Pee FOURN AL OF EXPERIMENTAL ZOOLOGY EDITED BY WILLIAM E. CASTLE FRANK R. LILLIE Harvard University University of Chicago EDWIN G. CONKLIN JACQUES LOEB Princeton University University of California CHARLES B. DAVENPORT THOMAS H. MORGAN Carnegie Institution Columbia University HORACE JAYNE GEORGE H. PARKER The Wistar Institute Harvard University HERBERT S. JENNINGS CHARLES O. WHITMAN Jéhns Hopkins University University of Chicago EDMUND B. WILSON, Columbia University and ROSS G. HARRISON, Yale University Managing Editor VOLUME VII THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA, PA, 1909 CONTENTS No. 1—August, 1909 ErHet N. BROWNE The Production of New Hydranths in Hydra by the Insertion of Small Erotesies WVitOIKty-SIX-Rigutes, 6.4... : x. ad’nasdge me eee ee ee ae I M. Louise SHOREY The Effect of the Destruction of Peripheral Areas on the Differentiation othe Neuroblasts.. With Fifty-seven Figtikes.....70.. 2... oes 25 C. M. Curtp Factors of Form Regulation in Harenactis attenuata. II. Aboral Resti- tution, Hetermorphosis and Polarity. With Twelve Figures........... 65 “Francis B. SUMNER Some Effects of External Conditions upon the White Mouse. With POT CEU gE GOMER oes oa sls foe 38 SS Sia iene + Staley a biases pe ee 97 Cuares W. Harcitt Further Observations on the Behavior of Tubicolous Annelids.......... 155 No. 2—September, 1909 HERBERT W. RAND Wound Reparation and Polarity in Tentacles of Actinians.............. 189 V T.H. Morcan A Biological and Cytological Study of Sex Determination in Phylloxerans » and Aphids. With One Plate and Twenty-three Figures in Text...... 239 C. M. CuiLp Factors of Form Regulation in Harenactis attenuata. III. Regulation Mar aiuaes With: | hitty-one FigGres., got aesann aves saws Tene sexe 353 No. 3—October, 1909 Marion L. DursBINn An Analysis of the Rate of Regeneration Throughout the Regenerative PROCESS NV IHL OLIN BAULES 5 Sori. « 2 lsvk s)-baneecgune seks Oss Sie Ene ame oie Max Mapes ELLIs The Relation of the Amount of Tail Regenerated to the Amount Removed in Tadpoles of Rana clamitans. With Three Figures................ OrEN E. FRAZEE The Effect of Electrical Stimulation upon the Rate of Regeneration in Rana pipiens and Amblystoma jeffersonianum. With Two Figures... ‘/ CHARLES ZELENY The Effect of Successive Removal upon the Rate of Regeneration...... The Relation Between Degree of Injury and Rate of Regeneration—Addi- tional Observations and General Discussion.................eceeeee Some Experiments on the Effect of Age upon the Rate of Regeneration. . . No. 4—November, 1909 Sercius Morcutis Contributions to the Physiology of Regeneration. I. Experiments on ovate opscita: With seven hipubes Wil ext. ..o.s04. 01 +. ooase cee A. J. GOLDFARB The Influence of the Nervous System in Regeneration. With Twenty- BAT CCARIDEER opie Ncharcintna Sr. Salk niche ease ce van ea eae ts Cena te oe MicwHakEt I. Guyer Atavism in Guinea-Chicken Hybrids. With Four Plates.............. IsaBEL McCracken Heredity of the Race-Characters Univoltinism and Bivoltinism in the Silkworm (Bombyx mori). A Case of Non-Mendelian Inheritance. With Pour: Tables andy Pwo Diapramis: 00... 0.95 <)aesis wane oe a 397 421 457 477 513 563 595 643 THE PRODUCTION OF NEW HYDRANTHS IN HYDRA BY THE INSERTION OF SMALL GRAFTS BY ETHEL NICHOLSON BROWNE Wirx Six Pirates INTRODUCTION During the winters of 1906-1908, I carried on some experi- ments in grafting Hydra viridis for the purpose of throwing more light on the factors concerned in regeneration. ‘The work was done at the suggestion of Professor Morgan, whom I sincerely thank for his kind interest and help. In my first series of experiments, which were done with the ordinary green hydra, I tried to discover what material when grafted would give ,the necessary stimulus to call forth the develop- ment of a new hydranth. It was found by experiment that if a tentacle with a small bit of peristome tissue at its base was inserted into the body of another hydra, the stock would regenerate a whole new hydranth at the place of grafting, the grafted tentacle remain- ing as one of the new circlet of tenacles. The question arose whether the regeneration was due merely to the presence of foreign tissue of any kind or whether it was initiated by some special kind of material; and if so, what kind of tissue this was. To solve this problem, tissue was taken from different regions of the body of Hydra viridis and grafted into different regions of the body of other hydras of the same species. In my second series of experiments, I endeavored to find out the exact origin of the regenerating material. An excellent opportun- ity to decide this question was offered by the discovery of Mr. D. D. Whitney that the green color can be entirely removed from Hydra viridis by putting the animals in a .5 per cent glycerine solu- tion and leaving them for about three weeks. These artificial Tue JourNAL or ExPERIMENTAL ZOOLOGY, VOL. VIIB NO. I. 2 Ethel Nicholson Browne white hydras form perfect grafts with normal green hydras and the two parts remain distinct in color. In all my experiments, I used for operation a small watch glass coated over with paraffine on the bottom and filled with spring water. ‘The operations were all done under a binocular microscope which has the advantage of giving considerable enlargement and plenty of working distance. After operation, the animals were kept in watch glasses filled with either spring water or aquarium water, renewed daily. Part I GRAFTS TO PRODUCE A NEW HYDRANTH FROM THE STOCK The object of this series of experiments was to determine, first, what tissue when grafted into the body of a normal hydra would cause a new head to regenerate in the region of the graft; and, secondly, in what region of the stock the graft must be made in order that the new hydranth develop. ‘To test these questions, material was taken from various regions of the body and grafted at different levels along the bodies of other hydras. All these experiments were done with Hydra viridis as both graft and stock. Group A Tentacle with Peristome at Base as Graft For operation in the present group of experiments, two hydras were put into the paraffine-coated watch glass and one was cut just below the circle of tentacles. ‘This circlet was cut in one radius so that the tentacles extended out from the peristome in a straight line or slight curve. From this line of tentacles [ cut off one, being careful to leave some peristome tissue attached at its base. As quickly as possible I made a small transverse slit in the body wall of the other hydra with a small scalpel or sharpened needle, and inserted the prepared tentacle into this slit by means of a needle. If the operation was successful, the raw surfaces healed so that after a few hours there resulted a perfectly normal hydra with the exception of a tentacle projecting from some part Production of New Hydranths in Hydra 3 of the body. In this way I inserted a tentacle with a small bit of peristome tissue at its base in different regions of the stock hydras with varying results. Series I Graft Made in Middle of Stock Result r. ‘The usual result following the graft of a tentacle with a bit of peristome tissue at its base into the middle region of a hydra was the outgrowth of a new hydranth from the region of graft, the grafted tentacle persisting as one of the new circlet. The result can be best shown by a specific instance. On February 19, a tentacle with peristome tissue at its base was grafted in the middle of the body of a healthy green hydra. After a few hours, the wounded surfaces had healed and there resulted a hydra perfectly normal except for the protrusion of a tentacle from the middle of its body (Fig. 1). Onthe next day, a slight outpushing of the body wall around the tentacle was observed. On February 21, two days after operation, this outpushing could be distinctly recognized as a new hydranth. ‘The new hydranth consisted of a short body protruding from the body of the stock, having at its distal end the large grafted tentacle and three very short tentacles (Fig. 2). On the following day these new tentacles were distinctly longer but not so long as the grafted tentacle. On February 23, the fourth day after operation, a small fifth tentacle had appeared and the grafted tentacle was still distinctly larger than the regenerated ones (Fig. 3). The difference in length of the tentacles was gradually lost by the further growth of the regenerated ones until on February 28, the hydra appeared as a double-headed hydra, one head bearing six tentacles and the other, the regenerated head, five; the original head was still con- siderably longer than the regenerated one (Fig. 4). That this new hydranth was functional and similar in all respects to the old hydranth was proved by the fact that it could capture and ingest food with as great ease as the oldone. ‘This double-headed hydra remained in the condition described without further elongation of the new hydranth until it died on March 14. Similar results have been obtained in ten other similar grafts. In some cases, how- 4 Ethel Nicholson Browne ever, the new hydranth grew longer than in the case described, so that it was of about the same length as the original head. The axial relations assumed by the new head in reference to the common foot were not constant for even the same hydra at differ- ent times. [he new head was sometimes at right angles to the stock hydra, it sometimes formed a right angle with the foot and a straight line with the original head, it sometimes formed an obtuse angle with the foot making a Y-shaped structure, and it was some- times at an acute angle with the foot forming a X-shaped figure. The new hydranth never showed any tendency to travel toward the aboral end of the stock hydra. ‘This is of interest in connection with Miss King’s experiments in grafting whole heads into the side of stock hydras. As a result of her work, she found that the graft either separated from the stock in from 14 to 22 days, or migrated toward the foot region and separated in from 5 to 7 weeks. These two modes of separation are due, she concludes from her experiments, to the “axial relations assumed by the components of the graft.” If the graft remains at right angles to the trunk, separation takes place without migration; if the graft forms a Y-shaped structure with the stock, it migrates toward the aboral end before separating. [he new hydranth that is formed in my experiment seems to act quite differently. It has no definite axial relations with the stock and does not migrate toward the foot. As to the final separation of the regenerated hydranth from the old stock, I can say nothing, for I have not succeeded in keeping these grafts more than a month after operation, and they have not separated within that time. — Result 2. In one case grafted February 19, after a slight outpushing of the body wall around the grafted tentacle and the formation of three additional tentacles (Fig. 5), a process of absorp- tion set in. On the seventh day after operation, the projection of the body wall to form a new hydranth was no longer visible, the four tentacles emerging directly from the body of the stock (Fig.6). Four days later, one of the four tentacles was absorbed, and just below the place where the tentacles were present, a small protru- sion on the body wall was noticed (Fig. 7) which on the following day could be definitely determined to be a bud (Fig. 8). The Production of New Hydranths in Hydra 5 other three tentacles were absorbed one by one, while the bud was developing, until on March 12, three weeks after grafting, the whole regenerated material had been absorbed and the bud had pinched off leaving the stock as a normal hydra. The departure from the usual result may be accounted for in this case by the formation of the bud. It may be that the stimulus for regeneration was present and excited the growth of new tissue as is evidenced in the outgrowth of three new tentacles, but when the bud developed, the available material was used for its develop- ment, thus depriving the regenerating hydranth of any means of growth. Result 3. In only two instances, no regeneration was stimulated by the graft of a tentacle with its basal tissue into the side of the stock hydra. But a gradual process of absorption set in, so that after two weeks no trace of the tentacle was left. It is possible that this result was due to the fact that none of the peristome tissue was left at the base of the tentacle, this bit of tissue being accident- ally broken off from the tentacle in the process of grafting; or, the results obtained in these two instances may have been an exception to the usual result. Series II Graft Made in Foot of Stock Result tr. The usual result following grafting a tentacle with a small bit of peristome tissue at its base into, or in the region of, the foot of a normal hydra, was the outgrowth of a diminutive hydra at the point of grafting. ‘The grafted tentacle was partially absorbed as regeneration took place, till it assumed the proportion proper for the small hydra. ‘This small hydra pinched off from the stock in three or four days after grafting, sometimes with only the grafted tentacle present, sometimes with one or two regener- ated ones. Following the history of one of these grafts, we find it as follows: On March 5, I inserted a tentacle with a bit of peristome tissue at its base into the foot of a normal green hydra (Fig. 9). On the next day, there was a very slight outpushing of the body wall of the stock carrying the grafted tentacle with it. On March 7, this 6 Ethel Nicholson Browne outpushing had increased and two new tentacles were just begin- ning to grow out from its distal end beside the grafted tentacle (Fig. 10). ‘The grafted tentacle had decreased in size noticeably. This outgrowth was distinctly different from that observed when the tentacle was grafted in the middle of the stock, being much smaller in circumference. On the next day, March 8, the out- growth could be clearly recognized as a minute hydranth with two short tentacles and one longer one, the grafted tentacle, which had become still further reduced in size by absorption (Fig. 11). In the afternoon of the same day, this small hydranth pinched off from the stock. On March 11, owing to the further absorption of the grafted tentacle and the further growth of the two regenerated tentacles, this hydra appeared as a typical hydra of very minute size, with three tentacles of about equal length arranged about the hypostome (Fig.12). “The volume of the small hydra was not more than one-tenth that of the stock from which it pinched off. In several other similar experiments, the same history followed except that the small hydra separated with only two tentacles present, the grafted one (reduced in size) and one regenerated one. In these cases, a third tentacle usually appeared after the hydra had pinched off. In still other similar experiments, there were no new regener- ated tentacles formed, but the small hydra pinched off with only one tentacle, the grafted one (Fig. 13). One or two tentacles were formed, however, after the small hydra had separated. Result 2. In two out of about twenty grafts in the foot region, the tentacle grafted was slowly absorbed. Result 3. In two other cases, abnormal hydras resulted from the graft of a tentacle in the foot region. An outgrowth from the foot carried the grafted tentacle along with it, as though to form a new hydranth (Fig. 14). But the tentacle was absorbed while the outgrowth enlarged, and no separation took place before the end of two weeks when these abnormal hydras died (Fig. 15). Series III Graft Made in Circlet of Tentacles of Stock The insertion of a tentacle into the circlet of tentacles of a normal hydra in no instance induced the formation of a new hydranth. The tentacle either remained as grafted or instigated the outgrowth Production of New Hydranths in Hydra 7 of one or two additional tentacles. In one case two new tentacles grew out two days after grafting a tentacle into the circlet, making ten tentacles in all (Fig. 16). A week later, three of these became fused at the base (Fig. 17). ‘The process of fusion spread until the three tentacles formed a single tentacle, leaving eight in the circlet, one more than the original number. Unless, however, as in this case, an abnormally large number of tentacles was present, no later absorption took place after regeneration. Series LV Graft Made Below Circlet and Above Middle of Stock Result x. In the majority of cases, this graft instigated the outgrowth of a new hydranth. ‘The history of these cases is similar to that which followed the graft of a tentacle in the middle of the body. A double-headed hydranth resulted from the graft, but the two heads were considerably shorter in proportion to the body than in the case of the middle-region grafts (Fig. 18). Result 2. In one case, after a slight outpushing of the body wall and the regeneration of two new tentacles (Fig. 19), this new hydranth coalesced with the old hydranth a week after grafting (Figs. 20, 21). There thus resulted a single-headed hydra with an extra number of tentacles. Result 3. In another case, no new tentacles were regenerated, but the grafted tentacle was included in the old hydranth by the shifting downward of the other tentacles (Fig. 22). Result 4. In one case, the grafted tentacle was absorbed slowly so that after two weeks, no trace of the graft was left. Series V Graft Made Between Foot and Middle of Stock Result 1. In most cases, this graft followed the same history as a foot-region graft. A minute hydra developed from the body wall of the stock, and pinched off in three or four days after graft- ing, when possessing only one or two tentacles (Fig. 23). Result 2. In several cases this graft gave rise to a new hydranth similar to that which is found in the middle region. Result 3. In one case, the grafted tentacle itself became a small new hydranth by enlarging; and on this former tentacle small 8 Ethel Nicholson Browne tentacles grew out (Figs. 24,25). By further increase in size and the formation of additional tentacles, there resulted a double- headed hydra, one of whose heads was somewhat longer and larger than the other (Fig. 26). Result 4. In one case, the grafted tentacle was slowly absorbed without any regeneration. From the foregoing group of experiments (Group A), it is evi- dent that in every region of the body, the graft of a tentacle with a bit of peristome tissue at its base may cause regeneration on the part of the stock. Moreover, the regeneration in every region except in the circlet of tentacles takes the form of a new hydranth, of normal size in the anterior and middle regions of the stock, and of minute size in the posterior and foot regions. Group B Two Tentacles with Peristome at Base as Graft in Opposite Sides of Stock In this set of experiments, I endeavored to find out whether a second tentacle with peristome tissue at its base inserted soon after the first in the same region of the body on the opposite side would cause a second new hydranth to regenerate. In some cases the second tentacle was inserted about two hours after the first, and in others it was inserted a day later, but both methods gave practically the same results. Result tr. In two out of six cases, both grafted tentacles caused the outgrowth of hydranths. One of these was grafted on March 20 (Fig. 27); two days later, there was a slight outpushing of the body wall of the stock at the base of each tentacle, a new tentacle having developed on one of the outgrowths (Fig. 28). On March 25, the hydra had two quite well developed but small hydranths in the middle region, each having two tentacles, a grafted and a regenerated one (Fig. 29). The smaller one of these was gradually absorbed, till on March 30, no trace of it was left, the hydra having only two heads, one with seven tentacles and the other, the regener- ated one having four tentacles, one longer grafted one and three Production of New Hydranths in Hydra 9 shorter regenerated ones (Fig. 30). By April 1, these four ten- tacles were of the same length, and by April 5, a fifthtentaclehad appeared, so that the hydra was a typical double-headed animal (Fig. 31). In the other similar graft, after the outgrowth of two small hydranths at the base of each grafted tentacle, they migrated from opposite sides so as to be adjacent (Fig. 32). A process of fusion then set in, so that ten days after grafting, the two regener- ated hydranths, each with three tentacles, were separate only at their distal ends (Fig. 33). ‘This fusion was complete two days later, resulting in the formation of a double-headed hydra. Result 2. In two other cases of similar grafts, only one of the grafted tentacles gave rise to a new hydranth, the other one being gradually absorbed (Fig. 34). Result 3. In the other two cases, both tentacles were absorbed without any regeneration. From this group of experiments it is evident that each of two tentacles with peristome tissue at their base, when grafted intothe middle region of a hydra, may cause the regeneration of a new hydranth. However, a process of fusion or absorption sets in sooner or later, so that only one of the regenerated hydranths remains. Group ‘& Tentacle Without Peristome at Base as Graft Having obtained so definite a response on the part of the stock to a grafted tentacle, I next tried to find out what part of the grafted tentacle was responsible for the regeneration. In this group of experiments I grafted just the tentacle without any peristome tissue at its base into the stock. This operation was somewhat difficult as the raw surface of the tentacle healed over rapidly and would not then adhere to the cut surface of the stock. About eight times, however, by performing the operation very quickly, I was successful. All the experiments gave the same result, whether the graft was made in the middle or foot region of the stock. There was no outpushing of the body wall of the stock, no regeneration whatever, but, on the contrary, the grafted tentacle IO Ethel Nicholson Browne was slowly absorbed so that after about ten days no trace what- ever of the graft was left. These experiments show that the graft of just the tentacle without peristome tissue at its base does not stimulate the regener- ation of a hydranth. Group D Peristome at Base of Tentacle Without Tentacle as Graft Methods In one set of these grafts, I cut off a tentacle with the peristome tissue at its base and inserted it into the body of the stock as described under Group A. ‘Then after the raw surfaces of the stock and graft had healed, I cut off the grafted tentacle close to the body wall, thus leaving grafted in the stock the bit of peristome tissue that was at the base of the tentacle. In another set, I cut off a circlet of tentacles, then cut off a few tentacles at their base, close to the circlet, and used a small piece of the remaining ring of peristome tissue. ‘This I grafted quickly into a previously pre- pared cut in the body wall of the stock. Both methods gave the same result. Series I Graft Made in Middle of Stock Result r. In three out of five cases tried, the graft in the middle region of the body gave rise to a new hydranth. ‘The outgrowth was similar to that initiated by the graft of a whole tentacle and peristome tissue at its base. The body wall of the stock at the region of graft pushed out and new tentacles were formed on it (Figs. 35-37). Result 2. In two other similar experiments, absorption took place and no regeneration occurred. Series IT Graft Made in Foot Region of Stock Result t. In two out of three cases in which the peristome tissue at the base of the tentacle was grafted into the foot region of the Production of New Hydranths in Hydra mp stock, there grew out a minute hydra, similar in formation and appearance to that instigated by a graft of a tentacle and basal tissue into the foot (Figs. 38, 39). Result 2. In the other case, the tissue was absorbed and no regeneration took place. CONCLUSION This group of experiments shows that the whole tentacle is not necessary for the production of a new hydranth on the part of the stock. But merely the peristome tissue at the base of the tentacle is sufficient, when grafted, to instigate the outgrowth of a new hydranth. Group E Tissue Anterior to Circlet of Tentacles as Graft The amount of tissue anterior to the circlet of tentacles is very small and it was found very difficult to cutit off without getting into the region of the tentacles. By waiting, however, until a large hydra was fully expanded and very quickly bringing the scalpel just