conyucant Dean, ConjuGANT SHows Drab OF THE OTHER ALIVE bike ghia GREATER VIGOR OtHER* Number of Per- Number of Per- |Numberof| Per- Number | Per- Pairs centage | Pairs | centage Pairs | centage | of Pairs centage After 7 | days | 6 6 20 22 2y | 33 = = After 20 days | 13 13 25 31 28 35 = = After 30 days 28 30 27 41 21 | 32 48 74 * Exclusive of cases where the progeny of one exconjugant had died. It may be broadly stated that of the sixty-five pairs which I have observed one conjugant either died or left a weak strain in which the descendants were half as numerous and much less vigorous than those of the stronger exconjugant. This striking difference in the restored vitality of the conjugants and their descendants gives strong grounds for the belief that conjugation as seen among these infusoria is really incipient fertilization as seen among the higher forms of life. Here we have indications that one gamete gives up its vitality to and loses its individuality in the other just as the spermatozo6n loses its identity in the egg where its presence forms a stimulus to development analogous to the rajeunissement and greater activity in cell division which follows conjugation. There is little reason to doubt that a physiological and perhaps a Rejuvenescence as a Result of Conjugation 89 physical difference exists between the two unicellular organisms which unite in conjugation and a difference of the same nature as that expressed morphologically in the case of Adelea ovata, where the male gamete does not fuse with the female but dies after delivering one of its four pronuclei. Baltimore August, 1906 nd me PiPICIAL PARTHENOGENESIS IN THALASSEMA MEE LULA BY GEORGE LEFEVRE Wirnh Six Pirates I Introduction........ cit dtherd Ono aaa tc Boia ettcn & CRORE eee Te ene CI eri ae Raat gl MEP th cia ibarthenocenesismvAnnel OSs. 930s aise se qewweaan so view ae weve ss aegis aetee 93 boll iWdgtemell eine] IN IS GdINo so 8. Serie St tie Gt eee actin Oe eee ne ee eee ee ere a AR 97 IV Experimental Results ....... IMEI evete siora. ag Soe ects Se een isle. a, Nica bakes a Meee 98 me CI gras) Pav COCNOREMELIGAPENES 1. jae ern ciesole se ciel swwisyile s eve's isk vl aycceew vole eevemels Jaumels 99 2 Artificial Membrane-Formation and Parthenogenetic Development ................. 104 Tae serraousolithe Livin Materials. crys ccteloncaesss ¢ 205,46 ia oe gieeein deseo ene enlaces 109 THEM Ge CIMeM ted aC Chmbc metrak Aout re cu aogpivie le ie oh WA ooh older ween ae eels 109 PET A PIODE OME Ol ALPES OUIES cia 51 cis siete cs coda) 06 He -o gpioyayoy 0 eK) oi Syevevare foc eet ces eee seep pete eee 137 ue Ouipmlontie @leavayerCentrOsOmesy-n secretaries citer rnersisieleiee era eitesieeer te 140 Ae Numerical Relationsofthe| Chromosomes: .\.fvae s/o. eileis aslo sae wise scleiriodine eere 142 RATT IgV PP Te EP e Lays padzyspats.2 6.4) 3:5 fa wAcikt Ne b-oahavaistate Mrevels wid ieie}a Vax Seetere ganeig nays Me ede 144 oa -alcwire (CARad) 5.5 ce Sado lave SOS eka RE aes eae aon CR Deore ae epee ee RE 146 I INTRODUCTION In two brief abstracts (05, 06) I have published a few of the results which have been obtained from a study of artificial par- thenogenesis inthe echiuroid, Thalassema, mellita (Conn). I have Tue JourNaL oF ExPeRIMENTAL ZOOLOGY, VOL. IV, NO. I. g2 George Lefevre there shown that the eggs of this worm can be induced to develop into actively swimming trochophores, in the absence of sperm, by exposure for a few minutes to dilute solutions of acids, both inorganic and organic. Nitric, hydrochloric, sulphuric, carbonic, acetic and oxalic acids were used successfully, and in favorable experiments from 50 to 60 per cent of the eggs thus treated de- veloped into swimming larve that could scarcely be distinguished from normal trochophores of a corresponding stage. Continued and more detailed examination of the material has yielded many points of interest which are described at greater length in the present paper. The experimental part of the work and the observations on the living material were made at the laboratory of the U. S. Bureau of Fisheries, at Beaufort, N. C., during the summer of 1904, while the cytological study was completed in the following summer at the Marine Biological Laboratory at Woods Hole, Mass.! The development and life history of Thalassema mellita were first described by Conn (’84, 86), but as most of his observations were made upon the living material alone, his account is superfi- cial and inadequate, and, as has been recently shown by Torrey, many of his descriptions are radically wrong. A careful study, however, of the maturation and fertilization of this worm has been made by Griffin (’96, ’99), while the early embryology has been very accurately described by Torrey (’02, ’03). With the ex- ception of a brief communication by Kowalevsky (’72), and a note by Cowles (’03) on the rearing of Thalassema trochophores into the young worms, there exists no further literature on the develop- ment of this genus. Hitherto, the egg of Thalassema has been known to develop. only after fertilization by sperm, but my work has shown that it may readily be induced to develop parthenogenetically. It is, moreover, a particularly favorable object for experimental work of this kind. 1T wish to express my thanks to the Hon. George M. Bowers, U. S. Commissioner of Fisheries, for the privilege of occupying a table in the Beaufort Laboratory, and to Dr. Caswell Grave, Director of the Laboratory, for many courtesies extended tc me. My thanks are also due the Carnegie Institu- tion for the grant of a table in the Marine Biological Laboratory at Woods Hole in 1905. Artificial Parthenogenesis in Thalassema Mellita 93 My purpose in undertaking this research has not been primarily to analyze the physiological processes involved, but rather to study the morphological phenomena concerned in artificial parthenc- genesis, and especially, by a careful cytological examination of the material, to compare, as far as possible, the development arti- ficially produced with the normal events lez ding up to the forma- tion of the larva. My attention, therefore, has not been mainly directed to an investigation of the nature of the action which parthenogenetic agents exert upon the egg, nor to an exhaustive study of the conditions under which such agents-act, although in the course of the experimental part of the work a number of in- teresting facts have been brought out and noted. After having discovered that acids in dilute solutions could cause the formation of swimming larvz, I did not attempt to extend the method, as it seemed adequate for my purpose. II ARTIFICIAL PARTHENOGENESIS IN ANNELIDS Artificial parthenogenesis has been observed in the case of several other annelids by a number of experimenters. Loeb (or) first succeeded in causing development of the unfertilized eggs of Chzetopterus by increasing the osmotic pressure of the sea-water, by the action of KCl and other potassium salts in the absence of the osmotic effect, and by exposure to dilute solutions of HCl. By all of these means he obtained swimming larve which he states presented an appearance exactly like that of normal trocho- phores arising from fertilized eggs. Inasmuch as the changes leading up to the formation of these swimming structures were totally different from normal, developmental phenomena, and as the trochophore stage was apparently reached without visible signs of cleavage, Loeb concluded that normal cell lineage is an entirely secondary phenomenon. ‘The structure of these swimming larve of Chztopterus was afterwards carefully examined by Lillie, whose observations will be referred to below. In the same paper, Loeb states that he also produced certain changes in the unfertilized eggs of Phascolosoma and Podarke, the former dividing into 30 to 60 cells, while in the latter only the first cleavage occurred. 94 George Lefevre Observations on the annelids were next extended by Fischer ('02, 03) to Amphitrite and Nereis. Mathews (’or) had previously shown that artificial parthenogenesis could be produced in the starfish by mechanical agitation of the eggs, and this Fischer proved to be also true in the case of Amphitrite. He found that the eggs of this worm are extremely susceptible to mechanical shock, and can be brought to the “trochophore stage” by squirting them from a pipette after a residence in sea-water of from one-half toone hour. Fischer, furthermore, found that Ca (NO:): 1s capable of inducing parthenogenetic development in Amphitrite, a result which he attributes to the specific effect of calcium ions, although in Nereis he thinks the essential factor 1s the abstraction of water from the egg caused by the increased osmotic pressure of the sea- water. The morphological phenomena concerned in the develop- ment of the unfertilized eggs of Amphitrite and Nereis are nearly as widely divergent from the normal as in the case of Chzetopterus. The calcium eggs of the former show a totally different appearance from that of eggs fertilized by sperm. Although rarely cleavage may occur in a more or less normal manner as far as eight or twelve cells, the majority of eggs that divide do not go beyond the two-cell stage. Since a larger percentage of eggs reach the swim- ming condition than undergo cleavage, Fischer was inclined to believe that the formation of the trochophore could take place in the absence of segmentation, a result in harmony with Loeb’s conclusion for Chetopterus. In Nereis, parthenogenetic de- velopment is likewise far from normal in character, although cleavage, for the most part very irregular, seems to be of commoner occurrence than in Chetopterus and Amphitrite. The true morphological nature of these supposedly normal looking parthenogenetic larve of annelids has been clearly eluci- dated by Lillie (’02), who has shown that the unfertilized eggs of Chzetopterus, after exposure to salt solutions, pass, without segmen- tation, through certain phases of differentiation, resembling some of the normal processes, although the resulting ciliated structures are widely different from trochophores arising from fertilized eggs. Since my results in Thalassema are utterly unlike those of Lillie, it is necessary to refer to his observations more fully. Artificial Parthenogenesis in Thalassema Mellita 95 Mead (’95, ’98b) had already shown that the germinal vesicle of Chetopterus breaks down when the egg comes in contact with sea- water; the first maturation spindle forms, and the chromosomes pass into the equatorial plate, but the mitosis is not completed unless fertilization takes place. He also discovered the important fact that the addition of a small quantity of KCI to the sea-water produces the same effect as the spermatozoon, causing the extru- sion of the polar bodies, the formation of the yolk lobe, and other changes in the egg preparatory to the first cleavage (’98a, p. 213). Lillie, however, extended these observations, and found that, after exposure for about one hour to solutions of KCl in definite con- centration, the unfertilized eggs of Chatopterus may undergo a process of cytoplasmic differentiation unaccompanied by cell division, and in about twenty-four hours after the beginning of the experiment give rise to ciliated structures which in some cases more or less simulate the appearance of trochophores. They usually contain but a single nuclear area, and the cytoplasm is differentiated into a ciliated ectoplasm and a yolk-laden endoplasm which are comparable with the ectoderm and endoderm of the trochophore. Since the KCI solutions cause a disintegration and ultimate disappearance of the cell membrane, the naked cytoplasm is left unprotected, and fusion-phenomena between different eggs are of common occurrence, as Loeb (or) had previously observed, the agglutination being greatly increased bythe addition of a small quantity of CaCl, to the K-containing sea-water. In harmony with the results of Loeb, Lillie also observed that the period pre- ceding differentiation of the cytoplasm is characterized by amce- boid movements which may exhibit an astonishing degree of activ- ity. If cleavage takes place at all, as it does in some eggs, it rarely goes very far, and only in a small proportion of such eggs does cell division approximate the normal. Division of the cytoplasm unassociated with nuclear division is of common occurrence, the non-nucleated portions of protoplasm always fusing sooner or later with the general mass. A comparison of my own results with the observations and conclusions of Lillie will be made further on. Results essentially similar to those of Lillie have been obtained by Treadwell ('02) in Podarke obscura after treatment of the 96 George Lejevre unfertilized eggs with solutionsof KCl. Inthis case differentiated, ciliated structures may also arise without segmentation, and pseudo- cleavages, involving only the cytoplasm, are of frequent occurrence. Ciliated embryos, however, may also be produced as the result of a cleavage process, in which both cytoplasm and nucleus are con- cerned, but here cell-division is quite abnormal. Fusion of sepa- rate eggs was observed in Podarke, but it israrer than in Chetop- terus and fewer eggs unite into a common mass. Bullot (’04), experimenting with the eggs of an annelid, Ophelia, has obtained results which are not in accord with those just cited, in that he has shown that the parthenogenetic larve of this worm, produced by solutions of KCl and NaCl, arise only from seg- menting eggs. How nearly normal the processes of cleavage and differentiation are in this case cannot be determined from the inadequate figures and description given, although he states that “the divisions go on regularly into four, eight, sixteen, and more cells,” and that later a blastula of characteristic shape is formed. Lastly, Scott (’06) has studied the morphological phenomena of parthenogenesis in the eggs of Amphitrite which were subjected to the action of salt solutions, especially solutions of Ca(NO:).:, and to mechanical agitation. Usually from 5 to 25 per cent of swim- ming structures were obtained. He found that certain differen- tiations may occur with or without cleavage and with or without the formation of polar bodies. Cleavage of the egg may take place, but it is generally abnormal, always so in later stages. A’ ciliated body is produced, which may show more or less extensive cell divisions but usually exhibits no true segmentation; the mass may, however, contain many nuclei. In no instance was anything remotely approaching a normal larva obtained, although certain cytoplasmic differentiations were present, as the develop- ment of an ectoplasmic layer, the growth of cilia, and the appear- ance of vacuoles and pigment. It is clear, then, that the previous work on artificial partheno- genesis of annelids, with the methods which have been employed, has shown little in common with the processes of normal develop- ment, and that at best a ciliated structure has been produced which exhibits certain specialized regions of the cytoplasm but no nor- Artificial Parthenogenesis in Thalassema Mellita 97 mally differentiated organs. [hata far more normal result, how- ever, has been obtained in the parthenogenetic development of Thalassema will be pointed out in the following pages. x III MATERIAL AND METHODS Thalassema mellita inhabits the dead tests of the sand-dollar Mellita pentapora (Gmelin) and occurs abundantly onthe exten- sive shoals in the harbor of Beaufort, N. C. The egg is a parti- cularly favorable one for experimental work in artificial partheno- genesis. With a little experience one has no difficulty in distin- guishing the sexes by the difference in color of the sexual products which show through the semi-transparent body wall, the sperma- tozoa appearing a milky-white and the eggs a light golden-yellow. Males and females may, therefore, be separated, and it is not necessary to touch the former during the course of an experiment. The full-grown o6cytes are contained in the segmental tubes which fill a large part of the body cavity, and upon rupture of the tubes great numbers of eggs may be obtained in perfectly clean cultures, entirely free from immature eggs, slime and débris of all kinds. Every precaution was taken to avoid contamination by sperma- tozoa, and it may be stated at the outset that in all experiments the control eggs were absolutely negative and never showed a single case of cleavage or differentiation of any nature whatever. The female worms were first separated from the males and kept by themselves in a dish of sea-water over night. Before using, they were thoroughly washed in fresh water, as were the dishes and instruments employed in the experiment and the hands of the operator. ‘The body wall was then slit open, whereupon the tubes, gorged with eggs, burst out through the opening. The tubes were first rinsed in sterilized sea-water, in order to remove the blood adhering to their surface, and then snipped off and dropped into a dish of sea-water which had previously been raised to a temperature of 70° or 80°. The eggs, when first removed from the tubes, are collapsed and pressed out of shape, as a result of close packing in the confined space, and do not become spherical until after fertilization by the 98 - George Lefevre sperm or treatment with the solutions. The unfertilized control eggs, however, for the most part retain the compressed form when allowed to remain in normal sea-water, and finally die in this con- dition. The eggs, furthermore, when taken from the tubes, are naked, and the failure of the control eggs to ever form a membrane furnishes an additional check on the experiments, for, as will be shown beyond, every egg subjected to the action of the acid solu- tions throws off a fertilization membrane in all respects identical with the membrane which appears upon the entrance of the sperma- tozoon. For the study of sections, eggs and embryos were killed in Wilson’s picro-acetic mixture (2 per cent acetic) which gave ex- cellent results. Osmic acid (1 per cent), followed by prolonged immersion in Miller’s fluid, proved very satisfactory for later stages, while weak formalin and a mixture of 24 per cent formalin and 50 per cent alcohol were useful for the demonstration of cilia in total mounts of the older embryos and larva. Sections were stained in iron hematoxylin, with or without a counter stain, while Conklin’s Delafield’s hamatoxylin gave the best results for whole preparations of the cleavage stages. Most of the sec- tions were cut 5 in thickness, and all of the drawings were made with the camera. IV EXPERIMENTAL RESULTS At the beginning of the investigation, the attempt was made to induce parthenogenetic development by the use of salts, and MgCl,, Ca(NO,)., KCl,and NaCl were tried. All, however, gave negative results, except in a few cases an irregular fragmentation of some of the eggs was produced, but it never led to the formation of swimming larve. In many of these experiments the osmotic pressure of the sea-water was increased, and it would seem, there- fore, that parthenogenetic development of Thalassema cannot be produced by subjecting the unfertilized eggs to the action of hypertonic sea-water. It should be stated, however, that the range of salt solutions employed was not exhaustive, since early in the work it was found that acids gave promise of better results, Artificial Parthenogenesis in Thalassema Mellita 99 and experiments with salts were discontinued. It is not 1m- possible that favorable solutions of these and other salts might have been found which would have caused development, had the attempt been made to investigate the action of salts in greater detail. I was also unable to obtain parthenogenetic development by mechanical agitation or by exposure of the eggs to low tempera- tures. Unlike Asterias and: Amphitrite, the eggs of Thalassema show no changes whatever after either gentle or violent agitation, and the low temperatures which Greeley (’02) found to be capable of producing development of the unfertilized eggs of the starfish were utterly ineffectual in bringing about a similar result in the case of Thalassema. On the other hand, parthenogenetic develop- ment, which was strikingly normal in a great many experiments, took place after treatment of the eggs for several minutes with dilute solutions of certain acids. Nitric, hydrochloric, sulphuric, carbonic, acetic and oxalic acids were employed, and all yielded about equally successful results. 1. Acids as. Parthenogenetic A gents The method of causing artificial parthenogenesis by the use of mineral acids was first elaborated by Loeb and Neilson (’01), who employed a solution of 3-5 cc. Jo Morganic acid + 100 cc. sea-water, with an immersion of 3-20 minutes, and in the case of Asterias succeeded by this means in bringing about 20 per cent of the eggs toa gastrula stage. ‘They ascribed the result to the specific action of H-ions. Since these initial experiments, inorganic acids have been used with some success as parthenogenetic agents by Loeb and others on the eggs of echinoderms and worms. The use of CO, in artificial parthenogenesis is due to Delage ('02, 04) who found the method to be remarkably successful with the eggs of the starfish. The eggs were placed in sea-water charged with CO, by means of a “sparklet,”’ and after an immersion in the charged sea-water for about an hour, practically every egg developed into a swimming larva. The larve were kept alive for three and one-half months, and were reared to the beginning of 100 George Lefevre metamorphosis. ‘The oldest larva, which were still in an active and healthy state, were accidentally killed, but not before the early stages of the transformation into the starfish had already made some progress. Delage, moreover, observed that partheno- genesis in Asterias is independent of the formation of polar bodies and occurs when one, two or no polar bodies have been extruded, a result in harmony with what I have found to be true in Thalas- sema. He states, however, that the treatment must be given at some time during the process of maturation, as it is not effective either before the breaking down of the germinal vesicle or after the formation of the egg pronucleus. It will be seen that this limitation does not hold for the eggs of Thalassema which were in every case subjected to the action of the acid while in the germinal vesicle stage. In my experiments I not only found that the inorganic acids, which had been used in a few cases by previous workers, were eficient parthenogenetic agents, but that certain organic acids as well, namely, acetic and oxalic, were equally successful, if not superior. A considerable variation was observed in the behavior of the eggs of Thalassema in the solutions employed, and, owing prob- ably to differences in internal conditions, possibly in the degree of ripeness, the same solution and the same duration of immersion did not in all cases produce the same result, either in the character of the development or in the percentage of eggs involved in the process. ‘This variability, moreover, was independent of tempera- ture and could not be controlled. It is true, however, in the case of each acid employed, that an optimum solution and an optimum duration of exposure were found which could be relied upon to yield satisfactory results in the majority of experiments. In favorable experiments, which were the rule and not the exception, from 40 to 60 per cent of the eggs underwent development and gave rise to actively swimming trochophores which closely coin- cided with normal larve in appearance. The following solutions, with the time of immersion, are those that gave the best results; they were, in consequence, most fre- quently used for obtaining embryos and larve: Artificial Parthenogenesis in Thalassema Mellita IOI m Minutes. Recreate RO 179 COSEA-WAEED oa coy 2 alates ale, Semen oe een ee 5 IO m Me eer EMIOW -J9O'5 CE1 SEA-WALEE, occ yr sip. v aie ayy s\anaiaternps elniee@ sv ae 5 Ke) : m Meee eda Oal-}- GOCE, SEA-WATED oc <5 feca eee culeieee eos 0 ees 8 20 m . . Pore a Oxalic acid --}~ 88 CC. SCA-WAUEF . 05s oe se ote es eee vere s © 8 20 m . . icone ACHE sacid 4-65 CC.S€a-Watel .....0g 6s e4 veep aces ene 5 Io In the case of CO,, the gas was passed from a generator into sea-water for ten minutes and the eggs immersed in the charged water for one hour, after which they were transferred to pure sea- water.2. The result was very satisfactory and usually about 50 per cent of swimming larve were obtained by this method. Although a wide range of solutions and exposures were tested in the case of each acid, in the table on page 12 are placed a few results which are selected from a great many experiments and which will serve as characteristic illustrations. After determining by experiment the optimum solution and exposure in the case of each acid, satisfactory results were usually obtained by adhering more or less closely to such conditions as experience had proved to be the best, but an examination of the following table will show that the expectation was not always fulfilled. For example, in Nos. 3 and 4, when the same solution of HNO: and the same exposure were employed, one experiment yielded 40 per cent of swimming larva, while the other gave only 5 per cent; and again, in Nos. 5 and 6, 60 per cent and 25 per cent were obtained, respectively, from an equal exposure to the same HCl solution. It is difficult to assign causes to this seemingly capricious difference in the relative proportions of developing eggs in experiments carried on under conditions as nearly identical as possible. In addition to the variability of the results obtained in different experiments, where the same solutions and exposures were used, 2A “sparklet” apparatus was not available at the time my experiments were made. 102 George Lejevre No. Solutions Employed | m i ||)t7 cc, — NG; -- 83 (ce..5.We 10 | m z | 18 ce. —HNO; = 82 cc. S_W- 10 m 3 | 18 cc. —HNO3 + 82 cc. S.W. 10 m 4 18 cc. —HNO3 + 82 cc. S.W. 10 m 5-. | 15 cc.— HCl +-.85 ce. S.W. 10 1 m 6 15 cc. — HCl -— 85 cc. S-W- 10 m 7, |iGec— HCl S4ec> SW 10 m 8 | 18 cc. — HCl + 82 cc. S.W. 10 m 9 10 cc. —H»SOx4 + go cc. S.W. 20 m H.SO,z + 88 cc. S.W. TO) 4 |, 32\cC: 20 m II | 15 cc. —HSO, + 85 cc..S.W. 20 m 12 cc. — Oxalic + 88 cc. S.W. 20 m 13 | 15 cc. as Oxalic + 85 cc. S.W. m 14 10 cc. —Acetic + go cc. S.W. 10 , m 15 15 cc. — Acetic + 85 cc. S.W. 10 16 15 cc. ee Reptic Se eS SAE 10 17 COz passed into water for 1o minutes. 18 CO, passed into water for 20 minutes. Time of Exposure 5 minutes 3 minutes 4 minutes | 4 minutes | | 5 minutes | 5 minutes | | - 5 minutes } 4 minutes oo i=] o o n | 8 minutes 5 minutes 8 minutes | 6 minutes 7 minutes 5 minutes 6 minutes 1 hour | | y hour Percentage of Swimming Trochophores 55 17 40 “7 14 35 10 5° 45 3 Artificial Parthenogenests in Thalassema Mellita 103 I was greatly struck with the marked difference in results observed when the strength of the solution or the duration of the immersion was varied bya very slight degree. For example, ina given experi- ment 60 per cent of the eggs developed into actively swimming trochophores, which could not be distinguished from normal larva, after five minutes’ exposure to the following solution: 15 cc. = HC1+85 cc. sea-water. Another lot of eggs from the same female, treated with the same solution, but for 6 minutes instead of 5, yielded only about 5 per cent that underwent any develop- ment at all, while in none of the eggs did this proceed beyond the early cleavage stages. Here a difference of but one minute in the time of exposure gave rise to a profound difference in the result, in the one case the solution being adequate to initiate the developmental processes in a majority of the eggs, which then pro- duced apparently normal larve, while in the other case only an abortive early development was induced in a very few eggs. Such differences, however, in the relative proportion of larva were by no means constant; in Nos. 15 and 16 of the table it is seen that a difference of one minute in the exposure to the same solution of acetic acid had no effect upon the percentage of larva obtained. The following table illustrates a similar variability in cases where the duration of immersion was constant, but the solutions differed very slightly in the degree of concentration: | ‘ | Percentage No. | Solutions Employed ELS of Swimming | Expoome Trochophores | m | I 17 cc. —HNO3 + 83 cc. S. W. 5 minutes 40 10 | | m 2 | 18 cc. —HNO, + 82 cc. S. W. 5 minutes 3 | - Io ii | m | ae |. 614 cc.—Acetic + 86 cc. S. W. 6 minutes ° fe) m eas) CC. —— Acetic | 85 cc. S. W. 6 minutes 55 10 104 George Lefevre In each of the two cases cited above, the eggs were taken from the same females and placed in the solutions at the same time, the only difference being that the second solution was stronger than the first by I cc. of the dilute acid. 2 Artificial Membrane Formation and Parthenogenettc Development The unfertilized eggs of Thalassema after transference from the acid solution to normal sea-water, throw off a membrane iden- tical with that which is formed upon entrance of the spermatozo6n. The artificial production of a membrane has been observed by former experimenters. O. and R. Hertwig (’87) first discovered that, by the addition of chloroform to the sea-water, the unfertilized eggs of the sea-urchin may be caused to form a fertilization mem- brane which is entirely normal in appearance. Herbst (’93) later confirmed the result obtained by the Hertwigs, and found that not only chloroform but several other substances, namely, clove oil, creosote, xylol, toluol and benzol, act in a similar manner, the best results being given by benzol. More recently Herbst (’04) has obtained a normal membrane formation by the use of silver salts. Loeb (05d, ’o5e) tested the action of hydrocarbons in this respect and found that the ripe eggs of Strongylocentrotus and Asterina, when put into 50 cc. of sea-water which has been shaken with I cc. of benzol or amylene, immediately form membranes which are identical in appearance with the normal fertilization membrane. By subjecting unfertilized eggs of Strongylocentrotus purpuratus to a 24 to 14 7 NaCl solution or to a 24 m cane sugar solution, he also succeeded in causing a mem- brane formation, but in these experiments the osmotic pressure was so high that the eggs were greatly injured and underwent cytolysis without subsequent development. Solutions of lower osmotic pressure caused development, but not membrane forma- tion (05a, p. 79). It should be mentioned that Wilson (or, p- 533) states for Toxopneustes that “some of the magnesium eggs showed a faint ragged membrane, but others were absolutely devoid of a Bae Ee ” although he gives no details of his observations on Artificial Parthenogenesis in Thalassema Mellita 105 this point. Hunter (’04, p. 214) also records the presence of a membrane surrounding unfertilized eggs of Arbacia after treat- ment with MeCl,. Loeb (05), in a series of recent papers, has published the results of experiments which have confirmed my observations on the formation of a membrane after exposure of unfertilized eggs to acid solutions. Although our observations agree as to the power of acids to call forth a membrane formation, certain marked differences occur in our results, and it may be well to compare his experiments and my own in this place. By the use of an improved method, Loeb has succeeded in closely imitating the process of normal development in the unfertilized eggs of Strongy- locentrotus purpuratus. If the eggs are treated with hypertonic sea-water alone, no membrane is formed, and only a small per- centage undergo any development at all. ‘The rate of develop- ment of these is much slower than in the case of fertilized eggs and the larve arising from them do not rise to the top but swim at the bottom of the dish. By first exposing the unfertilized Z n eggs, however, to 50 cc. of sea-water to which 3 cc. — of a fatty 10 acid, e. g., formic, acetic, propionic, butyric, valerianic or caproic acid, are added, for from } to 14 minutes, they forma character- istic fertilization membrane when put back into normal sea-water. The membrane was not produced as long as the eggs were left in the acidulated water, nor was it formed when they were taken out a little too early or too late. Eggs treated with the acid alone do not develop, but in a few hours begin to disintegrate, and after twenty-four hours practically all are dead. Subsequent treat- ment, however, with hypertonic sea-water produces a surprising result. If, after the appearance of the membrane, the eggs are placed in 100 cc. of sea-water, to which 15 cc. of a 24 1 NaCl solution has been added, for from 20 to 50 minutes, go to 100 per cent of the eggs develop with the normal rate of segmentation. “A large percentage of the blastule originating from this com- bination of methods looked perfectly normal, and rose to the sur- face of the sea-water. Their further development into gastrulz and plutei occurred with the same velocity as that of the control 106 ; George Lefevre eggs, which had been fertilized by sperm; and the larve showed an equal degree of vitality” (Loeb ’06, p. 168). He did not find, however, that all acids were effective in causing the membrane to appear; with the exception of CO.,, all the membrane forming acids were monobasic, organic acids. Mineral acids and dibasic or tribasic organic acids, such as oxalic or citric, were not suitable for the purpose. The acids that gave the best results were the lower representatives of the fatty acid groups. A comparison of the foregoing with my own observations will show some rather striking points of difference. In the first place, the acids which I found to be capable of calling forth the mem- brane formation are of a more widely different character, while the response to the acid treatment of the egg of Thalassema in throwing off the membrane does not seem to be as narrowly restricted as in the case of Strongylocentrotus. Although but few acids were available for use when my experiments were made, those which I employed represent a series of considerable range. Unlike Loeb’s experience, I found that the mineral acids, HNOs, HCl, and H,SO,, were quite successful with the eggs of Thalas- sema, and, as a matter of fact, some of my best results were ob- tained with HCl. Among organic acids, furthermore, the mem- brane forming power was not limited to monobasic acids, for one dibasic acid, at least, oxalic acid, yielded about as good results as did the monobasic acetic acid. Contrary to the conclusion to which Loeb came in the case of Strongylocentrotus, in Thalassema, at all events, it would seem that the acid effect is essential to membrane formation. It should also be pointed out that the limit between the minimal and maximal exposure of the eggs of Strongylocentrotus required for membrane formation is narrower than in Thalassema. Al- though the minimal exposures given by Loeb (05, p- 122)MaE several different solutions of acetic acid correspond quite closely at the same temperature with my observations, the eggs of Thalas- sema will form membranes after being subjected to stronger solu- tions and for a much longer time than is true of the sea-urchin. By referring to the table on p. ror, it will be seen that the opti- mum treatment in my experiments would have been a decided a Artificial Parthenogenesis in Thalassema Mellita 107 over-exposure for Strongylocentrotus, according to the figures given by Loeb. The most striking difference, however, between the behavior of Strongylocentrotus and that of Thalassema 1s seen in the subse- quent events, after the membrane has been produced by the acid treatment. Whereas, in the former the unfertilized eggs undergo no further development unless exposed to hypertonic sea-water, in Thalassema the action of the acid, or perhaps the changes which the egg undergoes as a result of the membrane formation, is suficient to lead the egg on to cleavage and the ultimate forma- tion of a swimming larva. Treatment with hypertonic sea-water is, therefore, unnecessary for the further development of the unfertilized eggs of Thalassema. During a brief stay at Beaufort, N. C., in June, 1906, I had an opportunity of repeating Loeb’s experiments while testing the effect of his combination of methods upon the eggs of Thalassema. Although the membrane was formed, as it always is in Thalassema after treatment with an acid, the subsequent use of hypertonic sea-water not only did not give rise to larva, but actually inter- fered with the development which would otherwise have taken place with the acid alone when used in the proper concentration. When the unfertilized eggs, after an exposure to an acid, were placed in a solution of NaCl, either no developmental changes occurred and the eggs early disintegrated, or at best only a few early cleavages, usually irregular in character, were produced. I at first tried the same solutions and exposures as Loeb had used with Strongylocentrotus, and, after finding them to be futile, I varied the different factors between wide limits. “The acids were, of course, used in solutions strong enough to call forth the mem- brane formation, but too weak to cause the eggs to develop, in order to test the ability of NaCl solutions to bring about develop- ment after the membrane had been previously produced by acids. In no case, however, did I succeed in discovering a combination that could cause a normal development and the formation of swim- ming embryos. The following solutions gave the best results , m ! obtained: 10 cc. a HCl + 90 cc. sea-water for five minutes, and 108 George Lefevre 15 cc. 25 m NaCl+ 85 cc. sea-water for from 40 to 50 minutes. The percentage of eggs that divided, after treatment with this combination, was low (about 10 per cent), the cleavage was abnormal, and in no case was a ciliated embryo produced. It should be stated that such developmental changes as did occur in experiments like this must be ascribed to the effect of the hyper- tonic sea-water, since control eggs showed that the same acid solution, when used alone, was too weak to cause any develop- ment. It was clear from a great many experiments which I made that the use of hypertonic sea-water combined with the acid treatment is not only not an improved method of artificial parthenogenesis for Thalassema, but, on the contrary, yields immeasurably poorer results than the method employed in my original experiments. I might also add that the combination of methods gave absolutely negative results with the eggs of the sand-dollar Mellita pentapora, which I subjected to a similar treatment. Loeb ('05e) also found that the eggs of a starfish, a species of Asterina occurring in the Bay of Monterey, form a membrane after exposure to solutions of a fatty acid, although a higher con- centration is necessary for this form than is required for Stron- gylocentrotus. The eggs, however, do not have to be subsequently treated with hypertonic sea-water, as the process of artificial membrane formation is sufficient to cause development in Aste- rina, and in this respect the case is similar to that of Thalassema. In harmony with the observations of Delage (’02c) on Asterias, Loeb also found that the eggs of Asterina, which maturate nor- mally upon being placed in sea-water, could not be made to develop by exposure to an acid solution until after the breaking down of the germinal vesicle. Thalassema eggs, on the other hand, differ markedly in this respect from those of the starfish, for, unless they are fertilized by sperm or acted upon by a par- thenogenetic agent, the germinal vesicle remains intact, however long they may lie in sea-water. Artificial Parthenogenesis in Thalassema Mellita 10g VY OBSERVATIONS ON THE LIVING MATERIAL 1 The Unsegmented Egg In normal fertilization the spermatozo6n enters the odcyte and lies more or less quiescent in the cytoplasm during the changes involved in the process of maturation which are initiated very shortly after the appearance of the sperm inside the egg. Almost immediately the egg throws off the membrane which soon draws away from the surface and becomes completely detached, while at the same time it becomes spherical, probably as a result of the absorption of water. In the parthenogenetic eggs, the same changes take place. After exposure to the acid solutions, the eggs were in all cases transferred at once to sterilized sea-water, when a typical fertilization membrane became apparent in a very short time. The eggs do not round out as quickly, however, as they do after the entrance of the sperm, but usually in about 30 minutes from the time they are placed in normal sea-water they assume the spherical form, although it occasionally happens that this change is considerably delayed and an hour or more may elapse before they recover from the flattened, compressed con- dition in which they are when taken from the tubes of the female. After treatment with the acid and transference to sea-Wwater, the membrane formation was of absolutely universal occurrence and was exhibited by every egg, whether sojourn in the acidulated water wzs of the proper duration to produce subsequent cleavage and development, or not. The eggs of my experiments exhibited the phenomena which simulate the appearance of “‘spinning”’ activities and which have been described by Torrey ('03) in the normally fertilized eggs of Thalassema. If the perivitelline space be examined under a high power, shortly after the appearance of the membrane, excessively fine protoplasmic threads may be seen passing from the surface of the egg to the membrane and varying from time to time in thickness and constitution. ‘These delicate strands persist during the cleavage stages, while some appear to be attached to the polar bodies and give the impression of holding them in place, as Torrey has described. ‘The connections with the polar bodies, I1O George Lefevre however, are soon interrupted, for the latter usually break away from their original position, and either pass into the cleavage space or float freely about in the perivitelline fluid. Torrey ex- plains the presence of the threads by supposing that, when the membrane separates from the egg, the protoplasm adheres to the corrugations on the inner side of the membrane, and, because of its viscid nature, is drawn out into the threads. If the mem- brane is merely the denser surface layer of the egg which is mechanically lifted up as a result of the expression of a liquid secreted by the egg, as Loeb (’05e, p. 155) is inclined to believe, the formation of the threads can be readily explained. Fig. 1, which is drawn from the living egg, shows these strands of denser super- ficial cytoplasm in an egg which has already maturated and formed the first cleavage amphiaster. 2 Formation of Polar Bodies The first visible change in the interior of the living egg, after transference from the acid solution to normal sea-water is the bodily migration of the germinal vesicle from a position near the center to the animal pole of the egg. ‘This change in location, which may be seen by comparing Figs. 2 and 3, drawn from sec- tions, does not always take place, but it has been observed in a great many cases. In the normal egg the migration does not occur, but the germinal vesicle breaks down near the center of the egg and the first polar spindle later rotates into a radial position, with the outer aster close to the surface, in the usual manner. In the parthenogenetic eggs, however, owing to the outward movement of the germinal vesicle, the spindle arises, as a rule, considerably nearer the surface of the egg. The extrusion of the first polar body takes place in from 45 to gO minutes after removal from the acidulated water, thus showing a great retardation of the maturation, since in the normal egg the first polar body is formed in about 20 minutes after the entrance of the spermatozodn. It soon moves away from the surface of the egg and the second arises immediately under it and very shortly afterward, although the interval between the appearance of the two is more variable than in normal maturation. Artificial Parthenogenesis in. Thalassema Mellita III In the living parthenogenetic egg, the polar bodies are abso- lutely indistinguishable in position, size, form and other character- istics from those formed after fertilization by sperm. Conn (’86) described the almost invariable division of the first polar body, but as his observations were limited nearly entirely to the living egg, he did not follow the details of the mitotic phenomena. Griffin (99), however, from a careful study of sections, determined that the first polar body divides by a complete and typical mitosis, although in certain minor details there are signs of degeneration. The same mitotic phenomena and division occur almost without exception in the parthenogenetic eggs, and will be described be- yond. Fig. 4, showing the divided first polar body, is drawn from a living egg which had been exposed to the action of acetic acid; the first polar body in this case appeared in 45 minutes after removal from the solution, and the second 15 minutes later. It is of interest to note that the polar bodies respond to the same divisional stimulus supplied by the acid solutions as does the egg cell itself, as both bodies have been frequently seen to undergo several cleavages. This revived activity of the degenerate polar cells, which may be regarded as an abortive parthenogenesis, results in the formation of a miniature, morula-like cluster of minute cells which, however, soon break away in a mass from the surface of the egg and may persist for some time in the space beneath the membrane. Exposure to the acid solutions restores to a certain degree the energy of division in these rudimentary germ cells and an attempt at development follows. As many as sixteen cells have been counted with certainty in the miniature embryos, and, although I have not been able to determine it in all cases, it 1s certainly true that some at least of these subsequent cleavages of the polar cells take place mitotically. In Fig. 1 is seen an instance in which at least ten cells have been formed, and Fig. 62 is drawn from a section which has passed through five cells of a cluster; two of these show indications of mitotic activity. This parthenogenetic development of the polar bodies in Thalassema should be compared with the observations of Fran- cotte (97) on a turbellarian, Prosthecerzeus; here the first polar 112 George Lefevre body, which is often abnormally large, is occasionally fertilized by a spermatozoon and dev elops into a small gastrula. Although in the great majority of experiments normal polar bodies were eased by the unfertilized eggs, this was not always true, for in a number of cases the eggs divided and eventually gave rise to trochophores without any external indication of a previous maturation, or after the formation of but a single polar body. It occasionally happened that eggs in the same dish would show all of these conditions, 1. €., some would mature normally, some would extrude only one polar body, while others would form none; yet all of these classes of eggs might undergo development and produce larve indistinguishable from each other. It was more often the case, however, that in any one dish, in which the eggs had been exposed to the same solution, all the developing eggs would either maturate normally, or else throw off only one polar body, or again none at all. As the egg of Thalassema is quite opaque, the internal phenomena involved in these changes can only be examined in sections, and their description must, therefore, be reserved for the portion of the paper dealing with observations on the preserved material. 3 Cleavage After the formation of the polar bodies, there are no further signs of change visible in the living egg for some time. The first cleavage does not take place at a definite period after the eggs have been subjected to the action of the solutions, but it may appear at any time, varying from 2 to 34 hours, the shorter interval, how- ever, being the more frequent one. ‘The appearance of the first cleavage is not correlated with the time of extrusion of the polar bodies, and a delayed maturation does not necessarily mean a corresponding postponement of segmentation. As the first cleay- age occurs after normal fertilization in about 50 or 60 minutes from the time the spermatozoOn enters, it is seen that the activi- ties which lead to segmentation are called forth much more slowly in the parthenogenetic eggs. The early cleavages are closely similar to the normal in a great many cases and in favorable experiments where the optimum Artificial Parthenogenesis in Thalassema Mellita 113 conditions were present, the segmenting eggs could not be dis- tinguished from controls fertilized with sperm, except for the lack of uniformity in the rate of division exhibited by the former, especially during later stages. “he rhythm of division is generally more or less disturbed in eggs developing parthenogenetically, and the intervals between successive cleavages are, therefore, less constant than in eggs normally fertilized. All gradations, how- ever, are encountered from cases in which the rate of segmentation closely approximates the normal to those exhibiting nearly every stage of cleavage in the same dish at a given time. As in the case of the normal egg, the first furrow begins at the upper pole and cuts in somewhat more rapidly here es at the lower, resulting in the formation of two equal blastomeres (Fig. 5). The second cleavage is also equal and gives rise to four blastomeres of exactly the same size. By comparing Fig. 6 with Torrey’s Fig. 1B (03, p. 173), it will be seen that the same relations exist here in regard to the polar furrows as are present in the normal egg. [he two upper blastomeres do not quite touch, and one of the polar bodies has passed into the space thus left between them. The four cells, constituting the first quartet of micromeres, are formed at the time and nearly equal the macromeres in size (Fig. 7). The cleavage space from now on increases rapidly in size and frequently one or two polar bodies may be seen lying in it (Fig. 40, pb). The origin of the second quartet of micromeres by division of the macromeres, and the unequal division of the first quartet to form the primary trochoblasts, whereby the 16- cell stage is established, as Torrey has described, may be clearly followed in the parthenogenetic eggs and found to be perfectly normal in a great many cases. But beyond this stage I was un- able to observe the cleavages with any degree of certainty in the living egg on account of its opacity and the flattening down of the blastomeres during the resting period, a difficulty which Torrey also encountered. A 16-cell stage, drawn from the living egg, 1s seen in Fig. 8. 114 George Lefevre 4. Formation of Later Embryo and Larva Later cleavage stages, the formation of the blastula and gas- trula, and the differentiation of the early trochophore will be con- sidered in the chapter dealing with the observations on preserved material, as most of the developmental changes, which can be de- termined with accuracy in the parthenogenesis of Thalassema, must be made out from an examination of sections and total preparations. As Torrey has stated (op. cit., p. 187), cilia first appear on the normal blastula with great eiiery at four and one-half hours after fertilization, and simultaneously on the prototroch and rosette. Although I have occasionally observed the appearance of the cilia of the prototroch and the apical flagella on the parthe- nogenetic embryos at the same time, 7. ¢., four and one-half hours after removal from the solutions, it is more usually the case that their formation is considerably delayed and they are not seen for from sixto nine hours. In the few experiments where the cilia were first observed at the normal time, the rate of development from the first cleavage onward coincided very closely with the normal rate. In favorable experiments, the cilia show perfectly typical relations; those of the prototroch are at first short and delicate and form a broad band completely encircling the embryo, while the apical cilia soon become quite long and project in front as a pencil of rather stiff straight flagella. “Torrey’s description of the normal cilia at this stage corresponds in all respects with the observations which I have repeatedly made on the parthenogenetic embryos, and may be quoted: “At first the cilia on the prototroch are uni- form in size, but during the differentiation of the trochophore, there appear two bands of longer cilia—one at the upper edge which beat actively, and a narrower one on the lower edge which hang down and move more slowly and indefinitely. Between these two rows the shorter cilia are retained. The long flagella, borne entirely. by the rosette, are about 20 in number and when the embryo is actively swimming are carried stretched out in front and bunched closely together, quite as in a pilidium larva. When the animal is at rest the flagella wave about slowly. In the trocho- Artificial Parthenogenesis in Thalassema Mellita 115 phore of about twenty-two hours they are replaced by very short inactive cilia” (p. 190). Compare my Figs. 9 and to. All degrees of abnormality, however, may be seen in the ciliation of the parthenogenetic embryos and larvae. In some cases the apical flagella are entirely wanting or reduced in number, while the cilia on the prototroch may as often fail to form a complete band and appear in irregular clumps or patches, sometimes occurring only on one side; or they may lose every trace of a band- like arrangement and cover more or less uniformly the whole pre- trochal region of the animal (Figs. 11 and 12). In regard to their activity, they may beat in the same way and with about the same vigor as they do in the normal embryo and larva, producing the characteristic spiral movement of the latter. Associated with abnormalities in form and distribution, their movements may also depart widely from those that are typical and lack all apparent coordination. ‘he enfeebling of the stroke may be so pronounced as to render the cilia incapable of causing any bodily movement of the embryo. I gained the impression that functional derange- ment of the cilia was always correlated with morphological abnormalities of the embryo, and that the farther the embryo departed from the normal in structure, the more erratic were the movements of the cilia. : As Torrey has described, the embryo is at first entirely separated from the egg membrane, but later through the elevation of the surface in these regions, the rosette and primary prototroch become pressed against the membrane which is then punctured by the cilia (p. 187). In the parthenogenetic embryos, especially when the relations are more nearly normal, the same perforation of the membrane by the cilia takes place. “This, however, is not always the case. Sometimes the perivitelline space is so greatly enlarged and the membrane in consequence so far removed from the sur- face of the embryo, that the cilia only touch it at their outer ends, if they reach it at all; or, again, although the membrane may lie close to the embryo, the cilia fail to puncture it and are bent down, being held in this condition until they are released by the rupturing of the membrane (Fig. 13). As already stated, the movements of the cilia are usually quite 116 Geo rge Lefevre normal in character and their stroke apparently as vigorous as in embryos and larve produced from fertilized eggs, yet I have never observed a single instance where the Goche piers rose to the sur- face of the water. They invariably swim close to the bottom, if not in actual contact with the dish. This peculiarity in behavior has been observed with remarkable constancy by most experi- menters on artificial parthenogenesis, although Delage in the starfish, after treatment with CO,, and Loeb in the sea-urchin, by the use of hiscombination of methods, has succeeded in obtaining larve that rise in the usual manner to the surface. In Thalas- sema, however, none of the methods which I have employed has produced trochophores that are free from the abnormality in question. Unless the parthenogenetic larve were isolated, they would rarely live over 36 hours, for the eggs which had failed to develop were soon attacked by bacteria, much sooner in fact than the con- trol eggs which had not been exposed to the solutions, and con- taminated the dishes to such a degree that the larve speedily succumbed. It was, moreover, extremely difficult to separate the larva, as they did not rise to the surface, and had to be picked out individually from among the non-developing eggs with a very fine pipette. The trochophores, however, that were isolated were kept in separate dishes and the water frequently changed, yet in no case did I succeed in rearing them longer than a little over three days. The oldest trochophore raised lived about 80 hours. There was no appreciable advance in development, however, beyond the second day, as the larve seemed to enter upon a station- ary period at that time. As they grew older, their movements became more and more feeble and irregular, and they gradually disintegrated, the body finally rupturing and the protoplasm flowing out. 5 Abnormalities aa Cleavage and Behavior Although in experiments where the optimum conditions were present the great majority of the dividing eggs segmented in a normal manner, an endless variety of irregular cleavages were encountered, especially in cases where the strength of the solution Artificial Parthenogenesis in T halassema Mellita I Py employed or the duration of immersion was not such as to yield the best results. Many of these abnormalities are similar to those which have been described by others (cj., especially Wilson, ’or) and need not be spoken of in detail here. ‘That such abnormal cleavages lead to the formation of ciliated structures is clearly indicated by the fact that the percentage of all eggs dividing, both normally and abnormally, nearly agrees with the percentage of swimming embryos which are later found in the same culture. Many such counts were made, and the correspondence in per- centage was found to be remarkably close. I have, moreover, frequently isolated abnormally segmenting eggs and_ directly observed them to develop into ciliated, cellular structures. It is extremely doubtful, however, whether eggs that divide irregularly ever undergo a later regulation and produce normal embryos, for all the ciliated structures which were raised from isolated, abnormally segmenting eggs departed more or less widely from normal forms. Previous experimenters on artificial parthenogenesis of annelids have observed that ciliated structures may arise from unsegmented eggs by a process of progressive cytoplasmic differentiation whicn takes place in the entire absence of cleavage [Loeb (’o1) and Lillie (02) in Cheetopterus, Fischer (’02, 03) in Amphitrite and Nereis, Treadwell (’02) in Podarke, and Scott (’06) in Amphitrite}. In Thalassema, on the other hand, it is undoubtedly true that eggs which do not divide never undergo further differentiation and never form ciliated structures (c7. Bullot, ’o4). Without excep- tion every swimming embryo observed possessed a well-marked cellular structure, and eggs which had failed to segment after ‘residence in acidulated water, when afterwards isolated, in no instance gave rise to a differentiated body. Certain common types of abnormal cleavage were constantly met with, e. g., every degree of inequality in size of the blastomeres formed by the first and second cleavages was found, while eggs were frequently observed to fall at once into three or four cells at the first division. ‘The trefoil stage, in fact, seemed to be character- istic of certain solutions, in which a great preponderance of eggs exhibiting this abnormality was noted. An unequal division of the eggs at the first cleavage was occasionally followed by division 118 George Lefevre of the larger blastomere in a plane parallel to the first, so that three nearly equal cells were formed in a row, as shown in Fig. 14. One of the commonest abnormalities, and one which might be observed at any stage of the development, was due to the failure of one or more cells to undergo cytoplasmic cleavage when the nucleus divided, and, as a result of this condition, large multi- nucleated cells might be found in embryos of all ages. Some examples of such cases as these will be referred to below. Owing to departures from the normal type of cleavage, young blastulz are occasionally produced in which the cleavage cavity is not closed; this results in the formation of a cylindrical embryo which is open at both ends. An optical section of this modified blastula, showing the cells radially arranged around the cavity, is drawn in Fig. 15. Although grastrulation usually takes place in parthenogenetic development in quite a normal manner, disturbances in the pro- cess do occur and produce a great variety of pathological embryos. Many degrees of incompleteness in gastrulation may be seen, the extreme case being one in which only a very few entoblastic cells sink into the cleavage cavity. Such embryos, although they may be ciliated and externally resemble trochophores, are found upon sectioning to be merely large, hollow, spheroidal bodies in which the enteron is represented by a few scattered entomeres. Ameeboid activities, which have been observed so commonly in the parthenogenetic eggs of echinoderms and worms, are very rare in the eggs of Thalassema, and when they occur at all, the formation of pseudopodia is not extensive and takes place very sluggishly indeed; Fig. 16 represents about as pronounced a case as I have seen in my experiments. Fusion phenomena, which were so conspicuous in the experi- ments of Loeb (’o1) and especially in those of Lillie (’02), after treatment of the unfertilized eggs of Chztopterus with KCl and CaCl, solutions, were entirely absent in Thalassema. In Cheetop- terus, eggs may fuse into masses which show certain differentiations and form giant, ciliated structures and double, triple, quadruple, etc., monsters. Lillie, for example, has described fusion masses, into the composition of which about 100 eggs had entered, and Artificial Parthenogenesits in Thalassema Mellita 11g ‘ these through amceboid activity gave rise to a “veritable wilder- ness of pseudopodia.” Such fusions are rendered possible by the destruction of the egg membrane, after treatment with salt solu- tions, especially solutions of CaCl,, and, with the heightening of amoeboid movements, the naked masses of protoplasm show a strong tendency to adhere whenever they come in contact. The acid solutions, however, which I have used, never cause disap- pearance of the membrane in Thalassema eggs, and this may explain, in part at least, why fusion phenomena are entirely absent. The eggs in which a membrane formation has been called forth never adhere, but during the later stages of disintegration they may become attached by their membranes to form clusters like frog spawn; here, however, there is no protoplasmic fusion. VI OBSERVATIONS ON THE PRESERVED MATERIAL The internal changes involved in the artificial parthenogenesis of Thalassema have been followed as closely as possible, and by means of sections and total preparations many stages have been examined throughout the entire period of development covered by the material. In the study of the preserved material, however, serious dif- ficulty is encountered in the attempt to determine the proper sequence of changes taking place in the eggs and embryos, since, as has been pointed out, the rate of development varies so greatly in different eggs, even in the same culture, that a large number of stages may be represented in each sample lot preserved. Fur- thermore, the living egg can only be used as a control to a limited extent, as it is too opaque to allow of any detailed examination of internal changes, and an attempt to reconstruct successive stages of development from a study of preserved material alone is attended with more or less unsatisfactory results. Nevertheless, with respect to most points, this uncertainty is reduced to a mini- mum, and in cases where the parthenogenetic embryos coincide in structure with those produced from fertilized eggs, it is fairly safe to take for granted that the observed stages of development have followed in their normal sequence. It becomes much grezter, 120 George Lefevre however, when we are dealing with unusual or abnormal conditions for which no control can be had from living eggs, and here con- clusions can only possess a greater or less degree of probability according to the circumstances of the particular case. t The Oécyte and the Maturation Divisions The full-grown odcytes, when removed from the tubes, possess a large nucleus placed somewhat eccentrically. The cytoplasm is heavily charged with yolk spheres, which stain a deep black with hematoxylin and are present throughout the entire cell body, except for a peripheral and a perinuclear layer of yolk-free cytoplasm (Fig. 2). The nucleus is filled with a granular reticu- lum which takes the plasma but not the chromatin stains at this time. The large vacuolated nucleolus persists until the breaking down of the germinal vesicle, when it gradually dissolves and dis- appears. ‘The reduced number of chromosomes (12) lie scattered in the nucleus, usually near the membrane, and appear as coiled or twisted granular rods which in some cases are clearly seen to be double or to have the form of rings, or loops, or crosses. Asa rule, they stain quite black with haematoxylin and stand out con- spicuously in the lighter, granular reticulum (Fig. 2). During the early prophases of the first maturation mitosis, the chromo- somes undergo a concentration, take up the stain more intensely, and assume the variety of tetrad forms which have been minutely described by Griffin (’99). In this condition they approach the developing asters and finally enter the spindle. The formation of the membrane and the filling out of the egg to the spherical condition, after treatment with acidulated water, have already been described. It has also been mentioned that in many cases the nucleus migrates bodily to the animal pole, after the eggs have been returned to normal sea-water (Fig. 3), although this change in position does not always occur. The development of the amphiaster for the first maturation division can only be faintly observed in the living egg, but the process may be followed in great detail from an examination of sections. The earliest material I have in which the asters are unquestionably present was fixed 15 minutes after the eggs were exposed to an Artificial Parthenogenesis in Thalassema Mellita TZX acid solution. Since they first become evident in normal eggs about 3 minutes after entrance of the sperm, their appearance 1s much retarded in the parthenogenetic eggs. ‘They are situated very close to the nuclear wall, and each contains at its center a deeply stained centrosome or centriole (Fig. 17). As soon as the two asters can be discovered at all, they are invariably found some distance apart, and, like Griffin, I have been able to find no evi- dence that they arise by division of a single aster whose products afterwards diverge. In some cases Griffin (op. cit.,p. 590) observed a bipartite condition of the centrosomes even at this.early stage, but in my acid treated eggs I have not found them divided until a somewhat later stage. I have also been unable to discover in eggs which have been exposed to acid solutions any trace of the “secondary asters” which Griffin (p. 590) has described as being present in both fertilized and unfertilized eggs, but disappearing in the former in about three minutes after entrance of the sperm. The rays of the asters are at first few in number and excessively delicate, but, as they develop, they appear to grow into the nuclear mem- brane, flattening it and throwing it into folds and wrinkles, until they finally rupture it and enter the nuclear area (Fig. 17). As the formation of the amphiaster for the first polar mitosis in nor- mal maturation has been carefully described and figured by Grif- fin, and as these changes are usually the same in every detail in unfertilized eggs which have been subjected to the action of acid solutions, an elaborate account is not necessary here. During the continued invasion of the nucleus by the astral tays, the nuclear membrane gradually dissolves away, and the granular reticulum, now staining a dark blue with hzmatoxylin, is left in the cytoplasm where it afterwards disappears. Some of the rays, which are thicker and stain more deeply than the rest, appear to attach themselves to the chromosomes, the so-called “ trac- tion fibers.” The centrosomes are now distinctly double, as may be seen in Figs. 18 and 19, which show stages in the formation of the amphiaster. ‘The definitive spindle (Fig. 20) carrying the chromo- somes in the equatorial plate, now swings into a radial position in such a way as to bring the outer aster close to the surface of the egg. The outer rays of this aster shorten, while those situated more 122 George Lefevre laterally curve backward and cross the corresponding rays of the inner aster in the plane of the equator of the spindle, as in the preceding figure. “The chromosomes (tetrads), which show during the prophases the varied forms described by Griffin, undergo a concentration, and when seen in the equatorial plate present a more or less uniform type consisting of a cross, with a pair of thick broad arms in the equatorial plane and a pair of narrower perpen- dicular arms. A split is frequently seen in the latter, and some- times the transverse arms also appear to be divided in the middle by a faint line. According to Griffin, the split in the transverse arms of the cross corresponds to the original, longitudinal division of the spireme segment, and the first maturation division is, there- fore, equational or longitudinal. ‘This may be the case, but after a careful examination of the tetrads in both unfertilized and fertil- ized eggs, I am unable to find any definite basis for a determina- tion of the character of the first division. The variability of the form of the crosses renders the identification of either the equatorial or the polar arms with the longitudinal axis of the primary rods extremely uncertain. At the first mitosis, the crosses are drawn out into ellipses, which then divide, the dyads passing to the poles either as V’s or double rods. The separation of the V’s at the apex takes place at the second maturation division, which Griffin interprets asareducing division. Theouter pair of centrosomes and group of dyads become contained in a small projection or knob of clear, yolk-free cytoplasm, which by constriction around its base is eventually cut off as the first polar body (Figs. 21 to 23). Before the close of the first mitosis, however, the two inner centrosomes diverge in a direction nearly transverse to the axis of the original spindle; a minute spindle appears between the two, and around each centrosome a new system of delicate rays is formed (Fig. 21). ‘The dyads left in the egg lie at first on the outer side of the spindle which now rapidly elongates and rotates through about go° of arc until it assumes a radial position with the outer aster immediately under the point on the surface where the first polar body was cut off (Fig. 22). The dyads, in the form of double rods, arrange themselves in the equatorial plate with the long axis across the spindle, as seen in the preceding figure, and Artificial Parthenogenesis in Thalassema Mellita 123 during the anaphase the halves of the dyads move to the poles (Fig. 23). Acomparison of Figs. 21 to 23 with Griffin’s Figs. 18, 21, 25, will indicate their close identity. ‘The second polar body is constricted off in the same manner as the first and con- tains twelve single chromosomes, while the same number is found in the egg. I have occasionally observed a doubling of the inner centrosome of the second spindle, as described by Grifhn; Fig. 23 shows a case in point. ‘The twelve chromosomes left in the egg (Fig. 24) pass at once into vesicles which at first only par- tially fuse to form the egg nucleus, while every trace of the inner centrosome and its accompanying rays soon disappears (Fig. 25). As described by Conn (86) and by Griffin (99), the first polar body invariably divides, and, as the latter has determined, it does so by a complete mitosis. ‘The same division has been repeatedly observed in the parthenogenetic eggs, as shown in Fig. 4, which is drawn from a living egg, and also in Figs. 21, 23, 25 and 45 where the minute spindles and rudimentary asters with their centrosomes are clearly seen in section. In the fertilized egg, the second polar body never exhibits an attempt at division (Griffin, p. 615), but observations on living eggs, which have been exposed to acid solutions, show that all three polar bodies continue to divide until a cluster of miniature cells have been formed, as already described (Fig. 1). Although I cannot state that all of these divisions take place mitotically, I have repeatedly observed distinct indications of spindles in sec- tions of the groups of polar cells; Fig. 62 shows, for example, a spindle in two of the five cells present in the section. Up to this point the history of the parthenogenetic egg 1s identi- cal with that of the egg fertilized by sperm, except for the absence in the former of the spermatozo6n and its aster, and all of the details of the process of maturation which Griffin has so carefully worked out may be easily verified on the eggs of my experiments. From now on, however, a special description of the changes taking place in the parthenogenetic eggs becomes necessary, for the absence of the sperm nucleus and asters radically alters certain subsequent events, and a special account, quite different from that of the usual phenomena, must be given of the origin of the cleavage nucleus and amphiaster. 124 George Lefevre 2 Origin of the Cleavage Nucleus and Amphiaster After extrusion of the second polar body, the egg nucleus moves toward the center of the egg and soon loses its irregular con- tour by the complete fusion of the chromosomal vesicles which entered into its composition. ‘There now follows a long resting period, and during this pause, which may last for from one to two hours, or even longer, the nucleus increases considerably in size (Fig. 26). . The first indication of the preparation for cleavage is seen in the simultaneous appearance of two delicate asters which are situated opposite each other and immediately outside the nuclear membrane (Fig. 27). They lie in a plane approximately perpen- dicular to the axis of the egg, and in the center of each aster a dis- tinct centrosome 1s visible. ‘There is not the slightest evidence that the cleavage asters, for such these are destined to be, are derived by division of a single primary one; when they are first seen they lie at a cousdeea BIE distance from each other, usually on opposite sidesof the nucleus. ‘There is, furthermore, no doub- ling of the centrosomes at this time, and no intermediate stages are found which would indicate a division and separation of the asters. Fig. 27 illustrates the almost invariable condition of the egg nucleus when the asters are first discovered. My observations agree with those of Wilson (01) on the magnesium eggs of Tox- opneustes in that they showclearly that the cleavage centrosomes lie outside the nuclear membrane but immediately upon it, although Wilson is strongly inclined to believe, from the evidence presented by his sections, “that the cleavage amphiaster e1ses by division of the single center” (p. 564). He first observed a single centrosome, surrounded by a conspicuous aster, lying outside the intact nuclear membrane at one pole of the nucleus; all inter- mediate conditions were found between this stage, through a monaster stage, to the complete establishment of the dicentric figure, and, while recognizing the possibility of being misled by the nature of the material, he felt justified in regarding the amphiaster as the product of division of the primary monaster. As the evidence on which he bases his conclusion for Toxopneustes 1s not presented Artificial Parthenogenesis in Thalassema Mellita 125 by Thalassema, I have not been able to arrive at a similar inter- pretation for the eggs which I have studied. It should also be mentioned in this connection that Hunter (’04) gives an account of the origin of the cleavage asters for the parthenogenetic eggs of Arbacia which is in harmony with Wilson’s view. In a brief note he merely makes the statement that “a small aster with central dark body (parthenocentrosome) appears in contact with exte- rior of nuclear membrane. ‘This divides to form the amphiaster”’ (p. 214). The question of the orgin of the cleavage centrosomes of Thalas- sema will be referred to in a later section of this paper, but | have little doubt that they are new formations. ‘There is not the least indication of a continuity between the egg center and the cleavage centers, the former totally disappearing at the close of maturation, and I am not inclined to accept the view that the egg center persists, without at least some evidence of the fact. 3 The Cleavage Stages The completion of the first cleavage takes place entirely in accordance with Grifhin’s description of the process in the fertilized egg, except that only 12 chromosomes (the reduced number) are derived from the spireme of the nucleus. After the stage drawn in Fig. 27, in which the nucleus is seen somewhat drawn out toward the asters, the nucleus elongates still further and becomes fusiform. As the membrane disappears at the poles, the inner rays of the asters invade the nuclear area and form the spindle, some of the fibers appearing coarser and wavy and attached to the chromosomes; these are the so-called “traction fibers.””. With the development of the rays and spindle, a centrosphere surrounding each centrosome appears, and, as it rapidly increases in size, it becomes reticulated (Fig. 28). At the time of metaphase the centrosomes divide and pass toward the outer side of the spheres. Twelve rod-like chromosomes are derived from the spireme thread, which is left in the equatorial plane of the spindle after the disappearance of the nuclear membrane. ‘Their number, which can be readily determined in sections of the equatorial 126 George Lefevre plate, is shown in Fig. 29. The chromosomes divide typically, pass to the poles, and in the telophase are converted into vesicles from which the daughter nuclei arise by fusion (Fig. 30). The division of the egg into two equal blastomeres now takes place. During the anaphase, the centrosomes at each pole begin to sepa- rate, and soon delicate radiations begin to arise around each asa new center before the original rays have disappeared (Fig. 30). The difficulties in the way of tracing the cell lineage of the parthenogenetic eggs are so great that a detailed comparison with the normal cleavage and a determination of the origin and fate of. the constituent cells of the embryo are rendeed practically im- possible. Owing to irregularity in the rate of division and the possible presence of greater or less abnormalities at all times in development, the arranging of stages that appear to be normal in their proper sequence can only be attended at best by uncertain results. I shall, therefore, not attempt a continuous account of the development as far as I have been able to follow it, as any reconstruction of stages would be quite arbitrary and would rest almost entirely on the identifiction of individual embryos with normal ones of a corresponding degree of differentiation. Al- though in favorable experiments, where optimum solutions were used, I have met with no such “carnival of development” as many other experimenters in artificial parthenogenesis have observed, and although the proportion of embryos that are either normal or nearly so is very large, nevertheless abnormalities of all kinds are of not infrequent occurrence, and the determination in a given case as to whether an embryo is normal or not is by no means easy. But in spite of these difficulties, | am convinced that the partheno- genetic eggs in a great many cases undergo a normal development and exhibit the usual processes of differentiation leading up to the formation of the swimming larva. ‘The only differences in such cases between the organisms which are produced parthenogen- etically and those arising from fertilized eggs are to be found in the numerical relations of the chromosomes, as I shall point out be- yond, in the rate of development and in the fact that the partheno- genetic larva do not rise to the surface of the water when they begin to swim. Artificial Parthenogenesis in Thalassema Mellita 127 The establishment of the second spindles in quite a normal man- ner may be seen in Figs. 31, 32 and 33, and in the latter the cen- trosomes of one of the cells are clearly double. ‘This bipartite condition of the centrosomes has been observed repeatedly during mitosis at all stages of the development (Figs. 35, 36, 40), but I have not been able to demonstrate their persistence through the resting period from one generation of cells to the next. After the nuclei return to the resting condition, every trace of asters and centrosomes disappears, and their presence is not revealed again until the early prophase of the next mitosis. I am inclined to be- lieve that, although the centrosomes may divide and diasters arise at the close of a mitosis, these totally disappear, and that the asters for the next cleavage are new formations brought into evidence when the activities of the cell which bring about division are renewed. Fig. 34 is drawn from a four-cell stage during the resting period, and Fig. 35 from an eight-cell stage, showing four of the cells in mitosis; the early formation of the cleavage cavity is seen in the latter figure. Later stages in the formation of the blastula are illustrated in Figs. 36 to 40, in each of which one or more cells are taken in mitosis. Fig. 39 is probably somewhat abnormal, as it shows an unusually large blastoccel, a not infrequent condition. In Fig. 36, the section has passed through the equatorial plate in the cell indicated at a, and here the twelve chromosomes can be easily counted. As will be emphasized beyond, the reduced num- ber of chromosomes has been repeatedly determined in cells of blastulz and gastrule (Fig. 41, a) and the evidence 1s, therefore, perfectly clear that, except under certain abnormal conditions which will be referred to, a restoration of the normal number of chromosomes does not occur. 4. Gastrulation and the Formation of the Trochophore Gastrulation, which according to Torrey (’03) takes place about seven hours after fertilization, is considerably delayed in parthenogenetic development and occurs at from eight to twelve hours after exposure to acid solutions. ‘The process is of the embolic type, and the insinking of the entoblastic plate can be 128 George Lefevre observed in the living embryo as well as followed from an exam- ination of sections. It seems to be perfectly normal in charac- ter in a great many parthenogenetic embryos, although I have not been able to compare the origin of the entoblastic plate with its normal cell lineage. Sections, however, show a close corre- spondence with Torrey’s description. ‘The lower pole of the blas- tula flattens, while the upper becomes somewhat arched, and the entoblastic cells elongate and sink bodily into the cleavage cavity which soon becomes nearly filled withthem. Early stages of gas- trulation are illustrated by the sections drawn in Figs. 40 and 41, which closely correspond with Torrey’s Fig. 10 (p. 232). After sinking in, the entomeres multiply rapidly and form a rounded mass which withdraws a little from the body wall. ‘The enteron now consists of a thick epithelium surrounding a small lenticular cavity which later becomes greatly enlarged, while the cells com- posing its walls are densely filled with yolk spheres. Fig. 42 shows the entoblastic mass with its early cavity, and, although the section does not pass through the blastopore, it is very similar to Torrey’s Fig. 6A (p. 207). The blastopore, an elongated slit (Fig. 9), shifts from the lower pole, where it first appears, to the future ventral side, until itcomes to lie close under the prototroch. Although I have not followed its history in all particulars, the changes which it undergoes in the parthenogenetic embryos evidently agree with the account given by Torrey, who describes the enlargement of the anterior end of the slit-like opening and the persistence of this portion to form the mouth, while the posterior portion closes by approximation of its sides. After gastrulation has taken place, an invagination of the ecto- dermal cells around the blastopore occursand gives rise to the cesoph- agus (Fig. 43), the blind end of which abuts against the closed enteron. In the formation of the cesophageal invagination, three large cells, the cesophagoblasts, are distinguished from the rest by their size and come to occupy a definite position in the wall of the cesophagus, as described by Torrey (p. 205). The continuous history of these cells has not been followed in my material, but that the same specialized cells are present in the parthenogenetic Artificial Parthenogenesis in IThalassema Mellita 129 embryos is clear from an examination of Figs. 41 and 43 (cells designated at oes) which should be compared with Torrey’s Figs. 10D and 6B, respectively. The enteric cavity becomes secondarily divided into stomach and intestine by a partition, consisting of a double row of cells, which grows across from the dorsal wall and completely divides the stomach from the intestine except on the ventral side where an opening persists. The division of the archenteron into two _ cavities by this septum is well shown in Fig. 44, although the sec- tion is a horizontal one and does not show the cesophagus and the communication between the stomach and the intestine. The ectodermal cesophagus secondarily acquires an opening into the stomach, but the anus, which is not formed until a very late stage in the normal development, has never been observed in the par- thenogenetic larve. ; Fig. 44 is drawn from a CO, trochophore, which was killed 31 hours after the eggs had been treated with carbonated water. Although larve were raised for a considerably longer time than this, differentiation rarely proceeded beyond the condition indi- cated in the preceding figure. In the trochophore here repre- sented, in addition to the digestive tract which is differentiated into mouth, cesophagus, stomach and intestine, the prototroch is present, and the high columnar cells bearing the apical flagella are distinctly shown. A_ few cells of the larval mesenchyme (ectomesoblast), stippled in this and the two preceding figures, are seen scattered between the body wall and gut, to both of which they are adhering. I have not observed the definite ventral neural ciliated region described by Torrey, but occasionally more or less irregular patches of cilia have been found in this portion of the larva, as they have been, in fact, in other regions as well. 5 Rudimentary Cells The remarkable rudimentary cells whose origin and fate Torrey has so accurately described in Thalassema, are also recognized in the parthenogenetic embryos, although I hardly think they are as numerous here as in embryos raised from fertilized eggs. “These 130 George Lefevre cells arise, according to Torrey, early in the cleavage, and, wander- ing into the blastoccel, are seen lying on the cells of the entoblast into which they soon sink (Figs. 40, 41, rc). Here they become quickly absorbed and disappear. As long as they are present they may be easily recognized by their smail size and the con- tracted condition of their chromatin. In Fig. 40 is also seen a small cell pb, which is probably a polar body that has passed into the cleavage cavity; its cytoplasm has a faintly radiate appearance suggesting mitotic activity. 6 Abnormal Maturation Phenomena In an earlier part of this paper it has been stated that in some instances the unfertilized eggs of Thalassema, after an exposure to acid solutions, failed to extrude the second polar body, and in still other cases that neither body was formed. It was frequently observed, however, that such eggs segment and produce embryos and larve indistinguishable from those arising from eggs that have thrown off both polar bodies. These abnormalities were not associated with special solutions or external conditions of the experiment, but they appeared at any time, and even with the optimum solutions. It was of importance to determine the internal phenomena present in eggs showing these abnormalities of maturation, and an examination of sections from material preserved when the unusual conditions were met with has brought to light several interesting facts. In many of these cases, it is undoubtedly true that either the first or the second polar mitosis, or possibly both in some instances, may take place entirely inside the egg and without accompanying cytoplasmic cleavage, and in this submerged con- dition give rise to resting nuclei. a Absence of the Second Polar Body In the eggs which extrude only the first polar body, the second spindle fails to assume its usual position, and instead of rotating in such a way as to bring one pole to the surface, it sinks down 3 Cf. the observations of King (06) on the retention of one or both polar bodies in unfertilized starfish eggs after compression. hoa Artificial Parthenogenesis in Thalassema Mellita 131 until it comes to lie much nearer the center of the egg. In this unusual situation, it completes the mitosis and gives rise, without much doubt, to two resting nuclei. Fig. 45 shows a fortunate section in which the first polar body is present and the second spindle lies deep within the egg; the chromosomes which are single rods are ina late anaphase and have already reached the poles. ‘The spindle, which is cut throughout its entire length, as both centro- somes are present in the section, 1s perfectly normal in all re- spects save its position and unusual length. ‘The objection can- not be made that this might be the cleavage spindle, as the two are entirely different in appearance. Although the proper sequence of stages cannot be determined with absolute certainty from a study of the preserved material, there can be little doubt that this particular mitosis, and many others like it which I have seen, is taking place inside the egg. All of the eggs on which I have based this conclusion belong to a single experiment, No. 29 of my notes, in which not an instance of the formation of the second polar body was observed while the eggs were alive; sample lots from this experiment were killed at intervals during the maturation period, and I have sectioned several stages of the series. My attention being attracted to the peculiarities of maturation at the time of the experiments, I was especially careful to make detailed notes of each case and to look over a large number of living eggs which showed unusual conditions. From the data given in my notes [| feel confident that none of the eggs of experiment No. 29 extruded the second polar body, and it is quite unlikely that the spindle shown in Fig. 45, which is one of many similar cases found in the same material, would have later assumed a normal position and thrown off the polar body. Although I have not found all the intermediate stages, numerous cases like Fig. 46 occur in the same series of eggs, and it would seem highly probable that the two nuclei present are the result of the submerged mitosis and repre- sent the egg nucleus and the nucleus of the second polar body. In spite of the absence of indisputable proof, there can be little doubt that the two nuclei fuse to form a cleavage nucleus, since in later stages of the series many eggs are found which contain a single large nucleus accompanied by two asters, while still other eggs 132 George Lefevre show different stages of the first cleavage mitosis. If my inter- pretation of these conditions is correct, the case is identical with the rarer type of parthenogenesis described by Brauer (’93) for Artemia, where the chromosomes of the second polar body are retained by the egg and give rise to a reticular nucleus, as earlier described by Boveri (87) in Ascaris, which acts like a sperm nucleus and conjugates with the nucleus of the egg. Brauer’s observa- tions apparently confirmed Boveri’s conception that “ Partheno- genesis is the result of fertilization by the second polar body” (J.c., ‘p. 73). Of course, if the fusion of the two nuclei takes place, one would expect to find 24 chromosomes in the equatorial plate of the first cleavage spindle, but unfortunately I have not been able to obtain a favorable section in which the number of chromosomes could be accurately determined, but there are clear indications that it is greater than 12. The difficulty in establishing this point is not surprising, as the material from this experiment, which was preserved for the later stages, was quite limited in amount. b Absence of both Polar Bodies A more frequent abnormality is found in the failure to extrude both polar bodies, yet eggs which show this peculiarity may develop into swimming larve. ‘The condition was observed in a num- ber of experiments when after careful search not an egg was seen to give off the polar bodies, but the percentage of developing eggs in such cases was as high as the average. Fig. 47 shows a sub- merged first maturation mitosis in a late anaphase; the centro- somes are clearly seen to be double and the chromosomes are in the form of dyads and not single rods. It is exceedingly difhcult to reconstruct the successive stages in the internal maturation of such eggs, and here again positive proof cannot be furnished. Two resting nuclei are undoubtedly formed, as they occur in many eggs, but I am inclined to believe that these usually fuse at once to form a cleavage nucleus without the occurrence of a second maturation mitosis. Although I have not succeeded in finding a complete series of stages, there is some evidence, however, that in rarer cases the two nuclei just referred to may divide again mitot- ically, with the resulting formation of four smaller nuclei. In Artificial Parthenogenesis in Thalassema Mellita eG, Fig. 48 is shown an egg in which one of the two nuclei present 1s accompanied by two minute asters, while a single aster is seen lying close to the other nucleus, evidently in preparation for a sec- ond division. Fig. 49 shows four small nuclei which have pos- sibly arisen through two maturation divisions occurring internally. Of course, such a condition as this might be explained as the result of the formation of a tetraster without subsequent cytoplasmic division, but polyasters are entirely absent in the eggs from which these cases are taken, and if this were the correct interpretation, some evidence of the occurrence of multipolar mitosis would cer-* tainly be present. It is impossible to determine whether the four nuclei fuse, or not, to form a cleavage nucleus, yet that they do conjugate is indicated by the fact that in later stages of the same material where the cleavage amphiaster is present, a careful search fails to disclose accessory nuclei outside in the cytoplasm. In a few sections of the equatorial plate of the first cleavage spindle, found in the same series of eggs, the number of chromosomes is clearly more than twelve, although I have never counted an exact multiple of that number; in Fig. 50 twenty-three chromosomes are present in the section, and this would seem to prove that at least two of the nuclei had united. ‘ In a number of experiments in which it was noticed that none of the eggs extruded polar bodies, besides the cases of internal ma- turation already described, a still more unusual condition, which may be appropriately referred to in this connection, is apparently present in certain eggs. ‘These are cases in which the first matura- tion spindle is formed in the ordinary manner, but instead of rotating into a radial position it becomes elongated and placed symmetrically across the center of the egg. These large spindles (Fig. 51), clearly showing the crosses of the tetrads and the double centrosomes, predominate in the eggs of certain experiments, and since later stages of the same material show many eggs divided into two cells, it can hardly be doubted that they give rise to an equal or nearly equal segmentation of the egg, the products of which have, therefore, the value of odcytes of the second order. Fig. 52 shows an anaphase of such an egg, in which the distinct dyads are seen passing to the poles of the spindle. I am unable to 134 George Lefevre state, however, whether the next cleavage behaves like the second maturation division or not, but it would be of interest to know if the dyads reappear at that time, as might be expected. 7 Abnormal Mitoses It is not my intention to describe in detail the endless variety of abnormalities of mitosis that have been encountered in the study of the parthenogenetic eggs of Thalassema. Most of the unusual conditions which I have found are quite similar to those which have been described by other observers, especially by R. Hertwig (’96), Morgan (’96, ’9g, ’00), and Wilson (’or1), in the wifertitized eggs of echinoderms after treatment with salt solutions and other agents. It may be well, however, to refer briefly to the more characteristic abnormal forms. a Multipolar Mitoses Of these, the formation of polyasters, with resulting multipolar mitoses, was perhaps of the most frequent occurrence and was ob- served at all stages of development from the first cleavage onward. It has already been stated that the two cleavage asters, when the parthenogenetic eggs develop without abnormalities, appear simultaneously on the nuclear membrane, and give rise to the usual dicentric figure. When three or more asters appear, instead of two, they also seem to arise in situ, lying close to the nuclear membrane and at quite a distance from each other. I have never observed a doubling of the centrosomes in these cases of multipolar mitosis,or the least indication of division of the asters. Fig. 53 shows a typical case; the egg nucleus, which is here rather larger than usual, is accompanied by four small asters, showing central bodies, and the membrane is giving way in front of three of them. As the spindles form, the nuclear area is usually drawn out at the points where the asters lie, as seen in one of the cells of Fig. 54, and with the dissolution of the membrane, forms like these give rise to multipolar spindles. Such eggs may divide at once into a corresponding number of blastomeres, as I have frequently observed living eggs in which triasters and tetrasters eer a mt Artificial Parthenogenesis in Thalassema Mellita E35 were faintly visible to fall into three or four cells, respectively, at the first cleavage. Cytoplasmic division, however, does not by any means always follow a multipolar mitosis which may occur repeatedly in one and the same cell. That this is true 1s clearly proven by the large number of chromosomes often present in such cells. Fig. 58 shows an abnormal embryo in which one cell has evidently failed to undergo cytoplasmic cleavage and in which a multipolar mitosis is taking place. A similar condition is seen in Fig. 55 which presents a maze of spindles and asters in the unseg- mented egg. Cytasters in the acid treated eggs of Thalassema are of rare occurrence, and only occasionally does one find an aster which 1s not associated with nuclear material. In fact, I have never been able to thoroughly satisfy myself that I have observed a true cytaster, but in cases like Fig. 55, the small asters lying near the periphery of the egg, some of which show a central granule, may possibly be of this nature. At the close of a multipolar mitosis, the numerous chromosomes are either gathered into a single large nucleus, which is usually polymorphic in character (Fig. 56), or apparently many separate smaller nuclei may be formed which increase in size during the resting period (Fig. 57). Frequently the latter are lobed or con- stricted, as if dividing amitotically (Fig. 57), and it is also probable that they fuse at times into a single large nuclear area. [am inclined to believe that such eggs rarely segment, as they are especially numerous in dishes in which the developing eggs have reached late blastula and gastrula stages. 6 Giant Bipolar Figures In more or less abnormal embryos, I have quite often found large cells, in which segmentation had evidently not kept pace with nuclear division, that were characterized by the presence of a single giant spindle bearing an enormous number of chromo- somes. Fig. 59 illustrates a case in point. In the cell marked a, the spindle is abnormally large, while the equatorial plate is densely packed with small chromosomes which are greatly in excess of the usual number. In cell } of the same figure, a similar spindle, but wqoe George Lejevre smaller and containing fewer chromosomes, has been cut trans- versely in the middle. It is difficult to determine how figures like these have arisen, but since the number of chromosomes is greatly increased, nuclear division has undoubtedly occurred re- peatedly without accompanying cleavage of the cytoplasm. ‘Their origin could be accounted for by supposing that, after several nuclear divisions have taken place, probably through multipolar mitosis but without cleavage of the cytoplasm, all of the chromo- somes have been gathered into a single large nucleus which, at the next mitosis, is converted into a bipolar figure by the appearance of only two asters in connection with it. c Monasters Mention should be made of another common type of abnormal forms, the striking single radiate systems or monasters which have been described by previous observers, notably Hertwig (96), Morgan (’00) and Wilson (’o1). In Thalassema the monasters are only found in unsegmented eggs, the closely set rays forming a beautiful corona around the nucleus. I have never observed the monaster to resolve itself into a bipolar figure, nor to produce a segmentation, but the same alter- nating phases of activity, involving the rhythmic disappearance and reappearance of the rays and successive division of the chromo- somes, as have been described by Wilson (’o1, p. 546) for Toxop- neustes, also occur in the monasters of Thalassema. These periodic changes are undoubtedly comparable, as Wilson maintains, with the progressive transformations of the nucleus in dividing eggs. I have frequently been able to observe in the livingegg the periodic changes in the rays of the monasters, although they cannot be seen very distinctly. Sections show that the chromosomes usually lie at the center of the aster in a clear, hyaline area, from the border of which the rays diverge (Fig. 60). In a few cases, some of the chromosomes are found scattered among the rays, although this con- dition does not seem to be the common one, as it isin Toxopneustes, and in none of the monasters in my material have I found such figures as Wilson has described, where the center is formed by a spongy centrosome from which the rays radiate (/.c., Figs. 40, 41). a Artificial Parthenogenesis in Thalassema Mellita 137 In Fig. 60 the chromosomes are very numerous, and several divisions must have occurred to produce this condition. Some of my sec- tions (Fig. 61) show cases similar to those of Wilson in which the chromosomes are actually found to be splitting longitudinally, and there can be no doubt, when the number of chromosomes in the monaster is greater than in the egg nucleus, that division of the chromosomes has occurred at each active phase of the cycle. Not only are the chromosomes found dividing among the rays, but more frequently the longitudinal splitting takes place in the central clear area of the aster, as seen in the last figure. Here one is lying on the rays and is evidently in the act of dividing, while the rest, grouped in the center, have either split to form double rods or are taken in some stage of theprocess. By counting the chromosomes inthis and the other sections of the same egg, it was found that about 24 double rods were present, so that in all probability the division which is occurring at this time is the second one in the recurring transformations of the monaster. Fig. 62 shows a monaster in which the rays do not quite reach the central area; it is probable that the egg was killed just as the radiations were beginning to reappear. VII GENERAL CONSIDERATIONS I Differentiation Without Cell Division It has been shown in the preceding pages that certain solutionsof acids furnish an efficient stimulus for the development of unfer- tilized eggs of Thalassema into embryos and larve, and, further- more, that this development in favorable experiments closely approximates, if it is not identical with, the normal processes of differentiation leading up to the formation of the swimming trocho- phore. Previous experimenters on artificial parthenogenesis of annelids have obtained with the methods which they have employed only abortive attempts at development, and their embryos and larve have widely departed from the normal in almost all respects. Especially aberrant is the case of Chaetopterus (Loeb, ’o1, Lillie 02) i in which the unfertilized eggs may be caused to Gadecee cer- tain cytoplasmic differentiations in the entire absence of cell divi- 138 George Lejevre sion and produce ciliated structures faintly simulating the appearance of trochophore larve. My observations on Thalas- sema, however, bring the annelids into closer accord with the re- sults which have been obtained with echinoderms, especially with the starfish, where parthenogenetic development has been shown to be far more normal in character than in other groups of animals experimented with. The differences between the parthenogenetic embryos of Thal- assema and the structures which have been obtained from the unfertilized eggs of other annelids, notably Chztopterus, are most marked and involve important considerations. Differentiation without cell division has never been observed in Thalassema, but on the contrary progressive differentiation of the embryo in this case depends upon cell division at every stage. The formation of the differentiated masses of ciliated proto- plasm which he observed in Chetopterus, Lillie insists, must be interpreted as just “as truly a process of development as the forma- tion of the trochophore”’ (’02, p. 493). ‘The eggs, without dividing into cells “pass through well defined phases of differentiation, the yoik accumulating in a dense mass in the interior, and the periph- eral cytoplasm becoming vacuolated and ciliated. The ciliated ectoplasm and the yolk laden endoplasm are analogous to the ectoderm and endoderm of the trochophore, and the phases of differentiation resemble some of the normal processes, though the resulting object can by no stretch of the term be properly called a trochophore” (/.c., p. 477). Furthermore, “in some cases it is even possible to homologize the regions of these unsegmented ciliated eggs with the regions of the trochophore”’ (p. 495). The process of cell division, as such, Lillie concludes, “is necessary neither to growth, differentiation, nor the earliest correlations; but it is accessory in Metazoa, to all three as a localizing factor, often from the earliest stages”’ (p. 494). These results are in sharp contrast with my observations on Thalassema, as it has been shown that the processes of differentia- tion, if they do not depend upon cell division, nevertheless, do not occur in its absence. In this form, if the egg remains unsegmented, no differentiation of the cytoplasm takes place and a ciliated em- Artificial Parthenogenesis in Thalassema Mellita 139 bryo is never produced. I have never found a single instance of the occurrence of the pseudo-trochophore described by Lillie and others. It would seem true, therefore, that in the development of Thalassema at all events, cell division 1s something more than a mere “localizing factor,” it 1s rather, on the contrary, funda- mental and essential to all processes of differentiation and cor- relation. Not only does the conclusion just expressed seem to be the cor- rect one, but it is also undoubtedly true that the more closely the course of progressive cellular differentiation follows in the path of the normal processes of development, the more nearly normal is the resulting trochophore, both structurally and function- ally. Abnormalities in cleavage at any stage seem to perma- nently disturb the organization of the embryo, while the resulting defects and deficiencies do not appear to be made good later on. It is of interest to remark in this connection, that while differen- tiation of the egg does not occur in the absence of cell division, cytoplasmic cleavage in Thalassema does not appear to take place without preceding division of the nucleus. From the most care- ful examination of my sections, I have found no evidence that the cytoplasm segments if unassociated with mitotic phenomena involve a distribution of chromatin. ‘This fact is not in accord with the observations which have been frequently made on other eggs, and may possibly be correlated to a certain extent with the absence, or, at any rate, the very rare occurrence of cytasters in the eggs of my experiments. Finally, it may be remarked that, since the parthenogenetic trochophores of Thalassema possess a highly normal organization with a differentiated digestive tract, etc., a physiologically self- sustaining organism would seem to be a possibility in the partheno- genetic development of this worm, and that success in rearing the larve to maturity must depend upon the finding of satisfactory means of nurture for tiding the animals over the critical period. This accomplishment undoubtedly lies within the bounds of experimental investigation. 140 George Lefevre 2 Origin of the Cleavage Centrosomes The formation de novo of the centrosome, at first rejected by Boveri (’01) but later accepted by him (’02, p. 40) on the evidence furnished by Wilson’s experiments (01), has been recently at- tacked by Petrunkewitsch (’04), who has attempted to defend the continuity of the centrosome while gratuitously assuming that the ovocenter, although invisible and undiscoverable after maturation, persists in the parthenogenetic egg and later gives rise to the cleay- age centrosomes. His contention that the centrosomes of the multiple- -asters are not new formations but arise by division of the primary egg center, as do all asters containing central bodies, and, furthermore, that the cytasters of egg fragments do not possess central bodies, has been adequately criticised by Wilson (’04) and shown to be utterly unsupported by evidence. The obser- vations of both Wilson (’o1) and Yatsu (’04, ’05), that asters containing centrosomes can be artificially induced in egg fragments, in which there 1s no possibility of the presence of an egg center, completely sets at rest the question of their formation de novo, and in the light of these facts the probability of a similar origin of the cleavage centrosomes in parthenogenetic eggs seems to me to be very great. Since it has been experimentally proven that centrosomes may be induced as new formations, it is very difficult for me to conceive of the centrosome and its associated radiations as anything more than an expression in cell substances of forces or activities tending to produce cell division; that is, as an effect rather than a cause—an opinion, I believe, which is becom- ing more generally prevalent. With the destruction of the older conception of the centrosome as a persistent cell organ, any attempt to rescue even a shred of the former theory, in maintaining a physiological unity for the centrosome as an active stimulating agent in cell division, must, it seems to me, be futile. The independent origin of the cleavage centrosomes in the parthenogenetic egg of Thalassema suggests the possibility that in the normally fertilized egg they may not be derived from the sperm center, but that they, too, arise as new formations in the cytoplasm. Since Grifiin’s work on this egg, Thalassema has Artificial Parthenogenesisin Thalassema Mellita 141 generally been regarded as furnishing strong evidence of the per- sistence of the cleavage centrosomes and their direct continuity with the sperm center. ‘Ihe doubt expressed above has led me to examine with great care the entire period in the normally fertilized egg from the first appearance of the sperm aster to the establish- | ment of the cleavage amphiaster, and the results which I have ob- tained from these observations are not in accord with Grifhn’s description. He was confident that he had traced the sperm cen- trosome continuously into the cleavage centers, and stated his conclusion as follows: “In Thalassema, ‘the pause’ is of short dura- tion, and while the asters are a trifle less distinct, they nevertheless show clearly throughout, and the persistence of their focal cen- trosome is easily demonstrated. ** * In most instances the presence of the centrosomes can be made out with comparatively little difh- culty. Withthe commencing fusion of the nuclei, the centrosomes take up a polar position, and immediately become the centers of renewed activity, for many additional rays commence to start up about them. From the above it is quite evident that the cen- trosomes persist entire throughout the whole critical stage where, in so many forms, they have been lost sight of’ (99, p. 598). The material upon which I have made my observations was preserved at intervals of one minute throughout the entire period in question, and the sections made from it are as nearly perfect as sections can be. Contrary to Griffin’s observations, I have found a stage, just before the fusion of the pronuclei, when the © rays of the asters become exceedingly faint, if they do not disappear entirely, and the most careful search fails to reveal the presence of centrosomes. A little later, upon the reappearance of the rays, the centrosomes of the cleavage amphiaster can be demonstrated. I am convinced, therefore, that a critical stage exists in Thalas- sema, as in many other forms, and at this time the continuity of the sperm centrosomes cannot be followed. Kostanecki (06), in a very recent paper, has undertaken an elaborate attempt to prove the universality of the origin of the cleavage centrosomes (centrioles) from the centriole of the spermatozo6n, and has con- vinced himself “das im befruchteten Ei samtlicher Metazoen die Centriolen der ersten Furschungsspindel die direkten Abk6mm- 142 George Lefevre linge des von Spermatozoon eingeftihrten Centriols sind. Die von dieser Regel statuierten Ausnahmen erweisen sich bei gen- auerer Priifung als unhaltbar”’ (p. 429). In order to arrive at this conclusion, he rejects all conflicting observations of others, in many cases on entirely insufficient grounds. Griffin’s account for ‘Thalassema is emphasized by Kostanecki as furnishing strong support for his position, yet an examination of the same egg has led me to seriously doubt the genetic continuity even in this case between the sperm center and the centers of cleavage. In the light of my observations, therefore, it is difficult for me to avoid the suspicion, at all events, that in the normally fertilized eggs of ‘halassema, as in those which develop parthenogenetically, the cleavage centrosomes arise de novo and are caused to appear in the egg cytoplasm upon renewal of those activities which lead to the division of the cell. 3 Numerical Relations of the Chromosomes It has been seen that the number of chromosomes (24) character- istic of the fertilized egg of Thalassema is not restored during parthenogenetic development, but that the reduced number is retained throughout and has been repeatedly counted even in late blastula and gastrula stages. ‘This result is in accordance with the observations of several others, who have determined the per- sistence of the changed numerical relations of the chromosomes when their number has been altered, as in the fertilization of enucleated egg fragments and in artificial parthenogenesis, or under other experimental conditions. ‘This fact has been shown to be true by Morgan (’95), Boveri (95, 05), Wilson (’o1) and Stevens (’02). The parthenogenetic eggs of Uhalassema, therefore, bear out the contention, so strongly made by Boveri, that a restitution of the normal number of chromosomes does not take place when the number has been either increased or diminished by unusual con- ditions. ‘The fact, however, is opposed tothe results which Delage (99, ’01) has drived from his experiments on merogony and arti- ficial parthenogenesis in the sea-urchin and which have led him to maintain that the normal number of chromosomes is a specific character and is restored when it has been disturbed. Boveri Artificial Parthenogenesis in Thalassema Mellita 143 (02b, ’05), on the other hand, has not only proved that Delage erroneously determined the number of chromosomes in the fertil- ized eggs of Strongylocentrotus to be 18, whereas it is 36, a blunder which entirely vitiates his conclusion with respect to the chromatin relations of his parthenogenetic sea-urchin larve, but he has also shown itto be highly probable that, in his experiments on merogony, Delage was dealing with abnormal conditions which might easily have led himinto error regardingthenumber of chromosomes present in his embryos. Morgan (’95), moreover, several years before Delage’s work on merogony, had found the reduced number of chromosomes persisting in early cleavage stages of fertilized enu- cleated fragments of the sea-urchin’s egg, an observation, how- ever, to which Delage makes no reference. In a later paper on artificial parthenogenesis of Asterias glacialis, Delage (‘02c) states that a preliminary examination of his prepara- tions has led him to believe that the number of chromosomes in the morula and blastula is the same as in embryos arising from fertil- ized eggs, that is, 18, but as he offers no evidence in support of the statement, it cannot be accepted without further investigation. Although it is safe to conclude that the initial number of chromo- somes persists, it must be borne in mind, as Boveri, Wilson, and others have pointed out, that abnormal conditions may intervene to disturb these relations and lead to a multiplication of chromo- somes in a single cell, as may be seen, for example, in the case of monasters and other pathological mitoses where longitudinal splitting of the chromosomes may occur without accompanying cleavage of the cytoplasm, as well as in eggs that have been entered by supernumerary spermatozoa. ‘The unusual conditions which I have described in the matura- tion of the unfertilized eggs of Thalassema, whereby the cleavage nucleus probably receives, not only the chromatin of the egg nucleus, but that of one or both of the polar bodies as well, pre- sumably lead to an increase in the number of chromosomes throughout subsequent mitoses. ‘This may be also true of Asterias glacialis, since, according to Delage, the eggs of the starfish, like those of Thalassema, may at times fail to extrude one or both polar bodies. I have not been able thus far to prove that later 144 George Lejevre embryos arising from these abnormal eggs possess the additional chromatin, although I have found indications of such a condition, but I shall attempt to investigate this point at a future time when adequate material may be at hand for the purpose. SUMMARY 1 The unfertilized eggs of Thalassema mellita may be induced to develop parthenogenetically into actively swimming trocho- phores by immerison for a few minutes in dilute solutions of acids, both inorganic and organic. 2 After transfer from the acid solutions into normal sea-water, the egg throws off a typical fertilization membrane, the germinal vesicle breaks down, and maturation and cleavage follow. In successful experiments, which were the rule, from 50to 60 per cent of the eggs developed into swimming larve that could scarely be distinguished from normal trochophores of a corresponding stage. 3 The parthenogenetic development, in the majority of cases, involves a strictly normal maturation, a normal cleavage, at least in the early stages, and the usual processes of differentiation that occur after fertilization by sperm. 4 Gastrulation takes place in the normal manner, and the par- thenogenetic larva possesses a digestive tract, differentiated into mouth, cesophagus, stomach and intestine, and the prototroch and apical plate bearing the normal arrangement of cilia. 5 After maturation, the egg center disappears, and the cleavage centrosomes arise de novo, probably without division of a single primary center, and, when first seen, lie on opposite sides of the egg nucleus which becomes the first cleavage nucleus. 6 Cell division occurs mitotically throughout development, and division of the nucleus is usually accompanied by cytoplasmic cleavage. 7 The number of chromosomes characteristic of the fertilized egg is not restored, but the reduced number (12) 1s retained and has been counted repeatedly, even in late stages. 8 The rate of division is not as rapid, nor as regular as in normal segmentation, and the parthenogenetic larvz, although swimming Artificial Parthenogenesis in Thalassema Mellita 145 vigorously at the bottom of the dish, do not rise to the surface of the water. g After exposure of the eggs to acid solutions, the polar bodies may continue to divide mitotically and form a morula-like cluster of minute cells, thus exhibiting an attempt at parthenogenetic development. 10 In some experiments, the eggs extruded only one polar body and in others neither polar body was formed. In such cases, either one or both maturation mitoses take place inside the egg, with the resulting formation of resting nuclei which probably fuse to form a cleavage nucleus. In still other cases, there is evidence for believing that the first maturation spindle may directly become the first cleavage spindle, across which the egg divides into equal or subequal cells. ‘The numerical relations of the chromosomes in these cases have not been definitely determined. Eggs exhibit- ing these abnormalities of maturation give rise to larve indis- tinguishable from eggs which maturate normally. 11 An endless variety of abnormal cleavages, similar to those described by others, have been observed. Such cleavages lead to the formation of ciliated, cellular structures which, however, depart more or less widely from normal embryos. 12 Abnormalities of mitosis, as polyasters and monasters, are not infrequent, and when nuclear division is not followed by cleavage of the cytoplasm, chromosomes in excess of the usual number (12) may be found in a single cell. 13 Cytasters are either absent or exceedingly rare, and cyto- plasmic cleavage without preceding nuclear division has not been observed. 14 Amceboid movements of the egg are rare, and when they occur, are not extensive; “fusion phenomena” are lacking. 15 Cell division would seem to be a fundamental and essential factor in differentiation, since in no instance was a differentiated, ciliated structure observed which was unsegmented; the partheno- genetic pseudo-trochophores, which have been described for Chzptopterus and other annelids, are entirely absent. Zodlogical Laboratory University of Missouri August 15, 1906 146 George Lejevre LITERATURE CITED Boveri, Tu., '87—Zellen-Studien, I. Jena. °88—Zellen-Studien, II. Jena. *90—Zellen-Studien, II]. Jena. ’95—Ueber die Befruchtungs- und Entwickelungsfahigkeit kernloser Seeigel-Eier, etc. Arch. f. Entw’m. d. Org., II, 3. *99—Die Entwickelung von Ascaris meg. mit besonderer Ricksicht auf die Kernverhaltnisse. Festschr. f. C. v. Kupffer. Jena. ‘o1—Ueber die Natur der Centrosomen. Zellen-Studein, IV. Jena. *o2za—Das Problem der Befruchtung. Jena. o2b—Ueber mehrpolige Mitosen als Mittel zur Analyse des Zell-kerns. Verh. der Phys., med. Ges. Wiirzburg, N. F., xxxv. 05—Zellen-Studien, V. Jena. Brauer, A., ’93—Zur Kenntniss der Reifung des parthenogenetisch sich entwick- elnden Eies von Artemia salina. Arch. f. mikr. Anat., xlii. Butiot, G.—’o4 Artificial Parthenogenesis and Regular Segmentation in an Annelid (Ophelia). Arch. f. Entw’m. d. Org., xviii, I. Conn, H. W., ’84—Life History of Thalassema (Abstract). Stud. Biol. Lab. Johns Hopkins Univ., ui, 1. ’86—Life History of Thalassema. Jb1d., ii, 7. Cow tes, R. P., °03—Notes on the Rearing of the Larve of Polygordius appendi- culatus, etc., Biol. Bull., iv, 3. DELAGE, YVEs, ’99—Etudes sur la mérogonie. Arch. Zool. Exp. et Gén. (3), vii. *o1—Etudes expérimentales sur la maturation cytoplasmique et sur la parthénogénése expérimentale. bid. (3), ix. ’o2za—L’acide carbonique comme agent de choix de la parthénogénése expérimentale chez les Astéries. C. R. Acad. Sc. (Paris), cxxxv. ’o2b—Sur la mode d'action de l’acid carbonique dans la parthénogénése expérimentale (Asterias). b1d., cxxxv. ’o2c—Nouvelles recherches sur la parthénogénése expérimentale chez Asterias glacialis. Arch. Zool. Exp. et Gén. (3), x. ’o4a—Elevage des larves parthénogénétiques d’Asterias glacialis. Ibid. (4), iv, I. o4b—La parthénogénése par l’acid carbonique obtenue chez les oeufs apres l’emission des globules polaires. Ibid. (4), il, I. Fiscuer, M. H., ’02—Further Experiments on Artificial Parthenogenesis in Anne- lids. Amer. Journ. Physiol., vii, 3. ’03—-Artificial Parthenogenesis in Nereis. Jbid., ix, 2. Francotre, P., ’97—Recherche sur la maturation, etc., chez les Polyclades. Mem. cour. Acad. Sci. Belg., 1897. Artificial Parthenogenesis 1n Thalassema Mellita 147 GREELEY, A. W., ’02—Artificial Parthenogenesis in Starfish Produced by a Lower- ing of Temperature. Amer. Journ. Physiol., vi, 5. GrirFINn, B. B., ’96—The History of the Achromatic Structures in the Maturation and Fertilization of Thalassema. Trans. N. Y. Acad. Sci., xv. *99—Studies on the Maturation, Fertilization and Cleavage of Thalas- sema and Zirphea. Journ. Morph., xv, 3. Hersst, Curt, ’93—Ueber die kiinstliche Hervorrufung von Dottermembranen an unbefruchteten Seeigeleiern nebst einigen Bemerkungen iiber die Dotterhautbildung tiberhaupt. Biol. Centralbl., xiii. ’04——Ueber die kiinstliche Hervorrufung von Dottermembranen an unbe- fruchteten Seeigeleiern. Mitth. a. d. Zool. Stat. z. Neapel, vi. Hertwic, O. u. R., °87—Ueber den Befruchtungs- und Teilungsvorgang des tierischen Eies unter dem Einfluss ausserer Agentien. Untersuch. z. Morph. u. Physiol. d. Zelle. Heft 5. Jena. Hertwic, R.,’96—Ueber die Entwickelung des unbefruchteten Seeigeleies. Festschrift f. Gegenbaur, 11. Hunter, S. J., °04—On the Morphology of Artificial Parthenogenesis in the Sea- urchin, Arbacia. Science, N. S., xix, 475, p. 213. Kine, H. D., ’06—The Effects of Compression on the Maturation and Early Development of the Eggs of Asterias forbesii. Arch. f. Entw’m. ad Org., xx; 2: KostanEckI, K., ’02—Ueber kiinstliche Befruchtung und kiinstliche partheno- genetisch Furchung bei Mactra. Bull. Acad. Sci. Cracovie. Classe d. Sci. math. et nat. Juillet, 1902. ’04—Cytologische Studien an kunstlich parthenogenetisch sich entwickeln- den Eiern von Mactra. Arch. f. mikr. Anat., Ixiv. ’o6—Ueber die Herkunft der Teilungscentren der ersten Furchungs- spindel im befruchteten Ei. Jbid., Ixviii, 3. Kowa .evsky, A., ’72—Mittheilungen tiber die Entwickelung von Thalassema. Zeitschr. f. wiss. Zool., xxii. LEFEVRE, GEORGE, ’05—Artificial Parthenogenesis in Thalassema mellita. Sci- hice, N-75-5 5X15 532; P. 370: °o6—Further Observations on Artificial Parthenogenesis. b1d., xxiii, ’ 588, p. 522. Lititz, F. R., ’02—Differentiation without Cleavage in the Egg of the Annelid Chztopterus pergamentaceus. Arch. f. Entw’m. d. Org., xiv, 3-4. °06—Observations and Experiments Concerning the Elementary Phe- nomena of Embryonic Development in Chetopterus. Journ. Exp. Zo6l., ili, 2. 148 George Lefevre Logs, J., o1—Experiments on Artificial Parthenogenesis in Annelids (Chztop- terus) and the Nature of the Process of Fertilization. Amer. Journ. Physiol., iv. ‘o5a—On Fertilization, Artificial Parthenogenesis, and Cytolysis of the Sea-urchin Egg. (Translated from Pfliiger’s Archiv, ciii, 1904.) Univ. of Calif. Pub., Physiology, ii, 8. ‘o5b—On an Improved Method of Artificial Parthenogenesis. /bid., ii, 9. °o5c—On an Improved Method of Artificial Parthenogenesis (Second Communication). Ibid., ii, 11. °o5d—On an Improved Method of Artificial Parthenogenesis (Third Communication). Ibid., ii, 14. ’o5e—Artificial Membrane Formation and Chemical Fertilization in a Starfish (Asterina). Jbid., ii, 16. °o6—The Dynamics of Living Matter. Macmillan, New York. Logs, J., Fiscner, M. H., and Nettson, H., ’01—Weitere Versuche tiber kiinst- liche Parthenogenese. Arch. ges. Physiol., Ixxxvii. Matuews, A. P., ’o1—Artificial Parthenogenesis Produced by Mechanical Agita- tion. Amer. Journ. Physiol., vi, z. Meap, A. D., ’95—Some Observations on Maturation and Fecundation in Che- topterus pergamentaceus, Cuvier. Journ. Morphol.,x , 1. ’°98a—The Rate of Cell Division and the Function of the Centrosome. Biol. Lectures, Woods Hole, 1896-97, Boston. *98b—The Origin and Behavior of the Centrosomes in the Annelid Egg. Journ. Morph., xiv, 2. Morcan, T. H., ’95—The Fertilization of non-nucleated Fragments of Echino- derm Eggs. Archiv f. Entw’m. d. Org., ii, 2. °96—The Production of Artificial Astrospheres. Jb1d., iii. *99—The Action of Salt Solutions on the Unfertilized and Fertilized Eggs of Arbacia. Jbid., viii, 3. *oo—Further Studies in the Action of Salt Solutions and other Agents on the Eggs of Arbacia. Jbid., x, 2, 3. PETRUNKEWITscH, A., ’04—Kiinstliche Parthenogenese. Zool. Jahrb. Suppl., vii; Festschr. f. A. Weismann. Scorr, J. W., ’o6—Morphology of the Parthenogenetic Development of Amphi- trite. Journ. Exp. Zo6l., iu, I. Stevens, N. M., ’02—Experimental Studies on Eggs of Echinus microtuberculatus. Arch. f. Entw’m. d. Org., xv. Torrey, J. C., ’02—The Early Development of the Mesoblast in Thalassema. Anat. Anz., xxi, 9. el Artificial Parthenogenesis in Thalassema Mellita 149 Torrey, J. ., Co3—The Early Embryology of Thalassema mellita (Conn). Annals N. Y. Acad. Sci., xiv, 3. TREADWELL, A. L., ’°02—Notes on the Nature of “Artificial Parthenogenesis”’ in the egg of Podarke obscura. Biol. Bull., in, 5. Witson, E. B., ’01—Experimental Studies in Cytology, I. A Cytological Study of Artificial Parthenogenesis in Sea-urchin Eggs. Arch. f. Entw’m. d. Org., xii. *o4—Cytasters and Centrosomes in Artificial Parthenogenesis. Zool. Anz. XXViN, I: Yatsu, N., ’04—Aster Formation in Enucleated Egg Fragments of Cerebratulus. Science, N. S., xx, 521. 7o5—The Formation of Centrosomes in Enucleated Egg Fragments. Journ. Exp. Zodl., 11, 2. Pirate I All of the figures of this and the succeeding plates were drawn from parthenogenetic material, with the aid of the camera. The drawings were made under a magnification of 1000 diameters, except Figs. 5-14, as well as Fig. 16, in which the magnification was 700 diameters. The dotted line, which is drawn in many of the drawings surrounding the nucleus or mitotic figure, encloses the yolk-free area. All figures are from living material, except 2 and 3 Fig. 1 Living egg, showing delicate protoplasmic threads extending from the surface of the egg to the membrane, and also multiple polar bodies. Fig.2 Section of full-grown odcyte immediately after treatment with acid but before breaking down of germinal vesicle; the membrane, which was present, is not represented. Fig. 3 Section of egg in which the germinal vesicle has migrated to animal-pole after acid-treatment. Fig. 4 Living egg showing three polar bodies; the figure was drawn immediately after the division of the first polar body. Figs. 5-8 Two, four, eight and sixteen-cell stages drawn from living eggs. Fig. Young HCl trochophore, 15 hours old, showing apical flagella, cilia of prototroch and blastopore 5/. Fig. 10 HNOsztrochophore, 36 hours old, showing the larger cilia on upper and lower border of prototroch and the mouth mo; the cavities of stomach st, and intestine im, are seen at a deeper level. Figs. 11-12 Young trochophores, showing abnormal ciliation; in the former, short cilia cover the entire pre-trochal region, while the latter lacks the apical flagella. Fig.13 Abnormal embryo, in which the cilia have failed to puncture the membrane mb, and are pressed down by latter. Fig. 14 Abnormal 3-cell stage, formed at second cleavage. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA GrorGE LEFEVRE PLATE 1 THE JourNAL or ExPERIMENTAL ZOOLOGY, VOL. IV, NO. I Pirate II All figures are from sections, except 15 and 16 Fig. 15 Optical section of living, abnormal blastula which was cylindrical in form and open at both ends. ; Fig. 16 Egg showing ameeboid activity after treatment with acidulated water. Fig. 17 Simultaneous appearance of the two asters, with centrioles, on wall of germinal vesicle. Fig. 18 Breaking down of germinal vesicle and formation of first maturation spindle; the centro- somes are double and the fibers passing to chromosomes are thicker than rest. Figs.19-20 First maturation spindles, fully formed , the latterin its definitive position; the chromosomes (tetrads) are in the form of crosses, in some of which the perpendicular split may be seen. Fig.21 Formation of second maturation spindle; dyads are seen in both the egg and the first polar body; two centrioles with a few delicate rays are present in latter. Fig. 22 Second polar spindle in definitive position; the dyads are in the equatorial plane with the longer axis lying transversely to the spindle. Fig. 23 Anaphase of second polar mitosis; the inner centrosome is double. The first polar body shows mitotic activity. Fig. 24 Close of second maturation mitosis, showing 12 single chromosomes in egg. as well as several similar chromosomes in the second polar body. Fig. 25 Chromosomal vesicles partially fused to form the egg nucleus; centrosomes and rays have totally disappeared. Fig. 26 Egg nucleus during the ‘‘ pause.” Fig. 27 Simultaneous appearance of the two cleavage asters with their centrosomes on opposite sides of the egg nucleus. Fig. 28 Fully formed, first cleavage figure; the large, reticulated centrospheres contain divided centrosomes, and the spireme is segmenting into chromosomes. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA Grorce LEFEVRE a Tue Journat or Expertmentat ZodLoGy, VOL. Iv, No. I ea. PLATE Il Prate III All figures are from sections Fig. 29 Transverse section of equatorial plate of first cleavage mitosis, showing 12 chromo- somes, the reduced number. Fig. 30 Telophase of first cleavage, showing chromosomal vesicles and daughter amphiasters. Figs. 31-32 Prophases of second cleavage. Fig. 33. Anaphase of second cleavage, showing divided centrosomes. Fig. 34 Four-cell stage, resting condition. Fig. 35 Eight cell stage in section. Figs. 36-38 Sections of young blastule, showing cells in various stages of mitosis. In Fig. 36,a transverse section of the equatorial plate with its 12 chromosomes is shown at a. Divided centrosomes are seen in some of the mitotic figures. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA PLATE III GrorcGe LeFevre Tue Journat or ExperiMENTat Zo6LOGY, VOL. IV, NO. I Prate IV All figures are from sections Fig. 39 Young blastula with abnormally large blastoceel. Fig. 40 Section of an early gastrulation stage, showing the flattening of the lower pole of the embryo and the large entoblastic cells just before sinking in. Several rudimentary cells rc are seen in the cleavage cavity; one lies inside an endodermal cell. A polar body pb, showing mitotic activity, is in the blastoceel. Fig. 41 Section of a young gastrula, showing a mass of entomeres within the cleavage cavity; the cell marked oes is probably an cesophagoblast. At a the 12 chromosomes of an equatorial plate are seen. Rudimentary cells are shown at rc. Fig. 42 Later gastrula, showing entoblastic mass with beginning cavity en. The stippled cells in this and the next two figures are cells of the larval mesenchyme. Fig. 43 Later stage, showing a larger enteric cavity and also the ectodermal, oesophageal invagi- nation; the cells marked oes are probably cesophagoblasts. Fig. 44 Horizontal section of a CO, trochophore,31 hours old. The enteric cavityis now divided by a septum into stomach st and intestine in. The prototroch and apical plate are also seen. Fig. 45 Egg showing submerged second maturation spindle, with chromosomes at poles. The first polar body is preparing to divide. Fig. 46 Egg showing two resting nuclei, probably the result of the mitosis seen in last figure. The mitosis of the first polar body is in anaphase. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA PLATE IV GrorGe LEFEVRE AF) oye. XS d9 OX YO? FED) Pirate V All figures are from sections Fig. 47 Submerged first maturation spindle, showing divided centrosomes and dyads passing to the poles. Fig. 48 The two nuclei, which have probably resulted from the mitosis seen in the last figure, are preparing for a second mitosis. Fig. 49 Probably later stage of the same; the four nuclei may have arisen from a second submerged mitosis. Fig. 50 Transverse section of equatorial plate of first cleavage spindle, containing 23 chomosomes; the increased number of chromosomes is doubtless due to the fusion of two submerged maturation nuclei to form the cleavage nucleus. Figs. 51, 52 Enlarged first maturation figures, which probably behave as cleavage figures and lead to an equal division of the egg. In the anaphase shown in Fig. 52, the chromosomes are double and have the typical appearance of dyads. Fig. 53 Prophase of multipolar mitosis in undivided egg. Fig. 54 Abnormal four-cell stage, showing multipolar mitoses in three cells. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA PLATE V GrorGe LEFEVRE THE Journat or ExperIMENTAL ZOOLOGY, VOL Iv, No. I Pirate VI All figures are from sections Fig. 55 Undivided egg containing multiple asters and spindles; the separate asters near the peri- phery are probably cytasters, some of which show centrioles. Fig. 56 Single, large nucleus containing numerous chromosomes, which have probably been gathered together after a multipolar mitosis. Fig. 57 Undivided egg containing multiple nuclei, which have probably arisen through multipolar mitoses. Fig. 58 Abnormal embryo, showing one large cell with multiple spindles and asters. Fig. 59 Abnormal embryo, showing two giant spindles and their multiple chromosomes. Fig. 60. Monaster in undivided egg; numerous chromosomes are present in the central, clear area, while a few are scattered among the rays. Fig. 61 Monaster in unsegmented egg; the chromosomes in the central area are dividing or have divided, while one lying outside is also in the act of dividing. Fig. 62 Monaster in unsegmented egg; this is probably a prophase of a period of activity which will culminate in another division of the chromosomes. ARTIFICIAL PARTHENOGENESIS IN THALASSEMA PLATE VI Georce LEFEVRE \ \l M/ \yi AZ, \ \ Wii ilii Wiel YW ‘ S A\ il iy SZ, tr TNS MII ASS JTW WS / Tue Journat or ExpeRIMENTAL ZOOLOGY, VOL. IV, No. I CONCERNING THE THEORY OF TROPISMS BY JACQUES LOEB About twenty years ago! I began a physico-chemical analysis of the behavior of lower animals which had heretofore been explained in the anthropomorphic way characteristic of archaic science. My main efforts were directed toward the analysis of the role which light and gravitation play in the reactions of animals. I showed first, that the orientation and the direction of the pro- gressive motion of certain animals can be controlled unequivocally by the direction of the rays emanating from a source of light, and I showed, moreover, that this type of reaction is, as far as we can judge, in every point identical with the heliotropic reaction of plants. A few years later I showed that there exists another group of a imal reactions to light, which is not covered by the theory of tropisms, but which depends upon the rapidity of the change of the intensity of the light.*- This latter type of reaction I designated as Unterschiedsempfindlichkeit. “Those who are familiar with the terminology of the physicist will most readily understand the dif- ference between the two types of reaction if I state that heliotrop- ism depends upon the value / (7), where 7 is the intensity of light, while in Unterschiedsempfindlichkeit the reaction depends upon d1 ; ; : , the value of ai where ¢ is the time. Both forms of reactian may t occur in the same animal (e. g., Spirographis), but this 1s neither necessary nor the rule. 1My first two papers on animal heliotropism and geotropism appeared in January, 1888, and Not, as is often stated, in 1890 (Sitzungsber. der Wiirzburger Phys.-med., Gesellschaft, 1888). *T refer the reader to the following papers: Pfliiger’s Archiv, vol. 54, p. 100, 1893. (Studies in General Physiology, vol. 2, p. 286.) Pfliiger’s Archiv, vol. 56, p. 247, 1894. (Studies, vol.2, p. 345.) Pfliiger’s Archiv, vol. 66, p. 439, 1897. Comparative Physiology of the Brain. Dynamics of Living Matter, pp. 135-137. THE JouRNAL oF ExPpERIMENTAL ZOOLOGY, VOL. IV, NO. I. 152 Facques Loeb The reader will readily notice that I did not attempt to slow that all animal reactions are of the type of tropisms; on the con- trary, | was, as far as I am aware, the first to point out that there exists a type of reactions which are as different from tropisms as are quantities of the dimension of an acceleration from those of the dimension of a velocity. My aim was to analyze the behavior of animals from a physico-chemical point of view and substitute the methods of modern science for the anthropomorphisms of the metaphysician. In this attempt it made no difference to me whether the elementary components of the complex “Animal Behavior” were found to be of the type 7 (z), (e. g., tropisms) or of the type #($) (e. g., Unterschiedsempfindlichkeit) or of any other definite function. Moreover, I laid emphasis on the fact that it is necessary to control the animal reactions before explain- ing them, as only the control of the reactions offers a sufficient test for the correctness of our analysis. From this point of view I stated that for the control of heliotropic reactions the intensity of the light may remain constant during the experiment, while for the control of the reactions of the type of Unterschiedsemp- findlichkeit the intensity of the light must change with a certain rapidity during the time of the experiment. Whether it 1s due to mere carelessness or some other cause, a number of American authors have disregarded this discrimination, making their readers believe that the cases of Unterschiedsemp- findlichkeit are represented by me as examples of tropisms and then showing that the facts do not coincide with what they state to be my theory of tropisms. I will give a definite instance of this procedure. In 1893 I described a case of Unterschiedsempfindlichkeit in a tubicolous worm at Naples, Serpula uncinata, showing that when the rapidity of the decrease of the intensity of light reaches a cer- tain value a contraction of this worm 1s caused, while an equally rapid increase in the intensity of light causes no such reaction. In later writings I especially singled out this reaction to illustrate the typical difference between Unterschiedsempfindlichkeit and Concerning the Theory of Tropisms 153 tropism. To quote froma paper I published in 1897. “I noticed in the course of my investigation that besides the heliotropic effects there exists a recone kind of mechanical effects of light which is determined by the rapidity of the change of the inten- sity of light and which I designated as Unterschiedsempfindlich- keit * * * I found this type of reaction in tubicolous annelids- e. g., Serpula uncinata. ‘The gills of these animals protrude from the tube. If we move our hand between the animal and the source of light it rapidly withdraws into its tube as soon as the shadow strikes it. Inorder to find out whether positive and nega- tive changes in the intensity of light have the same effect I made the following experiment: A glass aquarium which was covered with a glass plate was placed on a table, about two meters from the window. When I rapidly closed the shutters the worms rapidly withdrew into their tubes, as a snail would upon a sudden touch. The shutters did not close tightly and it was sufficiently light in the room to observe the animals. If one waited a little the animals again stretched their gills out of the tube. When now the shutters were rapidly opened, no reaction on the part of the animals occurred. When they were inside the tube the opening of the shutters did not cause them to reappear. It is, therefore, only the decrease in intensity which acts as a stimulus upon the animals.” In the same paper I pointed out that in the physiological effect of the galvanic current physiologists discriminate between the effects dependent upon a constant current and the effects which depend upon the rapidity of the changes in the intensity of a cur- rent; and I showed that these differences correspond to the differences between a tropism and Unterschiedsempfindlichkeit. Since inductive effects depend upon the value of § the funda- mental importance of the discrimination between a tropism and Unterschiedsempfindlichkeit is at once obvious. Since the excised muscle is above suspicion of possessing a human soul it might be inferred that for the understanding of the corresponding type of reactions of lower animals physico-chemical data might suffice. $Pfliiger’s Archiv, vol. 66, p. 439. 154 Facques Loeb In a recent paper, “Experiments on the Behavior of Tubicolous Animals,” published in this Journal,‘ Hargitt repeats and confirms these simple experiments. He refers to my paper, just quoted, and then starts upon the following discussion: “There can be no doubt, therefore, that the reaction is not due to simply a difference of light intensity alone For whether in diffused or direct sunlight whether in natural or artificial light, the response is to the shadow, sudden diminution of light, a purely negative condition.. But it may well be doubted whether this can be properly designated as simply negative phototropism or helio- tropism” (Hargitt, p. 300). Whoever designated these reactions of tubicolous worms as negative heliotropism? Should Hargitt really, with my papers before him, state that I had done so? He does not leave his readers in doubt: “Furthermore it must be recalled in this connection that the particular stimulus involved in these observations, as previously pointed out, is not light at all directly, but the lack of light, or theshadow. Responseis, therefore, induced bya negative stimulus, if such an apparent paradox be tolerable in relationto phenomena of behavior. Of course, it is not overlooked that Loeb has desig- nated these and similar reactions as dueto ‘negative heliotropism.’ At the same time it is not clear that in the present case we are dealing with phenomena at all comparable with those associated with negative heliotropisms as ordinarily understood. For, as already observed, the phenomena are not in themselves negative. They are not dependent upon any given degree of light, or rather darkness, but to the suddenness of the change’ (Hargitt, p. 316). As was to be expected Hargitt concludes from this that the theory of tropism is no longer tenable and that we must return to the anthropomorphic viewpoint, for which, as he states, Jennings has already paved the way. “Under the later development of the theory of tropisms and its extension to the phenomena of animal behavior, its dominance has £Vol. 3, p. 295, 1906. Concerning the Theory of Tropisms 155 relegated the earlier views to the limbo of discarded anthropo- morphisms so-called. Without essaying any review of the pros and cons of this problem it may be said that already a reaction has taken place and frankness compels a reconsideration of some of these discarded and discredited views. Such a review has already been made by Jennings so far as it relates to the lower organisms, and his conclusion must, it seems to me, be equally true for many if not most higher animals as well” (Hargitt, p. 313). What Hargitt has done in one case, Jennings has done in a num- di ber of cases. He selects reactions of the type7 = , shows that these reactions do not conform with the theory of tropisms but fails to inform his readers that I had pointed out the existence of this type of reaction and their difference from tropisms long before he did. To give an illustration: Ina paper on the “Brain Phy- siology of Worms” I[ described experiments on Planarians and earthworms, showing that these animals are not or only slightly heliotropic but react to sudden changes in the intensity of light. Such animals become more quiet when the intensity of light 1s rapidly diminished, become more active when the intensity of light is suddenly increased. ‘The consequence 1s that places of a relative minimum in the intensity of light act like a trap upon them. To illustrate this effect and the difference of this reaction from that of heliotropic animals I mentioned the following experiment: One half of a glass vessel is covered with black paper, the other half left uncovered. If Planarians or earthworms are put into such a vessel they collect under the covered half. “They come to rest in those regions which are more weakly illuminated than the surrounding areas. ‘The direction of the rays of light is of little consequence.”* On p. 130 of his recent book on the “Behavior of the Lower Organisms,” Jennings describes the same experiment for Stentor and shows that these organisms will go from the light into the dark but not in the reverse direction. ‘*The essential point is the running back into the shaded region 5Loeb: Studies in General Physiology, vol. i, p. 360. Pfliiger’s Archiv, vol. 56, p. 247, 1894. 156 Facques Loeb without reference to the direction from which the light comes” (Jennings p. 131). Neither in this nor in any other case in which Jennings describes reactions which depend upon sudden changes in the 1 intensity of light (or any other form of energy) does he refer to my previous experiments on this type of reaction.* I think, however, that those who are working in this field should realize that reactions due to rapid changes in the intensity of light or any other form of energy were first recognized as being typically different from the cases of animal tropisms by the author of this latter theory; and that if new cases of Unterschiedsempfindlichkeit are found this does not contradict the existence of the reactions of the type of tropisms any more than the existence of the make and break contractions in a muscle contradicts the existence of electro- tonic effects in the same organ; or the existence of accelerations contradicts the existence of velocities. ¢ Although Jennings has attacked my views for years he is certainly not familiar with my papers. This is also evident from his erroneous statement of the theory of tropism on p. 94 of his paper, published by the Carnegie Institution in 1904. From the Rudolph Spreckels Physiological Laboratory of the University of California. THE MECHANISM OF THE GALVANOTROPIC ORIENTATION IN VOLVOX BY FRANK W. BANCROFT In a recent paper O. P. Terry (’06) has recorded the discovery that “if keptin the dark for two or three days, the [galvanotropic] response of volvox is changed from cathodicto anodic. ‘This may then be reversed at will by exposure to light.” He did not, how- ever, attempt to determine the mechanism by means of which this change in the direction of migration of the organism is brought about. As it appeared to me that the best opportunity for study- ing the nature of galvanic stimulation by the electric current was presented in those cases where a reversal of the galvanotropism was possible, I have investigated the effect of the current upon the flagella of volvox. 2 The only observation that I know of on the response of the flagella of this plant to the constant current is by Carlgren (09, Pp: 57), who was investigating volvox that was typically swimming toward the cathode. He added carmine to the fluid containing the organisms, and says that on several occasions he saw the cur- rent produced by the flagella stop at the anode while it continued at the cathode. METHOD The 110 volt power current was led through a water rheostat, milliammeter, pole changer and non-polarizable boot electrodes to the preparation containing the volvox. ‘The colonies were usu- ally examined in a small glass and paraffine trough. When the currents produced by the flagella were to be observed, india ink Tue Journat oF ExPERIMENTAL ZOOLOGY, VOL. IV, NO. I. 158 Frank W. Bancroft was mixed with the fluid in which the plants were swimming. ‘This made the determination of the direction of the effective stroke of the flagella an easy matter. ‘The preparation was covered with a coverglass resting on one or two thicknesses of filter paper, according to whether the organisms were to be held fast by the coverglass, or to be allowed to swim freely. The filter paper was connected with the boot electrodes by means of a hanging drop. The most useful modification of this method was to cut a hole in the center of a circular piece of filter paper about 8 cm. in diameter. This was placed upon a piece of glass, the volvox colonies were put into a hole in the center, covered with a coverglass, and exam- ined upon a revolving microscope stage. The boot electrodes were connected as usual with the filter paper by means of drops. By this means it was not only possible to move the preparation about, and follow the movements of swimming individuals, but the plants could also be turned through any desired angle, and thus the angle which their long axes made with the current lines passing through them could be varied to any extent without changing the current, and while the organisms were kept continuously under observation. The current density varied from about 200 to about 2503, from 40 to 100¢ being the strength usually employed. ‘The plants were always studied in the water in which they had been living when collected, or in tap water of similar composition. REVERSAL OF GALVANOTROPISM I have been able to confirm Terry’s results concerning the direc- tion of migration in all important respects. Plants that had been kept in the dark or exposed to diffuse daylight all went to the anode, while of those that had been exposed to dfrect sunlight for half an hour or more, 30 per cent or more went to the cathode. There appeared, however, to be considerable difference between my results and Terry’s as regards the intensities of light required for the anodal and cathodal galvanotropism. Thus Terry found that the usual response for volvox exposed to diffuse daylight was cathodic, while sometimes days of exposure to darkness were required to change the response to anodic. On the other hand, Galvanotropic Orientation in Volvox 159 the volvox which I examined normally showed strong anodal gal- vanotropism after an exposure to even bright diffuse daylight; and exposure to the direct sunlight was required to change the response to cathodal galvanotropism. BEHAVIOR OF THE FLAGELLA Volvox, as is well known, invariably swims with its anterior end in advance. No one has described a backward swimming of the colony as the result of any stimulus. It is accurately oriented by both light (Holmes, ’03, p. 320) and the electric current (Carlgren, ’00), but no one has worked out the mechanism of either of these orientations. Holmes, however (’03, p. 321), states for the helio- tropism that: ‘We are safe in saying that when volvox changes its direction it is because the flagella on the two sides of the organ- ism beat unequally.” Fig. 1 Diagram of volvox, showing the currents in normal locomotion. The feathered arrow indicates the direction of progression of the colony, the other arrows indicate the direction of the currents in the water. If india ink be added to a preparation containing volvox cur- rents are easily observed beginning at the anterior end and sweep- ing backward to the posterior extremity on all sides with equal intensity, as indicated by the arrows in Fig. 1. examined these currents around volvox colonies that were in the act of orienting themselves heliotropically but could make out no differences in the currents on the two sides. I do not doubt, however, that Holmes’ statement is correct. In the case of galvanotropic individuals, however, differences in the currents at the anode and cathode ends of the organisms are easily detected. [he most satisfactory way of studying these differences is with volvox mounted in a sufficient thickness of fluid so that it can swim freely. It swims so slowly that its motions 160 Frank W. Bancroft can easily be followed under the compound microscope, without the addition of any sticky substances to the culture fluid. If now a current be passed through such a preparation of volvox that has been exposed to weak light, strongly anodal colonies can be picked out and investigated. In such colonies it was invariably and frequently observed that the currents stopped or became very weak on the anode side of the colony while it was not changed much at the cathode side. Sometimes there was no change at the cathode side, and sometimes there was a decided increase in the activity of the flagella when the current was made, and during Fig.2 Diagram of volvox, showing the direction of the currents in the surrounding “medium when - the colony is swimming in a constant current. I, anode on the side. The currents stop on that side and the colony turns toward the anode. 2, anode at anteriorend. There are no currents at the anterior end, and the colony swims toward the anode. 3, the current has just been reversed, so that the anterior end is now cathodal. Currents have started at the anterior end, and have stopped at the posterior (now anodal) end. the flow of the current. ‘The same change in the behavior of the flagella was observed at the two poles no matter what the position of the organism. If it is swimming at right angles to the current lines then the flagella on the anode side stop, while those on the cathode continue and the organism is rapidly turned toward the anode (Fig. 2), and continues swimming in that direction. When Galvanotropic Ortentation in Volvox 161 the plant is swimming toward the anode there is seen to be no current at its anterior end while the current continues normally posteriorly (Fig. 2). If, now, the current be reversed a violent current is initiated at the anterior end and the current at the posterior end stops (Fig. 2). After this reversal the plant usually continues swimming straight forward to the cathode for a short distance, but soon swerves a little to one side or the other and rapidly orients itself to the anode again. If the colonies have been exposed to sunlight and made cath- odally galvanotropic the behavior of the flagella is completely reversed. Exactly the same pictures are presented to the observer but the current now stops on the cathodal side of the organism and the volvox colony becomes oriented toward the cathode. ‘This result is of some importance for it shows that in the case of volvox the direction of migration has not been changed by some quanti- tative changes in the action of a mechanism which remains funda- mentally the same; as has been described by Wallengren (’02, ’03) for Opalina and Spirostomum. In the cases described by Wallen- gren the forward stroke of the cilia was always produced by the galvanic current at the cathode side of the animal no matter what its orientation; and, as the forward stroke of the cilia is the best criterion we have for stimulation of infusoria (Bancroft, 105 )a it must be concluded that in all these cases the underlying nature of the galvanic stimulation was the same. In the case of volvox, however, the facts described indicate clearly that in the anodally galvanotropic colonies the pole at which the constant current pro- duces its characteristic effect is the anode, while in the cathodal colonies the same effect is produced by the current at the cathode. In other words the galvanotropism has been reversed by means of a reversal of the pole at which the electric current produces its characteristic effect. The only other case that I know of in which a reversal of the galvanotropism has been shown to be brought about in this way is that of Paramecium in which it was found possible to reverse the galvanotropism by chemical means (Bancroft ’o06, ’o6a). In this case it was made probable that the chemical conditions under- lying galvanic stimulation were the same no matter whether that 162 Frank W. Bancroft stimulation takes place at anode or cathode and dependent upon a certain definite ratio of the calcium ions to those of sodium, potassium, lithium, ammonium and other metal ions. The cur- rent can vary this ratio, as described by Loeb (’05), since many organic lons precipitate calcium. Lack of material prevented an investigation of the effect of chemical substances on the galvano- tropism of volvox. But it was found possible to change the pole at which the galvanic current stimulates by means of pressure. In all of my first experiments, which were made on the anodal colonies the coverglass was rested on but one thickness of filter paper so that the colonies were slightly pressed, and were thus pre- vented from moving. In all of these colonies it was uniformly seen that the cathodal flagella stopped contracting during the flow of the current. As the plants swam to the anode it was difficult to see how a stopping of the cathodal cilia could bring about this result, and much time was lost before it was discovered that the pressure just like the bright sunlight had changed the pole at which the current stimulates. What the pressure and the sunlight have in common is hard to see, unless it is that they both produce an intense stimulation and possibly a slight injury. DISCUSSION OF RESULTS It has been shown that the galvanotropic orientation of volvox is brought about by a cessation or great diminution in the stroke of the flagella at one pole of the organism. This diminution in activity of the flagella appears to be the only way in which volvox is capable of responding to stimuli. Nothing in the nature of a motor reflex has ever been observed in this organism so far as I know. ‘The flagella always strike most strongly backward. We have then the simplest possible kind of a mechanism for bring- ing about galvanotropic orientation. ‘The current diminishes the activity of the flagella at one pole of the colony and consequently the activity of the flagella at the other pole cause the organism to turn in that direction. We have here a tropism reduced to its lowest terms. ‘There is nothing of the nature of trial and error present at all. Galvanotro pic Orientation in Volvox 163 The fact that nothing of the nature of a motor reflex or reversal in the direction of the effective stroke of the flagella has been observed makes it very probable that, as indicated by Holmes, the heliotropic orientation is also brought about by differences in the strength of the stroke on the sides toward and away from the light. It would seem then that in the case of the orientation of volvox to light we have also a tropism pure and simple without any indication of orientation by trial and error. LITERATURE CITED Bancrort, F. W., ’05—Ueber die Giltigkeit des Pfligerschen Gesetzes fiir die galvanotropischen Reaktionen von Paramecium. Pfliger’s Archiv, vol. 107, pp. 535-556. °06—The Control of Galvanotropism in Paramecium by Chemical Sub- stances. Univ. of Cal. Pub., Physiol., vol. 3, pp. 21-31. ’o6a—On the Influence of the Relative Concentration of Calcium Ions on the Reversal of the Polar Effects of the Galvanic Current in Paramecium. Jour. Physiol., vol. xxxiv, no. 6, pp. 444-463. CaRLGREN, O., ’00—Ueber die Einwirkung des constanten galvanischen Stromes auf niedere Organismen. Engelmann’s Archiv, 1900, pp. 49-73. Homes, S. J., ’03—Phototaxis in Volvox. Biol. Bull., vol. 4, pp. 319-326. Logs, J., °05—On the Changes in the Nerve and Muscle which seem to Underlie the Electrotonic Effects of the Galvanic Current. Univ. of Cal. Pub. Physiol., vol. 3, no. 2, pp. 9-15. Terry, O. P., ’06—Galvanotropism of Volvox. Amer. Jour. Physiol., vol. xv, PP- 235 243- WALLENGREN, H., ’02—Zur Kenntniss der Galvanotaxis. I. Die anodische Gal- vanotaxis. Zeitschr. f. Allg. Physiol., Bd. 11, pp. 341-384. °o3—Zur Kenntniss der Galvanotaxis. II. Eine Analyse der Galvano- taxis bei Spirostomum. Zeitschr. f. Allg. Physiol., Bd. 11, pp. 516- 555: Se THE INFLUENCE OF EXTERNAL FACTORS, CHEM- feat AND PHYSICAL, ON THE DEVELOPMENT er PUNDULUS ‘HETEROCLITUS BY CHARLES R. STOCKARD With SEVENTEEN Ficures The following experiments have been undertaken to determine to what extent the form of the embryo and its manner of develop- ment might be modified by external influences. In a previous paper (’06), I have shown that lithium chlorid produces a definite effect on the development of Fundulus heteroclitus, as Herbst (92) had shown for the sea-urchin and Morgan (’03) for the frog. During the past summer I have been able to show that these abnormalities are not only definite but specific for the lithium ion In its action on this egg. It has also become desirable, owing to recent work on the sub- ject, to determine the permeability of the membrane of Fundulus eggs to the various salts; as well as to study the separate and com- bined effects of osmotic pressure and chemical actions on the development. The eggs of Fundulus, as has often been recorded, develop almost equally as well in sea-water, concentrated sea- water, fresh or distilled water, and even, as Morgan (’06) has recently mentioned, out of water.' Thus they furnish excellent material for a study of the actions of both hypertonic and hypo- tonic solutions. The experiments were performed at Wood’s Hole, Mass., while . occupying a table kindly furnished me by the Vassar Brothers’ 1Jn regard to the development of these eggs in fresh water, Loeb (’9q) states: ‘‘In fresh water the embryos hatch just as rapidly as in normal sea-water. The fish is able to live in fresh water.” On the contrary, I have shown (Stockard,’o6), that these eggs are always slower to hatch in fresh water and further that the newly hatched young soon die when left in this medium. Sumner (’06) has also shown that the adult fish is unable to survive in perfectly fresh water. Tue Journat or ExprriMeNntaL ZcOLoGy, VOL. Iv, No. 2. 166 Charles R. Stockard Institute of Poughkeepsie, N. Y. I wish to express my thanks to this Institution, as well as to the authorities of the Marine Biolog- ical Laboratory, for the working facilities furnished me while there. I am also glad to express my indebtedness to Prof. T. H. Morgan for many helpful suggestions, and to Prof. A. P. Mathews for assisting me in calculating the osmotic pressures of my solu- tions. METHOD AND MATERIAL As stated above the eggs of Fundulus heteroclitus lend them- selves peculiarly to such investigations as this. They are hardy and develop in different strengths of almost any solution applied. The action of a salt in sea-water solutions is generally more or less modified, owing to the presence of the salt constituents of the sea-water itself, but this effect can be controlled by the use of the distilled water solutions. The spawning season of this fish begins at Wood’s Hole about the middle of June and extends well into the first part of August, thus giving an opportunity to repeat the experiments many times and to test any uncertain points which might arise. In fertilizing the eggs | have found it advantageous to strip them from the female directly into a dry bowl saa then apply the milt from the male,stirring the eggs well so as to mix them thoroughly with the milt. ‘They are left to stand for five or ten minutes when water is added. ‘This method insures a larger percentage of fertilized eggs than will result if they are under water when the spermatozoa are applied. In those cases where the eggs were to be treated with solutions of salts in distilled water they were placed, after being fertilized “dry” directly in fresh water, thus little if any salts could have reached the eges after they had been removed from the body of the fish. The time elapsing between fertilization and hatching varies con- siderably with the temperature, season, etc., being from about eleven to eighteen days. ‘The salts used in these experiments were with few exceptions fresh Kahlbaum preparations. High percentage solutions were prepared and these were diluted to the proper strengths from time The Influence of External Factors on Development 167 to time as the experiments required. In discussing the experi- ments the percentage solutions are expressed in gram-molecular terms. The sugar solutions were prepared fresh for each experi- ment since they soon became acid witha fungus-like growthif kept for any length of time. Cane sugar inverts to some extent in solu- tions and may thus vitiate the calculations for osmotic pressures. It was found that the amount of solution and the number of eggs in the bowl affected to a greater or less degree the rate of development. A large number of eggs in a bowl almost full of liquid develop more slowly than fewer eggs in less liquid; this 1s due to a difference in oxygen supply as will be shown below. In the same experiment, therefore, approximately equal numbers of eggs and equal amounts of liquid were placed in each bowl. Since there is some individual variation in the eggs from different females the experiment and control were as far as possible from the same batch of eggs. PERMEABILITY OF THE EGG MEMBRANES A recent paper by Brown (’05), questioning the permeability of the egg membrane of Fundulus necessitates a discussion of this subject. The solution of this question is also essential in order to properly interpret my experiments. ‘ Loeb in 1893 showed that diffusion through the egg membrane of older embryos occurred very readily. ‘The addition of 3 grams of KCl to 100 cc. of sea-water brought the heart of a Fundulus heteroclitus embryo to a standstill in a few minutes. A consider- able amount of the salt must, therefore, diffuse through the egg membrane in a very short time. I repeated this experiment in the following manner: Seven five-day embryos were placed in a 0.67m (about 5 per cent) solution of KCl. Within ten minutes the heart action of one had ceased, another stopped in eleven minutes, and three others in twelve minutes. The heart’s action becomes first periodic and jerky and then gradually stops, though it will often continue to give weak irregular contractions at intervals of one or two. minutes for some time after it has apparently stopped. The embryos began to wriggle after having been in the solution only three or four minutes. When the hearts of all the individuals 168 Charles R. Stockard had stopped beaung, within fifteen minutes, the eggs were returned to sea-water; thirty-five minutes later one heart was contracting almost normally, while another was beginning feebly, the others had not recovered even after an hour. On examining the seven embryos the following morning all had entirely recovered. The result demonstrates the readiness with which salts permeate the membrane in eggs a few days old. In 1903 Brown recorded the results of experiments, showing the immunity of Fundulus eggs and embryos to electrical stimulation. For these experiments he used eggs at various stages but some were tried when in the two-cell stage. These experiments, as Brown concluded, go to show the permeability of the membrane during the first hours of development. He states that the most probable explanation of the immunity of these eggs to electric currents as well as to osmotic changes of the medium in which they live is that the membranes of the egg are so freely permeable to ions and possibly to neutral particles that no polarization can occur. “There is a gradual increase in susceptibility to osmotic changes and to the electric current as the embryo develops, the adult being readily stimulated by the current from a single cell, which is quite without action in the embryo.”’ With the above results and interpretations in view, Brown (’05) has since, from far less convincing experiments, arrived at opposite conclusions. He claims now that the membrane of Fundulus eggs is practically impermeable to salts and water during the first six or eight hours of development; since eggs placed in distilled water do not lose their salts during that period. It would be sur- prising if these eggs did lose their salts in distilled water as they are capable of normal development in this medium. Very probably the inorganic salts of this egg are held in combination in the proto- plasm so that they are not able to diffuse out in hypotonic solutions and the readiness with which the membrane is penetrated makes the osmotic pressure low. The fact that the conductivity of the distilled water containing the eggs increased after the first eight hours is probably due to an excretion from the eggs. They undoubtedly give off some waste products as an odor is often ob- served when a bowl containing eggs is uncovered after standing overnight. T he Influence of External Factors on Development 169 The extensive treatment of Fundulus eggs with salt solutions which Mathews (’04)? has recorded goes to show that the egg membrane is easily permeable. To demonstrate further the permeability of the membrane, during the first hours of development, I carried out the following experiment. Since embryos had been found to be affected in a definite manner by solutions of LiCl below the strength of +m, I determined to subject them for short periods while in the two-cell ‘stage to strong solutions of LiCl. Eggs were placed in a molecu- lar and a double molecular solution of this salt in distilled water; one hour and ten minutes later they were all quite abnormal, showing the lithium effect. Some of these were then transferred to sea-water and on examination, eleven hours later, still showed the lithium abnormalities. “Those left in the LiCl solution were all badly plasmolized or shrunken while the blastodermic cap was heaped up upon the top of the yolk, almost pinching away from it. Some eggs were removed from the double molecular solution after staying two and one-half hours in it, these failed to recover, and were in the same condition after sixteen hours as those still in the solution. “Those removed from the molecular solution after one hour were in the following condition after forty hours; many were dead but some had recovered and showed the embryonic thicken- ing forming on the egg. Of those that spent two and one-half hours in the solution one or two were still living though abnormal, the germ ring having descended only one half of the way down the yolk, and in one case an embryonic shield had formed. Those taken from the double molecular solution after one hour were, forty hours later, almost all dead, the few living ones being very abnormal. ‘Those that remained two and one-half hours in this solution were all dead. * Mathews tried with these eggs to ascertain the relation if any between the properties of the elements and their physiological action. He concluded that the poisonous action of any cation or metal upon the eggs varied inversely with the solution tension. ‘Those ions with a very low solution tension are very poisonous; those with a high tension are relatively inert. The poisonous action of any anion also follows this rule.” Further “there is an inverse relationship between atomic volume and poisonous action; and a direct relationship between equivalent weights and poisonous action. Poisonous action of the metals is a periodic function of their atomic weights. Elements which have a low atomic volume and high equivalent weight, as mercury, are more active than those with a high atomic volume and a low equiva- lent weight, as sodium.” Many exceptions to the foregoing were found. 170 Charles R. Stockard After seventy. hours those that spent one hour in the molecular solution produced some living embryos, and though well formed they were slower than the control in their rate of development. Those that remained in this solution for two hours showed only a few living eggs with badly stunted embryos. Of those from the double molecular solution of LiCl, after one hour, only one in twenty was alive, and this one was stunted. Solutions of double molecular and one and one-half molecular strengths were prepared with sea-water and an experiment similar to the above was conducted with like results, although the eggs recovered somewhat more readily after being removed from these solutions and returned to pure sea-water. This latter fact may indicate that some of the salts of the sea-water tend to counteract in part the effects of the LiCl. At first sight the above results seem to contradict my former statement to the effect that eggs removed from Li solutions in three, four or five hours showed no toxic effects in their later devel- opment. It is recalled, however, that the solutions then used were weak ones, while the above are strong enough to kill all eggs remaining in them. Such results as these can leave, I think, no doubt as to the fact that the membrane of Fundulus heteroclitus eggs is readily per- meable to salts during the first few hours of their development. The permeability of this membrane at later stages is also beyond question, and probably it becomes more readily permeable as development advances. THE DEVELOPMENT OF FUNDULUS EGGS OUT OF WATER The only reference hitherto made to the development of Fun- dulus eggs out of water is that by Morgan (’06), to the effect that these eggs will develop on a glass plate in a moist atmosphere. I undertook to rear embryos in this fashion to ascertain what abnormalities, if any, would result from such treatment. The most interesting result obtained was that although these fish develop to all appearances in a perfectly normal manner, except at a little faster rate, they are entirely unable to hatch while on the ‘moist plates. Eggs were kept from June 30 until August 2 in a The Influence of External Factors on Development 171 healthy condition without the embryos breaking through the membranes. Thus these fish remained enclosed in their egg membrane for thirty-three days while the control had begun to hatch after thirteen days. As there are some points of interest to be brought out in con- nection with the details of this experiment it may be briefly described. Eggs were placed, shortly after fertilization, on moist glass with all superfluous water removed, they were arranged so as not to be in contact with one another and then covered with a finger bowl to prevent the evaporation of the surrounding moisture. Other eggs were arranged in a similar manner when in the two-cell stage, while still others at this period were covered by finger bowls which had moist filter paper closely pressed in them, thus insuring a more moist atmosphere about the eggs. When the eggs were fifty-three hours old the control in sea- water showed embryos distinctly formed with the blastopores closed, those on the moist glass also had their blastopores closed and were slightly in advance of the control in their development. One of the eggs that had become dried and shrunken also showed anormalembryo. At three, four, six and ten days old the ones on the moist glass were continuing to develop normally though at a faster rate than the control. This more rapid development was in all probability due to the better aération out of water. When fourteen days old many of the control had hatched. At this time those that were under finger bowls without moist paper had become so dry, although moisture had been added several times, that a number of them died; in one lot fifty-four were dead and only twenty still alive, while another lot had forty-seven living and thirty-four dead. ‘This and other such cases indicate that it is important to keep the eggs moist. “The ones supplied with moist filter paper were all alive. None of these eggs out of water had hatched; twenty-three were then taken from the glass plate and _put into a finger bowl containing sea-water. Eleven hatched; in five minutes, eighteen were out after ten minutes and the entire lot, twenty-three, were swimming about after being in the sea-water for only eighteen minutes. These fish after hatching seemed further developed than the controls which had also hatched, evi= We Charles R. Stockard dently they had been prevented from hatching owing to their inability to break through the egg membrane when out of water. On the fifteenth day none on the moist glasses had hatched, with the exception of two which I feel sure came out on account of too much moisture or water having been put about those that were rather dry on the day previous. Thirty eggs were now put into sea-water, three of which hatched after ten minutes, thirteen were out in fifteen minutes, twenty-five in twenty minutes, while twenty- seven of the thirty eggs had hatched within twenty-five minutes. On hatching, the embryos swim directly to the top of the water and to the side of the bowl nearest the brighest light thus showing a negatively geotropic and a positively heliotropic reaction. At sixteen days old not an egg on the moist glasses had hatched, with the exception of the two which had hatched three days before. Twenty eggs were now placed in sea-water, the first one came out after ten minutes. It seems as a rule to require about ten minutes for the eggs to begin hatching and after this period they come out very rapidly. Fifteen were out after fifteen minutes and all were hatched after being in the water only twenty-seven minutes. On the morning of the seventeenth day two of those on the moist glass had hatched, but again an extra amount of moisture had been put about the eggs on the previous day. When twenty days old, six days after hatching had begun, all of the controls were hatched, those on the moist glasses were alive and well, though none had hatched. An excess of moisture was added to them while still on the glass and twelve embryos came out within twenty minutes, the moisture was then drawn off and hatching ceased. When twenty-six days old a few of the eggs began to die, three or four out of fifty died on one glass. ‘This was probably due to want of food as at this time all of the yolk had been absorbed for several days past as seen by comparing Fig. 1 of a newly hatched control embryo twelve days old, with Fig. 2 of a moist glass embryo hatched when eighteen days old (yk the yolk mass). Acomparison of these two figures will also show how the embryos on the moist plates have continued to grow and develop within the egg mem- brane. When eighteen days old one of these was equally as large The Influence of External Factors on Development 173 as an eighteen day control embryo which had been free swimming for four or five days. When twenty-seven days old, more embryos were dying, but those put into sea-water began to hatch within ten minutes, though they swam abnormally at first, going in a circular or spiral course. This was due to the fact that the fish had become cramped by its twisted position within the egg membrane and for an hour or so after hatching they were unable to straighten themselves, but they finally do so and their movements become normal. When at rest the crooked ones have a tendency to topple over on one side. Tonk \ ON : S TS SKS Ligokey A fy Man MMMM hhh Ld beige 2 ye Fig. 1. A normal embryo just hatched, twelve days after fertilization. yk, yolk mass. X17 diameters. Fig. 2 A newly hatched embryo which was developed out of water and made to hatch by being placed in sea-water when eighteen days old. yk, yolk mass. X 17 diameters. After thirty-one days they were still well and unhatched on the glass although a few had starved to death each day. When thirty- two days old, five were placed in sea-water, three came out within twelve minutes and all were out in twenty-three minutes after being put into the water. These had their bodies bent owing to the long cramped position they occupied within the egg membrane, though all soon straightened out and swam normally. 174 Charles R. Stockard THE OSMOTIC EFECTS RESULTING FROM DEVELOPMENT IN SUGAR SOLUTIONS In order to be able to discriminate between the osmotic effects and the effects of salt solutions it was necessary to make a careful study of the manner in which the eggs reacted to the physical changes; 1.e., to the osmotic pressures of the solutions. Loeb (°94) had found that these eggs are “remarkably independent of the concentration of the sea-water.’’ He found that the embryos develop in a perfectly normal fashion in fresh water as well as sea-water to which five grams of NaCl had been added to each 100 cc. of water. When seven and one-half grams of NaCl were added to each 100 cc. of sea-water a blastoderm was still formed, but rarely an embryo. When twenty grams of NaCl were added to 100 cc. of sea-water the power of development of freshly fertilized eggs was annihilated within three or four hours. The only abnor- fnohoes that Loeb observed when eggs were in concentrated sea- water were a shrinking of the yolks and a slower rate of develop- ment. These experiments fail to show whether the effect is osmotic or chemical, since by using NaCl one is dealing with a chemical poison as well as increasing the osmotic pressure of the solution. I have used sugar solutions prepared both in fresh and sea- water to determine the effects of osmotic pressure on these eggs. The sugar itself is supposed to exert slight if any chemical action. The eggs of Fundulus, since they develop normally i in fresh water, as ero by Loeb (? 02), Mathews (704), the writer (06), and others, are evidently insensitive to the lowering of the osmotic pressure of their surrounding medium. This is due probably in part to the inorganic salts of this egg being held firmly in combina- tion with its colloidal organic compounds in the protoplasm. Seven independent experiments were carried out with sugar solutions in distilled water of the following strengths: 1.53 m, 1.33 m, 1.0 m, 0.66 m, 0.5 m, 0.33 m, and 0.163 m. Solutions of sugar 1n sea-water were used of the strengths 1.33 m, I.0 m, 0.88 m, 0.66 m, 0.5 m, 0.33 m,ando.16m. A molecular solution of sugar is equivalent to about a 31.4 per cent solution and exerts an The Influence of External Factors on Development 175 osmotic pressure of 22.4 atmospheres. ‘The osmotic pressure of the sea-water at Wood’s Hole, as determined by Garrey (’05) is 21.918 atmospheres, thus being nearly equivalent to a gram mole- cular solution of sugar. DISTILLED WATER SOLUTIONS OF SUGAR Several of the sugar solutions employed exert an osmotic pres- sure lower than that of sea-water, yet owing to their high specific gravity the eggs float on them. We may first consider only the effects induced by the 1.53 m, 1.33 m and the 1.0 m solutions. ‘Two hours after being in the 1.53 m and the 1.33 m solutions the yolks have shrunken in a peculiar way, being circular in outline when viewed from above and oval when seen from the side, having such a form as a plastic sphere would assume if pressed between two horizontal planes. “In the molecular solution the yolk does not show this effect to any considerable degree and the rate of develop- ment is slightly ahead of the control. After about twenty-three hours all of the eggs in the 1.53 m sugar solution are dead, the membranes of many have burst and allowed theyolk to stream out, while in others the yolk is a small contracted and concentrated mass. At this time many eggs in the 1.33 m are also dead, although the few still alive are further advanced than the control, the blastoderm lies flat on the yolk andthe germ-ring is further down. Those twenty-three hours old in the molecular solution are alive and the germ-ring is further over the yolk than in the con- trol. The osmotic pressure of the molecular solution should be about 22 atmospheres which is almost equivalent to the pressure of sea-water, that of the 1.33 m is 26.2 atmospheres, only four atmos- pheres higher; while the 1.53 m has a pressure of 34.278 atmos- pheres or about twelve atmospheres above the pressure of sea- water. Eggs in the latter solution are seen to be fatally affected within twenty-four hours. We meet here with the same peculiar problem that I findin looking over Madame Rondeau-Luzeau (’02) and Morgan’s (’06) results. “They found the upper limit of NaCl on frogs’ eggs to be about 2 per cent, which exerts a pressure of 13.61 atmospheres, while Morgan found the upper limit for cane sugar to be 12 percent, with a pressure of 8.376 atmospheres. I have found 176 Charles R. Stockard as shown belowthat Fundulus embryos develop in a } msolution of MgCl, in sea-water. Here we have a pressure of about thirteen atmospheres above that of sea-water, while as indicated above a pressure of twelve atmospheres more than sea-water is fatal within twenty-four hours when exerted by a cane sugar solution. * A pos- sible explanation of such results is that some of the cane sugar becomes inverted and, therefore, exerts a pressure greater than that estimated, since two molecules are now present for each one calculated. The difference in activity which Jenkinson (’06) has lately recorded between isotonic solutions of cane sugar and dex- trose on the frog’s egg may also be due to the cane sugar having become partially inverted and, therefore, his solutions may not have been isotonic, as he thought; at least this explanation seems just as probable as the one he advances, that the membranes dif- fer in their permeability to the two sugars.° The augmentation of the effect when salts are added to sugar solutions that Morgan (’06) has recorded for the frog and as I shall show below for a fish egg may be largely due to the increased pressure or to the peculiar injurious effect caused by some action of the cane sugar. ‘he sugar used in most of my experiments was crystallized ‘ oc candy,” and was probably pure. In those solutions of sugar which exert a pressure lower than that of the sea-water the embryos develop normally and often at a rate faster than that of the control. ‘This acceleration probably is due to their floating and hence being better aérated. Another point of interest regarding eggs in these solutions 1s that the yolks often become swollen as would be expected in hypotonic solutions. This observation has never been recorded for these eggs in fresh water nor in any distilled water salt-solutions which were hypo- tonic to sea water. I have treated them with numbers of such solutions, always making careful study of the structural or form changes which resulted, but in no case have | observed a swelling of the yolk except in some, not all, of these hypotonic sugar solu- tions. ‘This swelling may be due to the sugar becoming inverted 3 This spring I have had an opportunity to compare the action of cane sugar with simple sugars, such as glucose and levulose, on the frog’s egg, and find that the effects of the cane sugar can be explained without assuming any inversion to take place. ae The Influence of External Factors on Development 177 after penetrating the membrane. If such should occur, the con- centration of the sugar solution within the membrane would be higher than that without and this might produce a strong en- dosmosis which would result in the swelling of the yolk. The fact that the yolk or egg does not swell in fresh water or in hypo- tonic solutions has been used as an argument to show that they are immune to osmotic effects. That they are not, however, entirely immune to such effects has been shown above. THE SEA-WATER SOLUTIONS OF SUGAR The yolks shrink in these solutions in the manner mentioned above, often within one hour after having been put into the solu- tion. The shrinkage of the yolk occurs so promptly that the outer membrane fits loosely around the egg, and often shows an indenta- tion on one side. There occurs below the blastoderm in many cases a bubble-like appearance which disappears, however, as de- velopment progresses. When about forty hours old many eggs die in the 1.33 m solution, which exerts a pressure of about twenty-six atmospheres more than ordinary sea-water. The dead eggs usually have a polar ball of protoplasm on the shrunken yolk. The living embryos have the tail end indistinctly indicated and the blasto- pore remains open much longer than is usually the case. If eggs were removed from these solutions and put into sea-water at any ‘ time before twenty hours they soon recovered and developed nor- mally with the exception of those from the stronger solutions, in which case the yolk rarely recovered its full size, and in consequence _the pericardium seemed abnormally large. As development con- tinues in the stronger solutions, the yolks become smaller and smaller, and the embryos are likewise much dwarfed with very weak heart contractions which begin only some time after the con- trol embryos have etsablished a free circulation. Fig. 3 shows many of these characters in an embryo when five days oldina mole- cular solution of sugar in sea-water. Fig. 4 shows the condition of a control embryo of the same age. The body of the embryo 1s often abnormally bent or twisted on the yolk. The sluggish circu- lation at times allows the blood to accumulate in certain vessels, commonly in the veins along the ventral line of the tail, as large 178 Charles R. Stockard red spots. When thirteen or fourteen days old the embryos in the 1.33 m sugar solution resemble the normal embryos of only four or five days. The eggs returned to sea-water after being twenty hours in this solution begin hatching when fourteen days old. Only in the 0.163 m solution of sugar in sea-water were the embryos observed to hatch. The little fish seemed almost normal although they swam peculiarly as their yolks seemed too large and heavy for them to carry. Fig.3 An embryo from a 0.88 m (30 per cent) sea-water sugar solution five days after fertilization, showing the greatly contracted yolk and the dwarfed condition. X 17} diameters. Fig. 4 A normal control embryo of thesame age. X 17} diameters. The above effects are those expected to result from an increased osmotic pressure, and they go to show that although these eggs are without doubt very resistant to such pressure, it nevertheless exerts an influence on their development. It will be noted by comparing the effects of the sugar solutions in distilled water with those in sea- water that a pressure more than double as strong in sea-water pro- duces a much less injurious effect on the eggs. A not impossible explanation of this peculiar fact is that sugar in the fresh water solutions becomes inverted much more readily than in sea-water.! The fresh water solutions of sugar were found to show an acid reaction within a day or two after their preparation, and this acid condition would cause the inversion of the cane sugar. On the * Dr. S. P. Bebee, of the Cornell University Medical College, has analyzed sugar solutions for me with the following results: A 15 per cent distilled water solution of cane sugar becomes partly inverted if allowed to stand for a day or so in the laboratory, while a 15 per cent sea-water solution kept under similar conditions showed no trace of inversion after three days. The Influence of External Factors on Development 179 other hand sea-water is neutral and the sugar would be much less disposed to invert under such a condition. A fungus-like growth often attacks eggs in fresh water, while I have never observed any such growth about eggs that were immersed in sea-water. ‘lhe fungus grows in a slightly acid medium and 1s often present in fresh water solutions of sugar. ‘The acid condition of a solution is in itself injurious. Owing to these conditions a frequent change of the sugar solutions was necessary. Solutions of glucose and glycerine were also used to test osmotic effects but these chemicals proved to be impure. THE SPECIFIC CHARACTER OF THE LITHIUM EMBRYO In my paper on the devlopment of Fundulus in solutions of LiCl, it was only possible to state at that time that lithium induced certain definite effects which were characteristic of this salt’s action. It was also suggested that these effects might be specific for lithium but such could only be known after a eae of other lithium salts, as well as a number of the salts of other metals had been used. I have since employed many other metallic salts and have found none of them to produce,with any constancy the abnormalities in development which result from the use of lithium salts.’ An experiment was first carried out in which eggs were subjected to solutions of LiNO, and Li,SO, to ascertain whether the resulting development would be similar to that found in LiCl solutions. Eggs were pees soon Ber fertiliZation in distilled water solu- tions of LiNO, + m and ss m, and Li,SO, 7s m and #5 m. The first four or five hours of development was Baa normal, but after eight hours those in LiNO, 2's m were beginning to send a projection of the periblast down into the yolk substance, a contin- uation of this process results in the large bubble-like segmentation cavity before described (’06). When twenty-two hours old the germ-ring is just below the equator in the control (Fig. 9). Those ° Madame Rondeau-Luzeau (’o2) and Morgan (’06) have found that the upper limit of LiCl that the frog embryo can stand is a 0.65 per cent solution, the osmotic pressure of which is 5.161 atmospheres, while the upper limit of NaCl is about 2 per cent and exerts a pressure of 13.61 atmospheres which is more than double that of the lithium solution. This comparison shows that the effects of the lithium salt are not due to its osmotic pressure alone. 180 Charles R. Stockard in LiNO, § ma few show the blastoderm as a polar ball with large bubble-like segmentation cavity (Fig. 5). Many eggs, however, have polar caps with peripheral germ-rings and a small embryonic shield just beginning to form (Fig. 6). A very few are more nearly normal with the germ-ring extending one-third over the yolk sphere. ‘This is the same condition found in eggs of this age in solutions of LiCl, which gives for the first few hours no evident effect, then retards the development, preventing the downgrowth of the germ-ring, and often causing the formation of polar proto- plasmic balls with the bubbles beneath, shown in Figs. 1, 2, 3 and 4 in my earlier paper (’06). It might be objected that at this stage of development there are only a limited number of ways that the eggs could be affected. This may be granted, but from what is recorded in other sections of this article it will be seen that there are several possible modifications that may appear, and no other substances have given, with any degree of constancy, the above modifications. Those in the +s m LiNO, solutions were in a simi- lar condition. After twenty-three hours in L1,SO, 7's; m most of the eggs are dead with polar balls of protoplasm. Of those still alive, most have polar caps with bubbles beneath (Fig. 7). Others have polar caps with peripheral germ rings and embryonic shields just forming (Fig. 8); a very few are more nearly normal and show the germ-ring one-fourth the way down the yolk. Li,SO, 3 m has caused similar though less pronounced effects. The control eggs when forty-six hours old show the embryos distinctly marked out with optic vesicles and lenses visible. In the LiNO, % m at this time a few eggs are dead with polar caps; their blastoderms showing a bubble-like appearance beneath. Many have embryonic thickenings forming in the polar caps simi- lar to the condition shown in my former Fig. 18. A few have their caps extending halfway over the yolk with short embryos formed, a few others have the blastopores almost closed although the head end of the embryo is abnormal. In LiNO, ss m many eggs are dead, a few have their blastopores closed though the head end of the embryo is abnormal with no eyes showing. ‘There are others with the blastoderms only one- -half over the yolk and form- The Influence of External Factors on Development 181 ing short embryos. In Li,SO, i's m some are dead, others have a polar cap with an embryo forming init. Of those in Li,SO, sc m some are dead, others are with the blastopore almost closed, but the embryos show no optic vesicles as yet, a few have polar caps with embryonic thickenings in them. | When seventy-two hours old, the eggs in LiNO, 5 m have some embryos almost as long as those of the control but with no circula- Fig.5 An egg from a 1 m LiNOs solution twenty-three hours after fertilization. _ sc, segmentation Cavity; od, oil drops. Fig.6 From a 3; m LiNOs solution when twenty-three and one-half hours old. Fig.7 From a =); m LigSQ, solution at twenty-three hours old sc, segmentation cavity. Fig.8 From a ;'; m LigSQx, solution at twenty-three and one-half hours old. Fig.9 A control egg of twenty-three hours. All magnified 17} dia. tion apparent; the heads are poorly formed and show no optic vesicles in surface view. In the LiNO, 3's m solutions all of the eggs are dead. In Li,SO, 1's m some have the germ-rings only one-half way over the yolk, with short embryos formed; in others development has stopped. In the Li,SO, sc m solution most of the eggs are dead though a few have badly twisted embryos with poorly formed heads and no optic vesicles. 182 Charles R. Stockard At ninety-six hours old, the LiNO, 3 m embryos show no eyes, the circulation of the blood can not be detected, while in the control it isvery distinct. ‘he pigmentspots are scarce. The Li,SO,7;m eggs have short embryos with their blastopores still open. When eight days old, those eggs in LINO, 5 m have short em- bryos with poorly formed eyes, they are pale in appearance, the heart beats slow the blood is colorless, and the tail unusually bent. Comparing this general description with the detailed one recorded for the development of this fish in solutions of LiCl it will be found that the development of the egg of Fundulus is as character- istic in solutions of lithium salts as is that of the frog under like conditions as recorded by Gurwitsch (’95, 96), Morgan (’03, ’06) and others. THE EFFECTS OF METALLIC €CHLORIDS ON THE DEVELOPMENT OF FUNDULUS EGGS A number of chemical solutions have been employed singly and mixed in order to further analyze osmotic and chemical action, as well as to distinguish if possible any definite morphological response that fers result from the action of any one salt. The notes on these experiments have become so voluminous that it is inadvisable to attempt to record them all. I shall, therefore, state as concisely as possible the factors involved and the chief results that followed. Loeb’s ('93) experiments with KCl were repeated:’ the con- centrations of the solutions used being } m, 3m and 3m in dis- tilled water and } m,? m, and molecular in sea-water. Many of the eggs in the stronger solutions died during their early develop- ment. ®° Loeb observed the interesting fact that Fundulus embryos would develop in solutions of KCl without circulation of the blood taking place. The heart was entirely still and the blood failed in consequence to move through the vessels. He stated that in these cases the blood system developed normally, the only peculiar point being that the pigment spots did not migrate to the blood vessels and arrange themselves along them as they usually do. I find that the circulatory system does develop to some extent but by no means normally as may be seen by a casual examination of the heart. Hence possibly the failure of the pigment cells to migrate. Loeb also found that embryos four to six days old were killed by remaining one hour in a 1.5 per cent solution of KCI as a result of the effect of this salt upon their heart’s action, while if put into a 5 per cent solution after fertilization they live and develop. The Influence of External Factors on Development 183 o) Solutions of KCI in distilled or sea-water caused the eggs to develop at a slow rate, the yolks were shrunken and bubble-like appearances were often seen below the embryonic shield. When studied in section these eggs showed the following conditions: The periblast beneath the embryonic shield had become vacuolated with huge cavities in it as shown in Fig. ro. It thus bulged into the yolk mass and the cavities produce the bubble-like appearance (erg as Paese Fig. 10 A section of the embryonic shield from an egg thirty hours old in a ? m sea-water solution of KCl, showing huge vacuoles v in the periblast beneath the shield. 