anterior to the tentacles, I found it possible in five cases to get this small bit of tissue cut off. This was grafted into the middle region of the stock. In none of the five cases was any regeneration instigated. ‘The grafted tissue was absorbed, so that the day after grafting no trace of the graft was to be seen. The tissue anterior to the circlet of tentacles will not give the stimulus necessary for the outgrowth of a new hydranth, when grafted into the body wall of a stock hydra. Group F Tissue From other Regions of the Body as Graft In this set of experiments, I endeavored to find out whether tissue from any region of the body other than at the base of the tentacle, would, when grafted, give rise to a new hydranth. For 12 Ethel Nicholson Browne this purpose, I cut off a small ring of tissue from various regions of the body, including the region just beneath the circlet of tentacles and grafted it into the middle region of a hydra. These rings in no case gave rise to a new hydranth, but were soon absorbed. If, however, instead of a small ring of tissue, a large ring was grafted, the result was different. “The experiments were done in the following manner: A hydra was cut in two beneath the circlet of tentacles; the foot was then cut off from the lower part. The aboral end of this band of tissue was then grafted into the side of anormal hydra. On the following day, the graft had developed tentacles. In this case, however, it must be noted that the regener- ation is entirely on the part of the graft and not of the stock, and the stock takes no part in the formation of new tissue. The conclusions drawn from these experiments is that no other tissue than that at the base of the tentacles is capable of so stimu- lating the stock as to cause it to produce from its body wall a new hydranth. . Group G Regenerating Tissue as Graft This series of experiments was undertaken for the purpose of finding out whether tissue which has begun, in the process of regen- eration, to be differentiated into tentacle-forming material, would, when grafted, influence the body wall of the stock to regenerate a new hydranth. ‘The method adopted was as follows: A green hydra was cut in two at about the middle of the body. The pos- terior half was ther left till the following day when the wound had healed and the process of regenerating a new hydranth had started, although no tentacles had formed. A very small piece was cut off from the oral surface, and this was grafted into the side of a normal hydra. On the next day an outpushing of the body wall had occurred and a day later two tentacles had formed on the new hydranth (Figs. 40, 41). By leavingthe regenerating piece different lengths of time before grafting part of it into the stock, it was found that about ten hours was the minimum that would give regeneration. If left only seven hours before grafting, the graft Production of New Hydranths in Hydra 13 was absorbed and no regeneration took place. The length of time necessary for regeneration to be far enough along for a piece of the regenerating tissue to call forth a hydranth from the stock when grafted, would of course depend on the rate of regeneration from an exposed surface, and this has been found to vary in differ- ent regions of the body. If, therefore, the original cut were made just below the tentacles instead of in the middle of the body, less than ten hours would probably suffice for the regenerating tissue when grafted to call forth a new hydranth from the stock. Group H Tissue of Bud as Graft In this experiment, a small piece of the anterior end of a young bud whose tentacles had not yet been formed, was cut off and grafted into the middle of the body of a stock hydra. There followed as a result of the graft an outgrowth of a new hydranth similar to that instigated by the insertion of regenerating tissue into the stock (Figs. 40, 41). That this outgrowth was a new hydranth and not an ordinary bud is shown by the length of its tentacles. These were long like those of a hydranth and not short like those of a bud (cf. Figs. 41 and 8.) General Conclusions to Part I The conclusions to be drawn from the whole set of experiments are as follows. A new hydranth can be formed by a hydra in any part of its body except in the tentacle region, when the necessary stimulation is given by a grafted piece. ‘The stimulus can be given by no other grafted piece than the material that lies at the base of the free tentacles, or that lies in a regenerating hydranth or a bud. The transformation of body wall material into hydranth material depends therefore not on the size of the piece grafted into it, but entirely on the differentiation of the grafted material. If the material grafted has been entirely differentiated into material lying at the base of a tentacle, or if it has been sufficiently differen- tated by regeneration or budding into this kind of material, this 14 Ethel Nicholson Browne tissue will cause the body wall tissue of a normal hydra to change its differentiation and function and become tentacle and hypostome tissue. Part II ORIGIN OF REGENERATING TISSUE AND FATE OF ABSORBED TISSUE In the second series of experiments, [ tried to find out the source of the regenerating material and the fate of absorbed material in the foregoing and other experiments. Attempts have been made to solve these questions in some cases by grafting Hydra fusca and Hydra viridis, a brown and a green hydra, but these attempts have proved unsuccessful, for the two species do not graft well. Although the graft has been made to stick for a day or so, it always pulls away from the stock before any results can be obtained. Miss King has attempted to solve the question in her experiments by using light and dark green individuals of Hydra viridis. She states that she is able to distinguish the two shades for two or three weeks, at the end of which time they fuse. In combinations between the artificial white hydras produced by Whitney’s method and the normal green one the contrast between the tissue of the stock and that of the graft is very distinct and remains so for about a month. Group A Regeneration of Hydranth In order to determine the exact source of the material forming the new hydranth in the preceding set of experiments (Part I), the following experiments were done. Series | White Tentacle with Base Grafted in Middle of Green Hydra The result of this graft was in six cases the outgrowth of green tissue from the stock carrying the white tentacle with it, and the later regeneration of green tentacles (Fig. 42). The new hy- dranth material, then, must come from the body wall of the stock, while the grafted tentacle remains as one of the new circlet. Production of New Hydranths in Hydra 15 Series [11 White Tentacle with Base Grafted in Foot of Green Hydra The result of this graft was in four cases the outgrowth of a small amount of green tissue at the base of the white tentacle to form a minute hydranth (Fig. 43). This pinched off as a small green hydra with one white tentacle (Fig. 49). Four days after operation a second tentacle appeared on it, composed of green tissue; some of the white material of the grafted tentacle seemed to have been absorbed into the anterior part of its body (Fig. 50). From these experiments the conclusion must be drawn that it is principally the material of the body wall of the stock and not the hydranth material of the graft that forms the new hydranth. The ectoderm and endoderm cells that have been body wall cells are therefore changed over into cells composing tentacles and hypo- stome. Group B Absorption In order to find out the fate of the material grafted into the stock hydra when no regeneration took place; whether the material was incorporated into the tissue of the stock hydra, or whether it was so absorbed that it no longer existed as such, the following experi- ments were performed. Series [ Green Tentacle Without Base Grafted into White Hydra As the green tentacle was absorbed, the green material spread along the body wall of the white hydra for a small area at the union of stock and graft (Figs. 44, 51). The two hydras on which the experiment was performed were unfortunately lost before com- plete absorption. But it was evident that the green tentacle material was being transformed into body wall material and incor- porated in the stock. Series If Green Circlet in White Hydra After the graft of a green circlet of tissue from a normal hydra into the middle region of the body of a white hydra, this tissue remained in the white body as a patch of green (Fig. 45). 16 Ethel Nicholson Browne We conclude then, that when a piece of tissue is grafted into a normal hydra and does not cause regeneration, the grafted tissue is incorporated into the body wall of the stock. The cells that have formed tentacle or body wall tissue are made over in the stock into body wall cells. The tissue is absorbed, not in the sense that it disappears, but in the sense that it becomes one with the tissue of the stock. Group C Grafts of White and Green H ydranths A few experiments were performed with green and white hydras to discover if possible whether a hydranth that was grafted in the side of another hydra kept its individuality or whether the tissues of the two hydras fused. In one of the two successful grafts in which a short green head was grafted into a white hydra at about the middle region, the graft retained its individuality and was of approximately the same size and in about the same position at the end of two weeks as at the time of graft (Fig. 52). As the hydra did not live until the graft pinched off, the final result was un- decided, but it seemed probable that it would pinch off at the line of union of graft and stock. The second graft was more interesting for the reason that the green grafted head increased in size until of equal length with the head of the stock and then mi- grated down toward the foot end of the white hydra (Fig. 46). The point to be noted is that in increase in size, the new material came not from the larger white hydra but was formed by the green hydranth. The graft not only kept its individuality but also completed itself by regenerating new material. A third experi- ment of somewhat different kind shows the same principle. A white and green head were grafted together by their aboral ends, the green one being somewhat shorter than the white one (Fig. 53). Both hydranths kept their individuality and the green one regenerated new tissue so as to become of equal length with the white one (Fig. 54). In this condition, the graft died, evidently just before separation would have taken place. From these experiments we conclude that the grafted hydranth, although intimately associated with the stock, keeps its individual- Production of New Hydranths in Hydra a7 ity. This conclusion agrees with that of Miss King who, in similar experiments used hydras of different shades of green as stock and graft. Group D Graft of Green Foot in White Hydra In four out of five cases in which the lower half of a hydra was grafted by its cut oral surface into the middle region of a normal hydra, the grafted foot was absorbed. The history of the other case was as follows: On February 22, a green hydra was cut in two, and the posterior half was grafted by its oral surface into the middle of the body of a white hydra. On February 25, the graft had moved down somewhat toward the foot of the stock (Fig. 55). On February 27 a new head was evidently being formed at the union of graft and stock, of graft material (Fig. 56). On February 28 the graft had quite completed itself, having formed several tentacles, but was still attached near the foot region to the stock (Fig. 47). On February 29, the green hydranth including part of the grafted foot and the regenerated head separated off from the white hydra, leaving, however, a small amount of green foot tissue attached to the stock (Fig. 57). From these experiments I should conclude that if the graft asserts itself sufficiently not to be absorbed by the stock, it maintains its individuality, forms its own tentacles and separates off as a com- plete hydra. Group E Reversed Polarity in Green and White Grafts Cases of heteromorphosis in hydra produced by grafting have been reported by Wetzel, Peebles,and King. A heteromorphic foot has been produced by Wetzel by removing the foot ends of the two hydras, grafting the two aboral ends together and cutting off one head close to the tentacles. A normal hydra was produced with a foot at the former oral end. Peebles performed the reverse experiment, cutting off the heads of two hydras, grafting them together by their oral surfaces and subsequently cutting off one 18 Ethel Nicholson Browne foot close to the line of union. Heteromorphic heads were pro- duced in five cases on the exposed aboral surface. King succeeded in obtaining heteromorphicstructures by cutting off both ends of a head-to-head or a foot-to-foot graft close to the line of union, leaving a ring of tissue with two aboral or two oral ends exposed. In most cases, she found that normal hydras resulted, one end having reversed its polarity so that a head was produced from the exposed aboral surface or a foot from the exposed oral surface. It has never been definitely shown whether these heteromorphic structures are really formed from material whose polarity has been reversed through the influence of the complementary structure or whether the material has rearranged itself so that polarity is not really reversed. That such a rearrangement is possible is shown by an experi- ment in which I| cut a hydra.longitudinally and reversed the two halves so that each free end consisted of half foot and half head with tentacles (Fig. 58). Two days later it was evident that the foot material of each end was migrating from its position near the tentacles to the middle of the body (Fig. 59). ‘That this structure was produced by a migration of the foot material and not merely by a split along the line of graft, separating the original half head and half foot, is shown by the fact that the free foot was not of equal length with the hydranth, but was only a small projection from the surface. ‘The final result of this graft was the separation of the two original half-hydras into two complete hydras (Figs. 60, 61). In this case, then, there has been a migration and rearrange- ment of material. Another instance of migration is the experi- ment described in Group C, where the grafted hydranth moved down along the stock (Fig. 46) from the middle to the foot region. Many similar cases of migration have been described by King and Rand and Hefferan. Is it not possible that in the case of the supposed heteromorphic heads, the aboral material of the graft has wandered in, leaving exposed the oral material of either stock or graft, so that the head really develops not from an aboral layer of tissue but from the oral layer? The answer to this question has been definitely determined, I Production of New Hydranths in Hydra 19 think, by the following experiments: A white and a green hydra were both cut a little beneath the tentacles and the two cut oral surfaces grafted together. A few hours later, after the graft had become secure, the green portion was cut off leaving a circle of green tissue with aboral end exposed and attached by its oral end to the oral surface of the headless white hydra (Fig. 62). In three cases, tentacles appeared at the free aboral end and these tentacles were formed of green material, the oral material of the white stock having no part in their composition (Fig. 63). Ina few other cases, one or two of the tentacles were formed of white material, the rest of green (Fig. 64). Whether this result was due to a somewhat oblique cut, so that the circle of green was minimal in one place, or whether the material rearranged itself, some of the green migrating posteriorly and the white anteriorly was not determined. In the former cases, where all the tentacles were green, there was no migration of the white oral material. These experiments did not show, however, that there might not have been a migration of material in the green tissue itself, the oral material going anteriorly and the aboral posteriorly. Inoneexperi- ment, it was conclusively shown that such migration did not take place. In this case tentacles formed not only on the exposed green surface, but also at the union of graft and stock (Fig. 48). More- over, both sets of tentacles were composed entirely of green mate- rial. One set must therefore have come from the oral material which still remained at the junction of graft and stock, and the other set from the aboral material which lay at the exposed surface. This was an undoubted case of heteromorphosis in the strictest sense of the word, and throws some light on the meaning of reversal of polarity in connection with heteromorphic structures. We find in this experiment, that although polarity has been reversed in so far as normal foot-producing material becomes head-forming mate- rial under the influence of the larger piece, the original polarity which determined that the head-forming material should be head- forming has not been lost. It would seem that the original polar- ity of the reversed piece is not altered, but that on account of the relation of the graft to the stock, a secondary polarity has been assumed. This secondary polarity is that of the stock which 20 Ethel Nicholson Browne asserts itself in the new material which now becomes a part of its own body. If this secondary polarity be conceived as prepon- derating over the primary polarity of the graft, this original polar- ity would be entirely submerged, and this is probably the case in the former experiments where only heteromorphic tentacles were formed. On the other hand if the primary polarity preponderates over the secondary we should have tentacles at the junction of graft and stock and no heteromorphic structure. Many cases of this result have been reported by Peebles and King. In three cases I succeeded in obtaining heteromorphic feet by grafting the two cut aboral ends of a white and a green hydra and then cutting off the head of the green end leaving an exposed oral surface of green tissue (Fig. 65). ‘This graft resulted in the formation of a normal hydra consisting of a head end of white material and a foot end of green material (Fig. 66). The foot end consisted of a true foot as evidenced by the presence of the characteristic sticky secretion which made the foot adhere to the substratum. These were evidently cases of heteromorphosis, the polarity of the stock so influencing the graft as to preponderate over its original polarity, thus calling forth a foot from its exposed oral surface. ‘The production of heteromorphic feet in these experi- ments confirms the like experiment of Wetzel and is opposed to the results of Peebles who obtained no heteromorphic foot in the fifteen cases tried. No hook-like processes such as described by Wetzel as preceding the formation of a heteromorphic foot appeared in my experiments. CONCLUSIONS As a result of the experiments recorded in this paper, the follow- ing conclusions may be drawn: 1. The fate of a graft in Hydra viridis depends on several factors. Rand states that “the fate of the graft depends upon its degree of specialization.” King states that “the fate of a graft depends, not upon its degree of specialization, but primarily on Its size and to some extent on its position in the stock.” From my experiments, I should conclude that the fate of a graft depends (1) primarily on its specialization. If the tissue grafted is a smal Production of New Hydranths in Hydra 21 amount of body wall tissue, or pure tentacle tissue, or pure hypo- stome tissue (tissue anterior to the tentacles) or a small amount of foot tissue, it is absorbed. If it is tissue lying at the base of the tentacle, whether it includes a tentacle or not, it becomes part of a new hydranth which under its stimulation grows out from thestock. The fate of a graft depends (2) on the size of the piece grafted. A large piece of body wall tissue with the oral end exposed in a lateral graft gives rise to a new hydranth, a small piece is absorbed. A large piece from the foot may produce a new hydranth when grafted laterally, a small piece is absorbed. ‘The fate of a graft depends (3) on the position it occupies in thestock. Ifinthecirclet of tentacles, a new hydranth is not produced by the graft of a tentacle with peristome tissue at its base, but in any other region of the body a new hydranth may be produced. In the foot region this hydranth is very minute; in the middle region it is of normal size. The fate of a graft depends (4) on its polarity. If a band of tissue is grafted by its aboral end to the oral surface of a half- hydra, leaving the oral surface of the graft exposed, anormal hydra is produced, the tentacles growing on the exposed oral surface. If a band of tissue is reversed and grafted by its oral end to the oral surface of a half-hydra, leaving exposed the aboral surface of the graft, unless sufficiently small, tentacles grow out at the line of union of the two components, and not at the free end. 2. No matter how specialized a tissue has become, it can, when grafted, be made over into a different kind of tissue and be incorporated into the body of the stock. For example, the mate- rial that had become differentiated into tentacle tissue can lose its differentiation and become body wall tissue and function as such. Likewise, differentiated tissue of the stock can, under the influence of special grafts, be made over into other kinds of differ- entiated tissue. For example, the foot tissue can, under the influ- ence of a grafted tentacle with basal tissue attached, be trans- formed into the body wall and tentacle tissue of a new hydranth. 3. A grafted hydranth and a grafted foot, when not absorbed, keep their individuality and do not become one with the stock. 4. Anew hydranth can be stimulated to grow out from a hydra by (1) thegraft of the peristome tissue at the base of the tentacle, with 22 | Ethel Nicholson Browne or without the tentacle itself, and by (2) the graft of the material of a regenerating hydranth and by (3) the graft of the material of a bud. Neither a wound nor the graft of any other kind of tissue will stimulate the stock to send out a new hydranth. 5. Areversal of polarity may take place in a graft, resulting in the production of a heteromorphic structure, if the polarity of the stock preponderates over that of the graft. Zodlogical Laboratory Columbia University Production of New Hydranths in Hydra 23 LITERATURE HEFFERAN, Mary ’o1—Experiments in Grafting Hydra. Archiv f. Entwicke- lungsmech. Bd. xiii. Kinc, HELEN Dean ’o1—Observations and Experiments on Regeneration in Hydra ; viridis. Archiv f. Entwickelungsmech. Bd. xiii. °o3—Further Studies on Regeneration in Hydra viridis. Archiv f. Entwickelungsmech. Bd. xvi. Morean T. H. ’o1r—Regeneration. *o8—Experimental Zodlogy. PEEBLES, FLORENCE ’00—Experiments in Regeneration and in Grafting of Hydro- zoa. Archiv f. Entwickelungsmech. Bad. x. Ranp, HersBert ’99—The Regulation of Graft Abnormalities in Hydra. Archiv f. Entwickelungsmech. Bd. ix. Wetzex G. ’98—Transplantationsversuche mit Hydra. Archiv f. Mikr. Anat. Bd. lii. Wuirney, D. D. ’07—Artificial Removal of the Green Bodies of Hydra viridis. Biol. Bul. Vol. xiv. No. 1. 