584 diameters, Fig. 11 A section of the blastoderm of an egg thirty-one hours old that had spent the first twenty- three hours after fertilization in a 7’; m MnClz solution. The central periblast, cpb shows much thickened, with many large nuclei accumulated in this region. c,Cavityinthe blastoderm. X 58} dia- meters, seen in living eggs. The embryos are always much dwarfed and pale. The heart never contracts although the embryo may remain alive for as long a period as two weeks. The pericardium is often puffed out and is unusually prominent, as also occurs in some other solutions, as shown in Fig. 12, pe for an embryo from a mixture of MgCl, and NaCl. 184 Charles R. Stockard Eggs that remained as long as thirty hours in 4 m distilled water solutions of KCl would recover if placed in sea-water. Other eggs were left for three days in KCl 2 m distilled water solutions, and afterward recovered, the heart beginning to beat, etc., when returned to sea-water. Normal embryos several days old were very readily killed if subjected to even weak solutions of KCI; their heart’s action being stopped. It thus seems as though an embryo may live and develop without its heart ever having con- tracted, but if it had once begun to contract any cause that may stop this contraction proves fatal. My results then in a general way agree with Loeb’s observations though I should take exception to his statement that the circulatory system develops normally even though the blood does not circu- late. ‘The major parts of the system do seem to develop but by no means nor- mally, the heart being small and weak and it is often only a straight tube with the balloon-like pericardium surround- ing it. Many clots of red corpuscles are noted in several of the sinuses. The above facts are also of interest in connection with Howell’s analysis of Fig. 12 An embryo from amixed _ the inhibitory action of the vagus nerve solutionof MgCl $m + NaCl5m on the heart beat as being due to the when forty-three hours old, showing : : : liberation of K-ions about the nerve endings. A mixture of a} m KCl and a molecular NaCl solution was prepared with 60 cc. of the former to 10 cc. of the latter. This mixture showed the same general effect on the eggs as the simple KCI solution; the blastoderm bulged up slightly and the yolks were shrunken. Many embryos, however, seemed stronger and better developed with more pigmentation and with larger red ewollen, pc, pericardium. blood clots. Some of these embryos were placed in sea-water — when seven days old but failed to recover. A CaCl, + m solution in distilled water proved highly toxic. The blastoderms flattened down, the cells apparently spreading unusually far apart. The eggs died within about twenty-four The Influence of External Factors on Development 185 hours. Eggs that were subjected to { m CaCl, one hour after fertilization were almost all dead within four hours, the living ones were abnormal, and all died after twenty-four hours. ‘This result further illustrates the readiness with which the egg membrane is penetrated during the first few hours after fertilization. A mixture of 60 cc. } m CaCl, and 10 cc. 1.0 m NaCl proved equally as fatal as the CaCl, alone had done, neither of the cations seem to exert an anti-toxic action toward the other. Solutions of NH,Cl in distilled water of concentrations ;'5 m, +m, +%m,im,4m and 4m were used; a molecular solution of NH,Cl is equivalent to about a 5.05 per cent solution. Sea-water solutions of } m, } m and ? m were also employed. ‘These solu- tions seemed to cause the yolk to shrink slightly, the blastoderm to thicken so that on examining the living eggs one would think that the segmentation cavity was abnormally large as it is in the lithium embryos. On studying sections of these, it was found that the cells are loosely connected, making the blastoderm unusually thick so that it projects down into the yolk. The segmentation cavity is, therefore, not abnormally large as in the lithium embryos. Many of the eggs die at various stages. he rate of development is retarded and the blastopore is slow to close.. Many of the embryos are short with their tail ending abruptly. In some embryos the heart beats slowly and the circulation is sluggish; in others there is no pulsation at all, and still others in the same solution may show a very good circulation. Embryos lived as long as eighteen days in such solutions but failed to hatch. Such short embryos as those above described seem to result from any cause that retards development and prevents the nor- mal down-growth of the germ-ring. When such embryos were removed from 4 m NH,CI solutions when forty-three hours old and placed in sea-water, they recovered in one day and hatched when fourteen days old. The embryos in NH,CI are always dwarfed, with poor circulation, lightly colored blood, and sparse pigmentation, having a pale appearance. he sea-water solutions of NH,Cl were much less toxic than the distilled water ones. Mixtures of NH,Cl + m + MnCl, s+ m, NH, Cl ¥m + MnCl, +s m and NH,Cl + m + MnCl, so m were tried. ‘The 186 Charles R. Stockard eggs lived better in these solutions than in either the NH,Cl or the MnCl,. The first twenty-four hours of development is almost normal though some eggs die, the rate of development after this period becomes retarded, the embryos have swollen pericardia in some cases, and are pale and small. Some of them continued to live in these solutions for fifteen days but were far from the hatching stage at this ttme. ‘The weaker toxicity of these mixtures when compared with the action of the salts used singly may be due as Loeb (’02) has claimed, to an anti-toxic effect of one ion on the other. The mixtures with less NH,C! were always less active. Loeb found the bivalent cations to show an anti-toxic action toward the monovalent ones. MnCl, solutions were used of the strengths »'5 m, 3's m, and 7m in distilled water; and 3. m, 7's m, and § m in sea-water. Eggs that were subjected to the action of such solutions responded in the following way: ‘Those in the distilled water solutions go normally for several hours, then when about eighteen or twenty hours old the blastoderm shows a dark central portion when viewed from above, while in side view it shows that the dark area protrudes downward into the yolk. A section of such a blastoderm is seen in Fig. 11. The dark area is shown to result from an unusual thickening in the center of the central periblast, cpb, and the accumulation at this point of a number of the large periblast nuclei. A slight cavity, c, is not uncommon near the surface of the blasto- derm. ‘This unusual thickening of the periblast seems to render difhcult the subsequent descent of the germ-ring, and development is thus slightly retarded. When about forty hours old many embryos have their germ-rings only one-half over the yolks, and a short embryo is outlined on the embryonic shield. Many of the eggs died in these solutions. “The embryos have a feeble pulse and the blood is often clotted in some of the larger vessels. When fifteen days old embryos hatched in the =’; m solutions but swam abnormally; one embryo was seen to hatch in a 75 m MnCl, solu- — tion but it was entirely unable to swim. The solutions of MnCl, in sea-water formed slight precipitates and the results are thus no doubt vitiated to some extent, neverthe- less the dark central portion of the blastoderm always showed. The Influence of External Factors on Dew lopment 187 Short embryos with open blastopores were formed in many cases. The heart was weak and tubular with feeble contractions and was surrounded by a swollen pericardium. One embryo when eleven and one-half days old hatched in a »'; m solution, but was unable to swim. In many of these embryos no heart beat could be detected and the yolks were badly shrunken. In one case a one-eyed embryo was noted, this is mentioned on account of the tendency of magnesium salts to produce such a condition, but the eye structures of this embryo were very imperfect and no lens was present, this condition will be found to differ entirely from that described below as caused by the action of MgCl. Eggs were subjected to MgCl, solutions of the following strengths 75 m, } m, + m, and }m in distilled water, and 0.238 m, 0.25 m, 0.286 m, 0.33 mand 0.5 m in sea-water; a molecular solution of MgCl,.6H.O being equivalent to about a 20.3 per cent solution. The early development in all of these solutions is strikingly normal considering the Jarge death rate which occurs during these stages. The salt seems especially toxic to the early embryo. At seventy- four hours someembryos are well formed, though behind the con- trol in their development, and the blood Shen arene is slow in some while others have a quick heart action. When ten days old all are weak and smaller than the control, the blood flow is slow and spas- modic; in some embryos the circulation has ceasedand the blood is collected in the sinus and heart and appears as a red streak in front of the head. Many of the livelier embryos wave their pectoral fins. In the { m and $ m distilled water solutions many embryos hatch when about fifteen days old, though they swim abnormally on account of their bodies being twisted. ‘The sea-water solutions cause the yolks to shrink and in these the embryos are also small with sluggish circulations. Although kept alive for twenty-four days none of the eggs in the sea-water solutions would hatch. The conditions cited above are general and occurred also in a number of different salt solutions, chee the condition which may now be considered seems peculiarly characteristic of the Mg salt. In the }m sea-water solutions one-eyed embryos occurred with sur- prising regularity in 50 per cent of the eggs. ‘This experiment 188 Charles R. Stockard was repeated three times and each time it so happened that exactly one-half of the embryos had only one eye. ‘These cyclopean fish were rather abnormally shaped though they were able to twist about and wave their pectoral fins vigorously. ‘The other embryos were apparently normal in all particulars, the magnesium seeming not to have affected them. In sections the one-eyed condition was found to result from the union or fusion of the Anlagen of the two optic vesicles. Cases were found illustrating various degrees in this fusion, it seemed as though the optic vesicles were formed too far forward and ventral and thus their antero-ventro-median surfaces fused. ‘This condi- tion results in one large optic vesicle which in all cases gives more or less evidence of its fused or double nature. As a rule but a single lens is formed, the size of which depends upon the size of the optic cup or more exactly upon the size of the ectodermalarea influenced by the optic cup toforma lens. Thislens formation is interesting in connection with the results of the experi- mental work of Lewis (’04) and otherson the lens development in Amphibians. I (07) have entered into a more detailed discussion of this subject elsewhere. The lens was found to show a double or fused structure in one case out of the ten embryos that were sectioned; the other portions of this eye were also more distinctly double than was usually the case. This condition represents the last step in the fusion of the two eyes, slightly greater fusion would result in a single eye. With no other solution has such a condition as the above been procured, and its abundant occurrence in sea-water solutions of MgCl, strongly indicates that this one-eyed condition is character- istic of the action of such solutions on the developing Fundulus embryo. ; Solutions of MgCl, 75 m + NaCl} m in distilled water, and MgCl, } m + NaCl 4m in sea-water were tried on the eggs with the results following: The distilled water mixture produced no effect on the develop- ment, nor do such strengths of the two salts employed separately. The sea-water mixture contained twice as much MgCl, as the distilled water one. The results are instructive. When eighteen T he Influence of External Factors on Development 189 hours old the blastoderms were raised up prominently on the yolks. Many eggs died during early cleavage, and altogether the eggs are decidedly abnormal. Neither of these salts acting alone would give such an effect. When forty-two hours old the yolks are shrunken and all of the embryos have a balloon-like pericardium in front of the head, Fig. 12, pc. Later, the circulation often becomes feeble. ‘This occurs also in simple MgCl, solutions. When nine days old the embryos are small and the yolks shrunken. All steps of the fusion of the two eyes into one are shown. ‘This condition makes it certain that the magnesium of the mixture has acted upon the embryos. After fifteen days the eggs are still alive, though small and pale. ‘Thus this double solution 1s more active than a simple MgCl, solution and produces magnesium effects with really less magnesium present than is necessary to give a like result when MgCl, acts alone in sea-water. It was stated above that a strength of } m MgCl, in sea-water was the weakest solution that caused the one-eyed embryo. ‘The fact that in the mixture a } m MgCl, sea-water solution gives a like effect may be due to the additional osmotic pressure exerted by the NaCl present as has been suggested by Morgan (’06), to explain similar phenomena in the action of salt solutions on frog eggs. It may also be suggested that the Mg ions act against the Ca ions of the sea-water and thus permit the Na ions to become more active, but this explanation will certainly not apply here, since the embryos show characteristic magnesium effects. Eggs were subjected to distilled water solutions of NaCl } m, tmand?m. During the first day of development many died in most of these solutions. When the eggs were forty-eight hours old the 3 m solution contained many dead eggs, although the few still alive were almost normal in appearance. ‘This solution contains only 2.19 per cent NaCl which is less than the amount in normal sea-water yet it is obviously toxic to these eggs. It is evident that other salts present in the sea-water counteract this toxic effect of NaCl. When fourteen days old all of the living embryos appear normal. ‘The ? m solution contained one hatched embryo which had a slow pulse and feeble fin movements, it lay at rest on one side but moved if pricked with a needle. In the } m solution of NaCl Igo Charles R. Stockard more embryos had hatched than in the control though all of these fish swim with a jerky motion often moving in a spiral course or even turning somersaults in the water. The salt seems to act either upon che nerves or muscle fibers of the embryo causing the nervous twitching or jumping movements. ‘The pectoral fins seem to lack their usual codrdination. This condition is not induced by the absence of some constituent of the sea-water since embryos hatched in distilled water swim normally. ‘The result is then undoubtedly due to the action of the NaCl. The embryos die within one or two days after hatching with their bodies pecul- larly curled or twisted. Jenkinson (’06) has lately recorded a simi- lar twisting and inability to swim for newly hatched tadpoles in NaCl solutions. Sea-water solutions of 2 m, 2 m and molecular concentrations of NaCl showed only a tendency to shrink the yolk. The develop- ment progressed almost normally and only a few eggs died. On their shrunken yolks the embryos when six days old were small and behind the control in their development. At fourteen days the embryos hatched in the m and 2 m solutions, those in the weaker solution swam normally while those in the stronger showed the same jerky motions described above. On comparing these effects with those in the distilled water solutions it is reasonable to suppose that some constituent of the sea-water is capable of counteracting the effect of NaCl up to a given point’ but when an excessive amount of the salt is present its action is not entirely checked. Eggs lived for twenty-four days in a healthy condition in the molecular NaCl solution although none of them hatched. Loeb kept eggs as long as five weeks in a NaCl solution in sea-water without hatching. Embryos three days old were subjected to a double molecular 7Tn 1902 Loeb found that Fundulus embryos would not develop in a solution of NaCl in distilled water equivalent to the concentration of NaCl in the sea; be then added a trace of calcium salt and found development to be normal. After a number of experiments the conclusion was reached that the salts of monovaleat cations with monovalent anions exert a toxic effcct at certain concentrations. This toxic effect could be annihilated through the addition of a small amount of a salt having a bivalent cation or by a still smaller amount of one having a trivalent cation. In other words, the antitoric effects of cations vary directly as the valence of the elements. It was also found that mono-, bi-, or trivalent anions were all unable tu produce a like effect. The Influence of External Factors on Development 191 solution of NaCl in sea-water and they continued to develop in an apparently normal fashion but with their yolks shrunken. Loeb (94) had found that embryos three or four days old might be placed into a 27} per cent sea-water solution of NaCl and continue. nor- mal development. None of these embryos, however, will hatch. In several of the NaCl solutions I found embryos that lacked all skin pigmentation thus appearing almost white, these were not true albinos, however, since their eyes showed pigment. Such pale embryos hatched when returned to sea-water. After a consideration of the foregoing results one must admit, it seems to me, as probable that some of the elements exert a specific stimulus on the fish embryo and cause it to develop in a character- istic manner. LiCl, KCl, MnCl, and MgCl, seem to induce rather constant and definite effects or types of embryos. The form of the embryo seems to be influenced by external factors in development as well as by internal ones; in other words, the chemi- cal environment of an egg 1s important in determining the final resultant of the factors in inheritance. It may be suggested as a probability that every element that forms a chemical union with the germ substance produces on the developing egg through its action definite anatomical and physio- logical effects, which of course will vary in different kinds of eggs. Thus since the normal form of an animal may be altered in a def- nite way by certain chemical actions of the elements, we may assume that the specific nature of any animal is a product of the chemical composition of the egg cell from which it sprang. THE ACTION OF MIXTURES OF SALTS IN SOLUTION: THE CHEMICAL VERSUS THE OSMOTIC EFFECTS The following experiments were conducted in order to deter- mine whether or not by increasing the osmotic pressure of the solu- tion through the addition of a chemically indifferent substance, such as sugar, the chemical action of salts might be augmented. In other words, will eggs become more susceptible to the chemical action of a weak salt solution if the osmotic pressure of this solution be increased? Morgan (’06) has performed similar experiments 192 Charles R. Stockard with frog’s eggs and concludes that in order to be effective the two pisos together must exert a higher pressure than the one pro- ducing its effect at the lower limit but less than for the other that produces its effect at a higher pressure. ‘These osmotic pressure effects are somewhat contr: dictory as I have above pointed outin mentioning Morgan’s results in which he finds the upper limit of NaCl to be about 2 per cent with a pressure of 13.61 atmospheres, while a like fatal limit for sugar was found to exert a pressure of only 8.376 atmospheres. It is also recalled that I described above a like contradiction in comparing the pressures of fatal sea-water solutions of sugar with similar solutions of MgCl,. As there stated, this See is possibly due to the fact that the cane sugar in solution becomes inverted and thus the actual pressure is really double that calculated. In working with Fundulus eggs, as has been already pointed out, the experimenter has the advantage of being able to keep them alive in solutions which exert pressures both above and below that to which the eggs are normally accustomed. ‘This fact has been of especial value in analyzing the results of the following experiments. To anticipate what is to follow it may be stated that on adding certain percentages of sugar to a distilled water salt-solution, the action of the salt was increased although the total pressure of the solution was less than the osmotic pressure of ordinary sea-water. Such a result may probably be due to the action which would take place if the sugar became inverted in the solution. The following distilled water solutions of LiCl+ sugar were employed in one experiment, LiCl 0.128 m, 0.096 m, 0.064 .m and 0.032 m with 0.44 m of cane sugar in each. All of these solu- tions exert an osmotic pressure less than that of sea-water, except possibly the first which has an almost equal pressure. After nine- teen hours the eggs in LiCl 0.128 m + 0.44 m sugar had polar caps with ‘“‘bubbles” beneath and many were dead, those in LiCl 0.096 m_ + 0.44 m sugar were in about the same condition. LiCl 0.064m + 0.44 m sugar hadalso produced polar caps and no germ- rings were formed, LiCl 0.032 m + 0.44 m sugar had caused half of the embryos in it to die, while the living ones had formed abnor- mal germ-rings. Eggs in a solution of LiCl 0.032 m are scarcely The Influence of External Factors on Development 193 if at all affected at this time, and those in 0.44 m sugar are normal. The eggs continue to show these graded abnormalities in the dif- ferent solutions and when sixty- akt hours old were as follows: All were dead in the three stronger mixtures, and a few short embryos had been formedand were still alive in the LiCl 0.032 m + 0.44 m sugar. In LiCl 0.128 m at sixty-eight hours many were dead but a good number of short embryos were present; in the LiCl 0.032 m 20 per cent of the embryos were almost normal. In the 0.44 m sugar solution the embryos were normal. ‘The result shows that sugar augments the action of the LiCl although the pressure of the mixed solution is less than that in which the eggs usually live. This conclusion seems to me correct for now I realize the improbability that the sugar may have inverted which would thus have exerted twice the pressure supposed; if this were true then all of the solutions would have a pressure higher than that of the sea-water, though still not high enough in themselves to cause any of the above effects as will be readily seen by comparing the pres- sures of sea-water solutions in which the eggs develop normally. A reverse experiment was conducted in which the amount of LiCl present in the solution was constant while varying amounts of sugar were added. LiCl 0.032 m was mixed with 0.293 m, 0.44 m, 0.586 m and 0.88 m sugar, and LiCl 0.016 m with 0.293 m, 0.44 m, 0.586 m, 0.88 m and 1.253 m sugar. The results of these experiments showed as one would expect from the above that the injurious action of the solutions increased with the amount of sugar present, and moreover the activity of the mixture was always stronger than that of either constituent when used alone. The last point is well illustrated by eggs of forty-eight hours in the solution of LiCl 0.032 m + sugar 0.586 m. All the eggs in this solution have the blastoderm in the form of a ball on the upper pole, only a few are still alive and in these the large bubble-like segmen- tation cavity is present. “The osmotic pressure of this mixture is lower than that of sea-water provided that the sugar has not inverted. At this time, forty-eight hours, eggs in 0.586 m sugar solution are all normal, and those in 0.032 m r LiCl almost all have their germ-rings three-quarters of the way over the yolks with short embryos formed; some, however, have the germ-ring only one- quarter or one-third of the way down. 194 Charles R. Stockard Mixed solutions of LiCl and sugar were also prepared in sea- water. A 0.293 m solution of sugar was added to 0.336 m, 0.256 m and 0.192 m solutions of LiCl. The general results agree with those described above for the distilled water solutions, although the contrast between the simple LiCl solutions, and the mixtures was not so sharp. Figs. 13 to 17 of eggs when twenty hours old serve to indicate very well the conditions caused by the solutions at this period. Fig. 13 shows the appearance of the majority of eggs in LiCl 0.256 m + sugar 0.293m. Fig. 14 shows the egg nearest normal in the same solution. Fig. 15 indicates the stage that the large majority of eggs in LiCl 0.256 without the sugar have reached at this time. A marked difference exists between this embryo and those in the mixture. Fig. 16 is the most abnor- mal one in the LiCl 0.256 m and Fig. 17 shows a control egg at this age. Eggs were subjected to the following distilled water mixtures of NH,Cl and sugar, NH,Cl 4 m, + m, and 75 m + 0.44'm sugar. ie development of the eggs in these different strength mixtures was as we would expect om the result shown above. ‘Those in the NH,Cl 3m + sugar 0.44 m were all dead within nineteen hours with their blastoderms in the form of balls of cells on the upper pole of the egg. At this time some of those in NH,Cl +m +sugar 0.44 m had the germ-ring one-quarter way down the yolk, the majority, however, showed the blastoderms as polar balls which had not flattened down; many were dead. ‘The weakest solution produced fewer abnormalities. The eggs in the 0.44 m solution of aS were normal at this time, nineteen hours, and those in NH,Cl + m had thirteen normal and seven dead. When forty-three hours old all of those in NH,Cl +m + sugar 0.44 m were dead, no embryos having been formed. ‘Those in the weak solution NH, Cl;'5 m + sugar 0.44 m were also dead at this time. Both of hte solutions exert an osmotic pressure less ee that of sea-water. After forty-three hours eggs in NH,Cl + were almost normal, and their condition in NH,Cl 75 m was é same, while those in the 0.44 m sugar were well up with the control. Thus again we see that the mixture exerts a far greater influence on development than either constituent acting alone is The Influence of External Factors on Development 195 capable of producing. It appears in this instance rather illogical to state that the extra pressure induced by the addition of sugar to the solutions of NH,Cl caused this salt’s action to become more pronounced upon the eggs, for as mentioned before the pressure of these mixtures 1s often below the usual pressure in which the eggs live, and from the experiment cited below we shall find that the Fig. 13 An embryo when twenty hours old in LiCl 0.256 m + sugar 0.293 m, the majority of eggs in this solution are in a similar condition. sc, segmentation cavity. Fig. 14 The least affected egg in the above solution. Fig.15 The majority of the eggs in simple LiCl 0.256 m solution show this condition. Fig. 16 The most abnormal egg in the LiCl 0.256 m solution at this time. Fig.17 A control egg when twenty hours old. All X 17} diameters. addition of sugar to sea-water solutions of NH,Cl, which are of course hypertonic, furnish rather indifferent results. One might argue on the other hand that salts of the sea counteract the effects of the NH, ion, but even if this does occur the high pressure does not particularly injure the eggs, and we are still in the dark con- cerning the question why the distilled water solutions of NH,Cl 196 Charles R. Stockard act more violently in the presence of sugar unless it be due to some action which might take place when the sugar molecules split if they become inverted in the solutions. Eggs are necessarily very delicate chemical indicators and it may be that an action hitherto undetected might be shown by them. Furthermore, Fundulus eggs are exceptionally adapted to the study of such questions as they are not necessarily subjected to abnormally high pressure in experimentation. Sea-water mixtures of NH,Cl 4m, 4 m, and 7 m + sugar 0.293 m were used with rather indifferent results. In each of the three mixtures the yolks were slightly shrunken and in the two stronger a small per cent of the eggs always died during the first day of development, but from this time until nine or ten days old they developed in a normal manner though somewhat slower than the control. When fifteen daysold in NH,Cl } m + sugar 0.293 m 95 per cent of the eggs were dead and the few embryos alive were small with feeble pulse. “They appeared as embryos should when seven or eight days old. In NH,Cl 4m + sugar 0.293 m 50 per cent were dead and the others were small and otherwise like those described above. The eggs in the NH,Cl 7s + sugar 0.293 m were all normal except for the small size of the yolks. None had hatched. In the 0.293 m sugar solution all had a nor- mal development; and in the sea-water solutions of NH,Cl } m and +; m development was almost normal except for the contrac- tion of the yolks. ‘The embryos were a little retarded in develop- ment and none of them hatched. These results lead also to the same general conclusion, that the mzxture acts more violently than would either constituent acting alone, although the difference in action here is not great. SUMMARY AND CONCLUSIONS 1 The membrane of the eggs of Fundulus heteroclitus is readily permeable to salts in solution as is shown in embryos a few days old by the fact that KCl will stop their heart action within a few moments. During the early stages the membrane 1s also easily penetrated since eggs subjected to the action of strong solutions of LiCl for The Influence of External Factors on Dev ‘elopment 1Q7 one or two hours do not recover from the effects of this treatment after being returned to sea-water. Many other facts go to show . the readiness with which this membrane is permeated. 2 Fundulus eggs develop normally, although at a somewhat faster rate, when kept on moist plates entirely out of water. The embryos developed out of water are unable to hatch while on the moist plates, but if at any time after the control has begun hatching some of the eggs are immersed in sea-water they will soon begin hatching, commencing usually in about ten minutes after being in the water and all coming out very promptly. On hatching the embryos show a positively heliotropic and a negatively geotro- pic reaction. Embryos were kept for thirty-three days, or twenty days after the control had begun hatching, on these moist plates without beginning to hatch. ‘The fish within the egg membrane grows in length and absorbs its yolk at about the same rate as hatched ones do. hey finally die of starvation after having assimilated all of their yolk, being still confined within the egg membrane. 3 Fundulus eggs are not entirely immune to osmotic effects though it has often been stated that they are. In weak cane sugar solutions the yolks were observed to swell, this has never been seen even in eggs developing in distilled water and may probably be due to some change taking place in the sugar after it has permeated the egg membrane. In concentrated sugar solutions the yolk shrinks in a somewhat definite manner. A 1.53 m distilled water solution of cane sugar killed the eggs within twenty-three hours. ‘The osmo- tic pressure of such a solution is calculated to be 34.278 atmos- pheres or about twelve atmospheres more than that of sea-water. Some salt solutions which exert even a greater pressure do not kill the eggs. This contradiction might be explained if the cane sugar becomes inverted in the solutions but from the evidence at hand this interpretation seems improbable. ‘There may possibly be an action of the new substances resulting from the inversion of the cane sugar molecule which is also 1 injurious to the eggs. Eggs hatch in 0.166 msolutions of sugar in sea-water. On com- paring the effects of sea-water solutions of sugar with distilled water solutions it was found that a pressure more than double as 198 Charles R. Stockard high in sea-water produced a much less marked effect. Such observations seem to indicate that the eggs were less resistant to chemicals when treated in fresh water, due possibly to a slightly weakened condition when out of their usual medium. ‘The fresh water solutions showed a strong tendency to become acid and a fungus-like growth was often present (see footnote, p. 178). It will also be recalled that the acid condition of the medium would in itself be injurious to the eggs. 4. Eggs that were subjected to the action of LiCl, LiNO, and Li,SO, were all affected in a similar manner, seeming to indicate that the cation common to the three salts was the active principle concerned. Of the large number of other salt solutions employed none of the metallic ions gave the same constant abnormalities which lithium induced. The lithium larva of Fundulus is as definite and well marked as those recorded by Gurwitsch and Mor- gan for the frog. 5 a It was found, as Loeb had already shown, that this egg will develop in solutions of KCI and live for several weeks without developing a heart beat. Loeb’s statement that the circulatory system develops normally is incorrect, since the heart itself is abnor- mal, the pericardium is often greatly swollen, and other portions of the system are defective. Although eggs will liveand develop in these solutions if placed in them soon after fertilization an embryo several days old will be killed in a few moments if treated in a like manner. Thus when the heart’s action has once become estab- lished the embryo can no longer withstand the action of KCl. b The effects of NH,Cl on these eggs were rather general, development was retarded, the blastopore was slow closing and many short embryos resulted. The circulation was poor. Some lived in these solutions for eighteen days though none hatched. In mixtures of NH,Cl with MnCl, eggs were less affected than in solutions of either of these salts used singly. ‘This fact may be due to the antitoxic action of one cation on another as Loeb has claimed to take place. c MnCl, solutions prepared in fresh water caused a thickening or concentration of the central periblast in early stages, develop- ment was retarded, and the embryo had a feeble pulse. Some of The Influence of External Factors on Development 199 the embryos in the weaker solutions hatched butswam abnormally. Solutions of MnCl, in sea-water induced similar effects. d Sea-water solutions of MgCl, caused the embryos to form one large single and almost terminal eye. ‘This single eye results from an early fusion of the two optic vesicles. “The optic cup is, therefore, abnormally large and the size of the lens in such eyes varies directly with the size of the optic cup. This condition is to be compared with that known in human monsters as Cyclopia. MgCl, when mixed with NaCl also caused this abnormality. e Eggs that were treated with NaCl solutions showed no abnormalities during their early development. In the weaker solutions many embryos hatched but were unable to swim in a normal fashion. The NaCl affects either the nerve or muscle substance of these fish causing them to swim with jerky motions, and to fall on one side when at rest. ‘The embryos would live for many weeks without hatching in very strong NaC] solutions. 6 Mixed solutions of salts and sugar act more violently on these eggs than either constituent would if used alone. Very small doses of a salt will give the effect of a much stronger dose, provided that sugar has been added to the solution. ‘The presence of the sugar thus seems to augment the activity of the salt. This may be due to the additional osmotic pressure that the sugar exerts, but such an explanation is not entirely satisfactory. Pathological Laboratory Cornell University Medical College New York City, December 1, 1906 LITERATURE CITED Brown, O. H., ’03—The Immunity of Fundulus Eggs and Embryos to Electrical Stimulation. Am, Jour. Physiol., ix, pp. 111-115. ’o5. The Permeability of the Membrane of the Egg of Fundulus Hetero- clitus. Am. Journ. Physiol., xiv, pp. 354-358. Garrey, W. E., ’05—The Osmotic Pressure of Sea-water and of the Blood of Marine Animals. Biol. Bull., viii, pp. 257-270. Gurwitscu, A., ’95—Ueber die Einwirkung des Lithionchlorids auf die Entwick- elung des Frosch und Kréteneier (Rana fusca und Bufo vulg.). Anat. Anz., xi, pp. 65-70. 200 Charles R. Stockard Gurwitscu, A., ’96—Ueber die formative Wirkung des veranderten chemischen Mediums auf die embryolane Entwickelung. Arch. f. Entw.-Mech., ili, pp. 219-260. Hersst, C., ’92—Experimentelle Untersuchungen iiber den Einfluss der verander- ten chemischen Zusammensetzung des umgebenden Mediums auf die Entwickelung der Thiere. I. Theil. Zeitsh. f. wissensch. Zool., iv, 3 pp. 446-518. ’93—Experimentelle Untersuchungen. II. Theil. Mittheil. aus der Zool. Station zu Neapel, xi, pp. 136-220. *96—Experimentelle Untersuchungen. III,IV,Vund VI. Theil. Arch. f. Entw.-Mech., 11, pp. 455-516. Howe 1, W. H., ’06—Vagus Inhibition of the Heart in its Relation to the Inorganic Salts of the Blood. Am. Jour. Physiol. xv, pp. 280-294. Jenxinson, J. W., ’06—On the Effect of Certain Solutions upon the Development of the Frog’s Egg. Arch. f. Entw.-Mech. xxi, pp. 367-460. Lewis, W. H., ’o4—Experimental Studies on the Development of the Eye in Amphibia. I. On the Origen of the Lens. Rana palustris. Am. Jour. Anat., ii, pp. 505-536. Logs, J., ’92—Investigations in Physiological Morphology. III. Experiments on Cleavage. Jour. Morph., vii, pp. 253-262. *93—Ueber die Entwicklung von Fischembryonen ohne Kreislauf. Pfliiger’s Archiv, liv, pp. 525-531. *94—Ueber die relative Empfindlichkeit von Fischembryonen gegen Sauerstoffmangel und Wasserentziehung in verschiedenen Entwick- lungsstadien. Pfliiger’s Archiv, lv, pp. 530-541. *95—Untersuchungen tuber die physiologischen Wirkungen des Sauerstoff- mangels. Pfluiger’s Archiv, Ixii, pp. 249-294. *00o—On Ion-proteid Compounds and Their Réle in the Mechanics of Life Phenomena. I. The Poisonous Character of a Pure NaCl Solution. Am. Jour. Fhysiol., ii, pp. 327-338. *o2—The Toxic and the Antitoxic Effects of Ions as a Function of their Valency and Possibly their Electrical Charge. Am. Journ Physiol., Vi, pp. 411. ’05—Studies in General Physiology. Univ. of Chicago Press. Matuews, A. P., ’04—The Relation between Solution Tension, Atomic Volume, © and the Physiological Action of the Elements. Am. Jour. Physiol. X, Pp. 290-323. Morean, T. H., ’03—The Relation between Normal and Abnormal Development of the Embryo of the Frog, as Determined by the Effects of Lithium Chlorid in Solution. Arch. f. Entw.-Mech., xvi, pp. 691-712. *06— Experiments with Frog’s Eggs. Biol. Bull., xi, pp. 71-92. The Influence of External Factors on Development 201 Ronpgavu-Luzeau, ’02—Action des Chlorures en Dissolution sur le Développement des ceufs de Batraciens. ‘Théses prés. Faculté des Sci. de Paris Univ. StocKarD, C. R., ’06—The Development of Fundulus Heteroclitus in Solutions of Lithium Chlorid, with Appendix on its Development in Fresh Water. Jour. Exper. Zodl., ii, pp. 99-120. °07—The Artificial Production of a Single Median Cyclopean Eye in the Fish Embryo by Means of Sea Water Solutions of Magnesium Chlorid. Arch. f. Entw.-Mech. xxiii, pp. 249-258. Sumner, F. B., ’06—The Physiological Effects upon Fishes of Changes in the Den- sity and Salinity of Water. Bull. U. S. Bureau Fisheries, xxv, pp. 53-108. a: MOVEMENT AND PROBLEM SOLVING IN OPHIURA BREVISPINA! O2 GGLASER Witnu Five Ficures INTRODUCTION The observations and experiments which [I shall describe and discuss in the following pages were made in the Marine Biological Laboratory at Wood’s Hole, for the purpose of testing Preyer’s conclusion that ophiurans are intelligent animals. In spite of the fact that there is still much difference of opinion as to what we mean by intelligence, all will agree, I think, that it involves at least the ability to learn and to modify behavior in accordance with experience. Jennings (06, p. 291) has formulated in the law of the resolution of physiological states, the way in which behavior is modified in experience: “The resolution of one physiological state into another becomes easier and more rapid after it has taken place a number of times.”’ I have attacked the problem of intelligence in Ophiura brevispina from the point of view afforded by this law of resolution. PROGRESSION Progression in ophiurans has been described by a number of observers, including Romanes (85), Preyer (86), von Uexkill (’05) and Grave (’00). These writers agree as regards the general method of locomotion in ophiurans, but they have not described all of the movements which these animals perform. All of these authors have noticed two types of progression, the first of which may be visualized by the aid of Fig. 1, in which the arms are numbered, and so distinguished by heaviness of line, that the most active is the widest, the least active the narrowest. 1 Contributions from the Zodlogical Laboratory, University of Michigan, No. 107. THE JouRNAL oF ExPERIMENTAL ZOOLOGY, VOL. IV, NO. 2. 204 O. C. Glaser In movements conforming to type /, Fig. 1, 4, the two arms 1 and 3 are used as a pair, whose strong backward stroke drives the animal in the direction indicated by the arrow. Arm 2, which projects forward rather stiffly, serves only the function of guiding, this being also the effect of 4 and 5, which are dragged behind. A slight modification of type J, 4, is found in type J, B, in which the distal end of arm 2 waves from side to side, and in this manner adds to the propelling force furnished by r and 3. Type /,C,is a further modification of J, 4, in which arm 2 instead of bending only distally makes a stroke as effective as either that of 7 or 3, and bends either to the right or to the left, so that the animal is Fig, 1 propelled by two arms on one side and one on the other. The course is zigzag if regular alternations in the direction of the stroke of arm 2 occur but if this always falls on the same side the course is circular. The movements that fall within this type are variable to an extent which has not been pointed out. J, 4, represents in its pure forms one of the two types which all previous writers have noticed, though Grave (’00) has also observed the modification B, of which C.is the extreme case. Von Uexkiill (05), who calls this type of movement Typus Unpaar voran, says: “Beim Bewe- gungstypus Unpaar voran, zeigt sich welch grosse Unterschied in der Bewegungsamplitude des ersten und zweiten Gangpaares besteht. Letzteres verhalt sich beinahe passiv. Doch kann es gele- Movement and Problem Solving in Ophiura 205 gentlich auch stirker in Aktion treten. ” Both Preyer (86) and Grave (’00) state that the “posterior” pair is dragged behind, and I have never observed more than insignificant movements in It. Type JI (Fig. 2), observed by all of the writers mentioned, and called Typus Unpaar hinten by von Uexkiull, may be described as two pairs of arms working synchronously, or alternately, the anterior pair initiating movement at one time, the posterior at another, or the movement may be begun by arms 2 and 4; by 1 and 3; by 2and 7; or by 3 and 4; the only constant factor is the behavior of arm 5 which is invariably dragged behind. A third type of movement, Fig. 2, ///, not previously recorded, involves the activity of all the arms in such a manner that the animal is forced forward by three arms on one side and a pair on the other. ‘This type may be thought of as a modification of J, C, IPs hh Fic. 2 in which arms 4 and 5 have become active, or as //, in which arm 5 has become active. Type J/J, is really 7, C, plus an additional pair, and as in J, C, the course is zigzag if arm 2 alternates regu- larly from side to side, circular if the stroke falls always in the same direction. It is not necessary to describe the finer variations to which these types of movements are subject; to point out, as has been done in von Uexkiill’s excellent paper (’05), how one may pass over into another, or how the course is affected by differences either constant or variable in the rate and strength of stroke of particular arms or particular combinations of arms. With the exception of type 206 O. C. Glaser Unpaar voran, in which according to von Uexkill effective move- ments occur in the two arms which are usually dragged passively behind, I have observed that Ophiura brevispina moves in practic- ally all the ways in which it is possible for a pentaradiate animal of its construction to move. INDIVIDUALITY The movements described are directly dependent upon the pentaradiate symmetry, but this symmetry does not exhaust the possibilities of behavior. A little observation shows that each animal is unique at any given time and that while its movements fall within the system of classification proposed, they have pecu- liarities that distinguish them from other movements of the same type. In general the movements may be either rapid or slow, and certain individuals seem on first acquaintance to be distinctly active or distinctly sluggish. More careful study shows, however, that very sudden changes of behavior occur, and that an active, rapidly moving animal may unexpectedly enter into a state of sluggishness that sometimes lasts for hours. I do not understand these sudden changes. ‘They are not due to the conditions in the aquaria; they occur with great suddenness and not in all of the animals; they are not due to either gentleness or roughness in handling because either may or may not be followed by a change in the behavior of the same individual in successive trials. Possi- bly any sortof handling may, in certain physiological states, cause a change of behavior, but what the physiological state in which this occurs is, is hard to ascertain. In certain experiments in which | encumbered the arms with rubber tubes, after the manner of Preyer (86), I frequently encountered the same sudden change from activ ity to passivity, and arms which were flexible and easily encumbered, would suddenly bend at their tips and stiffen, so that it was impossible to slip the tube overthem. ‘This stiffening might take place at the first trial, or some other one, and never again, or it might reoccur upon every attempt to encumber the arm. Periodic changes from activity to sluggishness also occur. Movement and Problem Solving in Ophiura 207 Thus, in June of the present year, more than half of the animals I studied were very active and quickly responsive to stimul1 during sometime of my acquaintance with them, but by August the whole race had changed. Perfectly fresh material brought into the laboratory in excellent condition and kept in large tanks of running sea-water, was so sluggish that I was forced to give up the experiments which I had planned for that month. None of the stimuli employed in June elicited reaction, and acids suth- ciently concentrated to attack the skeleton, as well as the electrical current, resulted in nothing but a few spasmodic contractions with no attempt at progression or escape. What the reason for this change was is not certain. A sluggish individual almost always has very large bursal openings; in fact, it is possible to predict with considerable certainty the behavior of an individual by examining its ventral surface. The enlarged bursal openings may be consequences of the spawning process, and the periodic change of behavior of the breeding activities. O. brevispina begins breeding in June and ends in August. Late in June many individuals have spawned, and many have the enlarged bursal open- ings; by the middle of August all have spawned (Grave ’00) and most of the individuals have the enlarged bursal openings. As the genital ducts lead into the bursae—which in some species are used as brood-pouches—their enlargement may very well be due to sexual activity, which is a drain upon the animals, and undoubtedly leaves them in a state of physiological depression. If this view is correct, the enlarged bursal openings are the indices of a lethargic state following the breeding season. Rapidity and sluggishness of movement have consequences of great importance in problem solving. Sluggish animals not only make fewer movements and take more time to perform them than active individuals, but they use in general fewer arms; their move- ments are less varied, and the arms very rarely come into contact with one another or cross. All this is very different with active individuals; their movements are quick and varied; they use relatively more arms, often move these through greater arcs than the sluggish animals; and, in addition, the arms touch and cross with great frequency. How “contacts” and “crosses” are related 208 O. C. Glaser to activity and sluggishness is easy to see. An active individual using four arms in progression has a much greater opportunity to make “crosses” and “contacts”? than if fewer arms were used. Very often when the animals move by means of two pairs of arms, the anterior pair is crossed by the posterior regularly. The same frequently happens when only three arms are used. Contacts and crosses also depend on the length of the arms, as the chances that they will occur in long armed individuals are greater than in short. How important arm length is, is indicated in the following table in which are summarized observations on three individuals which were active, but differed in the lengths of their arms and also in the manner of using them. ‘The effect of the latter factor emphasizes that of the former. The longest armed individual 4 used the “two pair of arms”’ stroke, only once in a total of 141 effective backward strokes, whereas the shorter armed individuals B and C, used this stroke eight times in 129 and four times in 126, respectively. TABLE I | | = No. of arms moved Individual |No. of Movements) ~~ ae Ber cent (joo ose | Contacts Crosses | I 2) 3 4 5 ee ea A 141 13 119 I | 26 | 28 | I | © | times moved a 129 II 16 o|29 13/| 8. o| timesmoved Cc 126 6 2 o | 16 | 26 | 4 © | times moved RIGHTING MOVEMENTS Two types of righting movements were observed, only the first of which has been described by von Uexkiill (05) in an excellent paper illustrated by means of kinetoscope photographs, and by Grave (’00), who says: “Iwo adjacent arms straighten out so that together they form a straight line. On these arms as an axis the body revolves, being pushed over by the three remaining arms, but mostly by the median one of the three.” This description, correct as far as it goes, is incomplete. At the bases of the straightened arms, and in the interradial portion of Movement and Problem Solving in Ophiura 209 the disc between them, movements occur whose effect is to bend the ventral surface in the direction indicated by the arrows. When this process, by which a small portion of the ventral surface is brought into the normal position (Fig. 3, 4), has proceeded far enough, the animal is righted suddenly by its own weight, since while the process de- scribed has been going on, arms 3, 4 and 5 have so ele- vated the dorsal surface of the disc that this falls into the normal position. In the second type of right- ing movement, Fig. 4, arm 2 curves nearits base, and bends under the disc which, as in the previous case, is elevated by the otherarms, particularly by 4 opposite 2. The disc thus rotates on the base of 2 as a pivot, and after it has been sufficiently elevated, the animal falls into the righted position of its own weight. The length of time required to execute the righting reaction was measured on eight individuals. I have summarized these results in Table II, in which are given the average time for each individual, as well as the maximum and mini- mum consumed. (See Table II.) These averages of course do not show the differences between the suc- cessive individual rightings of any of the animals used. ‘These differences Were in some instances very large, and have had a great effect on the averages. (See Table IIT.) ‘These measurements show that the variations from the mean may be very great; that because an individual has Tice righted itself very quickly a number of times is no reason for believing that it will continue to do so. In spite of those cases in which righting took place slowly, the Fic. 3 O. C. Glaser 210 “UNA, “XP IA] 64 ott jog’ He ja “UIT, “XPT off o1'S 4oo'F Ur, “XP JA or 2 of‘or jor gt w® 48 “UNA, “XP o1'6z o's jO0'F Kae Aifsioalls 512 ae? eh TONAL EIAT |) ORAL XPIAT || SOHAL “Ke TAT 09 IZ 09° St og’ t of tb ob'9 09'S oh 09 ‘OI ogt erg oS Sz LE-+ joey, {OS ¥1 ayn d e) Ea spouiup yydia fo saumr-Sunydts adviapy Il PTAVL i or or s[elny, “ON ‘say +9 *s1y QI “sry & fe o s[PAIOWUyT of Bz Lt Lz Lzaun{ arg PL IME Movement and Problem Solving in Ophiura TABLE III Individual righting-times of eight animals in many successful trials 12 17 uta) 39 22 20 12 34 19 9 «on 12 10 28 43 a5 33 14 Io II 10 A$) 31 Loa) ~ 35 17 on oO Io 37 29 198 wemnmntowo a Som 24,6) th. cnisen, Uy) SP xb foe) 12 4 4 27 12 ° Wot Soo oo 253 O. C. Glaser records when averaged show that these animals, on the whole, may be expected to right themselves in less than 45 seconds. One fact of considerable interest is clearly demonstrated by the averages as well as by the individual records—there is no reduction in the amount of time required to perform the righting act; in other words, under normal conditions, these animals do not im- prove by practice in the execution of their righting movements. PROBLEM SOLVING The expression “problem solving” is almost self-explanatory. Under this heading, I have placed such behavior as an ophiuran exhibited when stimulated by interference more or less unusual, and from which it was able sooner or later to escape. What I did was to observe the way in which the escape was made—the prob- lem solved—and how much time was consumed in doing it. The problem—the same asthat employed by Preyer—was to rid one or more arms of the small pieces of loosely fitting rubber tub- ing with which I encumberedthem. Inthe selection of individuals for experiment, my choice was guided by two considerations: whether all the arms were approxi- mately equal in length, neither rs broken, nor recently regenerated; and whether the individuals were not too active to make the obser- vations easy to record. When encumbered in the man- ner represented in Fig. 5, an ophiuran does many things, some | of which are recorded in von * Uexkill’s photographs. At first it may pass through a brief latent period, during which it lies motion- less on the bottom of the dish, and then it may crawl, dragging the encumbered arm behind it. Often the animal moves at an angle to the encumbered arm, orin rare cases inthe direction of it. The / Fic. § Movement and Problem Solving in Ophiura 213 progression may be of a very violent character involving many contacts and much crossing of arms, or the animal may simply writhe, without changing its location. If it does not move about, it usually waves one of its arms, especially the encumbered, in a horizontal plane, though the movements may also occur in a vertical plane and in circles. The encumbered arm is moved in a vertical plane oftener than the unencumbered ones; is frequently rubbed against the disc; against the adjacent arms; against the sides of the dish; and even against itself. Sometimes the encum- bered arm is waved over the disc, much as a man waves a long whip, and then is “cracked,”’ so that the encumbering tube moves nearer the distal end, and often slides off. When relieved the animal usually does not remain quiet, but continues its move- ments for a short time and makes several strokes that remove it from the place where the tube was gotten rid of. If at the instant of riddance the animal was not progressing, a short journey is begun at the moment of relief. When encumbered on more than one arm, the latent period is longer than when only one arm is encumbered; the first move- ments are not through as great arcs, nor are they so long continued in any direction. One movement is succeeded rapidly, not by its duplicate, but by another in a different direction, and this by still another. ‘The behavior changes constantly. If all of the arms are encumbered, the above changes in behavior cease very soon, and an entirely different kind of action is begun. Instead of movements in the usual sweeping manner, the arms quiver andtremble. In one case, one arm (the first to be rid of its rubber tube) in particular attracted my attention by quivering when the rest of the animal was perfectly quiet. “These quivering movements occur in a horizontal plane, and are so rapid, and many of them so slight, that it is impossible to record them accu- rately without special apparatus. Of all these movements, several are more effective than many others in bringing about riddance. The most effective are the “whip movement,” the “stripping movements;”’ certain of the “wavings,” and violent progression which involves a number of different movements. Of these the whip movement is the rarest; 214 O. C. Glaser the violent progression next, whereas the strippings and the wav- ings are the commonest of all. These observations open two ways in which the problem of resolution may be attacked; by studying the time taken to solve the problem and by noting the relative frequency of the most effec- tive movements. ‘The time and the frequency might both remain constant, or might change, or only one might change. As a reduc- tion in the amount of time taken to solve the problem need not necessarily be due to an increase in the relative frequency of the most effective strokes, these two must be considered separately, although an increased frequency of strokes best fitted to solve the problem would involve a reduction in the amount of time. Ifa reduction in the amount of time required does occur, it means that the physiological state produced by the rubber tube has been resolved intothe normal state more rapidly than it was resolved the first time. In other words, the animal has learned by experience. The following Table IV contains my measurements of problem solving time. In every case the animal was given the same prob- lem consecutively, viz: the same rubber tube was placed on the same arm, under the same conditions. As little time as possible . was lost between trials. TABLE IV* Trials Individual...... I 2 3 4 5 6 7 A BU tad 5/ 00” 2’ 00” 3/ 00” 6’ co” B o/ 30” 4’ 30” 1’ 00” ale 4/ 00” Gc o 45” 1’ 45” fod rid is co” V 30” D 2’ 00” 3/ oc” qAOr 3/ 00” 3’ 00” Py 4 3/0" E 12307 o! 45” 1207 1’ 40” Slisg ghegtex *These measurements include the latent periods. The number of trials recorded in Table IV is small. I was prevented from collecting more data by the sudden changes of behavior before alluded to. Other animals were tried but failed to react regularly even five times. The results as they stand, how- ever, are worthy of confidence; they are representative of the whole behavior which is varied and uneven; like the measurements of Movement and Problem Solving in Ophiura AUS righting time they neither increase nor decrease—the apparent increase being due to the failure to respond, for had this failure occurred sooner, some of the last measurements would have been smaller than the first. Fatigue played no part in the result, as the figures are too uneven. The objection might be advanced that these cases which I have called “problems,”’ were not such; that there was no reason why the animals should modify their behavior, and that what they did under the conditions of the experiment was nothing that they would not have done under normal conditions. ‘This eee is met satisfactorily I believe by the following experiments. A given arm was stimulated by encumbering it with a rubber tube, or by painting it with strong or dilute formalin or hydro- chloric acid of different strengths. ‘These trials, of which I made a great many, yielded very definite results. In only one case did an animal progress in the direction of the stimulated arm; in afew cases at-an angle to it, using it as one of the propellers, whereas in the vast majority of cases it moved in the direction diametrically opposite the stimulated arm. If the stimulus was strong, the movements were very violent, but no difference in direction was noted in the case of weak and strong stimuli. Under ordinary circumstances it is impossible to predict the direction in which an ophiuran, all of whose arms are of the same size, will move, but if one of the arms be encumbered the prediction that the animal will move away from the stimulus will be verified in the vast majority of cases. I think it is justifiable to assert that the direction of pro- gression has been determined in these cases, and if this is true there is a determining cause—a problem. My second line of inquiry—whether encumbered animals showed a noticeable increase in the number of movements best adapted to solve the particular problem given, was begun by find- ing the percentage of crosses and contacts in the same animals under the two conditions stated. The results are summarized in Table V. As contacts and crosses usually result from wavings I counted these in animal JJ unencumbered and with one arm encumbered. The results are summarized in Table VI 216 O. C. Glaser The general conclusion to be drawn from these experiments is that there is neither a decrease in the amount of time taken to solve the problem, nor an increase in the relative frequency of movements best fitted to solve it. In other words, the animals did not modify their behavior in accordance with the law of resolution, and consequently, so far as is objectively recognizable, learned nothing. TABLE V Animal Arms Encumbered Movements _ Per cent Contacts |Per cent Crosses Problems I ° 231 4.0 2.0 ° I I 202 3-9 | 1.9 3 I ° 267 13.8 10.8 ° II 5 553 9.2 11.2 5 TABLE VI ~~ + 2 a Se ee ee Animal | Arms Encumbered Movements Per cent Wavings II ° 117 26.4 Il I | 192 19.3 DISCUSSION The facts which I have brought forward in the foregoing pages agree with those of Preyer and von Uexkiill in showing that in problem solving the animal repeatedly changes its behavior, not persisting in a certain reaction when that is unsuccessful. If | venture to take issue with Preyer, and to assert that the behavior which both he and I observed does not warrant the conclusion that ophiurans are intelligent, I must rest my claim upon the validity of my interpretation of the facts, and this validity I shall now attempt to establish. The behaviorof Ophiura brevispina may be summarized by say- ing that this animal under normal conditions performs practically all the movements possible to a creature constructed as it is; that except for this limitation, its ordinary behavior is not predictable, hare. Movement and Problem Solving in Ophiura 2G) and that even the righting movements, because of their variety occupy a place between the ordinary behavior and reflex behavior, for though more definite than the former, they are less precise than those highly perfected types of response which gave us our first idea of reflex action. Regarding the manner in which ophiurans rid their arms of encumbrances, Preyer (’86, p. 125) says: “Aus den beschriebenen und ahnlich leicht zu variirenden Versuchen ergiebt sich zunachst, dass Ophiuren in 5-fach. verschiedener Weise sich gegen die beim Tasten und kriechen thnen sehr hinderliche Bekleidung mit einem Schlauche vertheidigen: (1) streifen sie ihn ab durch Reibung am Boden wenn er locker ist, (2) schleudern sie ihn fort durch geissel. formiges Hin und Herwerfen, (3) drucken sie ihn fest gegen den Boden mit dem freien Nachbararm, und ziehen den Arm aus dem dadurch fixirten Rohre heraus, (4) stemmen sie abwechselnd beide Nachbararme mit deren Zahnchen unten gegen dasselbe und schieben ihn ruckweise ab, (5) brechen sie durch Selbst- amputation den Arm mit der unbequemen Bekleidung ab. Hilft dass eine Verfahren nicht, dann wird das andere angewendet. Sehe ich hier von dem letzten, der Autotomie, ab, von der noch die Rede sein wird, so beweist schon die 4-fache Art der Abwehr bei einem und demselben Individuum unter denselben atisseren Verhaltnissen, dass hier kein einfacher Reflex vorliegt. Vielmehr besitzen die Ophiuren die Fahigkeit sich ganz neuen, von ihnen noch niemals erlebten Situationen schnell anzupassen.” “Wenn Intelligenz auf dem Vermogen beruht, Erfahrungen zu machen, d. h. zu lernen, und das Erlernte in neuer Weise zweck- massig zu verwerthen, so miissen also die Ophiuren sehr intelligent sein.” Preyer’s reasoning seems to be this: When encumbered on its arms the animal moves in different ways; failing to free its arms by these movements, it moves in other ways, and continues to change its movements until the encumbrances have been removed. The animal thus exhibits the process of discovery by elimination, learning in other words, and is therefore intelligent. If this indeed be learning, then all movements which any organism may under any circumstances execute are outward signs 218 O. C. Glaser of the process, for movements are never without cause, and the stimulus is aggravated, alleviated, or unchanged by them. What- ever be the result of the movement, the animal “learns” what has been the effect upon the stimulus, the cause of the movement. Two criticisms may be made of this point of view: In the first place, in behavior such as that of an ophiuran, movements which fail to solve a specific problem, or to contribute anything whatever to its solution, -are often repeated immediately. If the animal learned anything from them, it forgot what it learned at the instant of learning, for the intervals between two successive move- ments which fail for the same reason mzy be less than one second; to forget as rapidly as to learn, can be objectively recognized as neither. In the second place, in ophiurans at least, it is the excep- tion for an animal to perform only one movement at a time. Usually a considerable number, four, five, or six distinct move- ments are performed synchronously. — All of these, on the assump- tion I am criticising, result in learning, but the knowledge which they give may be of two sorts; some of the movements may tell the animal how to solve the problem, the others, how it cannot be solved. It is impossible for me to believe, without striking evi- dence to the contrary, that an ophiuran can learn at the same instant half a dozen facts, belonging some to one, some to the other of two distinct categories. If the idea that mere movement in various directions is a sign of learning, involves the serious difficulties which it seems to me to involve, we have nothing but behavior more or less permanently modified as the result of experience to fall back upon. I have shown that under ordinary circumstances Ophiura brevispina does not improve with practice, in its righting behavior, and in problem solving it shows no greater aptitude. Iam, therefore, forced to the conclusion that neither intelligence nor even learning have as yet been demonstrated in this animal. My experience with ophiurans also leads me to the conclusion that resolution will be very difficult to demonstrate, not only because of those sudden changes in behavior for which it is difficult to assign causes, but also because of the remarkable “action sys- tem” exhibted by these animals. This action system shows better Movement and Problem Solving in Ophiura 219 than many others, that behavior is structure in motion, and that complexity of behavior depends on the complexity of that which behaves. An act performed by one arm may also be performed by any of the others. ‘The arms may all do the same thing at the same time; some may do one thing and others another; and finally a single arm may execute different movements at different levels. As the disc itself may also execute varied movements, the number of possibilities is enormous. With this marked versatility to contend with, it is not surprising that resolution, demonstrated according to Jennings (’04, 05, ’06) for Protozoa, Ccelenterates, and other forms lower in the scale of complexity than echinoderms, or as low, should remain undemonstrated for ophiurans. The number of movements possible to an ophiuran is immense; if the animal only acts, the chances that it will perform movements fitted to relieve a certain physiological state are better than the chances that such will be the case in most other animals. If one of the many movements that will serve is not performed, another will be, and we should not expect to find resolution, unless the fit things to be done are few. Any of the problems presented might have been solved in a variety of ways. One or more of these Ways Were superior to any of the others, but all served the purpose. Where the variety of solutions to a problem is great, there is no need of resolution, and it does not occur. I have profited much by the elaborate criticism which Professor Jennings made of an earlier draft of this paper, and I take this occasion to thank him for his kindness. University of Michigan Ann Arbor, Mich. February 1, 1907 220 O. C. Glaser LITERATURE ‘CITED: Romanes, G. J., ’85—Jelly-Fish, Star-Fish and Sea Urchins. Kegan Paul, Trench & Co., London. 1885. Preyer, W., ’86—Ueber die Bewegungen der Seesterne. Mitth. a.d. Zool. Stat. z. Neapel. Bd. vi. von UEXKULL, J., °05—Studien uber den Tonus IJ. Zeitsch. f. Biologie. Bd. xlvi. Grave, C., ’00o—Ophiura brevispina. Mem. Biol. Lab., Johns Hopkins Univer- sity, IV. 5. Jennincs, H. S., ’04—Contributions to the Study of the Behavior of Lower Organ- isms. Carnegie Institution, Publication 16. °05—Modifability in Behavior. Journ. Exp. Zodl., vol. ii. °0o6—Behavior of the Lower Organisms. Columbia Univ., Biol. Series x. OCCURRENCE OF A SPORT IN MELASOMA (LINA) SCRIPTA AND ITS BEHAVIOR IN HEREDITY BY ISABEL McCRACKEN Laboratory of Entomology and Bionomics, Stanford University Witu One Pirate During the year 1904, early in the breeding season of the chryso- melid beetle, Melasoma scripta,' about 1000 pupe, and larve, in advanced stage, were collected from willows in the neighbor- hood of an artificial lake near Stanford University. Such of these as were not parasitized matured during the latter part of April and early May. ‘The adults represented the dichro- matic extremes of the species, the elytra being either spotted- brown (referred to in this, as in previous papers, as “S’’), or black (referred to as “B’’), the thorax in each case having a central black area widely emarginated with brick red. In the center of each red area and nearly adjacent to the central black region (sometimes approximating it) is a small black spot representing asingle punctation. (Figs. 1 and 2.) During the course of breeding through four generations from this collected material there occurred a number (four or five) wholly black individuals (Fig. 3), thorax as well as wing coy- ers being totally black (referred to in this paper as “AB”’). Since during the casual outdoor observations made throughout that 1A description of this beetle is given in the Journal of Experimental Zodlogy, 1905, vol. ii, pp. 117, 136, and vol. iii, pp. 320-336, where it is called Lina lapponica. It seems that this identification made for me isnot correct. The beetle is evidently the one figured by Riley under the name Plagiodera scripta (Fabr ) Ann. Rept. Agric. for 1884, pp. 336-340, pl. viii, Figs. 1 and 2; by Lintner, under the name Lina scripta (Fabr.), Rept. N. Y. State Entomologist for 1895, pp. 181-189; and by Felt as Melasoma scripta (Fabr.) in N. Y. State Museum, Memoir 8, vol. i, pp. 317-322, Pl. 16, Figs. 16-20. Tue JourNnaL or ExperRIMENTAL ZOOLOGY, VOL. IV, No. 2. 222 Tsabel McCracken season no such freaks were found, and those bred in the laboratory failed to mate, they were looked upon as representing possibly a pathological condition. However, in 1905 outdoor scriptas were kept under constant surveillance throughout the breeding season. At stated periods, four or five weeks apart, several hours were spent in the field, at which time several hundred individuals passed under inspection with the following result: First inspection, March 4. Thousands of beetles in a limited area feeding and beginning to breed. (These were in all proba- bility the hibernated individuals from the previous year.) No “all black”? (AB) individuals were observed. Several hundred individuals, representing each of the dichromatic extremes, were collected at this time for indoor controlled breeding. April1z: Many thousands of beetles observed; two AB females collected. May 14: Many thousands observed, two AB females collected. June 21: Many thousands observed, two AB males collected. July 28: Individuals in this particular feeding ground becom- ing noticeably fewer. Many hundreds of beetles observed, one AB female collected. August 21: Many hundreds of beetles observed, one AB male collected. Hence a total of five females and three males were collected in this locality during the five months the locality was under observa- tion and covering the breeding season of the beetle. “Three similar sports were collected during this time from poplar trees a half mile or so distant from this locality. During the progress of these outdoor inspections, indoor breed- ing was in progress from the collection of March 4, 1905, that 1s, a collection made up of “spotted” (S) (Fig. 1) and “black” (B) (Fig. 2) but no “all black” (AB). Four generations were reared to maturity from this collection. The following table gives the data in regard to the occurrence of the sport AB with the character of the lineage in each genera- tion. (The term “sport” is here used in the sense of a singular and decided variance from the normal type.) mata Sport in Melasoma and its Behavior in Heredity 223 There occurred, therefore, in the breeding room, a total of 20 AB, or sport individuals in a total of 11,369 individuals reared during the breeding season, most of these coming from the imme- diate collection of March 4. Inspection of the table shows that in the first generation, 168 matings were made, one brood only being reared from each pair. Parentage was represented by both SxS and B xB matings. The sport AB was found in the progeny of each series, ten in the former, six in the latter. In other generations matings were made in the S line only. In the second and fourth generations, single broods only were reared from each pair, as in the first generation, TABLE I GG a SS Now lane GGrand-| Grand- Total No. AB | Total | Parents. | Matings broods | Grand- | : A | parents. Bese parents:'| made. _|reared.| Se | bee | | Ist gen. | SxS 119 | 119 8 2\ | é I BxXB 29) 49) || 503400 seman 2d gen. | | | SxS SxS 45 45 TOSO) || 0) | I 3d gen SxS SxS SxS 42 180 4736 |3 0 3 4th gen. SxS SxS SxS SxS 19 i a, 549 emo. | ° LiGtial > 5. oR MOM OOS On USOC Oe mole rere 264 | 41z | 11,369 | | 20 while in the third generation, an average of five or six broods were reared from each pair (a minimum of two, a maximum of fifteen broods). With three exceptions, not more than one AB individual occurred in any one brood. In the first exception two individ- uals (a male and a female), occurred in a single brood, in the second exception five males occurred similarly, and in the third exception two males. We find, therefore, in a total of 264 matings, fourteen only producing sports. That the occurrence of the sport was normal, that is, was not due to laboratory conditions, is evidenced by the fact that but com- paratively few sports occurred (20 out of a total of 11,369 indi- viduals.) Since all broods in the breeding room are under prac- tically the same conditions, had an extrinsic influence been at work 224 Isabel McCracken tending to produce this variant, certainly more individuals would have shown the effect. The numerical results in second, third and fourth generations of the occurrence of the sport might have been different had con- tinued breeding of the progeny of the first B parents been carried on and if a larger number of broods had been reared from each pair. There was found to be no greater likelihood of the recurrence of a sport in a brood from parents in whose lineage sports had pre- viously occurred than in broods from parents in whose lineage no such sports were known to have occurred in so far as this point was tested. Numerous matings were made between individuals having AB sisters or brothers, but in no case was there a recur- rence of the AB character. That the sport is of occasional occurrence in the field has already been noted. The stable character of many sports or aberrant Variations in animals such as this appears to be is undoubted, since it has been shown several times, notably in the race of Ancon sheep from a single short-leggéd, long-backed ram, in 1791, and the production of the Mauchamp-merino breed of sheep in 1828 from a single ram with long, smooth, silky wool.2 Also more recently the production of a race of polled Hereford cattle in Kansas in 1889 from a single polled bull. The main purpose of the present investigation was to deter- mine, in case the new character should be found to be stable, its hereditary value in relation to the characters of the parent species. The hereditary value of each of the dichromatic extremes with relation to the alternative extreme had been previously deter- mined.! To test the hereditary value of the Melasoma sport, the first two sports occurring simultaneously in the first generation were used as parents for succeeding generations. Other sports were mated with the parent forms. ?Darwin, 1868, Animals and Plants, vol. i, p. 126. SGuthrie, 1906, Proc. Amer. Breeders Assoc., vol. ii, p. 93. *McCracken, 1906, Inheritance of Dichromatism in Lina lapponica. The Journal of Experimental Zodlogy, vol. ili, pp. 320, 336. Sport in Melasoma and its Behavior in Heredity 225 In the first matings of AB x AB, the male parent was of S par- entage, the female of B parentage. Five broods were obtained from this mating, a total of 130 individuals. Of these 77 indi- viduals were similar to the immediate parents, that is, were AB in character, and 42 individuals were similar to the grandparents on the female side, that is, were B in character. In eleven indi- viduals the wing. covers were black, while the character of the thorax was a mosaic of the thorax of AB and B (Plate, Fig. 4); that is, the emarginate area was in part black and in part red. The black blotch upon the red might be in any position, covering the anterior half, the posterior half or the median two-thirds, as indicated in the figure. We find here, therefore, a series of variations arising in the progeny of a mating between two similar sports or aberrant variations, so that if the latter had not arisen, first, it would have been considered but an ‘‘extreme variation”’ in a series. No such mosaic was observed in field collections. -There was no recurrence of the “S’’ type. The sport or parent character predominates somewhat in the offspring over the B Pype{i.8:1). The mosaic type appears to be a heterozygous form as in mat- ings of this type (I) the offspring always revert to the AB or B type with only an occasional [ individual. With the 77 AB, the 42 B, and the 11 [| individuals thus obtained, six categories of matings were made as shown in the following diagrams. ‘These diagrams show the pedigree for three genera- tions and the character of the brood produced in each. DIAGRAM I—Mating Category a First Matings : x | Boos First Generation S, B and AB sport B and AB sport Second Generation AB, B and I (7:3.3:1) | Third Generation AB ABand B (12:1) AB and I (6.9:1) AB, B and I. 226 Isabel: McCracken DIAGRAM Il—Mating Category b { { First Matings S-x-S$ Bx sE | | BxB | First Generation S, B and AB (sport) B and AB (sport) | | Second Generation AB, BI (mixed brood) B (all) Third Generation B (all) DIAGRAM III—Mating Category c First Matings Sx ‘| a = : % ' First Generation S, Band AB (sport) Band AB (sport) ! (all) S (all) Second Generation AB, B and I (mixed) S (all) Third Generation S (all) DIAGRAM IV—Mating Category d First Matings Se BxB a=] Ee First Generation S, B and AB (sport) B and AB (sport) Second Generation AB, B and I (mixed broods) Third Generation B (all) B and AB (mixed broods 3.6:1) Sport in Melasoma and its Behavior in Heredity oS | DIAGRAM V—Mating Category e First Matings [ x ' nee ? ? First Generation S; ae (sport) Band AB (sport) B-~ xX i Second Generation AB, B and I (mixed) B ae Third Generation B (all) Band AB (3.6:1) B,AB and I (mixed) DIAGRAM VI—Mating Category f First Matings : x | ea S, B and AB (sport) B and AB (sport) AB, B and B and AB (3:1) B and AB (with an occasional I) The following table gives the results in total as to the character of the individuals reared in the third generation in the different categories, from the first matings, the lineage for which is given in these diagrams. In the papers previously referred to on “heredity of dichromat- ism,’ it was shown that the color type S is dominant over the color type B, not at first completely so in every case, but showing from generation to generation an increasing prepotency in that direction. B was considered recessive in that it disappeared wholly or partially in the first generation of a cross between S and B, but bred true from the time of its reappearance in the sec- ond generation. Because B in its ontogeny passes through the S stage, it was suggested (’05) that S represents possibly the older, B the newer type. AB in its ontogeny passes through the S and B stages. Its normal infrequency of occurrence makes it appear 228 Isabel McCracken to be either the newest type or an atavistic form. Upon the assumption that it is a new type, we would expect its behavior to S and toBto parallel that of B to's. Our expectationis in the main fulfilled. Inspection of Table II] shows that in the progeny of AB xB (category b) all the offspring are B. In the progeny of TABLE IL (Third ae Mating Tot. No. Char. of Tot: No.) Lot. Tot ) Lot eee Parents / AB“B or Category Broods Broods Indy. |} AB} 8 I E AB :I a ABXAB 61 S; Sosastce ° V3 sesocc ° AB (all)..24 586 AB& B.. 6 133 123 | 10 12:1 AB&I .27 651 569 82 | 6.9: I AB,B&I 4 105 56 | 3x | ° 18) e5:6:gcn-res b ABXB 5 Biepecw 5 | 135(allB)| o ° G ABXS 17 SY ayes ae 17 448(all S) ° ° ° d BXB 16 Sieh carts ° (both parents Bi (all) eee 52 Of 52 from AB) AB&B .14 498 108 390 oO} 17326 e BXB 39 Syeeehe ° (one parent only B (all)...14 381 381 from AB AB&B .23 629 139 490 Ter AB,B&I 1 17 T] og I B&l I 30 29 I if SCT: 4 S) seceaiee ° (both parents AB(all) ..0 from AB.) Ball) 2 3 AB&B . 3 80 20 60 mised) AB, B&I 1 22 +) 15) oae AB xS (category c) all the offspring are S. That is, both S and B dominate AB completely in crosses between S and AB or B and AB. In these two categories AB is completely recessive, in the sense that it does not appear in the soma of any of the offspring. In BxB matings, extracted B (category d, Table II), both parents having been born of AB parents, two kinds of broods are Sport in Melasoma and its Behavior in Heredity 229 produced; that is, broods in which all the individuals are B and mixed broods of B and AB, individuals of B character predomi- nating. In these mixed broods the proportion of B to AB is 3.6 to 1. This parallels the history of the behavior of S toward B in matings of dominant S hybrids as previously determined. In BxB matings (category e, Table II), in which one of the parents was born of AB parents, two kinds of broods also appear; that is, broods in which all the individuals are B, and mixed broods of B and AB, or B, AB and | or B and I, the character B predomi- nating. It is noticeable that in this case, in which one-half as many AB ancestors are represented, the proportion of wholly B broods exceed that of the former cross by about 3 to 1, while the proportion of B to AB in mixed broods remains the same, 3.6 to 1. In neither case were any broods produced that were wholly AB in character. In I xI matings (category 7, Table Il) mixed broods result in which B is the predominating character, AB taking a second rank and I recurring but rarely, thus showing the heterozygous char- acter of I. The hereditary value of the character B in the first generation of a cross between B and AB (that is, B not previously contami- nated by AB) appears from these data to be, therefore, equiva- lent to the hereditary value of pure S to the character B. The hereditary value of pure S and pure B with reference to AB have the same equivalency. The stability of the sport character AB is absolute in neither first nor second generation matings, but comparison of data of first generation mating (Diagram I, category a) with Table II, a, shows an increased preponderance of the sport character in the latter case. AB in the second generation breeds true in more than two-thirds of all broods produced and in all mixed broods greatly predominates in the offspring. It is again noticeable that no S individuals appear in the offspring, although reference to Diagram I shows that as many S as B individuals are represented in the ancestry and reference to Table | shows AB appearing in broods of SxS parentage as well as those of B x B parentage. ‘This sug- gests that the type AB is after all but a new phase of the type B 230 Isabel McCracken and that its appearance in the progeny of S 0; for (U,—U,) =E = W in accordance with the First Law, and so might be set in equivalence with pv, and this substituted, with the result an identical equation as before. By the results of the computation in accordance with the results obtained from measurement, namely, that an original increase of pressure following fertilization and a decrease of this following each cleavage take place, while the temperature and volume remain constant, the quantitative value of the resultant energy- changes will, then, have been determined. But, in accordance The Energy of Segmentation 305 with this mode of determination, W and U,—U, will not have been kept apart; accordingly it is quite possible that during the event of cleavage heat is absorbed from the environment, that the internal energy will have increased by a certain amount. How- ever, to determine the exact value of such an increase would seem to be impossible. Yet experiment makes clear that there is a change in pressure. ‘lo what is this due? Equation [8] shows that in general such a change must be due to the fractional change in some other intensity; for, with U,—U,=0, and W=pv, by substitution in [8] we get That this change should be the result of a change in tempera- ture is impossible if the view is correct that this last intensity 1s constant; yet correlative change there must be, and the evidence is that this is in the chemical conditions. RECORD OF EXPERIMENTS The following is an epitomized record of the experiments per- formed,” showing what solution-concentrations were sufficient to inhibit the earlier cleavages; the results will be found to show a fairly close agreement. ‘The figures given here are simply those of the concentration, obtained in each case by starting with a cer- tain number of cc. of a 3 m. sugar solution in sea-water for each lot of eggs transferred, and then diluting this amount with a cer- tain measured amount of sea-water, plus the 14 cc. of sea-water necessary for transferring the fertilized eggs. ‘The pressures thus obtained are here not reduced to atmospheres; that is done only for the typical experiment in which the computation 1s carried out completely. In all of these the eggs in the control segmented very uniformly and well. Experiment I, July 15: First segmentation stopped in sol. Io cc. sugar sol. plus 11 cc. sea-water plus 1} cc. s. w.,'! and in 10The experimentation was done in the Marine Biological Laboratory at Woods Hole, in the summer of 1905. Us, w., sea-water. 306 E. G. Spaulding all stronger; continued in all weaker solutions. Second cleavage stopped by 10 cc. sugar sol. plus 13.75 cc. s. w., plus 14 cc.; the third by 1occ. sugar, plus 15 cc. s. w. plus 13; the fourth, by 10 ce. sugar plus 16.5 plus 1.5 cc. s. w. Experiment II, July 21: Made for purposes of refinement; found to be confirmatory of preceding. Experiment III, July 22: First segmentation stopped by 15 cc. sugar plus between 16 and 18 cc. s. w. plus 13 ce. Experiment IV, July 22: Inhibition point found to be between 16 and 16.5 cc. s. w. plus 15 cc. sugar plus 14s. w. Experiment V, July 24: First cleavage stopped by 15 cc. sugar plus 3 (16 plus 16.5) cc.s.w. plus 14.cc. This means that the (16), and all stronger, stopped, while the (16.5) and all “up” from this “allowed to proceed.’ This method of taking the “intermediate point” was subsequently adopted in each case. Second segmenta- tion stopped by 15 cc. sugar plus 4 (20.5 +21) s. w. plus 14 cc.; third, by 15 plus 3 (21.5 +22.5) plus 14 cc. Experiment VI, July 27: Temperature 23.5 C. Fine lot of eggs. Fertilized at 11 a. m.; repeatedly and frequently observed up to 3 p. m.; observations all confirmatory. First segmentation stopped by solution 15 cc. 3 mol. sugar sol. plus 16.75 cc. s. w. plus 1.5 s. w.; second by 15 cc. sugar plus 20.75 plus 1.5 cc. s. w.; third by 15 cc. sugar plus 21.5 plus 1.5 cc. s.w. Control; over 95 per cent of the eggs segmented uniformly. For the purposes of the computation to be made, the surfaces and volumes of each stage must be found. ‘This was done as fol- lows. In the one-cell stage the typical or “modal” sea-urchin egg is approximately spherical; accordingly the diameters of as large a number as the period of 55 to 60 minutes elapsing before the first segmentation allowed were measured by means of an ocular micrometer; from these data, widely divergent values being excluded, the average diameter was found, and the surface and vol- ume computed from well-known formule. In the two-cell stage the typical form is that of two oblate ellipsoids; the long and the short axis of each ellipse was accordingly measured for a number of eggs, the average for each taken, and the surface and volume computed. Difficulties in doing this for the four and eight-cell The Energy of Segmentation 307 stages were foreseen, but the results obtained for the first two stages showed that the volume after the first segmentation was the same as before it; it could, therefore, be assumed to be con- stant during the second and third and even subsequent stages, especially since general observation makes no change manifest. On the other hand, the increase of surface which was demonstrated to have taken place was shown to be not of direct significance in the computation made of the energy of segmentation. These measurements of diameters were made in Experiments V and VI on eggs taken, of course, from the control.” But it is evident that the numerical results thus obtained are in any com- plete computation to be combined with those obtained from the use of the inhibiting solutions on eggs of the same lot. Accord- ingly, it is the result of the complete computation from all the necessary experimental data, as taken in Experiment VI, that is presented below; and in connection therewith it may be remarked that, as between method and numerical result, it is the former rather than the latter that I would have regarded as the more worthy of emphasis. If the question be raised as to the accuracy of the numerical result, this can be estimated by considering the sources and probable limits of error introduced both by the method of compensating and by the fact that one is observing a “group” of eggs and must adopt the expedient of taking averages, etc. OBSERVED AND COMPUTED RESULTS IN EXPERIMENT VI Data: Segmentations stopped, the first by 15 cc., 3 m. sugar feeplus 10.75 plus 1.5 cc. s. w. at 23.5° C., etc. Now, it is well known that osmotic solutions in general follow the law for gases; and it is held, too, that there is no dissociation ina sugar solution. Accordingly the pressure of a mol. sugar sol. at o° C. is 22.4 atmospheres.” Miss Evis Berry kindly assisted me in the experiment in this way. 3An atmosphere is that unit of pressure which is exerted by a column of mercury 76 cm. in height of a density 13.596; this equals in C. G. S. terms 1013300 dynes, i.e., the pressure of such a column of mercury per sq.cm. The egg has of course an area which is only a small fractional part of a square centimeter. 308 E. G. Spaulding Making the correction for a room temperature of 23.5 C. in accordance with the formula, p, =p, Ci. 00367 t), the osmotic pressure of the above diluted solution, 7. ¢., the pressure sufficient to inhibit the first segmentation and therefore equal to the result- ant internal pressure, is 7.32 atmospheres. For the second segmentation, the numerical value of this inhib- iting pressure, as computed in a similar manner from the recorded figures above, is 6.53 atmospheres, and for the third, 6.40 atmos- pheres. MEASUREMENTS OF DIAMETERS AND AXES Average diameter of 20 typical eggs = .072 mm.; radius = .036. Area (4 zr’) = .0164 sq. mm. Volume (# z r°) =.000214 cu. mm. Two-cell stage: each cell an oblate ellipsoid. Average axes of 20 typical segmented eggs: Dong axis (ele ne che al ie eee coe 068 mm. DNOIEAORIS Wa oe he veh cs Brie eae 039 mm. J Area (cee xz ‘’) =.0118 sq. mm. for each cell; for both .0236 sq. mm. Volume (32d d” x 2) = .000206 cu. mm. for both cells together. From these values it is evident that, whereas the area has in- creased by .0072 mm., the volume has remained the same. It would now seem as if the data were at hand whose numerical values could be substituted in the “work integral” Be W = -{ udp Pi which becomes W = v (p, — p2) when the volume remains constant, as in this case. However, before doing this, the question must be answered, as to what may be the value of that pressure which is due to the tension of the surface film or membrane of the develop- ing egg. For it might seem that the resultant internal pressure before and after each segmentation was equal to, not alone the opposing osmotic pressure of the surrounding sugar sea-water The Energy of Segmentation 309 solution, but, rather, to this plus the pressure of the surface film or membrane. Accordingly, the numerical value of this must be found, that it may be known whether it is significant for our com- putation or not. The formula by which this pressure due to the tension of the surface," if this be only a film like the surface of a drop of water, may be computed, is Pia in which ¢ is the coeflicient of surface tension and r the radius of a sphere. ‘This ¢ is determined from the capillary action of a fluid in accordance with the formula, t= 4 grh D (g = action of gravity, r = radius of tube, / = height to which the fluid 1s drawn up, D = the density). Pfeffer gives this coefficient as .o1 g. cm. in relation to that of water as unity. Since other determinations are lacking, I made use of this, although, of course, it must be admitted that this coeficient might vary greatly with different kinds of proto- plasm. Substituting this value in the formula, 2 SSS r p = .0055 atmos. pressure For the two-cell stage, with each cell an oblate ellipsoid, the form- ula is more complicated: here ohn Vae-C tiAuz Yc ,—— tan-47 — : ao c area (a = long axis, c = short) 4T he best treatment of the general problem of surface tension, etc.,which I have found is M. Heiden- hain’s Die allgemeine Ableitung der Oberflachenkrifte, etc., in Anatomische Hefte, erste Abteilung, vol. xxvi. Wiesbaden. 1904. ' % Plasmahaut u. Vakuolen, Abhandl. d. Math.-phys. Kl. d. Sachs. Ak. d. Wis., 16, 185 (1891); cited by Hober, Physikalische Chemie der Zelle u. Gewebe, s. 38; Leipzig, 1902. This is the only determination I have been able to find. *For this formula I am indebted to Dr. C. R. MacInnes, of Princeton University. 310 E. G. Spaulding Substituting, we get p = .0063 atmos. Although, therefore, it appears from this that there has been an increase in the pressure which would result from the curved sur- face of the egg were this a film, it is also evident that this is of insignificant value in comparison with, the values, 7.32, 6.53, 6.40 atmos. found for the inhibiting solutions. It falls “outside the limits of error,’ and is, therefore, to be neglected in the applica- tion of the “work integral.” The fact, however, that this pressure has such a small compara- tive value, results, evidently, from the substitution of .or as the coeficient of surface tension of protoplasm. ‘The acceptance of this value is, of course, purely gratuitous; but if it be approxi- mately correct for the protoplasm of the sea-urchin egg, then the resulting small value of the pressure of the surface on the basis of the assumption that this is a film proves this assumption to be incorrect, and indicates that there must be a membrane, differ- entiated from the cytoplasm, to oppose the relatively high internal pressure as indicated by the strength of the solutions requisite to inhibit segmentation. There has been demonstrated, then, experimentally, an increase of 7.32 atmos. in the “resultant” pressure, as brought about by fertilization and the process following it up to the time of segmen- tation. As a result of these, the egg normally cleaves; it changes form, and it is now shown experimentally that as it does this the internal pressure therewith decreases; without fertilization these events do not take place. For the early segmentations, then, there are numerical data at hand from which the resultant energy change can be computed in accordance with the “work integral”’ We -{"vdp Pi which becomes, when the volume is constant, W—v (p,— pa) The Energy of Segmentation 311 Substituting in this the numerical values obtained for pressures and volume, we find that there has taken place, as a result of fer- tilization and processes subsequent to this, an increase in the energy of the egg of 732, 1,013,300 X<-.00000021 cu. cm: —1.567 eres; that, analogously, after the first segmentation, the energy is O57 << 1,013,300 X .00000021 Cu. cm. = 1.300 ergs. This means, that, as involved in or as identical with the first seg- mentation, there has been a resultant energy decrease, therefore, of .168 ergs; or that it has taken this amount of energy, about } of the total increase resulting from fertilization, etc., to bring about this cleavage. As bringing about the second segmentation, we find by sub- stitution: (6.53 —6.40) X I O13 300 X .000 000 21 =.028 ergs of energy to have been involved. CONCLUSION This completes the computation based on the measurements taken in Experiment VI. It could, of course, also have been made for some of the other experiments, and, had our purpose been to determine as accurately as possible the numerical value of the energy of segmentation, then a large number of both experiments and computations would have been necessary, in order from these to get a mean result. But it has been not the numerical result but rather the method, that is, the practicability, on an experi- mental basis, of applying the “work integral” and so the other equations of which it is a special case, that has seemed the more important and been deemed worthy of emphasis. ‘Thus would I forestall the point of the possible criticism that the numerical results themselves are meager, and that they have been found for only two segmentations on data obtained in one experiment. For, while these are the facts, nevertheless, on the other hand, Experi- ment VI and its results can be regarded as typical of further pos- sibilities, while on the other the limitation to two segmentations 312 E. G. Spaulding may be regarded as due simply to certain difficulties in experi- mental procedure which, of course, further refinements may over- come. Accordingly, I shall consider that that which was my immediate purpose, namely, the application to an organic event of the same general principles as are applied to inorganic events, has resulted ' successfully, and that thus a basis is furnished for answering the other questions which were propounded at the beginning of my paper. However, before that is done, an interpretation must be made as to just what the results obtained show as to the character of the energy-transfer which is involved in each cleavage. Here the principles stated in our introduction and developed in our formu- lation must guide us. In answering this question it must be said, in accordance, first, with what was shown as to the conditions under which our measure- ments must be taken, and, second, with the hypotheses formed as to the forces, etc., in the egg as a system, that the numerical values obtained for the energy-transfer in the two cleavages are the measure, first, of the difference between the energy decrease and its simultaneous increase, E = W+(U,—U,), during the event of cleavage; and, second, of the resultant, in energy-terms, of all those subsidiary processes and changes, morphological and other- wise, which contribute to the event; some of these must be identical with W, others with U,—U,; if there be any processes which do not so contribute either directly or remotely, then, of course, they are not included in this resultant. It is evident, then, first, that the result obtained allows for the possible increase 1n the internal energy of the ovum by the absorp- tion of heat, or other energy, though probably only the former, as simultaneous with a decrease in accordance with which work is done; and, second, that our result gives the measure, not of the entire energy of the cell, but only of that which, as an excess of the energy “lost” over that gained, is identical with the energy of cleavage. What, now, is the character of the energy-form in which there is this resultant decrease? ‘To this the answer is indicated, first, ~*~ T he Energy of Segmentation 313 by the hypothesis formed that in cleavage we are dealing with “forces”? which are efficacious only as perniidat pressures, and, second, by the nature of the factors actually determined by meas- urement, again pressures, that it is the * ‘volume energy” which is so concerned. ‘This “volume energy,” here the energy of the col- loidal solution, is a function, first, of the number of molecules or of particles, and of their velocity, and, therefore, second, of the chemical splittings and combinings, and of the temperature, respectively. With the temperature and volume constant, the decrease in vol- ume energy demands a correlative decrease in the number of mole- cules, or of colloidal particles, or of both, as accompanying cleav- age. ‘This decrease would take place as a result in turn of a com- bining, to a definite degree of course, of molecules and of particles, which chemical change would be accompanied by the passing of energy from the system (ovum) to the environment in the form of heat. At least part of the “resultant” decrease in the volume energy of the system is to be accounted for in this way. Con- cerning the remainder of the decrease the evidence shows that its reappearance is in the form of the increase in the energy of the surface and in the mechanical energy or work done in the moving of the “mass” surrounding the system as environment. Under normal conditions it is with the intensity of the “surface pressure” equal to the opposed intensity from within that an equilibrium of form continues. What now, finally, is the meaning of the fact that it has been possible to determine the energy of segmentation according to the method presented? ‘That meaning I propose to summarize, for I believe it stands firm, even on the basis alone of the limited numer- ical results obtained. As a first step in the demonstration it was necessary to state briefly the principles which it was my purpose to apply,etc. ‘These were then formulated and shown to be epitomized in the funda- mental equation W+(U,-U,) =1 - The “work integral” was then shown to be a special case of this 314 E. G. Spaulding formula, with the result that the successful application of the former to segmentation would mean also the validity of the latter and therefore of the Four Laws for this event. To it there would apply, then, the principles of Determinism, Potential Difference, Conservation, etc. But these laws are, seemingly, largely if not wholly quantitative, while on the other hand the organism, e. g., the ovum, is qualita- tive as well as quantitative. What, then, is the relation of these Laws to the qualities, and what are these? To answer the former question first, it may be said, that qual- ities in both the inorganic and the organic world are, at the same time that they are qualities, also quantities; and quantities are either extensive or intensive. Of qualities certain empirical laws are discoverable, while between these laws similarities are in turn found which lead to the Four Laws epitomized in equation [8]. Thus we get a “natural classification” of laws. From this it will be seen that the generic characteristics, so expressed, have the relation to that from which they are derived of being ultimately incorporate in the concrete qualities, and that they do not, although they are predominantly quantitative, simply exist alongside of these as a separate and distinct aspect. Rather, the Four Laws express the common quantitative aspects of these concrete qualitative- quantitative phenomena. This view is directly opposed to that which regards the Four Laws, because quantitative, as “not touching” the concrete qual- ities, and then finds that these last, because not so “touched,” furnish opportunity, especially in organisms, for Indeterminism, Regulation, Freedom, Entelechies, etc. But what are the qualities themselves? Are they not of things, events and relations? Our answer is: Let “thing” be equated with system; then system implies parts, and these may be either atoms, or coexisting energies, or both. In either case some of the qualities of the “thing” result from the codperation of the parts or elements, whose qualities are different from those of the whole which they form, the test being, that if isolated their qualities are found to be unlike those resultant ones; this bringing about by the parts of qualities which they themselves have not, may be called ciel The Energy of Segmentation 315 “creative synthesis.’’ Other qualities of the system are the same as those which the parts retain when isolated. ‘The latter give an additive result in the complex, the former do not. It is now possible to make a statement as to what the cell is, and, if we may generalize, to answer our major question as to just how different organic phenomena are from inorganic. According to our hypothesis the cell is a system, a complex of energies or of colloidal particles, etc. Some of these components can be isolated, and, with this done, are found to follow the usual inorganic laws; these they are therefore assumed to follow when in the complex. ‘The same assumption is also made for those components which cannot be isolated; that is, the contrary posi- tion that such an “exclusion” demonstrates the presence of an irreducible, organic, vital remainder is held to be incorrect in view of the successful application of the energy-laws to the organism as a whole. The qualities of the cell, are some of them, identical with the qualities of the parts and are the additive result of these, while others are the result of the “creative synthesis” of two or more constituent energies, etc. All these qualities are at the same time quantities, either extensive or intensive. Now, without it being necessary to treat either these energies or the qualities of the system analytically, it has been possible, since at least some of them act together to produce, or are identi- cal with, the event of cleavage, to measure this as a whole and bring it under the Four Laws. ‘Thus are the subordinate events which contribute to this resultant event also brought into the range of the validity of these Laws. The qualities of the organism—which are also quantities—are, accordingly, shown to be qualities which on this quantitative side have certain characteristics which are the same as those of the inorganic world—namely, those characteristics which the Four Laws formulate. Conversely, the Four Laws, as formulating these common characteristics, and as epitomized in equation [8], bring the concrete phenomena, both organic and inorganic, and the series of empirical laws into a “natural classification.” But this does not do away with the fact that here in the so-called 316 E. G. Spaulding . organic realm, as in the inorganic, there are specific qualities which differentiate each class of complex from every other class, or, indeed, each individual from every other. The organism may have, therefore, qualities which, as such, are specifically different from any found in the inorganic realm; a ““reduction”’ of these to inorganic being as impossible as is that of one inorganic quality to another. On the other hand, these very qualities, in that they are at the same time quantities, are like the inorganic in that they have in common with these the characteristics formulated in the Four Laws. In just this respect there is no difference between or- ganic and inorganic; they are in the same realm whatever that be called. The only difference between organic and inorganic which. still remains is, then, just that difference which persists between specific and specific, a difference which holds as good within the inorganic realm as it does between it and the organic. The only ground remaining for holding a distinction between the two realms is, that, taking the same level of classification or compari- son, the differences between certain complexes, called inorganic, is less than the difference between these and certain others called organic. But even this does not do away with the necessity of bringing all into one realm in which the principles of Conservation, Potential, Determinism, etc., are valid. I conclude, then, that all events, both cea and inorganic, take place in full conform- ity with these principles, and that there is no ground for holding or interpreting organic events, etc., to furnish contradiction or evasion of them. FACTORS IN THE REGENERATION OF A COMPOUND HYDROID, EUDENDRIUM RAMOSUM!} BY A. J. GOLDFARB Witu Two Ficures PERE ATYINTE AR VS CCl COLLIE DIE Pe fer crates epeeteie ecole eieTos ace @1S) svete ovale ethers) ole icie ofa cis) atrial eva rae Porters resol 317 2 Effects produced by removing lateral branches or pedicels. ............. 0000s eeeeeee eens 319 Guepbiecosacue tomrepionall differen Ces— age)... (aie. coe). oo sjniaie oe ole\s 9 cisias stsjoieie um ei) s rh) eboiisierele ae 320 Aue Causes and conditions underlying: heteromorphosis ...........-..6+-s-20: sss nseeesaes sine 322 Geecaosatc, its movements and mternal circulation’ . 20... 