1 - é icine b 4 2S) Dee shy CS . icles Sa Bea Ser has We $ ih? } ae niet shat ae. Fy 4 “0 4/ é ; it Se ‘ - ; ’ ~ Prate I Figs. 1-5 Graft of tentacle with peristome tissue in middle of stock. Figs. 6-8 Absorption of new hydranth. Figs. 9-13 Graft of tentacle with peristome tissue in foot of stock. PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE I Etuet NicHotson Browne Tue JourNAL or ExpeRIMENTAL ZoOLoGY, VOL. VII. Pirate II Figs. 14-1 5 Abnormal hydra produced by foot-region graft. Figs. 16-17 Graft of tentacle with peristome tissue in circlet of tentacles. Fig. 18 Graft of tentacle with peristome tissue below circlet and above middle. Figs. 19-21 Fusion of old and new hydranths. Fig. 22 Grafted tentacle included in old circlet. Fig. 23 Graft of tentacle with peristome tissue between foot and middle. Figs. 24—26 Transformation of grafted tentacle into new hydranth. PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE II Eruet Nicuotson BRowne 27. Tue JouRNAL cr EXPERIMENTAL ZOOLOGY, VOL. VII, No. I Pirate III Figs. 27-29 Graft of two tentacles with peristome tissue in opposite sides of stock. Figs. 30-31 Absorption of one new hydranth, and growth of other. Figs. 32-33 Migration and fusion of two new hydranths. Fig. 34. Absorption of one grafted tentacle. Figs. 35-37 Graft of peristome tissue without tentacle in middle region. Figs. 38-39 Graft of peristome tissue in foot. PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE III Etuet Nicuotson Browne Id, THE JouRNAL or ExpERIMENTAL ZOOLOGY, VOL. VII, NO. 1 Prate IV Figs. 40-41. Graft of regenerating tissue. Figs. 49-50 Small hydra produced by graft of white tentacle in foot of green hydra. Fig. 51 Graft of green tentacle without peristome in white hydra. Fig. 52 Graft of green hydranth in white hydra. Figs. 53-54 Graft of green and white heads. Figs. 55-57 Graft of green foot in white hydra. PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE 1V Eruet Nicnotson BRowNE Tue JourNAL or ExpERIMENTAL ZOOLOGY, VOL. VII, No. 1 a ~ - Fig. 42 Fig. 43 Fig. 44 Fig. 45 Fig. 46 Fig. 47 Fig. 48 Pirate V Graft of white tentacle with peristome in middle of green hydra. Graft of white tentacle with peristome in foot of green hydra. Graft of green tentacle without peristome in white hydra. Graft of green body tissue in white hydra. Graft of green hydranth in white hydra. Graft of green foot in white hydra. _ Heteromorphosis in reversed ring of green tissue grafted on white stock. _ PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE V EruHet Nicuotson Browne a * 1. del, iccio R Tue Journat or ExpertMeNnTAL ZoOLocy, VoL. vil. : . iuctas, vel a Figs. 58-61 Graft of reversed halves of a longitudinally split hydra. t Figs. 62-64 Heteromorphic heads in green and white graft. - ; Figs. 65-66 Heteromorphic foot in green and white graft. — eee PRODUCTION OF NEW HYDRANTHS IN HYDRA PLATE VI Etuet NicHoLtson BRowNeE Tuer Journar or ExperIMENTAL ZOOLOGY, VOL. VII. Pobre phere OF THE DESTRUCTION OF PERIPH- ERAL AREAS ON THE DIFFERENTIATION OF THE NEUROBLASTS. M.L. SHOREY ConTENTS i finit@ahiee@M. 26 éc0g 030000 4dao0 Caen go DBD OGaOnOC OMOEA nO MaaonoddG opndcriohcadosuaec 25 lee: KEIM TOM PECL CLGatctaratyiel geste) o's /a15/0)-\ejolaia «a 'e elnle: «aici oy osrersrela Biofelstere toler manatees mares feels 28 Al OR GEOMEE ATA Gad 3.6 Cac peOetettos DIRE EEE CEE Aee nea on comccdc pe Ans ese urade: 28 aC haracte pOMtAEG PCLA LOI c.1<'s\s/cis(s.0is 2, 2-2)= <0 cit/eis:s ciao vie leie a ele sisia)ala eye.e.s Claclels d.cislsialaeieeists 29 G iieineils..5- doe cbpaceadas be SAS eB ean Go DES DORA SAR OMe nae AO ucdae Se ocouosEOneSc 30 D Experiments on embryos after three days of incubation................0-eeeeceeeeeeees 32 1 Specimens preserved five or six days after the operation.............000eceeeeeeeee 32 2 Specimens preserved from four days to one day after the operation.............. o4a, 20) Petxpenmentsatlaterstapes Of developments ...% aes talaiat= » /m areloic/alole «/-1s0/4)feyei~)orstelsieipynnieTa tere 43 Ter AMET SINC Acts ON LNCUD ALONE. vcisie sp 58 D’ Experimentson Bufo americanus .: 2. 6.0 .. 6 secs ce ccs cceecies cc ccceersecorensrcasn= 60 Ee Remoyalofsomitesin Rana pipiens... 22-566. cee ore ecces sels ticleiesoisiselisie sis sis e's 61 F Summary of conditions found in amphibians..............0cesecceeecereecercceccees 61 @ Comalistiee.. Aan qntoboeeede ae OM GeO OO OneE as HRA Gadon ObOaSEd ODO ODO A Sen oGon cbc 62 I INTRODUCTION Observation of surgical cases in which the amputation of a limb has been followed by the degeneration of the nerve fibers innervating it, and finally of the nerve cells from which they arise, and cases of paralysis accompanied by the atrophy of the muscles Tue JouRNAL oF EXPERIMENTAL ZOOLOGY, VOL. VII, No. I. 26 M. L. Shorey concerned, have demonstrated that an intimate relation exists between the life of a muscle and that of its motor nerves. And at least since 1851, when Weber described an abnormal calf embryo with the nervous system entirely wanting below the second dorsal vertebra and an accompanying loss of the corresponding muscles, there has been a tendency to believe that there is likewise a depend- ence in development. Other teratological cases in which a defect in the nervous system has been accompanied by a defect in the organs normally innervated by tbe missing nerves have been recorded, and the effect of the absence of the nerves in experi- ments in regeneration has been adduced as evidence in favor of such a view. It is obvious, however, that neither of these condi- tions can furnish conclusive evidence regarding developmental processes, for, in the case of monstrosities it is impossible to know the nature of the original lesion, and in regeneration, while there are doubtless similarities between it and normal development, it is evident that neither the conditions nor the processes can be iden- tical. Bardeen, in 1900, made careful observations on the relations between the myotomes and nerves at early stages of embryonic development, and from the fact that there is no apparent contact between the two until a considerable degree of differentiation has been reached concludes that early development is independent. To quote, “The early development of the nerves is one of passive independence, without any immediate relations to the myotomes.” But it seems to me that such evidence must be regarded as incon- clusive, for the fact that two organs are not in direct contact does not exclude the possibility that the one may influence the other. To reach decisive conclusions regarding the interdependence of two organs or tissues in development it is necessary to study the behavior of one in the absence of all possible effects from the other. This can be done only by destroying some portions of the develop- ing organism, or by removing the part to be studied to a controlled environment. Both these methods have been used in studying the relation of the nervous system to peripheral organs, but in most previous studies the object of the investigation was to discover the effect of the destruction of nervous tissue on the organ inner- Differentiation of Neuroblasts 27 vated. A few experiments dealing with the effect of an abnormal environment on the developing nerve have, however, been made. Harrison (’06), experimenting on frog larve for the purpose of ascertaining whether other cells than those of the nerve center enter into the formation of the nerve fiber, obtained results which bear on the question under consideration. ‘The spinal cord of these embryos was extirpated, and small pieces of it transplanted under the skin of the abdominal walls. Small nerve trunks arise from these and run for some distance in various directions. From portions of the ganglionic crest thus transplanted small ganglia and nerves may arise. His own conclusions are as follows: “The nerve center (ganglion cells) is shown to be the one necessary factor in the formation of the peripheral-nerve. When it is trans- planted to abnormal positions in the body of the embryo it then gives rise to nerves which may follow paths where normally no nerves run, and likewise when the tissues surrounding the center are changed entirely nerves proceeding from that center may develop as normally.” In another experiment fragments of the medullary tube, isolated before visible differentiation, were placed in a drop of lymph, and it was found that these develop short fibers. From this experiment he concludes that the outgrowth of the fiber is largely independent of external stimuli, although its direction is doubtless influenced by a number of factors. ‘That the direction is influenced is shown by a number of experiments in which a limb bud transplanted to some other region of the body receives normal innervation from the nerves of the region to which it is transplanted. Braus (’06) removed one of the anterior limb buds of several toad larvez, and ten days later found that the brachial plexus was as well developed on the injured as on the uninjured side, although by far the greater part of the musculature had been destroyed, and at the time of the operation no nerve fibers running to the limbs were visible. Older operated embryos showed the brachial plexus present but diminished in size. From these experiments Braus concludes that “Der Befund am Plexus brachialis ist eine sehr deutliche Illustration dafiir, dass die Entwicklung der peripheren Nervenfasern unabhangig ist von den Endorganen.” 28 M. L. Shorey The work of the two last investigators has been published since, or about the time that I began to investigate the relation of nerve and muscle tissue in the chick embryo, and did not reach my hands until a later date. ‘The experiments to be described were undertaken for the purpose of studying the behavior of the develop- ing nervous system,when it is itself left quite intact and with all its relations normal except that a peripheral area to which nerves should be distributed is destroyed. The work was begun at the suggestion of Professor Frank R. Lillie, and I wish to express my indebtedness to him for proposing a problem which has proved to be of intense interest, and for suggesting the first methods by which it was to be attacked. I wish also to express my appreciation of his continued interest and helpful criticism. Acknowledgments are also due to Professors C. O. Whitman, C. M. Child, and W. L. Tower for suggestions in regard to illustrations, photography and material, and to Miss Sabella Randolph, Mr. Kenkichi Hayashi and Mr. C. E. Brues for aid in the preparation of drawings and photographs. Figures 36 and 37 are copied from Lillie’s Embryology of the Chick. il EXPERIMENTS (ON THE CHICK A Chotce of Material The chick embryo was chosen as furnishing a form favorable for these experiments; the work of Lillie (’04) in destroying various portions of the embryo had shown that no regulation or regenera- tion was to be expected and my own work has entirely confirmed these results. After two days of incubation, the earliest period at which I have operated, and in regions where no differentiation is apparent, the primordium for each organ 1s so definitely laid down that the destruction of it, or any considerable portion of it, results in a corresponding loss in the fully formed animal. Lillie’s work had also shown that comparatively large areas may be destroyed without interfering with the normal development of the rest. It would therefore be possible to destroy the primordia of definite muscles, or of sensory areas, before they had been penetrated by nerve fibers, and then determine the course of the nerves which Differentiation of Neuroblasts 29 should innervate them. ‘The rapidity of development, and the availability of material at practically all times of the year were also points in favor of the chick. B Character of the Operation It was decided that the first experiments should be the removal of the wing and after some preliminary tests the period at which the wing bud is just apparent to the naked eye as a small elevation on the body wall was chosen as the most advantageous for the opera- tion. The period of incubation before this stage is reached varies from 68 to 88 hours according to the time of year, the development of the embryo when the egg is laid, and fluctuations in the tempera- ture of the incubator, so that it is impossible to tell in exactly what condition the embryo will be found when the egg is opened, but Fig. 1, a photograph taken in this case after 74 hours of incubation, gives approximately the external appearance at the time when most of these operations were performed. ‘The wing, was never larger than this, and in the majority of cases it was smaller. Fig. 2, a photograph of a cross-section in the region of the wing, shows the relation and development of different parts at this time. Even if the wing bud is removed close to the body, and no part of the wing proper develops, some of the muscles, or portions of them, extending dorsally from it to the vertebrz, and ventrally to the sternum, always appear. It, therefore, seems evident that the pri- mordium for these has not yet grown into the wing, but is still con- tained in the muscle plate. Parts of the bones of the shoulder girdle will also form, so that only a portion of their primordium can be located in the wing bud at this period. Lillie has determined that the wing develops from the 17th, 18th, and 19th somites, and that it, with the other structures arising from these somites, is innervated by the 14th, 15th, and 16th nerves, the sensory root arising from the 12th, 13th and 14th ganglia. Later in development these three nerve trunks unite to form a plexus before their distribution to the muscles and sensory areas of the arm, but at this stage they have not yet reached the place of union and each appears as a distinct trunk extending toward the base of the wing but not yet reaching it. Neither is it in contact with the myotome. 30 M. L. Shorey ‘The neural tube exhibits but little differentiation; no definite cell areas are marked off within it, and only a thin layer of nerve fibers or axones has formed at the periphery. There is a small ventral root, but the cells from which these axones arise have not yet become localized as the ventral horn. By operating at this stage of development it was possible to remove the greater part of the wing without any direct injury to the nervous system. As stated above, some structures, or parts of them, belonging to the wing may still develop, and the intrinsic muscles of the vertebral column are entirely uninjured; still the peripheral area, both sensory and motor, is much decreased, and as the greater part of the peripheral nervous system is still to be developed, and most of the differentiation of the spinal cord still to occur, it is possible to study the effect of the loss of the end organ on the size and distribution of the peripheral nerves, as well as its effect on the cells of the central system with which they are con- nected, either directly or indirectly. C Methods During the period of incubation preceding the operation the eggs were turned every 12 or 24 hours to prevent the blastoderm from adhering to the shell membrane, but to insure the accurate location of the embryo a number of hours must elapse between the last turning and the operation. The method of opening and closing the egg was essentially the same as that described by Lillie (703) and earlier by Miss Peebles (’98), but instead of placing strips of egg membrane about the edge of the shell used as a cover for the opening in the operated egg I found it more effective to leave a margin of membrane attached to this cover. If then the curvature of the cover is the same as that of the egg to which it is applied, and there 1s no break in the membrane, a perfect closure is effected. Since the embryo lies on the left side (I found only two or three exceptions to this in the fifteen hundred or two thousand eggs used in these experiments) the right wing was always the one removed. For this purpose small spear-headed scalpels were first used and this method may be used with success if the wing bud Differentiation of Neuroblasts 31 has attained considerable size, but for the earlier stages, at which it was finally decided to operate, destruction by electrolysis is much more satisfactory. Number 12 sewing needles, one of them bent so as to penetrate the amnion and lie along the whole extent of the wing bud with as little tearing of the membrane as possible, were used for the operation. ‘These were held in electric handles con- nected with dry battery cells furnishing a current of about two volts. [he destruction is almost instantaneous, so that the current need be applied for but a very brief period. There existed the possibility that by this method of operation the electric current might penetrate beyond the region with which the needle was in contact, and thus lead to the direct destruction of cells within the spinal cord. Such a possibility is eliminated by the following facts:—in a number of embryos in which the wing was removed by scalpels the same defects appear as in those of the same age operated on by electricity. Moreover the defects always appear in the same region, and follow the same order, and it would scarcely be possible to suppose that an electric current, free to extend in any direction, should always follow exactly the same path and destroy the same cells, leaving those immediately sur- rounding them quite normal. And a third reason is found in the fact that until twenty-four hours after the operation the path of the current is well defined by means of the presence of the injured tissue, and in specimens killed at this period, and in which it is possible to determine by this means that the injury did not extend into or even close to the spinal cord, the nervous system is still defective and the defects are always of the same nature. I had at first many infected eggs, and occasional ones later, but as I have at other times operated every day for months with only one or two cases of infection, I believe that only sufficient care is necessary to avoid it entirely. ‘The surface of the egg was disinfected with mercuric chloride, and all instruments sterilized. There continued, however, to be a large percentage of loss from tears in the blastoderm, puncturing of bloodvessels, extensive injury to the amnion, and other causes whose nature was not determined. Probably most cases of death within a short time after the operation are traceable to injury to the circulatory sys- 32 M. L. Shorey tem, but when it occurs 5 or 6 days later, as it did in some experi- ments which I attempted to carry beyond that period, it seems to be in some way connected with defects in the amnion, for in all these cases which were examined carefully an aperture was found in the region of the operation. This would agree with the work of other observers who have shown that anamniote embryos may develop normally for a time, but that later this defect is fatal. That the amnion can, however, regenerate if not too extensively injured and form a completely closed sac was demonstrated by other experiments in which I knew it to have been injured by the operation, so that an improved technique would doubtless obviate this difficulty. Killing and fixing in sublimate zctic, and staining with Dela- field’s haematoxylin and Orange G was found to be entirely satis- factory for the end desired, and was the method used inthe majority of cases. As the effects would naturally be most pronounced after a con- siderable interval of time I began with a study of the oldest embryos, those killed between 5 and 6 days after the operation. When the conditions in these had been ascertained, specimens killed respectively 4 days, 3 days, 2 days, and 1 day after the injury was made were examined, and the results will be given in this order. Experiments on Embryos After Three Days of Incubation Specimens Preserved Five or Six Days After the Operation Experiment 538. This egg was incubated 84 hours and was approximately i in the stage shown in Fig. 1 at the time of the oper- ation. ‘The right wing was removed and the egg returned to the incubator for 5 days 5 ae so that when preserved the age of the embryo was 8 days 19 hours. ‘The only evidence of the right wing was a small rounded elevation scarcely noticeable on a casual examination. The wound was completely healed and the embryo vigorous and entirely normal externally except for the missing limb. Differentiation of Neuroblasts 33 It was killed, sectioned, and stained as given above, and the oper- ated and unoperated sides carefully compared. The bones and muscles were well defined and easily identified and the course of the nerves traced without difficulty. Fig. 3 shows the location of the chief wing muscles and Fig. 33, 4, the periph- eral distribution of the brachial plexus on the normal side. Since the branches of the 14th, 15th and 16th nerves which innervate the vertebral musculature are not affected, these muscles being quite intact, and the branches arising from the separate trunks before the formation of the brachial plexus, their distribution has been neglected. On the operated side the bones of the shoulder girdle are present but defective; the clavicle and scapula are, however, only slightly abnormal. Of the muscles the trapezius, rhomboideus, and sub- clavius are not noticeably affected; the pectoralis major, pectoralis secundus, coraco-brachialis, latissimus dorsi, teres et infraspinatus and subscapularis are present but defective, while the deltoides sup- raspinatus, biceps, pectoralis tertius, coraco-humeralis, scapulo- humeralis, and triceps are entirely wanting. Fig. 33, B, shows the peripheral distribution of the brachial plexus on this side. It will be seen by comparing this with 4 that wherever a muscle 1s missing the corresponding nerve 1s also missing; otherwise the distribution is the same as that of the opposite side. As A and B are drawn on the same scale they show some- thing of the quantitative loss, but measurements give this more exactly. It is difficult to arrive at the exact dimensions of the nerve trunks from sections without making reconstructions, but as a means of comparison each of the three nerves involved was measured directly below the union of the ventral root with the ganglionic root, and the number of sections in which each appears at this place counted. From the data thus obtained the following percentages of loss for the nerve trunks were calculated : WIDTH THICKNESS PERCENT PERCENT MALUITIEGVEH OP AA caste terstcitieit sve oie.tiple aloo siete els vletepucieie’s fiels rie? “erin aer 34.1 25 RIG ELLE GV ets tet taaiaN-Nal e/a olafo\a/ola)e.viz)rie/siai=)e\psle(ei= iniciafeta/elsiefelole = ce afela-tevo(eltnfet= 17.6 36.3 MALEMTACEN Crepe eerste TaicT oie) Se, os Sissel «cen ale ola ote avclinis ekeinnyelere a lannkelsveuctenimeel sya 35-8 30 34. M. L. Shorey Measurements of the ventral roots show this loss: WIDTH THICKNESS PER CENT PER CENT VAIN DADs 55 HhAo CabOno no oGUIboD DUdQHDasbeda Saud oud Odo acagaGuuenTE 31.9 45-4 TiN WEIN(ES oA be odo abnoonbe ano undoaodeeooocnoaoaumoceewan goae sour 8.1 33-3 GME Ouindsian sed ompS odo DadQdds Cpe dobacueubus onoEbAas50 Hedaboas™ 36.2 a9e4 Measurements of the dorsal roots, which are more difficult to take accurately, because in some sections the boundary between them and the vertebrz is obscure, show a less, but still a constant loss: WIDTH THICKNESS PERCENT PERCENT TA TESTER VE rs aol i sle/a\csers/erarsiwcfoisiers eiehe Grsnslataie aie sue. are aia, o) fol ale, tr ay loko Sys O-aVay 36 ° iia KERSS oon bb obo sesso coohoo taco cob ph dacoonenasupasoDOsaUgoOs I 22 Gio WIG. sod shuoodobohon dose oSbobocebad sb ebeoardbusos Saancadsune 38 II The ganglia of the operated side were also smaller, and this loss is expressed quantitatively in the following table: (In this it must be remembered that the 12th, 13th and 14th ganglia correspond to the 14th, 15th and 16th nerves.) LATERAL DORSO-VENTRAL DIAMETER DIAMETER THICKNESS PER CENT PER CENT PER CENT TZthiAn oO Mesenteric erent bie emer ce eiarestere t= 2A 16.7 20 WVHA Osc pogo oubuoc ope. pose soneocgopoESeudbea doe DPF) Bg 33-3 Tg thipant eon’ aicrisisltorsiccre ararsisiola sels she ate eee ctatolel als ecete sist olsy > 20.3 30.2 20 The loss in the spinal cord will be best shown by a photograph of a typical section (Fig. 14). This was taken in the region of the 15th nerve. The most evident difference here is in the ventral horn, this being decidedly smaller throughout the operated region. It is also noticeable that the abnormality is practically confined to the region of the 14th, 15th and 16th nerves, the first observable difference appearing three sections anterior to the beginning of ganglion 12, and although very careful measurements show slight differences to the eleventh section beyond the 14th ganglion it is markedly less immediately beyond that ganglion. ‘This loss in the ventral horn might conceivably be due to either of two reasons: the number of cells might be less, or the individual cells or part of them, might be smaller, or the loss might be due in part to both of these causes. Two methods of deciding this Differentiation of Neuroblasts 35 question were adopted. First the cells of the antero-lateral part of the ventral horn, this being the region affected, beginning with the section on which the 12th ganglion appears, and going to the section on which the 15th appears, were counted on each side. There is evidently much chance for error in this, particularly as the sections were not cut with this in view, and a change of focus brought different cells into the visual field, but the cells of each section were always counted a number of times, and the probability of error is equally great for each side. While, therefore, the figures obtained probably do not represent the actual number of cells, I think they may be regarded as giving a fairly accurate compari- son of the two sides. ‘The total loss for the whole area was 50 - per cent, and this was distributed to the three nerves as follows: PER CENT ACER TICTUE cteferal ste sottier eat orttatstarey Voy sr efess |e; aieie, «ia\s/s\s 4 : PA : . 