6.6 och eee cece ee cece sone 330 MESLOIONOLMNANON 4 <2 otc cescileaseciecesivcessseeie tae g Risser! atelosdVaigrd net abo are sober seomteeect ate 334 RI ACERG LACE PENEL AION ae Ye ria nels + Hele «-n.nyerciels, © ope. /0ie © siya eie oinid b aieisioumyebeioreis arse earl eres 335 MULTE CTSA BOL el Vad GY ere yotes foresees olelone mies ale ashen Biviedayaidios sisi e alda's'd) arame apamtore wee Reece 337 f Liheais OF GOMES Sado cobb cos poo eReaos Doe aap oa Ue ay OReeE BananaEnon onobS cabo oGeC 341 PEE EC OM ACKAOR OXY MEIN mic crore) cater ceree o's, 2.0 #/ Ais ielele 6) pe + oT oielaln elo: oc taiei ie oh eld wlls(ecorniater hele weerelions 343 MBE INMOUMCITECEASUIM I PL Garay. fate ote le!clorators jaye) s1ais\cie/eielta¥igs « stoiorejeieis (ote: tue a store aralene wtayellobeforalaavets 345 PEE ET ECLOMOLLEINDCLALULER er-yaloye.cre\efele 6. «ieisiere/s sio\aiaysiere s/se:seicinia elord\ 6 srstoue.s ie ¥ lela, Hiv opernlota ole are ole 346 13. Effects of repeated removal of polyps from the same lateral branches..............-+0+--005 346 BamEnects Of injuries to different parts of the stem... 2.0... 020s. ec wees o reser ones Aooe e460 347 Bpetitectsof diluted and concentrated séa-Wwater ... 200.6. 26 scenes denen s sere ce ceccce™ viele 348 Of CRANTSEG cdo 98 SOC UGE. Cligd pee CO Rice ae tte ee ann ee eee eS RP aaa AR ict hs 353 PRELIMINARY STATEMENT Loeb’s pioneer experiments on regeneration in hydroids, have stimulated a large number of investigators to study the effects of external and internal factors in these animals, especially upon unbranched or slightly branched forms like Tubularia. Most hydroids are affected by the same agencies, but not to the same degree; that while gravity is the determining condition in one hydroid, contact or regional differences or “ polarity,”’ determines 'T am deeply indebted to Prof. Thomas H. Morgan, who suggested these studies and who ren- dered much valuable advice and assistance to me throughout the course of these investigations. My thanks are due Prof. Edmund B. Wilson for the privilege of occupying the Columbia University Table at the Marine Laboratory at Wood’s Hole, Mass., and to Prof. C. W. Hargitt for many valuable suggestions. Tue Journat or ExperIMENTAL ZOOLOGY, VOL. IV, No. 3. 318 A. “f. Goldfarb the kind of regeneration in other hydroids. In the following study of Eudendrium ramosum, I have attempted to examine nearly all the known factors, external and internal, that enter into the life of this hydroid, especially those that take part during growth and regeneration. Eudendrium ramosum consists of one or more main stems, bear- ing pinnately arranged lateral branches which, in turn, branch again and again, hnally ending in pedicels each bearing a polyp. When kept in an aquarium the polyps disappear and regenerate period- ically. A few preliminary experiments made it clear that the method,” previously used, of adding the number of hydranths regenerated on a stem, each successive day after amputation of the polyps, did not give an accurate idea of the actual number of dijjerent hydranths regenerated in a given time. If, for example, one or more hydranths should regenerate on a branch at about the same time that an equal number of other hydranths degener- ated, the records would not show the formation of new polyps. In the following experiments the exact number of different hydranths produced each day was recorded by the aid of daily diagrams of each stem and branch showing the presence or absence of polyps, buds and stolons. Hydranths appear within two or three days after amputation. Later some or all of the regenerated hydranths may disappear to be replaced in part or in whole by new hydranths; or other cut ends, devoid of hydranths, may regenerate them now for the first time. In order to condense into the smallest space the data essential to an understanding of the phenomena, the number of hydranths that appear within three and six days respectively, after the removal of polyps, are quoted in the following tables, unless specifically mentioned to the contrary. When fractions are used the numer- ator represents the number of new hydranths formed in the time stated; the denominator indicates the number of lateral branches or pedicels removed. For convenience these fractions are usually reduced to per cent. It is nearly impossible to obtain stems absolutely alike in all "Light as a Factor in the Regeneration of Hydroids: Goldfarb, Journ. Exp. Zodl., 1906. Factors in Regeneration 319 respects. For practical purposes, stems that resemble each other in size, number and size of branches, that come from similar regions in the colony, and that are removed from their habitat at the same time, will be called “‘ similar stems.” EFFECTS PRODUCED BY REMOVING LATERAL BRANCHES OR PEDICELS Experiment 1. ‘(his experiment was undertaken to determine whether stems, bearing lateral branches but with the pedicels and their polyps removed, would regenerate a greater or less per cent of polyps than stems with all the lateral branches trimmed off close to the main stem. On one side of a large stem the branches were cut off close to the main stem, while on the other side only the pedicels were removed. From a second stem the lateral branches were amputated on both sides; and from a third the polyps only were removed. ‘The records for each stem were as follows: TABLE 1 Regenerated in 3 days 6 days No. of stems Pedicels only Branches Pedicelsonly Branches removed removed removed” removed zZ Z ey 3 MSN elias Voiclials ya's") o.<19)p «ev ois 6.0 eile sbee eve 6 q * 7 34 * 7 es UO ——__. —__’ EMG Tee eWay cia i) pistons) ich s eXe/enelaiere, 6 6s es @ It ath & 4 5 3 14 PEER eT olla eh ota lci.olicac) toc) | sce \ol Ss) ee) 0) site se) oles 36 Ts i> 15 iF 24 10 34 21 55 6 BEMESEU@Nsilel.s a le)lo, 0 so, «0 «) 6 00 ee + ee 8 ces e ce se 4 ay Ge 6 yf ‘TG so. 9 Go oR are Oly ee 24% 35% 42% 777% * The figures for each side of the stem given separately. The conclusion is obvious, viz: that colonies from which all the branches have been removed regenerate more hydranths than those from which nothing but the pedicels and their hydranths were ampu- tated. ‘(his conclusion was corroborated by later experiments. All the lateral branches were removed from stems used in the succeeding experiments. 320 A. f. Goldfarb EFFECTS DUE TO REGIONAL DIFFERENCES?® Experiment 2. Is the tendency to regenerate polyps more strongly developed in one region of the stem than in another, or is the same average number produced in all regions of the same size? Pieces from a series of large stems were compared and the number of polyps produced in each was separately estimated, viz: (1) The basal end of a stem, about one-tenth of the whole stem, (2) the basal half of a second stem, (3) an entire third stem, (4) and the two halves of this stem separately considered. TABLE 2 Number of hydranths regenerated on Distal half of Basal half of Entire stem entire stem entire stem Basal half Basal tenth 6 6 0 5 { x0 10 16 g ° lea 3 oO 205 4; daye...25% | 5 ; 8 10 2 13 12 aus O02 aE 12 12 10 ° Average, 36% 70% 3% 17% 0% dlls 14 Bie 6 ( 20 10 10 9 oats! 9 2 4 6 day Seah CROs y Té = = To [ 22 iW Som Hee 24 12 12 10 Average, 83% 133% 33% 44% 0% No regeneration occurred on the small basal pieces until the seventh day after amputation. Even then very few polyps ap- peared. ‘The basal halves regenerated 17 per cent, the entire stems much more, namely, 36 per cent. More striking, however, is the difference in the regenerative power of the basal and distal halves of entire stems, for 70 per cent regenerate on distal halves, but 3 per cent on basal halves. ‘The figures for six days reinforce these conclusions. Smaller stems, however, do not reveal this sharp contrast in the regeneration of the two halves of stems. Experiment 3. Similar stems were cut into three nearly equal parts. The distal thirds regenerated two days after amputation; — most of the middle pieces did not regenerate till the third day, and © the basal pieces, not till the third or fourth day. The question of ’Some very interesting facts in this connection are given by Gast and Godlewski in Die Regula- tionserscheinungen bei Pennaria cavolinii, Archiv f. Ent., Bd. 16, 1903. Factors in Regeneration 321 rate of development will be discussed later. For the present it will suffice to state that because of this difference in the rate of devel- opment, the latent period’ was not computed and the number of complete polyps produced two days after their first appearance (which may be the fifth or sixth day after amputation), and the number produced within the next three days, were recorded. TABLE 3 Regenerated Average on 2 days 5 days 2 days 5 days Distal thirds 24754 $crSFi/2 FFE REVERE 66% 100% Middlethirds #§ 538 § Set /FRMSERBRAS LE O1% 101% Basalthirds 7 $3 sSr01r 33s)? $93 F torr F FS 42% 777% We may conclude that the distal and middle pieces regenerate practically the same number of hydranths but far in excess of the basal pieces. Experiment 4. ‘The last experiment was modified to the extent of using not the main stem but the /ateral branches from the distal, middle and basal regions of the stem. ‘The distal branches were small and delicate, quite different from the middle and basal branches which owing to the smaller stems used, resembled each other closely. : TABLE 4 Regeneration on 2 days 5 days Branchestirom distal region; 62 .c'c sense sees ces es 37% 57% Branches from middleregion|... .. :..5..2+.es000005 50% 7970 iBranchessiLOMNbasallrePlOM oa) cle esre ss eee ss se 68% 104% Regeneration of lateral branches differs at different levels. It is greatest on branches taken from the basal regions (of small stems) least on branches taken from distal regions. An expla- nation of these phenomena will be attempted under the caption of “coenosarc.”’ From the evidence already cited we may summarize as follows: T he distal half of a stem regenerates a much larger number of polyps than the corresponding basal half. On the contrary, branches 4See Rate of Regeneration. ang A. “f. Goldfarb from the apical region give rise to fewer polyps than the branches from the middle and basal regions. While polyps at the apical end of stems are common, they are rare on the apical ends of branches.® HETEROMORPHOSIS The phenomenon of heteromorphosis® in hydroids, has been investigated by Driesch, Loeb, Morgan, Stevens and others, par- ticularly upon unbranched or slightly branched colonies. It was hoped that experiments upon a much branched hydroid like Eudendrium ramosum, would afford further insight concerning: 1 The conditions underlying the formation of heteromorphic polyps. 2 The effects of such polyps upon the regeneration of other polyps on the stems. 3 The rate of development at different levels of the stem. 4. The ccenosarc, its movements and its effects in the produc- tion of heteromorphic polyps. A great many observations on large stems of all kinds had shown that polyps are rarely produced at the basal region, partic- ularly at the basal cut ends, though common enough at middle or distal regions. Smaller stems or pieces of large stems—bear- ing from 15 to 20 branches—regenerated more basal hydranths than the much larger stems, though still less than the number of apical polyps. Experiment 5. Large stems with lateral branches removed were cut into three nearly equal pieces. Further details are given in Experiment 3. The number of hydranths regenerated at the oral and the basal cut ends of each piece, two and five days after their first appearance, are given in the following table: ®See Gast and Godlewski, loc. cit. °The following papers on Heteromorphosis and Polarity give various hypotheses to account for these phenomena in hydroids: Bickford *94 J. Morph.; Driesch ’92 Biol. Cent., ’96 Vierteljahrs-schr. Nat. Ges. Zurich, ’97 Archiv f. Ent., ’99 Archiv f. Ent.; Loeb ’91, Ueber Heteromorphose, Wiirzburg; Mor- gan or Biol. Bull., ’or Archiv f. Ent., ’04, ’05, 06 Journ. Exp. Zool.; Morgan and Stevens ’o4 Journ. Exp. Zodl.; Stevens ’o2 Archiv f. Ent. Factors in Regeneration 32% TABLE 5 Polyps regenerated on 2 days 5 days iGraltends seve ste ls oe ee ete eae 60% 60% Distal third of st ¢ pepo ot etem Nhrasaltendss. ite isos etree ce eke iene ee ne 2 30% 100% Bele eiediof stem Joral Celta: So aoe ACPA See ye rE ER a enn ae ats 80% 100% \ basal 7c heads ae en RA ce ee aK Aes MRT en ESI 40% 70% OL alentSmeereierercis coe oe cron ain earch oti ee 70% 100% Basal third of st f seo d/ot stem \ basal BIO (0 ES ca pkey a ee REO EE tac cee ea Bei 20% 30% With distal pieces excepted, the branchless parts of a stem regen- erate decidedly more hydranths at the apical end than at the opposite or basal end. Experiment 6. ‘The above results contrast sharply with those in this experiment in which only Jateral branches from different regions of the stem were used. TABLE 6 Polyps regenerated on 2 days 5 days (oralkend Simtetas eat. art, tine, % 37% Branches from distal part of stem Wiese age 370 \ basal (TYG i mea an Crees Sean ae a IS D5 gee 37% 50% Rend) eyed tee ee iti % q Branches from middle part of stem [ora Sieg me ee | basal ETS Rt Rese. eee edhe borers oe ra ae 62% 112% end sheets ol ce pk tiiiis aw nae eet % %, Branches from basal part of stem ie igs ie ee | basal CTO SEA Re tet see aie er nicks a ee 377%. 87% The lateral branches from any level of a stem produce a far greater number of heteromorphic than apical polyps.?- Under the caption “ccenosarc’”’ this will be explained. Experiment 7. In Experiment 6 all secondary branches were. removed. In the following experiment these were not removed nor were the polyps amputated. ‘The regeneration at the basal and apical ends only was recorded. While the polyps were dis- integrating, stolons were forming at all the basal cut ends, fasten- ing the pieces to the bottom or sides of the dish. Later stolons were present on several pieces at both oral and basal ends. ‘These stolons often grew to a remarkable length as long as or longer than the original specimen. Usually they gave rise to several branching stolons which regenerated one or more polyps, even as 7Driesch, Morgan and Stevens found that the apical ends of pieces taken from any region were first to regenerate and produced a greater number of polyps than the basal ends. 324 A. “f. Goldfarb many as nine polyps were present at the basal end of one piece. These hydranths did not appear before the third or even the fourth day after removal of the branches. ‘The following table gives some of the details of the experiment. TABLE 7 Polyps regener- No. of branches Polyps regenerated ated atoralends that regenerated at basal end of 3 days 6 days 10 days in 10 days basal polyps Distalibranches\..-... 2s ° 80% 160% 0% 90% Middle branches .......... ° o% 100% 10% 7°7% Basal beanches 5-7... 4-- = ° 50% 120% 20% 50% The presence of polyps on the branch retards regeneration but does not prevent the formation of polyps at the basal cut ends. The figures, particularly in the last column, of the above table seem to indicate a maximum tendency toward the production of heteromorphic polyps on distal branches, less and less on middle and basal branches, respectively. SUMMARY The mid and basal thirds of a stem behave quite differently from the distal third and from the lateral branches, in so far as the rela- tive number of hydranths regenerated at the oral and basal ends is concerned. The distal region of stems and the lateral branches from any region produce a greater number of heteromorphic and fewer apical polyps, than do the median and basal thirds of the stem. Lateral branches from which polyps had not been removed likewise produce more polyps at the basal than apical ends. Experiment 8. In the preceding experiments the presence of lateral free cut ends may have introduced disturbing factors not yet fully considered. In order to avoid these influences, some or all of the lateral cut ends were ligated ona series of similar stems as follows: a the apicalend only; . b the apical end and the two lateral ends on distal half of stem; c the apical end and all the lateral ends; d all the lateral ends; Factors in Regeneration 325 e all the lateral ends and the middle of the stem; 7 the middle of the stem only; g all the lateral ends, then stem was cut into two equal parts; h not ligated—control stems. The rate of development in the different series varied consider- ably. In the table the number regenerated for two and five days after the first appearance of polyps is given. TABLE 8 z E Z » 4 a a 4 Polyps reg. on in @ralfends)..:5)5---- 82% 66% 87% 88% Lateral ends...... 54% 25% 52% 44% ZGays).., 4 Basalends ....... 61% 75% 715% 100% 88% 100% 100% Total Average.. 66% 70% 62% 94% 70% 72% 100% Oraliendsize ......-1- 123% 116 125 155 Lateral ends ..... 94% 100 66 76 midayse-s-) eep enh spereiat as 9% 39% Basallendsie s:ione aes aoe meee cet elelerenre tien) atta iei l= 16% 52% In these pieces, each about 1 mm. long, more basal polyps are produced. A considerable number of internodes formed hy- dranths at each end, in some casessimultaneously. This fact indi- cates that some of these very short pieces had the potency to pro- duce more than one hydranth. The majority regenerated but a single polyp. Regional differences were not apparent in this experiment. It cannot be said, for example, that regeneration was retarded or accelerated, increased or decreased in one region more than in another; nor was the formation of basal hydranths pecu- liar to any one region or level. From these experiments we conclude that in the small preces mentioned, a greater number of heteromorphic than apical polyps are produced; that this increase ts not associated with any particular level of the stem; that oral polyps are rarely formed on distal pieces, though many regenerate on pieces from middle and basal regions of the stem. Experiment 12. Yo determine whether rapid growth first inone direction then in the opposite direction, could be effected, the new basal stems from a number of pieces were removed, viz: B pieces of the diagram. ‘There developed from the basal end 4, of these basal stems many hydranths which grew with remark- able celerity in the original direction toward a; there also regen- erated actively polyps at the oral (c) and the lateral ends of the B pieces. [here was no reason to doubt that the removal of the C pieces would result in the formation of hydranths at the new basal end of C. If lateral branches instead of basal branches be repeatedly removed the results are essentially the same. The actual figures are given below: TABLE 13 Polyps regenerated on basal pieces of B 3 days 6 days 9 days Oraliendse foo at 5 ce aon eer eee 0% 26% 53% (Basaliends xs scscis craic untscote ee otra aeterattcona pretote 33% 40% 53% Polyps regenerated on basal ends of lateral branches OraliendS: >.ctnciite cn oct eo eee ees 6% 36% 46% Basal end sic. a tog ee eee aseoac 0% 23% 50% Factors in Regeneration 329 The above tables give the number regenerated at the basal ends of pieces 3, 6 and g days after amputation. ‘The oral (and lateral polyps also) had not been cut off but disappeared within two days. Regeneration, therefore, at oral and lateral ends could not take place till one or two days after the basal polyps appeared. Regeneration in one direction, viz: from the basal end of B piece does not inhibit regeneration in the opposite direction at c; rapid growth and differentiation take place synchronously in op posite _ eer ral end 4 oral Ari eee re Mee basal. end oe ea et oe tee 1G ae end ee Fig. 1 directions. ‘This would lend weight to the view that regeneration is not dependent on internal changes of the entire stem or even parts of the stem but is rather due to reactions at the cut end only.’ An examination of Tables 14 and 15 will make itclearat a glance that gravity is a considerable factor in determining the increase or decrease of the number of hydranths at the upper ends (toward the zenith), regardless whether such ends be oral or basal. Erect 8Except for minute pieces, see also Experiment 25 on this point. 330 A. fF. Goldfarb stems regenerate in 3 days, 50 per cent at the oral and 45 per cent at basal ends. Inverted stems—with basal ends pointing toward the zenith—in the same time, regenerates 35 per cent at the oral and 60 per cent at the basal ends. ‘The figures for 6 days empha- size the point, viz: Erect stems regenerate 50 per cent at the oral, 80 per cent at basal ends, inverted stems 35 per cent at oral and 115 per cent at basal ends. Injury to different parts of a stem does not affect the regenera- tion at the axial ends of the stem. Nor does increasingly diluted or concentrated sea-water influence the number of hydranths at basal or oral ends. Heteromorphosis occurs independently of these influences. COENOSARC® The ccenosare may readily be observed in the more distal parts of a colony, for the perisarc there is thin, almost transparent and seldom covered with débris, polyzoans or hydroids, commonly found on the more basal regions. ‘The ccenosarc is com- posed of a hollow cylinder surrounded by the perisarc. “The lumen, within the ccenosarc is cir- cular in cross section and differs from Tubularia® in which the lumen 1s divided into two almost sepa- rate compartments by a central partition along the whole length of the stem. It has already been pointed out that the number of hydranths regenerated at the oral or at the basal ends, depends upon whether lateral branches or stems were used, and if the latter, upon the re- gion from which the piece was taken. Previous experiments and direct observations prove conclu- sively that the cenosarc withdraws from the distal Fig. 2 “Gast and Godlewski’s account of the coenosarc in Pennaria cavolinii is extremely interesting in this connection. Hargitt, G. T.,’03 gives an account of the histologic structure of the coenosarc. Archiv fiir Ent., Bd. 17. ‘Stevens, N. M.,’o2 records movements of the coenosarc in Archiv f. Ent., Bd. 15. Stevens ’or, ’o2, Archiv f. Ent., Bd. 13 and 15. Factors in Regeneration 331 ends of all branches, cut at any level, and from the distal ends of pieces from the apical region of the stem. The cenosarc also withdraws, as a rule, from the distal ends of small pieces from any part of the colony. In the diagram of a large stem a represents the oral, / the basal end. We can foretell with a fair degree of accuracy whether polyps will be formed at one or the other ends of pieces and the relative number regenerated at each end, provided we know the level at which the cuts were made. In pieces bc, no hydranths will appear at b, many at c; in pieces f h or g i more hydranths will appear at 7 and g, respectively, than at / andz. In region af the nearer to a the oral end of a piece lies the fewer the oral polyps; in region ; / the nearer to / the basal end of a piece lies the fewer the basal hydranths regenerated. The recession of the coenosarc from the distal end may extend only one internode or half a dozen or more, and the hollow peri- sarc thus produced often cracks and breaks off." In small pieces the coenosare may move not only toward the basal end but through the basal end entirely free from the perisarc, leaving the empty perisarc behind. Placing inverted stems in sand accelerates the basal movement of the ccenosarc, so that the parts embedded in sand become entirely empty; the coenosarc is found only in the basal regions surrounded by water. In erect stems under the same conditions the ccenosarc withdraws somewhat from the dis- tal end while at the basal end it either (1) does not withdraw at all, in the majority of cases, and in spite of the adverse conditions, (2) or slightly withdraws, (3) or disintegrates, the result of the ravages of large numbers of ciliate protozoa. Two counteracting tendencies may be said to be present at every cut end of a stem or branch, the resultant of which deter- mines whether a hydranth will or will not be regenerated, and whether regeneration will or will not be retarded; first, the movements of the ccenosarc from the distal cut ends already _ described, second—and I believe second in point of time—the regeneration of new tissue, which is negatively geotropic and, NGast and Godlewski, loc. cit. 332 A. Ff. Goldfarb therefore, tends to grow upward. The more ccenosarc present near the cut end, as determined by its width and density, the greater the regeneration. ‘The less impeded, the greater the movement of the coenosarc. Whether enough ccenosarc is pres- ent or is regenerated, to make up for the recession of the coenosare will determine whether or no hydranths will appear at the cut end. The basal movement brings additional ccoenosare to the basal end, condenses it greatly and thereby increases the number of basal polyps, on small pieces, on lateral branches and on pieces from the distal region of stems. The ce@nosarc in the middle and basal parts of the stem does not withdraw from the oral cut end and, as a matter of fact, there are more hydranths at the oral than basalends. In this hydroid atleast, we donot need to call to our aid “formative stuffs,’ and other hypothetical internal forces and materials to account for heteromorphosis, for the movements of the coenosare account for the presence or absence of polpys at some levels and not at others. The coenosarc of stems kept long in the aquarium often becomes fragmented. Fragmentation is the result of a splitting of the coenosare at one or another of the internodes; the coenosarc thins rapidly at these points, and finally breaks into pieces entirely independent of each other. Each part may move into the nearest lateral branch and give rise to polyps, or it may remain in the stem. Not infrequently in the latter case it contracts at both ends ulti- mately forming an ellipsoidal dense mass of coenosare near the basal end of the stem. It is believed that these hydroids winter over in this contracted condition. When stems are subjected for a long time to adverse conditions the coenosare forms this dense ellipsoidal mass. When such stems, which could no longer be made to regenerate, were cut into smaller pieces, polyps regen- erated provided the ccenosarc was injured. With a low power of the microscope the lumen within the coeno- sarc 1s seen to be filled with a colorless fluid in which myriads of | colorless granules float. These move slowly toward one end of The hypothesis of specific stuffs, moving in definite directions, was developed by Bonnet, later by Sachs, and still further perfected by Loeb. For criticism, see Morgan ’o1 Archiv f. Ent., Bd. 11, *o2 ibid., °04 Journ. Exp. Zoél. Factors in Regeneration 333 the stem; in pieces 1} inch long the trip from one to the opposite end takes from 1} to 3 minutes. As the granules accumulate at this end the stream moves more and more slowly until it finally ceases. [here may be a respite of a few minutes and then the stream courses in the opposite direction at first slowly then faster and faster, and finally slowly again as the granules pile up at the other end. The performance is repeated over and over again; the time for each trip may vary considerably. Now and then groups of granules are violently whisked about or “‘tremble,” the result of ciliary movement of the endoderm. Granules were never seen to pass out of the prostomium of the living polyp. After disintegra- tion of the polyps dark red masses" were frequently observed marking the spot where the polyps had been. The current in a lateral branch may be continuous with that in the main stem, moving at the same rate and in the same direc- tion, or it may be independent, and even contrary to the stream in the stem. The central stream may be continuous or divided into two or more independent streams. Though the ccenosare in branches and stem is continuous, the streams within the coenosarc of these parts may behave independently. ‘The contin- uity of the coenosarc may be permanently broken by pressing a stem firmly with the side of a needle. The ccenosare is cut into two parts which separate more and more from each other. If gently pressed, a dent is temporarily produced in the ccenosarc which slowly recovers its normal shape. The number of hydranths that arise from the basal cut end of pieces is closely associated with the amount of ccenosare near such ends. Other conditions being equal the more ccenosarc at or near a cut end the more hydranths produced. Several condi- tions must, however, be taken into account. A large stem does not necessarily contain more ccenosarc per unit of length than another one-half as long. Much depends on the more or less con- tracted condition of the ccenosarce. If it be attenuated, as in rapidly growing branches, it will have per unit of length less regen- erative potency than the more concentrated ccenosarc usually These dark red bodies are probably analogous with the red bodies resulting from the metabolic changes in Tubularia, studied by Bickford ’94; Stevens ’or and ’o2,and Loeb ’91, and Morgan loc. cit. 334 A. Ff. Goldfarb found near the basal region of stems. Careful measurements of the diameter of the ccenosarc at different levels of large stems actually shows that normally the diameter of the coenosarc de- creases toward the distal end and conversely increases basally. The relation between size and the number of hydranths regen- erated, particularly at the basal end is shown by the following ob- servation. Pieces less than I mm. long never produced a complete polyp, though they often regenerated shoots at one or both ends. Pieces as small as two-fifths mm. long developed shoots at one end. Larger pieces I to 2 mm. long may regenerate one polyp at each end though usually at the basal end only. Still larger pieces regenerated two or three polyps from one basal end, whereas larger median or basal pieces produced as many as nine polyps from a single basal end. STOLON FORMATION Little has been said concerning stolon formation, partly because of its comparative rarity, partly because stolons are often with difficulty distinguished from branches. A stolon,* root, or hydro- thiza is an outgrowth, positively geotropic and stereotropic in its reaction, which, when young, fastens itself by a sticky secretion to solid objects, and which does not directly give rise to polyps. In nature the stolon or stolon system anchors the hydroid. Less frequently, stolons may join two stems and sometimes the coeno- sarc of stolon and stem may fuse.* In the laboratory, stolons may appear at any of the cut ends. Though most frequently observed at the basal, they may appear at any lateral or even oral end, singly or in groups of branching stolons. ‘They may appear simultaneously at the oral and basal ends, or at one end only. The stolon may sometimes grow to a great length; in one instance a stem 27 mm. long regenerated a basal stolon 40 mm. long. Stolons may give rise to lateral branches usually pinnately arranged, which like pedicels end in hydranths. Some un- branched stolons after a time bend at their very ends and regen- 1¢The production of stolons has been shown in some species to be determined by gravity (Loeb ’91), in others by regional peculiarities (Stevens ’02), contact (Loeb), exhaustion (Driesch), by the kind of regeneration at the opposite end of piece (Morgan ’or). 15See Stevens ’o2. Factors in Regeneration 335 erate a polyp at the tip. In these, it is impossible to tell where the stolon ends and where branch begins. It is nearly always dificult to tell in advance whether a large growing shoot will ultimately become a branch or a stolon. Stolons in my experi- ments never regenerated stolons; when cut, only hydranths were produced at the cut ends, irrespective of the level at which the cuts were made. When stems are subjected to adverse conditions, the cut ends may regenerate new tissue, which becomes surrounded by a sticky perisarc, which does not differentiate into perfect hydranths. The new tissue is really a modified stem, which under these adverse conditions may increase in length and is then called a “stolon;” when it is amputated or when it grows into a favorable environ- ment the distal end frequently differentiates into a hydranth. When a large number of stems were placed in a shallow dish of water containing much débris, and the water was left undisturbed for many days, very few polyps appeared, while a remarkably large number of stolons were produced. When the water was frequently changed, however, the “stolons” invariably bore polyps. As the colony grows older, the perisarc of the stolons becomes thicker, less plastic and encrusted with débris. If the conditions remain constant, the “stolon”’ functions permanently as an anchor- ing organ and can no longer of itself produce polyps, though it has the ability to do so, if cut and removed to a favorable environ- ment. ‘Stolons” are not limited to any particular region nor is their formation influenced by gravity, size of the piece, kind of regeneration at opposite end, etc. “heir presence or absence, in this hydroid at least, is not an indication of the presence or absence of certain internal changes and is, therefore, useless as an index of the polarity of the stem. ce RATE OF REGENERATION®® The term rate is here used to designate the interval between amputation and that point in the differentiation of the regenerating 18Loeb (’91) determined rate of development and rate of growth for different hydroids. Morgan and Stevens ’04, Journ. Exp. Zoél., made careful observations on the rate of development at basal and oral ends of pieces. 336 A. Ff. Goldfarb tissues at which the tentacles of the new hydranth are clearly discerned. Under normal conditions two days is required. Un- der adverse conditions regeneration may not take place for four, five, six or more days. Regeneration does not extend over this entire period, at least in so far as visible changes are con- cerned. ‘There is a latent period, during which no visible changes obtain and which normally covers but a few hours, but which under unwholesome conditions may extend over several days. Once regeneration begins development proceeds normally in about 12 hours. So that adverse conditions increase not the actual period of regeneration but this latent period. In large stems the younger (distal) portions always regenerate polyps at least one day before the older (basal) parts. When stems were cut into three equal parts, Experiment 3, new polyps appeared on the different pieces in the following order: TABLE 14 Polyps regenerated on 2d day 3d day 4th day Distalithirds een © se. see sce secre 8 polyps additional polyp 1 additional polyp Middlethirdsiy: 2 ove secs ee Sees ee ° 9 polyps 1 additional polyp ‘Basalthirdst pm sree 2. Metres eee eee ° 6 polyps 4 additional polyps T he distal thirds regenerated polyps earlier than the middle pieces which in turn regenerated polyps before the basal thirds. To determine whether the two cut ends at any level of a stem regenerate at the same time, the above pieces of stems were used. A, B,C represents the distal, middle and basal pieces, re- A by spectively. At the cut 4 c, b is the basal end of the distal b piece, and c the distal end of the basal piece. It was found ; that the } ends on J pieces regenerated on the average in B F practically the same time as the c ends on B pieces, and | similarly for the b and c ends on the B and C pieces. It c was found that the } ends regenerate first, just as often C 7 as the c ends and as often as the b and c ends regenerated simultaneously. Although the distal region of a large stem regenerates polyps within two days or at least one day before the basal region, yet small pieces cut from the distal region (Experiment 10) never pro- duce polyps before the third day and often not till the fourth day. Factors in Regeneration Ze 4) This retardation occurs irrespective of the region of the stem from which the pieces were taken, and is due solely to the small size of the pieces. Medium size pieces (with 10 to 15 lateral branches) regenerate polyps at all the cut ends, including oral and basal ends, in approx- imately the same time. Ligating the distal end or ends acceler- ates basal regeneration. Ligating the distal end of small pieces taken from the distal region of a stem, does not however accelerate basal regeneration, because the ccenosarc withdraws from the distal end and the ligature does not affect it. When the single lateral branch of small pieces (Experiment g) was ligated the rate of development was uninfluenced. Regeneration on small old (basal) pieces taken from large stems, is very slow; it may take six or seven days before polyps appear. Embedding inverted stems in sand and to a less degree suspend- ing them, stimulates the early formation of heteromorphic polyps, but the rate of development at other cut ends was not at all or but slightly affected. Lack of oxygen or low temperature or contact with a solid body or greatly diluted or slightly concentrated sea- water retards the development of the first formed polyps. In the meantime the stems become more or less acclimatized to the new conditions and polyps thereafter are regenerated at a normal rate. Injuries of various kinds, such as lacerating and slitting the stems in many places or disintegration of coenosarc in some of the lateral branches, does not retard regeneration at the other cut ends. All efforts to accelerate the normal period of regenera- tion in less than two days, failed. EFFECTS OF GRAVITY" Experiment 13. A series of stems were suspended vertically in a dish of water, some with the distal ends pointing upward (toward the zenith), the “erect’’ stems; others in the contrary direction, the “inverted” stems. “The controls were placed hori- WLoeb (’91) believed that gravity and light were the external factors that determined the kind and direction of growth. Also see Driesch (’99). 338 A. F. Goldfarb zontally on the bottom of the dish. ‘The rate of development was practically the same in the three groups. TABLE 15 Polyps regenerated in 3 days Regenerated in 6 days on Oral Lateral Aboral Total Oral Lateral Aboral Total rect Stems. < ster oe oe 50% 54% 45% 53% 50% 86% 80% 83% Inverted Stems ss255-0.0- 65% 35% 43% % 44% 35% 75% 15% 75% Control Stems: $565 225-2: 33% 37% 44% 37% 33% 48% 66% 48% (In Contact) ov ob branches. sed) ase -2 ce eres cease mete ogee sania eee eee 574 An analysis of these figures gives some interesting details. In the first place, erect stems regenerate a greater total of polyps than the corresponding inverted stems. Secondly, erect stems produce a greater number of oral and a greater number of lateral polyps than do the inverted stems. ‘The lateral branches of inverted stems that bear polyps bend upward toward the basal end of the stem, and these branches are invariably longer than on the erect stems. Thirdly, by far the greatest number of basal polyps are produced on inverted stems, from which they grow upward, and directly opposite to the rest of the stem. If the basal branch of erect stems are long they too bend upward. No emphasis is laid on the control stems in this experiment, as they were influenced by contact with the dish, a disturbing factor at the time not fully appreciated. Experiment 14. Fine sand thoroughly cleaned was put in a dish of water. Erect and inverted stems were embedded in the sand to varying depths. The controls rested horizontally on the sand. The diagrams in the table show the position of the stems; the horizontal lines represent the level of the sand, the parts below it represent the parts embedded in sand, the parts above the line, surrounded by water. The conclusions from the previous experiment were more than corroborated. Regeneration on inverted lateral branches is largely inhibited. This is strikingly illustrated, on comparison of columns 2 and 4 and more particularly columns 3 and 5, in which the same number of lateral branches are free, but erect in the one series, inverted in the other. ‘The erect stems of series 4 tite rar, Factors in Regeneration 339 regenerates 33 per cent, the corresponding inverted stems of series 2 regenerates 8 per cent. Similarly, series 5 produces 36 per cent, the inverted ones 7 per cent. The branches in the two sets are too close together to warrant the belief that the difference in posi- tion on the stems can account for the large difference in regenera- tion. When inverted lateral branches do regenerate polyps, they invariable turn upward. Embedding stems in sand affects regeneration just as a ligature tied around the stem at the level of the sand would do, for in both TABLE 16 2 x : aA 4 2 3 ¥ s ‘ Polyps reg. on , in ( Oraliendsis. ach): . % % % 0% 42% 8% | Lateral ends ...... 8 7 33 36 13 3 days... { Aboralends ....... 33 58 66 25 MRO tale, foe ssre acs: 33 24 18 28 37 =) 714! Oraliend sya eirel- 50 58 33 Lateral ends ...... 38 21 41 55 28 Aboralends....... 74 140 158 58 6 days... — — — — — = eB Otaltysise.s waists 74 73 48 44 56 35 Mopaltmcnaberfostemse cer cesT aekcee cleo ele eee COTE ees 72 Mictalnum ber olbranchesi.q.q)0,00- + saens aemiede oe eeees bee cee: 281 cases regeneration at the free axial end (oral or basal) of the stem is stimulated. ‘The number regenerated at the upper ends of these stems depends on whether erect or inverted stems were used. Regeneration at the basal ends of inverted stems is very much greater than that at lateral or oral ends, for example, 58 per cent are produced at the basalends,8 percentat the lateral ends, o per cent at oral ends, in one series. Again 66 per centregenerated at basal ends, 7 per cent at lateral and 42 per cent at oral ends of a second series. Furthermore, the totalnumber of polyps produced ona stem 340 A. fF. Goldfarb depends on how much of it is embedded in sand. ‘The deeper the inverted stems are embedded, the proportionally larger total num- ber is regenerated and vice versa. On the contrary, the deeper erect stems are embedded the proportionally fewer polyps produced. Experiment 15. These conclusions were corroborated and extended by the following experiment in which the sand was banked at an angle of 45° to the horizontal. TABLE 17 Sa TR : Polyps reg. on in ( Oraliends).<5.--- 25% % 25% 50% % 66% 50% Lateralends .... 54 80 75 75 ° 66 56 3 days... 4 Aboralends ..... 50 100 50 100 83 { Wotalere = ere 50 83 73 7O 50 66 57 ( Oralends =~ =: =: 50 50 75 66 83 Lateralends .... 79 96 94 122 ° 83 92 6 days ... { Aboralends ..... 50 150 100 ~——- 100 116 | Total. 250 75 go 95 116 50 i) 93 SUMMARY. Hydranth bearing branches turn toward the zenith whatever the position of the stem, whether erect, inverted or inclined. Erect stems embedded in sand regenerate the largest total number of polyps, inclined stems less and inverted the least number. Regen- eration at the oral ends and lateral ends of similar inverted stems 1s largely inhibited, but at the basal ends is remarkably stimulated. It will be recalled that the coenosare withdraws from the distal cut ends of erect stems; but a ligature or its equivalent embedding in sand, causes the coenosarc to move upward and regenerate an oral polyp. On inverted stems the ccenosare under these influences, and particularly under the influence of gravity, with- draws not only from the distal end but from nearly all the lateral ends, toward the upturned basal end, resulting in an immense Factors in Regeneration 341 basal and almost m7/ lateral regeneration. Furthermore, the total number produced varies with the amount of free coenosare (not embedded in sand). EFFECTS OF CONTACT!® Experiment 16. Stems were suspended on two_ horizontal threads, others were in contact with the bottom of the dish. Hy- dranths appeared on the two series at the same time. TABLE 18 Polyps regenerated on 3 days 6 days No of branches SHS PEM CEC ESCELASH ats rocrersrie terse va.sahelatel sree aie 46% 68% 248 SPeMSiOn/DOLLOMUOL ISH) =. S...jof0 ectss ee mae oie 41% 48% III This experiment agrees with No. 13, in that stems surrounded by water and otherwise under identical conditions produce a larger number of polyps than those in contact with a solid body. Experiment 17. Shallow V-shaped grooves were made in a cake of parafhn of such a width that when stems were placed therein the lateral branches were in contact with the sides or bottom of the grooves. A record of each lateral end was made, whether (a) it pointed upward and out of the groove and therefore not in con- tact with the parafhin, (b) it extended sideways and touched the sides of the groove, (c) it pointed downward, and therefore in con- tact with sides or bottom of the groove. In six days only 23 per cent of all the ends regenerated polyps. Further details will more clearly illustrate to what degree contact suppresses hydranth for- mation. During the first six days but 7 per cent of all downward pointing branches (in contact), 8 per cent of all sideways pointing branches (in contact), 43 per cent of all upward pointing branches (free) regenerated polyps. “Two hundred and thirty-six branches were used in this experiment. Experiment 18. Stems were put within glass tubes, I to 1} mm. inside diameter. Oneor both apices of some stems extended beyond the tube, in others several branches protruded beyond the tube, or short glass rings were so arranged that some of the lateral branches were free between the rings. In nearly every instance 8Some hydroids react more quickly and more readily to contact than do other hydroids, Campan- ularia and Pennaria more than Eudendrium. See Loeb ’gr. 342 A. F. Goldfarb the lateral branches within the tube were in contact with the glass. The results were very decisive. TABLE 18 No. branches Polyps regenerated on 3 days 6 days removed Cutiends; withinithe tube <2. qacieciee aces 0% ° 88 Cntends; outside otjtube s/s. a. se eraee ae 56% 116% 34 In no case did regeneration occur on cut ends within the glass tubing. After six days these stems were removed from their tubes, and regeneration then proceeded normally at nearly all the cut ends. Experiment 19. Other experiments under somewhat different conditions illustrate further the inhibition due to contact. Each stem was rolled in a thick sheet of cotton, which was then immersed in water. No hydranths regenerated. Cotton abouthalf as thick was used and several hydranths appeared in three days, but by the sixth day they were gone and did not reappear thereafter. By using cotton one-half as thin again, a still larger number of polyps were produced, which disappeared less rapidly than in the pre- ceding case. ‘The results were similar when stems were placed between two flat layers of cotton,.so arranged that a stream of sea-water continually flowed through the cotton. When removed from the cotton, the stems regenerated readily—provided they had not been kept too long in it. Experiment 20. On one side of a dish of sea-water stems were loosely placed, on the other side a larger number of stems were crowded together into a groove 4 mm. wide by 5 mm. deep and somewhat longer than the length of the stems. ‘Table 20 gives the detailed results. TABLE 20 Indicates the total number of polyps observed on the following days 3d 4th sth 6th 7th 8th ogth roth 11th No.of stems 1 Crowded stems........ 2 2 I 2 ° ° 2 I 2 10 2 Notcrowdedstems.... 10 23 10 3 6 6 9 16 6 3, Crowded stems ........ 20 15 7 4 4 6 10 12 39\ = 12 4 Not crowded stems..... 40 35 20 27 36 22 20 20 19 7 *Stems had been scattered. Thecrowded stems regenerated only at the free unentwined branches on the top of the pile, and regeneration was less than in Factors in Regeneration 343 the scattered stems. When the crowded stems were removed from the groove and separated from one another, there was a mark- edly increased regeneration. Experiment 21. Dr. Louis Murbach kindly permitted me to use an ingenious apparatus devised by him, by means of which a stream of air bubbles was introduced at the bottom of a vessel so that the contained water was in constant agitation. Stems placed in the vessels were whirled around making about 35 revo- lutions per minute in one vessel and about 28 in the other. Very little regeneration occurred (about 8 per cent), somewhat more in the slower stream, slightly less in the faster one. “Two days after the appearance of polyps they were gone, and no more regeneration occurred. If the current of bubbles was stopped for 12 to 24 hours, some regeneration would take place. After ten days the stream was stopped altogether, and there resulted a constantly increasing number of hydranths. Whirling stems through water at a com- paratively rapid rate affects regeneration in practically the same manner as contact, already discussed. Contact 1s unfavorable to and more or less suppresses the develop- ment of polyps. It matters little whether branches touch each other, - or collide with a solid object, or whether contact is due to growth within a confined space. In the latter case there is an increasing pressure proportional to the amount of growth. Contact may be reinforced by pressure resulting from the weight of a superim- posed layer of wet cotton; or contact may be the impact result- ing from the whirling of stems through water. Whatever the nature of the. contact or pressure or weight or impact, regeneration of polyps is inhibited in proportion to the degree of pressure, weight, impact, etc. EFFECTS OF LACK OF OXYGEN!® Experiment 22. Sea-water was boiled to remove the oxygen more or less completely from it and the amount of water evapo- rated was replaced by an equal quantity of boiled tap water. An equal number of stems was placed in flasks filled to the brim with this deoxygenated sea-water. The smallest possible space Loeb ’g1 and ’95, Pfliiger’s Archiv., vol. 62. 344 A. “Ff. Goldfarb was left between the water and the cork, then the flask was hermete- ically sealed. The following tables indicate in per cent the vol- ume of water after boiling. ‘These figures will serve in a general way to indicate the amount of oxygen Saree TABLE 20a No. of hydranths observed on the following days No. of stems 2d 3d 4th:«=§th:)«=Sss6th:=Ss 7th «= 8th_~=—s oth_~—Ss roth 100 1 Norm sea-water in Openidishy. cee =e 40 18 8 7 8 Il 13 8 6 2 Norm sea-water in sealed! yessell mee 2. 14 13 12 3 5 6 5 ° 3 to 10; from 97% to77% oO ° ° ° ° T here was absolutely no sign of regeneration in the eight sealed flasks, Nos. 3 to 10 inclusive, from which the oxygen had been removed from the sea-water and its reabsorption prevented. Flask No. 2 which was also hermetically sealed, contained normal sea-water, and regenerated fewer polyps than the open dish No. 1. Regeneration in the former ceased altogether after eight days, while in the latter polyps continued to be produced. Regen- eration in the hermetically sealed flask No. 2 was inhibited, either because the supply of oxygen in the water was entirely appropriated by the previously developed polyps or because of the carbon dioxide produced by these polyps, or by both of these causes acting together. Additional evidence of a very interesting nature was obtained by filling wide mouthed jars about two-thirds full of oxygen—free sea-water as above. [here was, however, a layer of air over one inch deep between the water and the cover, which was sealed air- tight. TABLE 20b No. No.of min- Evap- No. of hydranths observed on the following days utes boiled oratedto 3d 4th sth 6th 7th th roth ith 12th 13th 14th I control control 8 2 3 7 18 7 7 2 3 3 4- 2 4 98% 2 fo) 9 II 7 2 3 2 2 I 2 3 8 823 4 4 9 7 4 2 4 3 3 5 6 4 10 81 I ° ° 3 8 2 3 ° ° ° ° 5 124 75 ° ° 2 c ° I I ° ° ° ° 6 15 73° I ° I 6 5 7 8 2 2 2 ° 7 174 70° ° ° 2 2 2 3 2 2 I ° ° * Probably an error. Factors in Regeneration 345 Polyps were produced in all the jars. ‘The greatest number appeared in the control dish No. 1. The previous experiment showed that no polyps are produced in water from which the oxy- gen had been removed. But in this experiment the water reab- sorbed enough oxygen from the overlying layer of air to supply the needs of the Ale eloping polyps. W rere large quantities of oxygen had been removed from sea-water as in Nos. 4, 5, 6 and 7 a much longer time was required to absorb enough oxygen to permit regeneration to begin; and the available supply of oxygen was more quickly exhausted in these than in Nos. 2 and 3, therefore, regeneration ceased earlier in the former series than in the latter. It follows that where much oxygen has been removed from sea-water, polyps ap pear later and disappear earlier than on stems kept 1 in water containing more oxygen. EFFECTS OF DIRECT SUNLIGHT”? Experiment 23. ‘his experiment unfortunately was not carried to completion. The apparatus was so placed that the sun shone directly upon the stems for about eight hours daily. The heat of the sun was guarded against by reducing radiation from the table and fixtures to a minimum and by surrounding the dishes by large volumes of water to which ice was sometimes added. Temperature records of each of the dishes were made at least three times daily. The stems were grouped as follows: 1 Dish was not guarded against the heat of the sun. 2 Dish was surrounded by 3000 cc. of water. 3. Dish was surrounded by gooo cc. of water. I, 2 and 3 were exposed to the direct rays of the sun. 4 Not exposed to the light, but kept in the shade of the room. 5 Not exposed to the light, but kept in a dark chamber. The temperature in 1 was, of course, several degrees higher than in any of the other dishes, especially about midday, when the temperature was often as high as 28° C. In 2, the tempera- ture was lower, while in 3, 4 and 5, which were practically the same, the temperature was lowest. Loeb ’92, 96; Driesch Zodl. Jahrb. ’g0; Goldfarb ’o6. 340 A. Ff. Goldfarb The greatest regeneration occurred in 1, then came 2 and 3, which regenerated about the same number and less than 1; there was a decided drop in 4 and 5 which regenerated least. There was a notable exception in one of the dishes of series 5 in which a surprisingly large number of polyps were produced. Dzurect sunlight stimulates stems to increased regeneration of polyps, even though the temperature of the water in which they are contained rises as high as 28°C. ‘There is, furthermore, an undoubted posi- tively phototropic bending of some of the new polyp-bearing branches. EFFECTS OF TEMPERATURE” Experiment 24. In the previous experiments bacterial increase was guarded against, particularly in the higher temperatures, and hydranths prospered in a temperature as high as 28° C. Without this precaution such high temperatures would be fatal to the stems. At what higher temperature polyps would be regen- erated was not determined. Up to a certain point the greater the warmth the greater the regeneration. When stems were placed in a refrigerator in which the tempera- ture varied from 10° to 16° C. there was no mistaking the inhibi- tory effects of the cold. A large number of stems never regener- ated at all, and the total number produced was exceedingly small. EFFECTS OF REPEATED REMOVAL OF POLYPS FROM THE SAME LATERAL BRANCHES” Experiment 25. As soon as polyps were distinctly differentiated at the cut ends, they were again amputated. ‘This daily removal of poylps was carried on for a period of 31 days, and daily records made of the number and position of the polyps removed. During this time there were regenerated: 1 At 15 cut ends, including oral and basal, 59 different polyps; 2 At 11 cut ends, including oral and basal, 28 different polyps; 3 At 15 cut ends, including oral and basal, 64 different polyps, making a total of 41 cut ends regenerating 151 new polyps or 368 - per cent in 31 days. *1Peebles ’98, on The Effects of Temperature on Reg. of Hydra, Zodl. Bull. “Hargitt, G. T.,’03, Reg. in Hydromeduse, Archiv f. Ent. ’03. os Factors in Regeneration 347 Many cut ends regenerated but once during this entire period, others as many as ten times. ‘The greater or less regeneration was not confined to any definite region of the stems. ‘The oral ends rarely produced any polyps for reasons already given. “The stems continued to regenerate polyps normal in every regard, until the last days of the experiment whenthey decreased appre- ciably in size. “Though observations ceased at the end of 31 days, regeneration of polyps would most likely have continued further. EFFECTS OF OTHER INJURIES TO STEMS If a stem or branch 1s cut at any level and the cut end 1s exposed to sea-water, a hydranth is normally produced. The question arose whether any severe injury except cutting a stem completely across would result in the formation of a polyp at the point of injury, and whether the regeneration of such adventitious polyps would effect regeneration at the neighboring cut ends. Experiment 26. All theinternodes of several stems were either. bored through with a needle, or severely lacerated or slit with a fine scissors. No regeneration at the injured internodes occurred except in two instances. [he wound appeared to close up imme- diately, only a crack in the perisarc marked the place of injury. In the two instances, above noted, the hydranths grew ‘out at right angles to the stem, in marked contrast to the hydranths on the rest of the stem, all of which pointed orally. The large num- ber of injuries on each stem did not reduce the number of lateral, oral or basal hydranths regenerated. Nor did the bending of large stems permanently into an acute angle, effect regeneration at any of the cut ends. When, however, one or two internodes on each stem were slit and the stems then bent so that the wounds were kept exposed to the sea-water, a large number of hydranths appeared from the bent ends. It made no difference whether the bend formed an acute or right angle. Nor was regeneration at any of the other cut ends, including oral and basal, effected by the formation of these adventitious hydranths. The coenosarc from each of the two injured ends of a slit would grow out directly in line with the axis of the stem, and then fuse into a single branch which would regenerate a hydranth at the 348 A. “f. Goldfarb distal end. Sometimes the ccenosarec from the two ends would fuse close to the wound, or each wounded end may independently regenerate a polyp, or less frequently the one or the other end only, develops a polyp. Whether the injury was at the distal, middle or basal part of the stem did not influence the regeneration. TABLE 21 No. of stems No. polyps reg. at experimented upon Nature of the Operation injured ends in 6 days 6 Stems punctured or slit at every internode I 8 Stems lacerated at every intrnode 2 6 Stems bent permanently ° 4 Stems bent and punctured at the bend ° 12 Stems slit and bent at point of injury II I} a sufficiently large area of cenosarc 1s cut, irrespective of the level of the stem, and the wounds are prevented from closing imme- diately, a relatively large number of hydranths are regenerated. These adventitious polyps appear at the same time as the polyps on the lateral branches, and seemed in no way to effect the regenera- tion of the latter. Experiment 27. After stems had been experimented upon for a long time and could no longer be made to regenerate, they were cut into small pieces. Sometimes the lateral branches were also cut close to the stem. ‘The pieces cut from the distal regions never regenerated, for the very obvious reason that coenosare is never present in the distal parts of stems. But the middle and basal pieces regenerated an incredibly large number of polyps at their cut oral and basal ends. Some of the lateral branches which had been cut close to the main stem also regenerated polyps. Here again injury to the ccenosarec accompanied by exposure to sea- water rejuvenated the pieces in so far as rapid and extensive regen- eration of polyps is concerned. EFFECTS OF DILUTED AND CONCENTRATED SEA-WATER”™ Experiment 28. Sea-water was diluted by the addition of tap-_ water so as to make a graded series, with differences of 5 and some- times 10 per cent, from normal sea-water to 50 per cent dilution. *8These experiments are based on Loeb’s 791. See also Snyder ’05, Archiv.f. Ent. Large numbers of stems and branches were used. te oui Factors in Regeneration 349 In the experiments of 1905 the total number of polyps was daily recorded. ‘These records agree with those of Loeb, that the num- ber regenerated increases with the increased dilution of sea-water. The maximum regeneration is reached in sea-water diluted 15 to 20 per cent; beyond this point, that is, in solutions more diluted, regeneration rapidly decreases; in 40 per cent, few polyps are produced, in 50 per cent, none. In 1906 the experiment was repeated, but the records indicate the number of different hydranths daily regenerated. ‘The results are practically in accord with the data obtained by the other method in 1905. ‘The rate of development in both series was the same in normal sea-water and in solutions diluted as much as 15 per cent but beyond this point the greater the dilution the greater the retardation. TABLE 22 No. of different polyps regenerated in No. of branches Solution 3 days 6 days used Normal 46% 82% 234 5% 56% 79% 15% 62% 84% 25% 28% 73% 35% 0% 34% 3 Stems of Pennaria tiarella behave in quite the same manner as Eudendrium in dilute solutions of sea-water. ‘The results in both hydroids are practically the same. Experiment 29. An effort was made to acclimatize the stems of Eudendrium and Pennaria to greatly diluted sea-water, and thereby to have them regenerate in solutions diluted 50 per cent or more. Sea-water was daily diluted 2} per cent more than the preceding day. Polyps appeared in all dilutions until 45 per cent was reached, beyond which regeneration ceased on Euden- drium stems, while Pennaria ceased at 50 per cent. In the very diluted solutions polyps were distinctly smaller than the normal polyps. ‘The experiment seemed to show thathydranths of Euden- drium and Pennaria could not be made to regenerate in solutions diluted more than 45 and 50 per cent, respectively, during the 25 days of the experiment. 350 A. }- Goldfarb Experiment 30. ‘Yhe stems from the preceding experiments were removed from their solutions to normal sea-water. Four to six days after the transfer, hydranths appeared, first, on stems from the least diluted sea-water, later on stems taken from 20 to 30 per cent dilutions while no regeneration occurred on stems kept in water diluted 40 per cent or more. The number regen- erated increased daily but not to the same degree, so that by the eleventh day after the transfer the stems taken from the greatly diluted water regenerated as much as those taken from the normal or slightly diluted sea-water. TABLE 23 Stems in dilute solutions for 11 days were transferred to normal sea-water Solution prior to Per cent regenerated in transfer 3 days 6 days g days 11 days Norm g* 20 32 40 5% ° 9 21 28 10% ° 22 36 47 15% 16* 22 32 38 20% ° 3 15 30 25% ° 10 14 27 30% < 5 ae 33 35% ° 6 26 40 40% ° ° ° ° * These polyps were present at time of transf r. The effects of concentrated sea-water will be taken up more fully in the following experiments. Experiment 31. A graded series was made by boiling and there- by concentrating sea-water. For example, 200 cc. of sea-water at 18° C. when boiled to a volume which at the original tempera- ture was 180 cc. constituted a go per cent concentrated solution. The water was filtered and aérated by thorough shaking. Every few days the water was replaced by water freshly concentrated to about the same per cent. TABLE 24a Polyps regenerated on the following days: No. of branches Concentration 2d 3d 4th sth 6th 7th 650 Norm 11% 50% 55% 52% 32% 5% 92% to 95% ° 5 15 15 15 4 85% to 89% ° 2 6 12 12 10 777% ° fo) ° I ° ° 65% ° ° ° ° ° ° 62% ° ° ° ° ° ° Factors in Regeneration 351 TABLE 24b Polyps reg. in Concentration 3 days 6 days No. of branches Norm 51% 91% 719 90% 14 51 87% : 7 32 85% 5 37 com 76% > ° ° 70%| etc The more concentrated the sea-water the fewer the polyps pro- duced. Slight increase of concentration results in large decrease in the number regenerated. ‘The maximum concentration beyond which no regeneration occurs is about 77 per cent. Retardation observed in diluted sea-water likewise takes place in concen- trated solutions. It is almost needless to add that the greater the concentration the slower as well as the fewer the polyps produced. Experiment 32. The maximum concentration at which regen- eration of Eudendrium stems will take place was more definitely ascertained by transferring them daily to solutions more concen- trated by about 24 per cent. Regeneration occurred in all con- centrations from norm to 574 per cent. Even in 50 per cent solutions, shoots were observed. ‘Thus by slowly increasing the concentration of sea-water, stems were made to regenerate polyps in a considerably greater concentration than had otherwise taken place, when stems were placed at once into the 574 per cent concentrated solution. Experiment 33. Stemsthat had been kept in concentrated solu- tions were after 18 days transferred to normal sea-water. TABLE 25 Stems kept in concentrated solution 18 days were transferred to normal sea-water. On the fol- lowing days there regenerated: Previous Concentration 2d 3d 4th 5th 6th 7th 8th gth roth Norm 0% 0% 2% 2% 4% 4% 6% 6% 10% 95% 2 2 2 2 4 2 4 II 28 89% ° $ 3 3 7 12 16 26 35 79% ° ° ° ° ° ° ° 73% ° ° ° ° ° ° ° ° 352 A. F. Goldfarb The greater the difference in salinity between the solutions in which stems had been kept and normal sea-water, the greater the regeneration after the transfer. But this is exactly what might have been expected from Experiment 28. Seventy-nine per cent concentration or thereabouts, marks the toxic point beyond which stems do not regenerate when transferred to normal sea-water. The percentage of salts present in sea-water determines whether regeneration shall or shall not take place. An excess, or on the contrary, too little salts present in the solution prevents regen- eration. Whether the effects produced are the result of differences in osmotic pressure or of the specific action of the salts or of both of these factors was not determined. Regeneration took place on the one hand in solutions diluted to 45 per cent and on the other concentrated to 58 per cent. From these two extremes the number regenerated increases to a maximum not in normal sea-water but in 15 to 20 per cent diluted sea-water. CONCLUSIONS AND SUMMARY It is extremely difficult, even approximately, to distinguish the external from the internal factors in regeneration. Both kinds of factors play important roles in the life history of Eudendrium and other hydroids. For convenience and for purposes of study each of the factors in these two series were separately considered, though it should be remembered at all times that this is an arti- ficial though convenient arrangement, and that these factors never act singly and independently of the rest. ‘These influences bring about various reactions, only some of which may be said to be adaptive. The following five factors may be said to result in adaptive changes in Eudendrium.”! 1 Gravity determines the position, at which regeneration shall more frequently take place, and the direction of growth. It does not determine the kind of regeneration, for with rare exceptions only polyps are produced when regeneration occurs at all. On erect stems oral and lateral cut ends regenerate pro- fusely. On inverted stems, regeneration is greatly stimulated, *4T oeb laid emphasis on but two factors, namely, light and gravity. Factors in Regeneration Bes at the basal end only, while polyp formation at the lateral and oral ends is largely inhibited. New stems and branches show a strongly negative geotropism, and grow upward irrespective of the position of the piece. 2 Sunlight. Stems or branches exposed to the direct rays of the sun regenerated a greater number of polyps than those kept in the shade of the room, the temperature in both cases being approximately the same. How much the increase was directly due to the effect of the actinic rays per se, or indirectly to the destruction of bacteria, or to the slightly increased temperature, or to all of these factors was notascertained. Manyof the stems and branches bend toward the sun, 1. e., they are positively heliotropic. 3 Temperature. Other conditions being favorable and equal, regeneration increases with increased temperature to the pe and decreases with the lowering of the temperature. At 10° C. regeneration is largely inhibited, while regeneration increases up to and including 28° C. temperature. One of the more important conditions, just mentioned, is exposure to sunlight for stems placed in water at a moderately high temperature and not exposed to direct sunlight produces far fewer polyps. 4 Any severe injury at any level of the colony, may cause polyps to regenerate, if the wound be exposed to sea-water. ‘The direction of growth of the pedicels and the rate of development of the polyps are subject to the external conditions mentioned and to the internal conditions to follow. 5 Contact, pressure, impact, etc., are inhibiting influences which tend to prevent complete development at those ends that come in contact or are pressed upon by a solid body. Shoots are often produced, but further differentiation is stopped. Con- tact determines, in some degree, particularly on very young branches, the direction of growth, which is away from any solid body, therefore, is negatively stereotropic. ‘lhe amount of inhib- ition is proportionate to the degree of contact, pressure, impact, etc. 6 Large variations in the concentration of sea-water probably never occurs in nature and the reactions of Eudendrium to differ- ently concentrated solutions can hardly be called adaptive. The maximum number of polyps regenerated does not occur in normal 354 : A. fF. Goldfarb sea-water but in solutions diluted with about 20 per cent of tap- water. The amount of salts present in this solution is most favor- able to regeneration. As the quantity of salts is increased by concentration or decreased by further dilution, the number regen- erated decreases until the mimimum is reached on the one hand at 45 per cent dilution and on the other at 58 per cent concentrated sea-water. Stems transferred from concentrated to norm sea- water which is equivalent to placing them in dilute solution, regenerate according to the principle laid down, viz: the more dilute the solution, to a certain point, the more hydranths pro- duced. On the contrary, stems transferred from dilute to norm sea-water, which is practically placing them into more concentrated sea-water, do not regenerate less than stems continuously kept in normal sea-water. Regeneration is not inhibited until the solu- tion contains more salts than that normally present in sea-water, while the stimulating effects of diluted sea-water occurs, when either concentrated, normal or dilute solutions are diluted to what is equivalent to a 20 per cent dilute sea-water. Before summarizing the internal factors, the behavior of the ccenosare under different conditions might perhaps more profit- ably be taken up. The coenosarc is circular in cross section, with no partition as in the case of Tubularia. Granules within the ccenosare stream alternately toward the apical and basal ends, either in the same direction throughout the colony or independently, in each branch, or even in different directions in the main stem. ‘The ccenosarc itself can move en masse within the perisarc. With the basal two- thirds of stems excepted, the cenosarc invariably moves toward the basal end of the piece, 1. e., in all branches, in the apical pieces of stems, and in small pieces from any region of the colony. It may even move entirely out of the piece, through the basal end. Now, as regeneration occurs only where the cenosarc 1s present, it follows that whether regeneration shall or shall not take place at a cut end 1s determined by the migration of the cenosarc. This movement basally crowds or concentrates the cenosarc at the basal end of the piece and if conditions are favorable at the end polyps readily appear there. ‘The migration of the coenosarc may be furthered in various ways, namely: Factors in Regeneration 355 1 Inverting pieces is almost certain to stimulate regeneration at basal ends; or, better still, embed inverted pieces in sand and a remarkable number of basal polyps appear on the free parts. 2 Tying a ligature at the distal end of a stem or, still better, ligature the lateral branches, then ligate the middle of the stem, and a greatly increased number of heteromorphic polyps result. 3. By cutting small pieces from any part of the colony a far greater number of polyps is produced at the basal than at the apical ends. Thus though the cenosarc 1s influenced by such external factors as gravity, ligatures, lack of oxygen, cold, and by internal factors, such as age, size of the piece, etc., the cenosarc normally behaves in certain definite ways which, without the aid of hypothetical “specific stuffs” not only accounts for the absence of polyps at cer- tain cut ends but accounts for their regeneration at other ends. Under a given set of conditions we can foretell with a fair degree of accu- racy the number and region at which regeneration will take place. Furthermore, it is not necessary to have recourse to the stimu- lating effects of necrotic tissues thrown into the circulation to account for the regeneration of polyps. In Tubularia it has been maintained that the breaking down of the partition near the cut end throws into the circulation material which stimulates regen- eration atthatend. Inahydroidresembling Eudendrium, namely, Pennaria, Gast and Godlewski believed that the disintegration of polyps supplied the circulation with material which stimulates regeneration. In Eudendrium there is neither a partition as in Tubularia, nor were the polyps permitted to disintegrate, for they were cut off at the beginning of the experiment: Yet hy dranths were formed within 48 hours-often at every cut end. Even when hydranths were daily removed as soon as formed, other polyps were regenerated. The following internal factors affect regeneration: 1 Age determines not the kind but the rate and number regen- erated. The younger the region the more numerous and the quicker do polyps appear. This statement is subject to special conditions already enumerated, such as ligatures, inversion of stems, migration of the coenosarc, etc. 356 A. ‘f. Goldfarb 2 The influence of the presence or absence of lateral branches and their pedicels. Pieces with the lateral branches cut off close to the main stem regenerate many more polyps than similar stems from which only the polyps have been removed, for the reason that the coenosare tends to withdraw from the pedicels whereas it does not do so from the lateral ends cut close to the stem. 3 This suggests another closely related factor, viz: the influence of size (1.e., the amount of cwnosarc) on the number and position of the regenerated polyps. Pieces less than 1 mm. long may regen- erate stems but never complete polyps. Pieces 1 to 14 mm. long regenerate but one polyp, more frequently however at the basal eae Sometimes one is produced at each end. On larger pieces two or three polyps may appear on an outgrowth at the basalend. Stull larger pieces may bear as many as nine basal polyps at one time, while it is rare for more than one apical poylp to be produced. 4 The influence of the old tissue on the kind of regeneration. Polyps are replaced only by polyps; stems if injured give rise to poly ps. Under certain unfavorable conditions proliferation of cells may take place but no differentiation into polyps occurs, and “stolons, or modified stems, result. If these are cut and removed or grow into a favorable environment polyps, not “stolons,” are regenerated. The rate of regeneration varies with the size and age of the piece. Large stems produce polyps quickest at the distal, slowest at the basal region. Medium size pieces regenerate at all the cut ends at the same time. Polyp formation is greatly retarded on small pieces even if the pieces are taken from the distal region of large stems. ‘The two cut ends at any level of a stem regenerate at the same time. ‘The presence of polyps does not prevent, but may retard, regeneration at the basal end. Unfavorable con- ditions, such as lack of oxygen, low temperature, greatly diluted and even slightly concentrated sea-water, gravity (on lateral ends of reversed stems) all retard development. Nothing availed to affect regeneration in less than two days. Zoological Laboratory Columbia University, New York New York City, March 1, 1907 STUDIES ON REGULATION xl FUNCTIONAL REGULATION IN THE INTESTINE OF CESTOPLANA BY Co Mee ELED With Twenty Text Ficures This Neapolitan form which has served for other experiments (Child ’o5a, ’osb, ’05c) is very favorable for the study of the intes- tinal changes which occur during form-regulation. ‘The intestine in normal animals is almost black in color and since other portions of the body are unpigmented is very distinctly visible in the living animal. Moreover, the regulatory changes are extreme and in some cases relatively rapid; and finally animals and pieces live for months in clear water without food so that it is possible to follow the intestinal changes during a long period. I THE TURBELLARIAN INTESTINE, ITS FUNCTIONS, AND FUNC- TIONAL FACTORS INVOLVED IN ITS DEVELOPMENT AND REG- ULATION This part of the paper aims to establish a general basis for interpretation of the experiments and observations to be described later. It precedes rather than follows the descriptive part because It is important, as well as economical of time and space, to be able to point out the bearing of the various experimental data, under each head instead of postponing interpretation to a general sec- tion where the chief points of the description must be reviewed. The basis of interpretation suggested here is, however, in part the result of these and other similar experiments, not a precon- ceived hypothesis with which the facts are to be brought into accord. As will appear also, it is in line with previous suggestions which I have made concerning the dynamic or functional char- acter of form-regulation (Child ’ 05a, 06a, ’o6b). Tue JourNAt or EXPERIMENTAL ZOOLOGY, VOL. IV, NO. 3. 358 C. M. Child rt The T urbellarian Intestine and its Functions The names by which the various organs of the lower inverte- brates are designated do not necessarily serve to indicate with any degree of exactness their functions. We commonly speak of the alimentary apparatus of such forms as the turbellaria as an intestine, a digestive system, etc., but strictly speaking the func- tions of this apparatus are not identical in all respects with those of the intestine of the vertebrates for example. It is of course a digestive system, but it is more than that. In the first place, the turbellarian intestine undoubtedly serves as a place of storage for undigested nutritive material. Any one who has observed turbellaria feeding can scarcely fail to recognize that this is an important function of the intestine, at least in cer- tain species. Food is often taken until not only the intestine but the whole body is greatly distended. In fact I have often observed the bursting of various species in consequence of rapid intake of food. ‘The opening in such cases is usually small and after out- flow of the excess of material soon closes. Under such conditions the intestinal walls must of course undergo great mechanical extension. Secondly, digestion is, at least in part, intracellular and the intestinal cells undoubtedly accumulate reserve material when food is abundant; in other words, when digestion proceeds more rapidly than material is removed. But besides the functions of digestion and accumulation of reserves the intestine in these forms is the chief means of distri- bution of the nutritive material to various parts of the body, 7. e., it is in greater or less degree a circulatory system, a fact which has been recognized by those authors who have termed it the gastro- vascular system. As a gastro-vascular system it contains fluid laden with nutritive substances. ‘This fluid moves to and fro, enters and leaves the various branches and regions according to the muscular contractions of the body-wall. ‘Thus the intestinal wall is subjected to the varying fluid pressures which, however, are more or less typical for each particular region since the muscular contractions are in general typical. A wide range of conditions Studies on Regulation 359 exists, of course, for each region, but the conditions in the terminal regions, for example, must be in general typically different from those in the middle region. To sum up: the turbellarian intestine as an organ of digestion and a store-house of reserve material is undoubtedly the seat of typical chemical reaction-complexes. As a reservoir for the tem- porary storage of undigested food and as a vascular system con- taining moving fluid it is undoubtedly subjected to a typical com- plex of mechanical conditions. ‘These two groups comprise, I believe, the most important functional conditions for the turbel- larian intestine. The intestine of higher forms, or at least some part of it, serves as a place of temporary storage for undigested food and often, as in certain birds and mammals, undergoes a high degree of special- ization in connection with this function. But in higher forms where a specialized circulatory system is present, the intestine does not function to any great extent as a system for the distri- bution of nutritive material and is not subjected to the mechanical conditions which must exist in such a system, although of course mechanical functional conditions are more or less important fac- tors in the functional complex in all cases. There can-be no doubt, however, that mechanical conditions constitute a much larger element in the functional complex characteristic of the turbellarian intestine than they do in higher forms. If functional factors play any part in development and regulation, we may expect to find the determining factors in the two cases different to a greater or less extent. 2 Functional Factors in Intestinal Development and Regulation It is a well-established fact that the mechanical conditions con- nected with the movements and pressure of fluid within the vessels jare factors of great importance in determining diameter, distri- bution, angle of branching and character of the wall of the blood- vessels. Since this is the case it is natural to expect that similar conditions will play a role of greater or less importance in develop- ment and regulation of the turbellarian intestine. 360 j. M. Child Judging from the form of the turbellarian intestine in relation to the form and structure of other parts it is difficult not to believe that functional and particularly mechanical conditions are impor- tant factors in its development. In the rhabdoccels where no strands of parenchyma or dorso-ventral muscles oppose it, it forms simply a sac, filling the pseudoccel in part or wholly accord- ing to conditions. In the polyclads, on the other hand, the intes- tine might be compared roughly to an elastic sac placed in a space in the axis of the body and then gradually distended so that parts of it are forced into the parenchymal spaces toward the periphery of the body. If the fluid contents of the intestine move and exert pressure in typical directions it seems to me that the effect of these movements must necessarily appear in the direction and size of the intestinal branches. All the facts seem to indicate that the general direction and arrangement of the intestinal branches in the various parts of the body is determined, at least in large part, by the mechanical conditions resulting from movements and pres- sures of the fluid contents. By altering these conditions the arrange- ment of the intestinal branches can be altered, as | showed for Leptoplana (Child ’o4a). In the triclads conditions are similar but the intestine develops in different form because of the position and form of the pharynx. ‘The almost infinite variations in type of the “normal” turbellarian intestine in a given species simply show, in my opinion, how largely its form as regards details is a matter of chance, determined often by the presence, absence, or position of spaces, or dorso-ventral muscular fibers in the paren- chyma, by slight individual differences in movement or consti- tution of other parts, etc. | If the functional conditions connected with the movements and pressures of fluid contents are essential factors in determining the form of the intestine, we may expect to find changes of form occur- ring when these factors change and the facts justify our expecta- tions. Starvation of a planarian results in degeneration and total disappearance of the most distal portions of the intestine in succes- sion: feeding results in the redevelopment of branches, but not necessarily in the same pattern, and increased distension of a normal animal results in the formation of new intestinal branches. Studies on Regulation 361 But it is in connection with experiments on form-regulation that the extreme plasticity of the turbellarian intestine becomes evident. ‘he changes in form and arrangement of the intestinal branches in the experiments of Lillie (or) and Bardeen (’o1, ’o2, ’03) are sufhcient to illustrate this point, although they do not demonstrate its correlation with the functional conditions result- ing from the movements and pressures of fluid contents. In most triclads and polyclads intestinal regeneration is usually much less complete than the regeneration of other parts when the animals are not fed. Moreover, and this seems to me to be a crucial point, it is much less complete in pieces without the ceph- alic ganglia than in pieces containing the ganglia (Child ’o4a). It can scarcely be supposed that there is any essential difference in nutritive conditions between pieces with and those without the ganglia. If anything, more nutritive material should be avail- able for growth in the piece without ganglia since it is much less active than the other. I do not believe, however, that such differ- ences in intestinal regulation can be due primarily to the differ- ences in nutritive conditions. The only reasonable basis for interpretation seems to me to lie in the differences in activity. In the piece without ganglia the movements of the intestinal con- tents are less frequent and less energetic, and consequently the stimulus to intestinal growth in the new tissue is less than in the piece containing the ganglia. Observation of two such pieces and of the movements of intestinal contents in their bodies shows very clearly that the intestinal pressures and tensions are much greater in the piece containing the ganglia than in that without them. All the data thus far available seem to me to indicate that the form and arrangement of parts of the turbellarian intestine is determined very largely by mechanical factors due to the presence and movements within it of fluid contents. ‘This statement is not to be interpreted, however, as signifying that nutritive factors play no part in determining intestinal form. ‘The form must be altered to a certain extent by the presence or absence of reserve material in the cells, by the general metabolic conditions, the relation between intake and output, etc. But I find it difficult to under- stand how such factors as these can possibly determine the general 362 C. M. Child outline of the intestine, and the direction and arrangement of its branches. Lack of nutrition may of course determine the degen- eration of a branch or of branches, but how can the presence of nutrition determine the position and direction of new branches? On the other hand, the conditions above mentioned do account readily for position, outline, and arrangement of parts and experi- mental data indicate that they are the factors chiefly involved. The development of intestinal branches is simply another illus- tration of the fact which I have mentioned elsewhere at various times, viz: that the stimulus to growth is not identical with the presence of nutritive material, but that, on the other hand, nutri- tive material goes where the demand is greatest even at the expense of reduction and disappearance of other parts where the demand is less. This relation between growth and _ nutrition seems also to show why such extensive intestinal reduction occurs in many turbellaria during starvation: the demand for nutritive material is greater in other parts than in the intestine, consequently material passes from it to them. In short, I believe the whole problem of the “self-regulation of metabolism”’ during starvation and indeed at other times 1s essentially a problem of relative func- tional activity in the broadest sense. In Cestoplana the axial intestine extends directly through the median region of the body from end to end. ‘The lateral branches are at right angles to the axial intestine in the pharyngeal region, but toward the anterior end gradually change their direction, and are directed more and more anteriorly: posterior to the pharynx exactly the reverse is the case (Fig. 1). The movements of intestinal contents in this species are briefly as follows: general contraction of the body forces the intestinal contents from both ends toward the pharyngeal region, and the axial intestine and the lateral branches of the middle region of the body become distended. General extension of the body forces the intestinal contents out of the middle region to a large extent and distributes them along the lateral branches even to the extreme terminal regions, if the contraction 1s strong. Under these conditions the intestinal contents move anteriorly in the prepharyngeal and posteriorly in the postpharyngeal region. Studies on Regulation 363 Local contractions and extensions of course cause local changes in the distribution of intestinal contents, but these follow the same rules as the more general movements. Evidently then, the intestinal branches in the middle region are filled and distended by the intestinal contents which accumulate in the middle region during contraction and the branches in the terminal regions by the contents during their flow away from the middle region. In the regions between the middle and end all intermediate conditions exist. “[hose branches which are filled and distended chiefly by the fluid accumulating in the pharyngeal region arise at right angles to the axial intestine since their for- mation is correlated essentially to lateral pressure of the intes- tinal contents, escape in other directions being impossible. But toward the ends of the body the intestinal branches are filled and distended by fluid which is moving anteriorly or posteriorly. If mechanical conditions are factors in determining the form of the intestine, the intestinal branches in these regions may be expected in accordance with the laws of hydrodynamics to be directed more or less obliquely in the direction in which the fluid is moving. The intestine in Cestoplana seems to me to possess exactly the form which might be expected if movements and pressures of fluid contents are the chief factors in producing it. The fact that a gradual change in direction of the branches between the middle and the ends of the body exists is due simply to the gradual change in conditions. In the regions between the middle and terminal regions the branches are filled and distended in part by the fluid moving away from the pharynx, and in part by standing contents escaping laterally from pressure in other directions. ‘The nearer the pharyngeal region, the more exclusive the latter condition of filling and distension, the farther away the more exclusive the former. Hence we may expect to find with increasing distance from the pharynx a gradual change in the direction of the branches from a position at right angles to the axis to one oblique toward the direction of movement of the contents. Similar conditions in general, with of course various specific differences, exist in other polyclads and triclads, and the form of the intestine as a whole and of each of the long branches in many 364 C. M. Child of the broader polyclads corresponds very closely to what may be expected if hydrodynamic factors play an important part in their formation. On the following pages the various regulatory changes in the intestine of Cestoplana under various conditions are’ described and their bearing on the above dynamic hypothesis of intestinal development is discussed. Perhaps it should be added in order to forestall objections that hydrodynamic factors are not considered as the only factors involved in determining intestinal outline and arrangement of parts in the turbellaria. It seems very probable that other factors must also play some part, though the facts seem to me to indicate that hydrodynamic factors are certainly of great importance. II THE NORMAL INTESTINE AND THE TYPICAL COURSE OF INTES- TINAL DEGENERATION IN THE ABSENCE OF FOOD I Descriptive The appearance of the intestine in newly captured animals differs to some extent, apparently according to the previously existing conditions. In Fig. 1 the terminal and middle regions of the intestine in a normal newly captured specimen are shown, somewhat diagrammatically. The intestine in this case is only moderately distended by its contents: in many cases it is so dis- tended that no spaces between the branches are visible and it appears as in Fig. 2. In uninjured animals kept without food a gradual reduction or degeneration of the intestinal branches occurs, though much more slowly than under certain experimental conditions. Intestinal reduction proceeds from the peripheral or terminal region of the intestine toward the middle. The first parts to disappear are the tips of the branches at the anterior and posterior end and as reduction of these branches continues branches nearer the middle region are affected until a condition resembling that shown in Fig. 6 is attained. In this case which represents a normal animal after about four and a half months without food, only short stumps of the lateral branches Studies on Regulation remain in the terminal regions of the body. With approach toward the middle the length of the branches increases until in the pharyngeal region they still retain their full length, though they are less distended than originally. At the beginning of the experiment the intestine of the specimen figured presented the condition indi- eared in Fig. 2. Undoubtedly intestinal reduction could be car- ried further in normal animals, but departure from Naples made it impossible to keep the speci- mens under observation longer. Intestinal reduction in this species consists in an atrophy and disintegration of the more distal portions of the intestine, not merely in a reduc- tion in size or contraction. Various stages can be more or less clearly distinguished, in most cases, though of course each gradually passes into the following. Starting with the normal well-filled intestine as in Fig. 2, or Fig. 1, the first changes consist in de- creasing distension, so that the individual branches become more clearly distinguishable. . Somewhat later the distal portions of these branches disinte- grate and form a longitudinal band of dark granu- lar substance, which appears somewhat like a longi- tudinal canal on each side connecting with the lateral intestinal branches (Fig. 3, also the pharyn- geal region in Fig. 6). Under high magnification, however, these longitudinal bands are clearly seen to be the débris of the disintegrated terminal regions of the branches. The lateral bands make their appearance first in the more terminal regions of the body and progress toward the “middle Tegions as the ends of the branches undergo degeneration. But a part of the products of degeneration 366 C. M. Child appears within the intestine. As degeneration of the branches proceeds a fluid crowded with dark granular masses appears in the intestine and may accumulate and distend the remaining parts of the intestine in pieces of certain sorts to be described in another section. In normal animals, however, this substance never accumulates to any great extent but undergoes resorption almost as rapidly as itis formed and undoubtedly serves as nutri- tive material for other organs which are still functional. As reduction continues the lateral bands become less conspicu- ous, the dark color gradually fading out as they undergo resorp- tion, and the lateral branches undergo further reduction until their tips no longer extend to the region occupied by the lateral bands. At this stage the intestine appears as in Fig. 4 or as in the regions a short distance anterior and posterior to the pharynx in Fig. 6. Somewhat later still the lateral bands disappear entirely | Fics. 2, 3, 4 AND 5 or break up into parts which sooner or later disappear. Various stages in the disappearance of the lateral bands are shown in Fig. 6. And finally, intestinal reduction may proceed so far that only the axial intestine remains (Fig. 5). In some cases, as in this hgure, the intestine still shows slight indications of the positions of the former branches, but often even these disappear and abso- lutely no trace or indication of branches can be discovered (Fig. 19). Often, as in Fig. 5, the lateral branches disappear before the last traces of the lateral bands, which may persist for a time as isolated groups of granules, presumably occupying the paren- chymal spaces originally filled by the ends of the lateral intestinal branches. The next stage is of course complete disappearance of the intes- tine from the regions concerned. ‘This stage is attained only in the terminal regions of the normal body and of pieces under cer- tain conditions. In Fig. 6 the intestine has disappeared almost Studies on Regulation entirely from the preganglionic region, in which it is present in normal well-fed animals (Fig. 1). In all observed cases of intestinal degeneration, except under certain conditions connected with form-regulation, the course of the process of degeneration is essentially the same and_ passes through the stages described above. ee Discussion According to the above account the intestinal degeneration begins at the extreme peripheral regions of the intestine and pro- ceeds “‘centripetally.”. The ends of the branches in the terminal regions are the first parts to disappear, and the last branches to undergo reduction are those immediately about the pharyngeal region. It can scarcely be supposed that the more peripheral branches or the more peripheral regions of each branch are less needed than the more central parts and so disappear first. Uhe peripheral portions of the intestine would seem to be just as essential as other parts for proper nutrition. ‘The head-region is the most active region of the body and yet the anterior end of the intestine dis- appears earlier than any other part of the prepharyngeal intestine. But when the course of reduction is considered from a functional standpoint interpretation becomes easy. In the first place the quantity of intestinal contents undergoes gradual decrease from the beginning to the end of the experiment. In well-fed animals the intestine is greatly distended (cf. Fig. 2) with food at first. This nutritive material is gradually used up, but as degeneration of the intestinal branches occurs a part of the products of degener- ation appearsin the intestine as a fluid crowded with dark granular masses. In normal animals this too undergoes resorption almost or quite as rapidly as it is formed, and gradually decreases in amount as time goes on. ‘hus even long after the food taken from without has disappeared the intestine is not empty, but the amount of intestinal contents is always decreasing. ‘he move- ments of this dark substance in the intestine can be readily ob- served and the following statements regarding their relation to the general muscular contractions are the result of direct observation. 