5 ee ‘ L ~~ ‘ ‘ Fig. 18 Cross-section of the spine cord ¢ at the level 0 the I sh nerve in Experiment my margin of motor nucleus marked by a cross. " ae : Fig. 19 Cross-section of the spinal cord i in Experiment 179, typical of the conditior * i ad 7 +s an 5 « : ’ “7 4 ‘s oh oe a? @ we 0 p Py ey “9 ~ an roe r Aa g . ‘ : J a he |S Fas nd A f ‘ standard deviation V number of terms in the series The reliability of an average thus varies inversely as the varia- bility of the series and directly as the square root of the number of terms. ‘The probable error of the mean, here employed as an index of reliability, is not to be confused with the probable error of @ series, sometimes employed as an index of its variability. This latter is a number of such magnitude that it is exceeded by exactly one-half of the deviations. It has the value: .6745 x standard deviation. The reliability of the standard deviation, or figure denoting the variability of each series, is indicated by the probable error of the standard deviation. ‘This is obtained by the formula: -6745 > standard deviation V2 > number of terms Effects of External Conditions 107 Since one of the primary objects of such investigations as the present is an inquiry into the effects of differences in the conditions of life upon supposedly homogeneous material, one of the principal points to be determined is the significance of any differences which may be discovered between the average values of a given character in two groups of individuals whose history has differed. For this purpose it is necessary to compare the probable error of the difference with the actually obtained difference between the two averages in question. The probable error of the difference 1s expressed by the formula: V/ (probable error of the first average)” + (probable error of the second average)” i. e, the square root of the sum of the squares of the probable errors of the two respective averages. Now the actually obtained difference is the most probably true difference and it is as likely to be too small as too large. Nevertheless the true difference may possibly equal 0, 1. e., be non-existent, in which case the ob- tained difference would be regarded as wholly “accidental.” From the table of the values of the “ probability integral” it may be calculated that the chances that a difference between two aver- ages is due to mere accident are: 250 out of 1000 when difference between averages = 1 % probable error of the difference. 156 out of 1000 when difference between averages = 1,5 X probable error of the difference. 89 out of 1000 when difference between averages = 2 % probable error of the difference. 46 out of 1000 when difference between averages = 2.5 X probable error of the difference. 21 out of 1000 when difference between averages = 3 XX probable error of the difference. g out of 1000 when difference between averages = 3.5 % probable error of the difference. 3 out of 1000 when difference between averages = 4 % probable error of the difference. than + out of 1000 when difference between averages = 4.5 X probable error of the difference, In proportion as the probability decreases that such a difference as been due to mere chance or accident (i. e., that it is the result of a multitude of independent causes having no relation to the conditions of the experiment), it is obvious that the probability increases that some constant modifying influence has been oper- ative in differentiating the two groups. It must be admitted however, that the probabilities here stated apply in full strictness 108 Franas B. Sumner only to cases where we are dealing with large numbers of indi- viduals. Davenport suggests 200 as the minimum number of individuals to be gathered for statistical treatment when the material is available; though he grants that much smaller num- bers may be employed to advantage, where we are restricted by circumstances. But it must be borne in mind, he says” “that the rules for determination of averages, probable errors, standard deviations and all the rest become less and less significant as the number of variants becomes smaller. Finally, in the region of twenty or so, the results can no longer be treated by mass statis- tics; twenty hardly makes a mass.” ‘To the experimentalist it must often happen, as in the present work, that the use of a large number of individuals, in any single series, is excluded by reason of the laboriousness of the methods employed. In such cases, we are told, no exact mathematical equivalent can be offered for the probability of a given result, even though the frequency distributions afford strong evidence on the subject. Of course the cumulative testimony of several independentseries of experiments is of highvalue. Ingeneral, it would seem that the experimentalist demands a somewhat different statistical tech- nique from the student of variation per se, and it is to be regarded as unfortunate that the methods at our disposal, have, thus far been developed mainly by the latter type of investigator, and with very little reference to the special needs of the former. To the experimentalist, who is studying the effects of artificial con- ditions, it is the significance of differences, and scarcely anything- else in the whole field of statistical theory, that is likely to be of interest. Regarding the accuracy of my computations, I can only say that every step has been gone over at least twice, and that, wherever possible, a different method of calculation has been employed in the repetition. RESULTS IN DETAIL Series of 1G00-1907 Owing to the small number of individuals used and the tentative character of my methods at the outset, this first series will not be 16 Tn a letter to the author Effects of External Conditions 10g discussed at all fully. The mean temperatures in the two rooms, during the period of the experiment, were 24.9° C. and g.1° C. (76.8° F. and 48.4° F.) respectively. No further analysis of the temperature conditions seems worth while in the present experi- ment. The humidity was not at any time determined. ‘[wenty- one mice (13 males and 8 females) were reared in the cold room; 20 mice (12 males and 8 females) in the warm room. ‘The ani- mals were not subjected to the differing temperatures until they were about three weeks old (21 + days). Previous to that time, the undivided lots had been reared under similar conditions. Each lot comprised individuals from 7 different broods, each of the latter having been divided into two portions destined for the warm and cold rooms respectively. About half the stock, consisting of the broods earliest obtained, were subjected to the experimental conditions for a period of 106 days, the remainder for a period of 83 days. At the expiration of these terms the mice were paired for breeding purposes, and the two contrasted lots were trans- ferred toa single room having a temperature somewhat interme- diate between those previously employed. None of the animals were killed immediately after this transfer, while the females were kept until they had reared their broods. ‘The interval between the discontinuance of the temperature differences and the killing of the animals varied from 15 to 55 days. ‘Thus the material was far from homogeneous. Nevertheless the figures obtained seem worth recording. (Table 1.)” The difference in tail length, in the males, at least, was often obvious without measurement, and it must be regarded as a signif- cant one, even without such overwhelming corroborative evidence as is offered later. In the case of the females, the difference is much smaller, though it is greatly in excess of the probable errors. No further analysis of this table seems called for. In the present series, the number of hairs per unit of skin area was computed for each mouse, according to the method described 7 Here and elsewhere the number of individuals measured for any given character is indicated by the figure in parentheses at the head of each column. In some cases individual tail measurements have been thrown out, where the tip of the organ had obviously been lost through accident. I trust, however, that it is not necessary to urge that no merely ‘‘exceptional” figures have been rejected B. Sumner 1S Franc TtO zo'h 6z°z | tof | righ 9g°z | gz°z ol’ 1 S7: | 1£S°z GUENe ‘Str oli" O70 oS +g Lg: LL o1'gg | Sz786 «| Sz-26 Lg: £6 | go'b6 | Lév-be | Leb-tz | z6L-be | zehl-bz Pe cman al ® | @® |e | @ ).@ | ® | @ | @/ @® | @ | & | @ PI°D UTE PIOO UIE POD ULE PIPD TEN PIPD UIE PI9PD UIE = sa]eula y sae sojewa J es sae so[e way sare 3 . eee HOINAT ‘TIwL HIONGT AGO" ay ea NI LHOIGM Lo61—gobr fo satsay I aTavs CIID Ot “UOTIRIAIP pIepueSg nel (e puatavaVeiarre a -eaeee Te TBOTAT Effects of External Conditions III TABLE 2 Series of 1906-1907: Number of hairs per unit of skin surface MALES | FEMALES Warm (12) | Cold (13) | es (8) | ahs (8) | a 2% | oe 310. | A; 304 a 462 ner | ssa fe ants | caspase yom | atta | Shae | Sho | et 25 ae 350 | tes, 326 | a 205 shi05 | hare | thas | bass | Be oe a a e) 306 ny: 305 oS age is 357 et 77 a 291 1) a32 | = 335 = 220 1s 310 eae. hae | ih aa es ase | | dhe PE ed oe anc.) 1-267.08 61a a a ay tacos mere Standard devition........| 31-37 4-32 | 39-434 5.22 | 50.12+ 8.45 | 57-404 9.67 Ti2 Francs B. Sumner above (p. 101). It seems worth while to present the individual figures obtained for these animals, since this is the only series with which the counting method was employed. In Table 2 each figure in the left-hand columns represents the number of hairs on a circular disk of skin 1.5 mm. in diameter. ‘Two figures are given for each individual, based upon two disks of skin taken at a dis- tance of I cm. to the right and the left respectively of the mid- dorsal line. The sum of these two is stated in heavy type. It will be noted that the mean number of hairs for the cold room males is 305.23 upon the two disks, that for the warm-room males being 267.08."* Here, then, is a difference of 14.3percent. Whe number of individuals is small, it is true, and the probable errors are large. Even granting the constancy, however, of such a difference, between two lots of mice thus treated, it is not neces- sary to conclude that there has been an actual increase in the number of (developed) hairs per unit of skin surface. If the warm-room individuals be supposed to have slightly thinner skins than those of the cold-room lot, the greater degree of stretching in the former (see p. 101) would result in a less dense distribution of hair upon its surface. But whether a difference so produced would be as high as 14 per cent may well be questioned. Again, it must not be supposed that I am urging this difference in the density of the coat of hair as an instance of permanent morpho- logical change. It may be due merely to a difference in the rate at which the hair is shed. ‘This point will be discussed later. The averages for the hairs of the female mice are not far from equal in the two contrasted lots. It has already been noted that the females exhibited a much smaller difference in tail length than did the males. Series of 1907-1908 During the winter of 1907-1908 the experiments were conducted on a much larger scale than previously, the conditions employed were such as were calculated to result in the production of greater 8 The number of hairs per square millimeter of skin may be readily computed, since the area of each disk was (approximately) 1.767 sq. mm. Thus the mean number for the cold room males is 86.4, that for all the mice comprised in the table is 85.0, etc. Effects of External Conditions 113 modifications, and the number of measurements applied to the animals was considerably increased. Temperature. ‘The mean temperature” of the warm room for the entire season was 26.30° C. (79.34° F.), that for the cold room 6.16° C. (43.08° F.). These figures correspond roughly to those for the mean annual temperatures of Key West or Porto Rico, on the one hand, and those for Eastport, Maine, or Minneapolis, Minn., on the other.2® But it would not, of course, be at all fair to compare the temperature conditions of the experimental rooms with those of the points named, still less to compare the climatic conditions as a whole. The mean daily range of temperature (1. e., mean difference between the maximum and minimum for each day) was 11.9? C. (21.4° F.)-in the warm room, 6.7° C. (12° F.) in the cold room. The maximum temperature reached at any time in the warm room was about 40° C. (for very brief periods), the minimum in the cold room being — 14.4°. But these figures represent exceptional occur- rences and have little significance. Curves have been con- structed (Fig. 1) showing the mean daily temperature in each room during the entire period of the experiment. It is plain, of course, that none of these figures can represent the actual temperatures to which the mice themselves were most of the time exposed. When huddled together in large numbers in a nest of cotton-waste, the temperature of the air immediately in contact with them, at least in the case of the cold-room animals, was doubtless very much higher than that in the room outside, i. e., that recorded by the thermograph. Nevertheless, we all know by experience the difference between sleeping in a cold room and sleeping in a warm one, even when the amount of bed- ding is varied to suit the circumstances. And it must be remem- bered that during part of the time the mice were feeding, exploring the cages, etc., and were then wholly exposed to the air. Humidity. The relative humidity in the warm room (see p. 100 above) ranged from about 22 per cent to about 66 per cent, the ” The mean temperatures here given are based upon four figures daily, these being taken from the thermograph sheets. 20 Report U. S. Weather Bureau for 1906-1907. 114 Francis B. Sumner mean for 12 determinations made during a period of four months being 39 per cent. This is somewhat less than the mean relative humidity at Phoenix, Arizona, for the year 1906, as stated by the U. S. Weather Bureau, that point having the dryest atmosphere, with a single exception, of any place in the United States for which records are given. In general, the humidity of the warm room varied inversely as the temperature, since no compensation was made by the evaporation of water; but the degree of saturation of the outside air must have been a factor. ‘The humidity of the cold room varied from about 49 per cent to about go per cent, the mean of nine determinations being 67 per cent. ‘This figure is a trifle less than that given for the mean humidity of Philadelphia during the year 1906. The physiological results of such differences in humidity must be far reaching. The quality of the air inspired must affect the processes of respiration, and the rate of evaporation from the body surface must differ widely in the two cases. ‘This last was shown by the eagerness for water displayed by the warm-room individu- als. To what degree the results which I offer have been brought about by differences in temperature and to what degree by differ- ences in humidity it is impossible to state. In these experiments it has not been practicable to control the humidity, independently of the temperature, and thus it has been impossible to decide this question definitely. The subject will be referred to later. Disposition of Stock. ‘Iwenty broods of young mice, aggregat- ing 135 individuals, were employed for this experiment. | In order to insure, so far as possible, a division into two lots of a similar hereditary endowment, one-half the individuals of each litter were subjected to the high temperature, the other half to the low. For this purpose, exchanges were made between the offspring of differ- ent mothers according to the following plan: One-half of brood A, plus one-half of brood B, were consigned to the care of the A mother in the cold room, the other half of each brood being con- signed to the care of the B mother in the warm room, and so on through the series. This disposition of the young was made while the latter were 2 to 4 days old (in a few exceptional cases as much as 11 days). “The members of each litter were marked at the “Vd 2 ra = 8 Ss) =) aes = ° ” Cal > oc SEE, 4 a @ 2 » af a > af a <= Sg < =) 2 | “ 2% 2 4 w a 52 .S) w eo 3 nw 3 i ined w a = w > 3 z YIGOLIO ii AS TTTTTES ZEEE Effects of External Conditions ies ea ioe wel (a LNA > Dy a HV bese aw tana ea ea ue rts TTT He REELS ASTI eAIE NEHGS: ical it BErsashect seeaill HD i Tt x cnitil ill of 1907-1908 and 1908-1909. These curves are based upon Fig. 1 Chart showing temperature conditions in warm and cold rooms during experiments The continuous lines represent the temperatures for the as computed from four daily readings at equal intervals. the mean temperatures for each day, (1) mean date of birth of h the 2}$-month and 3-month meas- The heavy vertical lines (continuous and broken) denote earlier winter’s experiments, the broken lines those for the later year. each year’s mice; (2) mean date at which the 42-day measurements were made in each year; mean date at whic y ; 4 y year; (3 urements were made. 115 116 Francis B. Sumner outset by a system of clipping the right ear and various fingers and toes of the right foot. Rather contrary to my expectations, the alien offspring were accepted by the mothers as readily as their own, so that little if any loss of life resulted from this procedure. A certain proportion of deaths occurred here as always during the rearing of these mice. The number of deaths in the warm room during the first six weeks was 6, giving a mor- tality of about 9 per cent. The number dying in the cold room was considerably greater, being 13, or about 20 per cent of the lot. During the next month of life no deaths occurred among the warm room lot, while 6 were recorded in the cold room; but scarcely any further deaths from natural causes occurred during the next four months, 1. e., until the end of the experiment. ‘The cold room individuals, throughout the earlier part of their life, at least, were much less active than those in the warm room. During the first few weeks they kept to their nests almost con- stantly. Nevertheless, when mature, they were of decidedly better appearance than the warm room lot, and, when paired, they reared a much higher percentage of offspring. It must be added, however, that the reproductive capacity of both lots was found to be distressingly slight—so slight, in fact, as to render futile any attempt to make a satisfactory test of the transmissi- bility of the modifications which had been produced. Among the 21 females in the cold room lot, 31 pregnancies are recorded for the 15 weeks during which they were kept with the males, while in the aggregate only 48 young were reared to the age of six weeks. Indeed, the majority of the litters either consisted entirely of still- born young, or of ones which died during the first few days after birth. In other cases, the young were apparently born healthy, but the mothers seemed unable to suckle them or perhaps lacked the instinct to do so. With the warm room lot the case was even worse. Of the young resulting from 50 pregnancies (doubtless over 200) only 35 individuals, or about 15 per cent, survived to the age of six weeks, while in the great majority of litters all the indi- viduals died either before or shortly after birth. I am still almost wholly at a loss to account for this failure of the powers of repro- duction. The mice were paired rather too young, it is true, being Effects of External Conditions 117 2% months old at the time. Many of them did not become preg- nant till they were much older than this, however, while it is well known that female mice may bear healthy young at an even earlier age. Again, when judged by most other standards, the animals appeared to be in perfect health. ‘They were active enough, and the fur was commonly in good condition, though the size of the females, at least, even when fully grown, was probably somewhat below normal. Moreover, after the earlier weeks of life, their mortality had been slight. That this damage to the generative powers must be set down as one of the results of the experimental conditions seems, nevertheless, probable. Measurements at 42 Days. ‘The weight and tail length of these mice was taken at the age of six weeks." No other measurements were at that time believed to be practicable with the living animals. The following table (no. 3) presents the mean and the index of variability for each of these measurements, the two sexes being treated separately. TABLE 3 Series of 1907-1908: Measurements at six weeks of age WEIGHT ‘ | TAIL LENGTH Males Females Males | Females | Warm Cold Warm | Cold | Warm | Cold | Warm | Cold (29) (31) (33) (23) Rag) (Gays sic! Ga) | Mean...............| 12:997 | 13.123 | 12.282 | 11.400] 68.83 | 54.16 | 69.06 51.91 | 40.270 | +0.374 | 40.179 | 40.301 | $0.35 |0.68 | +0.43:| +0.83 Standard deviation....| 2.156 | 3.088 M7 2.138 2.80 5 09" lgneg 5.89 | 0.191 | 0.265 | 0.126 | 0.212 | £0.25 | +0.48 | 0.31 | £0.59 Fig. 2 shows the distribution of weights for the cold and warm room groups (sexes combined); Fig. 3 shows the distribution of tail lengths for the two groups. From the table and polygons collectively the following facts may be gathered: *1 Owing to difference in the date of birth, this age was not attained simultaneously. The great majority, however, were born within the space of a week. 118 Francs B. Sumner 1 The tails of the warm room individuals are much longer than those of the cold room individuals. ‘This difference amounts to 27.1 per cent for the males, 33.0 per cent for the females, and 29.7 per cent for the sexes combined. Indeed, the two types are so distinct that, but for a single individual, there would be no over- lapping of the polygons; i. e. (barring this single exception), the longest tail in the cold room lot was shorter than the shortest in the warm room lot. ‘These differences were so patent to the eye that, had the two lots of mice been mixed together accidentally, I am (ee ae Gal ie TGisliC Sau) ell adel Fig.2 Series of 1907-1908: weight at six weeks of age (sexes combined). Abscissas denote weight in grams; ordinates denote number of individuals. Vertically shaded areas represent warm-room ani- mals; horizontally shaded. areas represent cold-room animals. i sure that I should have been able to separate them again with com- paratively few mistakes. Contrary to the condition in the first year’s series, a greater difference is here shown by the females. 2 The warm room males were on the average I per cent lighter than the cold room males; the warm room females, on the contrary, were 7.7 per cent heavier. Quite similar relations in respect to weight will be found in the series of the following year. The frequency polygons for weight show that two groups of individuals, having two different ‘‘modes,”’ are comprised in the cold room lot—a lighter and a heavier group. An analysis of the individual HN Mn | 7\ tail length at six weeks of age (sexes combined). Abscissas denote tail length in millimeters; ordinates denote number of individuals. ul ul — | ——— -— -————} _——| — —— _—— -— 60 6l 52) Sd C4 Oo OG morn oo) cso LA Mn MN I pitt went ul nm “ INI 45 46 47 48 49 50 SI ils Mi Hi 39 40 4) Shading as before. . Fig. 3 Series of 1907-1908 120 Frances B. Sumner figures shows that both the males and the females are divided quite sharply into these two groups. The number of animals is so small, however, that this phenomenon may be accidental. It will be referred to later. 