368 C. M. Child Under extreme conditions of intestinal distension with material all parts may be subjected to equal or nearly equal internal pres- sure but when decrease in the amount of intestinal contents occurs, —_——_—_—____/ bead delete haa gy Selatteteteta Fp: i} *, on a a es + . t ig 3 t t £ Fic. 6 as is the case when the animals are kept without food, the energy of the mechanical conditions connected with the contents must decrease more rapidly in the periph- eral than in the middle regions. ‘Thus, for example, if the intestine is only partly filled, the internal pressure on the walls in the extreme anterior and posterior regions is in general much less than in regions nearer the middle. In the first place, the fluid contents are forced into this region only during extreme extension and then appar- ently with much less energy than into regions nearer the middle. The consequence is that those regions in which the functional stimulus falls below a certain minimum gradually undergo atrophy and degeneration, and as the intestinal contents continue to decrease in amount this atrophy and degeneration gradually extend toward the middle region, which is the last to be affected. Size of the lumen of the various parts and friction between the contents and the walls must also play a part in determining movements and internal pressures of the intestinal contents and both of these factors tend to reduce the energy of the functional conditions more rapidly in the peripheral than in the middle regions. The lateral intestinal branches in and about the pharyngeal region persist longer than any others, simply because the functional conditions are less altered there than elsewhere. In the first place, contraction which drives the intestinal contents toward the middle is usu- ally sudden and violent, in consequence of sudden ex- ternal stimuli, while extension is usually much slower and less extreme. Consequently the intestinal contents are driven into the lateral branches of the middle regions with great force long after they have ceased to reach the extreme peripheral regions at all. “The normal movements of the animal, especially Studies on Regulation 369 the very frequent slight contractions of the anterior and posterior end all tend to keep the middle regions of the intestine more dis- tended than the peripheral regions. Very probably the other functions, 7. ¢., the digestive and storage functions, also play a part in determining atrophy. Of course absence of intestinal contents from any part of the intestine means absence of food to be digested and stored up. Hence the cells of this region may atrophy or change their character because of the partial or total absence of the stimulus to the digestive function or because of malnutrition: or again degeneration may occur because the demands upon these cells for nutritive material are so great in relation to the supply, that they are exhausted or forced so far from equilibrium that continued existence is impossible: degeneration from either of these causes would affect the peripheral regions first and proceed toward the middle. But as will appear below, in certain experimental cases it 1s impossible to account for the regulatory intestinal changes on any other basis than that,of mechanical stimuli from the contents. Ill INTESTINAL REGULATION IN CORRELATION WITH FORM- REGULATION OF PIECES The character of form-regulation in general in this species was described in an earlier paper (Child 05a). It will be recalled that regulation after removal of posterior pieces consists almost entirely in redifferentiation of the parts remaining, only a very small amount of new tissue being formed on the cut surface. Posterior regulation is qualitatively, 7.c., functionally, complete at all levels except anterior to, in, and immediately posterior to the cephalic ganglia. Regulation in the anterior direction, on the other hand, consists almost wholly of regeneration, except as regards certain cases of pharynx-formation, and is complete only at levels anterior to, in, and immediately posterior to the ganglia, being slight in amount elsewhere. As might be expected from these differences, intestinal regula- tion is much more extensive in correlation with posterior than with anterior regulation. But the most remarkable cases of intestinal 370 C. M. Child regulation occur in cases where return to the typical form of the species does not occur. In the earlier papers (Child ’o5a, ’o5b, ’o5c) dealing with this species the processes of form-regulation were interpreted as essen- tially cases of functional regulation, 7.e., “functional adaptation.” For example the redifferentiation into a posterior end of the pos- terior part of a prepharyngeal piece and the formation of the pharynx at a certain level of the old tissue was regarded as the result of a functional regulation in response to altered conditions, in consequence of which a portion of the body which had been functionally, as well as morphologically, prepharyngeal now became functionally posterior, 7. e., postpharyngeal, and in conse- quence underwent regulation, 7. ¢., functional adaptation of its structures to the new conditions. As I have pointed out repeatedly in different papers (Child ’o5a, ’o6a, ’06b), redifferentiation of old parts into parts similar to those removed can occur only when these old parts are capable in some degree of becoming the functional representatives or substi- tutes of the parts removed. If functional substitution for the part removed does not occur at all, form-regulation does not occur: if the substitution is confined to regions adjoining the cut surface the part is replaced more or less completely by regeneration, the completeness of replacement depending on the degree of functional substitution. As was shown in the earlier papers on Cestoplana (Child ’o5a, ’o5b, o5c), the phenomena of form-regulation in general can be readily and consistently interpreted on this basis and the differ- ences between anterior and posterior, preganglionic and post- ganglionic regulation, and regulation in the presence and in the absence of the ganglia, differences which on any other basis appear merely as isolated facts without special significance and without relation to each other, are clearly correlated and explicable. For the more complete discussion and interpretation of the experimental data in the light of this hypothesis the reader is referred to the earlier papers (Child ’o5a, ’osb, ’o5c, o6a, ’o6b). Since the phenomena of intestinal regulation are so striking in this species they were omitted from the preceding papers as deserv- Studies on Regulation 371 ing special consideration. As will appear, however, they afford strong support to the hypothesis which has served for interpreta- tion of the other phenomena: indeed a consistent interpretation seems scarcely possible on any other basis than that of functional regulation. I Intestinal Regulation in Correlation with Posterior Form- Regulation a Inthe Prepharyngeal Region The process of form-regulation in prepharyngeal pieces contain- ing the cephalic ganglia consists essentially (Child ’o5a) in the redifferentiation of the posterior part into a new postpharyngeal region and the formation of a new pharynx between this and the new prepharyngeal region. Regeneration is limited to the extreme posterior end of the piece and amounts to little more than the closure of the wound. ‘The length of the postpharyngeal region thus formed, and consequently the position of the new pharynx, depends on the level of the posterior end of the piece. If the piece includes only the most anterior part of the prepharyngeal region (Figs. 11 and 12), the new postpharyngeal region is short and the pharynx appears near the posterior end. With approach of the level of section to the original pharyngeal region the length of the new postpharyngeal region increases and the pharynx is formed farther from the posterior end (Figs. 7 to g). In all cases of regulation of prepharyngeal pieces containing the cephalic ganglia the lateral intestinal branches posterior to the new pharynx undergo complete disintegration within a short time after section, leaving only the axial intestine. ‘This is shown in Figs. 7 and 8 for a long piece, and in Figs. 11 and 12 for a short piece. In the first case the piece originally included that part of the body anterior to the line f in Fig. 1 and the new pharynx appeared at a considerable distance Son the posterior end of the piece. During the first few days following section the dark color of the ntestinal branches in the posterior part of the piece gradually fades. In six to eight days after section (Fig. 7), 7.e., after the S72 C. M. Child development of the new pharynx is well advanced, the intestinal branches posterior to the new pharynx are seen to be degenerating. The course of degeneration differs somewhat from that described above for normal animals. ‘The branches appear broader and further apart as if this part of the intestine had been stretched longitudinally, and in all probability a mechanical elongation of this part does occur in consequence of its function as a posterior end and region of attachment. A few days later the branches disintegrate completely and the débris, appearing as dark masses ~ Sarat rrerrenyerenenetss ewe aor Ns So Fics. 7, 8, 9 AND IO and granules on either side of the slender axial intestine, gradually undergoes resorption until after two weeks or more (Fig. 8) scarcely any traces remain. A slender axial intestine still persists, however. The difference between the postpharyngeal region and the remainder of the body is striking (Fig. 8) for in other regions intestinal reduction has as yet scarcely begun. ‘The sharp limita-_ tion of this peculiar process to the postpharyngeal region of the — piece makes it certain that the disappearance of the lateral intesti- nal branches is correlated in some manner with the “ redifferentia- tion” of this region from a prepharyngeal to a postpharyngeal — region. Studies on Regulation 373 Intestinal reduction in other portions of the body goes on in the same manner as in normal animals, though somewhat more rapidly. But, meanwhile, short and slender new lateral branches develop on the postpharyngeal intestine in many cases. ‘These never attained full Jeclomnne in the specimens observed, but there is no doubt that if the animals had been fed they would have developed and reached normal conditions. Fig. g shows the con- _dition seventy days after section of the piece from which Figs. 7 and 8 were drawn. Intestinal reduction in the pharyngeal and prepharyngeal regions has followed:the typical course but in the postpharyngeal region new branches have developed. In still later stages without food reduction of all parts of the intestine takes place almost equally until only very short lateral branches remain (Fig. 10, 143 days after section). How much longer such pieces may live it is impossible to say, for my observa- tions extended over only 143 days and many pieces were alive and active at the end of this time. In shorter pieces the process is essentially the same. Taking, for example, a piece including that part of the body anterior to the line } in Fig. 1, the new postpharyngeal region is short, and the pharynx appears nearer the posterior end and the amount of regeneration 1s somewhat greater than in a long piece like the pre- ceding. Fig. 11 shows this piece fifteen days after section. All traces of the postpharyngeal lateral intestinal branches have dis- appeared, only a very slender axial intestine remaining, which, however, extends a short distance into the regenerated tip. In the short pieces degeneration may begin in the redifferentiating region within four days after section, but in the long pieces does not usually appear for a week or more. As regards later stages a similar difference exists. In these short pieces intestinal reduction in other regions 1s always much more rapid than in longer pieces. In Fig. 11, for example, a stage fifteen days after section, reduction is far advanced in the pharyngeal and prepharyngeal regions, and in Fig. 12, forty-five days after section, scarcely any traces of lateral branches exist in any part of the intestine. These short pieces usually die from forty to sixty days after section, 7. e., much 374 C. M. Child earlier than the longer pieces. In consequence of the more rapid intestinal reduction and earlier death of these short pieces lateral intestinal branches never develop in the postpharyngeal region. These two pieces represent the two extremes as regards intes- tinal regulation in prepharyngeal pieces. The results in other pieces fall between these two extremes and differ in detail accord- ing to the part of the prepharyngeal region included in the piece. In every case, however, and my observations include some fifty cases, very rapid disintegration of the lateral intestinal branches took place i in the region posterior to the new pharynx and in the larger pieces a new system of lateral branches developed later. Fics. 11 AND 12 In prepharyngeal pieces from which the head-region and the cephalic ganglia have been removed the formation of a new post- pharyngeal region and pharynx takes place in the same manner as when the ganglia are present (Child ’o5c), the only difference be- ing that the new postpharyngeal regionis longer, the new pharynx farther from the posterior end and the process of degeneration somewhat less rapid than in pieces with posterior ends at the same level but containing the ganglia. As was pointed out in my earlier paper (Child ’osc), the only ground which suggests itself for this difference is the functional relation between prepharyngeal and postpharyngeal regions. Removal of the ganglia reduces the functional activity of the prepharyngeal region very greatly, but affects the activity of the postpharyngeal region to a less extent, hence in regulation the reaction to the altered conditions at the posterior end involves more of the posterior region of the piece than in cases where the ganglia are present, since the energy of Pty. Studies on Regulation 375 reaction is greater in proportion to that of the prepharyngeal reac- tion, when the ganglia are absent, than when they are present. In such prepharyngeal pieces without the ganglia the lateral intestinal branches in the region posterior to the new pharynx dis- appear in exactly the same manner as in the pieces already de- scribed, though apparently somewhat more slowly. “The absence of the ganglia, therefore, does not affect intestinal regulation in these pieces, except somewhat as regards rapidity. The products of degeneration of intestinal cells never accumu- late to any great extent in these prepharyngeal pieces. Appar- ently they undergo resorption almost as fast as they are formed, serving, doubtless as nutritive material for the various regulatory processes, and all parts of the intestine become more and more slender and delicate as time goes on. b In the Postpharyngeal Region The character of intestinal regulation after removal of a part of the postpharyngeal region differs according to the relative length of the part removed. If the level of section lies only a short dis- tance posterior to the old pharynx, e. g., at the line g, Fig. 1, the intestinal changes which occur in the region posterior to the old pharynx are essentially identical in character with those déscribed for prepharyngeal pieces and shown in Fig. 7 and 8. ‘This region, originally the anterior end of the postpharyngeal region, rediffer- entiates into a whole postpharyngeal region, and the lateral intes- tinal branches disappear in the same manner as in pieces where the postpharyngeal region is formed from a part of the prepharyngeal region. One important difference exists, however; the degener- ation is always less rapid in these than in prepharyngeal pieces, from three to four weeks being necessary for the disappearance of the branches. Similar changes occur in pieces with posterior ends at levels somewhat posterior to g in Fig. 1, but with increasing length of the old postpharyngeal region in the piece, the degeneration of the lateral branches becomes slower and less complete, until, when half (4, Fig. 1) or more of the old postpharyngeal region remains, the lateral branches do not disappear early as in the 376 C. M. Child pieces described above, but simply undergo reduction as in normal animals. c Discussion In the case of the formation of a new postpharyngeal region from a part of the old prepharyngeal region, or from the most anterior part of the old postpharyngeal region, all parts of the intestine except the longitudinal axial intestine degenerate com- pletely in much less time than that required for reduction in nor- mal animals. These cases present certain peculiar features: here the intes- tinal material is present, but apparently for some reason the lateral branches are unable to persist in the region which undergoes redifferentiation. In later stages in the longer pieces small new intestinal branches usually develop from the axial intestine in the redifferentiated region. When the new postpharyngeal region is formed from the anterior half or more of the old postpharyngeal region, no such intestinal degeneration takes place. How are these peculiar phenomena to be interpreted? ‘The hypothesis that the intestinal material of the redifferentiating region is used up as nutrition for the growth of this region may serve to account for the rapid disappearance of the products of degeneration, but it does not serve to account for the degeneration of the lateral intestinal branches alone, while the axial intestine persists. Moreover, the process cannot be regarded in the light of an adaptation, for it is certainly not economical of material and energy, neither does it fit the animal better in any way for contin- ued existence. On the contrary, it appears to be a useless destruc- tion of structures of great importance, a waste of energy, and in every way a process which must result to the disadvantage of the animal. But when we consider these cases from the functional stand- point they appear in an entirely different light.’ _In the functional redifferentiation of a part of the prepharyngeal region into a whole postpharyngeal region certain changes in the mechanical condi- tions must occur. After such redifferentiation contraction of the body forces the intestinal contents in this region in the anterior Studies on Regulation Bi direction and extension in the posterior direction, whereas the reverse was originally the case. ‘The intestinal contents now tend to enter the more anterior branches of the region during contrac- tion and the more posterior during extension, but the branches were previously subjected to conditions the reverse of these. These altered conditions must bring about a very different distri- bution of the pressures and strains on the various parts of the intes- tine in this region. If the outline, arrangement and direction of the intestinal branches is determined in any marked degree by mechanical factors connected with the presence and movements of fluid contents, it seems impossible to doubt that such an extreme change in these factors must result either in a transformation of the original structures or in their disappearance, for they are the product of conditions the reverse of those now existing. Appar- ently the change is too great to permit transformation and the old structures disappear. Moreover, if these mechanical conditions determine the intes- tinal changes in these cases, the persistence of the axial intestine is to be expected, for the functional conditions in it remain essen- tially as before, the direction of movement of the contents being merely reversed in each particular instance. Only slight quanti- tative changes, if any, are to be expected, therefore, in the axial intestine. As a matter of fact, the only change observed in the axial intestine in these pieces is a change in diameter in different regions. Instead of remaining larger as originally in case the piece was prepharyngeal, the posterior part becomes smaller than the anterior, a change which is doubtless correlated with the new functional conditions. But the fact that the pieces in which the new postpharyngeal region redifferentiates from a short anterior portion of the old postpharyngeal region show the same rapid disappearance of the intestinal branches may perhaps be regarded as an objection to this hypothesis. It may be said that in these cases the mechanical conditions are not altered in the same manner and degree as in the prepharyngeal pieces and that the intestinal degeneration cannot, therefore, be due to such alteration. This objection cannot hold, however, as a moment’s consideration will show. In these cases 378 C. M. Child a short anterior portion of the postpharyngeal region becomes functionally a whole postpharyngeal region and the change in mechanical conditions, although not a reversal as in prepharyngeal pieces, is without doubt great. The fact that when half or more of the original postpharyngeal region remains no degeneration, or practically none, except the usual slow process of reduction common to all specimens without food occurs, points in the same direction. ‘The larger the part of the postpharyngeal region from which the new whole region is formed, the less the change in functional conditions associated with the functional regulation and the less the degeneration. The facts as to rapidity of degeneration also support the func- tional hypothesis. The lateral intestinal branches of the rediffer- entiating region disappear most rapidly in short prepharyngeal pieces, where the new postpharyngeal region is formed from a region not far posterior to the cephalic ganglia. ‘The rapidity of degeneration decreases as the level of the region from which the new postpha ryngeal region is formed approaches the old pharynx. In pieces without the cephalic ganglia the rapidity of degeneration is somewhat less than in pieces with the ganglia. In those pieces in which the new postpharyngeal region redifferentiates from a short anterior portion of the old postpharyngeal region the dis- appearance of the intestinal branches is still less rapid than in the longer prepharyngeal pieces and,-as noted above, in those cases where half or more of the old postpharyngeal region remains, the branches persist and undergo reduction in the usual manner. It is not difficult to understand from the functional standpoint why these differences in rapidity of degeneration should occur. The change in the mechanical conditions in the intestine must be greatest when a region originally just posterior to the cephalic ganglia redifferentiates into a postpharyngeal region and least when the new posterior end is formed from a large part of the old postpharyngeal region. Between these two extremes the change is intermediate in degree. Evidently then the rapidity of degener- ation in these cases is, as might be expected, parallel to the degree of change in the mechanical functional conditions. In the pieces without the ganglia movement is somewhat less energetic and less pq Studies on Regulation 379 frequent, hence the change in conditions in the region undergoing regulation is less extreme than when the ganglia are present, and degeneration is therefore somewhat less rapid than in pieces with ganglia. As will appear below, however, this is true only for headless prepharyngeal pieces of considerable length in which but little of the anterior end posterior to the ganglia has been removed. In short pieces neither a new postpharyngeal region nor a new pharynx is formed and the intestinal changes are very different from those described above. The development of new short and slender intestinal branches in the postpharyngeal region after redifferentiation in the longer pieces is 1n all probability also a response to a functional stimulus. These branches correspond in arrangement and direction to the branches in a normal postpharyngeal region (Fig. g). Their failure to appear in the shorter anterior prepharyngeal pieces is undoubtedly due to the fact that in these pieces the intestinal con- tents are used up more rapidly than in longer pieces, probably in consequence of the extreme activity which is characteristic of the short pieces: perhaps also the terminal region of the intestine con- tains less reserve material than other parts. ‘Uhus the intestine becomes almost completely empty and very thin-walled after about two months in pieces including only the anterior fourth of the pre- pharyngeal region, and the pieces die, while in pieces including the anterior three-fourths of this region this condition is not reached after about five months. ‘Thus in the short pieces there is proba- bly neither sufficient nutritive material available nor sufhcient intestinal contents to furnish a stimulus to the formation of new intestinal branches in the redifferentiated postpharyngeal region. It is of interest also to note that when new intestinal branches appear in the redifferentiated region they never develop to larger size than the intestinal branches of other regions which are under- going reduction. It seems difficult to account for this early ces- sation of development on any other than a functional basis, but according to this hypothesis it is difficult to see why development should proceed farther, for the functional conditions connected with the presence of fluid contents are similar, quantitatively, in this region as elsewhere. 380 C. M. Child In short, when we consider the various features of this peculiar regulation as primarily “functional adaptations” or better as func- tional regulations, the morphological changes and results are readily interpreted. Moreover, I fail to see any other possible basis for interpretation. ‘That additional factors may be involved, which have not been recognized, is extremely probable, but the facts themselves seem to me to indicate that mechanical conditions play a large part in determining the character of the functional regulation, which, in my opinion, 1s the basis of the morphological changes. Undoubtedly the process is, at least in large part, a complex physiological reaction, not a simple mechanical distortion. 2. Intestinal Regulation in Correlation with Anterior Form- Regulation Anterior regulation is complete only at levels anterior to, through, and immediately posterior to the cephalic ganglia. The parts replaced are replaced chiefly by regeneration from the cut surface. In the case of regeneration of the head the ingrowth of the intestine into the new tissue requires no special consideration here, since it is similar to intestinal regeneration in various other species of turbellaria. In certain cases, however, 1n which the anterior end has been removed posterior to the ganglia and no new head is formed, cer- tain features of interest appear and these are considered briefly below. Pieces from which the anterior end has been removed at a level not far posterior to the cephalic ganglia (a, Fig. 1) behave and react more like normal animals than pieces from which more of the anterior end has been removed: they are more active and react to slighter stimuli than the other headless pieces, but do not regen- erate heads, although they produce more new tissue anteriorly than the others (Child ’o5a, ’o05c). In such pieces the only visible changes in the intestine consist in reduction of a type resembling that observed in normal animals. Fig. 13 shows a piece from which the anterior end was removed at a level corresponding to a, Fig. 1. The specimen was originally somewhat smaller than that Studies on Regulation 381 drawn in Fig. 1, so that the difference in size of the piece in Figs. 1 and 13 is not wholly due to reduction. Fig. 13 repre- sents a stage 143 days after section. Comparison with Fig, 6, a normal specimen kept for about the same length of time without food shows that reduction is somewhat more advanced in the head- less piece than in the normal animal. In such pieces, however, the axial intestine, especially in the prepharyngeal region, appears to be more or less distended by the dark colored products of degen- ciaPit pre Sa CANALS Oe ee “ eiehtn ater) 13 14 Fics. 13 AND I4 eration whose movements can readily be followed. ‘This sub- stance accumulates in headless pieces to a greater extent than in pieces with heads, undoubtedly because of the fact that these pieces, being less active than normal animals and pieces with heads, require less nutritive material and so do not use up the prod- ucts of intestinal degeneration as rapidly as do the other pieces. Consequently the products accumulate in the intestine and, since the movements do not force the intestinal contents into the lateral oe C. M. Child branches as frequently nor as strongly as in cases where the head is present, their effect appears chiefly in distension of the axial intestine. In the case of the pieces shown in Fig. 13 the pre- pharyngeal axial intestine is about the same diameter throughout its length, while the postpharyngeal axial intestine decreases in diameter posteriorly, because with the loss of the head the char- acteristic movements of the anterior end disappear to a large extent and conditions are much the same throughout its length, while the postpharyngeal region still exhibits the same regional functional differences as before, though its activity is somewhat decreased. In cases where a somewhat longer portion of the prepharyngeal region is removed a second smaller pharynx appears (Child 05C). The pharynx varies in position according to the level of section. Where most of the prepharyngeal region remains it is usually a considerable distance posterior to the old pharynx, but in cases where most of the prepharyngeal region is removed it may be almost identical in position with the old pharynx and in such cases its formation involves the degeneration of the old pharynx. For the discussion of these cases in relation to functional regulation the reader is referred to my earlier paper (Child ’o5c). The point of importance for the present consideration is that a new phary ngeal region 1s formed in these cases and a new pharynx arises in it. Bak in many cases the old pharynx persists, at least for a time, so that conditions are different from those in other pieces. In Fig. 14 a piece of this kind is shown at a stage 143 days after section. The level of section corresponds to c in Fig. 1. The old pharynx lies some distance anterior to the new and the intestine shows certain features of interest. Anterior to the old pharynx the axial intestine 1s large and much distended with the products of degeneration, but only short stumps of the lateral branches remain. Between the two pharynges, however, lateral branches are present, but both these and the axial intestine are very slender. For a short distance posterior to the second pharynx both axial intestine and lateral branches are large and filled with the dark substance, while in the more posterior regions they show the usual — Studies on Regulation 383 features of reduction. “The condition of the intestine and the visible movements of the intestinal contents in these and similar pieces in the later stages indicates that in the course of reduction in size the intestine sooner or later becomes occluded in the region of the pharynx or pharynges. During reduction in size the old pharynx is not reduced proportionally and in pieces which have been without food for several months it 1s often so large in propor- tion to other parts as to cause a bulging of the body-wall dorsally and ventrally in its region. It 1s probable that in such cases the pressure upon the intestine in the pharyngeal region is sufhcient to prevent to a large extent the passage of intestinal contents through it. Similarly the development of a new pharynx, posterior to the old, as in Fig. 14, may likewise sooner or later occlude the slender axial intestine in this region. ‘This being the case, the intestinal contents in the segues ngeal and poeta ngeal regions do not enter the “interphary eel region to any great extent, if at all. Consequently the course of intestinal regulation in this region is largely independent of that in other parts of the body. In the case shown in Fig. 14 this region contains but little fluid and both axial intestine and branches are slender, but since removal of the anterior end does not modify the muscular activities in this region except quantitatively to some extent, the branches have not entirely disappeared. In the prepharyngeal region (Fig. 14), on the other hand, con- ditions are widely different. Here the axial intestine is greatly distended with a large amount of the dark substance. If Fig. 14 be compared with that portion of Fig. 1 posterior to the level c, which represents approximately the proportions of the piece at the time of section, it will be observed that the prepharyngeal region of the piece has decreased in size much more than the post- pharyngeal region, doubtless, as was suggested in an earlier paper (Child ’o5c), because the energy of functional conditions in this region underwent a greater decrease with the loss of the head than in the postpharyngeal region, and so the former region has served IN part as nutritive material for the latter. But the effect of this reduction in size on the intestine has been to hasten degeneration of the lateral branches in this region, since the movements of intes- 384 C. M. Child tinal contents have decreased in frequency and strength with the similar decrease in muscular activity. Moreover, the reduction in length of this region has resulted in confining the fluid contents which remain within a smaller space and so in filling this portion of the intestine more completely, since the products of degeneration form more rapidly than they undergo resorption. This _pre- pharyngeal region of the body after removal of the head shows little differentiation of function, 7. e., the anterior end retains only in slight degree the characteristic motor reactions, hence the func- tional conditions are very similar throughout as regards the intes- tine, so that intestinal reduction shows no marked regional differ- ences. In the region posterior to the second pharynx, however, func- tional conditions remain much the same as in the normal animal (Child ’os5c), for the removal of the head affects the activities of the posterior end but little. Consequently contraction forces the intestinal contents anteriorly until they reach the region of the second pharynx, which they cannot pass, and so are forced into the lateral branches of this region and distend these. “The second pharynx appears rather late and before its development the intes- tinal banches just posterior to it often undergo more or less re- duction and after it appears enlarge again, very evidently in response to the altered functional conditions. ‘The movements of the dark substance can be observed very clearly in this part of the intestine and the distension, accompanying contraction of the body, of the lateral branches just posterior to the second pharynx is very evident. The fact that the products of degeneration accumulate in the intestine to a much greater extent 1n headless pieces than in normal animals and pieces with heads is a point of considerable interest. This accumulation cannot be simply the consequence of greater degeneration in these pieces, since in many cases it occurs in stages where degeneration is less advanced than in pieces with heads where no such accumulation exists. In normal animals and pieces with heads these products undoubtedly undergo more rapid resorp- tion than in other pieces and are used to a greater or less extent as nutritive material for other parts, as has already been noted. Studies on Regulation 385 Their accumulation in headless pieces must be due to the fact that these pieces use the nutritive material less rapidly than those where the head is present. ‘This is to be expected from the differences in activity between headless pieces and others. Moreover, it will be shown in the following section that these products accumu- late more rapidly and to a greater extent in the intestine as the activity of the piece decreases. The absence of correlation between intestinal degeneration and the accumulation of the prod- ucts of degeneration within the intestine indicates very clearly that the degeneration or persistence of the intestinal branches does not depend primarily on nutritive conditions. If movements are slight, intestinal degeneration may proceed more rapidly in pieces where a considerable quantity of the detritus, which undoubtedly possesses nutritive value, is present than in cases where the intes- tine is almost empty. Thus, for example, in the case just dis- cussed (Fig. 14) the intestinal branches in the prepharyngeal region have undergone much more complete degeneration in 143 days than in a normal animal (Fig. 2), although the products of degener- ation have accumulated in the headless piece to a much greater extent than in the other. In this case then, as in those discussed above, the visible regula- tory changes in the intestine are very evidently primarily func- tional regulations and are much more closely associated with the mechanical than with the nutritive conditions, 7.e., they are func- tional regulations in response to mechanical stimull. When a larger portion of the prepharyngeal region 1s removed, the second pharynx appears nearer the old pharynx, until in cases where the level of section is not far anterior to the old pharynx, this may persist, or it may degenerate and a small pharynx appear in approximately the same position (Child ’o5c). These cases present no new features of special importance as regards intestinal regulation. In some pieces the occlusion of the intestine by the pharynx or pharynges appears to be less complete than in others, and in such cases the peculiar conditions shown in Fig. 14 are less marked. When the level of section lies immediately posterior to the old pharynx, a new pharynx is often formed at the anterior end of the 386 C. M. Child piece (Child ’o5c): in such cases the intestinal changes are essen- tially similar to, though more rapid than those in the postpharyn- geal region of normal animals. Headless pieces entirely without a pharynx are discussed in the following section. IV) INTESTINAL REGULATION IN PIECES WITHOUT A PHARYNX As was pointed out in an earlier paper (Child ’o5c), the isolated postpharyngeal region possesses the power of functional regulation only in slight degree. When the plane of section is immediately posterior to or near the old pharynx a new pharynx is often formed at the anterior end of the piece, but there is no visible rediffer- entiation of a part of the piece into a prepharyngeal region. In many such pieces, however, and in all postpharyngeal pieces in which the level of section is any considerable distance posterior to the old pharynx, a new pharynx does not appear, 1. ¢., these pieces do not possess sufficient power of functional regulation to give rise to any of the other regions of the body. ‘The same is true of pieces below a certain length from the prepharyngeal region pos- terior to the ganglia. But these pieces, although they remain wholly without a pharynx and show practically no regeneration beyond wound-closure and no regulatory formation of other regions by redifferentiation, do present certain remarkable fea- tures as regards intestinal regulation. Such pieces show few of the typical reactions (Child ’o§5a, ’05c): they do not usually attach themselves to the substratum, but are merely propelled through the water by their cilia: they rarely extend to full length ale in course of time become greatly short- ened and al and show almost no muscular activity beyond slight contractions and extensions and peristaltic waves which pass from one end of the body to the other. Many such pieces, however, were kept under observation during 143 days and the experiments were concluded only because of my departure from Naples. ‘Two such pieces are selected for description: all others observed are essentially similar. ‘The first of these was a short prepharyn- geal piece, including approximately the region between the levels Studies on Regulation 387 d and ein Fig. 1. Fig. 15 shows the piece twenty-six days after section. It is much reduced in size and degeneration of the intes- tinal branches has been very rapid. At this stage the axial intes- tine was distended by a large quantity of the products of degener- ation. ‘his rapid degeneration and the accumulation of the prod- ucts of degeneration certainly cannot be interpreted as the result of lack of food. ‘The intestine was well filled at the time of section and by no means all of its contents were lost through the wound; moreover, the motor activity of the piece is very slight and its need for food is therefore less than that of more active pieces; it has not formed extensive new parts either by redifferentiation or regener- ation. If nutrition is the primary factor in determining the degen- eration or persistence of the intestinal branches, we should cer- tainly expect that they would degenerate very slowly in such pieces. Yet they degenerate more rapidly than in any other case except 146) 17. Fics. 15, 16 AND 17 2 those in which a part of the prepharyngeal region redifferentiates into a postpharyngeal region. About forty days after section no trace of the intestinal branches remained. If the experiment were carried no further, such pieces might be regarded as cases of “‘reversal of development,” similar to those described by various authors as occurring during star- vation. But after sixty-five days the piece presented the appear- ance shown in Fig. 16. A complete new set of very slender and delicate intestinal branches had developed and these persisted as long as the piece was kept under observation, but underwent gradual reduction as the size of the piece decreased in size (Fig. 17,130 days). The history of postpharyngeal pieces without a pharynx is essen- tially similar. Figs. 18 to 20 give three stages in the history of a piece corresponding to that portion of the body posterior to the 388 C. M. Child level g in Fig. 1. Fig. 18 shows the condition of the piece forty- Event iio 19, sixty-five days after section. At the latter stage all traces of intestinal branches have disappeared and the wall of the axial intestine, which is greatly distended with the products of degeneration, 1s as smooth as that of a rhabdoccel intestine. But forty days later, 7. e., 105 days after section (Fig. 2c), new slender branches had developed along the whole length of che axial intestine, and these underwent gradual reduction, but were still visible when the experiment was concluded at 143 days. The development of the new intestinal branches in these pieces usually requires some twenty days or more. ‘The various stages were examined with great care, in many cases under pressure, and there i ah 1 ne rt, \ 4 ip ; Nee dani 4 & 6 ok Lo Av J Yee Goes fh Salk ‘4 os yr fade i ’ pat “4 } aes tA APs ib yg b soho a | 4 Fics. 18, 19 AND 20 is no possibility of error. Their development follows in every case the great distension of the axial intestine with the products of degeneration. Similar intestinal changes were observed in every piece incapable of forming a new pharynx, provided it did not die too early. In general the rapidity with which the changes occur increases with TES in the length of the piece. This difference in rapidity is well shown in the two pieces selected for descrip- tion. In the first, which is considerably shorter than the other, all traces of the intestinal branches were lost within forty days and the new branches were present after sixty-five days, while in the second, 7. e., the longer piece, the degeneration of the branches required sixty-five days and the development of the new branches, forty days more, in all 105 days. Studies on Regulation 389 In these pieces then, although there is no other visible regula- tion except wound-closure, the original intestinal branches undergo complete degeneration and a new set of shorter and more slender branches develops in their place. Moreover, the development of new intestinal branches occurs only after two or three months, during which time the pieces have undergone considerable decrease in size. Evidently it is not correlated with other processes of form- regulation in these pieces. These remarkable phenomena seem to me to constitute a prac- tical demonstration of the hypothesis which has served as the basis for interpretation of the other phenomena of intestinal regulation in this species. As a matter of fact they play an important part in the development of the hypothesis in my mind. A brief dis- cussion will serve to show clearly their correlation with mechanical conditions. In the first place muscular activity is relatively very slight in these pieces, consequently the movements of the intestinal contents are also relatively slight. Under these conditions the intestinal contents accumulate, as can readily be observed, in the axial intestine and enter the branches but little, except when more powerful contractions are induced by artificial stimulation. ‘The old intestinal branches, not being adapted to these conditions, undergo very rapid degeneration and only the axial intestine remains. ‘he persistence of this part of the intestine is to be expected, since all muscular contractions cause movements of its contents, and since these are accumulating as time goes on. These pieces require little nutrition in consequence of their rel- atively slight activity, hence the products of degeneration do not undergo resorption as rapidly as they are formed, but accumulate in the intestine to such an extent that they distend it greatly, and finally bring about the formation of a new set of intestinal branches, which are adapted to the new conditions, and which undergo gradual reduction as these conditions change in following stages. Only in pieces where the axial intestine becomes distended with the products of degeneration do these new branches appear. In general, as the length of the piece decreases, the rapidity of degen- eration of the old branches increases and the distension of the 390 C. M. Child axial intestine and the development of the new branches occur earlier. This difference in the rapidity of change is also exactly what might be expected according to our hypothesis, for the char- acteristic muscular activity and consequently the movements of the intestinal contents into the branches decrease as the length of the piece decreases, hence the shorter the piece, the greater the change in mechanical conditions affecting the intestinal branches. If the mechanical conditions are the determining factors, it follows that the rapidity of degeneration must increase with decreasing length of the piece. But increased rapidity of degeneration results in more rapid accumulation of the products of degeneration in the axial intestine and so in earlier development of the new branches. This interpretation seems to me the only one possible. “These cases show very clearly that the factors which determine the degeneration or the development of a structure are not necessarily associated primarily with nutrition or its absence. Development without energy is of course impossible and this energy must come from nutritive material of some sort. But the mere presence of the material does not necessarily determine that a given structure shall develop. ‘That, as I have endeavored to show in most of the papers of this series and in others as well, is determined by func- tional conditions in the widest sense. VY THE RAPIDITY OF GENERAL INTESTINAL REDUCTION UNDER DIFFERENT CONDITIONS Intestinal reduction in the whole body or piece proceeds with very different rapidity in different cases. The rapidity of reduc- tion in certain special cases has already been discussed in the pre- ceding sections, but a general comparison of the various cases presents certain features of interest since it shows very clearly that nutritive conditions are, at least in certain cases, not the only, nor even the most important, factors in determining the rate of intestinal reduction. In the first place intestinal reduction proceeds more slowly in the normal animal without food than in headless pieces of any size. This is evident from a comparison of the figures. Fig. 6 repre- Studies on Regulation 391 sents the condition of a normal animal after about four and one- half months without food; Fig. 13 shows a piece including almost the whole body except the head after the same length of time without food. Shorter headless pieces differ still more widely from the normal animal; Fig. 15 shows a short, headless, pre- pharyngeal piece after twenty-six days of starvation and Fig. 18 a headless postpharyngeal piece without pharynx after sixty-five days. In these pieces total, or almost total, disappearance of the intestinal branches has occurred in a period of time from less than one-sixth to about one-half that necessary for intestinal reduction in the normal animal to the condition shown in Fig. 6. Such differences as these cannot be due to differences in nutri- tive conditions. [he normal animal is more active and must use a greater amount of nutritive material in proportion to its size in a given time than a headless piece such as that shown in Fig. 13, and its activity and nutritive requirements must be many times greater than those of the headless pieces shown in Figs. 15 to 17 and 18 to 20, in which movement 1s slight, yet in all of these pieces intestinal degeneration is more rapid than in the normal animal. Moreover, in headless pieces the rapidity of intestinal reduction increases with decrease in the length of the piece. ‘The long piece in Fig. 13 (all that part of the body posterior to ain Fig. 1) reaches the condition shown in 143 days, the piece shown in Figs. 18 to 20 (that part of the body posterior to g in Fig. 1) loses all traces of intestinal branches in sixty-five days, and the piece shown in Figs. 15 to 17 (that part of the body between d and ¢ in Fig. 1) loses all traces of intestinal branches in less than forty days (Fig. 15, twenty-six days). In these pieces, and in all similar pieces observed, the rapidity of degeneration of intestinal branches is in general inversely proportional to the length of the piece. At the time of section the amount of nutritive material in these various pieces must be about the same in proportion to their size. Of course some differences exist in this respect and there is more loss from the wound in some cases than in others, but contraction is usually so rapid that loss from the wound 1s slight. “The two pieces whose history is given in Figs. 15 to 17 and Figs. 18 to 20 were taken from the same worm; nutritive conditions must there- 392 C. M. Child fore have been very similar in both at the time of- section. But the muscular activity of the headless pieces decreases in general with decrease in length. ‘The longer pieces must therefore use up nutritive supplies more rapidly than the shorter pieces and if degeneration were due to lack of nutrition it must occur earlier and proceed more rapidly in the longer than in the shorter pieces. But exactly the reverse is the case. Moreover, the formation of new intestinal branches several months after section and after the old branches have undergone complete degeneration shows very clearly that sufhcient nutritive material is present to allow the development and maintenance of the intestinal branches, when the proper stimulus is present. It seems impossible, therefore, to escape the conclusion that nutritive factors are not the most important in determining the rapidity of intestinal degeneration. But when we consider the dynamic conditions resulting from the presence and movements of the intestinal contents, it at once becomes evident that the rapidity of degeneration is in general proportional to the change in these conditions. In the normal animal these conditions remain most nearly normal and after the food taken from without has disappeared from the intestine it still contains a certain quantity of fluid, which moves about in the char- acteristic manner, though its effect must be quantitatively less than when the intestine 1s well filled. In headless pieces the movements differ more or less from those of the normal animal and are always less energetic and less frequent, hence the functional stimulus from the contents must be less than in normal animals and intes- tinal degeneration must occur more rapidly in such pieces than in normal animals if it is correlated with decrease or absence of these stimuli. Moreover, motor activity of all kinds decreases with decreasing length of the headless pieces and the mechanical stimuli arising from the intestinal contents must decrease similarly, espe- cially in the lateral branches, since the less powerful the muscu- lar contractions, the less frequently do the intestinal contents enter the branches. Consequently degeneration of the intestinal branches must occur with increasing rapidity as the length of the piece decreases, if it is connected with these conditions. As shown above, the facts correspond exactly with the require- Studies on Regulation 393 ments of the hypothesis and [| fail to see how any other interpre- tation of them is possible. ‘They all indicate that the rapidity of degeneration of the intestinal branches is dependent, at least in large measure, on the degree of change in the mechanical func- tional conditions connected with the presence and movements of the intestinal contents, irrespective of their nutritive value. But when we compare normal animals with pieces which pos- sess heads the case is somewhat different. ‘The very rapid intes- tinal degeneration in the redifferentiating regions of such pieces has already been discussed in Section III c, and does not concern, us here, but the rapidity of intestinal degeneration in other parts of the body differs from that in normal animals and also differs according to the length of the pieces. In such pieces the activity remains the same as in normal animals, or in short pieces includ- ing little besides the head-region, is apparently even greater than normal. Consequently these pieces must require as much nutri- tive material in proportion to their size as do normal animals, or probably even more in the case of short pieces. Moreover, these pieces undergo qualitatively complete form-regulation, producing a new postpharyngeal and pharyngeal region with a new pharynx. ‘These changes must also require nutritive material. In such pieces the intestinal contents decrease rapidly in amount—thé shorter the piece, the more rapid the decrease—and those portions of the intestine remaining never contain any considerable amount of the products of degeneration as do those of the shorter headless pieces. These products appear to undergo resorption almost as rapidly as they are formed. Consequently the quantity of intestinal con- tents becomes very small and the axial intestine and all other parts become very slender (Compare for example Figs. g and 10 with Figs. 15 and Figs. 18 and 19). As noted above this difference indicates that the products of degeneration serve as nutritive mate- rial. Since this does not accumulate to any extent in the pieces with heads the intestine becomes almost empty and, notwithstand- ing the normal movements of these pieces, the mechanical stimu- lation of the intestinal walls must be very slight and must decrease centrifugally. Consequently the branches disappear and the rapidity of degeneration is determined, at least in part, by the 394 C. M. Child rapidity with which the intestine is emptied of iis contents in con- sequence of the demand for nutritive material. ‘Therefore intes- tinal degeneration increases in rapidity with decreasing length of the pieces, as is the case in the headless pieces. According to this interpretation the increasing rapidity of degen- eration with decreasing length is due in headless pieces primarily to decreasing movement of the intestinal contents in consequence of decreasing muscular activity, while in pieces with heads it is due primarily to decreasing quantity of intestinal contents in conse- quence of the great demand for nutritive material. In the headless pieces nutritive material arising from the degeneration of the intes- tinal branches accumulates in the remaining portions of the intes- tine, but the branches disappear in spite of its presence. - In the pieces with heads, on the other hand, this material is used up almost as rapidly as it is formed and the branches disappear because the intestine is nearly empty. In the first case the degen- eration is apparently due largely to lack of movement of the intes- tinal contents, in the other to lack of intestinal contents to be moved. Thus the data concerning the rapidity of intestinal degeneration serve still further to support and confirm the conclusion that intes- tinal regulation in this species is in large part a functional regula- tion in response to mechanical stimuli. VI CONCLUSION AND SUMMARY The phenomena of intestinal regulation certainly afford strong support to a dynamic or functional hypothesis of regulation and in this respect are in accord with various other phenomena in this and other species, which I have described and discussed in previous papers. The intestine retains its typical form, or returns to it, only when dynamic conditions are, or become similar to those which give rise to the typical form. Extensive intestinal regulation may Occur in the absence of other form-regulation, or intestinal regulation may fail to occur, while other parts undergo more or less complete regulation. ‘The results in each case are correlated with the dynamic conditions in the intestine, particularly the mechani- Studies on Regulation 395 cal conditions, and can be interpreted only on the basis of this correlation. [here can be little doubt that intestinal regulation in various other species of turbellaria will prove to be similarly dependent on mechanical conditions. The history of the pieces without pharynges shows how little significance there is in description or discussion of “‘reversal of development” without consideration of the dynamic factors in- volved. When these dynamic factors act in reverse sequence and direction from that typical of normal development, then, and not otherwise, does reversal of development occur. ‘There is no law, such as certain authors seem to postulate, that causes an organism to return more or less completely to an earlier stage of develop- ment if deprived of food, or under other changed conditions. The so-called “return” usually consists simply of the loss of previ- ous differentiation, but this does not necessarily constitute a rever- sal, for the method of loss may be very different from the method of acquirement, as in the present case. Moreover, the loss of the original structure or differentiation may be merely the first step in the development of something new in response to altered con- ditions, as is the case in the pieces without pharynges. These pieces are in no sense returning to an earlier stage of development or “embryonic condition,” because they lose the old intestinal branches, but are merely undergoing a process of functional adap- tation or regulation. The gradual simplification in intestinal structure, which occurs in various planarians in the course of starvation and reduction in size, is undoubtedly essentially a func- tional regulation just as truly as is the appearance of new branches under other conditions. Objection to my interpretation of the facts may perhaps be made on the ground that the recent experiments of Babak (’06) with amphibia indicate that chemical factors are much more important than mechanical in determining intestinal regulation. It can scarcely be doubted, however, that the amphibian intestine differs greatly from the turbellarian in function. As I have pointed out in Section I, the turbellarian intestine is much more than a digestive organ, being both a storage-reservoir for excess of undigested food- material and to a considerable extent a circulatory system. It 396 Ge Ms Giidd would be remarkable if mechanical factors were not much more important functionally in the turbellaria than in the amphibia. In order to interpret regulatory phenomena it is of the utmost importance to consider all the functions of an organ or structure and not merely one, or the most conspicuous. Only in this way shall we attain complete interpretation. It must be borne in mind that the name assigned to a part does not always indicate fully or exactly its functions, nor is the function commonly assigned to it necessarily its only function: in most cases it 1s merely a small part of the aed function. In the present case the changes in mechanical conditions are to a certain extent visible and accessible to experimental methods and, as 1 have endeavored to show, the processes of regulation in the intestine are evidently closely correlated with them; indeed it is impossible to account in any other way for certain of the changes, such as the rapid degeneration in a postpharyngeal region formed by redifferentiation and the development of new intestinal branches in pieces without pharynges after months of starvation. More- over, while other factors, such as the character of the food and the digestive activity, doubtless affect the structure of the cells and very probably their number, it is difficult to understand how fac- tors of this kind alone can determine the form, arrangement and direction of intestinal branches. These elements of the intestinal form must, it seems to me, be determined mechanically, at least in large part, and it is with these that the present paper is primarily concerned. The most important results are briefly stated in the following summary: I In normal animals kept for several months without food extensive intestinal degeneration occurs, beginning in the peripheral regions and proceeding toward the pharynx. ‘This degeneration involves chiefly the lateral branches and affects the axial intestine only in the terminal regions. 2 In pieces undergoing regulation without food in which a postpharyngeal region is formed by redifferentiation from a part of the old prephary ngeal or the anterior part of the old postpharyn- geal region, the old lateral branches of the intestine undergo rapid Studies on Regulation 397 and complete degeneration in the redifferentiating region and are replaced in the longer pieces by new branches, corresponding i in arrangement with Had of a normal postpharyngeal region. 3 Th headless pieces which have not sufficient regulatory capac- ity to give rise to a new pharynx (short prepharyngeal pieces and most postpharyngeal pieces) the old intestinal branches undergo rapid and complete degeneration, but after two months or more a new set of short and slender intestinal branches arise, which per- sists, but undergoes gradual reduction as time goes on. 4 In all other pieces undergoing regulation without food intes- tinal reduction occurs and usually proceeds from the peripheral towards the middle regions, though special modifications occur with special conditions. 5 The intestine of polyclad and triclad turbellaria 1s not merely a digestive organ, but functions also as a reservoir for the tempo- rary accumulation and storage of undigested food-material and also, to a considerable extent, as a circulatory system. Its con- tents are largely fluid and undergo movement in consequence of the muscular contractions of the body-wall. The presence and movements of these contents must produce characteristic mechan- ical effects upon the intestinal wall. 6 ‘The facts of intestinal regulation indicate that these mechan- ical conditions play an important role in determining the outline of the intestine and the direction and arrangement of the branches. Total disappearance of the old branches occurs when the mechan- ical conditions are widely altered, even though nutritive material be present in excess. [he rapidity of degeneration depends on the degree of change in the mechanical conditions. ‘The develop- ment of new branches after degeneration of the old is determined primarily, not by the presence of nutrition, but by mechanical con- ditions, though of course nutritive material is necessary for such development. Undoubtedly certain features of intestinal regulation are deter- mined by other functional factors, but the general outline and the arrangement and direction of branches are very evidently closely correlated with mechanical factors. 308 C. M. Child BIBLIOGRAPHY BaBAk, E., ‘06—Experimentelle Untersuchungen tiber die Variabilitat der Ver- dauungsrohre. Arch. f. Entw-mech., Bd. xxi, H. 4, 1906. BARDEEN, C. R., ’o1—On the Physiology of the Planaria maculata with Especial Reference to the Phenomena of Regeneration. Am. Jour. Physiol., vol. v, Igol. ‘o2—Embryonic and Regenerative Development in Planarians. Biol. Bull., vol. ii, no. 6, 1g02. °03—F actors in Heteromorphosis in Planarians. Arch. f. Entw-mech. Bd. xvi, H. 1, 1903. Cuitp, C. M., ’o4a—Studies on Regulation. V. The Relation Between the Central Nervous System and Regeneration in Leptoplana: Pos- terior Regeneration. Journ. Exp. Zodl., vol. i, no. 3, 1904. °o4b—Studies on Regulation. VI. The Relation Between the Cen- tral Nervous System and Regeneration in Leptoplana: Anterior Lateral Regeneration. Journ. Exp. Zodl., vol. i, no. 4, 1904. ‘o5a—Studies on Regulation. VIII. Functional Regulation and Re- generation in Cestoplana. Arch. f. Entw-mech. Bd. xix, H. 3, 1905. ‘o5b—Studies on Regulation. IX. The Positions and Proportions of Parts During Regulation in Cestoplana in the Presence of the Cephalic Ganglia. Arch. f. Entw-mech., Bd. xx, H. 1, 1905. ‘o5c—Studies on Regulation. X. The Positions and Proportions of Parts During Regulation in Cestoplana in the Absence of the Cephalic Ganglia. Arch. f. Entw-mech., Bd. xx, H. 2, 1905. ’o6a—Contributions Toward a Theory of Regulation. I. The Signifi- cance of the Different Methods of Regulation in Turbellaria. Arch. f. Entw-mech., xx, H. 3, 1906. ‘°o6b—The Relation Between Functional Regulation and Form-Regula- tion. Journ. Exp. Zodl., vol. ii, no. 4, 1906. Liu, F. R., °o1—Notes on Regeneration and Regulation in Planarians. II. Am. Journ. Physiol., voi. vi, 1got. THE BEHAVIOR OF LOXOPHYLLUM AND ITS RELA- TION TO REGENERATION S. J. HOLMES Wirtn Seven Ficures GENERAL CHARACTERISTICS OF THE SPECIES The general form of Loxophyllum meleagris, the species studied, is flattened and leaf-like, and tapering toward the anterior end which is turned toward the dorsal margin. The anterior third or fourth of the body is flatter and less granular than the hinder portion and the margins of the body are thinned out, especially along the oral side. “The middle and posterior regions are more convex and may be considerably distended filer gorged with food. he body is ciliated on the right side on which the ani- mal usually glides. ‘The cilia are arranged in rows which extend in a longitudinal direction except near the anterior end of the body where they curve toward the dorsal side. The whole oral margin is also furnished with cilia, but none could be detected on the left side of the body. The body wall is traversed with myonemes, both on the right and the left side, which extend longitudinally for the most part, but curve dorsally, like the rows of cilia, near the anterior end. ‘hey are more conspicuous and apparently thicker near the anterior end of the body, and they are especially well developed near the oral side. ‘Trichocysts are abundant along the entire oral margin and around the anterior end of the body, forming a uniform series closely set at right angles to the surface. On the dorsal side the trichocysts are mainly confined to the small promi- nences, a dozenor more in number, which give that side its crenu- lated contour. Numerous trichocysts occur also on the right or ciliated side. 400 S. Ff. Holmes The contractile vacuole is nearly spherical in form and is situ- ated near the dorsal side of the body a little in front of the posterior end. ‘There is a fine canal extending ‘from it anteriorly along almost the entire dorsal side. A short canal may lead into it from behind. The meganucleus is composed of numerous rounded masses (over twenty in some individuals) scattered through the larger part of the body. The anterior third or fourth of the body, how- ever, 1s usually free from nuclear material. Although the mouth of Loxophyllum is an inconspicuous slit near the edge of the body the animal is nevertheless able to ingest comparatively large forms. Rotifers form a common article of diet. I have often seen Loxophylla containing specimens of Anurea cochlearis and other rotifers of as large size, the body. being thereby much distorted in shape. In ejecting the lorica after the rotifer had been digested the body is much lacerated, but its power of rapid regeneration soon causes it to assume its nor- mal outlines. I have never been able to obtain Loxophyllum in abundance. Like many other predatory infusoria it thrives only in compara- tively pure water and quickly disappears 1 in the presence of putre- fying material. It is found on aquatic vegetation, and sometimes appears on the walls of aquaria, especially those supplied with run- ning water. A favorite situation is on the side of an aquarium just below the surface of the water. NORMAL MOVEMENTS The normal movements of Loxophyllum, compared with those of most infusoria, are sluggish, a circumstance which makes it easy to study the precise way in which they are performed. ‘The creature glides along the substrate on its right side, moving its anterior end about slowly as if feeling its way. Its usual mode of locomotion is as follows: It elongates the body, swims forward a short distance, then contracts, swims backward, turns toward the oral side, and then elongates, and swims forward in a new direc- tion. As it generally swims but a short distance before jerking back, the organism circles about toward the oral side in nearly the T he Behavior of Loxophyllum 40! same situation. I have often observed specimens on the side of an aquarium that remained over an hour within a few millimeters of their original position, although continually moving about. It is an interesting fact that the motor reflex or avoiding reaction in Loxophyllum takes place in a direction | just the reverse of that of Paramecium and many other infusoria; the turning 1s always to the oral instead of the aboral side. Most of the turning, how- ever, occurs after the infusorian has ceased to swim backward which makes it probable that the anterior cilia are relatively more active at this time. While swimming forward the direction of movement is at the same time more or less toward the aboral side, and the backward movements are more or less toward the oral side, but the principal change of direction occurs at the close of each backward movement. ‘The infusorian thus continually cir- cles about to the oral side. When swimming backward the body is generally bent over to the oral side, often throwing the oral margin into one or more folds. The movements of the body vary considerably in rapidity accord- ing to the degree of excitement of the animal, but I have never seen an individualin a state of absolute quiet. There is a certain regu- larity or rhythm of the forward and backward movements which is fairly constant for a long period. Most of the individuals in a dish move at a tolerably similar rate if some of them have not been more disturbed than others. At times Loxophyllum may glide forward for a considerable distance without reversing its direction, but it does this, I believe, only when in a comparatively high degree of excitement. its body is then strongly elongated and nearly straight. In the short forward movements which are performed in its usual circling about near one place the body is not so greatly elongated and it is bent over more strongly toward the aboral side. The extension and straightening of the body during its more rapid gliding aid the animal in maintaining a more direct course, although it com- monly veers around somewhat to the aboral side. The body of Loxophyllum is very mobile and it is able to change its shape in many ways by contracting locally in different regions. It may contract to half its maximum length, bend up or down or 402 S. Ff. Holmes to either side, or twist about on its long axis. It is almost con- stantly bending and writhing about in various ways. The ante- rior extremity is the most active as well as the most sensitive part of the organism. It is continually executing small movements, bending back and forth or up and down as if attempting to explore its environment. ‘The oral margin of the body at times performs a sort of undulating movement, usually when it is lifted up free from the surface. This motion when the animal is largely free from contact with the substrate may become a vigorous and rapid one and serves to turn the body about in the water. When slight the fluttering movements are confined to near the anterior end of the body but when more decided they involve a considerable part of the oral margin. When turned over so as to lie on its left or unciliated side Loxophyllum may right itself in several ways. At first it writhes about for a little while, but it is usually only a short time before one of its methods of turning over is hit upon. One common method is to raise‘up the ends of the body more or less, twisting about the anterior end until its right side touches the bottom. The rest of the body is then pulled over much as in the common righting movements of a planarian. Often, but not always, this is accompanied by a rapid undulation of the oral margin which apparently aids the turning. Generally Loxophyllum raises the oral side and twists about aborally, but it not infrequently turns over in just the reverse direction. Frequently the body is twisted about when the two ends are free in the water, the turning begin- ning at the anterior end and continuing until the whole body is twisted about. When placed on its left side Loxophyllum sometimes bends the anterior end of its body upward at right angles to the long axis, raising it until it stands erect, and then toppling over upon the opposite side. In one instance | saw both ends raised up to about the same extent until they nearly met, forming a sortof hoop; then the animal rolled over, through the force of ciliary action until the anterior end touched the bottom, when it attached and glided ahead, thus straightening out the body into its normal position. Loxophyllum is, as a rule, rather reluctant to swim through the rm, The Behavior of Loxophyllum 403 water, but when in a state of unusual excitement it may do so quite readily. It swims in a spiral course like most infusoria, circling about in a clockwise manner and at the same time rotat- ing on its long axis in the same direction. ‘The spiral course is maintained not so much through the natural asymmetry of the body, as by the fact that the body is curved tow ard the inner side of the spiral and held at a slight twist. By means of its spiral movement Loxophyllum 1s ane to travel in a nearly straight gen- eral course for a considerable distance. REACTIONS TO STIMULI Mechanical. In experimenting on the reactions of Loxophyl- lum to mechanical stimuli a glass rod was used which was drawn out into an exceedingly fine thread at the tip. By using a Braus- Driiner binocular microscope it was possible to apply stimuli of various degrees of intensity to any part of the body and readily observe the result. When the anterior part of the body is stimu- Jated the animal contracts longitudinally, swims backward and to the oral side, and bends its body orally at the same time. After this it extends the body again and swims forward. Stimuli applied to the tip of the body most readily produce this reaction.- It may be produced even without contact by moving the rod about a short distance in front of the anterior end. Stimulating either side of the body back to a considerable dis- tance produces the same reaction. It is obvious, therefore, that when the animal is stimulated on the oral side it turns directly toward the stimulus instead of away from it. Repeated appli- cations of the stimulus to the oral side will cause the animal to keep turning toward the stimulus, notwithstanding the unadaptive, or even injurious, nature of the response. The facility with which a stimulus evokes a response diminishes toward the posterior end of the body. Stimuli applied between the middle of the body and the posterior third, especially after the second or third trial, frequently produce no response, even when quite strong. The animal may often be poked about in this way, almost to the point of producing mutilation, without suffering any interruption of its usual activities. 404 S. Ff. Holmes If a stimulus is applied to the posterior end of the body, or a short distance in front of this on either side, the usual motor reflex is not produced. The animal swims directly forward. With repeated stimulation of the posterior end it may be kept swimming forward for a long time. If the stimulus is applied during the progress of the animal the rate of movement isaccelerated. Essen- tially the same behavior has been found by Jennings to occur in Paramecium in response to weak stimuli, and | have often ob- served the same phenomenon in this and several other infusoria. It indicates the first step toward reacting in a specific manner to the localization of a stimulus. The reactions of Loxophyllum are quickly modified by suc- cessive responses to stimulation. ‘This, I believe, is in large part due to the dulling of the sensitiveness of the organism through the repetition of stimuli. A very slight stimulus to the anterior end of the body suffices at first to produce a reaction. With repeated poking the anterior end becomes so dulled that the organ- ism may continue to swim forward in spite of frequent stimulation at this point. Recovery, however, is quick, for in a few minutes the responsiveness is as great as ever. A similar result is more quickly reached by the application of stimuli to the sides. The motor reflex may be elicited v ery readily for a few times, but it soon requires much stronger stimuli to bring it about. If the stimuli are applied quite far back it requires fewer stimulations before the animal refuses to respond at all. There seems to be a tendency for the organism to resume its usual activity which asserts itself when its sensitiveness becomes dulled so that it does not react so readily to stimuli. It may be kept from going forward, for instance, by repeatedly stimulating the anterior end of the body. But sooner or later the tendency to normal activity predominates and the animal may go forward in spite of considerable stimulation. Notwithstanding its rapid habituation to stimulation Loxo- phyllum exhibits certain features of behavior that seem referable to the summation of stimuli. Repeated stimulation may induce a condition of unusual excitement which may be manifested in con- tinual and quite rapid swimming, increased writhing movements, Pr. The Behavior of Loxophyllum 405 or in increased rapidity of its ordinary back and forth movements. The reactions may be less easily evoked, but the spontaneous activity of the organism 1s increased. Chemical. Owing to the small number of specimens available few experiments on the reactions of Loxophyllum to chemicals were tried, since they frequently produced fatal effects. Several drops of water containing specimens of Loxophyllum were spread out on a slide and a minute grain of salt or a small drop of weak acid was placed at one edge. “The Loxophylla showed no ten- dency to go directly away from the diffusing chemical. Some- times they would go toward it. In many cases they would move about irregularly until overcome by the chemical in case it was strong enough to be injurious. Tne majority of the individuals, however, usually succeeded, sooner or later, in getting away to a safe distance. When swimming toward the chemical the anterior end is more strongly stimulated and the animal swims backward, turns, and goes in some other direction. ‘The length of the back- ward course being dependent on the strength of the stimulus received, the animal is apt to go back further when pointed to- ward the stimulus than when pointed away from it. Also excur- sions toward the stimulus are more quickly checked than those in other directions. In consequence of these reactions the animal works its way farther from the stimulating substance. ‘The proc- ess is a slow one, especially since, owing to its natural rhythm of movement, Loxophyllum frequently changes the direction of its locomotion. Even when pointed directly away from the chemical it does not usually go very far before backing up and turning in another direction, and thus much of what was gained is lost. The whole process of negative chemotaxis in this form is a very slow, uncertain, and bungling one. In one experiment I placed a lot of Paramecia with several Loxophylla, and a drop of acid was introduced at the edge of the liquid containing them. ‘The Paramecia showed a very quick and marked negative reaction. “The Loxophylla were incompar- ably longer in getting away from the chemical. Some went toward it and were killed, while practically all of the Paramecia got safely away from the injurious substance. 406 8. 4. Holmes THE BEHAVIOR OF PIECES OF LOXOPHYLLUM It has been shown by Jennings and others that pieces of infu- soria react in much the same way as the entire organism so far as this is rendered possible by the shape of the parts concerned. The observations which I have made on the behavior of pieces of Loxophyllum confirm the general results obtained by other observers and add a few points of interest, especially in relation to the subject of regeneration, which is treated in a subsequent section. A specimen was cut transversely in two at about the anterior third. ‘The two pieces swam rapidly apart, the anterior one going forward, the posterior one backward. For some time the anterior piece swam about in a circle toward the aboral side. After this it began to move alternately forward and aborally, and then back- ward and orally. At times the oral margin would be raised up and moved in an undulating manner like the edge of a flag, and some- times the piece would turn completely over on its left side, but it usually glided along on its right side with little or no marginal motion. Biter about six minutes its backward and forward excur- sions became limited to about the length of its body. In going ahead there was a slight extension of the body, while in going backward the body was always widened. A little later its motions became confined to nearly the same spot. It would go forward, then backward, turning through twenty or thirty desu and then go forward again. Its behavior had become, therefore, much like that of the normal animal under usual conditions. When the anterior end of the piece was stimulated by contact with a fine capillary glass rod it would swim backward and turn toward the oral side. When the posterior end of this piece was stimulated it would not react nearly so readily, and often quite strong stimuli produced no effect. When the response did occur, however, it was manifested in two different ways. At times the piece would flatten and swim backward, especially if the stimulus were strong. At other times the body would elongate and swim forward. In the first case it is probable that the animal was pushed ahead so that the more sensitive anterior end was stimu- The Behavior of Loxophyllum 407 lated, and this would naturally produce a backward movement. The second response is like that of the normal animal when irri- tated at the posterior end. When the piece was swimming through the water stimulation of the posterior end frequently resulted in a marked acceleration of its speed. The righting movements of the anterior piece were much like those of the entire animal. Generally the oral margin would be raised up and waved rapidly back and forth, a movement which probably causes the oral side to be elevated until the piece topples over upon its right side. Considerable variation occurs in the method of turning over in the pieces as in the whole organism. In another experiment in which attention was paid mainly to the movements of the posterior piece a Loxophyllum was cut in two near the middle. ‘The posterior piece swam backward quite rapidly for about three minutes. After this its movements became slower and it would swim forward occasionally. In a few min- utes the forward movements began to increase, and after a while the infusorian settled down to moving forward and backward to about its own length. Sometimes it would raise itself from the bottom and tumble over on the other side only to quickly turn again into its normal position. At one time it left the bottom and swam 1n a spiral course for a considerable distance. When moving backward the piece would widen out, especially at the anterior end. When moving forward, on the other hand, the piece would become elongated and strongly drawn together at the anterior end. ‘These changes of shape were invariably associated with the different directions of movement. Stimulation of the posterior end of the piece causes it to pinch together in front, elongate, and swim forward. When the anterior cut end is stimulated the piece spreads out and swims backward. Comparatively strong stimuli are required, however, in this case as the cut end 1s considerably less sensitive than other regions of the body. By stimulating either end of the piece it may be caused to swim continually forward or backward as the case may be. In about thirty minutes after the cut was made the regeneration of the anterior end of the piece was well under way. ‘The slight waving motions of the anterior border were visible, but not so 408 Ss: 5 i Holmes apparent as in the entire organism. ‘The piece kept going for- ward and aborally a very short distance, then backward and orally, circling about in nearly the same spot, with a regular, incessant, rhythmical movement. In going forward the body became not only elongated but curved toward the aboral side. When con- traction occurred during its reversed movement, it was greater on the oral than the aboral side as it 1s in the entire individual. ‘The behavior of several other pieces taken from the two ends of the body was essentially like those described. Sometimes the pieces would swim about in a spiral course through the water for several minutes, but eventually they all settled down to the same regular back and forth movements. Pieces cut from the middle of ihe body showed the same rhythmical movement, extending and bending slightly aborally as they went forward, and contracting more on the oral side as their motion was reversed. RHYTHMICAL ACTIVITY OF LOXOPHYLLUM When first observing the activities of Loxophyllum I came to the conclusion that the frequent reversals in the direction of its move- ments were due to reactions caused by minute objects with which the sensitive anterior end of the body came into contact. But further observation showed that these reversals were due to inter- nal rather than external causes. When specimens of Loxophyl- lum were placed in water as free as possible from small particles the same regularity of reaction was found to continue. When gliding on the upper side of the surface film of a drop of clear water Loxophyllum reverses its movement about as often as when in the midst of objects with which it is continually colliding. But any doubt concerning the inherent rhythmicality of its movements is removed when we consider that the pieces into which the body is cut move back and forth at about the same rate as the whole animal. The cut anterior ends of the pieces of Loxophyllum are com- paratively insensitive to mechanical stimuli, and there can be no doubt that the movement of these pieces is a manifestation of rhythmic activity comparable to the beating of the heart muscle of higher animals. There seems to be no constant difference in The Behavior of Loxophyllum 409 the rate of the back and forth movements between a piece contain- ing the sensitive anterior end of the body and a piece from any other region. Even very small pieces show the same rhythm. On account of its rhythmical activity Loxophyllum does not have to wait for something to turn up in order to acquire new experiences. Its life is one of continual trial. Only to a compar- atively slight extent is its activity, under usual circumstances, directed by external conditions at all. It goes forward, back, turns orally, goes forward and back again, and so on, repeatedly, through its own inherent activity. In many of the lower organ- isms behavior mainly consists in more or less direct responses to external stimuli with little spontaneous movement, but unless something unusual affects it Loxophyllum keeps circling about near the same place for a long time. When tt meets with a strong or injurious stimulus it has its own methods of getting out of He way, but its ordinary behavior is mainly pied ie internally initiated impulses. COMPLEXITY OF BEHAVIOR From the preceding account it is evident that the behavior of Loxophyllum is considerably more varied than that of Para- mecium and many other infusoria. Paramecium, for instance, has a very few stereotyped modes of behavior, such as spiral swim- ming, the motor reflex, acceleration of forward motion when lightly stuumulated at the posterior end, the thigmotactic response, and bending the body when crowded among obstacles. Loxo- phyllum has not only all these responses, but several others in addition, 7. e¢., gliding movements, regular changes in the form of the body accompanying forward and backward movements, small feeling movements of the tip of the body, undulations of the oral margin, twistings, turnings, and contortions of the body under various conditions, special movements involved in swallow- ing large objects, and several kinds of righting movements. This greater complexity of behavior is probably a consequence of the fact that most of the creature’s life is spent in contact with solid objects. It appears to be a general rule that the behavior of the 410 S, 7% Holmes free swimming infusoria is more simple than that of the creeping or the permanently attached forms. REGENERATION While studying the behavior of pieces of Loxophyllum I found that regenerative changes set in soon after the animal was cut in two. A good opportunity was thus afforded for watching the regeneration of the animal which takes place so rapidly that one might almost be said to actually see it going on. ‘To determine, so far as possible, the exact method followed in regeneration is always a matter of interest and importance; and a form in which 6 Fig. 1 Showing the course of regeneration of a piece from the posterior end of Loxophyllum. The dotted line in this and the following figures indicates where the cut was made. the process can be watched under the microscope and followed step by step is especially favorable for this purpose. A Loxophyllum was cut in two near the middle by a slightly oblique cut (Fig. 1). In the anterior piece the sides near the cut end were drawn inward and soon met, thus closing in the cut portion of the margin and giving the piece much the appearance, except in its relatively greater width, of the normal animal. In the posterior piece to which attention was mainly directed, since much greater modifications were necessary to restore the normal form, the first step in the process of regeneration was the closing in of the sides at the anterior end. ‘The piece continued to swim forward and backward, undergoing the regular changes in form a - The Behavior of Loxophyllum 411 that accompany these movements which have been previously described. Soon, apparently as a result of these stretching out movements that accompany its swimming forward, the piece began to acquire a more narrow and elongated form. With each for- ward motion the sides would be pushed around more toward the middle of the cut end which gradually became reduced in extent. With each movement ahead it could be seen that not only the body elongated but that it elongated more on the oral than the aboral side, causing it to bend toward the aboral side at each advance. ‘This bend is not the result of the contraction of the aboral side, as one might very naturally suppose, but the exten- lod a a“ 6 ~ ae 4 iD, Fig. 2 Showing the regeneration of the posterior part of the body when cut off obliquely. sion of the oral side. Soon the oral side begins to grow longer than the aboral and to become pushed Gere the anterior pee of the body. The striations which originally ran in a longitu- dinal direction are now bent around the anterior end of the body more on the oral side than on the aboral. ‘There is no formation of new tissue here, and no differentiation of new cilia on the cut surface, but the oral margin becomes stretched around the anterior end of the cut piece. Both sides of the body extend and contract, the movements being greater toward the anterior end. ‘This end becomes (in consequence of these movements ?) more narrowed and more like that of the normal individual. ‘The cut end of the body is closed in by the gradual extension of the sides which fin- 412 S. F. Holmes ally meet, the point of union being carried by the greater exten- sion of the oral side so that it finally comes to lie on the aboral side some distance behind the anterior end. ‘The method of regeneration here followed in restoring the exter- nal form of the body is the simplest and most direct that can readily be imagined. ‘The elaboration of new structures is re- duced to a minimum. The part of the infusorian behaves much as an entire individual, narrowing the body as it advances and stretching the oral more than the aboral side; and this behavior seems to help mold the part into the final form. In order to find out how small a part of the differentiated oral margin could be stretched out to form the entire oral margin of the new individual a piece was cut obliquely across the body / 2 3 (\ ( ( ge He a Fig. 3 Regeneration of small pieces from the anterior half. (Fig. 2), so that the oral side of the posterior portion was consider- ably shorter than the aboral. In this case the general method of regeneration was much as before. Both sides curve in to close the cut end, the piece elongates and becomes narrowed; the oral side 1n the movements of the animal is extended more than the aboral, and we can see that it 1s gradually stretched forward; and finally it is pushed around the anterior end. The middle part of the cut surface which is becoming more and more reduced is apparently drawn back, but this appearance is due, I think, not to its being pulled back in the center, but to the extension of the two sides around it. The length of the ciliated oral side when regeneration 1s complete is considerably greater than at first. There was no extension of cilia in this case upon the cut surface. The anterior limit of the oral margin was very distinct and could ee The Behavior of Loxophyllum 413 be followed in its course without any difficulty. ‘The short oral margin of the posterior cut piece was simply stretched out to form the whole oral and anterior margins of the regenerated individual. Essentially the same method is followed in the regeneration of comparatively small transverse pieces (see Fig. 3). The experiment was then tried of reducing the oral margin still more. By making a cut across the anterior end and a longitudinal cut near the oral side the whole oral margin was removed except a small part near the posterior end of the body (see Fig. 3). In this case the amount of cut surface exposed was very much greater than in the previous experiments, so much so that it seemed incred- 2 3 4 | 4 7 \ / | Fig. 4 Regeneration of a specimen from which the anterior end and most of the ciliated oral margin was cut off. ible that the small remaining part of the ciliated margin could be extended so as to stretch over it, especially since the part remain- ing is one that gets little stretching during the usual activities of the animal. As might be expected, although it was stretched around the oral side to a certain extent, this part failed to give rise to any but a small part of the new oral side. ‘The new oral margin with its differentiated structures had therefore to be pro- duced by a new method. ‘The piece began to elongate and become narrowed and rounded in front. Owing to the lack of the con- tractile and extensile elements of the oral margin the character- 4I4 S. Ff. Holmes istic pushing ahead of the oral side did not occur. ‘The aboral side in fact began to be pushed ahead of the oral which accounts for the form of the pieces shown in the figures. After several hours the oral margin became thinner and clearer and the gran- ules of the endoplasm came to lie further from the edge. A slight transverse striation could be detected in it such as occurs more plainly in the normal individual, and soon short cilia began to be put out here and there chiefly toward the posterior end. As the clear margin became broader it showed a longitudinal striation and soon began to extend and contract more during the move- ments of the body. As the oral margin slowly acquired its char- acteristic differentiation it began to push ahead and extend around the anterior end where its striations assumed the usual bend. Fig. 5 Regeneration of a piece from the middle of the body from which the oral margin was removed. In this case regeneration was very slow compared with the two preceding experiments. ‘The piece was larger in size than the others but more differentiation had to be accomplished. Not until the oral margin became furnished with its cilia and its differ- entiated contractile elements so that it was capable of performing its usual réle in the movements of the animal was there any marked progress in molding the body into its final shape. ‘The anterior end, although it had become narrowed and rounded soon after the operation, did not take on any of its characteristic structural features until the oral margin became differentiated and began to be stretched around the front as in the cases of regeneration just described. The development of the new cilia extended gradually forward from the small part of the ciliated margin that remained and the The Behavior of Loxophyllum 415 possibility suggested itself that the new cilia which were devel- oped arose through the influence of the old ciliated margin or of material which were developed arose through the influence of the old ciliated margin or of material which might be derived from it. ‘To test this possibility a specimen was cut as is shown in Fig. 5 so as to leave no part of the cialiated oral margin remain- ing. The general course of regeneration is indicated by the Figs. 2-6. It will be seen that the aboral side, as before, extends at frst more than the oral, but after the oral margin becomes dif- ferentiated in its characteristic fashion it pushes around more than the aboral and produces the usual curvature at the anterior end of the body. ‘The cilia made their appearance in scattered groups about fourteen hours after the cut was made. iG UV Fig. 6 Regeneration after removal of the dorsal half. When Loxophyllum i is cut in two longitudinally the process of regeneration is comparatively slow. ‘The usual form of the body may be approximately reached in a comparatively short time, but the differentiation of the structures characteristic of either margin requires several hours. Fig. 6 represents a specimen from which the aboral half was removed by a longitudinal cut. ‘The posterior end of the first became bent aborally and was brought forward so that the two parts of the cut margin met and fused together. ‘The body as a whole shortened and widened; the injuries that were incidentally made near the anterior end of the body were repaired, and while the cut margin so far as could be ascertained seemed to close by the approximation of the upper and lower edges it was over twelve hours before the groups of trichocysts character- istic of the aboral margin made their appearance. 416 S. Ff. Holmes In a specimen cut longitudinally part way through the body (Fig. 7), regulation was effected by the meeting and fusion of the cut surfaces. [he movements of this specimen were of interest. As it swam forward the two sides became crossed. During back- ward swimming on the other hand, they diverged very widely. This is doubtless due to the fact that in the lengthening and short- ening that respectively accompany the forward and backward movements of the organism the marginal regions of the body are more active than the middle. The mechanism of the extension of the sides I have not ascertained. THE ROLE OF MOVEMENTS IN REGENERATION The foregoing experiments make it probable, as Child has attempted to show in other forms, that the role of the movements Fig.7 Regulationin a specimen cut longitudinally as shown in 7. In 2is shown the shape assumed when the animal is swimming forward; 3? shows the form during backward swimming. of the organism in bringing about the normal form of the body is an important one. There are, in fact, few cases in which the eficacy of the factor of movements seems more manifest. To a considerable extent at least the organism seems to pull itself into shape. It has certain ways of acting which, as observations on the behavior of the parts have shown, are characteristic of the behavior of even quite small parts in much the same way as they are of the whole. A small piece cut from almost any region of the body shows the same rhythm of back and forth movements, the same correlation of extension with forward movement and of con- traction with backward movement, and to a certain extent the same oral and aboral bendings as the entire animal. And when one carefully follows the course of regeneration it seems evident SPS, 1a The Behavior of Loxophyllum 417 that these movements are gradually working the part into the form of the whole. In the regeneration of the posterior half, for instance, one may see the oral margin extending and extending, growing a little longer with successive stretchings, until it curves about the anterior end of the body, and its striations are bent around so as to give the characteristic appearance of that region of the normal animal. The same kind of action is apparently instrumental in producing the same kind of form. But precisely what is the relation of the movements of the organ- ism to its regeneration does not, however, lie on the surface. It seems evident that the movements have an important part in shaping the general outline of the body. But are they the funda- mental causes of this change of shape, or agencies which merely assist or accelerate the action of other formative factors? “The experiments performed, while they indicate the importance of behavior in regeneration, show, I believe, that this factor is of a secondary or subordinate nature. It will be instructive to con- sider the course of regeneration in those experiments in which most or all of the oral margin was removed. Here regeneration was forced to follow a very different method from that adopted in the cases first described where a part of the oral margin was stretched out into the whole. ‘The new margin had to be formed entirely de novo. There were involved the thinning out and clearing up of the oral side, the differentiation of new contractile threads, new trichocysts, new cilia, a complicated ordering of newly differentiating structures. ‘The gross movements of the body could have had very little to do with all this. Until these differentiations were made the movements of this side of the body were not of the usual kind. Commonly the oral side extends and contracts more than the aboral, but when the marginal elements of this side were removed the opposite side was the more active. The oral side did not extend so rapidly as the aboral until the structures characteristic of the oral margin were established. If in the experiments first described the general form of the body seemed to be produced by the characteristic behavior of the animal, the characteristic behavior in this case had to wait until its structural basis was established by comparatively slow differ- 418 S. Ff. Holmes entiations of new parts. If form seems in some cases to be molded by function, function in turn is apparently the result of organ- ization. [he modifications of form and function, of course, go on pari passu, and are after all but different aspects of the same process. The gross activities of the organism are largely depend- ent on the finer organization of the animal since they are carried on in a very similar way, even by comparatively small pieces of the body. Where, as in the experiment cited, the animal is cut in such a way as to modify certain of its grosser movements the finer differentiations go on until a structure is produced which then undergoes the movements characteristic of the whole when the external shape is rapidly assumed. The processes of building up the finer structures of the body, the formation of new myonemes, cilia, etc., are really the fundamental features of regeneration. Pulling the body into shape 1s a sort of secondary matter in which the gross movements play an important part, to be sure, but these are themselves dependent upon the finer differentiations. In certain cases among the infusoria, such as some of the Hypotricha, the comparative rigidity of the body excludes the factor of movement from playing a very important role in shaping the outlines of the regenerating organism. Yet these forms regenerate with great readiness. Whees the factor of movement is of importance in the regeneration of the infusoria, it is, | believe, rather in the nature of an aid to other formative factors than an essential and fundamental factor itself. Zodlogical Laboratory University of Wisconsin REGENERATION AS FUNCTIONAL ADJUSTMENT BY S. J. HOLMES With One Ficure In a previous paper' I have ventured to outline a general theory of form regulation, based on the conception of an essentially symbiotic relation between the parts of an organism. ‘The con- ception is, of course, nothing new, but, so far as I am aware, no one has hitherto attempted to deduce from it a theory of regener- ation and other processes of a regulatory nature. The theory may be stated in brief as follows: The various parts of an organism are supposed to stand in such a relation to each other that each part derives some advantage or is helped to perform its normal functions through the materials and stimuli it receives from other and especially the contiguous parts of the organism. Each part in turn contributes something to the normal functioning of the parts surrouningit;the relation is one of mutual dependence. Being mutually dependent, the parts of an organism tend to settle into a condition of functional equilibrium. When a part of the organism is removed and tissue of an undifferentiated nature 1s produced in its place, this new tissue develops in the direction of the missing part because this line of development is favored through the influence of the surrounding parts. Whatever advan- tages accrued to the part formerly in this position from its relations to the parts around it would also accrue to this tissue in so far as it differentiates in the same way as the part removed. The new tissue differentiates according to the functional demands upon it and its line of specialization may be regarded as a case of func- tional hypertrophy. Regeneration of the missing parts, therefore, may be interpreted as an expression and act of getting back into a condition of functional balance. ' Archiv fiir Entwickelungsmechanick, xvii, Bd., p. 265, 1904. Tue JourNAL oF EXPERIMENTAL ZOOLOGY, VOL. IV, NO. 3. 420 S. Ff. Halmes Che process might be illustrated by the case of a social organism composed of animal cells and sy mbiotic algze which | described in my former paper. “‘We may suppose that both animal and plant cells tend to grow and multiply as far as circumstances permit. As these cells depend upon each other to a certain extent, neither kind of cell will tend to preponderate over the other, but they will all adjust themselves to a condition of approximate equilibrium. Now suppose that a considerable number of the alge of this com- posite organism be removed. ‘There is a functional demand by the rest of the organism for the products of the alge and an excess food supply for those which remain. ‘The algae, therefore, are supplied with exceptionally favorable conditions for growth and multiplication, and will be stimulated to regenerate their missing number. | By supplying the functional demand of the animal cells they indirectly beneft themselves, because by producing more oxygen they enable the animal cells to produce more of the sub- stances which they utilize as food. If we suppose that in our hypothetical organism there are, in addition to the two kinds of cell mentioned, indifferent cells which are able to develop into either animal cells or alga, it seems probable that, in the event of the removal] of the algz, the indifferent cells will differentiate so as to take the place of the missing numbers. * * * “For the sake of a simple illustration we have described an organism consisting of but two kinds of cells, but there is no reason to doubt that in a complex organism consisting of many varieties of cells standing in a symbiotic relation there would be a similar regeneration of any part that is removed. Let us imagine an or- ganism made up of a number of differentiated 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 dependent. 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 mutilation. We may represent our hypothetical organism graphically by the following Regeneration as Functional Adjustment 421 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 utilized by B, G and F, and derives something in return from each of these sources. Now suppose 4 is removed; the indifferent cell lying near by, no longer held in check by the same stimuli, begins to grow and develop. What line of differentiation will it most naturally take. Owing to the symbiotic relation subsisting be- tween 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 re- Oe HO® PO generation of the missing part.” For applications of this theory here set forth in barest outline to morphallaxis, heteromorphosis, physiological regeneration, and other modes of regulation, refer- ence may be made to my former paper. In some recent articles the problem of regulation has been ap- proached from points of view somewhat similar to my own. _ Jen- nings” has attempted to show that regulation in behavior is funda- mentally similar to other forms of regulation. The method of trial and error, which is so pronounced a feature of the behavior of lower organisms, and one through which they secure a large part of their adaptations to external conditions, is assumed by Jennings to be followed in the various processes occurring in ? This Journal, vol. ii, and Behavior of the Lower Organisms. New York, 1906. 422 S. Ff. Holmes the regulation of organic form. “A disturbance of the physio- logical processes,” he says, “results in varied growth activities. Some of these will relieve the disturbance; the variations then cease and the processes are continued.”” The result of the selec- tion of those growth processes which relieve the disturbance—or we might say make for functional equilibration—is the restoration of the lost part. Jennings has not attempted to develop a theory of form regulation in detail farther than to show the fundamental similarity of the method of regulation in various fields, but he holds that my own point of view so far as form regulation is con- cerned is in “essential agreement” with his. Child® has recently outlined a theory of regulation which, as he states, is “somewhat similar to that adopted by” myself, although differing in certain important particulars. According to both Child and myself, regeneration and other formative processes are the result of functional activity, or more specifically, func- tional equilibration. ‘Tissue differentiates in the direction of the missing part because it takes on the functional activity of the miss- ing part. To cite an illustration by Child, “after removal of the anterior or posterior end in Bipalium or Planaria maculata the terminal regions of the piece remaining are subjected to condi- tions somewhat similar to those existing in the terminal regions of the part removed. ‘The anterior end of the headless piece of Planaria is subjected to external conditions more or less similar to those to which the old head was subjected; moreover, its rela- tion to the other parts of the body is more or less like that of the head. Stimuli resulting from forward movement affect it first, and are transmitted from it to other parts, etc. Functionally speaking, it serves in some degree as a head. ‘The case is similar as regards the posterior end. After removal of the original pos- terior end, the posterior region of the piece functions in some degree as a posterior end, or to put the matter more strictly, its functional relations with other parts are more or less similar to those of a ‘tail’ or posterior end. * * * ‘Redifferentiation occurs as a result of a functional substi- 3 Archiv fiir Entwickelungsmechanick, xx Band, 1906; and this Journal, vol. iii, 1906. Regeneration as Functional Adjustment 423 tution of a larger or smaller part of the old tissues of the piece for the part removed; the substitution may be imperfect or incom- plete at first, and gradually attain completeness. In consequence of this functional substitution, the structure of the part involved is altered until it comes to resemble more or less closely that of the part removed.” Where regeneration through the formation and differentiation of new tissue occurs, it 1s this tissue which becomes the func- tional representative of the old part. In the regeneration, for instance, of the arm of a star-fish, or the leg of an arthropod or amphibian, the new tissue ““must be subjected to many condi- tions—internal and sometimes external—similar in a greater or less degree to those to which the part removed or some portion of it was subjected.” The factor of the exercise of a part which Child has regarded in many cases of so much importance, 1s not always necessary for regeneration. ‘‘The growth of the new leg is not the result of the attempt to use the leg which is missing. ‘The growing tissue begins to develop into a leg because its relations to the other parts of the system are in some degree similar to those of the leg removed. As it grows, the conditions approach much more and more nearly those to which the normal leg is subjected, 7.¢., there is a gradual return of the functional conditions to the normal.’’ The application of Child’s theory to the subject of polarity, heteromorphosis, and many other regulatory phenomena, it will not be necessary for our present purpose, to discuss. ‘The funda- mental idea of the theory 1s that form regulation is a result of func- tional regulation. In certain cases Child attempts to show that this supposed functional relation actually occurs as a consequence of mechanical conditions. ‘The anterior end of a planarian gets stimulated by the water and by the impact from foreign bodies much like the head does. The posterior end of Stichostemma is used by the animal in locomotion much as the tail is. At first the external stimuli affecting the ends are much the same whether they are anterior or posterior. In the movements of the animal both re- ceive frequent impact from contact with various objects. ‘There 424 S. f. Holmes is a difference perhaps in the mechanical stimuli received, but granting that these start the course of differentiation in different directions, they are entirely inadequate to account for the whole process of differentiation, as Child himself would probably admit. Where the factor of movement is absent, Child has recourse to the supposition of some other form of functional substitution, but he gives no clear account of why the substitution should occur. To say that the end of an arthropod’s appendage is regenerated because of the functional activities that occur within it, that mor- phallaxis occurs when the part readily takes on the function of the whole, and that regeneration takes place because the functions of the missing part are imposed upon the new tissue that is developed in its place, may all be very true so far as it goes, but until some principle for the explanation of this functional adjust- ment is brought forward the explanation of regeneration is far from complete. If form regulation is a consequence of functional regulation, as Child and | agree that it is, the interpretation of functional regulation 1 is the next obvious step. The inadequacy of Child’s theory is, that it does not contain any general principle of explanation for that functional substitution and equilibration upon which it is assumed that form regulation depends. ‘This, however, is a matter of incompleteness rather than error. But I suspect that when his theory comes to be developed so as to sup- ply this missing element, it will involve, to make it workable, the assumption of some such symbiotic relation between the parts of an organism as | have assumed. It is a strong point in favor of the theory of symbiosis that it affords to a certain degree an expla- nation of physiological adjustment; in fact, it is primarily a the- ory of physiological equilibration. ‘This physiological adjustment brought about through the symbiotic relations of the parts may, as | have attempted to show, be explained, or at least much of it may be explained, as the outcome of a tendency toward chemical equilibration, To the extent that this is true, we have an explana- tion of the regulatory activities of an organism in terms of famil- iar chemical phenomena. The conception of something like a symbiotic relation between the parts of an organism which is involved in my own theory Regeneration as Functional Adjustment 425 of regulation, Child rejects, but I think on insufhcient grounds. I have assumed that, according to the symbiotic relation of the parts of an organism, upon removal of a part, such as 4, in the hgure, the undifferentiated tissue in the region of 4 will differen- tiate in the direction of the missing part because of the functional demands, or for what, for want of a better term, | have called social pressure, upon that tissue. “ This,’’ according to Child, in refer- ring to the particular case illustrated by the diagram, “is exactly what will not occur under these conditions. If all the cells 4—F are symbuiotically correlated then removal of one of them, 4, must affect all the others, 7. e., the whole complex is altered by removal of one of its members. It is perfectly clear that the “social pres- sure’ of the altered complex will not be in the direction of differ- entiation of the indifferent cell into something like 4 but in some other direction, in other words, the indifferent cell cannot replace A but will form something different. Moreover, since all the cells were dependent upon 4 in some degree, the removal of 4 will probably render continued existence impossible for some of them and their place will be taken by the undifferentiated cells, but these will also develop into something different because the ‘social pressure’ is altered. It is perfectly Salem that no regulation in the sense of replacement of a missing part could occur in such a complex.” Now this conclusion may be perfectly clear to Dr. Child, but I must confess—perhaps I[ am blinded by my bias in this matter— that it is far from being so to me. According to Child, since the removal of 4 would alter B, G, F, etc., not only something differ- ent would be developed in place of 4, but the whole complex, according to my theory, would be profoundly altered. Now, I admit that the removal of 4 tends to alter B, G and F, etc. How far this tendency will result in a modification of these cells depends on the plasticity of the organism and the degree of mutual depend- ence of the parts—factors of course which vary in different organ- isms. But Child overlooks the fact that according to the sym- biotic relation assumed, the other cells C, D, E, etc., tend to keep B, F, G in their original condition. In so far as these remain in their original state, their influence on the indifferent tissues in the 426 S. Ff. Holmes region of 4 will tend to mold it in the direction of the missing parts. In so far as B, G and F are modifed through the loss of the missing part, their influence on the tissue in the region of 4 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 modifica- tion 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. If the loss of 4 tends to modify B, F and G, the presence of E, 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. How the process works out depends naturally on the degree of specification of the parts, whether or not new tissue is formed in the place of the missing part, and perhaps other factors. If the organism 1s plastic and its parts have not acquired an irretrievable set which prevents further modification, it may be entirely worked over in consequence of the disturbance of its social pressure in the - vicinity of the missing part, thus leading to redifferentiation, or morphallaxis. Whether we have morphallaxis or regeneration in a narrower sense may depend, among other things, upon the degree of specification of the parts. As I have suggested in my for- mer paper (p. 288), and as Child has maintained more at length, regeneration as opposed to “redifferentiation, increases as func- tional specification of the tissues increases or, in other words, the greater the degree of differentiation—the visible result of functional specification—the less likely is extensive functional substitution and consequent redifferentiation. ” This, it seems to me, is very much what one might expect ac- cording to the theory of regeneration | have outlined. Replace- ment of 4, according to Child, “can occur only when the relation is largely one-sided, 7.e., when 4 is dependent on B-F, but these latter are not to any marked degree dependent on 4. In this case, and in this case only, will the social pressure force the undiffer- entiated cell to differentiate into something like 4.” Where redifferentiation from new tissue is concerned, as in the present ie Regeneration as Functional Adjustment 427 case, It is not the relation of 4 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. -—F are relatively fixed, the tissue in place of 4 young and plastic, and more dependent so far as the direction of its differentiation is concerned, upon BF, than these are upon it. We may grant that, when regeneration occurs, the relation of depend- ence between the old parts ay 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. ‘Uhey may be in a symbiotic relation, nevertheless, each part contributing in 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 certain degree the quality and quantity of the activity of the surrounding parts, without producing extensive modifications of structure or function. If the parts B—F were more plastic, absence of 4 would natur- ally tend to cause greater changes in them, especially if new tissue 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 be a progressive 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 equilibrium 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 regu- lation, through morphallaxis, regeneration and the various com- binations of these processes 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. The difficulty pointed out by Child that the process of differen- tiation of a new part seems to begin at the tip and work back to the base is one which gave me some concern when developing my theory, but | think the difficulty is by no means a fatal one. When a developing limb shows first those structures character- 428 S. }. Holmes istic of its distal end we should bear in mind the possibility that the differentiation which first appears is not necessarily that which first occurs. What we know of developmental processes renders it very probable that a great deal of differentiation is going on in the rudiment of the limb before it is manifested by any external signs. Between maturation and cleavage an ovum may show no outward sign of differentiation, but experiment shows that this period is one in which developmental processes are rapidly tak- ing place. Before any external features are produced in the devel- opment of a limb, the main outlines of its differentiation may have been established through influences proceeding from its basal part, after which the tip might differentiate more rapidly than the intervening portion, and the other visible features of structure appear successively toward the base. I do not suggest this merely to save my contention by a retreat into the invisible, but there are certain considerations that make such an interpretation more or less probable. In many cases the visible differentiation is cen- trifugal rather than centripetal. ‘The tail of a tadpole cannot be said to differentiate from the tip toward the body as it regenerates. Zeleny has shown that in the early regeneration, as in the embry- onic development of the antennz of Mancasellus the formation of segments proceeds at first from the base to the tip; later new seg- ments are formed in the reverse direction. 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 environment causes the extremity of an organ to begin to differentiate 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. In my illustrations of the course of differentiation under the guidance of social pres- sure, I have taken the old part asa starting point, but if we have an undifferentiated mass of cells, it is conceivable that, if, for any reason, differentiation should start at the distal extremity of the mass, it might work back under the guidance of social pressure toward the base. ‘The distal differentiation would have to get Regeneration as Functional Adjustment 429 started in the right direction, or something else than the missing organ would be produced. ‘That cases of heteromorphosis some- times occur might be interpreted as the result of such failures. But the comparative rareness of heteromorphosis makes me sus- pect that the beginnings of visible differentiation that frequently appear at the tips of regenerating organs do not occur without any relation to the basal part. ‘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 regen- erated, and forms a congruent union with the basal part, is indica- tive of close interaction of the various parts of developing organ with the body of the organism at all stages of the process. I am inclined to think that neither centrifugal nor centripetal differentiation, expresses the entire truth of the matter, but that the new part differentiates as awhole, much as organs do in embry- onic development, and at all times in intimate functional relations with the old part, differentiation becoming accelerated in one part or another, according to special conditions. If differentiation began at the tip of the rudiment of an organ, and proceeded cen- trally, the whole might be differentiated before the body was. reached, leaving a mass of unused tissue between; or differentia- tion might reach the body before all the immediate parts were produced. If differentiation proceeded in the reverse direction, similar imperfections might arise. We must look upon a regen- erating mass of tissue as one in which incipient developmental tendencies are proceeding in various ways, modifying each other, and gradually working into a condition of physiological equilib- rium with the basal part and with the environment before much outward evidence of differentiation makes its appearance. It is probable that the main elements of a regenerating appendage of an arthropod, for instance, are blocked out before any external marks become visible. Even during the early stages of prolifera- tion of the cells of the regenerating appendages, it is not improb- able that incipient differentiations are becoming established. And the basal part notwithstanding the fact that the visible differen- tiation may take place in a centripetal direction, may exercise a guiding influence at all times over the regeneration of the part, 430 S. f Holmes and determine that it forms in harmony with the rest of the organ- ism. Such a conception is entirely congruous with the symbiotic theory, and is, I believe, consistent with the various observed facts of regeneration. If we explain form regulation as an outcome of functional regu- lation, we make little progress until we have some interpretation of the latter process, and any theory of form regulation which offers nothing in this direction, makes no more than the first step toward an explanation of the phenomenon. In his criticism of the theory of form regulation which I have outlined, Child has advanced arguments which are, I believe, by no means fatal to it, and he has not brought forward any other explanation of func- tional equilibration, which both of us regard as the basis of form regulation. Perhaps this may be supplied in further developments of his theory which Child hints are to be made in the future. While functional adaptation may occur independently of any sym- biotic relations, especially in the direct adaptation of parts to exter- nal conditions, the mutual adaptation of parts which forms so important an element in formative processes are, I believe, for the most part, dependent on symbiotic relations. At present I am unable to see how any general explanation of functional equili- bration among the parts of an organism can be reached unless we assume that the parts are, to a considerable degree, interdependent. Perhaps some other interpretation of functional regulation may be advanced which does not make use of this idea. “That remains, of course, to be seen. But the theory of the symbiotic relation of the parts of an organism has the merit of enabling us to interpret form regulation and functional regulation as the outcome of ordin- ary physiological activities, and hence to give, in a measure, a causal explanation of the teleological behavior which is manifested in so striking a degree by formative processes, and which forms the strongest support of some recent vitalistic theories. So far, at least, I hope it is in the line of progress. Zodlogical Laboratory University of Wisconsin From the Wistar Institute of Anatomy and Biology, Philadelphia SOME FACTORS IN THE. DEVELOPMENT OF THE AMPHIBIAN EAR VESICLE AND FURTHER EXPERI- MENTS ON EQUILIBRATION BY GEORGE L. STREETER, M.D. Associate Professor of Neurology at the Wistar Institute Wirn Srx Ficures In a previous paper concerning experiments on the developing ear vesicle’ it was shown that the group of cells forming the primi- tive epithelial ear cup or ear vesicle of the tadpole is specialized to that degree that although removed to an abnormal environ- ment the cells still continue to differentiate themselves into a struc- ture possessing many of the features of a normal labyrinth. Re- cently it has been shown by Lewis? that even earlier, while still an uninvaginated plate, the ear anlage is already capable of a cer- tain degree of independent differentiation. In the following paper additional evidence will be given of the high degree of develop- mental independence possessed by the early labyrinth cells. It will be pointed out that individual parts of the vesicle may develop independently of the rest of the vesicle. It will also be shown that the process of differentiation extends to the difference existing between a right and left-sided organ. A left ear vesicle trans- planted into the empty pocket left by the removal of the right ear vesicle develops into a labyrinth that is perfect in general form and in its relations to the brain, with the exception that it main- tains its left-sided character; the anterior semicircular canal is found on the caudal side toward the vagus group, while the pos- terior canal lies toward the eye, and likewise the lagena which 1 Streeter, G. L., 06: Some experiments on the developing ear vesicle of the tadpole with relation to equilibration. Jour. of Experimental Zodl., vol. iii. 2 Lewis, W. H., ’07: On the origin and differentiation of the otic vesicle in amphibian embryos. Anatomical Record, No. 6, Amer. Jour. of Anat., vol. vii. THe JourRNAL oF EXPERIMENTAL ZOOLOGY, VOL. IV, NO. 3. 432 George L. Streeter, M.D. normally buds out from the caudal border of the saccule in these cases is found extending forward toward the prootic ganglion. The ear vesicle, however, is not in all respects independent of the surrounding structures. Some experiments which are reported below, indicate that its position in reference to the brain, ganglion masses and the surface of the body is determined by the environ- ment itself; it may be rotated in any direction, and nevertheless it eventually develops in the normal attitude, with the saccule toward the ventral surface, the semicircular canals toward the dorsal surface, the lateral semicircular canal being toward the lateral surface, and the endolymphatic appendage toward the brain. The experiments were carried out on larve of Rana sylvatica and Rana pipiens, and the operating stage was the same that was used in previous experiments.* ‘The time is just at the close of the non-motile stage, and the epithelial ear consists of an invag- inated cup-shaped mass of cells just in the process of being pinched off from the deeper layer of the skin, with the edges turning in to form a closed vesicle. For simplicity the term “ear vesicle’’ will be used even though the closure is not yet complete; the attempt to distinguish between auditory cup and auditory vesicle does not seem to be justified for the present purposes. The technique of the operations was also the same as that described in the previous paper. Notes were made on the behavior of the animals, and at the end of from four to six weeks the specimens were preserved in a chrome-acetic mixture, cut 1n serial sections, and stained with hamatoxylin and congo red. With certain specimens the ear vesicle, adjacent ganglia, and a portion of the central nervous sys- tem were reconstructed after the Born wax plate method. Eleven such models were made, and photographs of some of them are reproduced in Figs. 2, 3 and 6. With the aid of these models it was possible to identify relations and detailed features of the laby- rinths that otherwise could not have been recognized. The morphological features of the experiments will be first con- sidered, and the behavior of the animals and its relation to equilib- rium will be treated separately in the latter part of the paper. 3 Streeter 06: /. c., Fig. 3, p. 547- WA wv, Development of Amphibian Ear Vesicle 433 DETERMINATION OF POSITION OF THE EAR VESICLE The conclusion that the attitude of the developed labyrinth, the position of its canals and various chambers, is determined by its environment is based on seventeen experiments in which the ‘ear vesicle was loosened from its normal situation and placed in an abnormal attitude, and the specimen then allowed to continue in its development. At the end of a month examination showed that the labyrinth had become differentiated with varying degrees of completeness, and in each instance had developed in Aeeil rela- tion to the surrounding structures. Rotation in Two Directions. In eight of these experiments the ear vesicle was rotated 180° around both its vertical and transverse axes, so that it was turned face inward and upside down; or, inother words, its lateral or invaginated surface was toward the brain and its ventral border was where the dorsal border should be, the maxi- mum displacement. After this procedure the wounds healed within a few hours, and the larva were reared up to the fourth or fifth week, when they were killed and cut in serial sections. ‘Uhe labyrinths of five specimens were reconstructed. Before describ- ing them reference should be made to the normal condition of the labyrinth at this age. A reconstruction of a normal one with its adjacent structures 1s shown in Fig. tf. From the reconstruction of a normal specimen it can be seen that the three semicircular canals have individual characteristics by which they can be separately identifed; such as the Y-shaped union of the anterior and lateral canals, and the overlapping of the caudal end of the lateral canal by the posterior canal, and the junction of the posterior and anterior canals to form the crus com- mune. ‘The differentiation between utricle and saccule is not yet complete, but the part that is to become saccule is so labeled. From the caudal border of the saccule can be seen a small pocket budding out which constitutes the lagena or primitive cochlea. Directly median to the crus commune ts the endolymphatic append- age, consisting of a small duct leading from the main labyrinth chamber up between the labyrinth and brain to a rounded pouch, the saccus endolymphaticus. In their histology, as well as in 434 George L. Streeter, M.D. their general form, the various parts of the labyrinth exhibit at this time individuality. (See Fig. 4.) The ventro-median portion of the vestibular sac and the ampullar ends of the semicircular canals possess high columnar cells forming the neuro-epithelial maculz which are supplied with fibers from the acoustic ganglion, lying against the medial wall of the labyrinth. The endolymphatic sac has cuboidal cells, and the lagena has intensely staining col- umnar cells like those seen in the macular regions. The lagena is further characterized by its sharply rounded outline, and by the fact of its being compactly surrounded by ganglion cells and fibers, and cartilage forming cells. These features are so definite that Crus commune Ce. anterior Sac.endol., C.sc. posterior : - P \ /Mesen. \Dien. ‘ ‘ Gang. prootic. \Sacculus C.sc. lateralis Fig. 1 Reconstruction showing the form and relations of the membranous labyrinth of a normal tadpole (Rana pipiens) one month old. The labyrinth, adjacent ganglia and part of the brain were reconstructed after the Born method, and the remainder of the figure was drawn from a dissec- Lagena ’ tion of a tadpole of the same age. Enlarged 35 diameters. the various parts of the labyrinth can be recognized without difh- culty, even though they happen to be incomplete, or out of their normal relations. Now if one examines the models of the operated specimens, photographs of three of which are reproduced in Fig. 2, it is seen that the individuality of the semicircular canals can at once be identified. In model a, the canals are practically normal; in model b, the anterior canal is small, and the lateral canal consists only of a pouch which has not been pinched off from the main cavity; in model c, the posterior canal remains a simple pouch, while the Development of Amphibian Ear Vesicle 435 anterior and lateral canals arenormal. In considering the posture of the canals it is to be noted that the surrounding structures have been left out in Fig. 2, to avoid unnecessary duplication; the three models are all represented in the same relative position as that of the labyrinth in Fig. 1, 7. e., the cephalic end is on the right, the caudal end is on the left, the ventral surface is below, and the dor- sal surface is above. ‘Thus it will be seen that the lateral canals in all three models are in the same plane; likewise the posterior canals all form the dorso-caudal border of the labyrinth, and the anterior canals form the dorso-cephalic border. The fact that the anterior canal is small in model },‘ and the posterior canal is small in model c, gives rise to a false impression of a backward sac. endolymph. lagena } a b ( Fig. 2 Reconstructions showing the form and posture developed by three labyrinths one month old’ which while primitive ear vesicles were rotated from their normal position so as to lie face inward and upside down. The models are placed so that their planes are parallel with those in Fig. 1. Thus they present a lateral view with the cephalic end toward the right, caudal end toward the left, dorsal surface above, and ventral surface below. Enlarged 50 diameters. and forward tilting of the vesicle. ‘The saccule and lagena have the same position as in Fig. 1, and the lagena points caudally as it should do. ‘The endolymphatic appendage lies on the median side of the crus commune; 1n models } and c it is small, but the tip 4 This may be due to injury received at the time of operation. Such localized defects are frequently seen. They may involve any part of the labyrinth, and they vary greatly in the extent of the labyrinth wall affected. In one case the entire labyrinth was defective, with the exception of the endolymphatic appendage, which was normal in structure and position, and presented a curious appearance, being attached to the small irregular vesicle representing the labyrinth. Such localization of abnormal development is evidence of the high degree of specialization of the cells forming the primitive ear vesicle. 436 George L. Streeter, M.D. of it can be seen in model a. ‘The acoustic nerve and ganglion are attached to the median and ventral surfaces of the labyrinth, and the nerve connection with the brain appears to be normal. The conditions found in the three specimens pictured in Fig. 2 are typical of what is found in the other five specimens examined. They vary in the completeness of their differentiation, some of them consisting of only a vesicle with perhaps a single canal pouch, but in all cases the acoustic ganglion is present on the ventro medial surface, and the macular areas can be recognized. The lagena is present in seven out of eight cases. [he endolymphatic appendage developed in six out of eight cases. As regards posture, the rule is that the more perfectly the labyrinth is developed the more accurately its posture corresponds to the normal relations. But even in the most imperfect specimens when the endolym- phatic appendage appears it 1s on the medial surface, and the ten- dency to canal formation is always on the dorso-lateral surface, and the saccule and lagena appear on the ventral surface. This condition of course applies only to vesicles that have been im- planted in the acoustic region as was done in all the above cases. Rotation in One Direction. In four experiments the ear vesi- cle was rotated 180° around its vertical axis, 7.c, turned face inward. These specimens were then reared as in the preceding instance, and eventually cut in serial sections. A reconstruction model of one of them is reproduced in Fig. 3, and if it is compared with Fig. 1 it will be seen that although the vesicle was started in its development with invaginated side toward the brain yet the com- pleted labyrinth has the normal posture. A section of the same specimen is reproduced in Fig. 4, showing the labyrinth surrounded by developing cartilage. ‘The acoustic ganglion is connected in normal manner with the brain and sends peripheral fibers to the thickened floor of the saccule. “The endolymphatic sac is in its normal position, and the narrow duct can be seen connecting it with the main chamber of the labyrinth directly median to the crus commune. ‘The series through this specimen show that his- tologically it is practically perfect. Of the other three specimens one was almost equally perfect, another showed some abnor- malities in the formation of the canals and the lagena, and the Development of Amphibian Ear Vesicle 437 fourth was quite imperfect, consisting of only a large vesicle with a thickened epithelial floor connected by a few nerve cells and fibers with the brain. Fig. 3 Reconstruction of a tadpole labyrinth one month old, which when a primitive ear vesicle was rotated from the normal position 180° in one direction, so as to lie with invaginated side toward the brain. A section through the same labyrinth is shown in Fig. 4. Enlarged 55 diameters. Fig. 4 Section through the membranous labyrinth shown in Fig. 3. It shows that though originally turned face inward it has developed inthe normal attitude. _e, endolymphatic appendage; c.c., crus com- mune; sacc., saccule; ¢. Jat., lateral semicircular canal; gang. acust., acoustic ganglion. Enlarged 55 diameters. Transplanted Specimens. ‘he irregularity of form of the six specimens transplanted to the region between the eye and nostril, previously reported,’ is so great that they give no assistance in solv- Streeter ’06, /. c., p. 557. 438 George L. Streeter, M.D. ing the question of posture. However, in five cases, which will be presently described, where the ear vesicle was transplanted from the left side to the right side into the place made vacant by the removal of the right ear vesicle, in spite of the fact that these ear vesicles were implanted with haphazard attitude toward the adjacent structures, they nevertheless in each instance developed right-side up, and with the median surface toward the brain, as can be seen in Figs. 5 and 6. Mesenceph. Sac.endolymph : Dienceph. : Gang prockic. a Lagena Fig. 5 Reconstruction showing the form and relations developed by a left ear vesicle when trans- planted to the right side; it shows that under such circumstances the ear vesicle retains its left-sided characteristics, though it otherwise normally adapts itself to its new situation. A photograph of the same specimen is shown in Fig. 6, c. DETERMINATION OF THE DEXTRAL AND SINISTRAL CHARACTER OF -THE EAR” VESICLE ‘The question as to whether the right or left-sidedness of the ear labyrinth is controlled by the environment, or is determined by some intrinsic character of its own constituent cells, is answered in favor of the latter by the fact that if the left primitive ear vesicle, before the time of its complete closure, is transplanted to the oppo- site side of the embryo it retains its original left-sidedness. In five specimens, at the usual operating stage, the right ear vesicle was removed, and at the same time the left ear vesicle was uncovered and lifted from its natural bed and then placed into the pocket Development of Amphibian Ear Vesicle 439 from which the right vesicle had been taken and allowed to heal. In making the transplantation no effort was made to place the ear vesicles in any particular posture. After keeping the speci- mens alive for one month they were sectioned and from three of them reconstructions were made of the transplanted ear vesicle together with the adjacent structures. “The three labyrinths are shown in Fig. 6, and model c is again shown in Fig. 5, with the brain included. It will be seen that in developing they have assumed the normal attitude toward the brain. ‘The endolym- sac. endolymph. lagena a b c Fig. 6 Reconstructions of three labyrinths which while primitive ear vesicles were transplanted from the left to the right side. They are all represented in the same position as the models in Fig. 2. The model ¢ is the same that is shown in Fig. 5. Enlarged 50 diameters. phatic appendage and the median side of the labyrinth is toward the brain, the semicircular canals are toward the dorso-lateral sur- face, and the saccule and lagena are toward the ventral surface. But it can at once be recognized that the saccule and lagena point forward toward the eye, and that the anterior and posterior canals are in reversed positions. We thus have a complete mirror image of the right labyrinth, 7. ¢., a left labyrinth. Model a possesses three semicircular canals and is almost a normally formed left labyrinth. In model 4 the lateral canal consists of a pouch whose walls did not undergo the customary approximation and central absorption. In model c the posterior canal is not pinched off. In each of the models the lagena, saccule, and endolymphatic 440 George L. Streeter, M.D. appendage are typical, and there is establishment of normal appearing nerve and ganglion connections. EQUILIBRATION It was found in the experiments performed a year ago that removal of one or both ear vesicles, just after they are pinched off from the skin, produces in the tadpoles definite disturbances in the development of their power of equilibration. It was found that when a tadpole is deprived of but one ear vesicle he is by virtue of the remaining one able to develop practically normal swimming abilities; but when both ear vesicles are removed the results are more serious, and in that case the tadpole never de- velops any sense of equilibrium and is never able to swim. ‘The loss is not compensated for by any other organ and the animal lies helpless on the bottom of the dish. With one ear vesicle the tadpole swims practically in normal fashion, and with no ear vesi- cle he cannot swim at all. The fact that one ear vesicle is sufficient for the maintenance of equilibrium greatly simplifies the study of this mechanism; it means that one side can be immediately eliminated, and the prob- lem is reduced from a bilateral one to a unilateral one. A series of experiments at once suggested themselves, in which the ear vesi- cle of one side was to be removed, and then various operative procedures undertaken upon the ear vesicle of the opposite side, and the test of its consequent functional ability was to be the very decisive one of whether the animal could swim properly, or whether it could not swim at all. In the paper referred to there is described the experiment of transplanting the ear vesicle into a subdermal pocket in front of the eye. When this was done the transplanted ear vesicle continued in its development, and in some instances established a nerve-gan- glion connection with the forebrain; but such specimens never gave evidence of functional activity. ‘The failure to functionate was not unexpected, inasmuch as the connections established were at an abnormal situation, and furthermore the vesicles though having developed many essential features of the normal labyrinth Development of Amphibian Ear Vesicle 441 were still quite imperfect in the formation of the separate cham- bers and thesemicircular canals. Sothis year in carrying out the experiments described in the first part of the present paper the behavior of the specimens was eagerly watched, and the endeavor was made to determine the amount of alteration in position and defectiveness in form that is compatible with functional activity, involving the problem of the correlation between function and morphology. The observations made in the different experiments have been arranged and condensed as follows: a Left ear vesicle removed; right ear vesicle loosened from skin and rotated, in six specimens around the vertical axis 180° and in eight specimens around both the vertical and transverse axis 180°. As has already been shown these ear vesicles developed into labyrinths of varying degrees of perfection, some being com- pletely normal in form and having apparently normal ganglion and nerve connection with the brain wall. (See Figs. 2, 3 and 4.) The behavior of all the specimens was uniform, both where the ear vesicle was rotated in one plane and where rotated in two planes; at the end of a week after the operation, when with a nor- mally functionating labyrinth they should be able to swim freely and directly, they instead exhibit only irregular movements or spin around in spirals or circles. “‘Vheir incodrdinate movements con- tinue, and at the end of a month there is no improvement; 1.c, they behave exactly like specimens with both ear vesicles removed. Evidently ear vesicles thus treated do not perform their natural function. b Left ear vesicle removed; right ear vesicle fragmented by teasing between the points of two needles, the fragments left in ‘place. ‘Ten specimens were treated in this way, and were kept under observation four weeks, during which time they gave no evidence of any sense of equilibrium. c Right ear vesicle removed; left ear vesicle transplanted to the empty pocket on the right side. Five specimens were oper- ated upon and observed for one month, at the end of which time they were cut in serial sections, and it was found that the ear vesi- cles had developed into fairly complete labyrinths, but had main- tained the characteristics of a left-sided organ. (Figs. 5 and 6.) 442 George L. Streeter, M.D. Throughout the whole period of observation they had exhibited incoordinate movements, and at the end of that time they were unable to swim. ‘This and the two previous operations indicated that rotation of an ear vesicle, or transplanting it from one side to the other, or fragmenting it was not compatible with the devel- opment of its function, in spite of the fact that the ear vesicle pro- ceeded in its development and had become to all appearances almost a perfect labyrinth. In the next experiments less severe treatment was tried. d Left ear vesicle removed; mght ear vesicle uncovered and carefully lifted out and then immediately placed back in its orig- inal position, the effort being made to do a minimum amount of injury. Of six specimens all exhibited symptoms of the absence of all sense of equilibrium. In the experiments a, b, c and d there was the possibility of injury to both the nerve-ganglion connection and the ear vesicle. In the following experiments the effort was made to restrict the injury to one or the other. e Left ear vesicle removed; mght ear vesicle uncovered and a fragment cut from the cephalic portion of its wall, care being used not to otherwise disturb the vesicle. Eight such specimens were kept five weeks, and none of them developed any sense of equilib- rium, or were able to swim. j Left ear vesicle removed; right ear vesicle uncovered and a small piece cut from its caudal border, any further disturbance being avoided as ine. Eight specimens were operated upon, and after keeping them four weeks none of them could swim properly. g Left ear vesicle removed; longitudinal incision made through skin on right side just dorsal to ear vesicle, and needle passed down between the neural tube and ear vesicle and moved back- ward and forward so as to sever its nervous connection without otherwise disturbing the ear vesicle or loosening it from the skin. None of the four specimens studied swimmed properly, though one of them could swim somewhat, but was easily confused by any excitement and then made wild and ill directed movements. It was thought that the ear vesicles in these cases would escape injury; but examination of the specimens when cut in serial sections Development of Amphibian Ear Vesicle 443 showed that they were not perfectly normal. ‘This experiment might be repeated on a larger number of specimens and still greater care used in severing the nerve connection, in which case a perfect labyrinth could doubtless be obtained. h (Rana catesbiana) Left ear vesicle transplanted into an- other specimen, in a subdermal pocket in the region of the pro- otic ganglion between the right eye and ear vesicle, thus the host had three ear vesicles, two being on the right side. “Iwelve days after the operation three out of four specimens so treated exhib- ited incoodrdinate movements. Here we have to consider the crowding out of position of the normal right ear vesicle by the one transplanted near it. i Left ear vesicle removed; fine needle passed through the skin so as to make a small puncture in the right ear vesicle; on with- drawal of the needle the edges of the wound immediately close and there is no lossof cells from underneath or from the skin itself. Of four specimens at the end of one month three were able to swim, and this demonstrated the functional ability of an ear vesicle thus treated. 1 Left ear vesicle removed; small section of the covering skin removed so as to expose the right ear vesicle, but otherwise it is not disturbed and the nerve ganglion connection is left intact. Five specimens were kept under observation for one month, and four of them behaved throughout like those possessing one untouched normal ear vesicle; except for slight incoordination brought out by excitement they could swim properly. On bringing together the results of these experiments, it becomes immediately apparent that almost any operative procedure car- ried out on young larve in the region of the ear vesicle seriously interferes with the development of the function of that organ. It is possible to lift a skin flap and expose it, and to make a needle puncture in it without destroying its subsequent usefulness; but any operation involving a loss of part of its wall or disturbing its position and nerve-connection with the brain causes apparently complete loss of function. ‘The functional disturbance is out of al) proportion to the histological condition. There may be a laby- rinth that to all appearances is perfectly formed and that seems to 444 George L. Streeter, M.D. have a normal nerve ganglion connection with the brain at the proper place, and yet the specimen may not have given signs of any functional activity on the part of that organ. Spemann’ is doubtless mistaken in attributing the disturbance in equilibrium simply to the alteration in the planes of the canals. He reports some experiments in which at an early stage a skin flap was turned back, and the ear vesicle taken out and replaced in various positions; and in such specimens he observed faulty equi- librium, and on sectioning his material the vesicle seemed to lie in an abnormal position, and this he assumes to be the cause of the abnormal movements observed. On the one hand, wax plate reconstructions of misplaced ear vesicles show that in my cases they regain their proper position, and the canals eventually lie in their normal planes; the specimens nevertheless continue to make inco- ordinate movements. On the other hand, in those experiments where the normal position of the vesicle, as regards the planes of space, was undisturbed the results were equally serious. My own experiments suggest that the difficulty lies not so much with the end organ as with the central connections, and perhaps further experiments in that direction would furnish additional infor-. mation upon this subject. CONCLUSIONS The primitive ear vesicle of the tadpole may be loosened from its normal position and rotated in various directions, so that its axes lie in abnormal planes, and notwithstanding such interfer- ence it eventually develops into a labyrinth which 1s right side up and exhibits the normal relations to the brain and the surrounding structures. When transplanted to the opposite side of the body, if placed in the acoustic region, it likewise assumes a normal posture. Judging from these facts, the posture of the labyrinth is controlled by its environment. The “laterality” of the labyrinth is determined before the clo- sure of the ear vesicle. When the left ear vesicle is transplanted 6 Spemann, H.,’06: Ueber embryonale Transplantation. VerhandI. der Gesell. Deutscher Naturf. u. Aerzte. 78 Vers. Stuttgart. Development of Amphibian Ear Vesicle 445 to the right side it retains its characteristics as a left-sided organ, though it otherwise adapts itself to its new position in a normal mannet. | The functional disturbance, in experiments on the ear vesicle, is out of all proportion to the histological appearances; any opera- tion carried out in the acoustic region involving a loss of part of the wall of the ear vesicle, or disturbing its position, or nerve con- nection with the brain results in faulty equilibrium; absence of function was observed in cases where the labyrinth and its nerve connections seemed to have attained perfect histological develop- ment. COMPENSATORY MOTIONS AND THE SEMI- CIRCULAR CANALS BY BENJ. C. GRUENBERG With Two Ficures Meveactions Olthe ior to Movements OlrotatiOney. he & B 1 Aur pressure lymph, etc.) = 2 Inertia (of viscera, z IN Centrifugal action Ww 4 Friction of support ea image ee Vv > <— 2 e 6 “Spin” hOdPEL SOs > h4 4) a rotated to the right (clockwise) on the ordinary turntable; in those represented under II the eccentric arrangement was used and the results given are for a portion of the revolution only, since a continuation of the rotation beyond 180° is virtually equivalent to a reversal of the motion. (It is of course understood that the rotations in the reverse direction gave corresponding results but in the opposite sense.) The arrows indicate the directions in which the respective factors are supposed to act. Response Compensatory Motions 461 On comparing the arrows in the four columns it will be seen that whereas all in II seem to be related to the direction of the head turning, none in I are so except 5 and 6 (retinal and spin impres- sions). In experiments on the turntable (I) the factor 5 could be elim- inated in a variety of ways: By surrounding the vessel containing the frog with some opake material, or placing it in a tall opake cylinder; by covering the eyes with the opake non-irritant mix- ture already referred to, or with a pad of absorbent cotton mixed with vaseline and lampblack; by placing the source of light on the turntable with the animal. In all cases the turning of the head in response to rotation was the same as in the usual rotation as to direction; but frequently it was less in degree. In other words, while the displacement of the retinal image can and does set up the compensatory response, the eye is not the sole sense organ through which such movements can be initiated. This leaves the spin as the only other factor to be further con- sidered. According to the results indicated in columns A and B of the table, the spin is the only factor (of those considered) in addition to vision that can constantly set up the head turning. In the experiments on the eccentric (columns C and D) where the spin is already eliminated, the further elimination of sight results in a total loss of the response. The slightest amount of spin is sufficient to set up a perceptible amount of head turning; considerable displacement of the retinal image is required to bring about the same amount of response. It is possible to move the frog in a right line without the animal giving any response whatever; but if the movement is not smooth, that is, if there is vibration, or very slight turning in a horizontal plane, the head responds at once. ‘That the response to the spin is quicker and greater in amount is also certain; the two factors may be caused to operate in opposite directions in the following manner: A dish holding the frog on a horizontal plane and facing the observer, is swung about slowly by the observer at arm’s length. The head will be seen to turn in the same direction as the move- ment of translation; that is, in a direction opposite to what we 462 Benjy. C. Gruenberg should expect on the anthropomorphic view of the animal “seek- ing to keep the same vision in sight.”’ But the turning of the head is Opposite to the direction of the spin that the observer uncon- sciously imparts to the dish in moving his arm outstretched, which is thus in the radius of a horizontal rotation. That the perception of spin or rotation is located in the organs of the inner ear seems likely from the fact that the response is eliminated when the semicircular canals are destroyed or removed, or when the acoustic nerve is cut. “That the sensation concerned involves a factor of rotation or turning is indicated by the fact that rectilinear acceleration does not yield the same constant response. It may, therefore, be concluded that the compensatory move- ments of the frog’s head set up by rotation arise in response to two distinct sets of stimuli, visual and dynamic; that the response to the visual stimulus is relatively feebler and slower than that to the dynamic stimulus; that the organ for the perception of the dynamic factor is probably located in the internal ear; and that the dynamic perception involves a rotation or turning element in the stimulus, as distinguished from an acceleration or movement in a single direction. 6 SUMMARY 1 There is apparent contradiction between the various re- sponses of the frog to rotation on the turntable and any theory of mechanical stimulation of peripheral organ as the origin of the responses. 2 There is considerable contradiction among various experi- ments that have been made in connection with the relation of the semicircular canals and compensatory movements. 3. A reéxamination of the compensatory movements and of the conditions under which they arise shows the presence of a mechanical factor, the “spin,” the significance of which in this connection seems not to have been considered before. 4 From an examination of the results obtained by earlier observers, a repetition of some of their experiments, and new experiments made in the course of the study, the following con- clusions are drawn: ~ C ompensatory Motions 463 a ‘The compensatory movements of the frog’s head set up by rotation arise in response to two distinct sets of stimuli, visual and dynamic. b ‘The response to the visual stimulus 1s relatively feebler and slower than that to the dynamic stimulus. c The organ for the perception of the dynamic factor is prob- ably located in the internal ear. d The dynamic perception involves a rotation or turning element in the stimulus, as distinguished from an acceleration or movement in a simple direction. I wish to express my sincere thanks to Prof. T. H. Morgan for the helpful suggestions and encouragement that have made this study possible; and to Prof. A. P. Wills for his assistance in eluci- dating to me the mechanics of rotation. 7] REFERENCES AND BIBLIOGRAPHY 1828 FLourens, Marie JEAN PieRRE—(1) Expériences sur les canaux semicir- culaires de |’oreilles dans les oiseaux. Memoirs de |’acad. roy. des sci., 9:455-466. 1830. Read before the Academy, 11 Ag., 1828. (2) Expériences sur les canaux semicirculaires de l’oreilles dans les mammi- feres. Memoirs de l’acad. roy. des sci. 9:467-477. 1830. Read before the Academy, 13 Oc., 1828. 1842 FLourens, M. J. P.—Recherches expérimentales sur les propriétés du systeme nerveux. Paris. p. 438. 1846 PRrecHTL, JOHANN JoseF—Untersuchungen iiber den Flug der Vogel. Wien, Carl Gerold. _ p. 212. 1866 Manoyer, M.—Recherches expérimentales sur la locomotion chez les poissons. Ann.desSci.natur. Ser. V. 6:5-15. (Function of the fins and of the air bladder in locomotion and in equilibration; specific gray- ity relations.) 1868 Lrypic, F—Ueber Organe eines sechsten Sinnes. Dresden. 1869 Goxtz, Fr.—Beitrage zur Lehre von den Functionen der Nervencentren des Frosches. Berlin, Hirschwald. pp. 130. 1870 Gortz, FriepricH—Ueber die physiologische Bedeutung des Ohrlaby- rinths. Pfliiger’s Archiv. 3:172—-193. SCHULZE, Franz E1tHARD—Ueber Sinnesorgane der Seitenlinie bei Fischen und Amphibien. Arch. Mikr. Anat. 6:62-88. Taf.4-6 (The sense 464 1873 1874 1875 1877 1878 Beny. C. Gruenberg organs of the lateral line serve for the perception of water waves of lower frequency than those perceived by the ears.) Cyon, E. von—Cours de physiologie. St. Petersburg. vol. ii. Breuer, J.—Ueber die Funktion der Bogengange des Ohrlabyrinths. Med. Jahrb. pp. 72-124. Brown, ALEXANDER CRuM—(r1) Preliminary Note on the Sense of Rotation and the Function of the Semicircular Canals of the Internal Ear. Proc. Roy. Soc. Edin. 8:255. 19 Ja., 1874. (2) On the Semicircular Canals of the Internal Ear. Proc. Roy. Soc. Edin. 8:370-371. (3) On the Sense of Rotation and the Anatomy and Physiology of the Semi- circular Canals of the Internal Ear. Jl. Anat. and Phys. 8:327-331. Cyon, Ernst von—Ueber die Function der halbcirkelférmigen Canile. Pfliiger’s Archiv. 8:306—326. Breuer, |.—Beitrage zur Lehre vom Statischen Sinne (Gleichgewichtorgan, Vestibulapparat des Ohrlabyrinths.) Med. Jahrb. pp. 87-156. Macu, Ernst—Grundlinien der Lehre von den Bewegungsempfindungen. Leipzig. Cyon, E. von—(1) Methodik der physiologischen Experimente. St. Petersburg. pp. 540-547. (2) Rapports physiologique entre le nerf acoustique. Compt. Rend. Acad. Sci. 82:856. Cyon, E. von—Les organs periphériques du sens de l’espace. Compt. Rend. 85:1284. Tomaszewicz, ANNA—Beitrage zur Physiologie des Ohrlabyrinths. Zurich. Inaug. Diss. Brown, ALEXANDER CrumM—Cyon’s Researches on the Ear. Nature. 18 :633-635; 657-659. (Historical summary and abstract of Cyon, 1878). Cyon, E. von—Recherches expérimentales sur les fonctions des canaux semicirculaires et sur leur role dans Ja formation de la notion de l’espace. Thesis. HeEnseENn, V.—Physiologie des Gehérs. Hermann’s Handbuch der Physiol- ogie. III (2). SpaMER, C.—Experimenteller und kritischer Beitrag zur Physiologie der halbkreisformigen Kanale. Pfliiger’s Archiv. 21:479-590. Bacinsky, BENNo—(1) Ueber die Schwindelerscheinungen nach Ohrver- letzungen. Monatsb. Berl. Acad. (2) Die Funktion der Bogengange des Ohrlabyrinthes. Biol. Zentralb. 1:438-446. 1882 1883 1884 1885 1886 1887 1888 1889 1890 Compensatory Motions 465 KiessetpacH—Zur Funktion der halbzirkelformigen Kanale. Arch. fur Ohrenheilkunde, 18. Sewatt, H.—Experiments upon the Ears of Fishes with Reference to the Function of Equilibrium. Jl. Phys. 4:339-3409. Buppe, E.—Ueber metakinetische Scheinbewegungen und tiber die Wahr- nehmung der Bewegung. Arch. Phys. 126-152. Bacinsky, BeNNo—Zur Physiologie der Bogengange. Arch. Anat. u. Phys. 253-266. STEINER, J.—Untersuchungen tiber die Physiologie des Froschhirns. Braunschweig, Vieweg. pp. 143, 32 cuts. Macu, E.—Beitrage zur Analyse der Empfindungen. _ Jena. STEINER, Is.—Ueber das Central nervensystem des Haifisches und des Amphioxus, und tiber die halbcirkelformigen Canale des Haifisches. Sitzungsb. Berl. Akad. 20 My., 1886. pp. 495-499. Detace, YvEs—Sur une fonction nouvelle des otocystes comme organes d’orientation locomotrice. Arch. Zool. Exper. et Gen. Ser. Il. 5:1- 26. (Function of otocysts in orientation of invertebrates. ) ENGELMANN, TH. W.—Ueber die Funktion der Otolithen. Zool. Anzeiger. 439-444. ScHAFER, Kart—Ueber die Wahrnehmung einiger passiven Bewegungen durch den Muskelsinn. Pfliiger’s Archiv. 41:566-640 (30 wood cuts). ScuraperR, Max FE. G.—Zur Physiologie des Froschhirns. Pfliiger’s Archiv. 41:75-9o. Cyon, E. von—Gesammelte physiologische Arbeiten. Berlin, August Hirschwald. p. 338. DELAGEs, YVES AND AuBERT—Physiologische Studien tiber die Orientirung. Tubingen. Logr, Jacgues—Die Orientierung der Tiere gegen die Schwerkraft der Erde. Sitzb. phys-med. Gesel., Wurzburg, Jahrg. 1888, 5-10. Breuer. J.—Neue Versuche an den Ohrbogengingen. Pfliiger’s Archiv. 44.:135-152. Brown, ALEXANDER Crum—Our Sensation of Motion. Armitstead lec- ture, in Dundee. Nature, 40:449-453. (The mechanics of possible stimulation through motion). STEINER, J.—Der Meniere’sche Schwindel und die halbzirkelformigen Canale. Deutsche med. Wochensch. No. 47. 958-960. Leipzig. Ewatp, R.—Physiologisché Untersuchungen tiber das Endorgan des Ner- vus Octavus. Berl. Klin. Wochensch., Nr. 32. Reprinted, Wiesbaden, 1892. Marey, E. J.—Le vol des oiseaux. Paris, G. Mason. 466 1891 1892 1893 1894 1895 1896 _ Benj. C. Gruenberg Breuer, J.—Ueber die Funktion der Otolithen Apparate. Pfliiger’s Archiv. 50:195-306, Pl. ITI-V. Krerpi, ALois—Beitrage zur Physiologie des Ohrlabyrinths auf Grund yon Versuchen an Taubstummen. Pfliiger’s Archiv. 51:119—150, PI. VII. Lors, JacguEs—(1) Ueber Geotropismus bei Tieren. Pfliiger’s Archiv. 49:175. (Studies in General Physiology, Chicago, 1905, pp. 176-190.) (2) Ueber den Antheil des Hohrnerven an den nach Gehirnverletzung auf- tretenden Zwangsbewegungen, Zwangslagen und assoziirten Stellungs- anderungen der Bulbi und Extremitaten. Pfliiger’s Archiv. 50:66-83. Ayers, Howarp—A Contribution to the Morphology of the Vertebrate Ear with a Reconsideration of its Functions. Jl. Morph. 6:1-360. (With bibliography of 295 numbers.) (Author’s abstract, as lecture, in Wood’s Hole Biol. Lect. for 1890.) Ewa.p, R.—Bedeutung des Ohres fiir normale Muskelkontraktionen. Zentralbl. fiir Phys. 5:4-6. Grrarb, H.—Recherches sur la fonction des canaux semicirculaires de Voreille interne chez le grenouille. Arch. de Physiol. 24:353-365. KreipL, ALors—(1) Zur physiologischen Bedeutung des Ohrlabyrinthes. Neurol. Centralbl. 11:222-223. (2) Weitere Beitrage zur Physiologie des Ohrlabyrinthes. Sitzungsb. Wien Akad. Wiss. Bd. 1o1, Abt. III, 469-480. 1893. 102 (I1):149-174. Ler, FrepErick S.—A Study of the Sense of Equilibrium in Fishes. Jl. Phys: 45:311- Breer, THEODORE—Die Accomodation des Fischauges. Pfliiger’s Archiv. 58:523-650; 35 fig. 3 pl. (With bibliography of 65 numbers.) Betue, ALBRECHT—Ueber die Erhaltung des Gleichgewichts. Biol. Zen- tralb. 14:95-114, 563-582. Ler, F. S—A Study of the Sense of Equilibrium in Fishes. Jl. Phys. 17:192—210. SCHAFFER, Kart L.—Funktion und Funktionsentwickelung der Bogen- gange. Zeistchr. f. Psych. 7:1-g. (Whirling produces no effect upon tadpoles; but after the canals become developed, they show signs of dizziness. ) STREHL, H.—Beitrage zur Physiologie des inneren Ohres. Pfliiger’s Archiv. 61:205-234. WarreEN, H. C.—Sensations of Rotation. Psych. Rev. 2:273-276. BECHTEREW, W. von—Ueber die Empfindungen welche mittels der soge- nannten Gleichgewichtsorgane wahrgenommen werden, und uber die Bedeutung dieser Empfindungen in Bezug auf die Entwickelung un- serer Raumvorstellungen. Arch. f. Phys. 105-141. an 1896 1897 1898 1899 Igol 1093 1904 S/S) 1906 a7 Compensatory Motions 67 j 497 Ciark, GayLorp P.—On the Relation of the Otocysts to Equilibrium Phe- nomena in Gelasimus pugilator and Platyonichus ocellatus. J]. Phys. 19 :327-343- Cyon, E. von—Bogengange und Raumsinn. Arch. f. Phys. 29-111. Ler, F. S—The Function of the Ear and the Lateral Line in Fishes. Am. Jl. Phys. 1:128-144. (Reviewed in Zool. Zentrabl. 6:409-411.) Devirz, ].—Ueber den Rheotropismus bei Tieren. Arch. f. Phys. (Suppl.) 231-244. WHEELER, WiLL1AM Morton—Anemotropisms and Other Tropisms in Insects. Arch. fiir Entwickelungsmechanik. 8:373-381. Lyon, E. P.—A Contribution to the Comparative Physiology of Compen- satory Motions. Am. Jl. Phys. 3:86-114. Prentiss, C. W.—The Otocyst of Decapod Crustacea; its Structure, De- velopment and Functions. Bull. Mus. Comp. Zool. Harvard. 36. Tu iiperc, T.—Das Labyrinth der Fische, ein Organ zur Empfindung der Wasserbewegung. Beihang: K. Svenska Vet-Akad. Handlingar, Stock- holm, vol. 28, no. 15. 25 pp. Lyon, E. P—On Rheotropism in Fishes, I. Am. Jl. Phys. 12:149-161. YerKEs, Ropert M.—Inhibition and Reinforcement of Reaction in the Frog. Jl. Comp. Morph. and Psych. 14:124. Parker, G. H.—The Skin, Lateral-line Organs and Earas Organs of Equili- bration. Proc. Am. Zool. Soc. Science, 20. Hap ey, Puitie B.—The Relation of Optical Stimuli to Rheotaxis in the American Lobster. Am. Jl. Phys. 17:326-343. STREETER, GEORGE L.—Experiments on the Developing Ear-vesicle of the Tadpole with Relation to Equilibration. Jl. Exp. Zod]. 3:543-558. Logs, JAcgues—Ueber die Summation heliotropischer und geotropischer Wirkungen bei den auf der Drehscheibe ausgelésten compensatorischen Kopfbewegungen. Pfliiger’s Archiv. 116:368-374. A STUDY OF THE SPERMATOGENESIS OF TWENTY- TWO SPECIES OF THE MEMBRACIDA, JASSIDZ, CERCOPIDA AND FULGORID, WITH ESPECIAL REb ERENCE FO. THE BEHAVIOR OF THE ODD CHROMOSOME! ALICE M. BORING With Nine PLates IKE HE CINCO. Cone engtin OHO e Ded Ueber o CASO DOUG OSE Oth ¢- ODOR An REA GD Maar 6 Scorn one 470 PAN SeoIGe EME VME Wavelets, -etev tsi pate aleieis.ctoteySreyebe sl ae seal eleieis'aiaie, sreceieue ates Glee ahelesatahe! cuss cae sayelepe «Ghee 470 Vastra ATTCNETAC LLVO Cl Sips eyalewse Peters serchel 2 orevayelcecreremetore nin atgttts ciel si ale,cheie were oyniayd rele sale adele) sl fieye setae sl 478 MBE IIS ata RENO LSM Ph sha eg, 5 os oss guava «Porat eva a eid oVey0y 0) wna aeas So ort afte asertovdas AG aisiyudbncn nts yew Sunes gee ebedtte 480 Uy esrra is ticteel dae fears yoonc' ctsislsysiensunvopstniesevsser eaters Hie ove soskereionstere ays clave! wis ierscbie) sid aserduertave an wtaysrerermyeuere we eleva 480 ! DUNT SEE ee elena ho.c Otis G0 Dice IOC ere ERO ID cle fo RIO TO cam rie Raia er oic ir acne CRO 481 MVE Ze ARAL CUA tar cvsicreveteras cic) stove eine tie cictters (ore taviovate Sytnolens ete wile clesete eases cieecl sh ccemlord sles eines alae 486 Geresaptarimii arc a iap.tatatereyai tet sususts wuttcse fcr, aia ele alele’s ajuisin whe, o dlare cm iotw one sais tagate al alaieheveletecdorepsrae Gs 487 Gexesamonbal uss tee clave ctcteesicte sec aie.s, conve ote rein srs ae, ciaveie a smrauictete oA caeep einiererete cust nes 488 WBer SANT CELO Sip alctes Specs) rotevayeiendl stacey t loisr sl etter dais, miavara cl tei So chore aca uey tue ca, a) 206 o Byer apeyecovevave abe reket 489 JN A/TUME) CARN GAROO0 doles ae PUM TOds PAONG dO Dn DOpon oT Op oOUEE oan WOO IoD S OcdEn MoRuC 489 Gamipylenchiarcurviata ce ect qostue 2 a leneuctett 496 ANSI eROyplNvORel Gevehaetey hl ELatSand Supe obe ape een ato Ubomeoten orn soca ocosouMb oun Se ueinoe ce 497 PAP UNO NOL ApA MOLALLA tare specatats, cia) ohctayelerelcrakeledsctars, 25) ciSgya. eas shaven tee kek are onerovesehereiece aust Manele eisie sere 498 Std peor ceomreteyetetereree tots to nieveey oe eeers Va oie header ales Sealn'a Ha lapesevans Sieber cae oma yermien 6 A1aES Pera rovene stares 498 Hoscilopteraseptemthionalisie 4. spereratvere x ersweveieie ats wists ciate ais) ainys a eIsIS eye eyors witrele aoe roe ere olane 499 JP esiBid BY ohSUrA oy AUELANOISES 3 See oS Ole On orn aes OO OIL Ree OR EIR err Oe RO Chiesa bot in ome 500 PILL OSG Pag Miva Cul ta meg teria eters telnet iahe creteie axaza saw Bice ele ces cette wae ieee eeegs ies Sol Poeciloptenauivatia tae tn) aeines ie sera a voarsesie) crise ivlevs clare 9 eists ua cre etoeletere siaueistela Aeeteraie 501 ‘Ulnaorancail Cone Glaritasane otcdoboocmes cad hoadeae de copsdtnoe cions comnts < ae aaRoncee ner 502 'SXOTAIREIN IGS 5 5.005 SRS ORE EIDE Coe Ogee ola a Ra er eRe Oe ales tee er a eee eer Peete 506 [SHS TO RAPLN. 5 py oa OES OLED GoeLe GPL Acie CONOR Cae Sette ons oio Morte te cist car RCI aera antame 509 LD ERG palm OF BENGE AS Coane nee mee oapaciod ten SoG bon olson ato ree ecc COME eae paramo 513 1A dissertation presented to the Faculty of Bryn Mawr College for the degree of Doctor of Philosophy. Tre JourNat or ExPERIMENTAL ZOOLOGY, VOL. IV, NO. 4. 470 Alice M. Boring INTRODUCTION The purpose of this investigation is to extend, to some families of the Hemiptera Homoptera, the studies of McClung, Stevens, Wilson and others on the relation of the accessory or odd chromo- some to sex determination. Except for the aphids, which have been extensively worked out by Stevens (’o5a, ’o6a), Cicada tibi- cens (Wilcox ’95) and Aphrophora quadrangularis (Stevens ’o6b) are the only species of this group whose spermatogenesis has been previously described. This study covers eight species of the Membracidz, six of the Jasside, four of the Cercopide and four of the Fulgoridz. My work was begun at the suggestion of Dr. N. M. Stevens at Woods Hole in the summer of 1905, continued under Prof. E. G. Conklin, at the University of Pennsylvania, in the year 1905-06, and completed under Dr. Stevens, at Bryn Mawr College, in the year 1906-07. To both Dr. Stevens and Professor Conklin I wish to express my appreciation of their valuable suggestions and constant help and inspiration. I wish also to thank Dr. Herbert Osborn of Columbus, Ohio; Mr. E. P. Van Duzee, of Buffalo; Mr. H. C. Barber, of New York City, and Dr. H. Skinner, of Philadelphia, for the identification of material. HISTORICAL REVIEW Most of the work on the spermatogenesis of the tracheate arthro- pods has been done since 1890. Such studies as those of Bitschli (71), La Valette St. George (’85), Platner (’86), Verson (’8Q), and Sabatier (’85) were concerned only with the formation of the spermatozoa, the arrangement of the cells of the testis into cysts, and the general mechanics of karyokinesis. The work of van Beneden (84), Boveri (’87) and O. Hertwig (’90) on Ascaris, and Mark (’81) on Limax, turned the interest in the study of the sex cells to the chromosomes, while Weismann’s daring hypothesis (’87) as to equational and reducing divisions added to the interest. By 1899, practically all investigations on spermatogenesis centered around the chromosomes in the spermatocyte divisions, and in S permatogenesis 471 that year we find the first statement that one chromosome behaves differently from the others (Henking ’go). Unfortunately there 1s the greatest confusion in the results for the next decade; but since Montgomery’s suggestion (ola) that synapsis means the conjuga- tion of homologous maternal and paternal chromosomes, and its confirmation by Sutton’s work on Brachystola (’00, 02, ’03), there has been greater accord. As a consequence of this, certain funda- mental theories are coming to reston a firm foundation. ‘The chro- mosomes are shown to keep their individuality from one cell gener- ation to another. ‘The real reduction in number is proved to be brought about by the joining of each paternal to a corresponding maternal chromosome in synapsis. It is found to make no differ- ence whether the reducing or equational division comes first, but the distinction between these two divisions is constant, the one being the separating of the individual spermatogonial chromo- somes, the other a simple splitting of these univalent chromosomes. In addition to this, recent work indicates that there is usually present throughout the Tracheata an odd chromosome in the spermatogonia, which behaves differently from the other chromo- somes throughout its history. Still later work seems to establish the fact that this chromosome has no paternal mate, does not join any other chromosome in synapsis, divides in only one sperma- tocyte division, and enters only half of the spermatozoa. In some forms, a small chromosome is present as the paternal mate of this odd chromosome, but dimorphism of the spermatozoa results in either case. The following review takes up the different observations on the Tracheata since 1890, and attempts to show how each helps to establish, or differs from, the above mentioned theories. Arachnida Wallace (05) finds an even number of spermatogonial chromo- somes, 40, two of these being larger than the others and different in behavior. ‘They are condensed in the spermatogonial rest stage, and take an eccentric position in the equatorial plate. They remain separate from each other in the spermatocyte growth 472 Alice M. Boring period and do not divide in either spermatocyte division, as the other 19 chromosomes do, thus appearing in only one quarter of the spermatozoa. Wallace concludes that all the spermatozoa degenerate except those with the two odd chromosomes. Montgomery in Lycosa (’05) finds an even number of chromo- somes in the spermatogonia. ‘lwo of these he calls heterochromo- somes, although the only characteristic that justifies this name is that they remain condensed in the growth period. ‘They conju- gate like the other chromosomes and divide in both divisions, all of the spermatozoa receiving one-fourth of the heterochromosome tetrad. The results of neither of these investigators agree with the more recent work on the odd chromosome in spiders and other forms. If, as Wallace states, no spermatozoa develop except those con- taining the two odd chromosomes and the nineteen ordinary chromosomes, the eggs must all contain only 19 chromosomes, as the spermatogonial number is 40. Suppose each egg to have 19 chromosomes; fertilization by a spermatozoOn with 19 +2 chromo- somes would give all the offspring 38+2 (19 +2 in the reduced number), whether male or female; but according to Wallace’s con- tention, the egg can have only 19; therefore it is impossible that all the spermatozoa, except those with the two odd chromosomes, degenerate. According to Montgomery, the heterochromosome in the spermatocyte is bivalent and divides in both divisions. Berry’s work (06) brings the odd chromosomein the spider into line with the odd chromosomes in other forms; it is a single chromo- some in the spermatogonia, and divides in only the second divi- sion of the spermatocytes, resulting in dimorphism of the sper- matozoa. M yriapoda Blackman (’o5a, ’o5b) finds in Scolopendra heros and S. sub- spinipes an uneven number of spermatogonial chromosomes. Synapsis takes place in the late anaphase of the last spermatogonial division, all of the chromosomes uniting in pairs except the odd one. The odd chromosome divides only in the second spermato- cyte division. ‘The peculiarity here is that the other chromosomes S permatogenesis 473 seem to undergo their reducing division when the odd chromosome is dividing equationally, but this is only a further mark of the indi- viduality of the chromosomes, and does not furnish any evidence against Montgomery’s theory of synapsis. Medes (’05) finds a similar condition in Scutigera forceps. Orthoptera Neither vom Rath (’g1, ’92) nor Wilcox (’95) noticed an odd chromosome in Gryllotalpa or Caloptenus, although both mention a nucleolus in the spermatocyte growth period which may be the same structure. hey both insist that there are two reducing divisions; that is, two divisions that separate whole chromosomes from each other. ‘This is probably due to a confusion in the use of the word chromosome. If we use the terminology suggested by McClung (oo), univalent chromosome in the spermatogonium, bivalent chromosome in the spermatocyte, and chromatid for each unit of the tetrad, the discrepancies inthe work of vom Rath and Wilcox are cleared up. Vom Rath finds 12 spermatogonial chro- mosomes. In the growth period, the spireme splits into six rods, each of which forms a tetrad, or divides into four “chromosomes,” as he expresses it. As he calls each chromatid a chromosome, he considers that he has two divisions which separate chromosomes from chromosomes; and therefore must be reducing; while in terms of the original spermatogonial chromosomes, one division is reducing and one equational. Wilcox falls into the same difh- culty; he finds 12 spermatogonial chromosomes, and then the spireme divides into 24 “chromosomes,” which form 6 tetrads. He had, in reality, 24 chromatids, and only one reducing divi- sion. McClung (00, ’02a) has described the odd chromosome in the Acrididz and Locustidz. He worked on a number of forms and obtained uniform results. In the Orthoptera, this chromosome can be traced back into the spermatogonial rest stages. It divides only in the first spermatocyte division, giving dimorphism of the spermatozoa. In 1901, McClung suggested the theory which has since that time received substantial corroboration, that the dimor- 474. Alice M. Boring phism of the spermatozoa corresponds to the dimorphism of sex. McClung considers that the longitudinal division always precedes the reducing division, and thinks that this is important on account of the failure of the second polar body to be extruded in parthe- nogenetic eggs; but the work in the other groups of insects shows that the reducing division probably comes first as often as the equational. Sutton’s careful work (’oo, ’o2) on Brachystola magna offers convincing evidence for the individuality of the chromosomes. Each pair of spermatogonial chromosomes becomes enclosed in a separate compartment of the nucleus, while the odd chromosome ts in a vesicle shut completely off from the others. He suggests the application of Montgomery’s theory of the union of maternal and paternal chromosomes in synapsis to Mendelian inherit- ance. The observations of de Sinéty (’o1) on the odd chromosome in one of the Acridida and in several Phasmidz are entifely in accord with those of McClung; this chromosome divides in only one sper- matocyte division, producing dimorphic spermatozoa. In one of the phasms, he finds a chromosome complex similar to that de- scribed later by McClung (’o5) for Hesperotettrix, where the odd chromosome attaches itself to one end of a tetrad, forming a hexad which divides along the transverse axis of the tetrad, thus sending the odd chromosome and two chromatids of the tetrad to one cell, and two to the other. Unfortunately de Sinéty interprets both of the spermatocyte divisions as longitudinal, but on this point he is in the minority among the workers on Orthoptera. Baumgartner (’04), in Gryllus domesticus, finds the odd chromo- some in a separate vesicle as Sutton did for Brachystola, but he finds it dividing in the second division instead of the first. Stevens (05a) in Stenopelmatus and Blatella germanica, and Otte (’06), in Locusta viridissima, find that the odd chromosome divides in the second division instead of the first. Evidently there is no fixed rule as to where the odd chromosome shall divide. Voinov (’03), Montgomery (’05) and Zweiger (’06) all hold a different view as to the valence of the orthopteran odd chromo- some; but as each has studied only one species of the order, while : S permatogenests 475 the work of McClung, de Sinéty, Sutton, Baumgartner and Stevens covers numerous species in several families, we have a right to question the views of these other three observers. All three hold that the heterochromosome which they describe is formed from two spermatogonial chromosomes and divides in both spermato- cyte divisions. Moore and Robinson (’05) claim that the odd chromosome in Periplaneta americana is only a plasmosome which dissolves before each division and is reconstructed after it. Odonata The paper of McGill (04) on Anax junius seems to show the same confusion which Wilson has discovered in Paulmier’s work on Anasa tristis. McGill finds an even number of chromosomes in the spermatogonia, two of them small. ‘These she identifies with the chromatin nucleolus of the rest stage and the odd chromo- some, which divides in the first division and not inthesecond. If it could be shown that there are only 27 chromosomes in the sper- matogonial plate, and that the odd chromosome is one of the larger ones, this form would fall into line with other work. Lepidoptera The early investigators in this field, Platner (’86) and Verson (94) paid no attention to the chromosomes. I have not been able to read Toyama’s papers, but the references to them by McClung indicate that the work is not very satisfactory. Stevens (o6b) gives a few figures for two species. There are two con- densed bodies throughout the growth period, which fuse in pro- phase like the m-chromosomes in Alydus (Wilson, ’o5c), and this body divides in both divisions like the equal “idiochromosomes’’ of Nezara. Coleoptera The only work on the Coleoptera which deals with the hetero- chromosomes is that of Stevens (’osb and ’o6b) and of Nowlin (06). Some of the beetles have an odd chromosome and others have an unequal pair in which the large member of the pair is the 476 Alice M. Boring maternal homologue of the odd chromosome, and the small mem- ber is the paternal mate which is lacking with the odd chromo- some. In the Coleoptera, the reducing division comes first, the equational second. In this order of insects there is substantial proof of McClung’s sex determination theory, as the odgonial equatorial plates have been shown to have the large chromosome, while the spermatogonial plates have the small one, and there is the same difference between the somatic plates of the males and females. ‘The theoretical bearing of these facts will be discussed later. Hemuptera ‘The chromosomes in this group are so large and few in number that they have attracted many workers, but in spite of this fact, there have been greater discrepancies than in almost any other group. Henking (’go0) in working on Pyrrhocoris apterus, was the first to notice that in one spermatocyte division, one chromo- some does not divide, thus causing a dimorphism of spermatozoa. He counted 24 chromosomes in the spermatogonia, and thought that this odd chromosome had the same valence as the others. He observed a large darkly-staining nucleolus in the growth period, although he did not associate a chromatic nature with it, or con- nect it with the odd chromosome of the spermatocyte mitoses. He formulated no theory to account for the dimorphism of the spermatozoa. Wilcox (’95) records that there are 12 spermatogonial chromo- somes in Cicada tibicens, and 24 spheroidal bodies in the sperma- tocytes, instead of a reduced number, results similar to those on Caloptenus femur-rubrum. In Anasa tristis, Paulmier (’99) describes two small sperma- togonial chromosomes, which form first the chromatin nucleolus in the growth period, then a tetrad which divides in the first sperma- tocyte division, and not in the second. Because this chromosome is small and appears in only part of the spermatozoa, he regards it as degenerating chromatin. Wilson (’o5c), working over the same field, finds that Paulmier has confused two bodies, inas- much as the two small chromosomes form a tetrad and divide in le, = y ae ae S permatogenesis NG, both divisions, while the odd chromosome, which divides only in the first division, is the chromatin nucleolus of the rest stage and one of the large chromosomes of the spermatogonia. He main- tains that Paulmier made a mistake also in the spermatogonial number, which is always odd. Foot and Strobell (’07), by the use of smear preparations and photo-micrographs, have attempted to show that Wilson is in error in his observations on the spermato- genesis of Anasa. ‘They find that the odd chromosome acts essen- tially like any other chromosome, is made up of two spermatogonial chromosomes and divides in both spermatocyte divisions, its only peculiarities being that it does not appear as a tetrad in prophase and occasionally divides later than the other chromosomes in metaphase. They attempt to show that the chromatin nucleolus of the rest stage is not a chromosome, but dissolves before meta- phase like a plasmosome. Wilson (’07) has carefully gone over his preparations and still thinks that his former conclusions are correct. “There is need of more work with smear preparations to test their reliability. Gross (’04), in his work on Syromastes, apparently confuses the m-chromosomes with the odd chromosome much as Paulmier did. In Pyrrhocoris apterus (’06) he finds the odd chromosome bival- ent but dividing in only one spermatocyte division. —- Montgomery (ota) calls the odd chromosomes of the Hemip- tera “chromatin nucleoli’? and considers that they may vary in number and valence. He explains them as chromosomes on the way to disappearance during progressive evolution. His results show many discrepancies which have since been explained by Wilson (’os5b and ’o05c). Wilson groups the Heteroptera into three classes, those with an unequal pair of heterochromosomes, those with an odd chromo- some and m-chromosomes, those with an equal pair of hetero- chromosomes. In the first class, the chromosome number in the second spermatocyte is one less than in the first spermatocyte. This is due to the fact that the conjugation of the unequal pair does not take place until after the first spermatocyte division. ‘This is the most direct evidence yet found for Montgomery’s synapsis hypothesis, for the small chromosome can be proved to be paternal, 478 Alice M. Boring and the large one, maternal. In the second class, the odd chromo- some is homologous with the large maternal element in the unequal pair. [he m-chromosomes are a pair, whose synapsis 1s delayed until just. before the first spermatocyte division. ‘The third class includes forms where there is neither an unequal pair, nor an odd chromosome, and therefore no visible dimorphism of the sperma- tozoa, but the fact that the equal heterochromosomes do not con- jugate until after the first spermatocyte division, relates this class to the first class, and suggests that there may be a masked dimor- phism, the equal heterochromosomes representing different char- acters, possibly, as truly as the unequal heterochromosomes where there is a visible dimorphism. Wilson cites a great deal of evi- dence for the individuality of the chromosomes, finding the same size relations between pairs of spermatogonial chromosomes as there are between single chromosomes in the spermatocytes. He elaborates McClung’s sex determination theory, brings forward much evidence for the dimorphism of the spermatozoa, and shows that there is a corresponding dimorphism in the somatic equatorial plates of the male and female of several species of the Hemiptera heteroptera. MATERIAL AND METHODS My material was collected at Woods Hole in the summer of 1905, at Cold Spring Harbor in the summer of 1906, and at Bryn Mawr in the fall of 1906. ‘The insects were caught in the usual sweep net, and the testes dissected out as soon as possible. Each testis consists of a group of several follicles, each attached by a separate duct to the vas deferens. The testes from the larvz just ready for metamorphosis, and from the adults soon after meta- morphosis, in most cases give all stages from the spermatogonia to the mature spermatozoa. Before putting up material of any species, Schneider’s aceto- carmine proved to be a quick and efficient reagent for determining whether the testes contained all the important stages. This fixes and stains the material at the same time. ‘The testis is put on a slide in a drop of the stain, and the cells separated by press- ing down the coverglass. The preparation is made air-tight S permatogenesis 479 with vaseline, and in a few minutes, the chromatin is stained a deep carmine. The entire spermatogenesis might be worked out in such preparations, the only disadvantage being that the achro- matic structures are not well fixed, and the preparations are not permanent. Camera drawings made from the aceto-carmine material, compared with those from sections of material fixed in the usual reagents, show the chromosomes in the former much larger in size. (Compare Fig. 198 with Fig. 205, and 201 with 207.) This difference is largely due to Shemales in the usual fixing fluids and alcohols. Pie relative sizes and positions of the structures are the same in both kinds of preparations. If the material showed the right stages, it was put up in various fixing fluids: Gilson’s mercuro-nitric, Flemming’s strong chromo- aceto-osmic, Hermann’s platino-aceto-osmic, and Carnoy’s acetic alcohol with sublimate. ‘The dissecting was usually done in the hxing fluid, but the small quantity of Petecial that was dissected in physiological salt solution and immediately transferred to the fixing fluid, showed just as good fixation, as is shown by the clear outlines of all the cell structures. A few cases of poor fixation were apparently due to the long time the insects were kept in captivity, as was sometimes necessary when the material was col- lected several miles from the laboratory , and immediate dissection was impossible. Gilson’s mercuro-nitric was the fixative used most frequently, because it gives excellent fixation of the chromatin and is a very convenient fluid to use, but nearly all material was also put in one or both of the osmic mixtures, as these give better fixation of the achromatic structures. “The Gilson was used for two to six hours, the Flemming and Hermann for twelve to twenty- four hours, followed by the same length of time in running water. The Carnoy was used but little. It does not fix so well as the Gil- son. Its real value is for material where an aqueous fixative can- not be used. After fixation, the material was run through the alcohols, cleared in xylol, and embedded in parafiine with a melting point of 52°C. Most of the sections were cut 5 « thick, a few 34 » and 5 fl. Many stains were tried. The three giving most satisfactory 480 Alice M. Boring results were Heidenhain’s iron hematoxylin, either without a counterstain, or With a slight tinge of orange G, thionin without a counterstain, and Auerbach’s combination of acid fuchsin and methyl green. With iron haematoxylin, the long method gave the best results. Preparations in this stain furnish the best outlines for camera drawings, but for work in spermatogenesis, there is the disadvantage that it often stains plasmosomes and chromosomes alike. ‘Thionin has proved a valuable stain for distinguishing between chromatic material and plasmosomes. With this mate- rial the best results are gained by leaving the slides in the stain from one to five minutes, rinsing off with water, and differen- tiating under the microscope with 95 per cent alcohol. ‘The basichromatin holds the stain as a navy blue or dark purple, depend- ing upon the material; while the plasmosome and oxychromatin either take a very pale blue, or hold no color at all. ‘The Auer- bach stain also gives differentiation between basi and oxychro- matin, the odd chromosome standing out bright green in the rest stage against the pink spireme or scattered oxychromatin. OBSERVATIONS Membracide In the Membracidz, the testes are situated ventrally, near the anterior end of the abdomen. ‘They are white in color, and each follicle is round. Such ripe spermatozoa as are present are found near the duct and the spermatogonia are situated on the opposite side. ‘The rest of the follicle is filled with the intermediate stages, grouped into cysts containing cells in about the same stage. ‘The succession of these stages is rather difficult to follow in the Mem- bracidz, because the follicles are spherical and no one longitudinal section gives all of the stages. ‘The only way to trace the develop- ment is to find cysts with most of the cells in one stage and a few in transition to the next stage. In this way, the links between the stages can be filled in. In the eight species from which my mate- rial was obtained, the general course of development is very simi- lar, with only here and there a striking difference. I shall there- S permatogenesis 48 I fore describe in detail one species, Entilia sinuata, and then men- tion the chief points of interest in the other species. Entilia sinuata This form was found in September, at Woods Hole, on the leaves of the Golden Glow, and later near Philadelphia, on the wild sunflower. The resting spermatogonia stain very lightly, as there are only a few basichromatin granules in the midst of much scattered oxy- chromatin (Fig. 1). When the cell is preparing for division, a heavy, rather darkly-staining spireme is formed with the chroma- tin aggregated at regular intervals along the linin (Fig. 2). A longitudinal split appears in this spireme, a slight indication of which can be seen in Fig. 2. The chromatin next becomes con- densed and segmented, but these segments still retain their linin connections. ‘The longitudinal split in each segment is also very conspicuous at this stage (Fig. 3). Condensation of the segments continues, there being first an elimination of the [enetadie split (Fig. 4), and then a shortening of the segments until they are about twice as long as broad, the form which they have as they enter the equatorial plate of the spindle (Fig. 5). “They appear in the plate with their longitudinal axis at right angles to the longitudinal axis of the spindle and with the linin connections still intact. ‘This division, therefore, is a longitudinal division, separating each chro- mosome into two parts along the line of the original longitudinal split, which appeared in prophase. A lateral view of the spindle in metakinesis also shows convincingly that this division is longi- tudinal (Fig. 6). The number of chromosomes in the spermato- gonial division is 21 but it is impossible to pick out the odd chromosome. The chromosomes become so_ closely massed together in anaphase (Fig. 7) that one cannot tell whether the linin connections still remain intact, or the conjugation of chromo- some pairs takes place here. By the time the cell division is com- pleted, the new nuclear membrane has been formed, possibly as Conklin (02) has suggested, by the joining together of the linin sheaths of the chromosomes after these have absorbed liquid from the cytoplasm (Fig. 8). A linin connection joining the chromo- 482 Alice M. Boring somes end to end 1s visible soon after they have lost their smooth contours (Fig. g). The last spermatogonial telophase is followed by a dense, darkly- staining contraction stage, which looks like a tightly wound spi- reme. Here the outlines of the chromosomes and their connections are entirely obliterated. The contracted mass occupies only a part of the nucleus, leaving a large clear space at one side (Fig. 10). This space appears in preparations where the fixation of other parts seems to be perfect, so it can hardly be looked upon as an artefact, as McClung (oo) at first claimed. I have used Wilson’s (’05b) expression, “contraction stage” as simpler than McClung’s “synizesis,’ for the most cipteeced period of “‘synapsis” as Moore used the term. The chromatin now goes through a series of changes comparable to those of Anasa tristis, (Wilson ’o5c): (1) an early postsynapsis, with a fine spireme, much twisted on itself, still staining deeply, but filling the nucleus much more com- pletely than in the contraction stage (Fig. 11); (2) a late post- synapsis, with the spireme filling the cell completely, less twisted, and staining unevenly (Fig. 12); (3) an early growth stage, with the spireme thicker, the basichromatin aggregated at regular intervals along the linin (Fig. 13); (4) a rest stage, where the spi- reme scarcely stains at all, and in the midst of the pale nucleus (in iron hematoxylin) there is one lens-shaped black body (Fig. 14), which, following Stevens, I shall call the odd chromosome. _ It is the “accessory of McClung, the “chromatin nucleolus” or hetero- chromosome” of Montgomery, the “chromosome spéciale’’ of de Sinéty, or the “heterotropic chromosome” of Wilson. From_ the action of similar bodies in related species, I am convinced that it must be present here in the postsynapsis and early growth stages, but the spireme stains so deeply and twists on itself so much that it hides the odd chromosome. In the succeeding stage, where the spireme becomes longitudinally split, the odd chromosome length- ens out and loses the smoothness of its outline, although not the intensity of its staining reaction (Fig. 15). The spireme next divides into ten segments, each retaining its longitudinal split (Fig. 16). Counting the odd chromosome, which remains closely applied to the nuclear membrane, there are now I1 chromatic S permatogenests 483 elements present in the nucleus. Just before the contraction stage, the spermatogonial chromosomes were joined end to end by linin connections, and out of the contraction stage there came a continuous spireme, which has passed through various stages and finally segmented. If the chromosomes conjugate end to end in the late anaphase (Fig. 8), as Fig. g might suggest, the longitudinal axis of the primary spermatocyte segments, or chromosomes, represents the longitudinal axis of the spermatogonial chromo- somes. ‘The presence of a massed anaphase and of the contrac- tion stage makes it impossible to prove that this 1s the case here. It has, however, been proved for other forms (Sutton) and the agreement of all other steps in the process points toa possible similarity in this respect also. The 10 segments next become tetrads by the formation of transverse arms which always remain a little shorter than the longitudinal arms, and thus make it always possible to distinguish between the longitudinal and transverse axes (Figs. 17 to 19). While the tetrads and dumb-bells are form- ing, the odd chromosome rounds up again and becomes a lens- ° shaped body, still applied to the nuclear membrane (Fig. 20). It is in. the dumb-bell form that the chromosomes usually enter the spindle (Fig. 24), but occasionally they are still in the form of cross- shaped tetrads (Fig. 22). This shows conclusively that, the longi- tudinal axis of the dum! -bell is the same as the longitudinal axis of the tetrad, and that the first spermatocyte mitosis is a transverse division. That it is probably a reducing division can be shown by tracing back the development, and working out the corresponding axes: the division between the halves of the dumb-bell (Fig. 24) corresponds to a division along the lateral arms of the tetrad (Fig. 17), and that to a transverse section of the spireme segment (Fig. 16) and that to the separation of one spermatogonial chromosome from another, if we assume that each spireme segment equals two spermatogonial chromosomes joined end to end. This may be further evidence against McClung’s (’00) contention that the reducing division is always the second. In the equatorial plate of the first spermatocytes the odd chromosome stands a little apart from the other 10 chromosomes, and is smaller in diameter (Fig. 21). It does not divide in the first spermatocyte division, but lags 484 Alice M. Boring behind the others in going toward the spindle pole (Figs. 25 and 27). [he chromosomes mass together in the anaphase, so that as soon as the odd chromosome joins the others, it is no longer possible to distinguish it (Fig. 28). The spindle fibers stand out very clearly, especially in the mate- rial fixed in Flemming or Hermann, and it is noticeable that the odd univalent chromosome 1s joined to only one pole by its mantle fibers, while the bivalent chromosomes are attached to both. During the telophase the granules of a “Zwischenkérper”’ can be seen on some (Fig. 25) or all (Fig. 26) the spindle fibers. These show only in iron hematoxylin preparations which have not been extracted very thoroughly. In such preparations the centrosomes of the first spermatocyte division can also be seen (Fig. 23). They divide during the anaphase of the first division (Figs. 25 and 27) in readiness for the second division which succeeds the first without any reconstruction of the nucleus. ‘The chromosomes rearrange themselves (Fig. 29) into a plane at right angles to the plane of the first division, and soon form a regular equatorial plate. Half of the second spermatocytes con- tain 10 chromosomes (Fig. 31) and the other half 11 (Fig. 30), that is, 10 plus the odd chromosome. In the cells containing 11 chro- mosomes, the odd one does not differ enough in size to make it any longer distinguishable. In this division, all the chromosomes in all of the cells divide. The reasons for this conclusion are: (1) the lateral views of the metaphase (Fig. 32) never show one undi- vided chromosome among the other dividing ones, (2) all the chro- mosomes are attached by mantle fibers to both spindle poles, and (3) in the anaphase, there is never a lagging chromosome near one pole without a mate at the other pole (Fig. 33). That this division of chromosomes is at right angles to the first, thatis, longitudinal and equational, is certainly conditioned by the formation of the spindle which is derived directly from that of the first division. The same fibers between the chromosomes and centrosomes remain intact, and as the centrosome divides, the chromosomes are pulled into an equatorial plate at right angles to the equatorial plate of the first spermatocyte division. ‘This second division therefore corresponds to the preliminary longitudinal splitting of S permatogenests 485 the spireme in the growth period. One spermatocyte division 1s reducing and the other equational. In the anaphase, the chromo- somes again become massed together (Fig. 34) and the nucleus is reconstructed by the formation of a nuclear membrane (Fig. 35). The “Zwischenkorper” is again noticeable in this telophase. In the young spermatid (Fig. 36), the chromatin is still massed together and stains deeply. ‘The spindle material remains as the ““Nebenkern,”’ as first described by v. La Valette St. George (’86) for insect spermatids. The chromatin soon scatters through the nucleus in definite clumps and it is evident that half of the sperma- tids contain a smooth round darkly-staining body (Fig. 37), while the other half do not (Fig. 38). Through several succeeding stages, this same fact is noticeable; 7. e., when the chromatin becomes more diffuse (Figs. 39 and 40), when it forms a pale net- work and the axial filament has grown out (Figs. 41 and 42), and even when the chromatin has begun to condense to form the head of the spermatozoon (Figs. 43 and 44). The method of deter- mining whether this body is in only half the cells or in all is as follows: cysts of spermatids in various places were picked out and the number of cells with and without this body were counted in each cyst. In studying sections, it must be remembered that parts of some cells are in another section, so even if this body (x) were actually present in all the cells, it would not appear in all in any one section of acyst. On the same principle, if it were actually in only half the cells, it would appear in less than half in any one section. In Entilia, this body appears in a few less than half of the sperma- tids. It always takes the chromatin stains, deep blue with thionin, and green with the Auerbach. As it resembles the odd chromo- some of the first spermatocyte rest stages in staining reaction and contour, and as it appears in not more than one-half of the sperma- tids, a condition which the odd chromosome necessarily fulfills from the fact of its not dividing in the first spermatocyte division, we seem to be justified in concluding that the body x of the sperma- tids 1s a derivative of the odd chromosome of the spermatocyte. There is nothing unusual about the formation of the spermatozoon. The “ Nebenkern”’ forms the sheath of the axial filament (Fig. 41), the acrosome differentiates from the cytoplasm at the apex of the 486 Alice M. Boring _ head, the head forms by condensation of the chromatin (Figs. 44 to 47), passing through one rather diffuse stage (Fig. 46). Vanduzea arcuata Vanduzea arcuata was found in abundance on the locust trees near Cold Spring Harbor in June. ‘The spermatogonial plates show 17 chromosomes, varying in size (Fig. 48). It is not possible to arrange them all in pairs, but at least two large pairs are well marked (a, and a,, b, and b,)._ In the growth stage, the odd chro- mosome appears as a long, darkly-staining body, without a smooth contour. It is at first bent upon itself in different forms (Fig. 49), and later lies at full length along the nuclear membrane (Fig. 50), resembling the same stage in Entilia sinuata (Fig. 15). In the equatorial plate of the first spermatocyte division, there are 9 chro- mosomes, two of which are larger than the others (Fig. 51, a and b), corresponding to the four large ones in the spermatogonial plate; ais slightly larger than 4 just as a, and a, were slightly larger than b, and b,._ This point certainly counts as evidence that each spermatocyte chromosome represents not an indefinite segment of the spireme, but two individual spermatogonial chromosomes. The odd chromosome can be recognized by i its eccentric position. Fig. 52 shows all the chromosomes but x in metakinesis, and in Fig. 53 x is passing to one pole undivided. Figs. 54 and 55 show variations in the position of x in anaphase; it does not always lag behind, but may even precede the other chromosomes to the pole. The second spermatocyte equatorial plates, containing g and 8 chromosomes, respectively, are shown in Figs. 56 and57. Each has one large chromosome a, one not quite so large b, and six small ones of about the same size. Fig. 56 has a ninth chromosome of intermediate size which must be the odd chromosome, as x in the first spermatocyte plate has a corresponding intermediate size (Fig. 51). All the chromosomes divide in this division, including the odd one, as is shown in all of the lateral views of the metaphase (Fig. 58) and of the anaphase (Fig. 59). Half of the spermatids contain the odd chromosome, and half do not (Figs. 60 and 61). S permatogenesis 487 S : Ceresa taurina. Three species of Ceresa were found near Cold Spring Harbor on the morning-glory vines and tall weeds, during the last three weeks of July. Unfortunately the chromosomes of the spermato- gonial plates in all three forms are too close together to make it possible to count them. ‘They all have the same reduced number of chromosomes and a peculiar deposition of chromatin on the nuclear membrane in the growth period. As this phenomenon is most pronounced in Ceresa taurina, I shall give the details for this form. In the contraction stage, the chromatin is massed at one side of the nucleus in a number of darkly-staining loops with their bases united in a dense flat chromatic plate, which stains more deeply than the loops (Fig. 62). As the loops spread through the nucleus, they stain less, making the contrast with the black plate more intense (Fig. 63). In the rest stage (Figs. 64 to 67), the reticulum does not take basic stains at all; the chromatin plate appears in various forms, sometimes continuous and sometimes broken up into two, three, or four parts. By the time a split spi- reme is formed, it has been almost entirely dissolved (Fig. 68), and in the prophases, no trace of it is left (Fig. 69). When these masses dissolve, the odd chromosome becomes visible as a round, smooth body (Figs. 67 and 68), which probably was concealed in the midst of the chromatic plate as far back as the contraction stage, but its presence was obscured by the similarity of its staining reac- tion to that of the other chromatin. As to the meaning of this deposition of chromatin on the nuclear membrane, it seems possi- ble that it is basichromatin thrown out from the chromosome loops in the contraction stage, and that it takes no part in the fur- ther formation of the chromosomes, since it disappears before the next division. ‘The only case at all similar which I can find in the literature is that of Gryllus campestris described by Voinov (’04). He claims that all the chromatin is gathered into the “corps nucle- inien double,” leaving the non-stainable achromatic substance spread through the nucleus, and that when thespireme forms, the chromatin is added to it again from this structure. He neglects the distinction between oxy and basichromatin, and thinks that when all 488 Alice M. Boring the stainable chromatin is aggregated 1 into one body, there is no chro- matin left elsewhere. The situation is much clearer if looked at from Conklin’s point of view (’02): although the nucleus in the rest stage does not take basic stains, it still contains chromatin in the form of oxychromatin; this has the power of changing into basichromatin to form the chromosomes for division. ‘The basi- chromatin masses of the rest stage, with the exception of the odd chromosome, which here again shows its individuality by a differ- ence in behavior, are apparently rejected substances, which dis- appear without playing any further role in karyokinesis. In the prophase, the odd chromosome lies close to the nuclear membrane as in the forms previously studied, and in the metaphase it has a somewhat eccentric position (Fig. 70). “The chromosomes here are so nearly of the same size that it is impossible to trace any individuals from cell to cell; but the odd chromosome, by virtue of its position and its univalence, can be followed until the second spermatocytes are formed. Figs. 71 to 73 show its varying behay- ior in metaphase; it may either follow or precede the other chromo- somes to the pole. ‘This fact is shown also by the two anaphase figures, 74 and 75. [he second spermatocyte equatorial plates show the two numbers of chromosomes 11 and Io (Figs. 76 and 77), but the odd chromosome can no longer be distinguished from the others, either in metaphase (Fig. 78) or anaphase (Fig. 79). In all the spermatids (Fig. 80), there appears one large body (7) taking the basic stains, probably analogous to the body in the beetle sper- matids called a chromatin nucleolus by Stevens (’o6b). It is impossible to decide whether the odd chromosome in half the sper- matids keeps its individuality as was observed in Entilia and Van- duzea, for all the chromatin stains deeply and in some stages 1s broken up into many separate masses (Fig. 80). Ceresa bubalus The only external difference between this species and the fore- going one Is its greater size and the different angle of the prothor- acic protuberances. ‘The only difference in the spermatogenesis as can be seen by Figs. 81 to g2, is that the mass of rejected chro- matin is not so conspicuous. In the bouquet stage (Fig. 81), the Per S permatogenesis 489 plate is not nearly so large as in the same stage of Ceresa taurina (Fig. 63). Fig. 82 represents one of the most extreme cases of the growth stage. Ceresa diceros The shape and size of this species is about the same as in Ceresa bubalus, but the coloring is different, being brown and white, instead of uniform green. ‘The spermatogenesis is practically the same, as Figs. g3 to 101 show, but a preparation from the testis of one could be distinguished from a preparation of the other, because the cells, chromosomes, and spindles of C. diceros are always smaller than those in C. bubalus. Atymna castanea This species was found on the chestnut trees exclusively, and was very abundant at the end of June and beginning of July. No spermatogonial plates in which the number of chromosomes could be counted were found. ‘The odd chromosome appears in the rest stage as a large round body with a smooth contour and an affinity for basic stains (Fig. 102). In lateral view of the metaphase of the first spermatocyte division, it is apparent that it does not divide (Figs. 104 and 105), and in the anaphase it has the position usually characteristic of this order, between the plates of chromosomes, but nearer one pole than the other (Fig. 106). The number of chro- mosomes in the first spermatocyte is again 11 (Fig. 103), two of them constantly larger than the others (a andb). These two large chromosomes appear in all the second spermatocyte plates, whether they have 11 or 10 chromosomes (Figs. 107 and 108). All the spermatids contain a chromatin nucleolus (Fig. 11), as in the genus Ceresa. ‘There being apparently no other basic-staining body in any of the spermatids, the odd chromosome in half of them must take part in the formation of the general reticulum like the other chromosomes. Campylenchia curvata Campylenchia curvata was found in sweepings from various weeds throughout July. The material showed all desirable stages. 