3. In the case of both sexes, the variability both in weight and in tail length is considerably greater for the cold room lot dian for the warm room lot. 4. The mean weight of the males (irrespective of history) is considerably greater (9.6 per cent) than that of the females; the tail length seems to bear no constant relation to sea! does the variability of either character. Measurements at 24 Months. At the age of 24 months the mice were mated. At about the same time (when 75 to 78 days old) the measurements previously made (weight and tail length) were repeated. ‘The number of females surviving to this age was 54 (33 warm + 21 cold). The number of males reserved for breed- ing was 13 (7 warm + 6 cold). There remained 43 males (22 warm + 21 cold) which were not needed for breeding purposes. They were therefore killed at this time, and subjected to a con- siderably greater number of measurements than had hitherto been employed. At this period, therefore, there are to be considered three groups of individuals (each group consisting, of course, of a warm room and a cold room half) : (1) the females; (2) the mated males; (3) the unmated males. Referring to Table 4, it will be seen that among the female mice the warm room lot are 4.2 per cent heavier than the cold room lot, while the tail length of the former is 26.6 per cent greater. Comparing the mean figures here given with those of the table for 42, days, it will be found that in the cold room lot the mean weight has undergone an increase of 33.1 per cent during the interval between the two measurements, while the mean tail length has undergone an increase of 16 per cent.” In the warm room lot, * Of course this increase in the mean has not exactly the same value as the average individual increase. The latter figure cannot be given for the present series, since the mice had not been marked so as to be individually distinguishable, although the members of each litter were identified by a brood number. Inasmuch as there had been but two deaths in the interval between the measure- ments, we are dealing with practically the same group in each case. Effects of External Conditions I21 on the other hand, the increase in weight and in tail length have been 28.8 per cent and 10.3 per cent respectively. ‘There has thus been manifested a tendency toward equalization in respect to both of these characters, but especially in respect to tail length. * This appendage has undergone a percentage increase which has been more than half again as rapid in the case of the cold room (i. e., the shorter tailed) as in the warm room (i. e., longer tailed) ani- mals. Further evidence for such a general tendency toward the neutralization of early differences will be offered later. As regards variability, the standard deviation for tail length, both in the warm and cold room lots, has undergone a slight absolute decrease, TABLE 4 Series of 1907-1908: Females 24 months old | WEIGHT TAIL LENGTH | | | | Warm (33) =| Cold (21) | Warm (33) Cold (21) se | WMeancaee ec Mere as 15.821 + 0.261 15.176+ 0.330 76.18 + 0.38 | 60.19 + 0.76 Standard deviation....) 2.221 0.184 2.2364 0.233 Tee he) ae Very ||" = ais) Se (o)slye| | | notwithstanding a considerable increase in the average for each lot. The standard deviation for weight, on the contrary, has under- gone an increase in each lot. The relative variability (1. e., ratio of standard deviation to mean) has increased in one case (warm room), decreased in the other (cold room). TABLE 5 Series of 1907-1908: Mated males 24 months old WEIGHT TAIL LENGTH | Warm (7) | Cold (6) | Warm (7) | Cold (6) Wea ninea reir ite fe li6842' = Ost OSy | 720. nI7 0.824 | 78.00 + 1.17 | 66.00 + 0.62 No discussion of the “mated males” (Table 5)” is worth while, owing to the small number of individuals comprised. ‘These 3 A certain degree of selection was exercised in picking out these males, the larger individuals of a brood being chosen for breeding purposes. This is shown by the difference in mean weight between the ‘‘mated males” and the ‘‘unmated males.” Francis B. Sumner JAP! ‘adv jo syyuoUT $z qe ,,sajvur payewUN,, Jo yISuRy {IV} :g061-Lobr jo saliag «9 “307 CB SSO GA aE ek eR SD Gi Ch WL Or NGS) 2 O90 £9 A I ‘adv Jo syjuour -§z We , safeur payeuun,, yo syyduaz Apoq :go61-Lob1 jo saiag $ “31 ‘adv JO syuoU-Fz We ,. saTeUT payeuTUN,, Jo 7YSIAM :go61-Lob6r jo sauag + “31 no» | | it ; PIS SILANE IS CIAO 09 16S SON LS 9G GON Ho) BG ueS iw Effects of External Conditions 123 TABLE 6A Series of 1907-1908: Males (unmated) killed at 2 months. WARM ROOM (22) | Body | eee | Hair in Weight | | Tail | verte- Ear Foot | length mg. | | brae | spe Ce 714, 32 13.2 16.9 328 sain a 76 qa 33 12.6 E7ato |. 21460 | 15.1 79 82 32 13.6 17.3 350 | 13.8 72 72 2H I My] 18.2 198 18.8 83 74 32 Arg 17.0 343 hey, 81 78 33 | 12.4 yfees 341 ele; 77 73 34 12.5 16.4 228 18.3 86 77 32 13.0 17.0 271 12.3 71 714 eed fo ea, 17.3 198 16.6 83 17 320 lg Tae 1735 289 iee2T'.6 88 79 B29 Ae enka? 17-5 371 | 21.7 84 79 qi eiges 16.9 318 | 17.3 | 81 74 | 31 Petignre: 17.9 233 16.8 | 81 PN ee 12.3 17-2 258 18.3 87 [79 33 13.6 17.6 279 18.7 85 ne ei 12.9 17.2 pe 16.1 81 76 32 1262, 17.6 229 eet 714 71 32 13.3 17.5 130 18.8 | 84 79 32 T3755 18.1 261 17.2 | 82 78 ated bag Vee) 17.4 234 18.7 86 81 32 1351 17.5 258 20.8 88 pone 31 13.0 17.6 306 Meanie | 16.964:| 81.41 | 75.73 | 32-09 | 13.005 | 17.359 | 264.59 | 0.346 | +0.69 | 0.45 | +0.11 | 40.056 | +0.055 | +8.98 Standard | 2.724 4-78 3-11 | 0.74, 0.389 0.382 | 62.48 pee ge o-2 77 £0.49 | £0.32 | 0.08 | Fo.040 0.039 | £6.35 124 Series of 1907-1908: Francis B. Sumner TABLE 6B (Males unmated) killed in 24 months. COLD Room (21) Standard 2.888 deviation... -+0.301 Body length Pee Caudal a | vertebra 59 31 60 31 59 31 59 31 60 32 61 31 57 [28]* 62 32 58 [29]* 64 33 60 32 62 32 64 33 65 32 51 3° 53 32 | 60 31 Pe zke 33 67 | 30 65 | 32 | 67 31 | 61.10 31.58 0.65 | +0.14 | | 4.41 0.88 +0.46 | £0.09 | \Hair in Ear Foot | ye ESD 12.6 16.4 462 12.8 16.8 | 349 Dag) 16.5 | 237 aly el 328 12.6 | 17.1 | 330 12.9 166,95) S257 2/15 16.2 | 228 12.2 | 17.4 | 233 12.0 16.0 | 205 12.6 16.9 | 293 1233 16.7 | 239 12.4 16.8 | 227 13-5 | 17-5 | 364 1369 | O75 ch. Oe Me ade i i 2 12.2 16.3 225 13.0 | 16.9 | 284 TA SO 762 2EG 13-4 17-3 284 rs | 16.7 383 iehsyfe IA a sigese) | 294 | . 12.786 | 16.800 | 294.76 +0.082 | +0.065 + 11.66 0.558 | 0.441 | 79-17 +0.058 | +0.046 | +8.24 125 Effects of External Conditions vege | St-9- L1°6L gh z9 99°11 | 968+ gl'v6z | 6S-bgz PICO. WIE SNVUSITIIN NI YIVH dO LHOIGM | | | gto'oF | 6£o'oF | gSo°oF | obo-o-F | b0'0F | goo | gho | tho | SS-oF | 6boF | 10b-oF | LizoF Zgt°o gSSo 6gt-o ggio tL: 1b'+ 11°f €£°S | gl BB z bzL'z Sgo‘o=F | SSo-oF | zgo°oF | QSo°0F | VIOF | ITF Sg'oF | Sb oF | glroF | 69°07 | Sth-oF ght-oF oog'gt | OSE*L1 | ggdsz1x | Soor tr gst 60°zt or'I9 EL°SL | 1L:0g | rb-1g | So6-Lr | ¥96°91 PICO. ULIE MA plod WIR AA, PISO WIP AA PIOD me | plod | ue | POD Wat AY =. | AVAIALUGA | Lood ava HIONGT TIVL HIONGI AGO LHOIGM avanvo widWaNn uonerAsp piepueys uostsvg u02 4of pasuvsiv saan ty *syuom ¥% I patty (pawwuun) sajvyy 09 GTdVL :9061-Lo6r fo satiag 126 Francis B. Sumner figures will be later combined, however, with certain of those given for the next group. The unmated males, measured at the age of 24 months, consti- tute the most important groupin the second year’sseries. ‘Table 6 (A, B, and C) presents the measurements for these 43 mice. These measurements were all made after killing. Comparison of these figures with those given for the males at 42 days is of course only possible in respect to two characters— weight and tail length. In order to determine accurately what changes have occurred in these, however, we must first combine the figures for the present group with the preceding group of “mated males,” since the two together comprise the entire col- lection of males which had been measured earlier in life.** ‘Table 7 accordingly represents the mean weight and tail length for all males at the age of 25 months. TABLE 7 Series of 1907-1908: All males at 24 months of age WEIGHT TAIL LENGTH | i | | Warm (29) Cold (27) Warm (29) | Cold (27) | | Meany tseeeel 17.176 | 18.396 | 77-41 | 63.33 | 0.331 | 0.396 | 0.44 | 0.56 Standard deviation... .| 2.646 3.053 | 3-54 4.28 +0.234 +0.281 +0.31 +0.39 Comparing these figures with those of the males at six weeks of age, we note that whereas the tails of the cold room lot have increased 16.9 per cent in the interval between the measurements, those of the warm room lot have increased only 12.5 per cent. There is thus seen to be a tendency to “catch up” on the part of the less developed organs, which has already been pointed out for 24 In combining these figures an allowance is first necessary. The tail length of the dead mice was, as stated above (p. 103), obtained by a different method from that practised upon the living ones. I have found that in living mice of this size the tail is stretched on the average about 1.5 mm. during the suspension. This amount has accordingly been added to the mean tail length of the unmated ~ males before combining with that of the mated males. The resulting figure represents approximately the tail length which would have been obtained had all been measured alive. Effects of External Conditions E27 the females and will be discussed later. It must be added, how- ever, that in the present case the difference in weight between the two groups has increased rather than diminished. In respect to variability, two of the four standard deviations comprised in the table show an absolute decrease, one of the others indicates a slightly lessened variability, while in the fourth case there is an increase, both absolute and relative. Not much importance is to be attached to these latter comparisons, however. Returning to a consideration of the figures presented in Table 6 it is seen that the weight is 5.3 per cent less in the warm room lot than in the cold room lot; while the body is seven-tenths of one per cent longer and the tail 23.9 per cent longer. Passing to the new measurements (not applicable to living animals), the aver- age length of the (left) ear is 1.7 per cent greater in the warm room lot; that of the foot 3.3 per cent greater. The average weight of hair (see p. 102) for the cold room mice is 11.4 per cent greater than that of the warm room individuals. It has seemed worth while to represent graphically the frequency distributions of these characters (Figs. 4 to 9). The difference in weight between the two sets of mice (Fig. 4) cannot with any certainty be regarded as a significant one. ‘The difference in body length (Fig. 5) is too trivial to be taken into consideration. In striking contrast, however, is the case for tail length (Fig. 6). Here there is no overlapping whatever of the polygons, while the modes are removed by a distance of 13 mm. ‘Two questions present themselves respecting this difference of tail length: (r) Does it involve an actual difference in the volume of the organ? And (2) does it involve a difference in the number of vertebrz ? In order to test the first question the diameter of the tail at its widest point was obtained by means of calipers for all the mice of this group. While this is a dificult measurement to make with any great accuracy, the figures are probably exact enough for present purposes. The averages for warm room and cold room animais are 2.94 mm. and 3.02 mm., respectively. There is thus the possibility of a slightly greater attenuation of the tail, accompanying its more rapid increase in length, but the former is certainly nothing like proportionate to the latter. There is * 128 Francis B. Sumner therefore an actual difference in the volume of the tail. The number of caudal vertebrz was likewise counted for all of this Fig. 7 Series of 1907-1908: ear length of Fig. 8 Series of 1907-1908: foot length of LP) ““unmated males” at 24 months of age. “Cunmated males” at 24 months of age. group. Ihe mean figures are given in the table* (6), where it is seen that the difference between the two groups (if significant 5 The number of vertebra was counted, from the sacrum to the last nodule of bone distinguishable under a low power objective. The task was comparatively simple and admitted of tolerable accuracy. The average error of observation was probably less than one vertebra per mouse. Since the skele- Effects of External Conditions 129 at all) amounts to only about half a vertebra, i. e., it is not sufh- cient to make any appreciable difference in the tail length. In fact there is very little deviation from the mean throughout the entire series. It would be surprising indeed if the number of vertebra had been found to be altered by temperature conditions, first because this number does jnot in all probability vary very widely in most species of mammals, and, secondly, because the definitive division of the embryonic tissues into vertebrae is prob- ably complete long before birth. The present difference in ear length is not convincing, but foot length, has, with practical certainty, been affected by the tempeature conditions. It must be remembered that we are here dealing with mice which differ among themselves widely in size, and that neither the mean weight nor the mean body length is quite the same for the two groups. In Table 8 are presented the relative magnitudes of cer- tain characters. Here we have the mean ratios between the length of tail,ear and foot in each individual and the length of the body; likewise the ratios between the weight of the hair and the square of the body length. Inthe case of the ear and the foot, the vari- ability of the ratios is found to be much greater than that of the absolute measurements. ‘This is due to the fact that these parts vary but slightly as compared with the size of the animals. Indeed their length is remarkably constant throughout each series, irre- spective of the body length of the individual. The case of the Aair deserves a rather full discussion, since this is the character, in particular, whose modification may be supposed to be of an adaptive nature. It is seen from Table 6 that the mean weight of the pelage for the cold room individuals is 11.4 percent greater than that of the warm room individuals. The variability is very high, to be sure, partly because the animals vary much in size, partly because they actually vary in the density of their hair ton was not thoroughly denuded of muscles, etc., (alcoholic specimens were used) it was not always easy to distinguish the termination of the sacrum and the commencement of the caudal series, and an error of one or two vertebre perhaps resulted occasionally from this cause. In a few instances, some of the minute terminal vertebre were lacking, owing to obvious injury to the tail. In such cases, the figures have been enclosed in brackets and have not been included in making up the averages. Francis B. Sumner T30 10100 oF glooo’ oF Sfrof | grroF 1g0°O-F 660°0 6S-oF oS oF | $1600°0 1LLoo'o Loz" Str glo gl6°0 Lo°S S6°+ St100'0 I1100°O 161°OF got’ oF Siro Iv1°O fg°o- Wiese 10S+0"o 6S6£0°0 S6g°0z 16€°17 1Lg°S1 £z0°91 93'S fz" £6 (17) plea (zz) WIR (17) plea (zz) wae A (17) plod (zz) war A\ (17) poo (zz) wae {(HLONAT aoa) ISNWWYD NI YIVH JO LHOIAM HIONAT AGOD SHLONAT LOO | HLONAT AGO ‘HLONGT ava HLONAT AGO sHLONAT TIVL ; “*** WORIAIp plepueryg yisua, Kpog fo sadvjuarseg ut passaadxa ‘sonvi—sajDu pajwuu lr) :9061—Lobr fo satsag 8 TIAVL Effects of External Conditions 131 coat. I have therefore made an endeavor to compute the relative amount of hair, making allowance for the area of the skin—that is, | have obtained the ratio between the hair weight of each mouse and the square of its body length. The mean of these figures is 0.04501 mg. per sq. mm. for the cold room lot, 0.03959 for the warm room lot. According to this com- putation, the cold room mice have a relative amount of hair 13.6 per cent greater (heavier) than the warm room ones. It must be admitted, however, that such a method of computa- tion is open to criticism. To estimate the relative skin areas of these mice by comparing the squares of their body lengths presupposes that they are, in the language of geom- Fig. 9 Series of 1907-1908: weight of hair (absolute) of ‘“‘unmated males” at 24 months (expressed in hundredths of a gram). etry, “similar solids,” which they are not. As a matter of fact, while the warm room mice have a slightly greater body- length than the contrasted group, they are lighter, on the aver- age, by nearly one gram, i. e., a difference of over 5 percent. They are probably somewhat less plump, therefore. I must add, however, that I do not believe any such slight difference of shape to be accountable for the difference in the amount of hair which is shown by the two groups of mice. The most serious criticism of these figures relates to the number of indi- viduals, which is confessedly too small to permit of our drawing any final conclusions in the presence of such high variability. 132 Francs B. Sumner Until further data are available, however, it seems worth while to offer them for what they are worth. Fig. 9 shows the distribution of hair weight for the entire lot, both the hot room and the cold room individuals. No fair com- parison can be drawn from these polygons, as has already been stated, since mice of very different sizes are represented. In Fig. 1o I have substituted in each case for the absolute weight the ratio of the hair weight to the square of the body length. While the modes lie at the same point, the centers of gravity are con- siderably separated. In view of statements cited by Lydekker® regarding certain modifications which are alleged to have been produced in cats by Fig. 10 Series of 1907-1908: weight of hair (relative) of “‘unmated males” at 24 months (= ratio of hair-weight to square of body length). 3 ake life in a cold-storage warehouse, I endeavored to determine for the present group of mice whether there was any appreciable dif- ference in the length of the vibrissae between cold room and warm room individuals. The measurements have necessarily been rough in the extreme. The longest hair among the “whiskers” of one side was measured by calipers, without straightening or removing it. The mean figures obtained were 26.8 mm. for the warm room lot; 25.4 for the cold room lot. Considering the crud- ity of the method, the individual measurements vary compara- - ® A Handbook to the Carnivora. Part I: Cats, etc.; pp. 158-163. Effects of External Conditions 133 tively little, though enough, probably, to deprive this difference of any significance. No such obvious modifications as has been alleged for cats 77 is here evident. Measurements at 7 Months. ‘The subsequent history of the female mice varied considerably with different individuals, accord- ing to the exigencies of the breeding experiments. Each female from either lot, as she became pregnant, was transferred to a room kept at a temperature somewhat intermediate between the hot and cold rooms. If, as was commonly the case (see p. 116), her brood did not survive, she was taken back to the room from which she came. Thus, in the interval between February 6 and April I, a considerable proportion of the females were transferred back and forth between the warm or cold room and the “intermediate” room, in some cases more than once. I[ have not thought it worth while to compute the average duration of each set of conditions for the lot. On April 1, all the mice were moved to the “intermediate” room and the sexes separated. On May 1, they were paired again, but with the same unfortunate results. ‘The entire lot of females was killed and measured between June 4 and July 6. All were about 7 months of age at the time of killing, save for a few mothers of broods, which were allowed to remain with the latter till they were old enough to take care of themselves. [hese somewhat older individuals (74 to 8 months) have, however, been included in the table herewith given. As they were, with little doubt, all fully grown, this proceedure seems fair. The mean figures obtained for each lot of females is given in Table 9g. It herewith appears that the weight in the warm room lot is 2.4 per cent greater than in the cold room lot, the body two- tenths per cent longer, the tail 14.9 percent longer,*the foot 4.1 per cent longer, while the average ear length is practically equal in the 27 Tt is true that in the case cited by Lydekker the elongation of the vibrisse was attributed to the darkness, rather than to the cold. 8 Tt seems probable that, so far as the tail at least is concerned, the cold-room mice have departed from the more usual or normal condition, while the warm-room individuals have been little if any modified. Fifty-nine adult female mice, of unknown history, which were received by me during the present winter, had a mean tail length of 92.8 mm.; i. e., their tails were considerably longer than even the warm-room females of Table 7. It must be noted, however, that they were larger mice, having a mean weight of 26.6 gms. Francs B. Sumner obb'oF | SLE-0F Lo1*Sz LSg: tz (9) (4) PICO WIE AA LHOTIM SACO O COOLIO A ey hy ZgtiO-+= | tSoro== | 1WIye== | cogto== || 9S-os- geo gS°o- 98°07 fgz° Li rl Ply Logo: +1 tbo +1 cecal LL 00° Lg 69°S6 11° $6 OF | COs ie: © 0) | & PIPD, UIE PIOD | WHE AA ie COA /AN PICO OTE Lood Uva (ONIAIT) HIONGT TIVE HLONAT AGOd plo syiuowm ZL inoqv .‘pajwu,, savpy %9061-Lobr fo satiag Ol ATAV.L ¥g0"O== || Sto:0== | zSo-o== | FEo:o== || SS*o== Loo Iv-oF | ob-oF £LS-o Zgt‘o tgb'o ggt-o 06"+ £6-z zg°f St-+ 160'0 | oSo'oF | tlo‘oF | gto'oF | gl-oF gt:oF gS*oF 9S'oF 1£9°91 zee: Li fol: €1 trl: tr Lg*SL 96°98 +g° £6 fo +6 (8!) (4z) (81) (47) (81) (42) (81) (47) PI°O UIE PIOD EAN PICO COIB AN PIPD ure (©NIAIT) Lood ava HIONAT Ado HLONAT TivL fgf-oF | Lit-oF Hate LSb° tiS‘of | 6h oF ZS 2 6So° tz (st) (47) PI°D UIE LHOIGM * WORIAIp prepueys abs we JOOHOOMAT EASY Fi 13 plo symou £ inogv sajpuag :go61-Lobr fo satsag 6 TIAVL Effects of External Conditions 135 two lots. None of these differences except those in the length of the tail and foot are to be regarded as having any significance. Fig. 11. Series of 1907-1908: tail length of females at 7 months of age (based upon meas- urements after death—see text.) Fig. 12 Series of 1907-1908: foot length of emales at 7 months of age. Comparing:the measurements for this age with those (weight and tail) made%upon these same mice?® when 24 months old, we find 79 Six of the warm-room lot and three of the cold-room lot had died in the meantime. 136 Francis B. Sumner that the weight has increased during this interval 48.4 per cent in the cold room (1. e., lighter) lot; 45.8 per cent in the warm room (heavier) lot. The gain in tail length for the cold room lot has been 25.7 per cent, as compared with an increase of only 14.2 per cent on the part of the warm room individuals. In respect to each character, therefore, but more especially in respect to the tail, there has been a very obvious tendency toward a diminution of the differences between the two contrasted sets. Referring to variability, a comparison of the standard deviations for tail length shows that there has been a slight reduction in the absolute and a considerable reduction in the relative variability in both the warm room and cold room lots. The standard devia- tions for weight have undergone an increase in each lot, though the relative variability has remained practically unchanged. Frequency polygons (Figs. 11 and 12) have been plotted for tail length and foot length in this group of females. The polygons for the former overlap to a very slight extent; those for the latter are sufficiently distinct to admit of no doubt as to their significance. The males used for breeding were kept under the extreme tem- perature conditions until April 1, at which time they were 4to 5 months old. They were then transferred to the “intermediate room” along with the females, and were killed upon reaching the age of 7 months. The measurements for this group are given in Table 1o. Comparing the present figures with those for the same mice at the age of 24 months, we find that the lighter warm room lot has gained to the extent of 32.7 per cent of its former weight, while the heavier cold room lot has gained only 25.1 per cent, again showing a tendency toward equalization. he same fact is evident in the growth of the tail. This has amounted to 17.9 per cent in the case of the cold room lot; 11.5 per cent in the case of the warm room lot. Levelling Down of Early Differences. Reference has more than once been made, in discussing partic- ular sets of measurements, to a tendency for these experimentally produced differences to diminish with growth. A table has been Effects of External Conditions 137 prepared (Table 11) which inciudes all the cases in which this point can be tested. TABLE 11 Series of 1907-1908: percentages of increase* in the intervals between successive measurements 6 WEEKS 2} MONTHS 7 MONTHS { ABSOLUTE MEASURE- | PERCENTAGES OF | PERCENTAGES OF MENTS INCREASE | INCREASE | . ; aa ae | : Weight Fal | Weight Tail Weight ea length length length SS ein { Warm.........