490 Alice M. Boring Many spermatogonial plates were found, some of which it was possible to count. It seems that there must be one short period in the arrangement of the chromosomes into the plate, when they are spread further apart than at any other time. Judging from the behavior of the chromosomes of the first spermatocyte in coming into the equatorial plate, this more open stage must occur when the chromosomes are first drawn into a flat plate from their scat- tered position in prophase. Later as metakinesis begins and the mantle fibers pull from the two poles, the chromosomes are drawn closer together and the diameter of the plate becomes smaller. Fig. 112 shows a very clear spermatogonial plate, with 19 chromo- somes. It is possible here to group the chromosomes into g pairs with one left over; only the two most distinct pairs are lettered, a, and a,, long and slender, }, and b,, a little shorter and thicker. The two chromosomes formed by the fusion of these pairs are designated by a and 4 in Fig. 114, the equatorial plate of the first spermatocyte, and in Figs. 117 and 118, the equatorial plates of the second spermatocytes. [he number of chromosomes in the equatorial plates are what would be expected after finding 19 in the spermatogonia; 10 in the first spermatocytes, and Io andg, respectively, in the second. In the rest stages (Fig. 113), x ap- pears as usual, but there are also present two other smaller bodies with the same staining reaction, m, and m,. I have called them m-chromosomes, as they have all the characteristics of Wilson’s m-chromosomes in the rest stage of the Hemiptera Heteroptera (o5c); they are of equal size and they take the basic stains like the odd chromosome. As unfortunately they are not enough smaller than some of the other chromosomes to be readily distinguished in the spermatogonial plate, or to be traced through the prophase of the first spermatocyte to the spindle, it is impossible to see whether they really represent one pair whose fusion has been delayed. The odd chromosome appears as usual in metaphase (Fig. 115) and anaphase (Fig. 116) of the first spermatocyte division, and as usual is not distinguishable in the metaphase (Fig. 119) or ana- phase (Fig. 120) of the second division. In the spermatids, a basic-staining body appears in half the nuclei (Figs. 121 and 122), and so must here (as in Entilia and Vanduzea) represent the odd S permatogenests 491 chromosome, rather than the chromatin nucleolus of the other Membracidz studied. Enchenopa binotata Enchenopa binotata was found throughout July at Cold Spring Harbor on the locust and wild cherry trees, on blackberry bushes and sometimes in general sweepings of weeds. Its spermato- genesis has been the most puzzling of any form studied and the following account is given tentatively, with the intention of going over the work as soon as more material can be obtained. ‘The first facts to be noticed are that all the chromosomes appear as dumb-bells in the metakinesis of the first spermatocyte (Fig. 128), there is no lagging chromosome in the anaphase (Fig. 130), and all the second spermatocytes have 10 chromosomes (Fig. 131), the same number as the first spermatocytes. In iron hematoxylin preparations extracted to the same degree as in other material, no darkly-staining body appears in the rest stage, but in those extracted for a shorter time, a long twisted body appears against the pale spireme (Fig. 124). This can occasionally be traced into a stage where the spireme has segmented (Fig. 125), but never any further, as it does not assume a compact rounded shape until the other chromosomes become condensed. ‘The question arises as to whether this body in the growth stage represents two sper- matogonial chromosomes and consequently divides in both sper- matocyte divisions as all bivalent chromosomes do; or whether it is univalent, analogous to most odd chromosomes in insects, but divides in the first spermatocyte division and not in the second, thus differing from all the other Hemiptera Homoptera studied and resembling most of the Heteroptera. ‘There were a few sper- matogonial plates in such a stage that it was possible to count the chromosomes, but these did not have the chromosomes as clearly spread apart as in most the other species studied. In five plates, 19 chromosomes were counted (Fig. 123) and in two, 20. One of those with 20 may, however, be deceptive; two of the chro- mosomes are much smaller than any in the other plates, the plate is at the surface of the section, and as x in Fig. 1231s V-shaped, it is possible that the bend of the V was cut off and the two small chro- 492 Alice M. Boring mosomes may really be but one. Other evidence for the univa- lence of one chromosome is its occasional appearance in early metaphase of the first spermatocytes when it has not yet assumed the dumb-bell shape (Fig. 129), and a few second spermatocyte metaphases where it apparently does not divide (Fig. 133). If it does not divide in the second spermatocyte division, the second spermatocyte spindle should always appear as it does in Fig. 133 rather than as in Fig. 132, unless the odd chromosome is usually in the center surrounded by the other chromosomes. ‘That this probably is true 1s indicated by several cases like Fig. 135, the two anaphase groups of one second spermatocyte spindle, a having g chromosomes and } 10. ‘There is a space in a corresponding to the chromosome marked x in b. This evidence is anything but satisfactory, but the possibility of such an exception to the general rule that the odd chromosome divides in the first spermatocyte division, is too interesting a fact to leave unmentioned. Here again one large chromosome in the first spermatocyte (Fig. 126) is represented by two in the spermatogonia (Fig. 123, a, and a,), and by one in the second spermatocyte (Fig.131,a). Fig. 127 shows an occasional first spermatocyte with 11 chromosomes, implying a delay in the fusion of one pair. Here we find the chromatin nucleolus in all the spermatids (Fig. 136). Fasside The testes of the Jassidz are pale yellow in color, and there- fore very easy to dissect out. ‘The follicles are about three times as long as broad; this makes it easier to trace the development from stage to stage than in the Membracida. My material includes six species, four of them caught at Cold Spring Harbor in July, and the other two, Agallia sanguinolenta and Phlepsius irrotatus, at Bryn Mawr in October. Chlorotettrix unicolor and C. vividus This material was fixed and preserved as belonging to one species, but study of the sections showed two different reduced numbers of chromosomes, 11 and 9g. ‘This led to a careful com- S permatogenesis 493 parison of my specimens with those in the collection at the Acad- emy of Natural Sciences, Philadelphia. There proved to be two species, C. unicolor and C. vividus, in which the only marked difference is the width of head and thorax. Some of my specimens are slightly narrower than others, so | have probably mixed the two species, and cannot state whether the g chromosomes belong to C. unicolor or to C. vividus. The resting spermatogonium has a reticulum of oxychromatin and linin and a plasmosome, which:stains black in iron hamatox- ylin, but shows its achromatic nature in thionin (Fig. 137). There were no good spermatogonial plates in the material with the smaller number of chromosomes, but a lateral view of the spindle is shown in Fig. 138, and the anaphase in Fig. 139. The chromatin then passes into a contraction stage which is very dense, but contains several clear vacuoles (Fig. 140). This has a very different appearance from the contraction stage of the Membracide. A spireme stage follows where the chromatin again fills the nucleus and still stains deeply (Fig. 141). The odd cchtontmeens is first visible in the rest stage (Fig. 142) where the chromatin stains least and is most scattered. It is closely applied to the nuclear membrane as was usually the case among the Membracide. ‘The spireme splits longitudinally (Fig. 143), and then becomes seg- mented (Fig. 144). In all stages the odd chromosome can be distinguished by its small size. In the prophase of the first sper- matocyte division, it can be recognized by its rounded contour; in the equatorial plate, by its eccentric position (Fig. 145); in the lateral view of the metaphase (Fig. 146), by its undivided condi- tion; and in anaphase, by its lagging behind at one pole of the spindle (Fig. 147). In the equatorial plates of the second sper- matocytes with g chromosomes, it can still be recognized by its small size (Fig. 149). As it divides in the second spermatocyte division, there is no indication of it in a lateral view of the meta- phase (Fig. 150), or anaphase (Fig. 151). Two of the g chromo- somes are larger than the others (a and bin Fig. 145), and they keep their individuality in the second spermatocyte (a and 4 in Figs. 148 and 149). In all the spermatids, there is one condensed body, which resembles the body called a chromatin nucleolus in 494 Alice M. Boring five species of the Membracidz. In the early spermatid, this 1s the only condensed body distinguishable (Fig. 152), but later when the chromatin becomes more diffuse, it appears that half the sper- matids have another smaller condensed body (Figs. 153 and 154), which is lacking in the other half. This must be the odd chromo- some, observed in the same stages of three species of Membracidz. In a still later stage, when the reticulum is arranged around a series of clear vacuoles, this difference is still to be observed; all the cells have the one large body, but only half have the small chromosome (Figs. 155 and 156). After this, both bodies disappear, the chro- matin reticulum becomes slightly more condensed at first (Fig. 157), the nucleus then elongates but keeps the vacuoles (Fig. 158), and finally condenses into the head of the spermatozoon (Fig. 159). The acrosome is differentiated from cytoplasm at the apex of the head. Fig. 160 is the spermatogonial plate of the species with the larger number of chromosomes. It contains 21 chromosomes, four larger than the others, not differing conspicuously in size among themselves (a,, a,, ,, b,). The first spermatocyte equa- torial plate has 11 chromosomes, and they show the same size relation as those of the other species, two large ones and one small odd chromosome in an eccentric position (Fig. 161). This plate simply has two more chromosomes of intermediate size than the other. The second spermatocyte plates again show the two large chromosomes (Figs. 162 and 163), the total numbers being 11 and 10, instead of g and 8. Diedrocephala coccinea A few scattered individuals were found in July in general sweep- ings, but in August an abundance of material was obtained from the blackberry vines. The spermatogonial plates show 23 chro- mosomes, two larger than the others (a, and a, in Fig. 164). In the postsynapsis stage, the odd chromosome is not surrounded by the spireme, as has been the case in the forms described above, but it stands out distinctly by itself in the clear part of the nucleus (Fig. 165). In the rest stage, it is still of the same size and in the same position, although the nucleus grows much larger and the. S permatogenesis 4.95 chromatin becomes scattered and diffuse (Fig. 166). The first spermatocyte shows the odd chromosome as a medium-sized body, eccentric in the plate of 12 chromosomes (Fig. 167), and not divid- ing in metakinesis (Fig. 168). In anaphase, it lags behind the others (Fig. 169). “The two large chromosomes of the spermato- gonia have fused into a single large one in the first spermatocyte (a in Fig. 167), and this keeps its individuality in the second sper- matocytes (a in Figs. 170 and 171). Half the second spermato- cytes have 12 chromosomes, and half 11. The spermatids all have the chromatin nucleolus, and half of them the odd chromo- some (Figs. 174 and 175), as in Chlorotettrix. Diedrocephala mollipes This species resembles Diedrocephala coccinea in shape, but not in color, being bright green instead of red and green striped. Its spermatogenesis is also similar (Figs. 176 to 185), but the cells and chromosomes are smaller (cf. Fig. 177 and 167). They both have the same number of chromosomes, 12, but Diedrocephala mollipes has no one chromosome markedly larger than the others. The spermatids have both a chromatin nucleolus and an odd chromosome. Phlepsius irrotatus : The spermatogonial plate contains 15 chromosomes, two larger than the others (a, and a,, Fig. 186). ‘These are represented by a in the first spermatocyte (Fig. 188a) and also in the second sperma- tocytes (Figs. 191 and 192, a). The growth period shows the odd chromosome (x) as a round body with even contour (Fig. 187). The univalent chromosome x has the peculiarity here that it never comes to lie in a flat plate with the other chromosomes in the first spermatocyte division, as is indicated in Fig. 189. ‘To get all 8 chromosomes, the equatorial plate must be drawn at two different foci (Figs. 188aand 188b). ‘The odd chromosome always precedes the others to the pole (Fig. 190), never taking the lagging position characteristic of the species previously described. We have noted that this sometimes takes place in other forms (Vanduzea arcuata, and the three species of Ceresa), but Phlepsius is the first form 496 Alice M. Boring where this position is invariable. “The second spermatocytes con- tain 8 and 7 chromosomes (Figs. 191 and 192). The spermatids all contain the chromatin nucleolus (Figs. 195, 7, and 196, m) and half of them, an odd chromosome (Figs. 195 and 196, 2). Agallia sanguinolenta No spermatogonial plates were found in this form. ‘The odd chromosome appears as usual in the growth period (Fig. 197). There are 11 chromosomes in the first spermatocyte (Fig. 198), and 11 and 1o in the second (Figs. 200 and 201). The odd chro- mosome does not divide in the first spermatocyte metakinesis (Fig. 199), but passes to one pole after the other chromosomes in anaphase (Fig. 206). The spermatids all contain a chromatin nucleolus, and half of them, the odd chromosome (Figs. 203 and 204). Figs. 205 to 207 are drawn from aceto-carmine preparations at the same magnification as Figs. 197 to 204. Cercopide The testes of the Cercopidz are situated near the posterior end of the abdomen. ‘They are white in color, and each follicle is round, with a comparatively long duct joining it to the vas deferens. The material comprises four species, and the spermatogenesis of none of them resembles very closely that of the species studied by Stevens (’o6b). Clastoptera obtusa This species was found on the alder at Cold Spring Harbor. The resting spermatogonium stains very lightly and has a plasmo- some (Fig. 208). In preparing for division, the chromatin forms a spireme, which becomes more dense, and then segments (Fig. 209). There are 15 chromosomes in the spermatogonial equa- torial plate, all of about the same size (Fig. 210). ‘The division 1s longitudinal as usual (Figs. 211 and 212). After the telophase, the chromosomes soon become joined by linin connections (Fig. 213), form a compact spireme in early synapsis (Fig. 214), a dense mass in the contraction stage (Fig. 215) and a spireme loosely wound on itself in postsynapsis (Fig. 216). “The odd chromosome S permatogenests 497 appears inthe contraction stage distinct from the dense chromatin mass, and remains so in postsynapsis and the early growth stage (Fig. 217). Itis from the first, a small, ovoid, pea boniaured body, and still shows clearly when the spireme has segmented and the tetrads are forming (Fig. 218), and when the ena bells are formed (Fig. 219). It takes an eccentric position in the equatorial plate of the first spermatocyte (Fig. 220). It does not divide in the first spermatocyte division (Fig. 221), and is the last chromo- some to reach the pole in the anaphase (Fig. 222). As there are 7 chromosomes, plus the odd one, in the first spermatocyte, so there are 8 in half the second spermatocytes (Fig. 223), and 7 in the others (Fig. 224). The odd chromosome behaves like the others in the second division (Figs. 225 and 226), and is not distinguish- able in the spermatids, all of which have a chromatin nucleolus (Fig. 227). In the development of the spermatid, the chromatin reticulum first becomes massed on the side of the nucleus toward the axial filament (Fig. 228), and then forms a dense U, leaving the rest of the nucleus clear (Fig. 229). “The nucleus then elon- gates, still leaving a clear space toward the apex (Fig. 230). The mature spermatozoon has a solid dense chromatic head (Fig. 231). Aphrophora quadrangularis This species was found on the grass and low bushes in July near Cold Spring Harbor. Originally a small quantity of material was collected and tried in aceto-carmine, as it was supposed to be the same species that Stevens (’06b) had found in Maine and described. But the reduced number of chromosomes proved to be 11 instead of 12, so material was fixed in Gilson and kept to be studied at a convenient time. ‘The material was obtained from two distinct localities, but not kept separate. [he sections showed follicles with 11 chromosomes and a few with 12. Whether this difference corresponds with the difference in locality it 1s unfortunately not possible to say. Another peculiarity is that the form with 12 chro- mosomes does not resemble, in some of its stages, the form with 12 chromosomes described by Stevens. ‘The most important stages of the form with 11 chromosomes are shown in Figs. 232 to 242. There are 21 spermatogonial chromosomes (Fig. 232) and 498 Alice M. Boring 11 and 10 second spermatocyte chromosomes (Figs. 238 and 239). The odd chromosome can be traced as an individual as far back as the contraction stage (Figs. 233, x,and234,x). Aplasmosome (p) also appears in the growth period, the thionin clearly bringing out the difference between the two. One of the 11 chromosomes is larger than the others, as is shown in Figs. 235, 238, 239. Uhe odd chromosome does not divide in the first spermatocyte division (Figs.236 and 237). ‘The spermatids all contain a chromatin nucle- olus (Fig. 242). A few stages of an individual with 12 chromo- somes are shown in Figs. 243 to 248. This series much more nearly resembles that of the other form from Cold Spring Harbor with 11 chromosomes, than that of the form found in Maine with 12 chro- mosomes. ‘The Maine form has no contraction stage (Stevens ’o6b, Figs. 240 to 249), while this form has a distinct one with the odd chromosome and the plasmosome outside of the spireme in the clear part of the nucleus (Fig. 243). The only possible con- clusion seems to be that three species (so determined by the differ- ences in spermatogenesis) have been up to this time grouped as one, and all called Aphrophora quadrangularis. Aphrophora 4-notata Aphrophora 4-notata is interesting especially in connection with Aphrophora quadrangularis, as being another case of differ- ence of chromosome number within the same genus. Aphrophora 4-notata has 14 chromosomes for the reduced number (Fig. 250) and consequently 14 in half of the second spermatocytes (Fig. 253) and 13 in the other half (Fig. 254). “The odd chromosome is pres- ent in the spireme stage (Fig. 249), and does not divide in the first spermatocyte division (Figs. 251 and 252). Fulgoride The testes of the Fulgoride are orange-colored and show through the thin white walls of the abdomen. The separate folli- cles are oblong. Of the four species in my material, three belong to the genus Peeciloptera, and one to the Amphiscepa, but according to the spermatogenesis, P. bivittata is much more like the Amphis- S permatogenests 499 cepa than like the other two species of Pceciloptera. P. septen- trionalis and P. pruinosa were found on the nettle and the other two species came from sweeping low grasses. In this material, the cells and chromosomes are large and the achromatic struc- tures especially well preserved. The material fixed in Flemming, and stained in thionin makes some of the clearest preparations included in this study. Poeciloptera septentrionalis The resting spermatogonia of this form are small and stain lightly (Fig. 256). In preparation for division, a spireme 1s formed, each granule of which splits longitudinally (Fig. 257). The chromatic part of the spireme segments, retaining the linin connections and also an indication of the longitudinal split (Fig. 258). There are 27 chromosomes in the spermatogonial plate, two longer than the others (a, and a, of Fig. 259). Fig. 260 shows distinctly that this division follows the preliminary longitudinal split. After the telophase, the chromosomes become more diffuse and join into a spireme (Fig. 262). “This spireme contracts into a small dense ball at one side of the nucleus (Fig. 263), and then the cell goes through a long growth period in which the diameter is at least doubled. The odd chromosome appears as soon as the spireme becomes pale enough to conceal it no longer (Fig. 264). Then a pair of m-chromosomes appears and a small plasmosome (Fig. 265). The plasmosome and odd chromosome both increase in size, the latter having a vacuole in the center (Fig. 266). ‘The odd chromosome has now attained its full size, but while the cell and nucleus continue to increase, the plasmosome keeps on grow- ing (Fig. 267). Even though it is now larger than the odd chromo- some, it stains scarcely at all, while the odd chromosome and the m-chromosomes stain a deep blue, thus demonstrating the valu- able differentiating powers of thionin. In the next stage (Fig. 268) the odd chromosome and the plasmosome are unchanged, but the spireme stains more deeply and shows a longitudinal split. “The m- chromosomes no longer appear, they have probably become indis- tinguishable from the other spireme segments. ‘The plasmosome and odd chromosome still keep the same relative size in the pro- 500 Alice M. Boring phase, while the tetrads are forming (Fig. 269), the plasmosome sometimes not being dissolved until after the spindle is formed (Fig. 271). There are 14 chromosomes in the equatorial plate of the first spermatocyte (Fig. 270), one of them being marked by its eccentric position, another by its large size. ‘This large chromo- some keeps its individuality in all the second spermatocytes, those with 14 chromosomes (Fig. 273), and those with 13 (Fig. 274). The odd chromosome does not divide in the first spermatocyte division (Figs. 271 and 272), but does in the second (Figs. 275 and 276). The development of the spermatid in this family is very peculiar. The nucleus stains quite deeply, so that nothing more can be made out than that there seems to be one condensed body in each spermatid (Fig. 279a). The “Nebenkern” goes through a complicated development somewhat similar at first to that de- scribed by Baumgartner (’02). First delicate fibers are formed in it (Fig. 277), then it appears as a long coiled fiber in a clear space, surrounded by a definite membrane (Fig. 278). ‘This space becomes separated by a partition into two tubes, each containing several shorter fibers (Figs.27gaand b). ‘These tubes and fibers both become elongated (Fig. 280). The tubes grow still longer and smaller in diameter, and at the same time twist around each other in an irregular spiral (Fig. 281a). Cross sections through different portions of these twisted tubes indicate that they must also be constricted in places (Fig. 281b). They finally become flattened, presenting some such an appearance as in Fig. 282a, and in cross section as in Fig. 282b. In this species, the chromo- somes in the female somatic cells could be counted, and proved to be 28 in number (Fig. 283), there being the same two long ones that appeared in the spermatogonial plate. The significance of the even number in the female, and the odd number in the male will be pointed out in the theoretical considerations. Pceciloptera pruinosa Pceciloptera pruinosa resembles the last described form exter- nally in every character but color, being a grayish purple instead ofa pale green. The principal stages are shown in Figs. 284 to 293, the only difference being that there are two large chromosomes S permatogenesis 501 instead of one, in the first spermatocyte equatorial plate (Fig. 285) and also in the second spermatocyte plates (Figs. 288 and 289). The chromatin in the spermatid nucleus does not stain so deeply, and here it can be demonstrated that there is achromatin nucleus in all of the spermatids (Figs. 292 and 293), and the odd chromosome besides in half of them (Fig. 292). Here also the female somatic chromosome number 1s 28. Fig. 294 shows some of the chromo- somes overlapping each other, but they are really entirely separate from one another, lying at slightly different levels; itis a late pro- phase stage of an egg follicle cell before the chromosomes are drawn completely into one plane. Amphiscepa bivittata All this material came from larve. ‘The different stages are shown in Figs. 295 to304. “he spermatogonial plates contain 25 chromosomes, two pairs of long ones, one pair longer than the other (Fig. 295). In the rest stage, there are no m-chromosomes, but two plasmosomes are present (Fig. 296). ‘The first spermato- cyte plate shows two large chromosomes, one larger than the other (Fig. 297), corresponding to the two large pairs of the spermato- gonium. ‘The plasmosomes here persist into the metaphase (Fig. 298). The odd chromosome is quite small (Figs. 297,-x, 298, x, 299, x) and does not divide in the first division. Chromosomes a and 5b of the first spermatocyte retain their relative sizes in the second spermatocytes, both those containing 13 chromosomes (Fig. 300), and those with 12 (Fig. 301). Peeciloptera bivittata Peeciloptera bivittata very closely resembles the last described species, even to the number and relative sizes of its chromosomes (Figs. 305-313). It has two plasmosomes in the growth period, and one or both of these persist in a most remarkable fashion even to the anaphase of the second spermatocyte division (Fig. 312). The size of the chromosomes and cells is greater than in Amphis- cepa bivittata. 502 Alice M. Boring THEORETICAL CONSIDERATIONS Individuality of the Chromosomes The theory of the individuality of the chromosomes was first proposed by Boveri (’88) as a result of his work on Ascaris. He found a constant number of chromosomes in each species, always half this number in the two maturation divisions, and the original number restored by fertilization. Every year adds to the number of species found conforming to these rules, and consequently making Boveri's theory more plausible. Beginning with Sutton’s work in 1goo, many species have been shown to give evidence of a more direct nature, and among these, the Hemiptera Homop- tera can be classed. In the first place, it is a sign of individuality, when we are able to pick out one chromosome in every equatorial plate by some characteristic size, shape or position. ‘This can be done for 14 out of the 22 species of Hemiptera Homoptera studied, the characteristic usually being the large size of the chromosome (see Pceciloptera septentrionalis, Figs. 270, 273, 274). Secondly, all evidence that supports Montgomery’s hy- pothesis of the union of paternal and maternal chromosomes in synapsis necessarily supports the theory of the individuality of the chromosomes. In Peeciloptera septentrionalis, the large chromo- some in the spermatocytes (Fig. 270) is represented in the sperma- togonia (Fig. 259) by two large chromosomes. Half of the chro- mosomes in each spermatogonial plate must have come originally from the spermatozoon, and half from the egg. Only one large chromosome could be received from the spermatozoon, according to Fig. 270, therefore the other large one must have come from the egg. As these two large chromosomes, one paternal and one maternal, are represented by a single chromatic element in the spermatocyte, this must be formed by the union of a paternal with a maternal chromosome of the spermatogonium. ‘Thus we see that the Hemiptera Homoptera are in accord with Montgomery’s hypothesis of synapsis and reduction. In the third place, the behavior of the odd chromosome supports Boveri's theory. In the Hemiptera Homoptera, the odd chromosome can seldom be identi- fied in the spermatogonia, but from the contraction stage to the S permatogenests 503 anaphase of the first spermatocyte, and sometimes to the meta- phase of the second spermatocyte (Figs. 56 and 149) its individ- uality is marked. It takes the basic stains when the rest of the chromatin takes acid stains; it frequently has a smooth round con- tour in the early prophase, when the other chromosomes are irreg- ular rods or tetrads; it usually is closely applied to the nuclear mem- brane until that is dissolved, and then keeps an eccentric position in the first spermatocyte equatorial plate; it does not divide in this divi- sion, and either precedes or follows the other chromosomes to the pele. In Vanduzea arcuata (Fig. 56), where it is intermediate in size, and in Chlorotettrix (Fig. 149), where it is the smallest chro- mosome, its individuality is still marked in the second spermato- cyte. Finally the facts that have brought about the dropping of the old discussion about prereduction and postreduction, speak for the individuality of the chromosomes, in that they show the essential point of reduction to be the separation of each maternal chromosome from its paternal mate, and their distribution to differ- ent spermatozoa. ‘The uselessness of insisting on prereduction or postreduction is shown within the order Hemiptera, where the odd chromosome may divide in either division; in the Heteroptera, it usually divides in the first, while in the Homoptera, the usual place of division is the second spermatocyte, but Archimerus and Banassa are exceptions in the former and Enchenopa in the latter. Value of the Number of Chromosomes in Taxonomy and Evolution McClung (’05) states that for Orthoptera, a certain number of chromosomes is characteristic for each family, the chromosome grouping marking the genus, and the relative size of the chromo- somes indicating the species. Unfortunately this is not true for the Hemiptera Homoptera as the number varies within the family and even within the genus, being constant for the species only. The case of Aphrophora quadrangularis may make this doubtful, although it seems more probable that two or three species have previously been included under one name, than that in the same species, the reduced number should be sometimes 12 and some- times 11, which would not accord with the simplest laws of heredity. 504. Alice M. Boring Montgomery has for many years endeavored to determine the stage of evolution by the number of chromosomes that a species possesses, those having few being considered higher in the scale than those with many. ‘The chromatin nucleoli were supposed to be degenerating chromosomes as a species evolves to a higher form. But he has recently collected data from all the scattered literature, tabulated the number of chromosomes and the species, and finds that there is no such correlation (’06b). Inthe Hemiptera Homop- tera there is no reason for considering Vanduzea arcuata, with 9 chromosomes, more highly evolved than Entilia sinuata, with 11, or Phlepsius irrotatus, with 8, more so than Peeciloptera septen- trionalis, with 14. Sex Determination We have seen in the historical review of the work on tracheate spermatogenesis, that the most recent and reliable work all points to a dimorphism of the spermatozoa in the forms with an odd chro- mosome or an unequal pair of chromosomes. McClung was the first to suggest that the one characteristic that most generally divides tlie anienad kingdom into two equal classes is sex, and that therefore, the dimorphism of sex and of spermatozoa may be causally connected. ‘There is need of careful statistical work on the proportion of males and females among different species of insects. In general collecting, however, one gets an impression of equality in numbers. McClung’s theory was a brilliant guess, which the work of Stevens and Wilson has substantiated. The Hemiptera Homoptera furnish additional evidence for this theory. Females of many of the species were sectioned for o6go- nial and somatic equatorial plates. Only two furnished the desired stages, Poeciloptera septentrionalis and Pceciloptera pruinosa. In both the spermatogonial number is 27, the spermatozoa pos- sessing 13 and 14 chromosomes, and the female somatic number is 28. Stevens and Wilson have shown that there is no difference between the somatic number and the unreduced number in the germ cells; in the female, both numbers are even, in the male, both are odd (or even, when a small chromosome 1s included). As the female somatic number in Peeciloptera is even, the odgonial S permatogenesis 505 number must also be even, and all the maturated eggs necessarily possess the same number of chromosomes, 14... Applying Wilson’s Co6b) formula for sex determination to the Pceciloptera, we have the following: I Egg (14 chromosomes) + Spermatozoon (14 chromosomes) = Female (28 chromosomes). Il Egg (14 chromosomes) + Spermatozodn (13 chromosomes) = Male (27 chromosomes). Here again it is possible to apply Castle’s (’00) theory of sex as a Mendelian character, which has been so fully elaborated and applied to the case of the odd chromosome by Wilson. _ It involves the assumption of two kinds of eggs, male and female, as well as the two kinds of spermatozoa which are actually to be observed. It also involves the assumption of selective fertilization: an egg bearing the female determinant must be fertilized by a spermato- zoon with the male determinant, while an ege¢ bearing the male determinant must be fertilized by a spermatozoon with the female determinant. In case II of the above formula when the egg 1s fertilized by the spermatozoon without the odd chromosome, the sex determinant must be introduced by the egg; and as in this case, a male is produced, the eggs fertilized by a spermatozoon without an odd chromosome must bear the male determinant, and the chromosome which has disappeared in the males must be the one with the female character. So in case I, where the egg 1s fertil- ized by the spermatozo6n with the odd chromosome, the sperma- tozoon must bear the male character and the egg the female; as this combination always results in a female, it 1s necessary to assume that the male character is recessive and the female domi- nant. ‘lhe above formule can be extended to show these assump- tions and will read thus: I ¢ Egg (14 chromosomes) + (%) Spermatozoon (14 chro- mosomes) = 9° (3) Female (28 chromosomes). II (#) Egg (14 chromosomes) + (0) Spermatozoon (13 chro- mosomes) = (<) (0) Male (27 chromosomes). This is the part of Wilson’s theory that deals with the case presented by Peeciloptera and presumably the other Hemiptera Homoptera. ‘The facts as far as they go are not at variance with the theory. 506 Alice M. Boring SUMMARY 1 An odd chromosome is present in the spermatogenesis of 22 species of the Hemiptera Homoptera, as shown in each case by some or all of the following facts: a [The spermatogonia have an uneven number of chromosomes. b A dense body takes basic stains in the growth period. c One chromosome stands in an eccentric position in the first spermatocyte equatorial plate. d In the metaphase of the first spermatocyte division, one chromosome does not divide, and has half the valence of the others, as shown by its spherical shape when the others are like dumb- bells. e In anaphase of the first spermatocyte division, one chromo- some at one pole behaves differently from the others, either pre- ceding or lagging behind. 7 Half of the equatorial plates of the second spermatocytes contain the same number of chromosomes as those of the first spermatocytes, but half contain one less. g Half of the spermatids contain a condensed body, taking basic stains, which is the odd chromosome. 2 The odd chromosome shows certain variations in behavior, either individual or specific. a Inthe anaphase of the division where it does not divide, in some cells it may precede the other chromosomes to the poles, while in others it lags behind them. This individual variation is a characteristic of certain species, the three species of Ceresa and Vanduzea arcuata, while most of the species studied have the odd chromosome always lagging be- hind, and Phlepsius irrotatus has it always preceding the others. b In Enchenopa binotata, it divides in the first division, and in the second division, where it does not divide, it neither precedes nor lags behind the others. c The shape of the odd chromosome in the growth period varies. It may be always spherical or ovoid with a smooth con- tour, as inthe Fulgoridz, Cercopidz, Jassida, and some of the Membracidz. It may be long and uneven in contour as in Van- duzea arcuata and Enchenopa binotata. S permatogenesis 507 It may pass through both forms in different stages, as in Entilia sinuata. 3. Inthe spermatids of 19 species; that is, all except three of the Membracide, there is a chromatin nucleolus in all of the sperma- tids entirely independent of the odd chromosome. In seven of these species, the odd chromosome is present also in half of the spermatids, in others there is no indication of it. In the three Membracidz without the chromatin nucleolus, Entilia sinuata, Vanduzea arcuata, and Campylenchia curvata, the odd chromo- some is present in half of the spermatids. 4 In the genus Ceresa, in the contraction stage some of the basichromatin is thrown out from the chromatin loops and per- sists through the growth period as a chromatin deposition on the nuclear membrane and finally dissolves without apparently taking part in the formation of the chromosomes for the first spermato- cyte division. 5 In three species, Campylenchia curvata, Poeciloptera septen- trionalis, and Poeciloptera pruinosa, a pair of m-chromosomes remain condensed in the growth period. 6 The number of chromosomes has no significance for group- ing species into families. In reduced number, in the Membracida, 5 species have 11 chromosomes -' 2 species have 10 chromosomes species has 9 chromosomes * _ in the Jassidz, 2 species have 12 chromosomes “ 2 species have 11 chromosomes ¥ I species has —g chromosomes I species has 8 chromosomes in the Cercopide, 1 species has 14 chromosomes I species has 12 chromosomes I species has 11 chromosomes I species has 8 chromosomes Vv in the Fulgoridz, 2 species have 14 chromosomes | 2 species have 13 chromosomes 7 ‘Vhe number of chromosomes has no significance for group- ing species into genera. 508 Alice M. Boring Chlorotettrix unicolor, 11 chromosomes Chlorotettrix vividus, g chromosomes Aphrophora quadrangularis, 11 or 12 chromosomes Aphrophora 4-notata, 14 chromosomes Peeciloptera septentrionalis, 14 chromosomes Peeciloptera bivittata, 13 chromosomes 8 The number of chromosomes is constant for each species. In the case of Aphrophora quadrangularis, where there have been found both 11 and 12 chromosomes, probably two species are pres- ent, which have not been separated in classification. g The only points in the spermatogenesis in which all of the species of one family resemble each other more closely than they do the species of the other families are the appearance of some of the growth stages and the transformation of the spermatid into the spermatozoon. 10 In fourteen of the species studied, the individuality of cer- tain chromosomes can be traced from the spermatogonium to the second spermatocyte, a pair of similar chromosomes in the sperma- togonium bearing the same size relation to the other chromosomes of the equatorial plate as a single chromosome bears to the others in the first and second spermatocyte plates. In all the species, the odd chromosome can be traced as keeping its individuality from the growth period to the anaphase of the first spermatocyte division, in Chlorotettrix and Vanduzea arcuata to the metaphase of the second spermatocyte division, and in Enchenopa binotata, from the spermatogonial plate to the telophase of the second sper- matocyte division. 11 In all 22 species, there 1s a dimorphism of the spermatozoa, which probably corresponds to the. natural dimorphism of sex. 12 Two species of Fulgoridz in which the female somatic num- ber of chromosomes is 28, while the spermatogonial number is 27, furnish further proof for the theory of sex determination advanced by McClung, Wilson and Stevens. Bryn Mawr College May 4, 1607 S permatogenesis 509 BIBLIOGRAPHY BauMGARTNER, W. J., ’02—Spermatid Transformation. Kans. Univ. Sci. Bull., ey oe ‘o4—_Some New Evidences for the Individuality of the Chromosomes. Biol. Bull., vin, p. 1. vAN BENEDEN, E.,’83—Maturation de |’Oeuf, la Fecondation et la Division Cellu- lave. . Paris. Berry, E. H., ’06—The ‘“‘Accessory Chromosome” in Epeira. Biol. Bull., x1, P- 193. Biackman, M. W., ’o1—Spermatogenesis of Myriapods. I Notes on Spermato- cytes and Spermatids of Scolopendra. Kans. Univ. Quart., x, p. 61. ’03—Spermatogenesis of Myriapods. II] On the Chromatin in the Spermatocytes of Scolopendra heros. Biol. Bull., v, p. 187. o5—Spermatogenesis of Myriapods. III] Spermatogenesis of Scolo- pendra heros. Bull. Mus. Comp. Zool., Harvard, xlvui, p. 1. ’o5b—Spermatogenesis of Myriapods. IV On the Karyosphere and Nucleolus in the Spermatocytes of Scolopendra subspinipes. Proc. Am. Acad. Arts and Sci., xli, p. 331. Bovert, T., ’°87—Zellen-Studien I. Jena. *88—Zellen-Studien II. Jena. Butscuty, O., ’71—Vorlaufige Mittheilung iber Bau und Entwicklung der Samen- faden der Insekten und Crustaceen. Z. wiss. Zool., xxi, p. 402. CastLE, W. E., ’03—The Heredity of Sex. Bull. Mus. Comp. Zeol., Harvard, xp: LOO: Conkuin, E. G., ’02—Karyokinesis and Cytokinesis. Jour. Acad. Nat. Sci., Philaz, xr, p. 1. Foot, K. anp E. C. Srroseii, ’07—The “Accessory Chromosome’’ of Anasa tristis. Biol. Bull., xu, p. 119. Gross, J., °04—Die Spermatogenese von Syromastes marginatus. Zool. Jahrb. f. Anat. u. Ont., xx, p. 439. ’o6—Die Spermatogenese von Pyrrhocoris apterus. Zool. Jahrb., xxii, p- 269. ; Henxinc, H., ’90a—Untersuchungen iiber die ersten Entwicklungsvorgange in den Eiern der Insekten. II Ueber Spermatogenese und deren Beziehung zur Eientwicklung bei Pyrrhocoris apterus. Z. wiss. Zool., li, p. 685. *gob—Ueber Reduktionstheilung der Chromosomen in den Samenzellen von Insekten. Internat. Monatsschr. f. Anat. u. Phys., vil, p. 243. Hertwic, O., ’90—Vergleich der Ei- und Samenbildung bei Nematoden. Arch. f. Mikr. Anat., xxxvi, p. I. 510 Alice M. Boring Mark, E. L., °81—Maturation, Fecundation and Segmentation of Limax cam- pestris. Bull. Mus. Comp. Zool., Harvard, vi, p. 173. McC ung, C. E., *98—A Peculiar Nuclear Element in Male Reproductive Cells of Insects. Zool. Bull., 11, p. 187. *oo—Spermatocyte Divisions of the Acrididz. Kans. Univ. Quart., ix, P-/3- *o1—Notes on the Accessory Chromosome. Anat. Anz., xx, p. 220. °o2a—Spermatocyte Divisions of the Locustide. Kans. Univ. Quart., x1, p- 185. °o2b—Accessory Chromosome—Sex Determinant? Biol. Bull., i, p. 43. °05—The Chromosome Complex of Orthopteran Spermatocytes. Biol. Bull., ix, p. 304. McGr11, C., °04—Spermatogenesis of Anax junius. Univ. Missoun Stud., II, 5. Mepes, G., ’05—Spermatogenesis of Scutigera forceps. Biol. Bull., ix, p. 156. Montcomery, T. H., ’97—A Preliminary Note on the Chromatin Reduction in Spermatogenesis of Pentatoma. Zool. Anz., xx, p. 457. *g8—Spermatogenesis of Pentatoma. Zool. Jahrb., xii, p. 1. *gg—Chromatin Reduction in Hemiptera. Zool. Anz., xxii, p. 76. ‘oo—Spermatogenesis of Peripatus. Zool. Jahrb., xiv, p. 277. ‘o1ra—A Study of the Chromosomes of the Germ Cells of Metazoa. Trans. Am. Phil. Soc., xx, p. 153. *orb—Further Studies on the Chromosomes of the Hemiptera Heterop- tera. Proc, Acad. Nat. Sci., Phila., lit, p. 261. *04—_Some Observations and Considerations on the Maturation Phenom- ena of Germ Cells. Biol. Bull., vi, p. 137. °o5—Spermatogenesis of Syrbula and Lycosa. Proc. Acad. Nat. Sci., Phila., lvii, p. 162. °o6a—Terminology of Aberrant Chromosomes and their Behavior in certain Hemiptera. Science, n. s., xxiii, p. 36. *o6b—Chromosomes in the Spermatogenesis of the Hemiptera Heterop- tera. Trans. Am. Phil. Soc., n. s., xxi, p. 97. Moore AnpD Rosinson, ’07—On the Behavior of the Nucleolus in the Spermato- genesis of Periplaneta americana. Quart. Jour. Mic. Sci., xlvii, p- 571- Now ttn, W. N., ’o6—A Study of the Spermatogenesis of Coptocycla aurichalcea and Coptocycla guttata, with especial reference to the Problem of Sex Determination. Jour. Exp. Zodl., 11, p. 583. Orre, H., ’o6—Samenreifung und Samenbildung von Locusta viridissima. Zool. ANZ; SXX,-p-.520- PauimierR, F. C., ’98—Chromatin Reduction in Hemiptera. Anat. Anz., xiv, p- 514. S permatogenesis 511 Pautmigr, F. C., ’99—Spermatogenesis of Anasa tristis. Jour. Morph., xv, sup- plement, p. 223. Pratn_er, G., ’86—Die Karyokinese bei den Lepidoptera als Grundlage fiir eine Theorie der Zelltheilung. Internat. Monatsschr. f. Anat. u. Histol., lll, p. 341. vom Ratu, ’91—Ueber die Reduktion der Chromatischen Elemente in der Samen- bildung von Gryllotalpa. Ber. d. Naturf. Ges. Freiburg, vi, p. 62. *g2—-Zur Kenntniss der Spermatogenese von Gryllotalpa vulgaris. Arch. f. Mikr. Anat., xl, p. 102. °95—Neue Beitrage zur Frage der Chromatinreduktion in der Samen- und Eireife. Arch. f. mikr. Anat., xlvi, p. 168. *g4—Ueber die Konstanz der Chromosomenzahl bei Thieren. Biol. Cen- tralbl., xiv, p. 449. ; SapaTIER, A., ’g90—De la Spermatogénése chez les Locustides. Compt. Rend. Acad., Paris, cxi, p. 797. DE StnETy, R. ’o1—Recherches sur la Biologie et Anatomie des Phasmes. La Cellule; xix, p= 117: StrEeveENS, N. M., ’o5a—A Study of the Germ Cells of Aphis rose and Aphis ceno- there) sours xp. Zool;,:11; p. 313: °o5b—Studies in Spermatogenesis with especial reference to the “‘ Acces- sory Chromosome.” Carnegie Ins., Wash., pub. 36. *o6a—Studies on the Germ Cells of Aphids. Carnegie Ins., Wash., pub. Rie o6b—Studies in Spermatogenesis II. A Comparative Study of the Het- erochromosomes in Certain Species of Coleoptera, Hemiptera, and Lepidoptera, with especial reference to Sex Determination. Car- negie Ins., Wash., pub. 36, 11. Sutton, W. S., ’oo—Spermatogonial Divisions in Brachystola magna. Kans. Univ. Quart., ix, p. 135. *o2—Morphology of the Chromosome Group in Brachystola magna. Biol. Bull., iv, p. 24. °03—Chromosomes in Heredity. Biol. Bull., iv, p. 231. vy. La VALETTE St. GEORGE, 85—Spermatologische Beitrage 1. Arch. f. mikr. Anat... %XV, p- 501. Verson, E., ’8g—Zur Spermatogenese. Zool. Anz., xii, p. 100. °94—Zur Spermatogenese bei der Seidenraupe. Z. wiss. Zool., lviii, Pp: 303. Voinoy, D. N., ’03—Spermatogénése d’été chez le Cybister roeselei. Arch. d. Zool. Exp. et Genl., 4 series, 1, p. 173. 512 Alice M. Boring Vornoy, D. N.,’04—Sur une Disposition special a la Chromatin dans la Spermato- génese du Gryllus campestris Reproduisant des Structures Observés seulement dans |’Ovogénése. Arch. d. Zool. Exp. et Genl., ¥ series, ii, Notes et Revues, p. 63. Wa tace, L. B., °0o—The Accessory Chromosome in the Spider. Anat. Anz., XVIll, Pp. 327. °o5—Spermatogenesis of the Spider. Biol. Bull., viii, p. 169. WEIsMANN, A., ’87—On the Number of Polar Bodies and their Significance in Heredity. Essays on Heredity. Oxford. Wicox, E. V., ’94—Spermatogenesis of Caloptenus femur-rubrum. Preliminary Notice. Anat. Anz., x, p. 303. ’95—Spermatogenesis of Caloptenus femur-rubrum and Cicada tibicens. Bull. Mus. Comp. Zool., Harvard, xxvii, p. 3. *96—Further Studies on the Spermatogenesis of Caloptenus femur- rubrum. Bull. Mus. Comp. Zool., Harvard, xxix, p. 193. *97—Chromatic Tetrads. Anat. Anz., xiv, p. 194. Witson, E. B., ’°05a—Chromosomes in Relation to the Determination of Sex in Insects. Science, n. s., xx, p. 500. ’o5b—Studieson Chromosomes. I ‘The Behavior of the Idiochrosomes inthe Hemiptera. Jour. Exp. Zodl., i, p. 371. ’o5c—Studies on Chromosomes. II The Paired Microchromosomes, Idiochromosomes, and Heterotropic Chromosomes in the Hemiptera. Jour. Exp. Zool., 11, p. 507. ’o6a—Studies on Chromosomes. III Sexual Differences of the Chro- mosome Groups in Hemiptera, with some Considerations on Deter- mination and Inheritance of Sex. Jour. Exp. Zodl., 1, p. I. ‘o6b—A New Theory of Sex Production. Science, n. s., xxill, p. 189. °07—The case of Anasa tristis. Science, n. s., xxv, p. IgI. ZweIGER, H., ’o6—Die Spermatogenese von Forficula auricularia. Zool. Anz., XXX, Pp. 220. © ¥ DESCRIPTION OF PLATES. The figures were drawn with the aid of the Zeiss-Abbe drawing camera, No. 111. Figs. 1-46 were drawn with a Leitz oil immersion obj. 7'z and a Zeiss compensating oc. 12, Figs. 47-313 with a Zeiss apo- chromatic oil immersion obj. 2 mm., oc. 12.. They have been reduced one-third, giving a magnification of about 1000 diameters. Abbreviations Used on Plates a and a= one pair of spermatogonial chromosomes. a= a bivalent primary spermatocyte chromosome representing a and ap. b; and bo= one pair of spermatogonial chromosomes. b = a bivalent primary spermatocyte chromosome representing },and bp. m, and mz = a pair of m-chromosomes. n = chromatin nucleolus. Pp; pis p2 = plasmosomes. x = odd chromosome. Fig. 1 Bip. | 2 Fig. se 194°) Xo SONY CN Un Fig. 10 Fig. 15 Fig. 16 Pirate I Entilia sinuata (Family Membracide) Spermatogonial rest stage. Spermatogonial spireme. Spermatogonium, segmentation of the spireme, each segment longitudinally split. Spermatogonium, condensation of the segments of the spireme. Spermatogonial equatorial plate, 21 chromosomes. Spermatogonial metaphase. Spermatogonial anaphase. Spermatogonial telophase, formation of nuclear membrane. Spermatogonial telophase, polar view. First spermatocyte, contraction stage. First spermatocyte, early postsynapsis stage. First spermatocyte, late postsynapsis, fine spireme. First spermatocyte, coarse spireme. First spermatocyte, rest stage. First spermatocyte, split spireme. First spermatocyte, spireme divided into 11 split segments. Figs. 17-19 First spermatocyte, early prophase, tetrad formation. Fig. 20 Fig. 21 Fig. 22 First spermatocyte, late prophase, dumb-bell formation. First spermatocyte, equatorial plate, 11 chromosomes. First spermatocyte, metaphase, chromosomes still tetrads. Figs. 23,24 First spermatocyte, metaphase. Figs. 25,26 First spermatocyte, anaphase, centrosomes divided for the second division. Fig. 27 Fig. 28 Fig. 29 Fig. 30 First spermatocyte, telophase. First spermatocyte, telophase, polar view. Rearrangement of chromosomes for the second spermatocyte division. Second spermatocyte, equatorial plate, 11 chromosomes. ———————— % ai eR eer ol A SPERMATOGENESIS PLATE I Auice M. Borinc MEMBRACID . a. HE JournaL or ExperimeNntTaL ZobLoGy, VoL. Iv, No. 4 Pirate II Entilia sinuata (continued) Fig. 31 Second spermatocyte, equatorial plate, 10 chromosomes. Fig. 32 Second spermatocyte, metaphase. Fig. 33 Second spermatocyte, anaphase. Fig. 34 Second spermatocyte, anaphase, polar view. Fig. 35 Second spermatocyte, telophase. Fig. 36 Spermatid, first stage. Figs. 37, 38 Figs. 39, 40 Figs. 41, 42 Figs. 43, 44 _ Figs. 45, 46 Spermatids, second stage, half with x, half without. Spermatids, third stage, half contain x, half do not. Spermatids, formation of axial filament, half contain x, haif do not. Spermatids, condensation of the chromatin, half contain x, half do net. Spermatids, later stages. Fig. 47. Mature spermatozoon. Vanduzea arcuata (Family Membracide) Fig. 48 Spermatogonial equatorial plate, 17 chromosomes. Figs. 49, 50 First spermatocyte, growth period. Fig. 51 First spermatocyte, equatorial plate, g chromosomes. Fig. 52 First spermatocyte, metaphase. Figs. 53-55 First spermatocyte, anaphase. Figs. 56,57 Second spermatocytes, equatorial plates, containing 9 and 8 chromosomes, respectively. Fig. 58 Second spermatocyte, metaphase. Fig. 59 Second spermatocyte, anaphase. Figs. 60, 61 Spermatids, half contain x, half do not. SPERMATOGENESIS Atice M. Borinc PLATE! . - $ are joao Sona uh sais AN pg Lirepebtes aay nano A. M. B. del. MEMBRACID HE JourNAL or ExprrIMENTAL ZOOLOGY, VOL. IV, NO. 4 Pirate III Ceresa taurina (Family Membracide) Figs. 62, 63 First spermatocyte contraction stage, a mass of rejected basichromatin at the base of the loops. Figs. 64-66 First spermatocyte, rest stage, showing rejected basichromatin. Fig. 67 First spermatocyte, rest stage, showing x in the midst of the rejected chromatin. Fig. 68 First spermatocyte, split spireme stage. Most of the rejected chromatin has dissolved, showing x plainly. Fig. 69 First spermatocyte, prophase. Fig. 70 First spermatocyte, equatorial plate, 11 chromosomes. Figs. 71-73 First spermatocyte, metaphase. Figs.74,75 First spermatocyte, anaphase. : Figs. 76, 77 Second spermatocyte, equatorial plates, containing 11 and 10 chromosomes, respectively. Fig. 78 Second spermatocyte, metaphase. Fig. 79 Second spermatocyte, anaphase. Fig. 80 Spermatid, with chromatin nucleolus. Ceresa bubalus (Family Membracide) Fig. 81 First spermatocyte, synapsis stage, showing rejected chromatin. Fig. 82 First spermatocyte, rest stage, showing rejected chromatin. Fig. 83 First spermatocyte, equatorial plate, 11 chromosomes. Figs. 84, 85 First spermatocytes, metaphase. Figs. 86,87 First spermatocytes, anaphase. Figs. 88, 89 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Fig. go Second spermatocyte, metaphase. Fig. 91 Second spermatocyte, anaphase. Fig. 92 Spermatid, with chromatin nucleolus. Ceresa diceros (Family Membracide) Fig. 93 First spermatocyte, rest stage, showing rejected chromatin. Fig. 94 First spermatocyte, equatorial plate, 11 chromosomes. Fig. 95 First spermatocyte, metaphase. Fig. 96 First spermatocyte, anaphase. Figs. 97, 98 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Fig. 99 Second spermatocyte, metaphase. Fig. roo Second spermatocyte, anaphase. Fig. tor Spermatid with chromatin nucleolus. Atymna castanea (Family Membracide) Fig. 102 First spermatocyte, rest stage. Fig. 103 First spermatocyte, equatorial plate, 11 chromosomes. SPERMATOGENESIS PLATE II Auice M. Borinc 97 | 101 A. M. B. del. MEMBRACID® E Journat or ExperimENTAL ZoGLoGy, VOL. IV, No. 4 Pirate IV Atymna castanea (continued) Figs. 104, 105 First spermatocyte, metaphase. Fig. 106 First spermatocyte, anaphase. Figs. 107, 108 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Figs ively. Fig. Fig. 109 Second spermatocyte, metaphase. 110 Second spermatocyte, anaphase. Ii! 112 113 114 115 Spermatid, with chromatin nucleolus. Campylenchia curvata (Family Membracide) Spermatogonial equatorial plate, 19 chromosomes. First spermatocyte, rest stage. First spermatocyte, equatorial plate, 10 chromosomes. First spermatocyte, metaphase. 116 First spermatocyte, anaphase. - 117,118 Second spermatocytes, equatorial plates, containing 10 and g chromosomes, respect- 119 Second spermatocyte, metaphase. 120 Second spermatocyte, anaphase. Figs. 121,122 Spermatids, half with x, half without. Fig. Fig. Fig. Fig. Fig. 123 Enchenopa binotata (Family Membracide) Spermatogonial equatorial plate, 19 chromosomes. 124 First spermatocyte, spireme stage. 125 126 127 First spermatocyte, early prophase. First spermatocyte, equatorial plate, 10 chromosomes. First spermatocyte, equatorial plate, 11 chromosomes, occasionally found. Figs. 128,129 First spermatocytes, metaphase. Fig. Fig. 130 131 First spermatocyte, anaphase. Second spermatocyte, equatorial plate, 10 chromosomes. Figs. 132, 133 Second spermatocytes, metaphase. Fig. Fig. Fig. in one, Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 33 x does not divide in this division. 134 Second spermatocyte, anaphase. 135a and b Second spermatocyte anaphase, two plates from the same spindle, 9 chromosomes 10 in the other. 136 Spermatid, with chromatin nucleolus. Chlorotettrix unicolor and Chlorotetirix vias (ami Fasside) Spermatogonial rest stage. Spermatogonial metaphase. Spermatogonial anaphase. First spermatocyte, contraction stage. First spermatocyte, spireme stage. First spermatocyte, rest stage. First spermatocyte, split spireme stage. First spermatocyte, prophase. First spermatocyte, equatorial plate, 9 chromosomes. First spermatocyte, metaphase. SPERMATOGENESIS Atice M. Borinc PLATE IV a os os vitae / 2% 0.x coe ( S85 See 141 142 145 MEMBRACIDZ AND JASSID SHE JourNAL or ExperiMENTAL Zo6LoGyY, VOL. IV, No. 4 4 ‘we... 146 A. M. B. del. Pirate V Chlorotettrix unicolor and Chlorotettrix vividus (continued) Fig. 147 First spermatocyte, anaphase. Figs. 148, 149 Second spermatocytes, equatorial plates, containing 8 and 9 chromosomes, respect- ively. Fig. 150 Second spermatocyte, metaphase. Fig. 151 Second spermatocyte, anaphase. Fig. 152 Spermatid, first stage. Figs. 153,154 Spermatid, second stage, half with x, half without. Figs. 155,156 Spermatid, third stage, half with x, half without. Figs. 157,158 Late spermatid stages. Fig. 159 Head of mature spermatozoon. Fig. 160 Spermatogonial equatorial plate, 21 chromosomes. Fig. 161 First spermatocyte equatorial plate, 11 chromosomes. Figs. 162, 163 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Diedrocephala coccinea (Family Fasside) Fig. 164 Spermatogonial equatorial plate, 23 chromosomes. Fig. 165 First spermatocyte, postsynapsis stage. Fig. 166 First spermatocyte, rest stage: Fig. 167 First spermatocyte, equatorial plate, 12 chromosomes. Fig. 168 First spermatocyte, metaphase. Fig. 169 First spermatocyte, anaphase. Figs. 170,171 Second spermatocytes, equatorial plates, containing 12 and 11 chromosomes, respect- ively. Fig. 172 Second spermatocyte, metaphase. Fig. 173, Second spermatocyte, anaphase. Figs. 174,175 Spermatids, half without x, half with. Diedrocephala mollipes (Family Fasside) Fig. 176 First spermatocyte, rest stage. Fig. 177 First spermatocyte, equatorial plate, 12 chromosomes. Fig. 178 First spermatocyte, metaphase. Fig. 179 First spermatocyte, anaphase. Figs. 180, 181 Second spermatocytes, equatorial plates, containing 12 and 11 chromosomes, respect- ively. Fig. 182 Second spermatocyte, metaphase. " Fig. 183 Second spermatocyte, anaphase. 5 Figs. 184,185 Spermatids, half without x, half with. SPERMATOGENESIS PLATE,V Auice M. Borinc 155 156 157 158 od 9) Gh «(Sa)» 180 181 182 183 184 185 A.M. B. del. JASSIDE Tue Journar or ExperimENTAL ZoOLoGY, VoL. IV, NO. 4 Prate VI Phlepsius irrotatus (Family Fasside) Fig. 186 Spermatogonial equatorial plate, 15 chromosomes. Fig. 187 First spermatocyte, rest stage. Fig. 188a andb First spermatocyte, equatorial plate, and the odd chromosome x. Fig. 189 First spermatocyte, metaphase. Fig. 190 First spermatocyte anaphase. Figs. 191, 192 Second spermatocytes, equatorial plates, containing 8 and 7 chromosomes, respect- ively. Fig. 193 Second spermatocyte, metaphase. Fig. 194 Second spermatocyte, anaphase. Figs. 195,196 Spermatids, half without x, half with. Agallia sanguinolenta (Family Fasside) Fig. 197 First spermatocyte, spireme stage. Fig. 198 First spermatocyte, equatorial plate, 11 chromosomes. Fig. 199 First spermatocyte, metaphase. Figs. 200, 201 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Fig. 202 Second spermatocyte, early anaphase. Figs. 203, 204 Spermatids, half without x, half with. Fig. 205 First spermatocyte, equatorial plate, aceto-carmine preparation. Fig. 206 First spermatocyte, anaphase, aceto-carmine preparation. Fig. 207 Second spermatocyte, equatorial plate, aceto-carmine preparation. Clastoptera sbtusa (Family Cercopide) Fig. 208 Spermatogonial rest stage. Fig. 209 Spermatogonial prophase. Fig. 210 Spermatogonial equatorial plate, 15 chromosomes. Fig. 211 Spermatogonial metaphase. Fig. 212 Spermatogonial anaphase. Figs. 213,214 First spermatocyte, early synapsis. P Fig. 215 First spermatocyte, contraction stage. Fig. 216 First spermatocyte, postsynapsis stage. Fig. 217 First spermatocyte, spireme stage. Fig. 218 First spermatocyte, early prophase, tetrad formation. Fig. 219 First spermatocyte, prophase, dumb-bell formation. , ?e Cae Fig. 220 First spermatocyte, equatorial plate, 8 chromosomes. - Fig. 221 First spermatocyte, metaphase. oo Fig. 222 First spermatocyte, anaphase. Figs. 223, 224 Second spermatocytes, equatorial plates containing 8 and 7 chromosomes, respect- a4 ively. Fig. 225 Second spermatocyte, metaphase. Xe SPERMATOGENESIS Atice M. BorinG JASSIDZZ AND CERCOPID/ Tue Journat or ExperiIMENTAL ZOOLOGY, VOL. IV, NO+ 4 PLATE VI 202 213 A. M. B. del. Pirate VII Clastoptera obtusa (Continued) Fig. 226 Second spermatocyte, anaphase. Fig. 227 Early spermatid, with chromatin nucleolus, Fig. 228 Spermatid, formation of axial filament. Figs. 229,230 Later spermatids. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 231 232 233 234 235 236 227) Mature spermatozo6n. A phrophora quadrangularis with 11 chromosomes (Family Cercopide) Spermatogonial equatorial plate, 21 chromosomes. First spermatocyte, contraction stage. First spermatocyte, spireme stage. First spermatocyte, equatorial plate, 11 chromosomes. First spermatocyte, metaphase. First spermatocyte, anaphase. Figs. 238, 239 Second spermatocytes, equatorial plates, containing 11 and 10 chromosomes, respect- ively. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 240 241 242 243 244 245 246 247 248 249 250 251 252 Second spermatocyte, metaphase. * Second spermatocyte, anaphase. Spermatid, with chromatin nucleolus. A phrophora quadrangularis with 12 chromosomes (Family Cercopide) First spermatocyte, contraction stage. First spermatocyte, equatorial plate, 12 chromosomes. First spermatocyte, metaphase. First spermatocyte, anaphase. Second spermatocyte, equatorial plate, 12 chromosomes. Second spermatocyte, anaphase. Aphrophora 4-notata (Family Cercopide) First spermatocyte, spireme stage. First spermatocyte, equatorial plate, 14 chromosomes. First spermatocyte, metaphase. First spermatocyte, anaphase. Figs. 253,254 Second spermatocytes, equatorial plates, containing 14 and 13 chromosomes, respect- ively. Fig. 255 Second spermatocyte, anaphase. SPERMATOGENESIS PLATE VII Atice M. BorinGc 237 238 239 240 241 247 A. M. B. de. CERCOPID& Tue Journar of EXPERIMENTAL ZOOLOGY, VOL. IV, NO. 4 Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Prate VIII Peciloptera septentrionalis (Family Fulgorida) 256 Spermatogonial rest stage. 257 Spermatogonial split spireme. 258 Spermatogonium, spireme segmented and condensed, segments split. 259 Spermatogonial equatorial plate, 27 chromosomes. 260 Spermatogonial metaphase. 261 Spermatogonial anaphase. 262 First spermatocyte, early synapsis stage. 263 First spermatocyte, contraction stage. 264 First spermatocyte, spireme stage. Figs. 265,267 First spermatocyte, rest stages, growth in size of nucleus and cell. Fig. Fig. Fig. Fig. Fig. 268 First spermatocyte, split spireme stage. 269 First spermatocyte, prophase, tetrad formation. 270 First spermatocyte, equatorial plate, 14 chromosomes. 271 First spermatocyte, metaphase. 272 First spermatocyte, anaphase. Figs. 273, 274 Second spermatocytes, equatorial plates, containing 14 and 13 chromosomes, respect- ively. Fig. Fig. 275 Second spermatocyte, metaphase. 276 Second spermatocyte, anaphase. Figs. 277,278 Spermatids, formation of fibers in the ‘‘Nebenkern.” Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 279a Spermatid, ‘“Nebenkern” separated by a partition into two tubes. 279b Cross section of ‘‘ Nebenkern” structure as in 279. 280 Spermatid, elongation of fibers and tubes. 281a Spermatid, irregular spiral of twisted tubes. 281b Cross sections of tubes of 281a. 282a Spermatid, further twisting and flattening. 282b Cross section of 282a. 283 Female somatic equatorial plate, 28 chromosomes. — SPERMATOGENESIS PLATE VIII Auice M. Borinc FULGORID Aa Tue Journat or ExPEerIMENTAL ZOOLOGY, VOL. IV, No. 4 - Prater IX Peciloptera pruinosa (Family Fulgoride) Fig. 284 First spermatocyte, rest stage. Fig. 285 First spermatocyte, equatorial plate, 14 chromosomes. Fig. 286 First spermatocyte, metaphase. ‘ Fig. 287 First spermatocyte, anaphase. Figs. 288, 289 Second spermatocytes, equatorial plates, containing 14 and 13 chromosomes, respect- ively . Fig. 290 Second spermatocyte, metaphase. Fig. 291 Second spermatocyte, anaphase. Figs. 292,293 Spermatids, half with x, half without. Fig.294 Female somatic equatorial plate, 28 chromosomes. Amphiscepa bivittata (Family Fulgoride) Fig. 295 Spermatogonial equatorial plate, 25 chromosomes. Fig. 296 First spermatocyte, rest stage. Fig. 297 First spermatocyte, equatorial plate, 13 chromosomes. Fig. 298 First spermatocyte, metaphase. Fig.299 First spermatocyte, anaphase. Figs. 300, 301 Second spermatocytes, equatorial plates, containing 13 and 12 chromosomes, respect ively. Fig. 302 Second spermatocyte, metaphase. Fig. 303 Second spermatocyte, anaphase. Fig. 304 Spermatid. : Peciloptera bivittata (Family Fulgoride) Fig. 305 First spermatocyte, rest stage. Fig. 306 First spermatocyte, equatorial plate, 13 chromosomes. Fig. 307 First spermatocyte, metaphase. Fig. 308 First spermatocyte, anaphase. Figs, 309,310 Second spermatocytes, equatorial plates, containing 13 and 12 chromosomes, respectively. Fig. 311 Second spermatocyte, metaphase. Fig. 312 Second spermatocyte, anaphase. Fig. 313, Spermatid. * Ss on ae » w le ae SPERMATOGENESIS PLATE IX Auice!M. Borinc A. M. B. det. FULGORID JHE Journat or ExperiIMENTAL ZoOLoGy, VOL. IV, NO. 4 q ah (oe ‘> CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. E. L. Mark, Director. No. 191. THE REACTIONS OF THE POMACE FLY, DROSO- PalitA AMPELOPHIDA LOEW, TO ODOROUS SUBSTANCES WILLIAM MORTON BARROWS With Five Ficures i Tnnaclici@hisy tage co coon pe ads Seolncson ete o SOO poet Oo An ees Gneme Comm notion omnia 515 [nl LUE SIGNISS aoa 66c0udgducuneod mop od odd pbonD Oo sod od pompoGn eo obdoound sonmminn oso 516 1 Preliminary experiments. .........- So SBOE Sc tos Gas Teme COCR ECM n cre aan ir 516 2 Experiments with alcohol, acetic acid and acetic ether............ 0. + esse eee eee e eee 519 3 Experiments on the directive effects of odorous substances...........-22.-++eeeeeeee 527 4 Experiments to determine che position and function of the olfactory sense organs.........530 MNO HED CCCI CISCUSSION «\ cemomn diac © 536 WY DISIG GER coer ene cnr ede bone Dar SOOSe Send belo manta Doc een actana hhcc. SODed ean 537 I INTRODUCTION Drosophila ampelophila is a small fly about three millimeters in length belonging to the family Drosophilide. It lays 1s eggs on fermenting fruit, which serves as food for both the larve and the adults. The ease with which large numbers of these insects can be reared in the laboratory during the winter as well as the summer, and the definiteness with which they react to many forms of stimuli, make them favorable subjects for experimentation. Since they find their food with great certainty even in the dark, a habit that seemed to involve the sense of smell, I was led to take up an investigation of their reactions to odorous substances. Where the flies were abundant, it was noticed that they often entered bottles and other receptacles containing alcohol. ‘The fact that the fermenting fruit upon which they feed is continually generating alcohols and other related compounds, led me to sus- pect that it was these substances that served to attract the flies, and that they therefore probably presented a clear case of chemotro- pism among air-inhabiting animals. Tue JouRNAL or ExPeRIMENTAL ZOOLOGY, VOL. IV, NO. 4. 516 William Morton Barrows The experiments recorded in this paper were undertaken to determine, if possible, first, to what substances Drosophila is chemotropic, and secondly, in what way the fly finds its food. The work was carried on in the Zodlogical Laboratory of Har- vard University, under the direction of Prof. W. E. Castle and Prof. G. H. Parker, to whom I am indebted for much valuable advice and careful criticism. II EXPERIMENTS 1 Preliminary Ex periments For the preliminary experiments, which were planned to ascer- tain whether certain substances were stimulating or not for the flies, the following apparatus was devised. A two-dram vial closed at the end by a cork stopper was arranged as atrap. Piere- ing the stopper and reaching nearly to the bottom of the vial was a glass tube with a caliber of about two millimeters (Fig. 1). The bore of the tube was large enough to allow a fly to creep through easily and yet small enough to make it difficult for the fly to turn around after having once started into the tube. The tube pro- jected beyond the cork on the exterior about three millimeters. If a fly once got halfway down the tube leading into the vial, the chance of its backing out or finding its way out later was very small About 1 cc. of the substance to be tested was placed on a piece of filter paper in the vial. Five vials thus charged with substances to be tested usually formed the set of traps. At least one of these was always used as a control in that it contained filter paper wet with distilled water only or with some other material used as a check. The traps, with their open ends directed toward the light, were placed in a vertical glass cylinder 20.5 cm. high and 17.5 cm. in diameter. ‘The bottom edge of the cylinder rested on a sheet of clean filter paper and the top was closed by a glass plate. The atmosphere in the cylinder was kept moist by the evaporation of distilled water exposed in a small vessel Many hungry flies, usually one hundred, were liberated in the glass cylinder and left there for twenty-four hours. By hungry flies is meant those which iis Reactions of the Pomace Fly to Odorous Substances 517 had been supplied with distilled water but had been kept from food for twenty-four hours. If they are kept without food much longer than this, they begin to die and few survive, sixty hours. After the flies had been allowed twenty-four hours in which to enter the traps, the experiment was discontinued and the indi- viduals in each trap were counted. In this way some idea of the influence of different substances on the movements of the flies could be ascertained. These experiments were preliminary, in that they aimed only to determine what substances called forth positive reactions. ‘The Fig. 1 Vial arranged as a trap. results, which are necessarily fragmentary, because of the difficulty of dealing with odors in a quantitative way, are given in Table I. From Table [ it is apparent that of the ten substances which were tested singly in aqueous solutions, the flies gave definite positive reactions to four, namely amyl, and sie ethyl alcohol, acetic and lactic acid (Experiments 1, 2, 3, 5,6). The remaining substances to which the flies did not react positively or did not react in numbers large enough to be significant, are propyl alcohol, butyric acid, valerianic acid, glucose, amyl valerianate and acetic ether. Experiments which were undertaken subse- 518 William Morton Barrows TABLE I Numbers of flies out of one hundred, which, in each of eight experiments, entered the several traps charged with odorous substances. The duration of each experiment was twenty-four hours No. of flies in No. of Experiment Substance in the trap trap { i “ACCHC agi 4 Per Clb... soem Fane ore oe EE ee ee ° | | 2: Aleohol-io/ percent. sc nko sn. faase sos e ne sree 4 WAC sess ste ree ake 4.1) 3: Water:(conttrol) sca. 3.5. tknlnts anon a ween aa | I | 4 Acetic acid 4 per cent with trace of acetic ether.......... 5 U § Glucose and‘ water-s a. sc. ce seine tee ee Ok cee eee : ° (| 1 Acetic acid 4 PERCENEs 5. ee oc a ers ee I I) 2. Menhel soipes cedh-eoe-. ans ers tod eae a eet eee 6 Bis 11-33 Water (contrel)).s) 36 nage see oe ee i eee 4 Acetic acid 4 per cent with trace of acetic ether. ..... aaa ee 61 | “5: Glucose:and Waters. : .- 32. J-22 ues asec oe occ eo ee | ° | 1 Alcohol 1o per cent containing acetic ether .o4 per cent... . . 85 | 2°Alcohol roiper cents fie~ ease ce as oe sion = sence on See 12 ze ‘‘heg” Water (control) 2 2 ood ce wyerew be acs oee eee 4 Acetic acid/4percentycsce- Ge ccte cee oe eee 2 \| 5 Acetic acid 4 per cent containing acetic ether .o4 per cent... .| 19 (| 1 Alcohol ro per cent with trace of butyric acid.............. 43 | 2 Alcohol ro per cent containing 2 per cent hydrochloric acid. | ° y ee Bi iscie se ee | 3 Alephol 10 percent (eontrel) © 52.2.2... 2. ae eee 26 | | 4 Alcohol 10 per cent with trace of valerianic acid............ 14 5 Alcohol ro per cent with trace of isoamy] acetate........... | ° 1 Water with trace of valerianic acid...................--- / ° | 2 Butyrieacid:2per cents <5 2.2 jos. ce techs eee ee | I } Lpepanebecencorcace? | 3. Water (control) -s-5.. cco son. ou ieee orem ea ee eee 2 4 Propy!l alcohol 2/per cent: 25-25. 3-52.-\. 26 = J oe ° 5 Water with a trace of amyl valerianate..................- ° f i Water with a trace of aceic ethers. -- 222.) eee eee eS } juambactic acide, percent...5..-2 «. -c.5- a at eee 6 ee ) 2) Water (control) 222s: seis en 20 oes aie ee ° [ 4 Amylaleohol2 percent: .- 2.02.6. siqo- ery oe ee 3 1 Alcohol ro per cent with a trace of acetacetic ester ......... | ° . 2 Alcohol ro per cent with a trace of isobutyl acetate ........ 9 "' Dee ees seen ence ee 3, Alcohol|xo'per.cent (control): .. << 3... Svcs te eel 5 4 Alcohol to per cent with a trace of methyl acetate.......... 29 a |. 5 Alcohol ro per cent with a trace of isoamyl acetate......... 8 : * 1 Alcohol ro per cent containing .o5 per cent osmic acid...... | 2 i; 2 Alcohol 1o per cent containing 2 per cent hydrochloric acid. ° Dire Soret acla ee ee te 3 Alcohol'ro. percent. (control): ba. 9-0. = nae et ee ee I 4 Alcohol ro per cent containing 2 per'cent nitric acid........ ° 5 Alcohol ro per cent containing 2 per cent acetic acid........ 8 Reactions of the Pomace Fly to Odorous Substances 519 quent to those under discussion, however, showed that the flies did enter or attempted to enter traps which contained acetic ether in aqueous solution, so that it is possible that under other conditions the flies might be positive to some of the substances to which in these preliminary experiments they seemed not to be. In Experiment 2 one trap was charged with 4 per cent acetic acid and another trap with 4 per cent acetic acid containing a small amount of aceticether. “he number of flies found in the trap which contained the mixture of acid and ether greacly exceeded that in the trap charged with acid only. A similar contrast was observed in Experiment 3. ‘his increase of flies in the trap containing the mixture 1s evidently due to the acetic ether. A similar phenomenon was seen when to 10 per cent ethyl alcohol, a small amount of acetic ether, isobutyl acetate, or methyl acetate was added (Experi- ments 3 and 7). Isoamyl acetate may possibly also be classed with these substances, though the results of Experiments 4 and 7 do not show an entire agreement in this respect. An increase was also obtained when acetic or butyric acids were added in small amounts to the alcohol. “This property may be slightly shared by valerianic acid and possibly by osmic acid, but it is certainly not a characteristic of hydrochloric and nitric acids which seem to be strongly repellent (Experiments 4 and 8). - All the organic substances tested in these preliminary experi- ments are found in fermenting fruits and the test conditions which gave the highest positive numerical results are probably those which simulated most closely the natural optimum conditions. 5 Experiments with Alcohol, Acetic Acid and Acetic Ether As a more complete analysis of the effect of some of the odorous substances used in the preliminary experiments was desirable, it was decided to test more fully acetic acid, ethyl alcohol and acetic ether. These were chosen because they are commonly found in fermenting fruit, the first two in quite large quantities and the third in traces. “lo make the tests more accurate, a new piece of apparatus was constructed in which two traps were so placed in the sides of a leaden trough that each fly as it passed 520 William Morton Barrows through the trough had a chance to enter the charged trap or the check trap or to react to the odor issuing from the former. The plan and elevation of the apparatus used are shown in Figs. 2 and 3, respectively. On a wooden base 4, some 30 cm. in length _ —_ Fig. 2 Fig. 3 Figs. 2 and 3 Plan and elevation of apparatus used in testing the reactions of Drosophila to odors. A, wooden base; B, leaden trough; C, zinc slide for closing exit from the receiving chamber, E; D, glass plate; E, inverted glass dish forming a receiving chamber; F, inclined way; G, inverted glass dish form- ing a collecting chamber; H, hole in glass plate serving as entrance to collecting chamber; J J’, entrance tubes to trans; F, suction tube. and 12 em. in width, was mounted a leaden trough B, which was 14 cm. long and about 2.5 cm. broad except at the far end (right in the figure) where it expanded into a chamber 3 cm. square. The Reactions of the Pomace Fly to Odorous Substances 521 passage in the trough was 5 mm. deep, and 11 mm. broad and extended from the near end, which was closed by a zinc slide C to the square chamber at the far end. ‘The trough was covered by a glass plate D, which fitted to the lead closely enough to make it practically air-tight. Against the near end of the trough was an inverted cylindrical glass dish £, which served to hold the flies to be tested. ‘This dish was raised to the level of the glass plate by a block of wood. Its chamber communicated with the passage of the trough by a short inclined way F, which allowed the flies to pass into the trough when the slide C was open. At the opposite end of the trough and resting upon the glass plate was another inverted glass vessel G, communicating with the trough by means of a small hole H through the glass plate. This vessel served as a reservior to hold the flies that had been tested. A fly creeping from chamber F along the trough to chamber G passes between the open ends of the two traps, which were inserted opposite each other through the walls of the trough at / and /’. A small glass tube piercing the far wall of the chamber at f was con- nected by a rubber tube with a suction apparatus, by means of which a current of air could be drawn through the trough at any desired rate. ‘The suction apparatus consisted of a large bottle filled with water, closed at the top by a stopper with-two holes. Through one hole was inserted a bent glass tube, which served asasiphon. ‘The other hole was filled by a short glass tube which connected with the rubber tube from and served to admit the air under external pressure to the partial vacuum formed by the siphon. Reference to Figs. 2 and 3 will show that when the siphon was allowed to run at a given rate, controlled by a clamp on the rubber tube, a current of air flowed from E through the trough, past the ends of the traps to the outlet . The aim was to have this air current carry all the escaping odorous particles away from the mouth of the trap. To test the apparatus, hydro- chloric acid was allowed to evaporate in chamber E and this gas was drawn by the air current along the trough and past one of the traps which was charged with ammonia water. White fumes of ammonium chloride were formed at the mouth of the trap J, and deposited along the path of the current. The dotted line in Fig. 2 522 William Morton Barrows marks the edge of the current of ammonium chloride, which shows that none of the ammonia moved against the current toward £. In using the apparatus it was so placed that the flies, which are positively phototropic, would creep under the influence of light from chamber £ to chamber G. A number of hungry flies were placed in chamber £; the current of air was then started; and the two vials were placed in position, one containing the substance to TABLE II The numbers of flies which reacted positively to each of eight different strengths of ethyl alcohol. In each experiment the number of flies used was one hundred g 5 Number of flies that Binbenob fiestas 1) g 2 | ie 8 r=] 2 OE turned toward but Sl ee Number of the rc 2 entered’ the did not enter the = : 5 r= : S experiment = & z = if a 3 A A eS sl Alcohol | Control | Centrsl | Alcohol | 3 S 313 s 3 a trap trap trap trap go |e I 100 | ° ° ° ° ° ° DI SAREE IS ICS IS ° I ° I ° z S IBACHBAR GORE Gad 50 ° 3 10 ° 10 3 Aree np s eisleniaei 25 2 ° 15 I \ - 15 Gosddorisdoaouse 25 3 2 3 ° (Onacbes Sareea. 20 7 ° 9 ° 16 ° TiceStoasba pocee 15 2 I 6 ° | i : addHapoc UpHcaS \ 15 2 ° 12 I Qaaancee ae isers 10 ° I I ° a : LOno saaniel eiits ahs 10 10 I 4 ° 1 a en ee is ° 3 ° I Lesage sce sec 5 2 ° 3 ° : or his oot p cece: 5 I I I 2 Tae keer eee 5 ° ° 5 ° J Aa epoca ace water ° I 2 2 2 3 be tested, the other containing water used as a control. ‘The slide was then opened far enough to allow a few flies at a time to pass down the trough toward the light. The flies, that reacted posi- tively by turning abruptly toward the traps, were counted as were those that entered either of the traps. After fifty flies had been admitted to the trough, the slide was closed and the traps were interchanged. Now fifty more flies were allowed to pass through Reactions of the Pomace Fly.to Odorous Substances 523 the apparatus and their reactions recorded. ‘These records added to those of the preceding fifty constituted the records of one experi- ment. When the different strengths of the same substance were tested on different days, the last strength used on the previous day was first tested in order to be sure that the hungry flies in stock had remained uniform in their response to this stimulus. Having ascertained this, the experiments were carried forward as though they formed a continuous series. ‘The results of these experiments are given in lables II to VI. TABLE III The numbers of flies which reacted positively to glacial acetic acid and to different strengths of this acid in water. In each experiment the number of flies used was one hundred | Number of flies that =) as} a ! S & Number of flies that ie ae ee Ss umber 0 1es at | s | g So | el | turned toward but | & ¢ 4 2 3 “th entered the Kee Qa Ss 2 a) Number of the | ‘S &, did not enter the B2é6 | 2 3 . a a Peay MAD) =| Px experiment 4d | FS sy Bogs . - (2) ay ge Acid Control | Control | Acid aa tlaes o Ss | | So Srila 5 iat ro) = ro) -e Ye # trap trap | trap trap = a | I glacial ° ° | 5 I 5 1 Diere 50 ° ° 6 ° 6 ° \ | 3--- 25 ° ° 12 oT] N a | P \ Aas 25 ° ° | 10 I Gis. : 20 ° ° 7 ° 7 ° | G32: 15 I I | 12) ° 13 I Weer 10 I ° 25 ° 26 ° Bees 5 18 | I 16 ° 34 | | | Gua. 3 ° 9 ° 12 | ° IOo.. 2 I eC z ° 3 ° From Table II it can be seen that the greatest number of positive reactions to alcohol were obtained at 20 per cent concentra- tion, while strengths above or below this grade show a decrease in the number of positive reactions. It will be seen from Fig. 2, that a fly in passing from the near end to the far end of the trough enters the area of stimulation obliquely. Consequently one side of the animal must be stimu- lated before the other. Many of the flies entering the odor in this way give a positive reaction by turning toward the stimulated side. 524 William Morton Barrows This reaction indicates that the flies respond to a difference in the intensity of the stimulus on the two sides of their bodies. The following peculiar response was very often observed. After flies had passed the traps they would often suddenly turn and run back with a characteristic zigzag motion to the mouth of the charged trap. While testing with 50 per cent alcohol, it was noted that about one-third of the flies reacted in this manner. . Evidently they are able to follow the current of odor back to its source. These responses will be further discussed in a subsequent part of this paper. TABLE IV The numbers of flies which reacted positively to each of three different strengths of acetic ether in water. In each experiment the number of flies used was one hundred ) + c . - oe | Number of flies that “S 2 3 Ee = Oo o J: ; S 3 8 Number of flies that | ened Ea he sae , 3 as xe o 9 . “ entered the : a 2) 35 eae Number of the o a did not enter the E 2 5 g 2 Py Z a & s o 5 experiment ar ae ar lis a Tae = aS 5 deer an oO = = cS Ether Control Ether | Control | = S Sires S S a o | SB RGA oH Ss = trap trap trap trap ag = ite 8 ° 7 33 6 | ‘ Pe Ag 7 ie 8 ° I 5 ° | \ 3 4 ° ° 9 ° 8 ; ( Weitere ce aeais erie 4 ° ° 7 2 o 2 ° ° 13 ° i 5 = Wey I OSh.02 sseese see 2 ° ° 10 2 J It is plain from an inspection of Table III that the largest number of positive reactions was, produced by 5 per cent acetic acid. Not only is this true, but, when the trap was charged with this strength, about half of the flies which at first failed to respond to the stimulus returned through the current of odorous material from the far end of the trough back to the mouth of the trap. These responses are not included in the table. Acetic ether is soluble in water only to the extent of about 8 per cent, and when used in such high concentration it affects the flies in a singular manner. ‘They show intense excitement and struggle at the mouth of the trap for a chance to enter, yet when one has succeeded in entering it backs out almost immediately. Reactions of the Pomace Fly to Odorous Substances 525 It is probable that the dissolved ether evaporates rapidly forming an almost saturated atmosphere inside the trap, and this 1s known to kill the flies in less than chree minutes. Hence the excessive stimulation probably causes them to back out of the entrance to the trap into which they had been enticed by the less concentrated vapor. Acetic ether is never so abundant in decaying fruit as in the weakest solutions (2 per cent) tested in these experiments. In order to make a mixture which should combine the optimum strengths of alcohol and acetic acid, equal volumes of 40 per cent alcohol and Io per cent acetic acid were mixed. ‘This mixture then contained 20 per cent of alcohol and 5 per cent of acetic TABLE V The numbers of flies which reacted positively to each of four different strengths of a mixture of equal parts of 40 per cent alcohal and 10 per cent acetic acid diluted with water. In each experiment the num- ber of fltes used was one hundred ets Number of flies that} “G6 &# 4 | G6 & & i o 8 eo Number of flies that ‘ go 3 2 re a tuned toward bur |] 896 9 | 8 6 x 3) entered the ; wey Sp |) ee) Number of the | did not enter the la 2 poled y a : a | 3 2 experiment § &, 3 = ane sats 2268 42 a ko = 2 = = os a Charged Control Charged | Control | § $5 53| 8 S @ fa) Pt oUF. US oF + 2 trap | trap | trap trap B a see 100 13 ° 17 4 30° 4 Davee 50 26 5 2 ° 28 5 eat 25 29 I 2 ° 31 I Aer 125 27 5 12 ° 39 5 1 acid. Table V shows the results obtained by testing flies with this mixture either pure or diluted with water. A solution of 124 per cent of this mixture, which is equal to a mixture of 24 per cent alcohol and % per cent acetic acid, gives a slightly higher number of positive reactions than is given either by 5 per cent acetic acid (Table III) or 20 per cent alcohol (Table II). The numbers of the positive reactions are not significantly large, yet it is probable that the mixture is uniformly more stimulating than alcohol or acetic acid alone. Table VI shows the results obtained by testing the mixture con- taining 20 per cent alcohol and 5 per cent acetic acid (Table V) to 526 William Morton Barrows . which had been added 8 per cent of acetic ether. Of the dilutions used, 12} per cent of the mixture induced the largest number of flies to react positively. It is probable that the experiments were complicated by the presence of a higher per cent of acetic ether than is met with under natural conditions. In Table I, Experi- ment 3, about .o4 per cent of acetic ether was added, respectively, to 10 per cent alcohol and to 4 per cent acetic acid, and in both instances there was an immense increase in the number of re- sponses as compared with the responses to those reagents alone; this increase must have been due to the slight amount of acetic ether present. We may safely conclude that acetic ether probably plays some part in the reactions of Drosophila to normal food. ’ TABLE VI The numbers of flies which reacted positively to each of three different strengths of the mixture of alcohol, acetic acid and acetic ether. In each experiment the number of flies used was one hundred 1 |Number of flies that os . = | ene 2 & |Number of flies th a 8 |eeeee So Number o ies that FS ae | turned toward but} § = | | & GS S — c | } = S08 entered the d in 2S oes Number of the x did not enter the EB 2 = E 2g z oO ae: = = experiment ao = aS5&| & 84 a 2 c 5) = S & | Charged | Control | Charged Control | 3 a = | ca zS Pe 5 S'|oeeaees 2 trap trap. | trap trap = l I Store) ° I | 4 ° 4 rg et ia, le pA 123 13 3 7 ° 20 3 | Bs. os scl aic ess 6} 4 ° 5 9 The foregoing experiments show that Drosophila is positively chemotropic to alcohol, acetic acid and under certain conditions to acetic ether. The optimum strengths of alcohol and of acetic acid are 20 and 5 per cent, respectively, while that of acetic ether is uncertain, but must be only a fraction of I per cent. Table VII, made up from data given by Leach (’05), shows that alcohol and acetic acid are commonly found in cider vinegar, fermented cider, and California sherry in per cents that are close to those which call forth the largest numbers of reactions in Droso- phila. Acetic ether is found in these fluids in very slight traces. Reactions of the Pomace Fly to Odorous Substances 527 3 Ex periments on the Directive Effects of Odorous Substances Having determined that these flies are chemotropic to fermenting fruit, | turn to the second question, In what way does the fly find its food ? To ascertain the accuracy of flight toward the food, experi- ments were carried out in the following way. About one hundred hungry flies were liberated in a large laboratory room. A few minutes after their liberation a tumbler, freely exposed on the top of a table and containing fermenting banana was opened. As the hungry insects in flying through the room passed near the table they eventually discovered the banana. When they were six or more feet from the tumbler they showed a rather characteristic TABLE VII Amounts (in per cent) of acetic acid, and alcohol found in cider vinegar, fermented cider, and California Sherry (Leach ’05) Acetic acid in per eent | Alcohol in per cent Number Substances —- —-—— —|— —— — : ; of samples Max. min. Av. | Max. Mins je eAv | Cidervinegar....) 7.61 Br24 4-65. | | 44 Fermented cider.) 6.59 24 3.18 6.85 1.1 4-72 16 California sherry | [FO m | sos 21.85 8.22 ‘ 66 vibratory flight. Short rapid excursions were made through the air, ‘up and down, forward and backward, right and left. Sometimes the fly came nearer the tumbler and under such circumstances it often remained in its vibratory flight in this new situation. As it approached to within about three feet of the tumbler the excursions shortened and the fly oriented more accurately to the source of the odor, though the flight still showed considerable vibration to right and left while the head of the fly was directed generally toward the tumbler. When about two feet from the tumbler, the vibratory movements grew less and less extensive and the flight became more rapid and more accurately directed toward the tumbler. ‘The last six or eight inches of the journey was made in nearly a straight line to the edge of the tumbler or to its base. 528 William Morton Barrows It is a source of continual surprise to see how accurately and quickly many flies will find food. Not only do the flies find food easily and certainly during flight, but they can also find it successfully when creeping. To test this the following experi- ments were tried. A small piece of fermenting banana was placed on a glass plate one inch square and the glass plate with the banana was put in the center of a square sheet of paper ruled into 25 squares each one an inch on a side (Fig. 4, ato b). The glass cylinder used in the first experiments (p. 516) was then placed over this paper in such a position that the four corners of the paper just touched the lower edge of the cylinder. To prevent air cur- rents from driving the odorous particles away, the chamber was closed by a glass plate above. Hungry flies were admitted singly at the bottom of one side of the chamber. ‘They flew as usual toward the light side and upward, but in the course of a minute started to creep down the glass toward the bottom of the chamber. As they came on the paper their course was carefully plotted on a duplicate sheet of ruled paper. The courses given in the diagrams (Fig. 4) show the characteristic paths traversed by twelve flies. ‘These paths showthat inmost cases the flies took the most direct route in order to reach the food, 7.¢., they did not merely run upon it by chance. ‘The paths are usually so direct that it appears as if the flies found the food by sight, but that this is not so, is shown by subsequent experiments, in which after the removal of the antennz the flies seldom found the food, though their eyes were intact. It is clear that both in flight and in creeping the movements toward food are at first irregular and afterward more accurately directed, so that the fly eventually takes an almost straight course to the food. ‘The beginnings of the courses in flight and 1n creep- ing are such as to suggest the trial and error method of response. This view is supported by what is sometimes seen in the zigzag course taken in the return of flies against the current of odor in the trough in the experiment described on p. 524. ‘The conclusions of the courses however in both flight and creeping are so accurately directed toward the food that trial and error can play no part in explaining this condition; one is forced on the contrary to assume Reactions of the Pomace Fly to Odorous Substances 529 DIRECTION OF LIGHT RAYS Fig. 4 The irregular lines represent the paths traversed by twelve flies in reaching a piece of fer- menting banana placed in the center of the area. The area was § inches square and the fly almost always started from some point on the outer edge of the area. 530 William Morton Barrows some such method of orientation as is implied in the theory of tropisms. ‘The more usual conception of the tropism theory, as advocated particularly by Verworn, is to the effect that when an animal is unsymmetrically stimulated it turns until it is symmetric- ally stimulated and either faces toward or away from the source of the stimulus and then moves in the appropriate direction. It is evident from this that symmetrical stimulation is an essential feature of this theory. Loeb has extended this view in the sense that he often implies that the stimulus acts directly on the loco- motor organs, but I do not regard this as an essential part of the tropism theory, and, as I shall show presently, this modification of the theory has no application in this case. The question, then, is, are the accurately directed responses of Drosophila dependent on symmetrical stimulation? If this is the case, one would natur- ally turn to the antennz of this animal as the symmetrical recep- tive organs of smell for such reactions. 4. Experiments to Determine the Position and Function of the Olfactory Sense Organs It is generally believed that in most insects the antennz ate the seat of the olfactory organs. In some species these organs are placed in pits, while in others they are exposed on the surface of the antenna. Mayer (’79) described a pit in the third segment of the antennz of a species of Drosophila which he considered an olfactory organ. I have found that Drosophila ampelophila has a large sac-like pit situated in the end of the termina! (third) segment of the antenna, whicn contains sense cones. Fig. 5 shows a front view of the head of Drosophila and the position of this pit in the left antenna. The following experiments on flies which had been deprived of the third segments of their antennz show that without doubt the olfactory organ in Drosophila is located in this segment. It was found by repeated trials that the antennz could not be satisfactorily covered with gum to keep out the stimulating odor, nor could they be burned off without considerable injury to the fly. The method finally employed was to place a fly, already etherized,on its back on a glass slide and to hold it down with a small camel’s- Reactions of the Pomace Fly to Odorous Substances 531 hair brush, which in turn was held in place by a small rubber band. Secured in this way under the lens of a high-power dissecting microscope, the third joints of the antennz were cut off by a pair of fine embryological dissecting scissors. ‘Uhere is a deep division between the second and third segments of the antenna, and the third segment was usually removed without injury to the second. After the effect of the ether! had passed away, the fly operated upon seemed in most respects perfectly normal. Such flies were left for twenty-four hours without food, but were supplied with Fig. 5 Front view of the head of Drosophila showing in dotted lines the position of the olfactory pit in the terminal segment of the left antenna. water. At the end of this time they were liberated singly in the large cylinder used in the experiments described on p. 516, and were carefully watched for five minutes in order to be certain that their behavior was normal. Having ascertained this, their ability to find a piece of fermenting banana in the cylinder was tested. The time required to find the food was recorded, or if the food was not found in fifteen minutes, the experiment was discontinued. 1Tt should be noted that the process of etherizing has no lasting effect on the ability of the flies to scent food, 7. e., normal flies after recovery from ether find food with as great certainty as they did before etherization. 532 William Morton Barrows Of fourteen flies which were thus tested all with one exception failed to find the food within fifteen minutes. In the exceptional case food was found in twelve minutes, but the insect’s course was such that it obviously came to the food by accident. During these experiments the flies were carefully watched, and although many of them came within 1 cm. of the food, they did not find it. To test this matter further, four of the flies with defective anten- nz were allowed to wander in a small glass tumbler until they found the food, which they apparently did by accident. When one foot of the fly touched the food or a small drop of water the tongue was immediately put down and the animal began feeding. It is not impossible that two transparent hairs which are found beneath the claws of each front foot, and have the appearance of sense hairs, may be instrumental in giving rise to this feeding reaction. It therefore seems certain that the sense of smell is absent, or at least greatly reduced, in flies which have lost the terminal joints of both antennz. In order to determine the rela- tive time taken to find food by flies with and without the terminal segments of the antenna, six normal flies were admitted to the chamber and the time recorded which elapsec before they reached the food. ‘These flies were then operated apon; the distal seg- ments of both antennz were cut off and they were allowed to rest twenty-four hours, when they were again tested and the time similarly recorded. ‘The results of these experiments are given in Table VII!. From this table it will be seen that the normal hungry fly finds the food in about two and one-half minutes, while the same fly after having beer operated upon seemed unable to locate the food. We may conclude: first, that Drosophila does not find its food by sight, but by smell, and when this setise is lost it reaches the food only by accident; and, secondly, that the olfactory sense organs—at least those which are concerned with finding food—are localized in the third or distal segment of the antenna. The fact that in Drosophila the antennz are the principal organs concerned in the reception of olfactory stimuli and that they are symmetrically placed on the body of the animal, leads to the conclusion that these flies orient to odorous centers in the way Reactions of the Pomace Fly to Odorous Substances 533 assumed by the tropism theory. Forward locomotion would be called forth by an equal stimulation of the two antennz and lateral movements by an unequal stimulation of these organs. If this view is correct, circus movements ought to result after the removal of one antenna even though the stimulating atmosphere contains a uniformly distributed odorous substance. It was, therefore, thought desirable to experiment upon flies from which one antenna had been removed in order to produce excessive unilateral stimulation. Before the operation the flies were tested for five minutes in pure air and five minutes in an atmosphere with odor, to make certain that they were normal, 7. ¢e., that they did not turn more TABLE VIII Records of the times which six flies took to find food before and after the terminal segments of their antenne had been removed Time in minutes in which the Numbers of the flies ~--—— —— Normal flies found the food Injured flies found the food 1... 5 SOC EO cee Diels 1. 03 ee 4.0 Chenin eee 1-75 Flies all failed to find the food at the doo Scqo ne Bae ce Ree | ans end of twenty minutes. - 2) a0 59 ORO eta 2.0 MPA SEY sia. e orese die So, 03 1.0 EMEA VER ie cs visser isch ath 2.45 frequently to the right than they did to the left or vice versa. The terminal segment of the right or the left antenna was then cut off from each fly. After the removal of this segment, the flies were fed and allowed to remain twenty-four hours, when they were again hungry. They were then admitted singly to the cylinder containing only pure air and watched for five minutes. Without exception they reacted as they did under similar circumstances previous to the operation. A little odor was now blown into the cylinder from a wash bottle partially filled with fermenting banana and the flies were again admitted singly and their movements carefully watched. If the fly moved in a circular path, as was 534 William Morton Barrows usually the case, a record of one circle was put down for the fly when it had turned continuously through an arc of 360 degrees in one direction. A circle was called positive if the fly moved with its normal antenna on the side toward the center and negative if it moved in the opposite direction; 7. ¢., a fly having its left antenna cut and moving always with its right side next the center would be said to be describing a positive circle, etc. “Table IX shows the records of twelve flies tested in the manner described. TABLE Ix The numbers of positive and negative circles made by twelve flies which had only one functional antenna. Each fly was tested singly for a period of five minutes in a uniformly odorous atmosphere | The number of circles made ina Numbers of the flies Normal antenna positive negative ‘ | direction direction Ab ba pede Boek YACE COR HOA Be OOOO ODER AS 6 aco right | 6 ° Pete esau cots Sect as Sat coboPaeegnonodtad duce left 4 ° 7[dos Co Oe AGOA NOG SHOE O a HOBOS Doe Ae aUios right 4 ° lipsocEbacccbhe Sotwennenas ce dina d bp ac Maser left 3 ° SOG SOTA Od son OHNE DOS AOE MODI IGensoDo SC left 5 I bdo Ee ans Pena s Oe ea ae SoU Sob suse S 4 a5e right 5 I JEW WS Gtne tice che Aigo Dab cane Soomon ame ee right 5 ° (SEA Ge am anc Ucoteo atau aoe eae left 2 ° 71 Rf CRE, fea RE I cert SER aOR left 2 ° Bee eee Liou beers ucis cree meee Rae vise eee aioe left 5 ° ace ea RELA Re io EAB OS OLA aA ee left 2 ° Oe ES BS i Oa aSin hac IIe Bern ry oon OIE left 3 2 TINGE Onset, Rate ce ear Oreo SARE SOOO PE BOCA tenner OOO Ate 46 4 From this set of experiments it will be seen that forty-six out of fifty of the circles, or g2 per cent, were made in a positive direction, 1.e., toward the normal antenna. As this antenna is obviously the one stimulated, it is clear that the flies must orient to unequal, unilateral stimulation.? 2 Since this paper was written Kellogg (’07, p. 153) has recorded circus movements in the males of the silkworm moth after the removal of one antenna and on exposure to odors. Reactions of the Pomace Fly to Odorous Substances ig Ww Wn III THEORETIC DISCUSSION The experimental results recorded in this paper show very con- clusively that the reactions of Drosophila to odorous mates:ais are by no means uniform, but vary in method under different circumstances. When the stimulus is very weak little more than random movements are excited, but when the stimulus is some- what stronger trial and error movements gradually prevail, where- by the fly becomes approximately oriented toward the odorous material, much as has been emphasized for many lower animals by Jennings (04). Finally the orientation to the odorous material becomes very accurate and the fly may be said to take an almost direct course to it. It is clear that the latter part of the course is accomplished by methods in the main free from anything that can be described as trial and error. Since under a like degree of stimulation flies, after the loss of an antenna, carry out circus movements with great regularity, it seems impossible to explain the movements under these conditions in any other way than on the basis of the tropism theory. This theory has been stated in several ways. As applied to chemical stimulation Verworm (’99, p- 429) declares: “The word chemotaxis is applied to that prop- erty of organisms that are endowed with the capacity .of active movement by which when under the influence of chemical stimuli acting unilaterally they move toward or away from the source of the stimulus. Where there is an approach to the source of the stimu- lus, there is positive chemotaxis, where there is a removal from the “source negative chemotaxis. Unilateral stimulation with chemical stimuli is only realized when the concentration of the substance in question gradually increases from the living object in one direc- tion.” The method by which Drosophila finds its food is directly com- parable to that observed by Harper (’05, p. 33) in the reactions of Perichzta to weak and strong light. ‘This earthworm orients away from the source of a weak light stimulus by frequent random movements, z.¢., by the trial and error method. But when the light stimulus is greatly increased the orientation is direct, random movements toward the light are suppressed altogether and the 536 William Morton Barrows worm appears to move directly away from the light without notice- able trial movements. It seems to me probable from experiments described by Pearl (03, pp. 623-670) that planarians may follow some such method in finding food. However, as the animal is not highly special- ized, the distance through which it can orient accurately is small and the result is not striking. Ivy SUMMARY 1 Drosophila ampelophila is a small fly peculiar for its fond- ness for fermenting fruit. 2 These flies are positively chemotropic to amyl and especially ethyl alcohol, acetic and lactic acid and acetic ether. 3 Acetic ether, isobutyl acetate and methyl acetate, when added in small amounts to Io per cent ethyl alcohol, greatly increase its attractiveness. A similar increase is noted where acetic or butyric acids are added to the alcohol. All these organic sub- stances are found in fermenting fruits. 4 The optimum strengths of ethyl alcohol and acetic acid as determined by the number of positive reactions given to different strengths is 20 and 5 per cent, ee while a mixture con- taining 24 per cent alcohol and 3 per cent acetic acid gives a slightly higher number of positive reactions than is given by either 5 per cent acetic acid or 20 per cent ethyl alcohol. Alcohol and acetic acid are commonly found in cider vinegar, fermented cider, and California sherry in per cents that are close to those which call forth the largest number of reactions in Drosophila. 5. Drosophila does not find its food by sight, but by smell, and when this sense is lost it reaches its food only by accident. The olfactory sense organs—at least those which are concerned with finding food—are located in the third or terminal segment of the antenna. 6 When one antenna is lost and the other antenna is stimu- lated by food odor, circus movements are carried out in such a way as to prove that the fly orients normally by an unequal stimulation of the antennez. Reactions of the Pomace Fly to Odorous Substances 537 7 Drosophila, when stimulated by weak food odor, first shows random movements, /. ¢., it attempts to find the food by the method of trial and error, but as the fly passes into an area of greater stimu- lation, these movements give way to a direct orientation. ‘This orientation is a well defined “tropism” response. ‘These reactions of Drosophila are paralleled by those of Perichzta to strong and weak light and possibly also by the food reactions of planarians. V BIBLIOGRAPHY Harper, E. H., ’05—Reactions to Light and Mechanical Stimuli in the Earthworm Pericheta bermudensis (Beddard). Biol. Bull., vol. x, no. 1, pp. 17—34- Jennincs, H. S., ’04—Contributions to the Study of the Behavior of Lower Organ- isms. Carnegie Institution of Washington, Publication no. 16, 8vo, 256 pp. Kettoce, V. L., ’07—Some Silkworm Moth Reflexes. Biol. Bull., vol. xii, no. 3, Pp- 152-154. Leacu, A. E., ’05—Food Inspection and Analysis. New York, 8vo, xiv+ 787 pp., 40 pls. Logs, J.,’97—Zur Theorie der physiologischen Licht- und Schwerkraftwirkungen. Arch. f. ges. Physiol., Bd. 66, Heft 9-10, pp. 439-466. Mayer, P., ’79—Zur Lehre von den Sinnesorganen bei den Insecten. Zool. Anz., Jahrg. 2, No. 25, pp. 182-183. : Peart, R., ’03—The Movements and Reactions of Fresh-water Planarians: A Study in Animal Behavior. Quart. Jour. Micr. Sci., vol. xlvi, pt. 4, PP- 509-714. Verworwn, M., ’99—General Physiology. Translated by F.S. Lee. London, 8vo, xvi +615 pp. c : [vs . Pd 7 seen Per asl is tata aL eet LA De 8 > #87 Rens ft 4.1 ii¢ ¢ 7 ir. a 4 Tr? AFM © 4 " , * i : ee, + a » N 2 ; - = # = - cs a . P a * a . i : 29) = a ) | / ; ; : "> . 4 ~ , RS , ‘ CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. E.L. Mark, Director. No. 192. ie EEPECT OF TEMPERATURE ON THE’ MIGRA- TOW OF THE RETINAL PIGMENT IN DECAPOD CRUSTACEANS BY EDGAR DAVIDSON CONGDON Witu Seven Ficures I INTRODUCTION The last few decades have witnessed a gradual accumulation of knowledge concerning photomechanical changes in the retinal pigment of vertebrates, cephalopods and arthropods. It has also been shown that light influences the melanophores of the reptile skin much as it does the retinal pigment cells. The effects of temperature upon the pigment migration of the melanophores, especially in Anolis and Phrynosoma, have been discussed recently by Parker (’06). Only two reports of the effects of temperature on retinal pigment have appeared and these both refer to the frog. Kiuhne (’79, p. 334) stated that in frogs which were subjected to low temperature in darkness, the retinal pigment extended farther toward the light between the cones than it did in those which had been subjected to high temperature in the dark. Herzog (’95) subsequently investigated this subject at greater length. He agreed with Kiihne that below 18°C. any decrease of temperature causes a distal migration of pigment and any increase, a proximal one. Above 18° C. he believed the result was the reverse. It was suggested to me by Prof. G. H. Parker that the decapod crustaceans would be favorable objects for the study of the influ- ence of temperature upon pigment migration because of the marked photomechanical changes often found in their eyes. ‘The prawn, Palemonetes vulgaris Stimp., and the crayfish, Cambarus bartonu Gir., were chosen as being easily obtainable species whose photomechanical reactions were well known. The conditions found in Palamonetes will be considered first. Tue JourNAL or ExPeRIMENTAL ZOOLOGY, VOL. IV, NO. 4. 540 Edgar Davidson Congdon II PALAEMONETES VULGARIS STIMP. The eyes of the prawn consist of a stalk and terminal bulb. The former contains a series of four optic ganglia, which are enlargements of the optic nerve. The bulb is made of numerous rod-like ommatidia, which extend somewhat radially from a base- ment membrane (Fig. 1, mb. ba.) near the end of the stalk to square facets (cta.) in the cuticula covering the bulb. An ommatidium may be roughly divided into a distal two-thirds, consisting chiefly of cone cells, and a proximal third containing the rhabdom. ‘The four cone cells (cl. con.) lie parallel to the axis of the ommatidium, and in their distal portions are closely associated to form the glassy cone. They taper proximally from the cone to the rhabdom whence they possibly extend as processes to the basement mem- brane. Between the cone and the facet are two small corneal hypodermal cells (cl. crn.) Six distal retinular cells full of black pigment together form a sheath around the cone cells (Fig. 3). They do not belong exclusively to one ommatidium, but each one serves as a partial sheath for three cones. ‘The lower third of the ommatidium contains seven proximal retinular cells (Fig. 1, cl. px.), lying close together and parallel to the ommatidial axis; they extend as long processes into the region proximal to the basement membrane. Distal to this membrane they unite to form the spindle shaped rhabdom (rhb.) ‘They contain black pig- ment. Associated with them are one or two whitish accessory pigment cells. Fibrilla from the distal optic ganglion extend up through the proximal retinular cells to end in the rhabdom. Distally, the central parts of the ommatidium are transparent and convey the light to the rhabdom, which is thus open to stimulation. The photomechanical changes of Palamonetes have been de- scribed by Parker (’97). In increasing light the distal retinular cells migrate as wholes in a proximal direction, thus restricting, as Exner (’g1) has pointed out, the amount of light that enters the deeper parts of the eye. The pigment of the proximal retinular cells is at the same time carried distally along the sides of the rhabdom, probably by protoplasmic streaming within these cells. This process also reduces the amount of light that can reach the Migration of the Retinal Pigment 541 thabdom. In decreasing light the distal pigment cells mvoe distally till they surround the cone and the proximal pigment 1s carried below the basement membrane. In consequence of the Uj Bigrrest? IB}: light 10°. light 30%. dark. 10%. dark. 80°. Fig. 1 Fig. 2 Fig. 3 Fig. 4 Figs. 1 to 4 longitudinal sections of the ommatidia of Palemonetes, magnified 200 diameters, show- ing the distribution of the retinal pigment under the following conditions: Fig. 1, in light and at 10°C. cl. con., cone cell; c/. crn., corneal hypodermal cell; c/. dst., distal retinular cell; c/. px., proximal retinular cell; con., cone; mb. ba., basement membrane; cta., tuticula; rhb., rhabdom. Fig. 2, in light and at 30° C. Fig. 3, in dark and at 10° C. Fig. 4, in dark and at 30°C. movement of the proximal pigment, the light-colored accessory pigment about the rhabdom is exposed and may serve, as Exner 542 Edgar Davidson Congdon observed, for areflecting apparatus. These pigment migrations plainly tend to protect the eye from over-stimulation by strong light and to increase its chances of stimulation in weak light. It is evident from the foregoing that light conditions must be taken into account in testing the effects of temperature. To do this, a series of experiments in the dark and another in the light were planned, each of which included three temperatures: 10°, 20° and 30°C. One extreme of temperature in each series would increase and the other decrease the effect of light, if indeed tem- perature is a factor in the migration of pigment. As the photo- mechanical changes in Palzmonetes are ordinarily completed in about two hours, [ extended my experiments to only two and a half hours. ‘The three experiments of each series were performed at the same time. Care was taken in the light series to have the three aquaria for the three experiments placed close together so as to receive equal illumination. Six to twelve individuals were taken for each experiment. All of the animals came from a com- mon supply and were similar in size and sex. One set of experi- ments was conducted twice, once with each sex. No difference in responsiveness was found between males and females. The animals were put into quart glass jars filled with water and these jars were placed each in a two-gallon cylindrical aquarium. The latter was filled with water at the desired temperature; thus the water in the inner jar containing the animals was brought in half an hour to the required temperature. Preliminary trials showed that 10° C. and 33° to 35° C. were not harmful to the prawns when thus gradually produced, though the higher tem- perature would cause at times the death of the animals if suddenly applied. In the dark series there was no easy means at hand for maintaining the desired temperature in the dark-proof box with- out admitting light. Consequently the box was kept closed dur- ing the whole of the experiment, the water thus being allowed to Phee gradually toward room temperature. This resulted in a variation of about 3° C. during the experiment, an amount not sufficient to predjudice the wee At the end of treatment the animals were plunged into water at 80° C. for a fraction of a min- ute, until fixation was accomplished. ‘The eyes were then prepared Migration of the Retinal Pigment 543 for examination by being hardened, cut into sections and stained with borax carmine. Eyes of the different experiments showed a very perfect series of migration stages both for the proximal and the distal pigment. In the dark series, the proximal pigment was always proximal to the basement membrane. At 10° C. (Fig. 3) it was close against the basement membrane; at 20° C. the distance of its distal mar- gin below this membrane was equal to about one-fifth the length of the rhabdom; at 30° C. (Fig. 4) the distance was two-fifths of the length of the rhabdom. In the light series at 10° C. (Fig. 1) there was a strong concentration of the proximal pigment into the ends of the cells just distal to the rhabdom. Only a little pigment could be seen proximal to this. At 20° C. more of the pigment was proximal in position and surrounded the rhabdom. At 30° C. (Fig. 2) the pigment was rather dense around the rhabdom yet not so abundant as at the distal ends of the cells. Both series show that with increasing temperature the proximal pigment moves proximally and with decreasing temperature it moves distally. The distal retinular cells can not be said to show as pronounced a response as the proximal ones did, yet the series was convincing. In all preparations from the light series the distal retinular cells were proximal to the cone; in the dark series they were at least partly surrounding it. At 20° C. in the light the distal pigment was in large part distributed evenly along the cone cells; a small part was collected at the top of the proximal retinular cells. Low temperature increased the effect of the light by massing all of the pigment just above the proximal retinular cells. “The high tempera- ture produced an even distribution of the pigment along the cone cells with no proximal accumulation. In the dark condition the distal pigment cells completely covered the cone at high tempera- ture (Fig. 4). Low temperature (Fig. 3) resulted in a proximal migration equal to one-third the length of the cone. In general, increased temperature causes distal, and decreased temperature, proximal migration in distal retinular cells. It may be said that in both types of pigment cells in Palamonetes the effect of increased temperature is opposite to that of incre: sed light. 544 Edgar Davidson Congdon Parker found a migration of the accessory pigment, but as it 1s not easy to determine this even in its relation to the light, I did not occupy myself with its relation to heat. III CAMBARUS BARTONII GIR. In the crayfish Cambarus bartonii the reaction of the retinal pigment to changes in light and in temperature were by no means so clear as in the prawn. Exner (’gI, pp. 108-109) long ago made a similar statement about Astacus, so far as light reactions were con- cerned. In experimenting on Cambarus the factor of light was elimin- ated by confining the temperature treatment to animals in the dark. The series of temperatures tested was 2°, 10°, 14°, 22°, 29°, 34°, 39° and 41°C. The high temperatures 39° and 41° C. resulted in the death of half the animals and produced variable conditions in the eyes of the survivors. [he experiments were discarded as probably dependent upon irregular moribund conditions. There was no reason to think that 2° C. was harmful. As a precaution against shock the more extreme temperatures were only gradually applied. Animals that were subjected to 2° C. were kept at this temperature for twelve hours, so that, if the cell processes were somewhat retarded by the cold, sufhcient time would be given for the reaction to become complete. ‘The animals were kept at the other temperatures for at least two and a half hours, as in the case of Palamonetes. The low temperatures were obtained by putting the animals in a shallow dish supplied by water cooled by ice. For the high temperatures, a tank was arranged in a light-proof box so that water of a desired temperature could be replenished without admitting light. After the experi- ments the eyes of the animals were fixed in water at 80°C. The tough cuticula was then removed and the retina sectioned and stained in borax carmine. The distal retinular cells of Cambarus showed in the different experiments a considerable diversity of positions and could not be reduced to a simple series conformable to the differences of tem- perature. I am inclined to ascribe this condition to some fault in _ Migration of the Retinal Pigment 545 my methods. ‘The average positions of the proximal pigment of different animals formed a natural series, although the average Non bet a ™ Figs. 5 to 7 longitudinal sections of retinular and sub-retinal portion of the eye of Cambarus, magnified 350 diameters, showing the distribution of the retinal pigment under the following conditions: Fig. 5, in dark and at 2°C. Fig. 6, in dark and at 22°C. Fig. 7, in dark and at 34° C. interval between the temperatures employed was only about 6° C. At 2° C. (Fig. 5) a considerable amount of the proximal pigment 546 Edgar Davidson Congdon was found clustered around the rhabdoms or betweenthem. Some- times a little was scattered in the region between the rhabdoms and the basement membrane. Eyes of animals that had been kept at 34° C. (Fig. 7) frequently showed no pigment at all above the basement membrane. In some cases a small amount of pigment was scattered near the rhabdoms. Although there were frequent individual differences, the remaining experiments yielded a natural series of results between the two extremes mentioned. The proximal pigment of Cambarus moves therefore, like that in Palazmonetes, distally with decreasing temperature and proximally with increasing temperature. Light and heat have opposite effects. As in Palamonetes, the movement in response to temperature is much less than in response to light. IV DISCUSSION Parker (’06, p. 410) in a recent paper summarized the effects of light on the melanophores of the reptile skin and on the proximal retinular pigment in the arthropods in the statement that all migration due to increased light is distal and so toward the source of light, and all migration due to decreased light is proximal. He also gives experimental evidence that increased temperature produces a proximal migration and decreased temperature a distal one in the melanophore pigment of reptiles. ‘The proximal pig- ment of the decapod crustaceans Palamonetes and Cambarus, as shown in this paper, falls under the same rule. Herzog’s obser- vations on the influence of temperature on the migration of the pigment in the frog’s retina agrees with this statement for tempera- tures below 18° C., but not for those above this point, where the reverse is said to be true. Aside from this observation of Herzog’s, which needs confirmation, all evidence points toward a general law for temperature the reverse of that for light; namely, increase of temperature causes proximal migration, decrease of temperature, distal migration. In most instances of the migration of retinal pigment, the proc- ess has an adaptive value in controlling the amount of light that reaches the receptive organs. Possible adaptations may also be Migration of the Retinal Pigment 547 easily pointed out in the migration of the melanophore pigment of reptiles. On the other hand, the migration of the retinal pigment in crustaceans as caused by change of temperature seems to me not to be adaptive, for it is always small in amount and it occurs at temperatures higher and lower than those commonly experienced by the animals. It seems reasonable that the migration is closely associated with the accelerating effect of heat on the chemical changes in the melanophore cell, and even if the migration were more marked, it 1s difhcult to see what advantage it would give to its possessor. A probably related instance of lack of adaptation has been described by Hess (’05, p. 423) in the cephalopod eye, which often required one to two days in which to complete che pigment migration. ‘This length of time renders the migration at best a very imperfect means of adaptation for sur- roundings 1n which light and darkness follow each other at half-day intervals. VY SUMMARY In both Palemonetes and Cambarus the proximal retinal pig- ment migrates distally when the temperature is lowered and proxi- mally when it is raised. In Palzmonetes the distal pigment migrates proximally when the temperature is lowered, and distally when it is raised. In all cases increased temperatures cause a pigment movement the reverse in direction to that produced by increased light. The effect of temperature is much weaker than that of light. In the eyes of crustaceans retinal pigment migration due to temperature changes 1s probably not adaptive. BIBLIOGRAPHY Exner, S., ’91—Die Physiologie der facettirten Augen von Krebsen und Insecten. Deuticke, Leipzig und Wien, viii +206 pp., 7 Taf. Herzoe, H., ’05—Experimentelle Untersuchungen zur Physiologie der Bewegungs- vorgange in der Netzhaut. Arch f. Anat. u. Physiol., Physiol. Abt., Jahrg. 1905, Heft 5-6, pp. 413-464, Taf. 5. Hess, C., ’05—Beitrage zur Physiologie und Anatomie des Cephalopodenauges. ; Arch. f. ges.” Physiol., Bd., 109, Heft 9-10, pp. 393-439, Taf. 5-8. 548 Edgar Davidson Congdon Ktune, W.,’79—Chemische Vorgange in der Netzhaut. M. L. Hermann, Hand- buch der Physiol., Bd. i, Theil 7, pp. 235-342. Parker, G. H., ’97—Photomechanical Changes in the Retinal Pigment Cells of Palemonetes, and their Relation to the Central Nervous System. Bull. Mus. Comp. Zoél. Harvard Coll., vol. xxx, no. 6, pp. 273-300, I pl. *g9—The Photomechanical Changes in the Retinal Pigment of Gam- marus. Bull. Mus. Comp. Zo6l. Harvard Coll., vol. xxxv, no. 6, pp. 141-148, I pl. *06—The Influence of Light and Heat on the Movement of the Melano- phore Pigment, especially in Lizards. Jour. Exper. Zool., vol. i, no. 3, pp. 401-414. From the Zoélogical Laboratory, Columbia University. OBSERVATIONS AND EXPERIMENTS ON REGENER- ATION IN LUMBRICULUS! S. MORGULIS "INO TUIISIOTILS5 00 Soe Oe dle DO ClO 6 OES Oetitid ARE OR A In cr resin ren or bear Ae Ee Se 549 ER EHNA OERLC DEO OU GHOD fe sia) taco lo'alstoiaivivs elereroter suas ol-Taveve sig ey sleicis/el saya ce areas sie sigs visitas vite 551 SPURTE 20 oS bathe BES Cen SO GRnIn Oo Sete eee eae. ameter a ose Otel ct 556 Mm allest pantGa pable OL repeneratiON yc .tacisto aie) sic sie Gis oreieiereie om late Auere elsiaie ciavedayel seetiokas -leesehetete 556 BREIL OA NOSLE LION LEP ENC AON s.(cai/eyonsic cis xis. 2vs is 35 3: 2, 2 had all regenerated a head and a tail. This experiment shows that pieces containing only two segments are capable of regenerating. Experiment II. a_ By following the same method, pieces from the anterior half of the worms, consisting of 3, 3, 3, 3, 2, 2, 2, I, I, 1 old segments also produced a new head and tail. This experiment showed very distinctly that a single segment is capable of maintaining its existence, and of regenerating a perfect head, consisting, as is usually the case, of six segments, and also a tail. b From the posterior halves of the same worms, | got thirteen pieces of three segments each; sixteen of two segments and twelve of one segment. Of these there regenerated eight pieces of three segments, seven pieces of two segments each, and none consisting of one segment only. B Rate of Posterior Regeneration The object of the following experiments was to determine whether the length of the piece or its relative position in the worm’s body is directly responsible for the rate of its posterior regenera- tion. Experiment III (October 25). About thirty-fhve worms were divided into seven parts each. ‘The seventh piece (the tip of the 8 Archiy. f. Entwickelungsm., xiv, p. 586, 1902. 558 S. Morgults tail) was not utilized in these experiments. Although the average length of these terminal pieces was somewhat over forty to forty- five segments they all died. Pieces of corresponding levels were kept in the same dish, so that they were all practically under like conditions. Accord- ing to the level to which the pieces belong, they will be named A,, A. As Ay A;, Ag At the end of two weeks all the tails that had been regenerated by these pieces, were cut off; the number of their segments as well as the number of segments in pieces, by which they were produced, was recorded, and the results aregiven in Table III. Inthistableis also given the average num- ber of regenerated segments per one old segment (in the last line). By this method of calculation all the individual variations were obliterated, and at the same time this, so to speak, “ideal old seg- ment” with the corresponding “ideal” number of new segments served to indicate the regenerative power at a given level, and served also as a basis for the comparison of the rates of regenera- tion. In accordance with our method of calculation, we find that to each of the original segments at the first level, there were on the average 3.2 new segments; at the second level 3.3, at the third level 3.1, at the fourth level 2.6, at the fifth level 1.8, and at the sixth level only 0.9 new segments. There is no evidence of a correspondence between the number of old segments and the number of new, regenerated segments, provided the old parts are not very different in length. Glancing over the Table III we can see at once that in pieces at the first level, those having twelve segments produced thirty-two or fifty new ones; those having thirteen old segments thirteen or forty-two. In pieces at the second level also, those of eleven old segments regenerated twenty-nine or fifty-six, and those of sixteen may regenerate eighteen or fifty-four new segments, etc. ‘The same lack of correspondence will be found in all parts of this table as well as in subsequent ones. On the other hand if we compare the smallest pieces of a cer- tain level with the largest pieces of the same level we may some- times find very considerable differences in the amount of regener- a!) Regeneration in Lumbriculus 559 ation. ‘Thus in pieces of the third level those of seven or nine segments regenerate thirty-five or thirty-eight segments respect- ively, whereas pieces of twenty-eight, twenty-nine and thirty-one TABLE III October 25, 1906/ November 8, 1906 Ai Ao As Ag | As Ag eee fe | se) be) ee] bs | se) fe] be) ea | as Behes | fo | 2b | fo | 2b | Bo | PE] F2 | fe | Pe | oe co et arse uO a acs tw uO ua Fl we ue ve wg eee icalsalse|o2ise| oa} se) ss| ss | 2 a Z eens Z 4 Z Z A | Z Zz | Z Z 10 22 8 25 7 35 II 42 | 8 40 II Io II 32 Io | 45 9 38 12 31 II 28 16 17 nm | 38 10 46 10 41 12 AQ i) 4: 20 18 he Be! r2- | 42 1X | 29 10 41 12 62 16 38 13)-| 24 12 44 II 35 10 54 15 45 17 42 20 8 12 50 II 36 Il 28 15 46 17 56 20 16 13 ra, i} 11 47 II 52 15 ey |) ae) ° 22 22 Hay |( 32 II 48 rite ° 18 34 19 44 Pate | ats 13 | 36 nip || AS 13 37 18 42 20 e4n | = 20 23 ma. ||. 42 12 50 13 44 19 50 20 42° | 24 10 14 Agee || 12 65 14 50 19 50 21 38 | 25 36 14 Bel] ah) ° 14 56 19 80 22 21 25 48 14 cre 13 42 14 66 | 20 45 22 35 +28 18 14 56 | 14 AZ eS 50 | 20 49 22 40 30 9 15 48 | 15 | 66 16 58 20 55 22 48 30 40 15 57 15 7o 16 6om a 20 58 23 43 32 30 16 40 16 18 17 47 20 60 216 21 42 | 32 20 58 16 49 17 48 21 ° 25 42 45 27 21 68 16 Cale orSerilt” 'ci5 22 | 38 26 42 23 64 17 Gr | ae 32 22 | 52 28 28 25 go 18 52 |} 24 53 22 53 28 42 19 66 | 28 68 22 54 28 58 24 65 29 60 23 49 30 45 26 65 31 77 24 55 30 50 | 26 66 31 a7 | 30 55 36 76 1a —— Si ois |e Cd EE 2 ae 3-3 | 3.1 2.6 1.8 0.9 segments regenerate as many as sixty-eight, sixty or ninety new segments. It is not in the least surprising, I think, that pieces so consider- — 560 S. Morgulis ably different in size should have different regenerative capacities; still these differences are of little account when we take into con- sideration the regenerative capacity of pieces intermediate in size. ‘Thus a piece of only fourteen segments belonging to the same level of the worms regenerated sixty -sIX segments, or six segments more than the piece i: twenty-nine segments. In view of this evidence we may safely assume that the length of a piece has no direct relation to the rate of its regeneration, and that the rate is dependent upon the position of a piece in relation to the worm’s head. ‘The pieces nearer to the head have the highest regenerative power, which gradually decreases as we pass from the front backward. I wish also to point out in this connection the great range of variability in the regenerative power of various worms, that can be easily seen on looking over every column of the table. eee | om Ae | As_| As | As | Asc Diagram showing the rate of regeneration at different levels. In the accompanying diagram an attempt is made to demon- strate the rate of regeneration at different levels by means of rec- tangles. ‘The horizontal lines represent the “ideal” old segments of different levels, whereas the vertical ones represent the “ideal” number of new segments, regenerated at these levels. In either case a segment is expressed by a line 1 cm. long. The worms were kept in clear, filtered water, so that they did not have any food for two weeks, but since the pieces had no heads they would not have been able in any case to feed for a greater part of the time. After fourteen days (November 21) the pieces A, to A, were examined again. It will be remembered that on November 8, Regeneration in Lumbriculus 561 the regenerated tails of these pieces had been cut off, so that since then the worms had been regenerating anew. Only pieces of three levels (A;, A,, A;) were alive. The data are given in Table IV. TABLE IV [rr November 8/ November 21 As Ag A; No. of segments No. of segments No. of segments No. of segments No. of segments! No. of segments in old part in new tail | in old part in new tail in old part in new tail 9 16 9 16 6 12 9 33 II | 27 Il 14 10 15 12 6 12 4 II 24 13 34. 15 | 22 12 10 14 24 15 23 14 22 15 46 15 | 25 14 27 16 22 18 14 32 17 | 17 18 20 14 38 17 | 22 18 21 15 25 17 28 18 22 15 32 18 20 19 17 16 26 18 22 20 6 16 34 19 37 20 9 17 28 20 20 20 12 19 26 20 26 21 26 26 41 20 27 23 "#38 27 43 Ze) 29 24 Hy 29 | 45 20 33 24 22 21 24 | 25 40 24 44 | 26 26 30 36 27 27 28 17 | 36 49 1.8 He 1.0 This table shows that in course of the second period of two weeks, the pieces regenerated about one-half as many segments as were regenerated for the first period of two weeks. ‘The regener- ated tails were cut off once more, November 21. (It should be remarked in this connection that not only the regenerated tails, but also the regenerated heads as well were cut off every time.) 562 S. Morgulis After a third period of two weeks (December 5) the pieces A,, A,and A, were examined. ‘The result is recorded in the Table Y. Again we see that the pieces have regenerated only about one- half as many segments regenerated for the second period of two weeks, and about one-fourth as many for the first two weeks. TABLE V November 21 / December 5 A3 As As No. of segments No. of segments|No. of segments No. of segments No. of segments No. of segments in old part in new tail in old part in new tail inold part — in new tail 9 16 | 10 2 12 2 10 6 10 13 12 II 7 14 3 15 10 12 ° 14 10 17 7 13 ° 16 7 17 9 13 9 16 9 18 4 14 12 18 22 18 5 15 10 19 13 19 6 15 18 21 19 13 16 II 20 17, 16 15 22 10 17 15 23 14 24 21 24 13 26 31 27 5 29 27 0.8 0.6 0.5 It is worth mentioning in this connection, that the ratio between the amounts of regeneration at various levels remains almost con- - stant during the three successive periods of two weeks each. ‘Thus the ratio between the amounts of regeneration of A, and A,, A, and A,, A, and A,, for the first period will be: Ei eA es a es Bf. toe eek ee oe The same for the second period will be: | et 1.8) ia 1.8 ne ae a oa jes 5) 1.0 Regeneration in Lumbriculus 563 And for the third period: 0.8 0.8 0.6 = 1225 -= 1.0; —— = I.2,or 0.6 Ons Oy TABLE VI November 8 November 21 December 5 087,22). SOR eee ea Tez 1.2 | 1,6) MONTAG sey sxiscccotets ess eos | ney] 1.8 | 1.6 “VA 660002 aR DOSE 1.4 igh 1.2 Experiment IV. Another experiment was started on November 29. A number of worms were cut into seven pieces each, as in the previous experiment. Pieces of corresponding levels were kept in the same dish. The dishes contained nothing but filtered water. No food was present. ‘The pieces will be named accord- ing to the level of the body B,, B., B,, B,, B,, By. The worms used in this experiment had been in the laboratory for some weeks, so that their regenerative capacity was consider- ably lessened. In general worms just brought in from the pond are the best as far as their regenerative power is concerned, and lose it when kept in the laboratory. I regret that I did not take the record of the amount of regeneration at the end of the first two weeks, as it would be very interesting in connection with the results obtained from this experi- ment. At the end of four weeks (December 27) the pieces were exam- ined, the segments counted, and their numbers are given in Table VII. This table shows practically the same result in regard to the relation of the regenerated to the old tissue, and that of the former to the level of the worm’s body, as the Tables III, IV and V. In this case, however, the regenerated tissue was not removed, and the pieces were left regenerating till January 10, 1907, or fora total of six weeks. “Table VIII gives the result of counting the segments at the end of six weeks. The number of new segments is the same as at the end of four ) 564 S. Morgulis weeks, 7.e., no new segments had regenerated posteriorly. The regenerated segments, however, grew larger, and all the microscop- ical ones at the tip of the tail became conspicuous, and their setz of considerable size. Slight differences in both these tables (VII and VIII) are probably due to some miscounts, which are almost inevitable. TABLE VII November 29, 1906/ December 29, 1906 Bi Bo | Bz Bz B; Bs $e] 34a] 82 | $a] ge | Ba | de | oa! Fe | dq | ee B2) 62/82/62 |62| 62) 62] 8" | 2) 6? | eo) oe Bu / 88 | 8u | 8b | 8a | 8B] 8a | FE) Bo | SE! ba | BE ool ee teal eal cal eel ool ce) eo) en S18 ha de Le be be 2” |e le | | | II 25 10 21 10 17 13 : 13 15 | 13 16, 0) ay 12 20 II 22 II 19 13 20 16 | ©. || * 65ers 12 40 12 24 12 30 13 30 16.| 2 ae "ie 5 13 35 13 26 12 ZO Is 225) || a6 23 | 20 12 14 35 13 32 14 12 16 i | 17 II 23 28 15 26 13 36 14 26 16 Sa Wi 817-0 gueas 24 7 iS) a2 14 36 os) 39 16 a9 ety) aes 2a ae 16 31 14 40 16 2 ae Ley 27 17 18 | ~29 18 16 40 15 26 16 43 17 28 18 17 17 30 15 33 17 ZY 18 26 18 21 17 32 15 44 17 34 18 32 | 19 12 17 45 18 36 17 39 20 30 | 19 | 42 18 30 18 43 24 24 20 46 20 15 18 38 18 45 24 40 22 25 21 17 18 45 20 32 22 42 21 21 18 45 21 50 23 35 21 | 30 8 15 27 44 2135 | 22, 254) | | e239 26 | } 2 ofr p60 ) 2a 26 2.26 2.26 1.80 Le7t | 1.09 | 0.62 Let us now consider this last experiment in connection with the previous one, and see what we can infer from them. In either case the pieces of the worms were regenerating without food for a period of six weeks. In the latter case, where the process was Pe Regeneration in Lumbriculus 565 going on undisturbed, the parts formed a certain number of seg- ments, then the formation of new segments stopped, and only growth, or increase in size took place. It seems as if these pieces of worm, brought out of equilibrium by the operation, completed themselves, attained again a state of equilibrium, and formed TABLE VIII Bi nth alo eee, ee, 2 l ge gs Be | ge) ge | ga | de | delde | Ga | Ge | Es ee) 62) 62/87/62) 82/82/87 | 68 | 82) ge) 22 Bu | 2b) eu | 2S) ez | BE) Be | BE | Be | SE | Ba | BE Seer so | 3s" | se {i Ob Se alisotase sill ese le Sean eau eno ener Pemeeeseiesed | 6.4) =. 48} 68 | ¢# + e383 7 eal os | od lea Z Za A a Z | 4 Z Z a | a A Z WE | «25 10 21 ro |b a6 13 13 15 13 16 7 12 20 Il 22 ir | 19 13 18 15 16 16 15 12 40 12 22 .|-30 | 13 30 16 ° 20 12 meas: ||") “13 26 14 12 13 32 16 21 23 28 14 35 13 33 14 26 16 18 17 Il 24 7 15 26 | 13 36 15 37 16 34 17 15 24 14 um | 40 14 35 16 | 22 16 40 17 15 29 18 aie 40 | 14 36 16 43 17 27 17 18 17 31 14 40 17 21 17 28 18 17 17 a2 aa 15 27 17 | 39 17 30 18 21 Z Tees, || 16 23 Zi) he day 18 28 19 12 TSP il kZO 18 36 24 40 18 32 19 45 iS | 34 18 47 20 46 20 15 18 38 18 49 22 27 21 17 18 44 20 32 22 42 21 22 19 Bee 27) |) 52 23 36 21 30 27 44 as 3 22 | 25 | | a es | | 24 36 | 27 26 2.24 2.24 eG 1.74 1.09 0.66 dwarf worms. ‘These were formed by pieces one-seventh the original length of the worms. If in accordance with Bulow’s data each worm divides spontaneously on an average into four pieces (3.5 pieces being the actual average), worms only twice as big as 566 S. Morgulis those obtained from my experiment, would result, and would also very likely remain dwarf worms. If my view in regard to the formation of dwarf worms from pieces of Lumbriculus is true, the worms if divided would become continually smaller and smaller, tll they would be reduced to single segments. As a result of the continual cutting off of the regenerated tissue more new material was produced than when only one cut was made through the same level. ‘Thus in the first experiment for the third period of two weeks there were to each old segment 0.5— 0.8 of a new segment, while the pieces cut only once did not produce even a single segment for the same length of time. C Regeneration of Regenerated Parts Tails that had been regenerating from parts of Lumbriculi for some time, were cut off so that none of the old tissue remained. Some of the tails had been regenerating for two weeks. After removal from the old part a new tail was allowed to regenerate in course of the following two weeks. ‘These were also cut off. a October 31. ‘Thirty such regenerated tails were obtained from pieces belonging to the third level, and thirty regenerated tails from pieces of the fifth level, of worms that were divided into eight parts each. ‘These tails had been regenerating since October 17. The tails regenerated by pieces of the third level we will call for convenience A, and those regenerated by pieces of the fifth level B. These sixty tails were kept for two weeks, and when examined on November 14, eighteen A tails and thirteen B tails, or half the original number were alive. ‘Those that survived regenerated new heads of five to seven segments, with one exception, of which I shall speak presently. ‘The little worms contained on an average, probably, about forty-five to fifty-five small segments. In the course of two following weeks they did not produce any new posterior segments. ‘That this is really so can be ascertained with certainty because in the regenerating tails the terminal five to six segments, which are the youngest, are microscopical in size. Dur- Regeneration in Lumbriculus 567 ing the time intervening between October 31 and November 14, all these microscopical segments increase in size very much, their setae became large, and the blood vascular system quite conspicu- ous, as in older segments; but no more ‘ anlage’’ of segments have made their appearance. ‘The last large segment was contiguous to the anal segment. This result shows that when a new regenerated tail regenerates a head, and thus forms a little worm, although the segments grow larger, no new segments are laid down at the posterior end. Whether the same holds true also for non-regenerated posterior ends of Lumbriculus, I do not know, since in all cases when pos- terior ends were removed they died. b November 14. Again twenty-eight regenerated A, tails, and thirty B, tails were obtained from the same pieces as the fore- going at the third and fifth levels, which had been regenerating from October 31 till November 14, or for two weeks. November 28, I found alive out of fifty-eight tails (A,, B,) eleven A, and seven B,. They all had regenerated new heads, although the number of segments in these heads was in some cases only four. Thus about one-third of all the regenerated tails produced new heads, and eighteen very small worms of about twenty-five to thirty segments each were obtained. In.this case also no new segments were laid down at the posterior end, but the segments became larger. c November 14, I cut in two all the thirty little worms formed by the regeneration of A and B tails, in order to determine the rate of their posterior regeneration. ‘The cut was made not quite in the middle, so that the posterior parts were somewhat longer than the anterior. After a lapse of two weeks, November 28, twelve anterior pieces which came from A, and eight anterior pieces from B, were found alive and regenerating new tails. None of the posterior pieces, although bigger than the anterior ones, had survived. (In other experiments a small percentage of the posterior pieces regenerated heads.) On the whole a very fair percentage (about 65 per cent) of the anterior parts survived the operation and continued to regenerate the missing tails. 568 S. Morgults In the tables given below are recorded the number of segments present in the anterior pieces left after the operation of November 14 (c) and the number of new segments they had produced at the end of two weeks, November 28. In accordance with our method of calculating the average num- ber of new segments per one old segment, to define the regenerative dower, we will have 0.11 segments standing for this power in pieces from A and 0.06 segments for that in pieces from B. TABLE IX November 14/ November 28 Anterior parts from worms formed Anterior parts from worms formed by regeneration of A by regeneration of B No. of old seg- No. of new seg- No. of old seg- No. of new seg- ments ments ments ments 16 bud* 16 bud 16 bud 17 bud 19 3 19 bud 19 5 1g 6 20 bud 20 bud 20 bud 23 4 20 26 bud 21 3 26 bud ve 5 20.5 Ted: 23 ah Average No. of new segments an: bud per one old segment, 0.06 25 3 20.4 228 Average No. of new segments per one old segment, 0.11 * Whenever the term “‘bud” is used, it indicates that an unsegmented portion in which the anus lies is produced only, but no segments are deposited between this regenerated organ and the old tissue. Of course these data are so meager that it would hardly justify much speculation, but it seems to me nevertheless very suggestive. If we compare the rate of posterior regeneration in the little worms formed from the anterior parts of the regenerated tail pieces (A and B), we find that they stand to each other in a ratio very much like that of the rate of posterior regeneration in the old parts, from which the A and B tails originate. Regeneration in Lumbriculus 569 eo The parts of the third level regenerated from October 17 to October 31 the tails A at an average of 4.4 new segments for each old segment; and those of the fifth level regenerated in the same time the B tails at an average of 2.6 new segments. During the next two weeks, from October 31 until November 14 they had regenerated at an average 2.4 segments and 1.7 segments, respect- ively. ‘The ratio between these rates of posterior regeneration is: a4 See inc = ie4 The ratio between the rates of posterior regeneration of the little worms formed from A and B is: Orr iat The eighteen A, and B, tails, which survived and had regen- erated heads were also cut intwo. ‘This operation was performed on November 28, or two weeks after they had been separated from the old tissue. At the end of two weeks again (December 12), only seven ante- rior pieces from A, were alive. ‘The fate of these seven pieces is shown in the following ‘Table. - TABLE X November 28 December 12 No. of old segments. No. of new segments. 14 bud 14 bud 15 bud 16 bud 17 bud 18 bud 27 bud This table shows that only the anus was formed, but no new segments were produced in these seven worms. Other experiments gave similar results and need not be recorded here. It is evident that a regenerated tail is not only capable of regenerating a head, from its anterior cut surface, but also of 570 S. Morgulis replacing its posterior part when cut in two, and that the property of regeneration passes over to the new tissue together with the protoplasmic material it is built of. D A Case of Heteromorphosis One of the thirty-one A tails of the previous experiment regener- ated a tail in place of ahead. ‘This is the only indubitable case of heteromorphosis in Lumbriculus of which we have record. ‘That this was a genuine tail and not merely a misformed head can be easily proved (1) by the number of regenerated segments, (2) by the position of the anal aperture; (3) by its ‘functional activities.” When an abnormal head develops it does not contain more than six to seven new segments. Here eighteen segments were regen- erated. ‘The segments gradually decreased in size and were micro- scopical near the distalend. ‘This sequence of segments is charac- teristic of the tail. In the regenerating head on the contrary all the material is laid down first, and then its segmentation appears. In this case the terminal aperture is not a mouth (which may some- times assume such an abnormal position) because it is round and not triangular in form, and lies in a knob of indifferentiated mate- rial, as in the case of the anus. The best proof of its being a tail is the direction of the contractions of the blood vessels, dis- sepiments and entire musculature. Whereas in a head contrac- tion takes place from before backward, here it was in the reverse direction, viz: in that characteristic for a tail. In the old tail the contractions were running in a direction opposite to that of the heteromorphic tail. The waves of contractions in both tails started at their distal end, ran toward each other gradually slowing down and vanishing altogether in the vicinity of the point of their union. ‘The contractions in the old tail were more vigorous, but they never passed beyond the old part. This heteromorphic tail developed from one of a number of pieces that had been kept in the same dish. It could not therefore be due to an external influence. In order to see whether the old tail would again produce a heteromorphic tail or a head, and also to see what the heteromorphic tail would regenerate, I severed Bien.» Regeneration in Lumbriculus A the heteromorphic tail from the old one, exactly at the line of their union. Unfortunately, both pieces soon died after the operation. E Some Comments on Anterior Regeneration Although this subject was pretty thoroughly studied by Bulow and v. Wagner, there are some points that have not been considered at all. In the formation of the new head, abnormalities are not infre- quent. Double heads, arising immediately from the cut surface; or from a. common stalk a little distance from the cut surface occur in about 5 to 10 per cent of the pieces. At the posterior cut sur- face, on the contrary, double malformations are of very great rareness. ‘The only instance I find recorded in the literature is that spoken of by Bulow, of a worm developing a double tail, and another similar instance which I found last summer. From the posterior cut surface there grew out two tails of somewhat differ- ent lengths, and the whole worm had a Y-shaped form. By observing the process of regeneration of an abnormal head one will be impressed by the constant movements that are going on inside the body-walls in a forward direction. ‘These exert a great pressure upon the delicate regenerated epidermis, causing it to protrude in many points. ‘This action may be largely responsible for these malformations, for if, on the other hand, the malforma- tions are supposed to be due to the operation, why should we not find abnormalities in the regenerating tail also? I have never observed any malformation of the head of Lumbriculi freshly caught in ponds. Another point that I wish to call attention to is the dissimilarity between a regenerated and a normal head. “Das Vorderende ist immer etwas grinlich oder griinschwarz, was von dem Pigmente herrihrt, welches namentlich die den Darm- kanal bedeckenden Driisen erfiillt.”* This green pigment is arranged segmentally in seven to eight very deeply colored bands, which give to the head a striped appearance. If from one to seven of the anterior segments are removed, the number removed will be restored, but the substituted segments lack the pigment, and are ° Fr. Vejdovsky. 572 S. Morgulis perfectly transparent. If six to seven segments are cut off, so that one or two more stripes of the greenish pigment are left in the old tissue, the head will be perfectly repaired but not a single gran- ule of the old green pigment will be seen in the new segments, though I watched it for four weeks. Thus the two adjacent head segments, one from the old worm, filled with pigment, the other produced by the worm anew, pale and pigmentless, lie side by side. Heads regenerating from any other level of the worm’s body, far away from the pigmented region, are also without pigment. If worms under natural conditions do reproduce themselves by dividing into several pieces (four being the approximate average, calculated from Bulow’s experiments), should we not expect from this, that the majority of the worms in nature would have heads without pigment? In fact, such worms with pigmentless heads are not very frequent. It is true that after a lapse of a considerable time the pigment is formed anew in the regenerated head, and begins to develop from the distal end of the regenerated head. ‘This delayed development of the pigment in the regenerated head needs however a more complete study. REGENERATION AND ADAPTATION The tip of the tails in Lumbriculi is almost always missing, or regenerating, which fact indicates that this portion of the worm’s body is easily injured and that these injuries are of a very frequent occurrence. Onthe other hand the power of regeneration is very feeble in this particular region as compared with that in the more anterior and less frequently injured regions. If we attempt to find a connection bentenn these facts, we shall, contrary to Weismann’s claims, reach the conclusion, already expressed in 1898 by Prof. T. H. Morgan in regard to the frequency of accidental injuries and the power of regeneration in the hermit crab, that “no such relation is found to exist.” This low capacity of regeneration in a region where regeneration is always going on, seems to contradict the view that the capacity a hee & ~ —_— = Regeneration in Lumbriculus 573 to regenerate is an “adaptation of the organism to definite demands made upon it by conditions of life,” and that it is “not the out- come of primary qualities of the living substance,”’ “not an inherent quality of the organism,” as it also contradicts the view that it is due toan ‘adaptation produced by natural selection.’”” SUMMARY 1 Pieces of Lumbriculus containing only a single segment are capable of regenerating both a new head and tail. 2 Regeneration from a posterior end takes place more rapidly in pieces from the anterior region of the body, and gradually decreases as the pieces are taken from the more posterior region of the worm. 3 Apiece ofa worm, when subjected tothe operation of cutting a few times will produce more new tissue for the same length of time than when subjected to cutting only once. 4. Norelation whatsoever between the number of old segments in a piece and its rate of regeneration can be found. 5 There is no relation between the available food and the rate of posterior regeneration at different levels of the worm. 6 In regard to its regenerative capacity each worm shows variations of its own. ; 7 Regenerated tails, when detached from the old part, are capable of regenerating new heads, but do not produce any new posterior segments. 8 Pieces of such regenerated tails are also capable of posterior as well as anterior regeneration, from the posterior and anterior cut surfaces. g The pigment of the regenerated head probably does not arise in connection with the old pigment, but develops anew. 10 In the case of the anterior regeneration, where only six to seven (eight) segments come back, the eighth (or ninth) to the tenth (or eleventh) segments of the old worm are dropped out. tr The experimental evidence, likewise that from observations, 10.4. Weismann: The Germ-Plasm, 1893. 574 S. Morgulis is opposed to the view that the breaking off of pieces of the worms with their subsequent regeneration, is a regular mode of reproduc- tion in Lumbriculus. LIST OF REFERENCES Bonnet, C., 1745—Traité d’insectologie. Seconde partie. Observations sur quelqueses pices de vers d’eau douce, qui coupés par morcoeux, devien- nent autant d’animaux complets. Paris, Butow, C., °83—Die Keimschichten des wachsenden Schwanzendes yon Lumbni- culus variegatus, etc. Zeit. Wiss. Zool., xxxix. *83—Ueber Theilungs- und Regenerations-vorgange bei Wurmern. Arch. Naturg., xlix. CuiLp, C., ’06—The Relation Between Functional Regulation and Form Regula- tion. Jour. Exp. Zodl., 111. DrrescH, Hans, ’06—Regenerierende Regenerate. Arch. Entwicklungsm., xxi. Grouse, E., ’44—Ueber Lumbricus variegatus Miiller’s und him verwandte Anne- liden. Arch. Naturg. Hesse, R., ’94—Die Geschlechtsorgane von Lumbriculus variegatus Grube. Zeit. Wiss. Zool., v, 58. Iwanow, P., ’03. Die Regeneration von Rumpf und Kopfsegmenten bei Lum- bricus variegatus Gr. Zeit. Wiss. Zool., 75. Lerpy, J., ’50—Descriptions of some American Annelida abranchia. Jour. Acad. Nat. Sc. (Philadelphia), 2 series, 11. Morean, T. H., ’o1—Regeneration, New York. °o2—Experimental Studies of the Internal Factors of Regeneration Arch. f. Entwicklungsm. xiv. °o6—The Physiology of Regeneration. Jour. Exp. Zodl., in. Mrazek, Au., ’06—Die Geschlechtsverhaltnisse und die Geschlechtsorgane von Lumbriculus variegatus Gr. Zool. Jahrb., xxi. Ranpo.pu, Harriet, ’02—The Regeneration of the Tail in Lumbriculus. Jour. Morph., vu. Wacner, F. von., ’00—Beitrage zur Kentniss der Reparationsprozesse ber Lum- briculus variegatus I. Zool. Jahrb., xin. ’o5—Beitrage zur Kentniss der Reparationsprozesse bet Lumbriculus variegatus II. Zool. Jahrb., xxi. WEIsMANN, A., ’93—The Germ-Plasm, New York. Veypovsky, Fr.,’84—System und Morphologie der Oligochzten. ZELENY, CH., ’03—A Study of the Rate of Regeneration of the Arms in the Brittle- star Ophioglypha lacertosa. Biol. Bul., vi. a, a. Se eS Eee ee | ; CORRELATION AND VARIATION IN INTERNAL AND EXTERNAL CHARACTERS IN THE COMMON TOAD (BUFO LENTIGINOSUS AMERICANUS, Le C.) BY WM bs KEL LicoOw dt sean: W.tuH Six Figures AND Twenty-two TABLES MPL EAE AON LUT CEN DED Sere overs oy ow 25! Shs rs, ss ot Foes a vaceraije PRES SS Sig Gehans NE atau s IG Tee om Meee ae eee 575 MARS TrMe NELLY Oey cts els 2a ckata jai aioid ss 0s. asd Dayovae aoe SIZE aca a b,c etate See lavas alebelals sievers g casi eeels emia 576 eNMOLTE RCL ALAM et acc he soon eo Pass apne ch a Ona aS TavS ote cca CE ed Metres os gangs ae SE Me is oe Ges aR are 577 a IM IEGET IIIS 8 ais ee ne ie Rene ne Rc cy owe tice ten CI hore ee ore ene Sion od mare Pore et ae 577 PRUNES LYCNLS 8 PE ope fore cy Tee cack ce hes She ef aw ale lene 5, NS nae ces ae ieee Meee ee Ne 578 Bee ernie nical ratio) betweenithe:SCXES |e. sok case sae UR seas 607 mee Coniparative couelatiom of the SexeSe. 5 = sae = ie ee See eee eee 607 b Comparative correlation of external and internal characters............-.....-... 608 “Yl Diteerea ranean Sarl See Ses Se ote es CEN A lee tetas eR mirc Poet oe 612 I INTRODUCTION The present study was undertaken with a view toward getting information first as to the degrees of correlation and variability of internal characters, such as the viscera and muscles, regarding which our knowledge is remarkably deficient, and second as to the general condition of correlation among a considerable number of characters in a single group of individuals. It has been pointed out frequently that selection acts not upon single characters or variations alone, nor as such, but upon the entire organism as the unit: that there results from this action a “balance of fitness”’ THE JouRNAL or EXPERIMENTAL ZOOLOGY, VOL. IV, No. 4. 576 Wm. E. Kellicott among a large number of characters—in fact, throughout the entire organization. ‘This condition of balance is to a certain extent measurable by coefhcients of correlation which are merely indices of the corresponding change in a given character accompanying a certain change in another given character—‘ relative’ and “sub- ject.” The ideal method of approaching this matter is of course through multiple correlations involving the relation of at least three characters but this method is not practicable because of the time involved in carrying through such calculations. Consequently the method adopted here is merely to calculate the coefficients of cor- relation between all of the measured characters in pairs. The material from which the data were drawn was unusually favorable for such astudy. ‘Toads were readily secured, were not affected by capture and were easily kept in nearly normal condi- tions during the brief time elapsing between their capture and measurement. ‘The group studied was a perfectly homogeneous _ one, all in excellent nutritional condition, with precisely similar environmental conditions, and quite isolated geographically. The number measured and weighed was not large absolutely (425) but the fact that practically an entre colony of the particular variety under observation was collected and measured is almost unique. ‘The results are therefore at least free from errors due to the sampling of a larger population, errors which easily may be so great as to vitiate results even though a most careful preliminary study of the organism and its habits may have been made and particular care taken to procure a “random” sample. The fact that the subjects were of various ages, the growth phe- nomena not being taken into account, renders the data given here of no value for certain special purposes, but for the relations which are being sought here this is not an objection. We are dealing here with the conditions of correlation and variation in a total natural society of normal individuals. II SUMMARY The material under consideration consisted of practically an entire colony of the common toad. Accurate measurements were 0 —— —————— eee Correlation and Variation in the Toad Wy: taken of thirteen characters, including both external dimensions and internal organs. ‘The numerical ratio between the sexes was found to be 658 males to 1000 females. ‘The sexes are perfectly distinct with respect to size, variability and correlation. ‘The females are on the whole about 24 per cent larger, 23 per cent more variable and 1o per cent better correlated than the males. ‘The internal characters are about four times as variable as the external. The ratios between the average values of pairs of characters remain the same in the two sexes. ‘The distributions are all skew and nearly all negatively, apparently the result of including individ- uals of all ages over one year. The correlation coefficients are all relatively high. ‘The exter- nal characters although less variable than the internal, are in the males 54 per cent and in the females 30 per cent better correlated than the internal characters. “Those individuals above the aver- age in any pair of characters show much less “scatter” about the regression line than those below the average. In the general discussion these results are compared with those of other observers. Some of the points mentioned are the relation between efficiency and mass or dimensions in external and internal characters; the extremely high variability of internal characters; the relation between variability and correlation. , The conclusion is reached that from the side of fitness or sur- vival conditions of correlation here seem to be more fundamental’ than conditions of variability, and that the general subject of cor- relation is of increasing importance. DTS EVE DATE I Material The subjects from which these datat were secured consisted of a society of 441 individuals of the Common American Toad (Bufo lentiginosus americanus, Le. C.) ‘The society was one found on Cedar Point, Ohio, a low sandy point 200 to 300 yards wide extending for six or seven miles obliquely into Lake Erie and 1The data were secured during the summer of 1905 at the Lake Laboratory of the Ohio State University. 578 Wm. E. Kellicott partly enclosing Sandusky Bay. ‘Toward the extremity of this point a colony of toads has become established the individuals of which differ in several minor respects from those of the main-land. The colony is fairly isolated geographically and inhabits the lake beach for a distance of about a mile and a half. During the day- time the toads lie buried about two inches below the hot surface of the sand and only a few feet or yards back from the water’s edge. Shortly after sunset they uncover and hop down to the water to pick up food carried in by the usually light wash. Food is more than abundant throughout the season and with no particular exer- tion all are able to maintain themselves in an extremely well fed condition. As it becomes fully dark they assemble in this fashion, sometimes just within reach of the water and at the beginning of the season were picked up easily in considerable numbers. Dur- ing the summer almost daily collections were made and toward the Jail subjects became so rare that ultimately only a dozen or two could be found during an entire week. The data therefore were drawn from practically an entire popu- lation occupying a uniform stretch of sandy beach about twenty feet wide and a mile and a half in extent where all were subject to conditions which were remarkably uniform though somewhat unusual for creatures of their kind. And while the total number of individuals measured was not large absolutely, yet it represents nearly 100 per cent of this homogeneous group and consequently the observations are free from errors such as result from the sampling of a larger population. 2 Methods All the toads one or more years old were collected: there is no way of determining their exact age. Since such characters as total weight or weights of viscera are liable to modification by the unusual conditions of confinement, care was taken not only to keep these conditions as nearly normal as possible but also to measure animals only recently removed from their natural sur- roundings. Immediately after collection in the evening the toads were placed in a large sand-box and no more were taken than could jew ee ee Correlation and Variation in the T oad 579 be measured the following day. In no case was any measurement taken from an individual which had been more than; twenty-seven hours in captivity. Fig.1 Ventral view of female toad in position for measurement. One-half natural size of toad weighing 41.0 grams. The characters measured and methods of measurement are briefly as follows: (Compare with Fig. 1.) 1 Total weight. The toad was brushed clean and weighed alive. There was a gradual loss in weight during captivity chiefly due to defecation and evaporation of water. The rate per cent at which this loss occurred was determined in four series of different sized toads and the proper correction made for each individual according to hours elapsed from time of collection. Usually upon opening the abdom- inal cavity the bladder and alimentary canal were found empty; in the few individuals where this was not the case, proper correction was again made. The toad was then pithed (brain and cord) a broad incision being made to permit free loss of blood. While taking the succeeding measurements the toad was placed on its back and the blood drained off completely. For the next measurements the legs were placed in the position shown in Fig. 1. 2 Length of body. From tip of nose to end of body between thighs. This of course includes the head. 3 Length of thigh. One-half the distance between the middles of the knee-joints when in the position shown in the figure. 4 Lengthofshank. Distance from middle of knee-joint to middle of ankle-joint. 5 Length of foot. Distance from middle of ankle-joint to tip of longest (second) toe when fully extended. 6 Length of leg. Sum of thigh, shank and foot. 7 Totallength. Sum of body and leg. 8 Width of mouth. Transverse extent between angles of mouth when closed. 9 Length of head. Distance from tip of nose to postero-dorsal margin of cranium exposed by pithing incision. 580 Wm. E. Kellicott 1o Weight of gastrocnemii. The ankle-joint was completely flexed and the tendo Achilles sectioned in a radius of the joint. The attachment of the muscle to the femur was then cut along the bone and the muscle removed. Both muscles were weighed on a balance sensitive to one milligram. They were handled only with forceps, by the tendo Achilles and were not allowed to dry. 11 Weight of liver. The abdominal cavity was then opened and the liver removed by cutting through its attachments and blood-vessels closely along its surface. It was then gently rolled in a towel until it ceased to stain and weighed as above. 12 Weight of ovaries. These were removed in same manner as liver except that being practically bloodless, they were not rolled. 13 Length of alimentary canal. The trunk was then completely divided, the mesentery cut through and the alimentary canal straightened out to its full extent by pulling gently. A condition is soon reached where no farther stretching occurs and the length was measured when this point was reached. This character may be considered as little subject to error in determination as any visceral character, since a number of tests showed that this method of measurement gave a very reliable datum, much more reliable than was expected. A résumé of the data will be given first as briefly as possible, discussion of their significance being deferred until they have been completely presented. 3 Numerical Ratio Between Sexes Preliminary examination of the data shows at once that the sexes must be treated separately. Of the total number measured 173 were males and 252 females but these numbers are not quite indicative of the actual ratio, as sixteen additional individuals were collected and used for purposes that might have affected the values of some of the internal characters and which therefore were not measured. Of the total 441 collected 175 were males and 266 females, giving a ratio of 658 males to 1000 females in the entire colony. 4 Variability The means and standard deviations of all the characters in the male and female series, calculated according to the usual methods? are given in Table I and the coefficients of variability in Table II. 2Formule and methods from Davenport (04). 2 WEF) a “= Ey = ££ 0.6745 —— Z Vn ISS o ~ (#*. f) Eg = + 0.6745 mera Ae = G Cc C 244 Cc = x Ico Eq = 4 0.6745 1 Sr 2 eee M V2n Hee Correlation and Variation in the Toad 581 TABLE I Means and standard deviations of all characters Total weight, grams... Total length, mm..... Length of body, mm. Length of thigh, mm . Length of shank, mm. Length of foot, mm... Length of leg, mm.... Length of head, mm.. Width of mouth, mm.. MEAN STANDARD DEVIATION On=173 2 n=252 On=173 | Yn=252 yigdonenaaea 33-54010.364 52.560+0.617 7-105 0.257| 14.526+0.436 tele cy skarayateke 160.405+0. 518/183 .690+0.628| 10.107-0.365) 14.783+0.444 Spree OR aiejerhs | 67.871-0.221) 78.1130.292) 4.319-0.156| 6.877+0.207 SNe Se eas 29.575+0.101 33-690+0.098| 1.970+0.071| 2.538+0.076 BteraNetes forte 23 .020+0.080 26.325-+0.094 1.560+0.056, 2.224+ 0.067 meiieea-!-'s)- =|) 405840=1-0.133) 46.500-- 0.155 2.594-0.094 3 -657+0.110 SoS To PARE 92.8120. 309) 105.990 0.355 6.0350.218| 8.359+0.251 sctuodasoooan 18 .000+0.045| 20.052-0.051, 0.872+0.032, 1.205+0.036 Seles ieiat = sistahe 23.2020.074| 26.650+0.087) 1.437-0.052| 2.044-+0.061 Length of alimentary canal, mm . . . 353-873 1-779/429.603+2.253, 34.67041.255, 53.0181.593 Weight of gastrocnemii, decigrams... 6.8320.086 10.635+0.131 1.673+0.060 3.0770.092 Weichtobliver, prams,....2..25. . 1.5900.027, 2.233+0.034) 0.530+0.019| 0.794--0.024 Weight of ovaries, grams.......... 5-253-0.209 | 2.795-0.145 TABLE II Coefficients of vartability Oo 1=173 | o n=252 c E. c le Bisset iv ONC eaeQot alae oye (otsier aides inia lei stetcicterd aeuaeve wae e155 21.18+0.80 | 27 .640.89 GigE 365 Soa ace O RB RRS Cece Seen 6.30+0.23 : 8.05+0.24 “LODgHNO! HOG) cody. ca SOs Se OnE DUDE OS ACRE pean Oae 6.360.23 | 8 .800.27 Heeneewiopetbiph. 26.2. cases. eNO See 6.66+0.24 : Tes ja=O-23 ME WCENGSHan operetta ater cee ase Me iclsiends siete ee Se 6.780.25 | 8.45+0.25 MPS ENO ROO Eafe 19 S70) 21 oye) oicyens’sre a anaiovare wists) ase e/ei b's 6.35+0.23 7-8610.24 PRE TE DEEN GM Deets a tctay si slsieiecthe)sfefspt adele vieitia, «le soin’s Slavs 6.5010.24 | 7.8940.24 “ERPHDO! NODS SoS nO Re ano Eee ee eee 4.8210.17 6.0o1+0.18 “a2 G) rn GROSS GOED See eee ae eae 6.20+0.22 7-6740.23 Meuriitondlimentary/canals...) 2-120 + sic ee nines g.810.36 12.34+0.38 Bier iotrastlOCnem ti. oiac-fs (5.5 cfs .si ese inte e's os se = 24.500.94 | 28 .940.94 TIRE TA Git: 5 Se igen he A eee 2334-8 33-552: 1.20 BELA E GW GIES Abd SOCEM EOE Cree clea 53 -233-39 These tables show several facts of interest. First, the absolute distinctness of the male and female series as regards size, the female showing the larger values in every character measured. On the whole the female is about 24 per cent larger than the male, the actual 582 Wm. E. Kellicott percentages varying from 11.4 in length of head to 56.7 in total weight. This same distinctness between the sexes was found in the calculation of all constants and consequently they were treated separately throughout. Second, there is a corresponding distinctness between the males and females with respect to the amount of variability. The coefh- cient of variability may now be accepted as a thoroughly reliable measure of comparative variability. As expressed by this coefh- cient the females show a uniformly higher variability. On the whole the females are about 23 per cent more variable than the males, the actual percentages varying from 6.6 in weight of liver to 38.4 in length of body. Finally, inspection of the coefficients of variability shows that the characters are separable into two quite distinct groups. The external dimensions show not only a comparatively low degree of variability but a remarkable uniformity, the limits of C being, with a single exception, 6.0 per cent and 8.8 per cent. The other characters, which we may call “internal’’ are several times more variable and show much less uniformity, the limits of C being 9.8 per cent and 53.2 per cent. Excluding the total weight, total length and length of leg as composite characters, we find that the average coefhcients of variability of the external and internal char- acters are as follows: rot 2 Per cent Per cent External characters = Jac. soa disc cee ots aie he aia POO ae 6.2 I Internal characters js ster oe dejcsjeteciowie sists sao Se Has esis ea Para 22.5 4245 That is, the internal characters are roughly four times as variable as the external characters. It is an important consideration that, in spite of the very consid- erable disparity in absolute size of the two sexes, the ratios between the average values of pairs of characters are remarkably similar in male and female series. Table III gives a few of these ratios between characters measured in similar units. The ratios are extremely close throughout, particularly those between charac- ters which are obviously quite closely related in function, the body and leg dimensions, for example, where the ratios are the same OL | ‘ | 5 § . Correlation and Variation in the T oad 583 through the second decimal place. ‘This holds true for all of the external dimensional characters. namely, length of body, thigh, shank, foot, leg, head, width of mouth and total length. ‘The ratio between weight of gastrocnemii and total weight is the same in both sexes but as related with the other characters the gastrocnemii are much lighter in the male. So too the liver is proportionately lighter in the male with respect to every character except total weight. ‘The alimentary canal is, on the contrary, longer in the male except as compared with total weight, weight of gastrocnemi and liver. ‘The ratio between total weight and other characters is quite inconstant: the males are lighter with respect to total length, length of body, foot, leg, head, and alimentary canal and heavier with respect to length of thigh, shank, width of mouth and weight of liver. TABLE III Ratios between means of certain characters in the two sexes Ratio between average rol 2g Hematmoimsodyand total lengths ojo. es oia.c-< oie) ers ctheyessrave.oso"pwleraiol a w/e oats syeiave 423 425 HeaneMo mT AnGitOtalleEN Pte. sasicle case's caei eyacsicne ies ss niere sees Sern seanets 184 183 MenoiMonshankcanctotalilengthy, i. vse sine es coor css atmncawsr boss 043 143 Menprloiiooband totalilengthyic. 5:6 ssc. :o:c ene «lesen aye siectiese oi Mieiereimieiete Fesgenes 253 255 Menernowlerianditatallengthy:. . cts ssisicud « ae:ecade)4 sis 3.5 ss aiviavete tie eudiene rset -578 -577 Menprneubead andtotalilen pth ...25 0.036 o's «vires ss = 2S nee Ta ase Sete 112 -109 Motalilenethandlenethof alimentary canal.........0.0...2.2 0 ccs eee e eee es ariel 428 menounctited drandi width Of MOUt 7.) 5.2% « 755 ibaa fair by earaoe 6a ods BOG eO doo Ade ec oe ama acnaT Sas man nNobero seas: -954 en atbihcad sme eiite sci haysrete oie siotere ele tere mists tetera neta ee a eee evag Mecano aiee ace cop eBncGna ta ooctn Soo mo oSUE OL oT cea Soa senouEnSS 764 Ni Grid tleree Giacolai (anil 4 age Oboe bnom aban SboNurc Gon obouc Donn psa 55n Gaur ses oc -813 \ ETT GIT aie on or Ge naa SAU ARP OnSUG Gute 5. oomo bo StmcToa one y eons ooosenhr -643 Length alimentary canal..............- MA Nei Man rer aeaC occ Oem o oo G a acre -366 IWret et OVarieS a sjoreicle 2 perey-¥= este = rates las oher eae aie ete wed aren ele msale dete alana tale Average all characters........00000 ec eee ec eee e ee cece cee case teens Cy/ fi) Average external characters..........0+0---eeecee cece cent ee cececscene -837 Average internal characters......-.- +--+ +++ esse cece eee nee tence eeees 505 te Correlation and Variation in the Toad TABLE XI Indices of correlation between length of leg and SPEED MrT ose oso eta (ore Seeherao gi Cavaiot cd a aie Di dees nae Doce ees Le BEGUN EST. G@ Cis ice 2S ea ean gn t eaon eR cP R n Sar t EAE a EMSS VEE PL Reh) on ct oleae oh, sloteys eins 1a Meg eee ota se ee ae LOREHD CTH Lage Oe ee ee SUSE Ea oe DineM wormed boAcde Wierceen en TEA TT Ccrteretetarey cast teu rcs) Ais cua e hc leid orale) cieucis Fineieaé eladea ee > Soeelen lees Lan DAD UE CE ocd ace COO ae a ee eet ee eee Re Nr SN ter PBC AUR ciao cine eciats fo ale ae dead cs ioe aise wsieia dhe MERE elev PEL AREAS TLC trent e UNE ANSS, eed onesie OPS leven t/a, sis hiaieiei cia etna eNO Acre ee Rarer SAB ELOGH EID Ic. eia) oi Sia\al \ea io sacs eiciuscis Glee ou'a rece ome ewe vee ol fe Neacd ess ONT LISTS 6 0.0.6 BESO eRe Ler Nes AA iI en ace Mae rere MALTA CHEAT VICANAl ce cysrs coc sins ais wsie's o veivis de sa ea oe ates ee Face Nene “ENTE DY BETES CoS SS ROG BE CESIAE REIN BISON RE OT Iran nee Rit eb eet MERA P ETAL CHALAGCEES eie'stsicisis ero aslo ie/-)4 Sioa cons SA ase Aes s wee eae ee FEV CLAG ERECTING CRALACLETS sc, o avai sicene'e ave & a cebeve:e aveseva Miyata: spaiehe aus olay -383 DMI OIE Serta si etais v.cfelcae sea eis sis atvtalacitin agaiaa angie: e can aela weveperste peters 2 PNM na ETA GIA GA CECT oc. rovcia) Fe %,ciei0 we ois sis{ss ers scn'e,o: ehetsniiasesaveyagel eters michel Seer 609 PAV ELAPCEXCEEMAl CHAT SGtGES s/<.c%fee as exe <5 si Sie co hb ele, eee niet ieee «eae 631 PMEVGR ReMUILET AN CHALACHEES® ox.'io:cteisle nels « 's's. bites e ole passe sleieciere a sii sleie te -383 TABLE XVI Indices of correlation between length of alimentary canal and ot SN Urest Tate Ucar Sialercisya sero; Warere abs stash tees o wiaveigie oS oie @.cicie siedvele a eaioaaiele nares -376 ARSENE RECA Tit MEER Toke A ToT Se staves sche oh Lopera cs avs, toa saeco obs tauah ease yotspnctevsleloueler ore tee 487 [Leal SOAK: 00 366 BA BEA SE Bb SUBD UNIO CC eRe eR ene Seve tis 472 SP reset era PRUE EEC Petes ee Foe Porc ee) sie eh Ta) voce, Seca Wy eters cite Bee nis, idee weeelaye ra eSNG Sie -499 MD eoep ea esEN AE oeeterereto ober ts lars cemeteer at lavas ade acs a, 3) 3. Ssb ta. di occas wave Die wo ntaate saree sis 452 ere RHO CMe Malte cision ciavese tetris fol tye sih ef sic,ce aha siete oveyurel Norm aaiaeee we aisle ele -366 MUSEO Ae PPM UN el 5) Sek Ty ay orto cl sisycie le a, ofsidia ne abe alone Roa a hse geeaec oe 466 MO Ser re ULTEAG Beeteeteso eran tare Svs LeP Fore says: 6. aste wr Nici ots, v's, siars io musi ciavatetom) gave Gama ese 325 Wi Ghlin DOUG. 63 co S BOOC EIB. 0 SAO ORICUL EOE DOA Se Smee Te Seti cence care 379 Wigigiit: Gest ROXen iis 68 ada Podned Goo DOROSOe Sen arn ORE er eBe bree rane oS .508 err Eva iy lament Te orera ooo otc ais ave ake cts) < nists ereldizcs, o's’ v.cie.Sisiandi ais wbeteclen sisters Oeis -383 VAMEh TEE OT ETUER. 5 Ac) OOS TENORS CREED Oar eae aaa nw PAU CuA MEAN CHALAGCELS. 45 02 =!ntetct-hajaucin oS 2s Go aiarsis.dele o siefaeieabeweyeaiet oie + cle -428 MELA TELL ACH ALACECTS :atojoiclaoyayclen« orc.>\e°s\ererare.eraiote cfeyer ative chart ee Mae 433 PAMELA Renin Cemal Chala CLerSecrerstate eee eiieieie sie oes Hite nidis ee eee ere Glasto s 383 O38 40 808 -738 -769 Sy pehi .671 750 731 -647 -695 -726 618 -702 -716 ay Sy) .660 594 Wm. E. Kellicott TABLE XVII Indices of correlation between weight of ovaries and 2 Weiphttotal fe. n sis utc cs tiptoe iste ei wt GRE ee RIO omae asians 2.7150 Oe aie er -838 EB enero AE etl oat Phare a steerage Sor phecactah AS Sap cee tga roe Maer age ie ee 82 shout se ie ene -699 Agen eth DOAY, ees eager e eh ei tiae te iter eet tee ee eee eee So c0 54-5 -756 Berit Ghnptn cys eset tiers ..5.02---2- 2002252 -ee = 38.21 Weipht of heart|(Greenwood)).4).- 55ers ae oe 17.42 Weight of Kidney (Greenwood)..........-..:-..0cs000--- 16.80 Weight of liver (Greenwood) o-20 e905 sac eee eer 14.80 Weight of heart (Reid and Peacock [Pearson]).............. 19.82 20.70 Weight of liver (Reid and Peacock [Pearson])............... 14.32 22.23 Weight of kidney (Reid and Peacock[Pearson]).......... : 20.49 2258 Capacity of lungs (breathing capacity) (Galton[Pearson])..... ° 16.6 20.4 Weight of brain (Reid and Peacock, Clendinning, Sims, Bis- ChOM, earcon jee s noes elias kos os cee ana setae eee 8.07t0 10.25 7.93 to 10.64 Weight ot/brain (Pearl).2sc.14. case nc | chen ee ae cies 7.59to 8.85 7.09 to 8.72 Swine— Number of Miillerian glands (Davenport and Bullard)....... 48.0 Toad— Weightot vastrocnemitecce. «sas sesame eee 2 oir 24.50 28.94 Weightobliverc:setcucte ios. eens «cat oee inca Seiten: 33-34 5455 Weiphtolovanesscesne tec ae sc ie ect ee eae aero 53-23 Hengthiof alimentary, canal .1.0-s5-cst so hoes ee ee ee 9.81 12.34 The great mass of data on variability is drawn from measure- ments of body-weight, stature, skull and other skeletal charac- ters of similar nature. Of these, excluding body-weight as a com- posite of both external and internal characters, the coefficients range from 5-57 to 3-15, an average of 48 different characters in both sexes giving 4.2. In other vertebrates we have essentially similar magnitudes ranging from 8.80 to 2.69. (Lonnberg (’93) gives body-lengths of 141 specimens of Petromyzon fluviatilis which upon calculation show a coefficient of 9.66.) Obviously there is an enormous difference in the variability of these external and internal*® characters. In the toad we have 5 The use of the convenient words “‘external” and “internal” to distinguish these two classes of char- acters is perhaps justifiable though not exact. As external we may include characters such as stature, length of limbs or limb bones, number of vertebre, indices and other skull measurements except per- haps capacity, the position, number or size of extermal parts, etc.; in general all such characters as func- Correlation and Variation in the Toad 603 seen that the external characters are about four times as variable as the internal characters, a relation similar in general to that in man where the internal characters are roughly four to five times the more variable. Brain characters hold an intermediate posi- tion being roughly only about twice as variable as external char- acters. It is also significant that the variability of pure functions, as far as they have been measured, is of the same general magnitude as the variability of these internal organs. “Table XXI summa- rizes data collected by Pearson (’97) from various sources. TABLE XxI Coefficients of variability of functional characteristics in man of 2 Rrghteencrotinands(POrter i. .je:- 3.0.20 wie stein ojdie ci ecw aid mnie eaie a9 eh los 20. 30t0R 7009) S2eA toa cess Synrexeoenrand’ (Galton)... aie. oss Pers osv's ctenieee 3 aay 13.4 iy t Squeeze of hand (Cambd. Anthrop. Com.)................2.... 13.64 (r) 18.42 (r) 14.55 (1) 18.78 (1) Strength of pull (Cambd. Anthrop. Com.).............2.....05- 15.58 16.72 Siena ai pull (Ceo) Soaaasen Ane Sore eee eeere ne 15.0 19.3 Suivemgiln air jell (QA WalS®) Bape eaocisns hares oer ira see ern eA tena 15.32 22.62 iseenunessomeyesint (Galton ).-.sc.00. sence nse tienes owen 28.68 a2 Keenness of eyesight (Cambd. Anthrop. Com.)...............-- 33-25 32-93 (1) 34-73 (r) Werusalteedsitivity, (GaltOn) ices acess es voles cede nese wee en 35-70 45-75 Srmiienecnratblonw, (Galton) 105 sor vactn acon fh esis sl se oases a 19.4 ce As an explanation of this difference in variability between inter- nal and external characters Pearl (’05) has suggested tentatively that the greater variability of the internal characters is due partly to the fact that they depend to a very considerable degree for their value upon the general metabolic condition of the organism as a whole at the time of measurement, and partly to the fact that in visceral characters the thing measured is not the thing with which natural selection, as far as it has acted at all, has had to do. With respect to the first suggestion it might be objected that while this tion ‘‘ passively,” 7. e., whose value to the organism lies chiefly in position, or in a numerical or dimen- sional relation. As internal we should include organs whose function is of a more ‘‘active” sort—mus- cles, nerve-centers, glandular organs of all kinds, etc., structures whose value to the organism lies in a metabolic rather than a mechanical relation. ‘The distinction is not precise, certain organs may possess values of both kinds, and yet the distinction is broad enough to be useful. 604 Wm. E. Kellicott may be true for such characters as the weight of the spleen or liver it would be less true for such as the weight of the heart or brain or length of the alimentary canal. And farther, the values of many external characters are easily subject to modification by the general metabolic condition of the organism, particularly during the early period of growth (e. g., Vernon (’95)) and that many oa these modifications may be overcome by compensatory growth should conditions of life change. The second sugges- tion carries more force. It amounts practically to saying that at present we have no means of measuring the actual value of these internal characters which approaches accuracy. A measurement of the mass of the liver, for example, gives no exact information as to its functional worth. There is no legitimate reason for suppos- ing that there is an exact ratio between the size of a viscus and its functional value, indeed the ratio between mass and efficiency may be inverse and the better the tissue is functioning the smaller need its mass be to carry on its work in the life of the organism. In the statistical study of the characteristics of animals it should be borne in mind constantly that it 1s the functional value of a character to its possessor which is the bearing point of natural or any other ‘form of selection. It makes no difference to a toad how long his legs may be but only how far or how fast he can jump. If the length of the legs or their segments is an exact indi- cation of their ability to function, then only are we justified in using their lengths as data of actual evolutionary significance. Simi- larly, to the individual toad the weight alone of the liver is a matter of no consequence, only its ability to secrete and metabolize in both qualitative and quantitative relations, and unless we can demonstrate a close relation between bulk and efficiency, again the data themselves do not afford material for study of the evolution- ary significance of its variability in weight. That there is a very close relation between the dimensions of external characters and their functional value is probably true but the fact of their low degree of variability is not sufficient alone to prove it. The segments of the legs, for example, form a system of levers whose action depends largely upon their relative lengths and proportions. ‘Their value lies almost wholly in their purely mechan- Correlation and Variation in the Toad 605 ical or dimensional relations. This is indicated by the fact that the ratios between the average values of external characters are the same among individuals of various sizes. Table III showed the exact correspondance between these ratios in the male and female toad, and Donaldson (’98) has thoroughly demonstrated the same fact in a series of bullfrogs of various sizes. In frogs of all sizes and both sexes the sum of the leg bones is a nearly con- stant fraction of the length of the entire frog, and the proportional lengths of the several bones are also nearly constant. So it is for most external characters. But with internal characters there is good reason for questioning the exactness of this relation between size and efhiciency. As a matter of fact there are almost no data bearing upon the subject from which to form an opinion. It is generally believed from the evidence so far produced that there is no exact relation between size of brain and intellectual ability. There is a fairly close correlation between size of head and size of brain but Pearson (02a) was unable to find any appreciable correlation between size of head and intellectual ability. The only data which I have been able to find comparing the value of a function with physical measurement of the organ functioning is in the case of muscle. Weber’s law that the absolute power of a muscle is proportional to its cross-sectional area (/. e., to the number of fibers) is true only in the most general way. Weber found the absolute power of the human gastrocnemius and soleus to be about 1 kg. per sq. cm., while others have found values of 6.25 and 8.0 kg. in other human muscles. In the frog’s muscle Weber gives 0.6 kg. per sq. cm. as the absolute power while Rosenthal (’67) gives 2.8 to 3.0 kg. for frog’s semimembran- osus and adductor magnus and 1.0 to 1.2 kg. for the gastrocnemius. Howell (05) summarizes the matter by stating that “the absolute power of a frog’s muscle of I sq. cm. cross-area is estimated at from 0.7 to 3.0 kgs.’’—rather wide limits. Indeed when we con- sider the number and complexity of both external and internal factors affecting such a comparatively simple process as muscular contraction it seems useless to attempt at present any exact com- parison between size and efficiency. Many of the phenomena of 606 Wm. E. Kellicott muscular contraction show that the efficiency of a muscle depends after all fully as much upon the activity of the nerve centers con- trolling it as upon the characteristics of the muscle itself. The coefficients of variability of strength of pull, squeeze of hand, etc., are measures of the variability of the central nervous system as much as of the muscles involved in the action. And when we come to consider the action of absorbing surfaces or secreting organs it is simply impossible with our present knowledge and technique to make any more than the most general statements about the relation between the size of such organs and their functional value or efficiency. One farther consideration suggested above bears directly upon this matter. It is well known that the action of the central nerv- ous system determines to a very considerable extent both the quantitative and qualitative results of the action of metabolizing organs. ‘The functional activity of the digestive glands for exam- ple, is thus constantly modified without there being any detectible physical alteration (Pawlow and others), both the nature and amount of their secretions depending upon the action of the nervous system. The accurate adjustments of variations in the activty of these organs depend not upon the physical characters of the glands but upon the modifying influence of the nervous system in producing slight modifications of their internal metabolic proc- esses. This of course is a factor entirely lacking when the efh- ciency of an organ is directly dependent upon its relatively fixed dimensions. We must conclude therefore that measurements of the mass and dimensions of internal organs give data which can be used in an exact study only of these organs themselves without reference to their functional value to the organism as a whole: that they do not furnish evidence as to the precise efficiency of the organs, nor as to the effects of natural or other form of selection upon them in their functional relation; in this respect such data have only a general, not a precise, significance. The extremely high degree of variability of visceral characters may then, it seems to me, be also in some part the result of a fact that lies at the basis of many of the practices of modern surgery, Correlation and Vartation in the Toad 607 namely, that such organs rarely function to the limit of their capacity. Itis well known that large portions of many of the inter- nal organs may be removed without causing any serious or some- times even any visible disturbance of the physiology of the organ- ism. ‘The entire spleen, nearly the entire thyroid, or ovary, one entire kidney, and even large parts of the brain may thus be removed without visible effect. “This it seems can only mean that ordinarily such organs are functioning only in small part, that they are working with a large margin of reserve; that their func- tional value is not determined by their size. 3 Correlation a Comparative Degrees of Correlation in the Sexes In the discussion of this subject we are again limited practically to human data. ‘The relation here is quite similar to that of variability, 1. €., there is no uniform difference between the sexes but in general the female is perhaps slightly more perfectly corre- lated than the male. Of 29 pairs of coefhcients among individuals of the same societies collected from various sources the females show higher correlation coefhicients in but 14 and the average degree in the female is .442 as compared with .439 in the male— probably a non-significant difference. In swine (Davenport and Bullard ’96) the coefficients between the numbers of Millerian glands on the right and left sides of the body are .783 — female and .772 — male,a barely significant difference. Among the fishes the difference is somewhat more marked. Duncker (Ver- non ’03) states that in 40 pairs of coefficients 17 were unaffected by sex while in 11 the males, and in 6 the females were the more highly correlated. In the toad however, in correlation as in variability, there is both a decided and a uniform difference between the sexes. In 63 of the 66 pairs of coefficients (in two of the three exceptions the difference is less than the probable error) the female coefh- cients are higher than those of the male, the average degrees being -779 and .727, respectively. We see then that in the toad the females are about 10 per cent better correlated than the males. Here then we have a relation 608 Wm. E. Kelltcott that may underlie the fact that in the same community where con- ditions of life are remarkably uniform, the females should be so much (23 per cent) more variable than the males. May it not be because the females are at the same time more perfectly correlated f Variations from the type of a given character are not disadvantage- ous because they are backed up by corresponding variations in other characters; the “balance of fitness” is maintained in the more variable individuals. The females remain more variable—are not selected down to the same level of variability as the males—simply because their variability involves not single parts or organs but eroups of organs, in fact the entire organization. In other words, the organization of the females, abmodal as well as modal is more nearly a unit, the elements are better organized, less independent of one another, 7. e., simply “better correlated.” Schuster (’03) found a relation between sex and correlation in his crab measurements (Eupagurus) but there it was the males which were the more variable and more highly correlated. There seems therefore frequently to be a relation between sex and correlation as well as between sex and variability. Whether the relation is actually between variability and correlation will be discussed presently. b Comparative Degrees of Correlation in External and Internal Characters In the toad the average degrees of correlation in both sexes of the external and internal characters are .815 and .579, respectively (Table XVIII), and it should be borne in mind that it is the inter- nal characters which are the more variable. In man, to which we are limited for comparative data among vertebrates, the rela- tion is similar, as shown in Table XXII compiled from a number of sources. And here too the internal characters are the more variable. The coefficients of the brain are similar in general to those of the viscera ranging between .17 and .40 while the skeletal characters, excluding the skull, show an average correlation of .70: the skull resembles the brain in this respect. In general these coefficients are comparatively high in the toad, both in external and internal characters: I know of no form in Correlation and Variation in the Toad 609 which the coefficients are uniformly as high as here. The relation between the correlation of external and internal characters bears out the general conclusions reached in the discussion regarding their variability. The relative correlation of the brain in man has been mentioned by Pearl (’05), who points out that its corre- lation coefficients are of the same general magnitude as those of visceral characters. It should be noted however that the skull coefhcients are of this same magnitude, quite unlike the other TABLE XXII Coefficients of correlatien—internal and external characters Human— Elem aval UGS era ton oe ee eee ae Se nya ed .278 Hear last alge enics ces x tts Crh tae « hat Pinay ays tet ee ee Do eae 265 Reet ea HO IRTOMG YS eye. Master st pies co eet! gen. eate tres crepe sea opel eRtva Bee eee . 400 Brain and various skeletal characters, excluding skull.................... a7 tO 36 average = .21 mand warionis SKUL CHaractersé.-5.A122 s)-claale sins Me utoe tee eee 36 to .55 average = .47 Various skeletal characters, excluding skull, average..................... -70 Merriottss Ul MGHarACtersAVELACE.o/. . ..c.ccat os chee ea ye aes es daseaeinee: a6 Toad— RSs OETA ANIG LiVEl nts oe cf-nnsis.oisraee oe = ays Uae eye ease oagens ence Pe 696 CApocneMiandialimentary Canal res... es. zest rlsac celts aoe clo . 568 iG LOCDERIANG OVATIES Wo. LITBRATURESCIEED Bumpus, H. C., ’97—A Contribution to the Study of Variation. Skeletal Varia- tions of Necturus maculatus, Raf. Jour. Morph., xii, 455, 1897. Cuénot, L., ’99—Sur la determination du sexe chez les animaux. Bull. Sci. de la France et Belgique, xxxii, 462, 1899. Original paper not acces- sible—abstract by the author in L’Année Biologique, xv, 212, 1gol. Davenport, C. B., ’04—Statistical Methods. . 2d ed., New York, 1904. OO ee Correlation and Variation in the Toad 613 Davenport, C. B. anp Buttarp, C.,’96—Studies in Morphogenesis. VI. A Contribution to the Quantitative Study of Correlated Variation and the Comparative Variability of the Sexes. Proc. Amer. Acad. Arts and Sciences, xxxii, 85, 1896. Donatpson, H. H., ’98—Observations on the Weight and Length of the Central Nervous System and of the Legs in Bull Frogs of Different Sizes. Jour. Comp. Neurology, vii, 314, 1898. Futton, T. W., ’91—Observations on the Reproduction, Maturity and Sexual Relations of the Food Fishes. Annual Report of the Fishery Board for Scotland, x, 11, 232, 1891. Greenwoop, Jr., M., ’o4—A First Study of the Weight, Variability and Correla- tion of the Human Viscera, with Special Reference to the Healthy and Diseased Heart. Biometrika, i, 63, 1904. GriesHEem, A., ’81—Ueber die Zahlenverhaltnisse der Geschlechter be: Rana fusca. Arch. f. d. ges. Physiol., xxvi, 237, 1881. Howe 1, W. H., ’05—Text-Book of Physiology, p. 38, Philadelphia, 1905. . Lonnsere, E., ’93—Ichthyologische Notizen. Ueber die Variabilitat bei Petromy- zon. Bihang till K. Svenska Vet.-Akad. Handlingar, xvi, Afd. iv, no. 2, 1893. Minor, C. S., ’91—Senescence and Rejuvenation. First Paper: On the Weight of Guinea Pigs. Jour. Physiol., xu, 97, 1891. Montcomery, T. H., ’96—Organic Variation as a Criterion of Development. Jour. Morph., xii, 251, 1896. Peart, R., ’05—Variation and Correlation in Brain Weight. Biometrika, iv, 13, 1905. y Pearson, K., ’97—Variation in Man and Woman. Inthe Chances of Death and Other Studies in Evolution. London, 1897. ’°99—Data for the Problem of Evolution in Man, III. On the Magni- nitude of Certain Coefficients of Correlation in Man, etc. Proc. Roy. Soc., Ixvi, 23, 1899. ’o2a—On the Correlation of Intellectual Ability with the Size and Shape of the Head. Proc. Roy. Soc., Ixix, 333, 1902. ’02b—Mathematical Contributions to the Theory of Evolution, XI. On the Influence of Natural Selection on the Variability and Correlation of Organs. Phil. Trans., A., cc, I, 1902: ’o6—On the Relationship of Intelligence to Size and Shape of Head and to other Physical and Mental Characters. Biometrika, v, 105, 1906. Pritcer, E., ’81—Einige Beobachtungen zur Frage tiber die das Geschlecht be- stimmenden Ursachen. Arch. f. d. ges. Physiol., xxvi, 243, 1881. 614 Wm. E. Kellicott oe gee M. J., °67—Note sur la force que le muscle de la grenouille peut dévellopper pendant la contraction. Comptes Rendus, Ixiv, 1143, 1867. Scuuster, E. H. J., °03—Variation in Eupagurus prideauxi. Biometrika, ii, 191, 1903. Stronc, R..M., ’o1—A Quantitative Study of Variation in the Smaller North American Shrikes. Amer. Nat., xxxv, 271, 1901. Vernon, H. M., ’95—The Effect of Environment on Development of Echinoderm Larve. Phil. Trans. Roy. Soc., B., clxxxvi, 1895. aa, anation in Animals and Plants. New York, 1903. We pon, W. F. R., ’01—Change in Organic Correlation of Ficaria ranunculoides es the Flowering Season. Biometrika, 1, 125, I1go1. 2852 4 : Dh vent sy aN QL The Journal of experimental 2 zoology J68 Ve4 COp.ee Biological & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET i UNIVERSITY OF TORONTO LIBRARY uit alae “A wee ttshelachit he