| 12.997 | 68.83 | 32.1% | 12.5% | 32.7% | 11.5% tet | Cor a ee 13.123 54-16 | 40.2% | 16.9% | 25.1% 17.9% | | = $$$. Femal (Wits cease acl 12.282 | _ 69.06 28 .8% 10.3% 45-38% 14.2% poe \ Cold ae ee | 11.400 I “SEs91 | 33.1% | 16.0% | 48.4% | 25.7% **See foot note on p. 120 above. tIn considering the figures for 7 months, it must be recalled that only 13 males (7 warm + 6 cold) have been kept till this time. In figuring percentages of increase for these males, therefore, the later figures have been compared with those for this same group of ‘mated males,” and not with those for the males in general. In the column for 24 months, on the contrary, the figures for “all males’ . are given. Under the “‘6 weeks” column, we have the absolute measure- ments of weight and tail length for males and females belonging to the warm room and cold room lots. In the column for 25 months is given for each group the percentage of increase of each of these characters, during the interval between the two measure- ments. In the “7 months” column are given the percentages of increase over the measurements for 24 months. For each sex are presented two horizontal rows of figures: those for the “warm” and the “cold” lots respectively. It is thus easy to compare the contrasted figures for any one character. In each pair of these contrasted figures, that one has been printed in heavy type which represents the rate of increase for the group which had previously shown a Jower mean value for the character in question. ‘This group should, according to the hypothesis, be expected to have undergone a more rapid rate of increase. As a matter of fact, it will be noted that in 7 out of the 8 pairs of contrasted figures, the LE-o=F te-oF gto 6f-0o gti'oF £S1:oF ol1‘o-F 10t:0=F © \ +++ yonelAap prepurig t9'+ 19°F oz’ g6°S tgl't Lo1'z oIe 611°t f oad tS:o- gh oF 1S'o SS:oF S61°o+ Lit‘oF 1vz-oF €gz°o-F i\ as eb arte ogi 6b° 6S $6" 89 11°09 61° Lo 6gg° 11 £99°7I ogi’ tr bog* ZI if (S£) poo (ch) wire Ay (Lt) plod (£5) wre Ay (gf) ploo (£+) we Ay (0S) pjoo (SS) wie jy So[ PU SoTeTAL SopPUlo Soe Francis B. Sumner HLONGT TIVL . LHOIEM 138 syaam xis JD yis.uay piPt pup 14319 4 :6061-g061 fo satsay a1 TIAVL Effects of External Conditions | 139 number in heavy type is the larger number.?° This decrease in bodily differences originally brought about by differences of temperature has not been dueto a withdrawalof thelatter. Indeed, during the interval between the “6 weeks’’ measurements and the ‘“‘24 months” measurements the temperature differences in the two rooms have increased rather than diminished. (See Fig. 1.) Later, it is true, the temperature differences gradually diminished, and commencing with April 2 they were abolished altogether. At the latter date, however, the mice averaged nearly five months old, and their growth was probably not far from complete. While the tendency toward a reduction of the original differ- ences between the warm and the cold room groups is thus pretty clear, the evidence for a reduction of variability within each group is not so certain. An inspection of the tables shows us twelve cases in which we can compare a later standard deviation with an earlier one for the same character. In six of these cases the later standard deviation is smaller, i. e., the decrease in variability has been absolute as well as relative. In two cases there has been a relative decrease, though not an absolute one; while in two others, the relative variability has remained practically unchanged. In only two of the twelve cases has there been any appreciable in- crease in the relative variability. In view of the lack of uniformity in these results, however, and the commonly high probable errors, too much significance must not be attached to them. It is worth pointing out, however, that the variability for tai] length has de- creased absolutely as well as relatively in five out of six cases. Series of 1905-1909 At the date of writing, the experiments of the third winter have not been carried very far. A first and second series of measure- ments upon the living mice have, however, been made, and the results seem well worth comparing with those of the preceding 39 Tt is likewise true that in all of these seven cases there has been a greater absolute gain as well as a greater percentage increase. Rate of growth seems more fairly expressed, however, in terms Of proportionate increase. 140 Francis B. Sumner year." The temperature conditions during the first four months are indicated in Fig. 1, in which a certain degree of comparison with those of the previous winter is made possible. The,mean temperature of the warm room during the first three months of i ee va (oF GN sive TG Fig. 13 Series of 1908-1909: weight at six weeks (sexes combined). the animals’ lives was 21.3° C., with a mean daily range of 12°|C. In the cold room the mean temperature during this same period was 3.0° C., the range of daily fluctuation being 6°. The temper- *' The mice here used are from an entirely new stock, none of them being descendants of those raised during the preceding year. Animals from several different localities, having independent pedigree, are comprised in the lot. In the present series individual litters were not divided into a “warm room” and “cold room” half, as had been done with those of the preceding year. ~ Effects of External Conditions 141 ature conditions will be further discussed after the first table of measurements has been presented. ‘The mean relative humidity in the warm room (12 determinations) was 38.5 per cent, a figure very close to that of the preceding year. The humidity of the cold room (14 determinations) was 76.5 per cent, being thus consid- erably greater than that of the preceding year. ‘This is accounted for by the fact that the cold room used during the present year has been situated in a building which is entirely unheated. With respect to weight there is a remarkably close agreement between each of the four averages here presented and the cor- responding one of the preceding year. In each year’s series, the males are larger in the cold room lot, while the females are larger inthe warmroom lot. From the magnitude of the probable errors, however, we cannot feel sure of the validity of these differences.” Regarding tail length, it will be noted that the differences, while considerable, are very much smaller than those to be found in Table 3. The warm room males have tails which are 11.8 per cent longer than those of the cold room lot. ‘This difference, in the case of the 1907-1908 lot, was 27.1. The warm room females show a mean tail length which is 15.9 per cent greater than that of the cold room individuals, as compared with a difference of 33 per cent in the earlier lot. Thus, as in the preceding year’s measurements, we find that it is the females which have been modified most in this respect. But the amount of modification tor each sex has been less than half that which was earlier observed. Regarding variability no deductions of interest are to be drawn. No such evidence of a higher variability in the cold room lot as was previously noted is here to be found. In fact, the reverse is possibly true. The salient fact brought out by this comparison is the relatively small degree of modification in the tal length of the present lot of animals. ‘This fact is not satisfactorily explained by a com- parison of the temperature conditions for the two years. We find, it is true, in the later series a somewhat smaller difference in tem- perature between the two rooms during the first six weeks of the 32 See supplementary note below. ‘A CET AOS TESTS UT AT ORI TT et tail length at six weeks (sexes combined). Fig. 14 Series of 1908-1909 Effects of External Conditions 143 animals’ lives. This difference here amounts to 14.5° C., as com- pared with 16.6° during the winter of 1907-1908; and it might be at once inferred that in this fact we had a key to the difference of results. It must be noted, however, that, while the warm room temperature has been considerably lower for this period, during the later year (19.8° as compared with 23.7°), the cold room tem - perature has likewise been somewhat lower (5.3° as compared with 7.2°). Now a comparison of Tables 3 and 12 shows us that while the warm room tails agree pretty closely in length in the two series, the cold room tails are much shorter in the earlier one TABLE 13 Series of 1908-1909: percentages of increase in the interval between successive measurements 6 wrexKs* 3 MONTHS ABSOLUTE MEASUREMENTS | PERCENTAGES OF INCREASE | Weight pail, \e Weta 2 |). erat MEN ee 12.604 67.19 Ohaktias 20.31 + 0.61 Males...... | | | yCold’s.<: eres | iginicte), 60.11 | 58.86+ ? 22.79+ 0.46 | SN EE EEE ol Soe SA ORs Neca Oem. a ee me f Wiatiirs cv en veal 12.663 68.95 on ACL OO RE | 15.81 + 0.42 m ai Saini i Cold eC ENG | 11.889 59-49 40.75 ? | 18.53 0.66 } | *See Table 12. Were these modifications simple functions of the temperature dif- ferences, we should not have expected such a state of things. Indeed the author has no explanation to offer for the striking difference between the two years’ results. A second set of measurements was made upon the same mice at the age of three months. The statistical treatment of these later figures is not yet quite complete. I have determined, however, the mean percentage increase in weight and tail length for each sex, both in the warm room and the cold room lots. Since every 33 This is the figure for the room in which the thermograph was kept, and in which most of the mic lived during the greater part of this period. For certain reasons, however, all of the animals were kept for a period of varying length (8 to 42 days—the last in the case of only one brood) in another room, having a mean temperature about 4° higher than that of the room first referred to. This fact of course complicates the situation somewhat. 144 Francis B. Sumner mouse in the present year’s experiments is identified by a mark of its own, it has been possible to compute the rate of increase for each animal individually. In Table 13 are presented the mean percentages of increase for each character during the period in question. It will be seen from this table that in all four cases. the figure expressing the increase in a character is /arger in that group in which the previous absolute measurement had been smaller. It must be added, however, that in neither case is the figure for weight of any significance in this connection, despite the differ- ences between the averages. ‘The variability in the weight-increase has been enormous (ranging fromg per cent to 143 percent), so that the probable errors (not yet computed) are undoubtedly very large. For the growth of the tail, however, the case seems fairly certain. It w.ll be recalled that in the preceding series it was the figures for tail length which bore the strongest testimony to the principle of the leveling down of original differences. Reference to the temperature curves shows that here, as previously, the differences in the conditions have increased rather than diminished during this period. As regards the increase or decrease in variability within each lot, nothing can be said here, since the standard devia- tions have not been computed for the 3-months measurements. Supplementary. A yet later series of animals (born March 1909), consisting of a larger number of individuals than any of the preceding lots, has yielded, after similiar teatment, the fol- lowing results: (1) a difference in tail length somewhat greater than that shown in the preceding series (16.7 per cent in the pres- ent case); (2) an indubitable difference (both absolute and rela- tive) in foot length, which applies equally to both sexes and fully confirms the earlier conclusions on this point; (3) a simi- lar difference in ear length, which, however, is of far less certain significance; (4) a difference in weight, both sexes being heavier in the warm room lots (cf. pp. 118,138 above),'although this differ- ence was greater for the females (11.3 per cent) than for the males (9.1 per cent). The temperature differences to which the animals were subjected had already begun to diminish at the time of the first measurement (42 days), and both lots were trans- Effects of External Conditions 145 ferred to the same room when at the mean age of II weeks. When measured later, at the mean age of 3 months, the differ- ence in tail length had diminished to 6.7 per cent. The data for this series have not yet been fully compiled. SUMMARY OF STATISTICAL DATA The more significant facts which may be disentangled from this mass of data may be briefly summarized as follows: 1 The tail length (whether absouute or relative) was found to be very much less in the case of those mice which were reared at the lower temperature. 2 This difference was not accompanied by any appreciable difference in the number of vertebre. 3. The foot length was likewise considerably less in all of the cold room lots, though the differences were not so large as in the case of the tail. : 4 In two series, at least, the mean ear length was found to be smaller at the lower temperature, but these differences are perhaps to be regarded as “accidental.”’ 5 Body length was not affected with any degree of certainty; while the influence of temperature upon weight, although evident in certain cases, was inconstant, and seemed to depend upon sex. 6 Differences in the average quantity (weight) of hair have been demonstrated for the only series in which the point has been tested. The cold room lot were found to have an average amount of hair which was 11.4 per cent greater absolutely and 13.6 per cent greater relatively (1. e., allowing for the dimensions of the animal) than that of the warm room lot. The number of indi- viduals was, however, small (43) and the variability was high. 7 In another experiment a considerable difference in the aver- age number of hairs per unit of skin area was found among the male mice, the number being greater in the cold room individuals. Here, likewise, the high variability and small number of individ- uals render any conclusions doubtful. 8 No constant differences in variability were observable be- tween the warm room and the cold room mice. g To what extent the modifications cited above have resulted 146 Francs B. Sumner from differences in temperature and to what extent from differences in humidity is not certain under the conditions of the experiments. 10 Comparisons of earlier and later measurements upon the same animals show that there is a distinct tendency toward the reduction of these experimentally produced differences during sub- sequent growth, even when the conditions remain unchanged. 11 A diminution in variability within each of the contrasted groups, during the course of growth, was shown to be probable for tail length, at least, and possible in the case of weight. In order to complete this summary, I will add by way of antici- pation: 12 The modifications thus artificially produced are such as have long been known to distinguish northern from southern races of mammals. GENERAL DISCUSSION OF RESULTS In the following comment upon the results of these studies, I shall commence with the last statement in the summary. Itisa most significant fact that the experimentally produced differences which have been discussed in the present paper are found to be of just such a nature as are recognized by mammalologists and orni- thologists as distinguishing the northern from the southern rep- resentatives (individuals or geographical races) of some species. J. A. Allen has repeatedly called attention to the “marked ten- dency to enlargement of peripheral parts under high temperature or toward the Tropics.”’* Baird and other writers had previously made incidental mention “of the larger size of the bills of southern representatives of northward ranging species’? of birds, but Allen offers some detailed examples of this.* He concludes that “an increase in the length of the bill is most frequent in long-billed species, while in short-billed ones the increase is in general size, without material change in its proportions” (p. 231). A greater curvature sometimes accompanies the increase in length. Like- 34 The influence of physical conditions in the genesis of species. Radical Review, I, 1877, pp. 108- 140. (Reprinted with annotations in Annual Report of the Smithsonian Institution for 1905.) 3 On the Mammals and Winter Birds of East Florida, etc. Bulletin of the Museum of Compara- tive Zodlogy, vol. 2, 1871, pp. 161-450, pl. iv—viil. Effects of External Conditions 147 wise, “there are well-known instances of an increase in the length of the tail” (meaning the tail feathers). Much more explicit statements are offered regarding mammals, both by Allen and by Coues, though it is stated by the former that the responses to climatic conditions are less evident in this class than among the birds. ‘In mammals which have the external ear largely devel- oped, as the wolves, foxes, some of the deer, and especially the hares, the larger size of this organ in southern as compared with northern individuals of the same species is often strikingly appar- ent... .. In Lepus callotus|[‘Lepus texianus and its subspecies’ —later note], for example, which ranges from Wyoming southward far into Mexico, the ear is about one-fourth to one-third larger in the southern examples than inthe northern. . . . . Among the domestic races of cattle those of the warm temperate and inter- tropical regions have much larger and longer horns than those of northern countries. . . . . Naturalists have also recorded the existence of larger feet in many of the smaller North American mammialia at the southward than at the northward among indi- viduals of the same species. . . . .”’ In his monograph on the Muridz*? Coues repeatedly makes similar statements. Refer- ring to a mouse, “Hesperomys leucopus”? (now Peromyscus leu- copus) (p. 66), he says: “The arctic series averages larger than the United States specimens, and has shorter feet and ears, as well as shorter tail,’ and he alludes later to “the well-known law of smallness of peripheral parts in Arctic mammals” (p. 83). Comparing the red-backed vole, “Evotomys rutilus gapperi,”’ a more southern “variety,” with the species “E. rutilus,’®* he finds that the vertebral part of the tail is, on the average, about a third of an inch longer in the former, while the foot is 72 hun- dredths of an inch in length, as compared with 70 hundredths of an inch in the northern form. Relatively, the differences are even % Op. cit., 1905, pp. 382-384. 37 Monographs of North American Rodentia.—Report of the U. S. Geological Survey, vol. xi, 1877, pp- I-I091. 38Tt is quite possible that two or more distinct species are here referred to. I am not sufficiently familiar with the classsification of the Muride to know the present status of the various species and varieties referred to by Coues. ‘‘E. rutilus gapperi” is now regarded as a true species, Evotomys gapperi. 148 Francis B. Sumner greater, since the northern animals are of larger size. Indeed, the authors cited dwell with equal emphasis upon the Jarger size of the northern representatives (individuals or varieties) of species, both of birds and mammals, as compared with the’ southern. Previously, Allen tells us, Baird had “explicitly announced a general law of geographical variation in size; namely a gradual decrease in size in individuals of the same species with the decrease in the latitude and altitude of their birthplaces.*® And Allen further affirms that “this is true not only of the individual repre- sentatives of each species, but generally the largest species of eacn genus and family are northern.‘ The foregoing statements were made before the days of exact biometry, and an examination of the tables of measurements offered us shows that in most cases they comprise relatively few individuals and that the material used was not homogeneous, i. e., it includes alcoholic and fresh specimens, as well as dried skins. For this reason most of these tables are not likely to be of very great use to the modern student of variation. In more recent years, a very extensive mass of similar measurements has been gathered by a considerable number of naturalists, but, so far as the writer has been able to discover few if any of these have been subjected to statistical treatment with reference to testing the generalizations of Baird, Allen and Coues.*! It would seem overskeptical, however, to reject the emphatic opinions of a num- ber of able naturalists upon these matters, particularly as we have no more satisfactory data at our disposal. Regarding the pelage, there can be little doubt that this likewise responds directly or indirectly to climatic conditions. “At the northward, in individuals of the same species, the hairs are longer and softer, the under fur more abundant, and the ears and the soles of the feet better clothed. This is not only true of individ- uals of the same species, but of northern species collectively, as compared with their nearest southern allies.”*? Both Coues 39 Op. cit., 1871, p. 230. 40 1905, p. 378. 417 state this on the authority of several of our leading students of mammalian distribution to whom I have appealed for information. # Allen, 1905, p. 382. Effects of External Conditions 149 and Allen cited many specific instances of this fact for mice, hares, squirrels and other rodents. Moreover, obvious seasonal changes are to be observed in some species. Speaking of the squirrel, Sciurus hudsonius, var. hudsonius, Allen says: “In summer the soles of the feet are naked, often wholly so to the heel; in winter they are wholly thickly furred, only the tubercles at the base of the toes being naked. ‘The general pelage is also much fuller, longer and softer in winter than in summer.” It may well be that the change in the quantity of hair which appears to have been produced in the white mice during the exper- iments above described was comparable to these seasonal changes, i. e., that the results were purely temporary, and would have dis- appeared with a cessation of the conditions employed.* Indeed, since the life of an individual hair is comparatively brief, it would be necessary to effect some permanent change in the physiological activity of the hair follicles, in order that differences such as these should endure. Whether or not these effects are permanent, it has been believed by many that various changes in the character of the hair coat occur in domestic animals as the result of trans- ferrence to an unaccustomed habitat. Darwin, indeed, tells us* that “Great heat, however, seems to act directly on the fleece; several accounts have been published of the changes which sheep imported from Europe undergo in the West Indies. Dr. Nichol- son of Antigua informs me that, after the third generation, the wool disappears from the whole body, except over the loins; and the animal then appears like a goat with a dirty door-mat on its back.” And again:® “It has been asserted on good authority [Isidore Geoffroy St. Hilaire] that horses kept during several years in the deep coal-mines of Belgium become covered with velvety hair, almost like that on the mole.’ The “classical’’ Porto Santo rabbit may be cited as another and perhaps more authentic instance of the modification of mammals through changed cli- 8 Monographs of N. A. Rodentia, p. 675. 44Tt is uncertain, to be sure, in how far the season changes of the hair coat of mammals are direct responses to climatic conditions. 45 Variation of Animals and Plants under Domestication, vol. i, p. 124. 4 Op. cit., vol. ii, p. 336. 150 Francis B. Sumner matic conditions. Here, not only the hair, but other features, were affected. So far as the present writer is aware, however, no such differences as have formed the principal theme of this paper have been pre- viously brought about by direct experiment or even produced under such circumstances as would warrant one in stating positively that they were the immediate results of external conditions. Lydek- ker, in the work already referred to, cites a case on the authority of “an American newspaper”’ (so notoriously infallible in matters scientific!) which would certainly be important if true. It is worthy of mention only because the modifications alleged accord so well, in some respects, with those which have been demonstrated for mice. In order to combat the rats in a cold-storage ware- house at Pittsburgh—so the story runs—cats were introduced. The first of these died. “One cat was finally introduced which was able to withstand the low temperature. She was a cat of unusually thick fur, and she thrived and grew fat in quarters where the temperature was below 30°. By carfeul nursing, a brood of seven kittens was developed in the warehouse into sturdy thick-furred cats that loved an Icelandic climate. They have been distributed among the other cold-storage warehouses of Pittsburgh, and have created a peculiar breed of cats, adapted to the conditions under which they must exist to find their prey. These cats are short-tailed [italics mine], chubby pussies, with hair as thick and full of under-fur as the wild cats of the Canadian woods. One of the remarkable things about them is the develop- ment of their ‘feelers.’ . . . . In the cold warehouses the feelers grow to a length of five and six inches. This is probably because the light is dim in these places, and all movements must be the result of the feeling sense.” I am informed by Dr. A. E. Ortmann, who has kindly taken the trouble to make some inquiries regarding this story, that he can find no foundation for it whatever. ‘Those who had heard of it at all did not take it seriously. Moreover, as Dr. Ortmann points out, it seems quite unlikely that cats could be forced to live in a cold-storage warehouse unless caged. It has, nevertheless, seemed worth while to cite this account, owing to the prominence given to it by Lydekker. Effects of External Conditions 151 Passing to the question of the adaptiveness of these experiment- ally produced modifications, that of the hair would surely seem to fall within this category. A complementary physiological explanation for the change would doubtless be likewise possible, had we a sufficient knowledge of the various processes concerned. The shrinkage or “drawing in” of the peripheral parts under the influence of a cold climate might also be regarded as adaptive, for the reduction of these thinly clothed surfaces would diminish, at least theoretically, the radiation of heat from the body. Here again a simple physiological explanation is likewise possible. We might either appeal to the effect upon the peripheral circulation (as does Allen) or to the direct influence of temperature upon the protoplasm of the growing parts. In the case of the feet of the mice in the above experiments, the greater activity of the young animals in the warm room, and the greater consequent exercise of the limbs, may possibly have played some part in bringing about the difference. To what degree the modifications which I have described have been due to temperature and to what degree they have been due to humidity is not clear under the conditions of the experiments. As has been stated, the two have varied inversely. Allen and Coues seem to regard such differences, when presented by mam- mals in nature, as due chiefly to the temperature factor. Never- theless, the former writer tells us, speaking of hares, that “there is also a marked tendency to an enlargement of the ears in propor- tion to the aridity of thehabitat.. . .. . . Inthis connection, also, attention may be called to the fact that all of the long-eared species of American hares are found exclusively over the niost arid portions ofthe continent. ©... .””.< And it may be added that the color of the pelage of mammals and that of the plumage of birds is well known to vary with the hygrometric conditions. In many species of birds the degree of pigmentation is said to be a function of the mean humidity of the habitat. “Tower, indeed, regards the humidity as being much more important than the temperature in the production of color changes in beetles. Until, 41 Monographs of N. A. Rodentia, p. 272. 152 Francis B. Sumner therefore, it is possible to separate these two factors 1n our experi- ments, we cannot state with any certainty to what degree each has been operative. A priori, it would seem, perhaps, that the changes in the mice have been such as could more reasonably be attributed to temperature. The fact that the same sort of differences as those which some- times obtain in nature between northern and southern species or varieties of animals have been produced by artificial conditions acting within the individual lifetime will be taken by some as evi- dence that these differences in nature are likewise entirely “onto- genetic’? or acquired independently by each individual. Con- versely, the neo-Lamarckian will perhaps argue—and with equal right—that here we have evidence that natural varieties and species have resulted from the accumulated effects of external conditions since the reality of such effects has been palpably demonstrated by the present experiments. Neither conclusion is justified by the facts before us. It remains to be settled experimentally (and thus only!) whether or not such modifications are transmissible. It has already been stated that no constant difference in size between the warm room and cold room individuals has been found to obtain throughout my series. Here, then, the reputed effects of natural climatic conditions have not been paralleled. It is quite possible that the cold was so severe during the early growth stages that some individuals were stunted. Indeed it has been pointed out for the 1907-1908 series that there was considerable mortality amongt the cold room lot in early life. Reference to the frequency polygons in Fig. 2 shows that there are two distinct modes among the cold room individuals; and I have determined that this is equally true of each of the sexes taken separately. The impression conveyed is that there are two pretty distinct groups, one of which was stunted by exposure to the cold, the other being favorably affected, so as even to surpass the warm room lot in size. It must be added, however, that no such effect is manifest in the 1908-1909 series. One of the most important general conclusions which seem warranted from an analysis of the foregoing results is the prin- ciple of the levelling down of experimentally produced inequali- Effects of External Conditions 153 ties, even while the conditions which gave rise to them remain in full force. A diminution in initial differences of size has been demonstrated by Minot* in the case of growing guinea-pigs. His findings upon this point are thus summed up: “The study of the individual variations yields two important conclusions: Furst, that any irregularity in the growth of an individual tends to be followed by an opposite compensating irregularity. Second, the variability diminishes with the age.’’ Thus, “if an individual grows for a period excessively fast, there immediacely follows a period of slower growth, and vice versa, those that remain behind for a time, if they remain in good health, make up the loss (at least in great partif not alwayscompletely)soonafter. . . . . It is probable that the same is true for man and that therefore the usual and even the severer illnesses of childhood and youth do not greatly affect the ultimate size of the adult.’’ Pearson,* likewise, has shown that the variability both of weight and of stature in man diminishes from infancy to adult life. And indeed it is a matter of common experience that an early handicap in the size or strength of a child is frequently “outgrown,’’ wholly or in part. The variability which the above-named writers have considered is doubtless in part due to blastogenic differences, 1n part to somato- genic ones, resulting from fetal or post-natal conditions of nutrition, etc. In my own results, however, the most noteworthy fact is not a reduction in the general variability of my stock, but the diminu- tion of differences whose cause is knownto be external—and this while the effective conditions remain unchanged. ‘The foregoing statement applies to the growth of the tail, both of the male and the female mice, between the age of six weeks and the age of 24 (or 3) months. It likewise holds, with some qualification, for che growth of the tail during the next interval between the measurements, 1. e., between 24 and 7 months. In the latter case, however, the data are fewer, and the allowance is necessary that about midway dur- ing this third period the temperature conditions were equalized 48 Senescence and Rejuvenation. Journal of Physiology, vol. xii, no, 2. 1891, PP- 97-153, pl. ii-iv. “8 Proceedings of the Royal Society, Ixvi, 1900, p. 23 (cited by Vernon, in ‘‘Variation in Animals and Plants”). 154 Francis B. Sumner for the entire lot of animals. The diminution of differences in weight between the contrasted groups of animals is less certain, though it appears tolerably clear in the case of the 1907-1908 females. It will be recalled, however, that the differences in weight were only very doubtfully regarded as results of the tem- perature conditions. This tendency toward a reduction of experimentally produced differences in the relative size of parts, should it prove to be gen- eral, is of considerable theoretic interest. It adds another to the many well-known examples of a “regulative” tendency in living things.*° After the initial shock of change, with its resulting effect in deflecting the organism away from its individual norm, there ‘would seem to be a continuous effort to regain the latter. Here we have a principle which might be said to bear the same relation to individual growth as Galton’s “law of filial regression” bears to stem-history, though the analogy may be merely superficial. In either case, however, we have to do witha “reversion to mediocrity.” The process in question is directly opposed to that conceived of by Weismann, in his theory of “germinal selection,” as occur- ring among the determinants of the germ-plasm. According to this hypothesis, a given determinant, if once handicapped by un- favorable nutrition, is more and more pushed to the wall by its more fortunate competitors until it may be totally annihilated. The disappearance of useless structures in phylogeny and the fre- quent orthogenetic trend of evolution is thus explained. If it be objected that this analogy of mine is out of place I can only reply that Weismann’s whole conception of a struggle among the determinants of the germ-plasm was derived from what was assumed to occur among the parts of the organism as a whole. Some evidence has been offered above for the existence of a tend- ency in the growing body quite at variance with the demands of that theory. ‘To many readers, on the other hand, it will doubt- % Vernon’s principle that ‘‘the permanent effect of environment on the growth of a developing organ- ism diminishes rapidly and regularly from the time of impregnation onwards” (op. cit., p. 199) would account for the failure of these differences to augment with the growth of the organism. But it cer- tainly would not in itself account for the absolutely greater increase shown by the more retarded organs (or organisms) mentioned on p. 137 above. . Effect of External Conditions 155 less seem quite frivolous to attempt any serious refutation of the “germinal selection”’ hypothesis. In justification, [ will but call attention to the fact that this theory is not only treated respectfully butis ably defended in the most recent general treatise on heredity.™ 51 T refer to Thompson’s admirable work, Heredity (G. P. Putnam’s Sons, 1908), which should be in the hands of every student of this province. FURTHER OBSERVATIONS ON THE BEHAVIOR OF TUBICOLOUS ANNELIDS! BY CHARLES W. HARGITT In a former contribution on the general subject above stated the present writer (’06)? described certain observations and experi- ments on several species of these interesting annelids. During the past summer these have been variously repeated and extended, and it is the purpose of the present paper to supplement the account given in the former by such additional facts as have been brought to light, and thus afford a more complete account of the behavior of these annelids than was made by the earlier contri- bution. The present account has to do with but one species, namely, Hydroides dianthus, except as there may be incidental occasion to refer briefly to others. The earlier observations were made entirely upon specimens kept in the aquarium. In the following account considerable emphasis will be placed upon observations made on specimens in their native habitat, and on the conclusions drawn from modes of behavior exhibited by tubes taken at various localities, and under greatly differing environments. It seems worth while to empha- size this point, as it is altogether probable that certain of our experi- mental results obtained from animals in cages, finger-bowls, etc., have given rise to more or less inadequate, if not misleading and erroneous conclusions. For example, it will be found in phases of the accounts which follow that the behavior of specimens taken from depths of ten to twenty fathoms shows considerable differ- ences as compared with that shown by specimens from shore waters, or from depths of two or three fathoms. Similarly the aspects of growth in specimens taken from rocky bottoms will 1 Contributions from the Zodlogical Laboratory, Syracuse University. 2 Journ. Exp. Zodl., vol. iii, p. 295. Tue JourNAL oF ExPERIMENTAL ZOOLOGY, VOL. Vii, No. 2. 158 Charles W. Hargitt show important contrasts when compared with those taken from muddy bays, or mouths of rivers. Again, in the former paper attention was directed chiefly to the reaction of specimens to light, and that too without particular efforts to test the variation or modifability of behavior under slightly differing conditions. In the present account will be shown a much wider range of tests, and some particular results as to individual differences of behavior not given before, as well as evidences of modifiability to which only incidental reference was made. Most of the experiments and observations which follow were prompted by some results obtained from a series of specimens taken from a depth of some twenty fathoms off Gay Head in a dredging expedition made by the Fisheries steamer, Fish Hawk, early in August, 1906. ‘These colonies were brought to the lab- oratory along with other material obtained. Upon arrival the specirmens were placed as usual in shallow aquaria containing freshly obtained sea-water and left for some time in order that they might expand and thus be ready for the usual tests. One of the first points of interest noted was the unusual slowness of the worms to emerge from the tubes. And when finally there were some signs of activity they yet seemed extremely wary, protruding only the merest tips of the gills, and frequently retracting them as if in fear. This [ took for a time to be due chiefly to the probable effects of the rough handling incident to the operation of the dredge and their subsequent transfer from vessel to vessel, etc. After some hours they at last became fairly expanded in the usual manner, though not fully. And when finally the tests with shadows were made they showed, to my great surprise, wholly negative results. “hey were left again undisturbed for some time and then further tested in the same way only to again fail to show any sensory irritability. They were then [left tll night, when they were tested with the electric light as in the experi- , ments of the former year* and here again the results were almost entirely negative. ‘This seemed so wholly peculiar, and so utterly contradictory of the almost uniformly positive reactions of former 3 Cf. op. cit., p. 300. Behavior of Tubicolous Annelids 159 experiments that I was greatly puzzled. ‘The specimens were left until the following day when the tests were repeated, and with similar results though with considerable variations among individ- uals. It occurred to me that these strangely negative results might be in some manner related co the unusual depths from which they had been taken. Accordingly I secured a single specimen from about the docks, and hence from an unusually shallow habitat, and placed it in the same aquarium with the others. ‘This speci- men behaved quite as had those of the former year, protruding its crown of gills promptly, and showed the same prompt and con- stant response to shadow stimuli as had been observed formerly. This suggested at once a series of comparative studies of speci- mens taken from various depths. ‘The results of these experi- ments are shown in part in the accompanying tables to which detailed reference will be made slightly later. Variously tested both by artificial and natural light as before, the reactions of the shallow water specimen were incomparably more acute and con- stant than any of the specimens from the deeper waters. In order to secure a more critical comparison of the relative sensory qualities of these specimens they were arranged in an aquarium on my laboratory table so placed in relation to both natural and artificial illumination as to enable one to easily con- trol either, as to relative intensity, at will. “othe ordinary shadow tests produced by the interposition of a screen as in the previous experiments, the behavior showed the same diversities as have already been noticed. In order to have the tests as nearly identical as possible in character, and at the same time within easy control, it was decided to use the stimulus of the suddenly extinguished electric lamp, as referred to above. By this means it was quite easy to have all the specimens in similar conditions of expansion when a test was made, and by the ease with which the light might be extinguished, and the equal ease by which it could be turned on again almost instantly, one may clearly determine at a glance what specimens had responded and any which had failed to so react. Usually a given interval was allowed to intervene between any 160 Charles W. Hargitt two tests. This interval was, on the average, about five minutes, sometimes less, or slightly more, as conditions might require. Of the signs employed in the tables, namely, plus, minus, and zero, the first two indicate simply positive and negative reactions of the specimens, i. e€., retraction into the tube or failure to retract. That of zero signifies that in a given test the specimen so designated was out of commission or, in other words, had withdrawn into its tube at the time the test was made. ‘This was not frequent, but as will be seen it happened occasionally with different specimens. The letters A, B, C, etc., arbitrarily indicate the record of given specimens throughout the series of experiments of a given table. The numbers at the left of the tables indicate the number of the tests of a given day or time. As a check against possible error which might be involved in a too rigid attention to results obtained from laboratory experi- ments some pains were taken to observe the behavior of specimens in their natural habitat. This was possible by selecting speci- mens to be found along the lower tide lines and in tide pools, and at favorable times when the surface was smooth on still days, thus making observation ‘practicable, tests were applied similar to those employed in the laboratory. Without taking time for ex- tended details it may suffice to state that the reactions obtained from shadow and tactile stimuli were essentially the same as the former, and confirmed them in almost every particular. Some differences as to individual reactions were perceptible, especially as to tactile responses. ‘This | am disposed to attribute to the peculiarity of the habitat. It seems not unlikely that the effects of the buffeting of wave action to which these specimens are more or less subject may account for these apparent differences 1n tac- tile reaction. Another feature was apparently quite different in the shore speci- mens, namely, the close adherence of the tubes to the substratum on which they grew. There was no tendency to growupright and free, as was the case in those from deep waters. And this again may be interpreted as a further expression of the conditions involved in the environment. It would be quite impossible for the tubes to withstand any such buffetings as the waves and pound- Behavior of Tubscolous Annelids 161 ing surf are constantly imposing upon objects under these con- ditions. ‘The tubes show throughout the closest adherence to the rocks, which in this environment forms the almost exclusive base of support, though one sometimes finds specimens attached to the shells of living snails occupying these pools. When in the process of growth the worm comes to a sharp angle of the rock it either turns and grows back upon itself, or turns sharply over the edge and along the opposite surface. In these conditions the tubes show a distinctly flattened surface along the line of contact which is hardly perceptible on those from deeper waters. All these facts lend still further confirmation of the suggestion of adaptation as the more probable explanations of the facts, rather than the more usual explanation of stereotropism, though the latter may not be wholly lacking. As already intimated in the beginning, pains were taken to secure specimens from as wide a range of depth and habitat as possible in order to detect any peculiarity of behavior which might be due to purely local, or environmental conditions. From result- ing comparisons marked differences were distinguishable in speci- mens dredged from a muddy bottom and those from rocky or sandy bottoms. ‘The former tended to grow in larger colonies, and in a more or less vertical aspect, while the latter were often isolated and independent, and closely adherent throughout to the substratum, as much asthose to be found intide pools or shore lines, referred to above. ‘The extreme contrast was found in colonies taken from the debris about river mouths, or enclosed bays which received the drainage or sewage of adjacent towns. In these con- ditions colonies of immense size were found, made up of hundreds, or even thousands, of specimens. In some cases these aggregations were apparently of considerable age and formed massive calca- reous accumulations, not greatly unlike coral heads in sub-trop- ical harbors. Ocgasionally colonies would be found on the most unexpected substratum. For example, the writer was shown one such taken by Mr. Geo. M. Gray at the mouth of the New Bedford river growing on an old granite kettle, covering it almost com- pletely, both inside and out, and forming a most picturesque object, especially when the hundreds of specimens of various size 162 Charles W. FHlargitt and color were fully expanded in the aquarium. Some further reference to this colony will be made in another connection. A point of more than passing moment is the fact that in almost every colony of these annelids, whether large or small, were to be found evident individual differences as to sensory response. In most cases there was more or less sharpness and uniformity as to reaction, while in others there was just as evident a degree of uncertainty and inferiority of response. And the same variation was evident in the matter of recovery time following a stimulus. In some individuals recovery was prompt and constant, while in others it was just the opposite, sluggish and uncertain. Again, many specimens react variously on different occasions. On some days there were noticeable and measurable differences of reaction to a given stimulus. Many of these facts are graphically portrayed in detail in the several tables, in connection with which they are briefly discussed. Attention may be called particularly to the striking individuality exhibited by specimens and colonies taken from different habitats and localities. And the wider one’s range of observation the more evident and significant does this feature become. PHOTIC EXPERIMENTS AND FATIGUE The following individual records of the reactions of three speci- mens tested under exactly the same conditions may serve to fur- ther illustrate the point here under consideration and certain. other features of similar character. ‘The specimens were taken just below tide line on the adjacent shore of Buzzards Bay on July 20, and for convenience are designated as A, B and C. The specimen A was tested by the shadow stimulus at inter- vals of one minute, and for fifteen times in succession. In every case there was a prompt and positive response, the creature with- drawing into its tube. Protrusion was somewhat variable as to the time concerned, varying from ten to thirty-five seconds, but the results were unmistakably decisive in every instance. In the case of B the tests were applied at intervals of 45 seconds, and, as before, fifteen times in succession. In this experiment there happened to be five specimens in a colony, and taken as a Behavior of Tubicolous Annelids 163 whole there was possible a total of 75 reactions. Of this number 71 positive responses were recorded, with four negative ones, or 94.5 per cent as compared with the Ioo per cent of A. Specimen C was tested as in the former cases, but at intervals of 30 seconds, and carried on beyond the fifteen tests of the former, with a view to determine at what point any signs of fatigue might be detected, or whether it was possible to distinguish any periodic recurrence of negative reactions. For twenty-five tests the speci- men responded promptly and positively. Following this reaction it remained in the tube for a period of exactly five minutes. At the end of this period it protruded and was promptly tested as_ before. At the eighth test the response was only partial. At the seventeenth test it failed to respond, but continued after this to react positively up to and including the twenty-ninth. The thir- tieth was negative, as was also the thirty-second. The thirty- third test was followed by a positive response in which the speci- men again remained in the tube for a period of five minutes. On emergence the tests were resumed and with the following records: The second and sixth responses were only partial withdrawals, the tenth and thirteenth tests failed toinduce response; fourteenth, partial. ‘The fifteenth, sixteenth, eighteenth, twenty-first, twenty- third, twenty-fourth and twenty-fifth were all negative. In this series of tests it seems strongly probable that we have something quite analogous to fatigue, especially in the final twenty- five tests. At any rate some condition had become operative in modifying in a marked degree the behavior of C as compared with A and B and the earlier ones of C. The fact that the tests were applied at very brief intervals, and, moreover, that they were long continued, go to strengthen the suggestion of fatigue. It is necessary, however, in this connection to point out that fatigue as here employed does not necessarily imply exhaustion of the bodily musculature of the worm, though this may be in some measure involved. But fatigue may be sensory, a condition anal- ogous to similar conditions induced by long-continued direction of the vision to a given test. ‘This point has been referred to in the previous paper* and has also been noted by several earlier 4 Op. cit., p. 302. 164 Charles W. Hargitt observers, among them Patten, whose comment was cited in the previous paper (p. 312). In an interesting article on the behavior of these annelids Ada W. Yerkes® has expressed doubt as to the question of fatigue, but seems to have misinterpreted the significance of the term as em- ployed in my paper, and has, [ think, misapprehended my con- clusions. For example, I did not attribute the failure to react under rhythmic stimuli to fatigue, as Mrs. Yerkes has suggested. What I did say was, “In connection with this matter of rhythmic shadows it was observed that where experiments were repeated with any considerable frequency specimens sooner or later be- came somewhat irresponsive, often failing entirely to react to any of the usual tests. Yhzs I am inclined to regard as the result of fatigue’ (p. 301). Concerning the failure of response under rhychmic stimuli my words were: “ May it not be possible that in these rhythmic shadows we have a simulation of the more or less rhythmic shadows resulting from the ripples of wave action.” As will be seen, therefore, my suggestion of fatigue had no reference to the matter of rhythm, but to the long-continued experimenta- tion which was designed to induce reactions repeated up to the point of fatigue. Furthermore, it was explicitly stated that under the normal conditions of rhythmic wave action such shadows might be assumed to have ceased to act as definite stimuli, the creatures having become accustomed to their effects by reason of their constancy. It seems to the writer that the above account of later experiments, with others of similar nature, goes far to confirm the earlier view. As one carefully observes such behav- ior it is difficult to escape the conviction that something akin to fatigue of some sort is involved. It may be that certain sensory cells only are affected, or it may be that certain central ganglia are involved, whose cells have to do with sensory and muscular coordination. Some recent work on the problem of fatigue in nerve cells by Drs. Smallwood and Rogers® has thrown new light on the subject and furnishes convincing evidence as to:changes in the cells them- 5 Jour. Comp. Neurol. and Psychol., vol. xvi, p. 442. ® Jour. Comp. Neurol. and Psychol., vol. xviii, p. 67, 1908. Behavior of Tubicolous Annelids 165 selves associated with experiments on fatigue. Some of their results are similar to phases of those here under review, and seem to confirm the suggestion of fatigue. TACTILE EXPERIMENTS In the former paper reference was made in several places to tactile reactions, but no details of experiments were given. ‘These I have subsequently repeated and extended and with such varia- tions as to suggest certain inferences and conclusions not previously considered. As might naturally be expected in creatures whose sensory powers are as delicately adjusted as the reactions to light already described, the tactile sense is also very acute. And interestingly enough it is found most highly developed in the gill filaments, the organs concerned with photic sensibility, one and the same organ therefore serving the double sensory function of touch and photic perception. But this is not peculiar to these organisms. Not a few of the lower invertebrates show very similar conditions, and suggest the inference that tactile and visual senses are more or less intimately correlated. Indeed, but for the conventional definitions of these senses, based to a large extent on the highly differentiated organs of higher organisms, it might be fairly allow- able to regard them as modified expressions of sensory processes due to stimuli of fundamentally similar nature. No time can be taken in this connection to follow out the suggestion further, but it seems well to call attention to the facts, barely hinting at the more or less obvious inferences concerned. While experimenting with shadow tests it was observed that now and then a specimen was found whose reactions were markedly inferior, or lacking entirely at times. Something of this will be noticed in the tables. In order to determine that in such cases the behavior was not due to some abnormal or pathologic condi- tion the specimens were subjected to tactile tests, and in almost every case were found to respond as promptly to such treatment as any of the others. And in not a few such cases it was found that a few tactile tests sufficed to awaken, as it were, the dormant pho- 166 Charles W. Hargitt tic sense, the specimen responding subsequently almost as readily and sharply asothers. ‘This was particularly the case where some pains were taken to apply the two stimuli successively for a few times. Something akin to this was found in the earlier experi- ments where attempts were made to excise the gills (p. 304). Individuality In this class of experiments there was evident the same marked individuality of behavior as has already been shown in previous sections. For example, a given specimen, X, was found to be unusually indifferent to slight tactile stimuli. At one time its gills were gently stroked 123 times, some of the latter ones quite vigorous, before it finally retracted within its tube. “The same specimen at a later hour the same day was tested again. At the first touch it responded with a sharp and decisive retraction. After pro- truding it was again touched with a glass rod, and at the second test again retracted. ‘The third time it retracted after the fifth test. The fourth time only after the twelfth. The fifth time it only responded after the 115th stroke of the rod. On the sixth test it responded after the thirty-seventh stroke; and in the seventh test only retracted after the gills had been stroked 237 times, some of which were decidedly vigorous, and even then the reaction was not sharp or violent as in some cases. An interesting feature was observed during the progress of these experiments, namely, the exhibition of what seemed to be a defi- nite sense of tactile discrimination. While observing a group of specimens it was noticed that those occupying adjacent tubes often in the act of protrusion thrust their gills against one already ex- panded, but without in any way causing a retractile response on the part of either specimen. It was also observed that a specimen of Sabella living among the same colony occasionally thrust its body out of the tube for considerable distance and swayed it laterally and variously, jostling the gills of Hydroides in the pro- cess, but with the same negative results as have been referred to in the previous case. ‘To make the test still more explicit a specimen of Sabella was removed from its tube and, being held with forceps was used as a tactile brush, the gills of other species were brushed Behavior of Tubicolous Annelids 167 more or less vigorously and somewhat indiscriminately, and again with substantially the same results as just cited. [ then tried a somewhat different test, namely, forming a bruslt out of a tuft of very delicate red alga; this was used as a tactile instru- ment. While it was possible by extreme delicacy of touch to gently jostle the gills of specimens, there was an appreciable dif- ference as compared with that associated with the foregoing. Tt was also found that by using a gentle current of sea water ejected from a pipette against the gills the specimens would frequently bear considerable disturbance without reactions. An attempt was made by this means to train specimens to allow similar jostling by means of a delicate glass rod, ejecting water from a pipette in one hand and with the other touch the gills with rod. But it was of small avail. In almost every instanec the slightest touch of the foreign body induced the usual sharp response of retraction. It would seem, therefore, that we have in this type of behavior an example of tactile discrimination of a qualitative nature, enab- ling them to distinguish as to the nature of the stimulus. It should be noted, however, that in this matter as in that of the shadow stimulus, there was marked individual variation of behavior, and furthermore that these differences varied more or less from day to day. It was pointed out in the former paper that following certain experiments a degree of shyness or caution was induced. The same thing was quite as marked in the tactile experiments here under review. And in the present case the tests were made with smooth and delicate rods which involved not the slightest injury, such as was involved by the clipping of gills with scissors; still there was almost invariably involved the development of a degree of caution which was quite marked. Following a given test the specimen would protrude only the tips of the gills, followed by a pause; then a further protrusion, then another pause, and often a slight retraction; finally a further protrusion and the expansion of the crown of gills as usual. 168 Charles W. Hargitt REACTIONS OF NAKED WORMS Both tactile and photic tests were made upon specimens which had been removed from the tubes, with a view to determine some- thing of the character of the responses under such radically changed conditions, and also to ascertain the relative sensitiveness of va- rious regions of the body. ‘The earlier experiments had sufficed to determine that the photic sense was limited to the gill regions. It remained, however, to determine whether there was any similar limits as to tactile sensibility. ‘This was attempted by gently touching the body with the blunt end of a delicate rod.. As might have been anticipated, it was found that the gills were by far the most highly sensitive. Next to this the immediate region of the head was more or less responsive. The mantle-like portion was found to be but slightly sensitive to touch, and the posterior, free margin of this organ was slightly if at all sensitive to any ordinary tactile stimulation. ‘This was likewise the case with almost the entire posterior portion of the body, only the prick of a needle or the pinch of a forceps inducing some slight muscular reaction. Touching the gills always aroused a prompt and vigorous reac- tion. If stimuli were applied rapidly, at five to ten seconds inter- vals for a few minutes the specimens evidently became fatigued, and soon failed to respond, or did so but feebly, the body having gone into a condition of muscular contraction resembling tetanus. This was followed by apparent efforts on the part of the specimen to relax, the gills expanding in spite of continued tactile agitation. This condition may have been due to something like a convulsion resulting from overstimulation, and perhaps induced in some measure by the unusual condition of nakedness. After a rest of from five to ten minutes the specimen would again respond in exactly the same way, but became sooner unresponsive, lapsing into the same condition of fatigue as before. Reactions to the shadow stimulus were more or less peculiar and uncertain. This was doubtless due in part to the almost — constant writhing and twisting of the body under these conditions involving what I have previously termed “mixed stimuli” (pp. 303, 316). ‘The very unusual conditions of nakedness and help- Behavior of Tubicolous Annelids 169 lessness involved were such as to preclude anything like simple or direct reactions to any given stimulus. In other words, the conditions involved the operation of a complex of peculiar and anomalous stimuli, so that the reaction to the shadow stimulus was only one of several acting upon the creature at a given time, However, taking a specimen when less disturbed by promiscuous movements of the body it was found that a sudden shadow often caused a sudden response, or jerk of the body, and a correlated contraction of the gills. This might recur upon a second, or even a third test, but after that it became practically indifferent. Re- peated after an interval of a few minutes a similar reaction was liable, but failure was also frequent. Excision of the gills of such specimens was invariably followed by total absence of photic re- action, and fully confirm the earlier experiments and conclusions, Some connected account of the facts portrayed in the several tables in addition to references made at various places in the paper may make clearer their significance and importance. Incidentally it may be stated that these data are but a small fraction of a very large number of observations made during two successive summers. To attempt to include all would unnecessarily cumber the paper with details which, while not without some value, are not vital to the fundamental aims of the investigation. Tables I and II show records of two series of specimens: A B CDE and HI J KL. Both were from deep water, eight to fifteen fathoms, and show a fair average of reaction of specimens from this environment. ‘The records are specific, that is, made from definite individuals. The really significant feature of these records is their almost exclusively negative character. An inter- esting detail of Table I is the zero record of specimen C, quite marked in the first, or 11 o’clock test, and its almost exclusively zero record for the 2 o’clock test. In this respect the record is somewhat unusual. A specimen may often exhibit such a be- havior at one time and an altogether different aspect at a subse- quent test. In marked contrast with the foregoing is the record shown in Table III. Another series of five specimens, A B C D E, taken from shallow habitat illustrate what hundreds of similar tests of 170 Charles W. Hargitt specimens from similar surroundings have confirmed, namely, the unmistakable predominance of positive reactions to the shadow test. Tables IV and V show the same thing, and also show the rather remarkable variation of recovery time at different tests by the same, and different individuals. For example, in Table IV the recovery time of F and G at the very first test is in extreme contrast, and at the ninth test almost the same contrast, but in TABLE I Showing reactions of specimens from deep waters August Q, II a.m. August 9, 2 p.m. Temperature 22° C. Temperature 22.5° C. yaad B c D Bo has Bie Daa Te = = = = = = kaees = = Die = = = = = as = en | - aye = +10 = ena as = +18 = +30 | +20 fea! = o | - =e = ~ sae - Bosses ° + + - — _ {2 = (Sie = = ° ° = = = | =) a 3 +12 +10 ° fo) = = = = = Sie = = ° = - = _ = = 9: —- = ° _ — _ = == 10 — a fo) ~ — -- — | = == ) ° ° | | | | OlCn OM Om ONO) Om 0) KOLO O MORON ONO) 10 | | reversed order. The record of the same specimens at 2 p.m. shows much less contrast in these points, though they are not lack- ing. In some cases the contrast between certain individuals was much greater than these. In one case a specimen after a given test retracting and remaining in its tube for nearly half an hour, in exact time twenty-six minutes. It does not seem possible to explain these as in any wise due to anything in the nature of the stimulus. Behavior of Tubicolous Annelids , 71 Table VI contains a brief record of the comparative reaction of two specimens to shadow and tactile stimuli. ‘The longer recovery time following the tactile test is naturally what might have been expected. In other records similar conditions were found to those alluded to in citing the contrasts shown in Table IV, and probably explainable on same general assumption. But there is more involved in the records of these tables than TABLE II Showing reactions of specimens H, I, 7, K, L, from deep waters AugustI0,Ila.m. ~ August 10, 1:45 p.m. Temperature 21° C. Temperature 22° C. H I J K L H I J K L I = +15 = = Yi We a | = = +10 +12 2 — = = = = Kees = +10 +10 = 3 = = = = = = at; 15 aa = = 4 +25 +25 = +12 = = = +79 = = 5 +40 = ae a = = = =. = 6 = = 2 ae = = a = ote 2 7 = = os = * ath = Be pt a 8 +28 = +10 = = = - +10 = +10 9 = = ar its) = = = a ar = = we ao ee lleatiieal [alee EE age Nm Leelee ey da it Soha Le SI heel ae Lh al a ee epee el [tell eal aS | | | | | | | | | | An | | | | | | + oo | | _ _ oon | | | | | ie) | | | | N=) | - Bee eas tsb 2 Wen dee : ie 2 | rs) re) | the simple facts of the reactions. As expressions of behavior the facts call for explanation and interpretation. A point of more than passing significance is that involved in the recovery time indicated in the various tables. Does the feature of protrusion, following a given retraction induced by shadow or tactile stimulus, sustain any such relation thereto as to warrant the conclusion that these several aspects are essentially parts of a common reaction 172 Charles W. Hargitt cycle? If so should there not be discoverable some definite law of relational sequence ? On the assumption that organisms of this type are definitely organized mechanisms adjusted to the normal play of chemico- physical stimuli there can hardly be hesitation in answering afhrm- atively the several queries. Organic machines of such character, and designed to illustrate just such principles, are well known in . TABLE III Showing reactions of specimens A, B,C, D,E (from shallow waters), to shadow tests; intervals of recovery from reaction indicated in seconds. The minus sign indicates failure to react. Temperature of water, 22° C. A B Cc D E Deeper een Mie te oP erervoe cra ncrsie te cisiehetobe nice oases 6 7 7 8 6 De Pe rene crs oberee rebar Tey aye ol cr MoyesercucNepogh elansts wcaje xual hey sas 5 9 16 16 5 Fete ae USN AG DCS2 OSG GO Od ooo ncaa Rt 10 10 15 15 | = 7 ae Re nriione Cited Chg Cera RA MCD MORIA TOO IS 9 10 ie 15a 5 Goins cusenucncd Ewe oe toto En mocmOg yaaad aos 9 15 i AP APL 6 (Sub adis cia aula pog.ce co tunGda dk wecomunood So 8 8 10 10 = I ara oe ian Ie Son OCe Cn a Sepa enoc ato aK) | = = = Aare neemcch pretences eather eee Aaceey army A ee 0 ABC pre Io | 10 10 — 60 QUIS chelegsnert viene: ceapcast Mavens a ten raboqee hel oretere een teat 8 10 10 20 6 Ley a AVIS CR ne MPR OG EEN eed ten it 7 Peete 8 9 | ~ fo 10 | 5 1A ERS Pree sec Fitton ean ReRRN RO yO OO pom orttar ERAS 9 9 16 ite 20 TORS alata Sa GRE. acti ob earmtc ene tn ome Sera neTe c = = = =a = [ele vagal tt pe Sn Oh nee ate Gp eae cheat Gate 8 8 10 10 6 MA fires ears ates eaters atata: nie eaters: ohn. cra aks totene)p Meese RON ANS = = ~ 5 SE ete Ja niGd Geer eR IOe Dan Ane e Ionae SCO eas 3 righ | = = = — ESE Soho or acind nite Boma a Seer tare 8 cutg GIGS. iit 15 15 15 112) Tey ten acetate isin wea sli baat Sekeen ee ees Io | 15 20 20 15 HE ie pete in. eR CEO See oe Orme T ae. SARS mck io 12 12 = 12 IC arty clair a Bh A OP cen haat ce ood cee Se USO OG an 7 12 15 20 12 os ERAS Onn OO GSO SER em nT ee Hert oain Roe 8 | 12 i = | 6 every physiological laboratory. ‘The familiar ‘‘muscle-nerve”’ preparation affords a good example of such. Under proper arti- ficial adjustment such type of vital machine can be made to react with some approximation toward the requirements of the mechani- cal theory. Other similarly devised examples are not unfamiliar, in which chemical processes take the place of the physical in the former. Behavior of Tubicolous Annelids 173 TABLE IV Showing reactions of two specimens F and G, tested at 10 a.m. and 2 p.m., temperature 22° and 23° C. F | G iy G (10 a.m. | Bena ate Wet, ago ERE Tec T sa ttotetare Da sitter a Siete seete race eoakas e'e oy Pies elolel weavers 40 | 360 | 15 20 25 bap Soe aOR oe TRO Oe re ee ane ee Cree 30 | 60 10 30 ES MRE ts one e ere rea eS Ps Pa ns va otal ess tag y“al tne inher Save ve teal 30 180 | 12 45 Did CoO HOBISA SERCO Raden’ ODIO S GIL De Aa DION iin Hip chc 33. | 60 |} 10 30 Ba ncneonen Sonangnoon Adds Adice votorcc Agno seo booocempdde 60 | 50 | 18 45 GRRE err yo a rtese-tay Seiebare fe Slskcih sce Sica siclole G'p he ce-a/Sieig/b late 45. |. 5b | 15 50 loa boddBrcdiatys pela thes ASE TeSAeIG DIE Org Drie eee) 100 | go | 12 35 SNE is BUR Se TA ee iraagiee 8 Uh Lid ots ssid ale Ra Tk de band — | 60 | 13 80 Tacavecoesoshebogds based onscNOCeeorn DoUCeUDF age ADOr 300 arr ul = 100 MR Peeters ule 2S logic cases necks weds «eee 40 zo) | | as 40 SE et ee BN ho oes cae aint noite ees ohne 30 zoo || 12 | 110 Tip 03 o RA CAPO E ae ACOA OR OR ORO y ee gyn, NRMOn Ors Serna | 180 | 60 | 45 go Tis coc ct cee cies em ene tine me cette aid ae se eccsereeweees 43 | go | 75 15 Mle. codéppbepbtossbiocg esp acghoor nie) Capoten ooo Maaeverr | 180 | 105 | 20 go TiN PERE Nal CN cer asain wes ryan cS tyeie'| 7g) apie | zo go Wh, cn aneqahodasogus Umuendp cmodeions Oooo COU Dna caneaahs | 75 | 130 | 12 105 jo o8 tis 85 OE GE hdmo he CORCAS ORM SOS non Udo ogn OC rian aaee 470 150 || = 80 A) S BAe ei end OS EO Cetbeo oe yet PRACT eR or PI RNC nn CER 50 150 i5. | 85 TG): o10 a od wed 76.8 GO OOR SRE De UO a CER eS OAc a tonite aaerrnt | 60 /= 22555 II 45 DOr. \3.0'g0 EUS OHIOOG & a ORICA C TIE ROTOR acc USES CRE 150 go 18 = TABLE V Showing reactions of four specimens, F, G, I, F. In the first column is shown the reaction and recovery time of F and G at 3 p.m., temperature 25° C. Inthe second column is shown the reaction and re- covery time of same specimens at 4 p.m. and with temperature at 20°C. Inthethirdis shown the reaction and recovery times of the other two specimens under same conditions at 4:20 p.m., and at FAI Ore | | Sino mes F ci | J WA Pee eeteh tec ses ke', (Sis frevoieaiter ercastoe die dare akie's 18 13 | 60 60 | 18 20 Deine Pest ais saichs Bia haps ea ee seis aaa rae ee | 15 15 | 30 | 60 30 25 > GOS aii D GO RSAC FORE TOMO SDS SA One 14 EES |) G40 | 50 25 25 ese OS ea cr ORI OFLER Te CTE OCEANIC eae 18 Toe hi 45 | 45 || 20 20 2. ove BR nee ae SP ee ot ee | II ed |e 25.6 | 10 16 Riera re kate tees thay bis aie sede we 15 Guilin a5 o | 7 20 Mae ere a rn eee deel he 3ee ote os» 10 10 45 ° 10 25 oe, SEGRE Sad NG Ect ober See ee 10 10 go. "6 20 25 Osc. 8 Ag gE eS PNen OREO Ete re eee SAPS 15 20 | ag Pe oe 15 = TOS nic ccc ccsitic cect eset cesesceres tu TO EO p= <4 Sy = 15 = | I 174 Charles W. Hargitt Now, how do the facts embodied in these tables compare with those exhibited by the methods just cited? Do the reaction phe- nomena conform in any fundamental respect with those of a muscle-nerve preparation, or with those induced in cardiac mus- cle by such chemical reagents as Ringer’s solution? ‘The briefest comparison of a record of one of the latter experiments with the tables under review will quite suffice to show that this is not the case. To further test the problem an attempt was made by my colleague, Dr. Rogers, to reduce the records to a curve by appro- priately plotting them, and thus make the results more easily TABLE VI Showing comparison of reactions under shadow and tactile stimuli, as indicated by the reaction time given in seconds. Two specimens, A and B, were used, an individual record being made of each specimen, tested successively by each method. A B | acs o, | Shadow Touch | Shadow Touch MERE Reet oh rd Taree eet me | 12 | 50 16 50 De (rie. siine ngage ciel esis Perera s eotes